Economics Uncut : A Guide To Life, Death And Misadventure Simon W. Bowmaker Death, Edward Elgar

User Manual:

Open the PDF directly: View PDF .
Page Count: 485

Download
Open PDF In Browser	View PDF

ECONOMICS UNCUT

Economics Uncut
A Complete Guide to Life, Death and Misadventure

Edited by

Simon W. Bowmaker
Academic Associate in Economics, University of Edinburgh,
UK, Adjunct Lecturer in Economics, Florida State University
and Visiting Lecturer in Economics, State University of New
York at Buﬀalo and University of Colorado at Denver, USA

Edward Elgar
Cheltenham, UK • Northampton, MA, USA

© Simon Bowmaker, 2005
All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system or transmitted in any form or by any means, electronic,
mechanical or photocopying, recording, or otherwise without the prior
permission of the publisher.
Published by
Edward Elgar Publishing Limited
Glensanda House
Montpellier Parade
Cheltenham
Glos GL50 1UA
UK
Edward Elgar Publishing, Inc.
136 West Street
Suite 202
Northampton
Massachusetts 01060
USA

A catalogue record for this book
is available from the British Library

ISBN 1 84376 362 1 (cased)
ISBN 1 84542 580 4 (paperback)
Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall

Acclaim for Economics Uncut
‘Sex, drugs, rock and roll and so, so much more – there is enough here to
persuade even the deepest skeptic of the analytical power and breadth of
economics. The material is presented with careful attention to the evolution
of economic ideas, as well as state-of-the-art economic theory and empirical analysis. Many strands of economic thought are represented. Any class
would be enriched by examples drawn from this expansive collection.’
– Alan B. Krueger, Bendheim Professor of Economics and Public Aﬀairs,
Princeton University, USA
‘If you thought you could hide your secrets from the prying eyes of economists, think again. From sex to drugs to gambling to crime, this book will
show you how the tools of economics can be used to understand just about
any human behavior. This book will assuredly be the unoﬃcial economist’s
guide to vice for the foreseeable future.’
– Steven D. Levitt, Alvin H. Baum Professor of Economics,
University of Chicago, USA
‘This exciting book shows that economics can explain a dizzying array of
real world phenomena and that economics can be great fun. I recommend
this book to anyone who wants to see the progress that has been made in
using the tools of neoclassical economics to understand a wide range of
seemingly irrational behavior.’
– Edward L. Glaeser, Professor of Economics, Harvard University, USA
‘For students who care about how the world works, microeconomics should
be one of the most relevant and exciting subjects they study. Yet many students view microeconomics as some kind of abstract theory, and fail to understand how it relates to the real-world decisions made by ﬁrms, consumers, and
governments. It is therefore essential that students be shown detailed examples of the application of microeconomics to practical problems – studying
the theory alone is simply not enough. This outstanding book provides those
examples, and would make a perfect supplementary text for any intermediate
microeconomics course. The topics and examples covered are extremely
timely and interesting, and are explained with great clarity. Instructors take
note: students will ﬁnd this book fascinating, and reading it will give them a
much deeper understanding and appreciation of microeconomics.’
– Robert S. Pindyck, Tokyo-Mitsubishi Professor of Economics and
Finance, Massachusetts Institute of Technology, USA

‘Economics recently has developed an imperialistic policy. Economists now
write on many, many things which were not included in the traditional economic curriculum. This book is an example, but the reader may think that it
should have been included in the traditional curriculum. All the activities
analyzed have economic implications and the study is a good example of
applying economics in new areas. The reader may be a bit shocked by applying economics to such things as prostitution, but I think he will conclude that
economics has something to say and that practical people trying to organize
government controls in many areas would gain by reading this book.’
– Gordon Tullock, Professor of Law and Economics,
George Mason University, USA
‘Who would not be curious to know what economics can tell us about
drugs, crime, sex, religion and sports? So read this book – it demonstrates
convincingly how much standard economics provides us with great new
insights.’
– Bruno S. Frey, Professor of Economics, University of Zurich,
Switzerland
‘Economics is all around us, and this terriﬁc collection highlights the role
of economic forces even in domains where the reader may not expect it. As
a collection, the book’s chapters make the case forcefully that economic
incentives are embedded in all domains of life, and that even sex, drugs and
rock ’n’ roll are amenable to economic analysis. A collection that showcases
modern economics as the lively discipline it is. It would be hard to be a
student and not be excited by the possibilities.’
– Justin Wolfers, Assistant Professor of Economics, University of
Pennsylvania, USA
‘This fascinating book shows that economics can be used to understand
almost every aspect of human behavior: the consumption of addictive
goods, divorce, pornography, sports, religion, suicide, criminal behavior,
abortion, and rock ’n’ roll. The conventional wisdom is that these behaviors should be studied by sociologists and psychologists. Simon Bowmaker
and his collaborators demonstrate the fallacy of this proposition with great
insight and clarity. Their analysis is provocative, penetrating, profound,
and always entertaining. The reader is quite likely to conclude that there is
only one social science. All in all a most remarkable book. Bravo!’
– Michael Grossman, Distinguished Professor of Economics,
City University of New York Graduate Center, USA

‘Those who think economics has little to say about the real world will
change their minds after reading the essays in this splendid collection. Well
written and authoritative, they show how the economic way of thinking can
shed light on an unexpectedly diverse range of issues.’
– Robert H. Frank, Professor of Economics, Cornell University, USA
‘Who said that academic economists need to get out more? This book will
certainly shatter the illusion that economic analysis has little to say on
topics of everyday interest.’
– Evan Davis, Economics Editor, BBC, UK
‘This wonderful new volume shows that economic tradeoﬀs inﬂuence all of
the decisions that we make. The book both educates and provokes its
readers by using the rational actor model to explain seemingly pathological human behavior like suicide, drug addiction, and prostitution.’
– David Laibson, Professor of Economics, Harvard University, USA
‘This is an important book that showcases the strength of economics for
understanding human behavior and provides a convincing case for why
economics should be reckoned as the social science. The chapter on prostitution provides the most comprehensive discussion of the topic to date.’
– Lena Edlund, Associate Professor of Economics,
Columbia University, USA
‘Sex, drugs, suicide, murder, football – instead of a “dismal science”, this
book makes economics seem popular, or rather notorious. A rational analysis of these topics contains something to oﬀend everyone. You may not agree
with these papers but they will challenge you to think more deeply.’
– Robert Cooter, Herman Selvin Professor of Law, University of
California, Berkeley, USA
‘Economics is most useful – and most provocative – when the spotlight of
its intellectual rigour is trained on social issues. This book demonstrates
how economists think about sex, drugs, and rock ’n’ roll. This is an illuminating approach to subjects usually obscured by woolly thinking and conventional platitudes. It should be compulsory reading for anybody devising
or legislating public policy.’
– Diane Coyle, author of Sex, Drugs, and Economics and
Paradoxes of Prosperity

‘Economics is not about forecasting interest rates, but about how we live
our daily lives. Simon Bowmaker’s wide-ranging analysis illustrates how
fruitful and unexpected that approach can be.’
– John Kay, Fellow, St John’s College, Oxford and Visiting Professor of
Economics, London School of Economics, UK
‘Economics is generally associated with the ﬁnancial pages of newspapers
apart from front page discussion of major topics such as inﬂation, budget
deﬁcits, or unemployment. However, the topics discussed in many of the
other pages of a typical newspaper, such as crime, divorce, or sport, are also
appropriate for economic analysis. Economics is concerned with decisions
and many important topics in today’s society involve taking drugs or committing a crime or getting a divorce, for example, and so can be examined
from an economic point of view. Many of these areas can be considered
from diﬀerent directions: legal, medical, political, religious, sociological, or
psychological, and economics can be added to this list. No single viewpoint
can be considered the dominant one, in most cases, but together a deeper
understanding of the area will result. This book provides deep and stimulating discussions of a dozen or more important and interesting areas in
accounts prepared by economic experts. It illustrates how what has been
called the “dismal science” can be applied widely and usefully.’
– Sir Clive W.J. Granger, Professor Emeritus, University of California,
San Diego, USA and 2003 Nobel Laureate in Economics
‘In this insightful and entertaining book, Simon Bowmaker introduces
readers to the fascinating side of modern economics that applies economic
analysis to a wide range of social issues from illegal drugs to religion and
everything in between. In this form, economics is anything but the dismal
science. This is a fun and enlightening book that shows readers what many
economists often forget – that economics is a powerful tool for understanding the world around them.’
– Kevin M. Murphy, George J. Stigler Distinguished Service Professor of
Economics, University of Chicago, USA

Contents
List of contributors
Preface by Simon W. Bowmaker
Acknowledgements
Bend it like Becker by Simon W. Bowmaker
Introduction
David D. Friedman
PART I
1
2
3

xi
xiii
xvi
xvii
1

SINS AND NEEDLES

Economics of drug addiction
Simon W. Bowmaker and Frank Heiland
Economics of drug prohibition
Jeﬀrey A. Miron
Economics of drug liberalization
Mark Thornton, Bruce L. Benson and Simon W. Bowmaker

11
44
68

PART II GUNS AND ROSES
4
5

Economics of crime
Bruce L. Benson and Simon W. Bowmaker
Economics of marriage and divorce
Leora Friedberg and Steven N. Stern

PART III
6
7
8
9

101
137

BODY AND SOUL

Economics of pornography
Samuel Cameron
Economics of prostitution
Peter G. Moﬀatt
Economics of suicide
Samuel Cameron
Economics of religion
Robert J. Stonebraker

171
193
229
264

Contents

PART IV
10
11

Economics of assisted reproduction
Sherrie A. Kossoudji
Economics of abortion
Leo H. Kahane

PART V
12
13
14

CONCEPTION AND REJECTION

315

FUN AND GAMES

Economics of sport
John Goddard and Peter J. Sloane
Economics of gambling
Robert Simmons
Economics of rock ’n’ roll
Simon W. Bowmaker, Ronnie J. Phillips and Richard D. Johnson

Index

291

345
367
389

423

Contributors
Bruce L. Benson, DeVoe L. Moore Professor of Economics and
Distinguished Research Professor of Economics, Florida State University,
USA
Simon W. Bowmaker, Academic Associate in Economics, University of
Edinburgh, UK and Adjunct Lecturer in Economics, Florida State
University, USA
Samuel Cameron, Professor of Economics, University of Bradford, UK
Leora Friedberg, Assistant Professor of Economics, University of Virginia,
USA
David D. Friedman, Professor of Law, Santa Clara University, USA
John Goddard, Professor of Economics, University of Wales, Bangor,
UK
Frank Heiland, Assistant Professor of Economics, Florida State
University, USA
Richard D. Johnson, Associate Professor of Finance, Colorado State
University, USA
Leo H. Kahane, Associate Professor of Economics, California State
University, Hayward, USA
Sherrie A. Kossoudji, Associate Professor of Social Work and Adjunct
Associate Professor of Economics, University of Michigan, USA
Jeﬀrey A. Miron, Professor of Economics, Boston University, USA
Peter G. Moﬀatt, Reader in Econometrics, University of East Anglia,
Norwich, UK
Ronnie J. Phillips, Professor of Economics, Colorado State University,
USA
Robert Simmons, Senior Lecturer in Economics, Lancaster University,
UK
Peter J. Sloane, Professor of Economics, University of Swansea, UK

xii

Contributors

Steven N. Stern, Merrill Bankard Professor of Economics, University of
Virginia, USA
Robert J. Stonebraker, Associate Professor of Economics, Winthrop
University, USA
Mark Thornton, Senior Fellow, Ludwig von Mises Institute, USA

xii

Preface
Simon W. Bowmaker

This book applies microeconomics to social issues. From drugs to divorce,
pornography to prostitution, sport to suicide, and religion to rock ’n’ roll,
we show that economics can reach into the strangest of places and shed
light onto the occasionally dark side of human nature. Most importantly,
we establish that with every issue of society the economist has a right to
speak, and can do so usefully.
The idea to put together this book originated at the University of
Edinburgh, where I ﬁrst taught in 1999. I quickly discovered that students
have a surprisingly healthy appetite for a diet of sex, drugs and economics.
My experiences in the United States, including spells lecturing at Florida
State University and the State University of New York, Buﬀalo, convinced
me that they (and their instructors) would appreciate such a volume.
However, I am also aware from discussions with wider social contacts that
people are interested in the issues we cover, but are still in search of an analytical framework for clarifying their opinions and decision-making. The
chapters that follow will be of great value to these individuals.
To motivate the various topics, our starting point is a conversation with
Gary Becker, the economist who has pioneered the application of microeconomics to whole areas of life that conventional wisdom may deem either
inappropriate or inaccessible. I sought his views on a broad range of contemporary social issues, including drug legalization, same-sex marriage,
capital punishment and suicide bombing.
We then proceed as follows. Part I, ‘Sins and Needles’, is a three-chapter
examination of illicit drug use and policy. In Chapter 1, Frank Heiland and
I analyse theories of addictive behaviour, the price elasticity of demand for
illicit drugs, and the welfare economics of drug prohibition. Next, in
Chapter 2, Jeﬀrey Miron shows that the unusual features of drugs markets
result from their legal status, not from the mind-altering and addictive
properties of drugs themselves. Finally, in Chapter 3, Mark Thornton,
Bruce Benson and I examine the economics of drug liberalization, including the alternative policies of government monopoly, regulation, sin taxes
and the free market.
Part II, ‘Guns and Roses’, looks at the contribution economists
can make to understanding crime and marriage and divorce. In Chapter 4,
Bruce Benson and I investigate the economic theory of the supply of
xiii

xiv

Preface

criminal oﬀences, the demand for crime prevention and law enforcement
services, the allocation of resources within the criminal justice system and
the possible reasons for the decline in crime rates in the USA in the 1990s.
In Chapter 5, Leora Friedberg and Steven Stern then examine the gains
from marriage versus living together and develop an economic model to
help explain why people marry, the nature of decision-making within marriage and the nature of the decisions to marry and to divorce.
Part III, ‘Body and Soul’, begins with two chapters on the economics of
sex. In Chapter 6, Samuel Cameron looks at research outcomes on the
eﬀects of pornography, the characteristics of the industry, Amartya Sen’s
argument about Paretian liberals as applied to porn, the nature of demand
for porn, government approaches to its regulation, the impact of the
Internet and the positive externalities arising from the consumption and
production of porn. Next, in Chapter 7, Peter Moﬀatt considers the economics of prostitution. He examines theoretical models that seek to
explain, among other things, how equilibrium earnings are so high in a profession with only rudimentary skill and capital requirements, analyses the
empirical evidence relating to prostitution and discusses the issue from a
policy perspective, including a description and evaluation of various
models of legalization. In Chapter 8, Samuel Cameron turns to the economics of suicide. He discusses Emile Durkheim’s socio-economic contribution to the subject, examines economic models of suicide and the
empirical work undertaken by economists, investigates the potential economic cost of suicides, including those of celebrities, and shows how
an anti-suicide bureau might operate. Finally, in Chapter 9, Robert
Stonebraker examines the economics of religion. He considers the demand
for religion, supply-side issues such as how providers of religious goods and
services cope with free-riders, the problems of risk and uncertainty in religious choice and how competition and government regulation shape the
structure and content of religious markets.
Part IV, ‘Conception and Rejection’, focuses upon the controversial economics of assisted reproduction and abortion. In Chapter 10, Sherrie
Kossoudji provides an introduction to the markets for assisted reproduction, discusses the economics of the demand and supply of children,
explores the development and size of markets in assisted reproduction,
examines the economic issues in the markets for eggs, sperm and surrogacy,
and undertakes a discussion of the ethics of assisted reproduction markets.
Next, in Chapter 11, Leo Kahane looks at the demand and supply of abortion services and demonstrates how these fundamental tools of economic
analysis can be used to understand the workings of the market. He reveals
how the availability of abortion services and their legality has had
ramiﬁcations for other related issues such as teen pregnancy rates, the

Preface

propensity for pre-marital relations and ﬂuctuations in crime rates across
time.
Finally, Part V, ‘Fun and Games’, examines the economics of sport,
gambling and music. In Chapter 12, John Goddard and Peter Sloane consider the objectives followed by team and league owners in sport, investigate the nature of product demand, the sports league’s role as a cartel and
issues relating to the eﬃcacy and desirability of revenue-sharing, salary
caps and collective selling of broadcasting rights. In Chapter 13, Robert
Simmons examines the fundamental economic question of why people
gamble and considers the economic reasons why US states and governments across the world may opt to legalize forms of gambling. He also
looks at some contemporary policy issues involved in regulation and deregulation of gambling, including the social beneﬁts and costs, the use of state
government revenues from gambling for particular purposes and the threat
to government tax revenues and government regulation posed by the
growth of Internet gambling. In Chapter 14, Ronnie Phillips, Richard
Johnson and I provide a brief history of the music industry, examine the
complex relationships between the major record companies and the independents, and analyse the economics of recording contracts. We also show
how technological change is aﬀecting the relationship between record label
and artist, investigate the presence of the ‘superstar phenomenon’ in music
and discuss the economics of copying music.
So, various dimensions of social behaviour and interaction are discussed
in every chapter, but this has only been possible because extensive use is
made of economic analysis. Indeed, long-standing economics concepts are
shown to have continued potency throughout. Marginal utility and price
elasticity of demand are crucial to the analysis of illicit drug addiction.
Traditional demand and supply analysis still has a lot to offer to the criminologist and to the student of abortion. Old and new economic theories
jostle to provide a framework for analysis. The sturdy theory of the firm is
useful in examining sport, prostitution and rock ’n’ roll, while game theory
helps the analysis of suicide and marriage and divorce.
Of course, inevitably linked to the presentation of a social problem is
consideration of a policy response. Some of the issues discussed come into
conflict with the criminal law, for example, sale of particular drugs, so the
economics of enforcement and prohibition are given weighty consideration. Long-established policies such as competition policy have relevance
in fields such as sport. Taxation policy is used to examine the economics of
drug liberalization, and a variety of forms of regulation are dissected, even
in unusual matters such as religion. First and foremost, this is a book about
economics.

Acknowledgements
Many people have provided assistance during the preparation of this book.
First, of course, I would like to thank the authors of the chapters for all
their hard work and commitment to the task. My colleagues in Economics
at the University of Edinburgh have also been tremendously helpful, particularly Stuart Sayer who supplied the initial encouragement to undertake
the project. I am very grateful for all his advice. In addition, Simon Clark,
Richard Holt, Colin Roberts and Donald Rutherford have been a source of
some excellent ideas and invaluable input.
At Florida State University, I would like to thank Jim Cobbe from the
Department of Economics for the opportunity to teach economics in the
United States. During my ﬁrst visit to Florida State, Bruce Benson, Gary
Fournier, Carlos Garriga, Frank Heiland and Tim Salmon were very kind
colleagues and a lot of fun to work with. They also spent a signiﬁcant
amount of time helping me with this book.
My gratitude to the following who provided constructive comments on
various chapters: Farasat Bokhari, Andrew Burke, Samuel Cameron,
Richard Caves, Dhaval Dave, Isaac Ehrlich, Rodney Fort, Donald
George, Barry Hirsch, Ed Hopkins, Gary Koop, Ziggy MacDonald, Dave
Marcotte, Patrick Mason, Jeﬀrey Miron, Robbie Mochrie, Stephen Platt,
David Rasmussen, Jószef Sákovics, Tim Sass, John Sawkins, Robert Wright
and Artie Zillante. David Friedman was extremely generous in providing
extensive comments on the entire manuscript. I recommend that readers
visit his highly entertaining website at www.daviddfriedman.com. Of course,
I would also like to thank Gary Becker for kindly agreeing to an interview
and for his subsequent time and care in answering my questions.
At Edward Elgar, I am grateful to Francine O’Sullivan for her patience
and good humour throughout the whole process.
A special note of thanks goes to Cornelia for her help and support, and
for the very best of times. Finally, I am greatly indebted to my parents for
everything that they have done for me over the years. This book is dedicated
to them.

xvi

Bend it like Becker
Simon W. Bowmaker

Gary Becker is University Professor of Economics and Sociology at the
University of Chicago and a Senior Fellow at the Hoover Institution,
Stanford University. He is best known for his work on microeconomic
issues including human capital, economics of the family, economic analysis of crime, discrimination and population. His books include, A Treatise
on the Family (1981 and 1991), The Economic Approach to Human Behavior
(1976), The Economics of Discrimination (1957 and 1971), Human Capital
(1964), Accounting for Tastes (1996) and The Economics of Life (1997). In
1992, he was awarded the Nobel Prize in Economic Science, ‘for having
extended the domain of microeconomic analysis to a wide range of human
behaviour and interaction, including nonmarket behaviour’.
I interviewed Gary Becker on Friday, 12 March 2004 in his oﬃce at the
Hoover Institution, Stanford University.
Early discrimination
This book applies microeconomics to social issues. I understand that when you
were ﬁrst introduced to economics as a Princeton undergraduate the subject
did not help you understand such issues. Why was this and what made you
change your mind?
Well, I didn’t believe it helped me understand social issues because of the
way price theory and economics was taught, particularly in those days. I
had a very respectable teacher, but it was taught like a series of simple
formal propositions with no connection with real world problems.
That bothered me a lot. I did well on the courses, but I felt there was a
disconnect with what I was interested in and what was being taught in
economics.
Now, I changed my mind really when I went to Chicago to Graduate
School. There is no doubt the major inﬂuence on me was Milton Friedman
because Friedman taught in those days a ﬁrst-year price theory sequence
and he taught it as a ﬁeld where the theory was useful in explaining all types
of phenomena. He had many examples, it was rigorous, and became the
basis for a book he put out called Price Theory. That really convinced me
that, yes, this could be a powerful tool to understand the problems I was
interested in.
xvii

xviii

Bend it like Becker

Once you began to apply economics to social issues such as discrimination you
received a fair degree of criticism. Were you conﬁdent that your approach
would eventually be accepted?
Well, conﬁdent is maybe a little too strong [laughs]. I did believe that what
I was doing on discrimination and subsequently later on other issues was a
useful way of looking at it. And it was important to me that I had the
support of some of the professors at Chicago who I admired the most,
including Milton Friedman, Greg Lewis, Theodore Schultz, to mention a
few, two of whom subsequently won the Nobel Prize in Economics. If they
had conﬁdence that what I was doing was signiﬁcant, my own belief made
me not conﬁdent that it would win out but more willing to take the risk that
this was a good way of looking at it and ultimately the profession would
come around. It took a while, but that was my belief. I needed that support
from the people I mentioned. That was very important.
Happy families
We devote chapters in this book to economic behaviour of the family,
speciﬁcally marriage, divorce and assisted reproduction. Why do you think
your work in this area has been the most controversial?
Well, it probably was the most controversial. I hesitate a little in agreeing
with that statement. Certainly, discrimination was very controversial for I
would say ten years before economists began to take these issues seriously
and merge into the ﬁeld. At that time, when I mention anybody working on
discrimination, including eminent economists, they didn’t think that was
economics. Some of the people at Chicago required a sociologist on my
committee to keep me in line because they were sceptical.
Even the Nobel Committee in 1992 mentioned discrimination, crime
and human capital with approval, but when it came to the family, it was
still considered controversial. It’s changed a little since ’92, but not enormously, although an increasing number of economists are using a similar
approach to marriage and divorce. There’s actually a kind of mini-boom
in that area and not long ago The Treatise on the Family was considered in
the top 20 books in sociology over the last 50 years. I think it’s had some
impact.
The reason it remains controversial is that one is dealing with very intimate questions when one talks about the family. There is a belief that economics is a cold discipline. How can you talk about marriage, love, having
children, divorce – issues of that type – which are emotional decisions at a
certain level, with this cold calculus? I try to tell people that it’s not cold calculus, it’s a theoretical way of looking at people’s behaviour, which has a lot

Bend it like Becker

xix

of common sense to it. People want to do as best they can, and when I
restate some of the language, I ﬁnd many begin to agree. But then they say,
‘that’s kind of obvious’, and so the question is to show them certain implications of the theory that surprise them, such as who is most likely to
divorce and why birth rates may have gone down.
Do you believe changes in divorce laws account for the relatively high divorce
rates in the USA and Europe?
No. In The Treatise on the Family, I did a little calculation for California
and some theoretical analysis which suggested that the movement toward
no-fault divorce would have basically no eﬀect on the divorce rate. That discussion produced a fair number of subsequent papers that are still continuing. On the whole I think it fair to say that even among those people who
say there was an eﬀect on the divorce rate believe it was a small part of the
total increase in the divorce rate. There is nobody to my knowledge who has
studied this question empirically who has concluded that the change in the
divorce laws was the major reason for the increase in divorce rate. I believe
if it was a factor, it was a very small factor and there are other forces such
as the increased labour force participation of women, the lower birth rates,
the higher earnings of women compared to men and a bunch of other
factors that one could mention that are far, far more important.
I understand you are in favour of marriage contracts. How exactly would
these work?
Well, I am in favour of marriage contracts. I think it would give families
more ﬂexibility. I have often said that the best way to do it is to obligate that
the family has the contract prior to getting married so that suggesting a
contract to your potential wife or husband would not look like you didn’t
have conﬁdence in the marriage. That’s a bad signal. The contract is a
requirement, everybody has to do it. What it would do I think is enable
people to tailor what they have in their contract to their particular circumstances. Circumstances vary a lot. In the present system, we have to rely on
courts and judges to make adjustments to take account of circumstances.
Since the people and their children are the most involved, they are the ones
that should make the contract.
The main provisions would be the custody of children because there are
third parties to this arrangement. The provisions would be ﬂexible such as
how to allocate property, for instance, and you could re-open the contract
every three years to look at changes in circumstances. You would give
people much more ﬂexibility and I think it would reduce the role of the

Bend it like Becker

courts and judges. Let the people involved be the deciders of the type of
marriage they want. They could have a contract which can’t be broken (you
can’t divorce), and others can have a contract which could be broken by
either party under certain penalties or whatever, like a business contract.
That’s what I think it should be. It would reduce a lot of the problems we
face under the current system.
What are your views on the same-sex marriage controversy in the USA?
I don’t see what all the fuss is about to tell you the truth. I’m generally sympathetic. If you had a contract as the rule, well, gay couples can already
write contracts, they can already adopt children. People say, ‘well, that’s
adoption’, but adopting children is legal by same-sex couples as far as I
know in the overwhelming majority of, if not all, states in the US. I have
known a number of couples who have not only adopted children but had
them through various ways. Let’s say there are two lesbians living together.
One of them would be the host of the child and they would have a sperm
bank at some other place.
You know, the jury is yet out on whether there are signiﬁcant eﬀects on
children brought up by same-sex parents. One could maybe argue that issue
and I don’t know the answer to that. I don’t think anybody does. We don’t
have enough experience.
The marital issue seems to me to be a much less important step. We
already allow people of the same sex to live together. We already in many
respects have moved to make them eligible for a lot of the beneﬁts in companies and so on. So, I’m not sure why this has hit a button someplace in
the so-called Right. Maybe you don’t want to call it marriage and everybody would be happy but basically we are allowing it to happen today in
terms of contractual arrangements. If we went to a marital system, I would
say, yes, people of the same sex should be allowed to enter into that kind of
relationship. I don’t see why not from my point of view. If there is a controversy, it is over the children issue. That I don’t have a strong opinion on
because I don’t think we really know enough about that. It seems to me
there is a mis-emphasis someplace in the discussions.
How do you see the family (and its economics) evolving over time?
Well, it’s changed enormously in the last 50 years, no doubt more rapidly
over the last 50 years than in the previous 500 years. I think we understand
not all of the reasons, but a lot. It will continue to evolve, although at a
slower pace. Let’s say 40 per cent of ﬁrst marriages now break up. I don’t
expect we will ever get into a situation where 80 per cent of ﬁrst marriages

Bend it like Becker

xxi

break up or anything like that. You will see the increase slow down dramatically in that regard.
Birth rates have come down a lot. They have low replacement levels
throughout Western Europe and much of Asia. I estimate that almost half
of the world’s population are in countries with below replacement fertility.
It’s been a remarkable change. Are birth rates going to go lower? They
could, but it is hard to believe they could go much lower in countries like
Italy without immigration. In 25 years, they are going to have a much
smaller population, in 50 years, an even smaller one. So I don’t expect much
evolution in terms of those countries. I do believe that birth rates will come
down in countries that have high birth rates like India and some in Latin
America and Africa. That will be an important step.
I think divorce rates will not rise much more radically, and more women
will come into the labour force. These are the directions in which I think the
family will evolve. We will get a much more complicated structure of families with people in second marriages and stepchildren mixing together. One
may or may not like that, but I think that is a clear response to basic forces
that, in a broadly deﬁned sense, are economic in nature.
Are we likely to see, particularly in the USA, more inter-racial marriages?
Probably, but that’s been slow to pick up. We have very low rates of interracial marriages, although they are perfectly legal.
Do you know why might that be?
Well, it’s not diﬃcult to understand why. I have studied that question
although I haven’t written or published anything on it. It’s still true that
blacks are considered in a lower-level position in the United States than
whites. And typically they are. So, for a white person to marry a typical
black person – I say typical because there are many blacks who are extraordinarily successful – they have the feeling they need some compensation
for that. If you look at inter-racial marriages, you will ﬁnd that the white
person marrying the black person will get some compensation. When they
are marrying a black man, they (the black man) will be more educated than
the man a woman of her characteristics would get if they married a white
person. It’s a compensating diﬀerential if you like. I say compensating not
for any intrinsic inferiority but because the social perceptions in the United
States of blacks are at a lower level.
Now, to the extent that this weakens, and it has weakened a lot in the last
30 or 40 years, you would expect to see an increase in inter-racial marriages
and we have had some increase. And we have had them by inter-religious

xxii

Bend it like Becker

marriages and so on. That has increased a lot. How far will it go? I don’t
know. It has been a slow process, and I think it will continue to move in that
direction. Will we ever get to the point where we are as likely to have an
inter-racial marriage as a marriage within the race? Not in my lifetime, but
who knows what the more distant future will bring.
We’ve changed a lot. There is still discrimination in diﬀerent areas
against blacks but it is totally diﬀerent to what it was when I wrote the economics of discrimination. When I was growing up, I remember the attitudes
we had as boys. We just didn’t know anything about blacks and there was
tremendous segregation. They weren’t in schools. When I was at Princeton
it was not until my second or third year that the university accepted their
ﬁrst black undergraduate. It’s hard to believe now, but they just wouldn’t
accept any. So, things have changed enormously – for the good. And I think
they will continue to change. That will aﬀect inter-racial marriages. How
rapidly those will increase, I don’t know. I don’t think it will be something
that will be very rapid in the next decade or so.
Crime and the fear economy
This book’s chapter on crime examines some of the factors that may have led
to the fall in crime in the USA in the 1990s. Do you place more importance
upon police enforcement and punishment eﬀorts relative to the strength of the
economy?
Yes, I think if you look at victimization studies, crime actually started going
down in the early eighties. So, I put a lot of emphasis on increased enforcement of the laws, increased conviction of criminals and the growing
number of people in prison. One of the unfortunate by-products of all this
is that we have around two million men and women in prison in the United
States. That’s a huge number, a larger fraction than the other Western countries as far as I know. But I think it has cut crime rates in the United States.
I would say that is one factor.
One of my colleagues at Chicago, Steven Levitt, has written an interesting and controversial paper that I think is in the right direction on the eﬀect
of allowing abortion on the crime rate. He claims – he is a very good empirical analyst – that the eﬀects of the seventies show up in the nineties, 20
years later when these kids get born and don’t get born. That was really
what he was arguing. I think that was a factor.
The prosperity was a factor, but I don’t think it was a dominant factor in
the United States. We’ve had countries like Britain that have been pretty
prosperous where the crime rate has been going up, not down – very
diﬀerent from the US. I think it is because they have adopted a diﬀerent
policy in terms of treatment of criminals. The US change-around in about

Bend it like Becker

xxiii

1980 was very crucial. We were also, so I recall, soft on crime in the sixties
and seventies and crime rates skyrocketed. People said it was due to changing morals and cultures. I argued against that position and for once I think
I was right [laughs]. It wasn’t changing morals. Criminals adapted to what
we were allowing.
The intellectual view that punishing criminals was unjustiﬁed was itself
the crime. Once we got away from that view and tightened up and toughened up, we brought our crime rates down enough so that in all areas of
crimes of violence, we are considerably below Britain per capita now. On
property crime, we are below quite a number of Western European countries. We still lead on violent crime, but even that is narrowing somewhat
compared to Western Europe. So, I think we showed it can be done and I
don’t believe that prosperity was the major contributing factor in my
judgement.
Are you in favour of capital punishment?
Well, that’s obviously a very controversial issue. I think the discussion of it
is often greatly obfuscated. I certainly would be strongly in favour of capital
punishment the more convinced I was that capital punishment reduced
murders. That’s the main issue, not whether the state has a right to take
people’s lives. That’s a mis-statement of the issue because if capital punishment did reduce murders, by not using it the state is indirectly taking
people’s lives by allowing innocent people to be murdered.
So, I think the issue is can we cut down murders signiﬁcantly with capital
punishment? I believe we can. To the extent I felt I was right about that, I
would support capital punishment. If somebody could convince me that
the evidence shows strongly that we can’t, then I go against capital
punishment.
It’s something you don’t want to have to use for lots of reasons. You don’t
want to take people’s lives if you don’t have to. You don’t want to go to war
if you don’t have to. But if you have to go to war, most people are willing
to. If you believe that capital punishment would reduce murders, I think
most people would come along. But people have been convinced that the
evidence shows that capital punishment has no eﬀect. I think that is a misreading of the evidence. I would be willing to accept the view that the evidence is mixed. If it is mixed, then maybe you want to take the minimal view
and not use it. I don’t have any problems with that. I’m not a crusader on
the capital punishment issue, but I believe – you asked me what I really
believe – that it does deter and as a result of that we should use it. But I
also believe one cannot demonstrate that conclusively with the evidence
available.

xxiv

Bend it like Becker

What was your response to the papers on capital punishment produced in the
1970s by one of your former students, Isaac Ehrlich?
I think Isaac did some very important, pioneering papers. He didn’t work
only on capital punishment. He worked in general on deterrence, but he did
come out with a very large number on capital punishment. Other people
have criticized them. The debate is still continuing. There are at least two
dozen papers or so on that subject since then. I’ve just read one recently
that reaﬃrms, if you want, again that capital punishment has a signiﬁcant
eﬀect. Others claim it doesn’t, so it’s an ongoing discussion, but Ehrlich was
the one who got that discussion going.
His number that every execution reduces seven or eight murders is probably too high in my judgement, but I don’t know what the number is. I think
it’s positive. It may well be more than one, although it doesn’t have to be
more than one to be in favour of capital punishment. When you are punishing people who are murderers, the lives you are saving are often innocent
lives. Somehow you have to weight those diﬀerently in my judgement.
In recent years, we have seen a rise in international terrorism. Do you think
there is something we can refer to as the economics of fear?
Yes, actually I have a paper in process on that with an Israeli economist.
They know a lot about terrorism. I think terrorists almost by deﬁnition
operate by spreading alarm and fear. That’s their purpose and they succeed
often in doing it. We’re looking at people’s willingness to ride buses in Israel
after a suicide bombing on a bus. We ﬁnd big falls and some permanent
eﬀects. We have a bunch of other evidence on people’s willingness to go to
restaurants and, from the United States, evidence on people’s willingness to
ride airlines. So, I think we can use an economic model to understand that.
That’s what we are doing in this paper. We’ve developed an investment model
where people invest in trying to control their fear and those who are doing
an activity the most have the biggest incentive to make that investment.
Look at mad cow disease in Europe. A study of that by a French economist showed that people eating meat the most were least likely to reduce
their consumption. Now, that seems surprising, but our view is that it is the
right view because most of the reaction to mad cow disease was excessive.
It was a lot of fear. I think in England there were 150 cases documented, in
Europe there was hardly one, yet there was a big decline in French and other
European consumption. If you are eating a lot of meat you have more
incentive to try to make that investment to control your fear. While if you
are just an occasional meat eater, you might as well just give it up and not
try to associate with it.

Bend it like Becker

xxv

We ﬁnd that with terrorism too. Those people who ride buses regularly
are less likely to cut their bus use than those people who ride it occasionally. We have a bunch of evidence consistent with this interpretation. So, I
guess there is an economics of fear. They (terrorists) operate by fear and I
think understanding the type of responses people make can maybe help us
in reducing the negative eﬀects of terrorist attacks.
What criminological insights or perspectives have you found stimulating or
interesting in recent years?
Well, I think (Steven) Levitt has been one of the leaders in this, two of
his papers in particular. The one which better documents that putting
people in prison in the United States has a positive eﬀect and the abortion paper, although I don’t think that was a major eﬀect. I think it was
a very interesting piece of work and convinced me that was a factor
involved in it (the fall in crime in the US in the 1990s) which I wouldn’t
have thought of beforehand, I must say. So, I think those papers by Levitt
were good.
Controversies over the use of guns have been interesting. I don’t think
we have any clear-cut answers yet on the controversy over whether guncontrol laws reduce crime or raise crime. I’m pretty familiar with the
main papers in the literature, although it’s confusing because people
reach some very diﬀerent conclusions about it. My own belief is that gun
control laws probably do raise crime. I have a very simple argument.
Most people who use guns to commit crime get their guns illegally, so gun
control laws don’t aﬀect their access to them. But the people who would
legally use them to defend themselves would have a tougher time doing
it. So, I have tended to favour a diﬀerent approach: we punish those who
use guns in committing crime much more severely than those who don’t
use such weapons in committing crime. I think that’s a more eﬀective way.
You hit the person using the gun for bad purposes directly rather than
the control of guns.
I think the fact that US crime rates are going down while European crime
rates are going up is a very interesting set of case studies provided to us by
the data. The question is, why this very diﬀerent change in societies that in
other respects are very closely linked? I guess they haven’t been studied
enough. I gave you some of my explanations before, so these are the sort of
things that I think have been important. I can’t say I’m up on every single
study of crime that is going on. Since I haven’t been working on it, I haven’t
followed it as closely as I might otherwise. But this is what I have concluded
from what I am aware of.

xxvi

Bend it like Becker

Drugs and despair
We have a chapter in this book which examines addiction (speciﬁcally, illicit
drug addiction). We discuss your 1977 paper with George Stigler on the relative theory of addiction and your 1988 paper with Kevin Murphy on rational addiction. What stimulated your initial interest in addictive behaviour?
How have your views on addictive behaviour changed over time?
[Laughs] Well, I know what stimulated my interest. George Stigler and I started
to write this paper on why preferences are stable and the same. Then we started
thinking about examples. I remember he came to me with a quote from Alfred
Marshall that the more good news that a person hears, the more they like it. I
think we have that quote in our paper. I started thinking about that. I said,
‘One could think that one has a stock of capital and that it inﬂuences one’s
desires for it’. So we wrote that section on addiction. We didn’t quite get the
formal analysis right in my judgement but we were in the right direction.
Then I had a student, Larry Iannaccone, who wrote a thesis on addiction. I thought it was a very good piece of work and that stimulated
Murphy and me to get involved. We worked out this paper which became
the 1988 paper on rational addiction which was the development of both
what Stigler and I did and what Iannaccone did.
Have my views changed since then? I must say, not a lot [laughs]. We were
lucky in stimulating a fair bit of literature. It surprised me at the time that
there would be a number of papers. But a lot of economists seem to have
gotten interested in that question, some of them testing the rational addiction model, including Michael Grossman and others, some supporting it,
some rejecting it, other people trying to come up with alternative models
based upon inconsistent behaviour over time. We even discussed some of
these types of models in our 1988 paper but they have formalized them a
little more in terms of hyperbolic discounting, for example.
What do you think of these extensions to the rational addiction model?
Well, I know those papers pretty well, particularly the hyperbolic discounting work and the work on happiness, such as whether addicts are
more or less happy and like high taxes. I haven’t been convinced that these
extensions help us understand behaviour better. I have had these people out
to workshops at Chicago. Gruber’s done quite a bit of work on it and
Laibson has done hyperbolic discounting.
I am very friendly with both of them and I look at papers of this type.
But I haven’t felt they have advanced if you look at the data, including
people who state that they want to quit but can’t. Most of these issues I
think we anticipated in the discussion in the 1988 paper. I think we can do

Bend it like Becker

xxvii

as well at explaining this data as they can. In fact, a couple of these papers
(the one by Gruber, maybe one by Cutler and Glaeser) admit that from a
positive point of view, rational addiction does well, but from a normative
point of view, these other models have diﬀerent implications. They do, but
the question is, are they the right models for this behaviour?
So, I would say I like this literature. I think it’s good to try to confront
the rational addiction model. Eventually, it will have to be modiﬁed in some
fashion, but I think it has stood up pretty well to a lot of onslaught.
Empirically, it has done pretty well. Have I radically changed my views on
it? No, I haven’t. Rational addiction captured an important part of addictive behaviour and when the smoke clears, that will be the judgement in this
literature. Though maybe it didn’t capture all of it, so we will have to
modify the 1988 paper in some way.
We also examine in this book the economics of drug prohibition and we look
at policy alternatives. Would you legalize drugs? If so, how would you regulate a market for drugs?
Yes, I believe we should legalize drugs for a lot of reasons. For a pragmatic
reason, I think it [the war on drugs] has been a very costly programme and it
has had a very low beneﬁt/cost ratio. Now, what would we do if we did legalize
them? For one thing, like alcohol, we would want to have an age restriction on
who gets access to it. You may want to tax it if you believe it has certain externalities on other people. People who become addicted may drive or come to
work and harm others. A better way to treat that is to punish more severely
those who are driving while under the inﬂuence of drugs or who are at work
under the inﬂuence. They would be the main sort of things I think would be
necessary. It would be a fairer system and less discriminatory against blacks and
poor people. They are the ones who suﬀer the most under the present system.
How does that work?
Well, it works in three ways. One, if you look at the people who are sent to
prison for drug convictions, a huge disproportion are blacks or other
minorities in the United States. I know the data well for the US and I think
it would be similar for Britain. [Under legalization] They wouldn’t be there,
at least not for those problems.
Second, the drug markets are usually centred in poor parts of the city, so
poor neighbourhoods suﬀer as a result of that. You get a lot of transactions going on there. And not surprisingly when they are centred there, a
lot of crime is more common in general. The police have their hands full,
so it is hard to control the drugs trade.

xxviii

Bend it like Becker

And third, because it is concentrated there, access to drugs for poor,
inner-city kids is easier than it is for rich kids. That is one reason why we
have the law. I think it’s really the middle-classes [laughs] battling against
the poor and putting more of the burden on the part of the poor. It is discriminatory, maybe not intentionally, but that’s the way it works out,
against the poor and against the blacks, Hispanics and other minorities.
A legalized system, with taxes, would raise the price of drugs and the
poor would be more responsive to a higher monetary price. The middleclasses are more responsive to a higher non-monetary price. For example,
you can’t get good jobs if it was known you took drugs. Look at the way
[Douglas] Ginsburg lost the possibility of getting on the US Supreme
Court because somebody reported that he smoked marijuana when he was
at college. That was a big price at the end of the day.
So, for all those reasons, I would legalize it, I would tax it if need be, and
I think that would have a lot of beneﬁcial eﬀects.
The economics of suicide is another topic featured in this book. Do you think
economists have a role to play in understanding suicidal behaviour? For
instance, do you think we can help explain the alarming rise in recent years of
young male suicide in the USA and Europe or the emergence of suicide
bombing in the Middle East?
I deﬁnitely think so. In fact, I have on-going work with Richard Posner on
the economics of suicide. We have a way of approaching it which I think
is interesting. We are collecting a fair bit of data and trying to test it against
a lot of things. As you say, young males have a high suicide rate, older
males are also high, older females are not so high. Females are much
higher in attempts and lower in actual suicides. There are a lot of regularities that are interesting. Blacks have relatively low suicide rates even
though low-income people tend to have a higher suicide rate than higher
income people. It’s an interesting question. I’m not sure I have all the
answers but it’s a challenge to try to explain that. And there are a lot of
other regularities that one sees.
We have a quote from Arthur Schopenhauer and David Hume, both of
whom said suicide is a rational response to circumstances. Some people’s
lives are miserable. You almost never see a happy person commit suicide.
What causes suicide? Well, suicides are caused by severe, negative events
that hit people like bad health, going to prison, losing your job, losing your
money, divorce – these are all suicide stimulators. Some people get over that
quickly, but for other people it is devastating to have that to overcome.
Their utility levels go down, you might say, below that threshold where they
begin to think they are better oﬀ dead than alive.

Bend it like Becker

xxix

There are other people who, biologically, have a propensity to depression. A number of suicides are depressed individuals, some of whom have
attempted suicide in the past. Now, one might say, ‘we have to help these
people’. We try and often we succeed. Anti-depressants I think have been
an important tool, but up until the nineties we had very few tools. Psychoanalysis was not that successful in my judgement in treating highly suicidal,
depressed people. These were people who go through many episodes of
being really down, and then they come out of it sometime and then come
into it again. They begin to think, ‘Well, this is what my life is going to be
like, it’s going to be pretty miserable’. These people commit suicide. Is that
rational? I think so.
Do you think a suicide bomber is a rational individual?
I think it’s useful to interpret suicide bombers as rational in this following
sense. They want to achieve a goal and they are willing to give up their life
to achieve that goal. The hero in war risks and often loses his life in order
to accomplish an aim. Do we call that an irrational act? Well, think of a
suicide bomber as engaged in a war. They are willing to give up their lives
to harm others. So, that’s not on the face of it an irrational act. They may
be indoctrinated, so we have to think of preferences being formed, but
that’s okay, that’s part of economics – the formation of preferences. You’ve
got to understand that process.
If you raise the cost of committing suicide, namely by reducing the likelihood they will succeed, you will cut down suicide bombing. Israel hasn’t
stamped it out, but by various tactics (some strong-arm, others by protecting various things and careful checking), they have cut it down. So, this
strikes me as some evidence that you do get responses in suicide bombing
to reducing the likelihood that you will succeed. And there are other
approaches you can use.
When 9/11 came out, one of the problems I set on the theory course I
teach was can you interpret suicide bombing in a rational framework. I
think you can. It’s also a useful way to help understand why young people
get involved in doing it, what their goals are and how to combat it. Those
things go together.
Behavioural economics and wackonomics
Does the behavioural economics approach of economists such as Thaler and
Camerer limit or extend your approach to the study of human behaviour?
It does both. I think behavioural economics has been useful in economics.
It is challenging some things that economists accepted too readily maybe

xxx

Bend it like Becker

and introducing some concepts that may turn out to be useful. Although it
has certainly not provided any alternative, systematic theory to compete
with neo-classical theory, it might ultimately enrich it, like hyperbolic discounting which is not Thaler or Camerer, but Laibson.
Now, Laibson formulates that within a rational choice framework. You
have people who are rational but it is just that they are inconsistent over
time. So, they have got to take rational actions to recognize that. I think for
some problems that might be useful. But again, I haven’t been convinced
that for much behaviour like addiction we need hyperbolic discounting. But
nor have I been convinced that it’s worthless and not going to be useful. We
will have to wait and see. That would enrich the theory of rational choice.
Now we have diﬀerent ways of discounting. For some problems we can
maybe show why people are going to be hyperbolic and others maybe not.
That would be the goal at least.
Similarly, in other areas, like in ﬁnance, some behaviourists have pointed
out some empirical diﬃculties in the orthodox theory of eﬃcient markets
and the like. Do they have an alternative theory? Not really, not much of a
successful one. But they have made people less theological about eﬃcient
markets, by having to see how we can amend some things and make it work
better. There is good work going on both by people like Thaler and others
but also by people who are much more strongly supportive.
So, my own overall evaluation of behavioural economics is that while it’s
been useful, it’s not been a revolution. It’s usually too focused at the individual level, not at markets. Economics is about markets, it’s about groups.
We have a theory of individuals to get at group behaviour, but we want to
know what an excise tax does to behaviour or what more punishment does
to crime. We are not so focused on the individual, but behavioural economics and experimental economics, much of it, not all, is focused on the
individual.
This is one reason why there is a disconnect between psychology which
is usually individual and economics which is group. Let me give you a
reason why these diﬀerences may be huge. Most people can’t make probability calculations as accurately as some behaviourists suggest. Those
people don’t go working for casinos and other places where these type of
calculations are important. They won’t survive in that kind of job. The
blackjack dealer when he is playing against the players has to know when
to stop. If you watch dealers, they are basically working out through experience what we would call optimal stopping rules. That’s been shown.
People who couldn’t do that wouldn’t be running blackjack tables.
That’s true in a lot of other areas. You’ve got to look at the specialization that we observe in markets and embed a theory of individual choice
within a market specialization framework. Now, when you do that you can

Bend it like Becker

xxxi

well understand why most people might not be good at something but they
don’t end up being doctors or lawyers or whatever it may be. And is the
market equilibrium we get close to what we think we get from a rational
framework? I think it is, although it’s not perfect. Behavioural economics
can maybe ﬁll in some of the edges there, but it certainly hasn’t given us an
alternative approach to understanding these problems in my judgement.
Finally, I think it fair to say that a number of people may see some of the
topics covered in this book and believe it is a case of ‘wackonomics’. Thirty
years ago, Alan Blinder wrote a parody paper published in the Journal of
Political Economy entitled, ‘The Economics of Brushing Teeth’. Are there
any spheres of human activity where the application of economics does not
belong?
[Laughs]. Take Blinder’s paper. I thought Blinder’s paper was very cute and
clever. Indeed he was probably right that waiters have a greater incentive to
brush their teeth and have good breath than bus-boys or people washing
the dishes. He was dealing with it in a humorous way with minor little
issues. But it is true that waiters want to get tips and so on.
Are there any areas where it (economics) is not applicable? Probably. Do
we know what they are? No. How do I reconcile those two statements? I’m
not saying that at the moment this approach to behaviour explains everything that we see in the world. A lot of topics have been diﬃcult, like war.
When a country goes to war people become religious. There is interesting
work going on in the economics of religion. But nobody, including leaders
like Iannaccone, would claim that they have solved all of the problems to
understand everything about religion.
So, with every area, there are still considerable gaps in our knowledge,
whether it’s the family or addiction. Some of the gaps may be because we
haven’t used the theory well enough. That’s my position, but I don’t know
if all of the gaps are due to that. Maybe we have to modify the theory in
certain ways. I have an open mind on that question.
All theories evolve over time. Look, Newtonian mechanics for certain
purposes were replaced by quantum mechanics and certainly economic
theory which is not at the level of Newtonian mechanics in terms of generality and power will be modiﬁed over time. That’s inevitable, that’s what
a good, empirically-based science does. Economics is an empirical science
and I am well prepared to see the theory being modiﬁed over time. Which
modiﬁcation, which areas, I don’t think we know that yet. But I am sure
there are areas out there that will need some important modiﬁcations to the
theories.

Introduction
David D. Friedman

To most people, economics is the study of the economy, important but
boring issues such as unemployment, economic growth and inflation. To
many economists, economics is not a list of issues but a way of understanding behaviour – all sorts of behaviour in all times and places. The
economics of suicide is not about its eﬀect on the GNP but about using
the tools of economics to understand why people kill themselves. The
economics of law – my current field – is not about how changes in the law
aﬀect the economy. It is about law as incentive, legal rules understood as
ways of changing how it is in the interest of individuals to act. Applications
of economics, across the whole range of human behaviour, share one
underlying assumption: that behaviour is best understood by assuming that
individuals have purposes and tend to take the actions that achieve them –
are, at least in that limited sense, rational actors.
From rationality we get the economist’s approach to value – the principle
of revealed preference. If I act to achieve my objectives, you can deduce the
objectives from the acts. If I buy an apple for a dollar but refuse to buy it
for a dollar fifty, I must value the apple at more than the former amount
but less than the latter. If I boast of my love for great novels but spend
my time reading thrillers instead, that tells you two things about me – that
I value thrillers more than great novels and that I value other people thinking I like great novels more than other people thinking I like thrillers. Values
are judged by actions. Viewed from the inside, as humans rather than as
economists, rationality seems an implausible assumption. Most of us
observe elements of irrationality and error in other people, occasionally
even in ourselves. Viewed from the outside, as observers trying to find a
pattern in the interacting choices of millions of people, it is the best clue
we have. None of us is perfectly rational. But the tendency to choose on
the basis of the ends we desire is a predictable element, perhaps the only
predictable element, in our behaviour.
Starting with this one fundamental assumption, economists have generated a useful toolkit of concepts for explaining and predicting behaviour:
supply and demand, prices, sunk costs, externalities, opportunity cost, marginal value, present value, risk aversion, joint products and many others.
Such concepts are usually developed in the context of explicit markets
where goods and services are exchanged for money. They apply equally well
1

Economics uncut

to the broader range of behaviour considered here and in the work of economists such as Gary Becker and James Buchanan. Sunk cost explains why a
firm that has already paid the cost of designing a product and building a
factory may continue to produce even after discovering that the price the
product can be sold at will never pay back the full cost – because losing part
of the cost of development and tooling is better than losing all of it, which
is what will happen if it scraps the project. It also explains why the customer
of a prostitute, having spent time and eﬀort finding someone selling the
services he wants to buy, may go through with the deal even after he discovers that the price/quality combination he is being oﬀered is not good enough
to repay the cost of finding it. The theory of joint products explains why
a rise in the demand for silver will tend to produce a fall in the price of gold –
the two being refined from the same ore. It also helps explain the economics
of sex, marriage and abortion, since, in a world with imperfect contraception, sexual pleasure and (probabilistic) pregnancy are jointly produced.
Or consider suicide – the subject of Chapter 8 of this book. When is it
rational to kill yourself ? The simple answer was given more than 200 years
ago by David Hume, an economist as well as a philosopher: it is rational to
kill yourself when life is no longer worth living. That answer implies
predictions of who will kill him or herself and when. It also suggests a
highly controversial conclusion – that preventing suicide is not necessarily
a good thing.
Hume’s explanation of suicide is incomplete; like many starting points,
it assumes away a number of possibly relevant complications. Consider the
decision to charge a machine gun, a form of suicide engaged in by a very
large number of young men early in the last century. The reason to go over
the top in World War I was not that life was no longer worth living but that
there was something to be gained by dying. It is rational to buy something
as long as its value is more than its cost.
If you are unhappy at viewing death in battle as suicide, I have another
example. One of my current interests is the legal system of Imperial China.
In that system superiors – familial as well as political – had enormous
advantages over inferiors. For a son to kill his father was one of the worst
crimes known to Chinese law, punished by death by torture. A father was
free to kill his son for what we would regard as trivial oﬀences. A senior
relative could order a junior relative around. Punishments, even for serious
crimes, were scaled up when the victim was senior, down when junior.
The weak did have one recourse. Driving someone – even a junior relative – to suicide was a serious oﬀence. Judging by accounts of actual cases,
it looks as though some suicides were committed in order to invoke that
legal rule. If life was not very much worth living, the prospect of getting
revenge on the person responsible might tip the balance.

Introduction

This example points out an ambiguity in the economic approach to
understanding behaviour. We assume that individuals have objectives and
act to achieve them, but we do not know exactly what those objectives are –
they might, for instance, include revenge. This is a problem because if we
make no assumptions at all about objectives, economics loses all predictive
power; you can make any behaviour consistent with rationality by assuming
that the behaviour was itself the objective. Why do I spend all my time sitting
in a chair doing nothing? Because I like sitting in a chair doing nothing.
The ambiguity is dealt with in practice by a balancing act. We assume
objectives that we observe most of us actually have – money to buy goods,
health and leisure to enjoy the goods, the welfare of those near and dear to
us. We add in other objectives sparingly, when the addition gives a large return
in our ability to make sense of the world for a small cost in increased ambiguity of the theory. And we try to avoid using ad hoc assumptions about what
people want, to explain away apparent inconsistencies between what they do
and what our theory predicts – in the hope that there may be more interesting explanations to be found. Better a theory whose predictions are occasionally wrong than a theory that can explain anything and predict nothing.
Suppose you agree that the economic approach as I have described it is
relevant to a wide range of behaviour – that this is a potentially useful
toolkit of ideas. What should we use it for?
One use is to make sense of the world around us. Important parts of that
world consist of the behaviour of other people – living, dying, buying,
selling, marrying, divorcing, getting pregnant or having an abortion, using
or not using recreational drugs legal or illegal. The tools of economics
provide one way of understanding that elaborate dance; the attempt to
understand it and to test that understanding against the observed facts can
be a rewarding and intellectually exciting project.
A second answer is that we use economics to figure out what we, and other
people, ought to do. At the individual level that can be as simple as deciding
whether to buy or rent, stay in school or go on the job market, look for a
husband or wife or continue to play the field. The project of advising the rest
of the world brings us to fields such as policy analysis and welfare economics, in which economics is used to help decide what laws should be passed,
what policies followed by governments, how the world ought to be run.
I am, as you may have suspected, an economist. I am, at this very instant,
advising someone on what to do. You are the someone, and the advice is to
read this book. Why?
The economics of economics
If you are one of the first settlers on a newly discovered and previously
uninhabited island you will, sensibly enough, choose the best location on

Economics uncut

the island – the most fertile valley if you are a farmer. If you are a miner
with a newly discovered prospect, you will try for the richest veins of ore.
If you are one of the first entrepreneurs in a new field – making money on
the Internet, say – you will try to grab the most profitable opportunities.
Perhaps, if you are very clever, you will invent Google.
The same principle applies to academic fields. Economics is about,
among other things, unemployment and inflation and growth rates and predicting next quarter’s statistics. But it has been about those things for a very
long time, long enough so that thousands of economists have made their
careers trying to add just a little more to what we know about them. The
island is thickly settled, the vein of ore largely worked out. It is still possible, if you are suﬃciently brilliant and suﬃciently lucky, to do something
new, interesting and important. But the odds are not very good.
Consider, in contrast, the economics of suicide, or abortion, or drug
addiction, or prostitution, or divorce. A great deal less has been done in
those fields – little enough so that, as you will see in the chapters that follow,
their authors can cite very nearly everything and summarize quite a lot of
it. It follows that there is a lot still to be done – fertile fields lying empty, ore
to be mined, puzzles to be solved, careers to be made.
As evidence I oﬀer a short list of puzzles gathered from reading the rest
of this book. All of them are implied by those chapters, none is satisfactorily answered. Solutions to some, perhaps most, should be within the power
of many of the people reading these words. And finding them should be
more fun than adding one more footnote to the next paper on the causes of
inflation.
1 The cost of suicide
Chapter 8 cites an attempt to estimate the cost imposed on a society by
suicide, based on the loss of the output that the suicide would have produced in the rest of his or her life. That approach reflects, in my view, a
serious misunderstanding of what economics is about. The output I
produce comes back to me as income – the payment I get for producing it.
The fact that I choose to kill myself is evidence that the value to me of that
income is less than the cost to me of continuing to live. From the standpoint of economics, that cost – whether the misery of rejected love or the
pain of a terminal disease – is just as real as the benefit produced by
growing another bushel of grain. Economics is not about stuﬀ, it is about
people.
How would you do the calculation correctly? How could one estimate, or
at least set bounds on, the net cost (or benefit) produced by people killing
themselves, taking into account the benefit to those people of not living as
revealed by their choice not to live?

Introduction

2 The value of drugs
Calculations of the social cost of drug addiction typically make a similar
mistake – they ignore the fact that recreational drugs, like other forms of
recreation, provide a benefit in pleasure to the user. Suppose you are willing
to take revealed preference seriously, as economists should, and so wish to
include that value in your calculations. Using available data and the usual
economic approach to evaluating the benefits from production and consumption – consumer and producer surplus – how might you estimate the
net cost imposed upon a society by laws making some recreational drugs
illegal? In calculating benefits to be compared to that cost, what things do
or do not legitimately go into the calculation?
3 Violence and drugs: which should we punish?
Suppose you believe that the use of some drugs can be a direct cause of
violent behaviour. There are two ways the law might deal with the problem.
It could penalize drug use directly or it could penalize it indirectly, by
punishing violent behaviour, whether or not drug use is its cause. What are
the relative advantages of the two approaches?
4 Addiction and marriage
Chapter 1 discusses a model of drug addiction in which current use of a
drug increases the future cost of not using it. Chapter 5, in discussing
cohabitation prior to marriage, treats it as an investment in information,
likely to make people less likely to get into marriages they will later want to
get out of. Suppose we apply the ideas of the former chapter to the content
of the latter. Living with someone is addictive behaviour – breaking up is
hard to do.
What are the implications for the eﬀect of cohabitation on divorce rates?
On whether norms that limit cohabitation outside of marriage make us
better or worse oﬀ? How are your conclusions changed if we assume that
cohabitation is addictive for only one partner – an assumption arguably
justified by arguments from evolutionary biology suggesting that men have
a greater taste for promiscuity than women?
5 Sex, drugs and rock ’n’ roll
Chapter 6 argues that the consumption of pornography can be viewed as
an addiction since, as with addictive drugs, the pleasure from current consumption depends in part on past consumption. By this definition quite a
lot of things are addictive, including music, friendship and literature. Can
you think of any economically interesting diﬀerences between ‘good’ addictions such as those and ‘bad’ addictions such as, presumably, heroin and
pornography?

Economics uncut

6 Profit maximizing cops
Chapter 4 discusses one issue of resource allocation in the criminal justice
system – decisions that may lead to overcrowding prisons. It does not
explore the wider question of what incentives police and prosecutors face
and what behaviour they imply.
One example is civil forfeiture. Under existing US law, property used in
the commission of certain sorts of crimes forfeits to the government even
if the owner of the property has not been convicted, or even charged, with
doing anything wrong. How does that aﬀect the incentives of police and
prosecutors? To what extent does it depend on where the forfeited property
goes – whether to the police department responsible for the seizure or to
help pay for the US public schools?
For another example, consider that being careful not to convict the innocent makes it harder to convict the guilty. In a world of limited resources,
the criminal justice system faces a range of alternatives – from a policy
heavily weighted against the risk of convicting an innocent defendant to a
policy of going after anyone the prosecutor thinks he can convict, innocent
or guilty. How do current rules aﬀect that choice? What evidence can we
find in the actual behaviour of police and prosecutors? How might one
change their incentives so as to reduce the probability that an innocent will
be charged and convicted?
7 The optimal age of prostitutes
Chapter 7 provides an interesting account of the market for sexual services,
gathered from an online database of reports on British prostitutes by their
customers. One puzzling feature of the data is that it implies two diﬀerent
numbers for the age at which a woman is most sexually desirable. One is
deduced from prices: the more desirable a woman is the more, on average,
she can charge for her services, so we can deduce an optimal age from the
age at which the reported price is highest. The other is deduced from reports
of consumer satisfaction. The two numbers are not the same.
One possibility is that this is simply a statistical aberration – that if we
had enough additional data, the two figures would converge. But I can
think of others. Can you? Do they explain other diﬀerences between determinants of price and determinants of satisfaction reported in the chapter?
8 The wages of sin
The same chapter presents a model of the market for prostitutes in which
their high wages reflect the opportunity cost to them of giving up the
opportunity for marriage. This assumes that the number of women who do
not want to get married is insuﬃcient to provide the prostitution services
demanded at a price corresponding to the ordinary market wage, since

Introduction

otherwise there would be no need for wages to rise high enough to bid
additional women out of (future) marriage and into prostitution.
Data provided in the chapter, along with your own observations and
other information readily available, ought to be suﬃcient to tell you whether
the explanation oﬀered makes sense. If not, what other explanations might
there be for high wages of prostitutes?
9 Prostitution as an inferior good
The author of the chapter conjectures that prostitution is more common in
poorer societies and oﬀers an explanation – that higher wages in alternative activities, including marriage, make prostitution a less attractive profession. This assumes that rising incomes have no eﬀect on the demand for
prostitutes – that richer customers are not willing to pay a higher price –
which seems implausible. Can you think of features of poor societies other
than their low average income that might make prostitution more common?
10 Is prostitution a victimless crime?
In most societies prostitution is either illegal, a violation of social norms,
or both. Why? Who is made worse oﬀ by the existence of prostitution?
Economics rejects the obvious answer: if women thought being a prostitute
was worse than some other alternative available to them they would not
choose to be prostitutes. Are there more plausible candidates, important
groups of people made worse oﬀ by prostitution whose opposition might
help explain its widespread negative image?
11 How to create a shortage
Chapter 10 mentions a shortage of children for adoption in the 1970s and
1980s. Under US law, then and now, payment by adopting parents to a
child’s natural mother in exchange for permission to adopt was illegal –
price control at a price of zero. Under those circumstances, what would we
expect to happen as supply and demand curves on the adoption market
changed? What would have happened if the parties were free to negotiate
on price? What benefits or costs can you see to shifting from one system to
the other?
12 Do accidents cause people?
Children born out-of-wedlock are commonly viewed as accidents, an
unwanted side eﬀect of intercourse. It was therefore widely expected that
better contraception and the legalization of abortion would sharply
decrease the number of such children. In fact, the pattern has been the
opposite. Improved contraception and increased access to abortion in
developed Western societies has been associated with a sharp, perhaps

Economics uncut

historically unparalleled, increase in out-of-wedlock births. Can you oﬀer
explanations for that consistent with the economic approach to human
behaviour?
I hope that, by the time you finish this book, you will have answers for some
of these questions – or further questions of your own devising, with
answers for some of them. If so, you may want to consider enlisting in our
enterprise, arming yourself from the economist’s toolkit, and setting out to
explore the vast wilderness of human behaviour – all behaviour, in all times
and places.

PART I
SINS AND NEEDLES

Economics of drug addiction
Simon W. Bowmaker and Frank Heiland

From the concept of a double-kinked demand curve to the theory of rational addiction, economics can provide significant insights into the causes
and consequences of illicit drug use. In 1890, British economist Alfred
Marshall touched upon addictive behaviour in Principles of Economics:
Whether a commodity conforms to the law of diminishing or increasing return,
the increase in consumption arising from a fall in price is gradual; and, further,
habits which have once grown up around the use of a commodity while its price
is low are not so quickly abandoned when its price rises again.

Phlips (1983) notes that with this statement Marshall captures the three
fundamental aspects of addiction: physical response (tolerance), irreversibility (withdrawal) and positive eﬀects of habits (reinforcement). But
most economists since Marshall have tended to view an addict as a myopic,
imperfectly rational individual whose behaviour is not conducive to standard economic analysis. Thomas Schelling (1978), in describing a smoker
who wishes to ‘kick the habit’, stated:
Everybody behaves like two people, one who wants clean lungs and long life
and another who adores tobacco . . . The two are in a continual contest for
control; the ‘straight’ one often in command most of the time, but the wayward
one needing only to get occasional control to spoil the other’s best laid plan.

The purpose of this chapter is to show that the tools of economics can be
employed in the study of drug addiction. We proceed as follows. Section 1
provides a brief overview of the nature and extent of global illicit drug use.
Section 2 then considers various theories of addiction including the aforementioned theory of rational addiction in which addicts are shown to exhibit
consistent, forward-looking and individually optimal behaviour. Next,
Section 3 explores the price responsiveness of illicit drug users from a theoretical perspective. Are such individuals likely to be more concerned with a
drug’s ‘money’ price or its ‘eﬀective’ price, the latter taking into account
factors such as the risk of acquiring and taking the substance? Might we
expect younger, lower-income individuals to be more price-sensitive than
older, higher-income individuals? Is it possible for us to conceptualize a
market demand curve for illicit drugs that has varying elasticities for
diﬀerent categories of drug users?
11

Economics uncut

Section 4 provides a review of the empirical work undertaken by economists in estimating the price elasticity of demand for illicit drugs. The
majority of these studies are from the USA, reflecting the greater availability and development of price data in that country. The primary conclusion
from our literature review is that illicit drug users are just as, or even more,
sensitive to price changes as cigarette smokers. Section 5 contains a brief
analysis of the social costs of illicit drug use and shows how a welfare economics framework can be used to assess the rationale for government intervention in illicit drug markets. Section 6 concludes the chapter.
1 Illicit drug consumption statistics
A recurring theme throughout this book is the diﬃculty in obtaining reliable
and consistent data relating to many of these subject areas. Illicit drugs are
no exception. By definition, it is of course a largely ‘hidden’activity, and aside
from the diﬃculty of observing the ‘market’price of an activity that does not
enter the oﬃcial market, the figures economists work with in this area tend
to be drawn from either drug users’ self-completion surveys or are derived
indirectly from other data. While we will not dwell on the problems involved
in collecting and interpreting data in this field, the figures reported in this
section ought to be ‘handled with care’ (see Reuter and Greenfield, 2001).
The oﬃcial source of data for global illicit drug use is the United Nations
Oﬃce for Drugs and Crime (UNODC). Table 1.1 shows that over the 2003
to 2004 period, the total number of illicit drug users in the world was estimated to be 200 million people, equivalent to 3.2 per cent of the global
population or 5.0 per cent of the population aged 15 to 64. The most commonly consumed substances were marijuana (around 160 million people),
followed by amphetamine-type substances (26 million people using
amphetamines and 8 million people using ecstasy). The numbers using
opiates and cocaine were of a similar magnitude, with 14 million people
using cocaine and 16 million people using opiates, of whom more than 10
million were reportedly using heroin.
Figures 1.1 and 1.2 use data from the European Monitoring Centre for
Drugs and Drug Addiction (EMCDDA), the National Household Survey
of Drug Abuse (NHSDA) of Australia and the National Survey on Drug
Use and Health (NSDUH) of the USA. They show that marijuana is the
most commonly used illicit drug in these countries. However, the figures
also suggest that ‘lifetime experience’ of marijuana use is considerably
higher than ‘recent use’ (last 12 months prevalence). The clear inference
from this is that consumption of the substance is generally occasional or
even discontinued after a certain period of time.
Figure 1.1 reveals that lifetime experience of marijuana use ranges
from around 7 per cent (Portugal) to around 30–40 per cent (Australia,

2.5
4.0

3.2

5.0

0.6

0.4

26.2

Amphetamines

0.2

0.1

7.9

Ecstasy

0.3

0.2

13.7

Cocaine

0.4

0.3

15.9

Opiates

0.2

10.6

Of which
heroin:

Source:

UNODC (2005).

Note: Annual prevalence is a measure of the number/percentage of people who have consumed an illicit drug at least once in the 12-month period
preceding the assessment.

160.9

Marijuana

200

Illicit drugs
of which:

Global illicit drug use (annual prevalence) estimates (2003–04)

Global
(million people)
In % of global
population
In % of global
population age
15–64

Table 1.1

)

Lifetime prevalence of illicit drug use in USA, Australia and Europe

EMCDDA (2004) and NHSDA, Australia (2003) and NSDUH, USA (2003).

Figure 1.1

Source:

Au
s
t
ra
lia
(
2
00
1)
D
en
m
ar
k
(
2
00
En
0)
g
la
nd
(2 & W
0
a
0
2/ les
03
)

% reporting use

2)
00
(2
d

e
nc
Fr
a

(2
0

00
G

r
m
an
y
(2
00
0)
Ita
ly
(
2
00
1)
N
et
h
(2 erl
0 a
0
0/ nds
01
)

Marijuana
Cocaine
Amphetamines
Ecstasy

1)
00
l(
2

9)
99
(1
ay
w
or

)
01
(2
0
ai
Sp

ga
tu
Po
r

it
e
d
S
( ta
2
00 tes
3)
U

m
en

)

1
00
(2

0)
2)
0)
es
al )
00
00
00
W 2/03
(2
(2
(2
y
e
& 0
nd
nc
an
la
nd (20
a
m
n
r
a
i
F
F
gl
er
G
En

(2
ly

Ita

Figure 1.2

Marijuana
Cocaine
Amphetamines
Ecstasy

)
)
)
s
s
te
99
01
01
nd )
ta 0 3 )
la /01
19
20
20
r
S
(
(
(
l
d 20
he 0
in
ay
ga
te (
et 00
w
pa
ni
tu
N (2
r
S
or
U
N
Po

Last year prevalence of illicit drug use in USA, Australia and Europe

EMCDDA (2004) and NHSDA, Australia (2003) and NSDUH, USA (2003).

l
ra

st
Au

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

Source:

% reporting use

Economics uncut

Denmark, England and Wales and the USA). Rates of lifetime experience of amphetamine use range from around 0.5 per cent (Portugal) to
12 per cent (England and Wales) and from around 1 per cent (France and
Denmark) to 6.5 per cent (England and Wales) for ecstasy use. Lifetime
experience of cocaine use ranges from around 0.5 per cent (Portugal) to
15 per cent (USA).
Figure 1.2 indicates that recent use of marijuana ranges from around
3 per cent (Finland and Portugal) to 11–13 per cent (England and Wales,
USA and Australia). Recent use of amphetamines, cocaine or ecstasy is
in general reported by less than 1 per cent of individuals, although rates
of recent use of these drugs are somewhat higher in England and Wales
and Spain. The EMCDDA (2004) reports that recent use of cocaine in
particular among young people has risen to some extent in Denmark,
Germany, Spain, the Netherlands and England and Wales. This may be
due to the fact that between 1997 and 2002, the average retail price of
cocaine was stable or fell in all European Union countries (EMCDDA,
2004) or, alternatively, it may be a consequence of cocaine’s perceived
safety relative to ecstasy, which has received considerable negative media
attention (Hammersley et al., 2001). The EMCDDA (2004) notes that
the increase in ecstasy use that occurred during the 1990s in a number of
the European countries featured in Figures 1.1 and 1.2 appears to have
stabilized.
2 Theories of addiction
It should come as no surprise that an analysis of illicit drug use involves the
issue of addiction. In Chapter 2, it is argued that illicit drugs, including
heroin, are not as addictive as commonly portrayed. This section does not
enter into this debate. Rather, we examine theories of addiction, which help
explain how an individual may cross the threshold from casual drug use to
consumption that may be deemed addictive.
Non-economists may define addiction as a behaviour over which an individual has impaired control with harmful consequences. West (2001), a psychologist, classifies theories of addiction into five categories. The first group
involves theories that, ‘attempt to provide broad insights into the conceptualization of addiction’. Thus, we may couch addiction in terms of biological, social or psychological factors, or a combination of these. For example,
Betz et al. (2000), in terms of a biological theory, examine whether a
common biochemical mechanism may underlie addictions.
The second group of theories attempts to, ‘explain why particular stimuli
have a high propensity to becoming a focus for addiction’. The positive and
reinforcing properties of addictive drugs have, unsurprisingly, been investigated by numerous scientists. Robinson and Berridge (2001), for instance,

Economics of drug addiction

suggest that these properties are enhanced rather than lessened by repeated
exposure.
The third group of theories examines, ‘why particular individuals are
more susceptible to addiction than others’. Some persons may be especially
responsive to a particular stimulus, whether biochemical, psychological
or social. Cheng et al. (2000) and Cunningham et al. (1992) investigate
whether this susceptibility may be genetic.
The fourth group of theories addresses, ‘the environmental and social conditions which make addiction more or less likely’. An individual may find
himself or herself in a particular situation that triggers the need for the eﬀects
of a stimulus, or in which those eﬀects take on a greater significance. In such
cases, addiction may well result. Kenkel et al. (2001) examine the role of economic factors in the initiation and progression of drug use in this context.
Finally, the fifth group of theories of addiction, ‘focuses on recovery
and relapse . . . some are broad perspectives, others focus on eﬀects of
withdrawal from particular stimuli such as drugs; still others focus on
individual factors and others seek to model environmental influences’.
An important contribution in this field is Prochaska and DiClemente’s
(1983) Transtheoretical Model, which identifies stages of change and other
factors that predict treatment outcomes. Prior to the model’s development,
treatment for substance abuse was believed to benefit only people who were
motivated to enter treatment of their own free will. However, Prochaska
and DiClemente address the situation where the individual is in the process
of change that allows for a more eﬀective treatment developed for the individual rather than a ‘one-size-fits-all approach’.
Economists seek to conceptualize addiction within the established
framework of consumer choice. A good (or activity) is considered addictive if an increase in the stock of past consumption results in an increase in
current consumption, ceteris paribus (Becker et al., 1994). Our principal
interest is in rationalizing the observed behaviour of an addicted individual. Rationalizing means to propose a rule that relates an individual’s
objectives, preferences and constraints to his or her behaviour. A rule is illuminating if it is simple and applies broadly, that is, it predicts behaviour correctly for a large group of individuals. For instance, if we meet a fellow
economist in the library basement ‘chasing the dragon’ (that is, inhaling the
fumes from heroin as it is heated on a piece of silver foil), we will only be
surprised for a few moments, if at all. We may think of the unique immediate satisfaction provided by the drug (objective and preferences), the fact
that the colleague recently received promotion (less income constrained),
and that the chance of bumping into a colleague in the library basement
has declined greatly since the arrival of online archives (reduction in the
‘eﬀective’ price of taking drugs).

Economics uncut

Our focus for the remainder of this section is to rationalize harmful addiction, that is, a behavioural pattern defined by increasing consumption of a
good or activity that has, for example, detrimental future eﬀects upon one’s
health. However, it should be noted that theories proposed by behavioural
researchers can explain not only harmful addiction, but also beneficial
addiction within the same framework. The repetitive reading of one’s
favourite novel could be perceived as a beneficial addiction if a deeper
understanding of the story results in even greater appreciation of its quality.
The initial step of all behavioural theories of addiction is the assumption
that the amount of addictive good used or consumed in the past has lasting
eﬀects of some sort. Higher past consumption may result in a modification
of the consumer’s preferences, that is, the extent to which he or she ‘likes’
to consume the addictive good. Continuing the earlier example, this notion
suggests that the fellow economist ingests more heroin when we next meet
him or her since past consumption of the drug has increased their tolerance
for the drug. In other words, they require larger amounts of the drug to
derive the same degree of satisfaction as experienced in the past.
2.1 Relative theory of addiction
Using the tools of modern microeconomics, Stigler and Becker (1977) were
among the first economists to rationalize the behaviour of increasing consumption of a good or activity. Their ‘relative theory of addiction’ stands
in the economic tradition of the common preference approach, that is, they
attempt to explain diﬀerences in behaviour across individuals by diﬀerences
in economic constraints, as opposed to the individual’s objectives (or preferences). Therefore, rather than assuming a relationship between previous
amounts of consumption of a drug and preferences to explain addiction,
Stigler and Becker present a theory that posits a feedback between consumption and its eﬀective price as perceived by the individual.
Stigler and Becker assume that individuals have a preference for the commodity ‘euphoria’. This specific state of mind is a good that the individual
produces (and consumes) using other commodities such as heroin and time
as inputs. Since the commodity euphoria is ‘home-produced’, that is, by
construction not purchased in a market, its price is not readily available.
However, the implicit (or full, eﬀective or shadow) price of one unit of
euphoria to the economizing individual can be derived from information
on the cost of the inputs used in the production of euphoria and on the
technology adopted to produce euphoria.
As the degree of euphoria obtained by the individual depends, for
example, on the amount of heroin used, the price for one unit of euphoria
also depends on the price of this input. It follows that the price of euphoria
will therefore increase if the price of heroin increases and, similarly,

Economics of drug addiction

a reduction in the risk of getting caught taking heroin will lead to a decrease
in the shadow price of consuming euphoria.
The production of euphoria also depends on individual-specific ability
to experience the sensation. Stigler and Becker (and later Becker and
Murphy, 1988, in the rational addiction model) refer to this ability as
‘euphoric capital’ or, more generally, ‘consumption capital’. The critical
step in the Stigler and Becker model is the introduction of a feedback
between past euphoria use and the euphoric capital stock at present: the
euphoric capital depends on previous amounts of euphoria experienced
(‘euphoria exposure’). In the case of harmful addiction, greater past exposure to euphoria reduces the stock of euphoric capital today, implying that
the implicit price of euphoria increases with past euphoria exposure.
Hence, the cost of producing euphoria in the future increases if current
consumption of euphoria exposure increases. In other words, it is assumed
that as states of euphoria are more frequent and longer-lasting, the individual’s ability to become euphoric is impaired.
The relative addiction theory predicts that under certain conditions, the
demand for the inputs in the production process increases while the euphoric
capital and amount of euphoria decreases. Therefore, the model can explain
an individual’s path of rising heroin demand while the consumption of
euphoria is falling and the price of heroin remains constant. The condition
for this harmful addictive behaviour to occur is a suﬃciently inelastic
(shadow price-insensitive) demand for euphoria, that is, there are no other
goods that serve as good substitutes for the commodity euphoria.
While it is plausible that euphoria is essential to the individual, one
may wonder whether there are substitutes for heroin-induced euphoria.
Stigler and Becker implicitly assume that heroin is a necessary input in the
production of euphoria, but addiction can arise even if there is a substitute for the addictive substance. Based on the model by Rachlin (1997),
we may assume that social activity (for example, going out with friends)
can produce the same kind of euphoria as heroin. Social activity can
also be considered a good that exhibits positive addiction as postulated in the Stigler and Becker model, that is, the more social activity the
individual was exposed to in the past, the lower the current price of social
activity.
Suppose initially the shadow price of social activity is lower than the
price of heroin so that only social activity is used to produce euphoria. If
the price of heroin becomes cheaper relative to social activity then more
heroin relative to social activity would be used to generate euphoria. As the
individual’s social involvement fades, the price of future social activity
increases relative to the price of heroin. In turn, social activity would be
further substituted by heroin use and as the exposure to euphoria increases,

Economics uncut

the shadow price of euphoria increases, stimulating demand for the inputs
in the production of euphoria. Therefore, the relative theory of addiction
in this framework supports a path of increasing consumption of heroin and
decreasing social activity while the shadow price of euphoria and social
activity increases.
2.2 Melioration theory of addiction (‘primrose path’)
An important step toward a theory that explicitly formalizes the transition from ‘normal’ levels of use of the addictive good to excessive ones
results from the application of the melioration approach (Herrnstein and
Vaughan, 1980). Melioration posits that choices are the result of common
behavioural rules applied to specific circumstances. The behavioural rules
work well for normal consumption goods and activities but ‘trap’ behaviour under irregular circumstances such as the addictive goods case.
Herrnstein and Prelec (1992) build on the melioration concept to explain
why individuals may deliberately choose a path of increasing consumption
of a harmful good or activity. To illustrate this mechanism, we refer to
Figure 1.3. Point A shows the instantaneous payoﬀ to an alternative activity (‘going out with friends’) if the individual has never experimented with
heroin. Point B shows the payoﬀ to using heroin if this is the individual’s
first experience of the drug. Given the higher payoﬀ associated with point

Instantaneous
payoff

payoff to drug given history of drug use
payoff to alternative activity given
history of drug use

B
C

weighted average payoff to all actions

D
State dependence (proportion of times using
drug or engaging in alternative activity)

Figure 1.3

Drug user on ‘primrose path’ to addiction

Economics of drug addiction

B, it is rational for the individual to choose to use heroin for the first time
and experiment with mixing the frequency of drug use and the alternative
activity.
Importantly, melioration theory considers that the payoﬀs over time to
these actions are path-dependent (that is, they depend upon the individual’s
history of heroin use). The dotted line illustrates the payoﬀ to the alternative activity (going out with friends) given the individual’s recent history of
drug use.1 It is drawn as a downward-sloping line to reflect the assumption
that the use of heroin erodes the pleasures to be had from engaging in the
alternative activity. The dashed line, meanwhile, illustrates the payoﬀ to
using heroin. Note that the individual initially enjoys increasing returns
from the drug but the associated euphoria soon begins to decline rapidly.
Finally, the solid line shows the weighted average payoﬀ to all actions
undertaken by the individual through time.
Point C on the solid line represents the highest possible average payoﬀ
for the individual. This economic solution occurs at the point where the
marginal benefit of heroin use equals its marginal cost. However, melioration theory suggests that point D is the equilibrium solution because
beyond point C, the instantaneous payoﬀ to heroin use is still greater than
the payoﬀ to the alternative activity. In other words, the individual has an
incentive to increase his or her frequency of heroin use until the payoﬀs
from both activities are equalized (point D). Yet, at this point the individual is worse oﬀ than had he or she not started using heroin. The average
payoﬀ from both activities at point D is lower than the payoﬀ associated
with point A. The individual may be said to have ventured down a ‘primrose path’ of addiction, initially believing there was little danger of losing
control but eventually becoming ‘trapped’. Note that while each single decision made by the individual along this path was rational, the sequence of
decisions was certainly not rational.
The Herrnstein–Prelec approach is criticized for assuming that the individual is unable to anticipate distant consequences of his or her activities.
How increasing consumption of a harmful good is possible among individuals who are fully aware of the future consequences of their current
choice is addressed by Becker and Murphy’s (1988) rational addiction
model, to which we now turn.
2.3 Theory of rational addiction
The rational addiction model is the most influential economic theory of
addiction to date. It builds on the model by Stigler and Becker (1977) discussed earlier, but assumes that instead of consuming heroin-induced
euphoria, the representative individual derives utility in each period
directly from consuming the heroin. Like the Stigler and Becker model,

Economics uncut

consumption of the addictive good has lasting eﬀects through the addictive
consumption capital stock.
The main achievement of the theory of rational addiction, however, is to
formally derive aspects of dynamic (‘over time’) consumption behaviour
with respect to an addictive good. As such, it is an extension of the standard life-cycle model of consumption and saving. In this framework, an
individual takes into account the consequences of his or her decisions on
future outcomes. In other words, an individual is seen to be ‘more rational’
than previous models that assume myopic (short-sighted) behaviour. The
obvious challenge in the presence of forward-looking consumers is to find
illuminating conditions that guarantee addictive behaviour, that is, increasing consumption of an addictive good over the life cycle. As it turns out,
the way in which past consumption aﬀects how the individual values
current consumption is crucial for addiction to occur.
The theory of rational addiction outlined below which is based on Becker
et al. (1994), assumes that the individual with an infinite life maximizes his
or her present value, V, of lifetime utility from consumption as follows:
V

t1U(Yt,Ct, St, et)

(1.1)

subject to a lifetime wealth constraint,
A0

t1(Yt PtCt)

(1.2)

where 0 1 indicates the subjective time discount factor of the individual,2 A is the present value of wealth, Pt is the price of one unit of the addictive good, and U(Yt,Ct,St,et) denotes the individual’s period utility gain
from current consumption of the addictive good, Ct, and all other goods,
Yt. St is the addictive stock that is inversely related to period utility, and et
represents the individual’s preferences and endowments. The stock of addictive consumption capital depends on the previous period’s consumption,
Ct1, and the non-depreciated previous consumption capital stock, St1,
St (1)St 1 Ct 1

(1.3)

where 0 1 measures how quickly the tolerance level, St1, decays.
In this model, past consumption of the harmful good plays two roles. On
the one hand, higher past consumption levels of the addictive good may
increase the stock of the harmful drug, St, which directly reduces the utility
of the individual. Thus, an individual with a larger addictive stock would
need to consume relatively more of the addictive good in the current period
to obtain a given level of utility, ceteris paribus. This is the tolerance eﬀect.

Economics of drug addiction

On the other hand, a higher level of past consumption and the addictive
stock may also aﬀect the way the individual values the consumption of the
drug. This is the reinforcement eﬀect, where higher past consumption of
the drug raises the marginal utility (MU ) of its current consumption. Thus,
an individual who consumed more drugs in the past period is positively
motivated to consume relatively more drugs in the current period.
Becker and Murphy show that for addictive consumption behaviour (for
example, an increasing path of drug use) to be observed it is necessary that
an increase in the amount of consumption of the addictive good in the previous period boosts the marginal utility of currently consuming that good
more than it raises the expected discounted stream of future marginal disutilities associated with a greater stock of addictive capital. In other words,
the Becker–Murphy model requires a strong element of habit formation
or ‘reinforcement’ in the consumption of the addictive good, that is,
MU/ Ct1 must be large.
To illustrate these mechanisms, compare the addicted consumer of
heroin to the ‘social’ consumer of heroin. The latter has accumulated an
insignificant amount of heroin during past consumption sessions and consumes in small doses, that is, his or her level of tolerance is small. The
heroin addict may be more miserable than the social consumer due to the
‘side eﬀects’ of having a large stock of addictive capital (high tolerance),
but the immediate relative boost from the consumption of additional units
of heroin is much larger (strong reinforcement) for the heroin addict than
for the non-addict.
Becker and Murphy discuss a situation where past consumption of the
addictive good aﬀects the earning power of the individual. Higher tolerance, that is, a greater St in equation (1.2) is assumed to depress the individual’s wage rate (not incorporated in the model shown). By making
income dependent on the consumption level as well, the eﬀective (or
shadow) price for the addictive good increases far beyond the simple cost
of obtaining the drug.
Given the reinforcing nature of previous period’s consumption on
current consumption in the Becker–Murphy model, it is not surprising that
those who discount future consumption heavily, that is, those who are
impatient, are particularly likely to become addicted. However, the model
also predicts addiction under certain conditions even if the individual cares
a lot about the future. Individuals who are forward-looking and who have
a tendency to experience a strong reinforcement eﬀect of past consumption
on present consumption can find themselves on the addiction path. This
might explain why some young individuals are so willing to consume large
amounts of addictive goods. Following a path of rising consumption and
the accompanying build up of the stock of addictive consumption capital

Economics uncut

may be rational, even though the utility gain over not consuming the drug
at young ages is relatively small, since the anticipated greater future gains
from a stronger reinforcement eﬀect along the addictive consumption path
outweigh the costs.
As we have shown, Becker and Murphy’s theory of rational addiction
provides a series of insights into the dynamics of addictive behaviour. In
summary, there are three main factors according to this theory that determine whether or not an individual experiences addiction:
1.

How much he or she values future consumption relative to current consumption. The more heavily the future is discounted, the more likely
that a consumption path is followed that leads to addiction.
How much the addictive capital stock depreciates over time.
Individuals with higher depreciation rates are less likely to become
addicted.
The full (or shadow or eﬀective) price of the addictive good, that is, the
price that reflects the monetary cost of the drug, preferences and
the opportunity costs of consumption such as lost life cycle wages. The
higher the eﬀective price of the drug, the less likely that the individual
will become addicted.

Finally, the theory of rational addiction also predicts that the long-run
eﬀect of a price change diﬀers from the short-run eﬀect. Becker and
Murphy show that the price elasticity of demand is smaller in absolute
terms in the short run than in the long run. This prediction is the result of
the positive intertemporal feedback (reinforcement) eﬀect of past consumption of addictive goods. Suppose that the price of drugs increases in
the current period. This will decrease current consumption, which in turn
reduces the marginal utility of consumption next period and thus also
reduces future consumption. By symmetry, an anticipated increase in the
price of the addictive good in the future should reduce consumption in the
present because the consumer is forward-looking.
2.4 Hyperbolic discounting and other extensions to rational addiction
The standard sequential optimal choice model in economics such as
the Becker–Murphy model has been criticized for assuming that an individual makes time-consistent choices. Ainslie (1975), a psychologist, and
Thaler (1981), an economist, were among the first to systematically study
the nature of simple choice behaviour over time and using experimental
data. The former observed behavioural patterns of pigeons that were
time-inconsistent, while the latter found evidence for time-inconsistent
decision-making among humans.

Economics of drug addiction

Becker and Murphy impose time-consistent preferences by assuming
that an individual discounts the value of future rewards according to an
exponential function. Formally, future utilities U(Xt) are discounted by a
weight t, which is an exponentially declining function of t (meaning that
the value of future utilities will decline by a constant proportion every
period). An exponential function assumes constant discount rates and presumes that an individual’s relative evaluation of two rewards will depend
only on the amount of delay between receiving them.
However, one way of ‘rationalizing’ time-inconsistent behaviour is to
introduce time preferences so that the individual discounts the value of
future rewards according to a hyperbolic function (for example, see
Laibson, 1997) where discount rates are declining.3 In this case, delayed
rewards are weighted by t, where represents the preference for immediate reward and expresses the preference for reward delayed t periods,
relative to a delay of t1 periods.
Hyperbolic functions have the potential to exhibit preference reversals.
For instance, in Figure 1.4 the problem for an individual is to get through

Summed
reward
value

Preference
reversal

Time (delay = Tn – t )

Source: Adapted from Loewenstein and Elster (1992).

Figure 1.4

Hyperbolic discounting and drug addiction

Economics uncut

the day without using heroin. The dotted line represents the present value
of scoring heroin, while the solid line is the present value of going out with
his girlfriend. Point T1 is the possibility of scoring heroin (good) and T2 is
the possibility of meeting his girlfriend (even better). Almost all day, the
individual prefers the idea of meeting his girlfriend, but because of the
nature of hyperbolic functions, it is possible that as the opportunity to score
heroin comes closer and closer, he will begin to value it more. Indeed, at T3,
a preference reversal takes place, illustrating the idea that drug addicts tend
to have a preference for ‘small-early’ rewards over ‘later-larger’ rewards. An
initial preference for a ‘later-larger’ reward over a ‘small-early’ reward
reverses as the latter approaches, ‘with an advance choice of virtue changing to an immediate choice of vice’ (Read and Van Leeuwen, 1998). This
explains the surrender to temptation that is often characteristic of an
addict’s behaviour.
Building on the work by Laibson (1997), Gruber and Köszegi (2001)
develop a version of the Becker–Murphy model, which incorporates timeinconsistent individuals. As with rational addiction, individuals are
assumed to be forward-looking. Moreover, rationality in the Gruber
and Köszegi model implies that individuals are aware of the fact that
they have time-inconsistent preferences. The model makes the same predictions as the rational addiction model in terms of how consumers
respond to price changes. The important diﬀerence between the two
models, however, is that in the Becker–Murphy model interference with
the individual’s decisions, such as government policy aimed at lowering
illicit drug use, would hurt the consumer. Individuals are time-consistent
and those who became addicted may be miserable but they would be,
according to Becker and Murphy, ‘even more unhappy if they were prevented from consuming the addictive goods’. The equally rational but
time-inconsistent addicts in the Gruber–Köszegi model, on the other
hand, incur additional costs that they would like to avoid but cannot due
to their own choices. Consequently, Gruber and Köszegi refer to these
costs as ‘internalities’ and conclude that their model justifies government
intervention.
A role of public policy in the form of educational programmes that
inform consumers about the risks of using harmful addictive goods
results from another group of addiction models, which also extend the
Becker–Murphy theory of rational addiction. Orphanides and Zervos
(1995) assume that consumers face uncertainty with respect to their
chances of becoming addicted. They have initial beliefs about their subjective risk of becoming addicted to a particular good, and by choosing to
experiment with the substance they update their beliefs. The model features
individuals who choose to consume the good that is potentially addictive

Economics of drug addiction

for them but do not realize how addictive (hence dangerous) it is before they
are already addicted. The uncertainty in this framework with a possibility
of learning can explain the phenomenon of addicts who regret having
chosen this path, as well as withdrawal behaviour that occurs when individuals are already far along the addictive path.
Daw Namoro (2003) more generally considers the consumption behaviour of individuals when consumption may reduce life expectancy of the
consumer. Assuming that higher consumption of the harmful good lowers
the individual’s chances of surviving to the next period, he shows that
addiction is unlikely if the individual perceives the health risks to be
growing at an increasing rate (for small increases in consumption).
These models show the need for policy-makers to provide consumers
with more information on potentially addictive goods and their health
side-eﬀects. Policies based on Orphanides and Zervos (1995) should focus
on informing individuals about the general dangers and personal risks
of becoming addicted to a good. Daw Namoro (2003) suggests that aggressive advertisement towards drug users who perceive no or only a weak
relationship between consumption of harmful substances and their life
expectancy could reduce addiction.4
3 Price elasticity of demand for illicit drugs: theory
The theoretical discussion in the previous section shows how economists
have contributed to the debate about the determinants of addictive behaviour. While all models point to the importance of price as a determinant of
consumption behaviour, the meaning of price varies greatly between them.
Moreover, the discussion so far has abstracted from any diﬀerences that
may exist between actual addictive substances and the markets they are sold
in. To understand the empirical evidence on the sensitivity of demand for
illicit drugs to price changes presented in the next section, it is therefore
helpful to elaborate on the measurement and identification issues specific
to prices and markets for illicit drugs.
3.1 Eﬀective prices
Typically, any assumption we make concerning price elasticity is likely to
be based upon the notion of a simple ‘money’ price. However, according
to Stigler and Becker’s relative theory of addiction discussed earlier and
work by Moore (1973, 1990), illicit drug users will be concerned with a
drug’s ‘eﬀective’ or ‘shadow’ price, which includes the money price, but
also other components such as the perceived risk of both acquiring and
taking the drug, the probability of getting arrested, convicted, fined or
imprisoned, as well as the ‘search time’ involved. This ‘search time’, which
will diﬀer from buyer to buyer, from market to market and from time to

Economics uncut

time, is essentially a measure of an illicit drug’s ‘availability’. As Kleiman
(1992) states:
At any given money price, some buyers who would pay that price if a particular
illicit drug were easily available will not expend the eﬀort and incur the risk
involved in looking for it (and the possibility of searching unsuccessfully).
Therefore, the higher the search time – again, at any given money price – the
lower the quantity consumed. Search time acts as a kind of ‘second price’ that
users pay for their illicit drugs.

While search time could be measured to a certain degree of accuracy,
other non-money prices such as risk may be more diﬃcult to measure.
Nonetheless, diﬀerent categories of drug users will respond in diﬀerent
ways, which in turn depends upon the knowledge they have to determine
an eﬀective price. Indeed, this may well be the case for recreational drug
users whose greater price sensitivity reflects their limited understanding of
the market (MacDonald, 2004).
Building on the rational addiction model, Becker et al. (1991) provide a
further theoretical insight into the possibility of diﬀerent drug users having
diﬀerent price sensitivity. They suggest that individuals with greater preference for the present (young, lower-income, less educated) will attach a
smaller monetary value to health and other adverse eﬀects of illicit drug
use than individuals with a greater preference for the future (older, higherincome, more educated). Therefore, younger and lower-income individuals
will typically react more to changes in the money price of an illicit drug,
whereas older and higher-income individuals are more likely to respond to
changes in future harmful eﬀects.
Another related and consistent explanation is that how sensitive an individual is to the money price depends on what fraction of the full price is
comprised of the money price. For teenagers or young adults, who have a
low opportunity cost of time and may not be as future-oriented, the money
price makes up a large fraction of the full price. Thus, a given change in the
money price represents a relatively large change in the full price leading to
a greater consumption response.
3.2 Double-kinked demand curve
We can extend the notion of diﬀerent categories of drug users having
diﬀerent price sensitivity by reference to the concept of a double-kinked
demand curve. Blair and Vogel (1973) argue that at low prices, we find recreational users as well as addicts in the illicit drugs market. If this observation holds true and should prices increase, the resulting fall in demand
consists of two eﬀects. First, there is a ‘substitution eﬀect’ as the recreational users exit the market to seek other pleasures. Second, given that life

Economics of drug addiction

cannot be sustained by drug use alone, we find addicts restraining their consumption towards maintenance levels. Beyond a specific price, however, the
market consists solely of drug addicts who exhibit price insensitivity.
White and Luksetich (1983) consider the eﬀect of very high prices on
illicit drug demand, suggesting that a drug’s price may eventually become
so great that addicts too will have to exit the market due to a very strong
‘income eﬀect’. Perhaps, faced with such prices, addicts can only sustain
their habit by committing crime and are forced to leave the market due to
arrest or conviction. Alternatively, they may voluntarily choose to enrol
into a drug treatment programme.
Wagstaﬀ and Maynard (1988) combine the views of Blair and Vogel and
White and Luksetich to produce a double-kinked demand curve for illicit
drugs at the aggregate level. This market demand curve is depicted in
Figure 1.5 and has two elastic sections, one at high prices and one at low
prices, and an inelastic section for the middle range of prices. Should this
be a realistic representation of the illicit drugs market, then it can be used
to guide public policy. For example, we would find that price hikes, and
hence punishment, would have no eﬀect on consumption when drug users
are on the vertical (inelastic) section of the demand curve. One possible
problem, however, is that Wagstaﬀ and Maynard do not suggest a range of
prices at which the slope of the demand curve may change.

Price
of
drug

Addicts exit market due to arrest/
conviction or enrolment on drug
treatment programme

Market consists
solely of addicts
Recreational drug users
exit market and addicts
restrain consumption to
maintenance levels

Quantity of drug demanded

Figure 1.5

The double-kinked demand curve for illicit drugs

Economics uncut

4 Price elasticity of demand for illicit drugs: evidence
We now turn to the empirical work undertaken by economists in estimating price elasticities for various illicit drugs. Research in this area is still
somewhat hampered by poor data quality. Nevertheless, the studies
reviewed in this section typically make use of two sets of data. First, illicit
drug prices are derived from local purchases made by undercover drug
enforcement oﬃcers in the USA. Second, these prices are combined with
self-reported measures of drug use from either localized datasets or
national surveys to estimate the price elasticity of demand.
The estimation procedure itself tends to involve dependent variables,
which are binary indicators of whether or not a particular drug is consumed. This means that estimated price elasticities are with respect to outcomes such as past-year and past-month participation in drug consumption
rather than actual consumption. Elasticity of participation will tell us by
how much the probability of use of a drug changes given a fall in price. For
a fixed population, this elasticity will also reveal the extent of the percentage increase in the number of users following the price reduction. A number
of studies also produce estimates for the number of times a particular drug
is used. They calculate the total price elasticity of consumption frequency by
adding together the elasticity of drug participation and frequency conditional on participation.
The approaches to economic modelling of illicit drug use tend to fall into
three main categories: conventional, myopic and rational:
1.

Conventional – the conventional approach assumes that the current
consumption of an illicit drug depends only upon current factors.
Therefore, increases in current price can be expected to reduce current
consumption, while increases in past price and/or anticipated increases
in future price have no eﬀect on current consumption. This approach
does not take into account the dependence of current consumption
decisions on past behaviour that characterizes the use of many illicit
drugs.
Myopic – the myopic approach extends (1) by allowing current consumption of an illicit drug to depend upon current and past factors.
Therefore, not only will an increase in current price reduce current consumption, but increases in past price, by reducing past consumption,
also reduce current consumption. This approach predicts that the
long-run eﬀect of a permanent price change will be greater than the
short-run eﬀect, but it ignores the future implications of addictive consumption when making current consumption decisions.
Rational – in comparison to (1) and (2), the rational approach allows
current consumption of an illicit drug to depend upon current, past and

Economics of drug addiction

future factors. We noted earlier that the rational addiction approach is
the most significant theoretical contribution to the understanding of
illicit drug user behaviour, and this is largely because it provides a series
of testable propositions that have been investigated empirically in
numerous articles on alcohol consumption (Grossman et al., 1998;
Waters and Sloan, 1995), cigarette smoking (Bardsley and Olekalns,
1999), coﬀee drinking (Olekalns and Bardsley, 1996), the demand for
cinema (Cameron, 1999), and gambling (Mobilia, 1993, see Chapter 13).
In the context of illicit drug use, the rational addiction model makes
three key predictions. First, drug consumption is negatively correlated with
past, current and future prices of the drug. Second, current drug consumption is positively correlated with past and future consumption. Third,
as noted earlier, the long-run price elasticity of demand is greater than the
short-run price elasticity.
We can illustrate the diﬀerent empirical approaches found in the literature from the perspective of the rational addiction model. Assuming complete depreciation (1) and a utility function that is quadratic in its
arguments, the theoretical model of rational addiction presented above
(equations (1.1)–(1.3)) yields the following current demand function for the
illicit drug:
Ct 0Ct 1 0 Ct 1 1Pt et, t 1

(1.4)

where 0 and 0 capture the eﬀect of past and future consumption on
current consumption respectively, and 1 is the eﬀect of a contemporaneous
(short-run) price change on consumption. This demand function, under
diﬀerent assumptions about the coeﬃcients, constitutes the basis of a
number of empirical studies of addictive behaviour. The key predictions of
the rational addiction model discussed above translate into the following
expected signs of the coeﬃcients: 0, 0 0.
Myopic models of addiction, on the other hand, assume that the individual does not take the future consequences of their actions into account,
which implies complete discounting of the future, that is, 0, while
the reinforcement eﬀect of past consumption remains, that is, 0 0.
Meanwhile, conventional approaches to the demand for addictive substances abstract from any reinforcement eﬀect of consumption. Hence,
conventional studies of the price elasticity of demand assume 0 0.
It is clear that all three approaches predict a negative price eﬀect, that
is, 1 0, but in rational addiction and myopic models, prices are more
elastic in the long run than in the short run as a result of the reinforcement
mechanism.

Economics uncut

4.1 Conventional demand studies
Price data relating to marijuana is particularly unreliable and this largely
explains why there is only a small number of studies estimating price elasticity for this drug. One of the earliest is by Nisbet and Vakil (1972) who
used data collected from an anonymous postal questionnaire of students at
the University of California, Los Angeles, USA. Students were asked to
derive their individual demand functions for marijuana at current prices
and after various hypothetical price changes were made. Using this information and other actual data, the authors estimated a linear and doublelog demand function and produced price elasticities of demand for
marijuana of between 0.36 and 1.51.
Silverman and Spruill (1977) investigated the relationship between a
price index for retail heroin and monthly-recorded crimes in Detroit, USA,
from November 1970 to July 1973. The assumptions underpinning their
model are that heroin expenditure is a function of the retail price and quantity consumed and an addict’s ability to adjust consumption as prices
change is a function of tolerance build up and the availability of suitable
substitutes for heroin such as methadone. Using least squares regression on
a log-linear crime model, Silverman and Spruill found that property crime
rates in Detroit were positively related to the price of heroin and estimated
the price elasticity of demand for heroin to be 0.27. One criticism of this
model, however, is the assumption that price changes are caused solely by
shifts in supply.
Another small, localized dataset is used by Bretteville-Jensen and Sutton
(1996) to estimate the price elasticity of demand for heroin in Oslo,
Norway. Data were used on 500 individuals collected via questionnaire
from attendees at a needle exchange service in the city. The objective was to
distinguish between ‘ordinary’ drug users and ‘dealer’ users. The data
included information on income (and its sources), heroin consumption and
dealing activity, in addition to information on some of the seemingly nonmeasurable factors we noted earlier (such as attitudes toward risk and the
eﬀect of arrest on status) when discussing the notion of an ‘eﬀective price’.
Bretteville-Jensen and Sutton estimate the price elasticities for dealers
and non-dealers to be 0.20 and 1.23 respectively. This suggests two tentative conclusions. First, non-dealers are more price-sensitive than dealers.
Second, we cannot treat an individual’s decision to deal in drugs as being
independent of his or her decision to consume. In other words, if a dealer
is a heavy consumer, he or she is likely to be less price-sensitive.
Interestingly, the results of Bretteville-Jensen and Sutton’s estimated spline
model (which allows diﬀerent sections of the demand curve to have varying
elasticity) oﬀers little empirical support for the theoretical notion of a
double-kinked demand curve.

Economics of drug addiction

Saﬀer and Chaloupka (1999a) used data on over 49000 individuals
aged 12 and over from the 1988, 1990 and 1991 National Household Survey
on Drug Abuse (NHSDA), and drug prices from System to Retrieve
Information on Drug Evidence (STRIDE) to estimate the price elasticities
for cocaine and heroin in the USA. The annual participation price elasticities for cocaine and heroin were found to be 0.55 and 0.90 respectively,
while monthly participation elasticities for cocaine and heroin were 0.80
and 0.36 respectively.
The same sample was used by Saﬀer and Chaloupka (1999b) to estimate these price elasticities separately for seven diﬀerent demographic
sub-groups. The cocaine price elasticity was found to be statistically
insignificant for blacks and Asians, 1.83 for Native Americans, and
between 0.5 and 0.8 for white males, Hispanics, women and youth.
The price elasticity for heroin, meanwhile, was estimated to be 1.63 for
white males, 0.62 for females, 0.36 for youth, and almost zero for all
other groups.
Saﬀer and Chaloupka (1999a, 1999b) also estimate demand functions
for cocaine, heroin and alcohol, which include the money price of each substance, with the exception of marijuana. They report negative cross-price
elasticities of demand for these substances, indicating that they are complementary goods, that is, an increase in the price of cocaine, for instance,
leads to a decrease in the demand for heroin.
DeSimone and Farrelly (2003) examined data from 1990–97 NHSDA
and reported past-year cocaine and marijuana participation elasticities
of 0.13 and 0.25 respectively for individuals aged 18 to 39, although significantly negative price eﬀects were not reported using similar models for
persons aged 12 to 17.
Chaloupka et al. (1999) used the University of Michigan’s 1982 and 1989
Monitoring the Future (MTF) survey to estimate price elasticities for pastmonth and past-year outcomes of participation in cocaine use. The MTF
asks a sample of between 15000 and 19000 high school seniors in the USA
about their use of illicit drugs. Chaloupka et al. reported a price elasticity
of past-year participation of 0.89, an elasticity of past-month participation of 0.98, and corresponding frequency elasticities of 0.40 and 0.45
respectively. In a similar MTF sample from 1977 to 1987, DiNardo (1993)
found that past-month participation by high school seniors is insensitive to
price changes.
Pacula et al. (2001) used actual price figures for marijuana combined
with data from the 1985 to 1996 MTF on high school seniors and estimated
past-year participation elasticities of demand of 0.33 when time eﬀects
were omitted. When time was entered quadratically into their model,
however, the participation elasticity dropped to 0.06.

Economics uncut

Another dataset used by economists to estimate price elasticities is the
Drug Use Forecasting (DUF) dataset, which is maintained by the National
Institute of Justice in the USA to investigate drug use levels among booked
arrestees in urban areas. Its advantage is that the indicators of drug use are
objectively assessed from urine specimens and therefore not subject to
reporting errors that may be prevalent in survey-based data. Arrestees represent hardcore users of drugs, and therefore the elasticity estimates are
interpreted as the price responsiveness of heavy use.
Caulkins (1996) combined arrest data from DUF over the period 1987
to 1991 for 24 US cities with price data from STRIDE to indirectly estimate
the participation price elasticities for cocaine and heroin. He separated the
price elasticities into the product of two elasticities, which includes the percentage of arrestees testing positive for either drug as an intermediate variable. In doing so, Caulkins modelled a relationship with market quantity
and estimated simple equations, which included the arrests of drug users
and non-users (both related and unrelated to drug use) and a function of
expenditure on illicit drugs. For arrestees, the price elasticity estimates
were 0.39 for cocaine and 0.28 for heroin, while general participation
price elasticities were estimated to be 1.5 for heroin and 2.5 for cocaine.
4.2 Myopic demand studies
Van Ours (1995) used data on opium consumption in the Dutch East Indies
(Indonesia today) during the Dutch colonial period of the 1920s and
1930s when the Dutch government monopolized the opium market in that
region in an attempt to reduce criminality, guarantee purity and, ultimately,
curb opium use. As part of the operation of this system (or Opiumregie),
data on opium consumption and revenues were collected on an annual
basis. Van Ours collected data on the number of users by ethnic group for
22 regions over the period 1923 to 1938 and constructed series for the real
opium price and real income. His two-stage least squares estimation
produced short-run and long-run price elasticities of demand for opium
of 0.7 and 1.0 respectively.
Dave (2004), using DUF data from 1988 to 2000, confirms that cocaine
and heroin prices have a significantly negative eﬀect on the probability of
use for arrestees. He estimated a cocaine participation price elasticity
of 0.23 and a heroin participation price elasticity of 0.08 for arrestees
and reported that the long-run price elasticities are about twice this magnitude for both drugs. His study also showed that heavy users respond negatively to other components of the eﬀective price of drugs as proxied by
the probability of arrest for drug possession. Like Saﬀer and Chaloupka
(1999a, 1999b), Dave’s estimated cross-price elasticities suggest that cocaine
and heroin are complementary goods.

Economics of drug addiction

4.3 Rational demand studies
Grossman and Chaloupka (1998) applied Becker and Murphy’s 1988
rational addiction model in a study of the price elasticity of demand for
cocaine. They used data from panels constructed from the MTF survey
responses of high school seniors between 1976 and 1985, with the last
follow-up undertaken in 1989. The latter provides information about past,
current and future consumption by allowing for lags and leads of the
middle observation to coincide with past and current consumption respectively. Information on prices was taken from STRIDE.
They estimated the total cost of cocaine by geographic location over time
and, as with consumption, employed lags and leads to generate past and
future real prices of the drug. Similar measures were used for time-varying
socio-economic variables and numerous estimates were presented corresponding to whether they employed either OLS or two-stage least squares
and according to the diﬀerent measures of drug use. They found that,
irrespective of model specification, the estimated coeﬃcient of future
consumption is always positive and statistically significant, while the
coeﬃcient of past consumption is generally positive and significant. These
results provide solid support for the first two predictions of the rational
addiction model.
Grossman and Chaloupka’s discount factor was calculated as a ratio of
future consumption to the coeﬃcient of past consumption. This reduced
the robustness of the model’s results. The estimated discount rates corresponded to somewhat implausible interest rates and discount factors of
between 3 per cent and 4 per cent and 1.03 and 0.98 respectively. However,
Grossman and Chaloupka’s reported short-run and long-run elasticities of
demand for cocaine of 0.96 and 1.35 respectively were consistent with
the third prediction of the rational addiction model, namely that long-run
price elasticities are greater than short-run price elasticities.
Table 1.3 summarizes the empirical estimates of price elasticities for
illicit drugs discussed in this section. From this, we can reach the following
conclusions:
1.

Contrary to popular perception, the demand for illicit drugs appears
to behave according to the law of the downward-sloping demand
curve. Substantial increases in money prices seem to lead to substantial falls in illicit drug use. Indeed, comparing Table 1.2 to Table 1.3
reveals that illicit drug users are just as, or even more, responsive to
price changes as cigarette smokers.
The price responsiveness of illicit drug users is inversely related to age,
that is, young people are more sensitive to changes in price than older
people. One implication of this finding could be that since most illicit

Heroin
Opium

1977

1995

1996

Silverman &
Spruill
Van Ours

Caulkins

Brettville-Jensen
& Sutton
Grossman &
Chaloupka
Saﬀer &
Chaloupka

2001

2003

Pacula et al.

DeSimone &
Farrelly
Dave

2004

1999

Cocaine &
Marijuana
Cocaine &
Heroin

Marijuana

Cocaine

Cocaine &
Heroin

1999a

Chaloupka et al.

Cocaine

1998

Cocaine &
Heroin
Heroin

Marijuana

1972

Nisbet & Vakil

Drug (s)

Date

Monitoring the Future
Survey (1982 and 1999)
Monitoring the Future
Survey (1985–96)
National Household Survey
on Drug Abuse (1990–97)
Monitoring the Future
Survey (1988–2000)

Questionnaire of
UCLA students
Monthly data (1970–73)
from Detroit
Government data
(1923–38)
Drug Use Forecasting
System and STRIDE
Questionnaire of
Norwegian addicts
Monitoring the Future Survey
(1976–85) and STRIDE
National Household
Survey on Drug Abuse
(1988, 1990, 1991)

Data

Summary of price elasticity of demand for illicit drugs

Author (s)

Table 1.2

0.13 (cocaine, past-year)
0.25 (marijuana, past-year)
0.23 (cocaine)
0.08 (heroin)

0.7 (short run)
1.0 (long run)
2.5 (cocaine)
1.5 (heroin)
0.20 (dealers)
1.23 (non-dealers)
0.96 (short run)
1.35 (long run)
0.55 & 0.80 (cocaine,
past-year & past-month)
0.90 & 0.36 (heroin,
past-year & past-month)
0.89 (past-year)
0.98 (past-month)
0.33 (past-year)

0.36 (lower bound)
1.51 (upper bound)
0.27 (long run)

Elasticity estimates

Economics of drug addiction
Table 1.3

Summary of price elasticity of demand for cigarettes
Lowest estimate

Mean estimate

Highest estimate

0.14

0.387

1.12

0.245

0.6

0.1

0.598

1.0

0.25

0.374

0.47

0.579

1.21

0.608

1.44

Aggregate consumption
data (22 US studies)
Aggregate consumption
data (6 international
studies; developed
countries)
Developing countries
(8 studies)
Individual consumption
data on adults
(7 US studies)
Youth prevalence
(16 US studies)
Youth consumption
(12 US studies)
Source: Goel and Nelson (2003).

drug abuse begins in teenage years, large and sustained increases in
price may achieve substantial long-run reductions in illicit drug abuse
across the population as a whole.
Estimates of the long-run eﬀects of price on demand for illicit drugs
will be biased should we overlook their addictive nature. Taking the
latter into account indicates that long-run decreases in demand resulting from a permanent increase in the price of illicit drugs will be significantly greater than short-run decreases. On the other hand, the
addictive nature of illicit drugs implies that temporary price changes
will have little eﬀect on demand.
The demand for illicit drugs is a function of non-myopic behaviour.
This suggests that increases in the future price of substances (new
information on the future health consequences of illicit drug use, for
instance) will lead to substantial falls in current use.
More myopic individuals will be more price sensitive than will more
far-sighted individuals. As suggested by Becker et al. (1991) and noted
in (2), younger, less educated and lower-income persons will be more
responsive to permanent changes in the money price of addictive illicit
drugs than will older, more educated and higher-income individuals.
Likewise, more far-sighted individuals will be more sensitive to changes
in the perceived future consequences of illicit drug use and abuse.

Economics uncut

Increases in the ‘eﬀective price’of illicit drugs will also lead to reductions
in their current and future use. Such increases may include increases in
the expected ‘punishment costs’ of illicit drug use and abuse within the
legal framework, and new information on the short- and long-term
health consequences of illicit drug use and abuse as noted in (4).

5 Drug prohibition: the welfare economics approach
Estimating the price elasticity of illicit drug use has a potentially important
role to play in the design and execution of drug policy. First, it helps confirm
that drug purchases are driven by market forces. Second, it assists in evaluating the costs and benefits of drug prohibition. Governments expend
resources on enforcement and apprehension to curtail illicit drug use and
one may think of reductions in ‘external costs’ (such as accidents, crimes
and neglect of responsibility) caused by illicit drug use as benefits of drug
policy. In an economics framework, enforcement and apprehension work
by raising the price of illicit drugs and consequently reducing consumption.
If drug users are not very price responsive, however, and consumption does
not change by much, perhaps the costs outweigh the benefits.
From an economist’s viewpoint, the external costs of illicit drug use
create a divergence, illustrated in Figure 1.6, between the marginal private

Costs/
Benefits
MSC

MEC

MPC
MPB

Q0
Quantity of drugs

Figure 1.6

The social costs of illicit drug use

Economics of drug addiction

costs (MPC) of the individual decision-maker (that is, the drug user) and
the marginal social costs (MSC) borne by society as a whole.5 Given the
empirical results reported in the previous section, we assume that individuals have a downward-sloping demand curve for drugs. Second, we assume
a constant marginal private cost of consuming the drugs, which reflects the
‘eﬀective’ costs faced by the individual drug user (such as the risk of arrest),
which he or she will equate with the marginal private benefits of consumption (MPB) such as a feeling of euphoria. This yields an optimal consumption level of Q0. The marginal external cost curve (MEC), meanwhile,
reflects the external costs imposed by an individual’s illicit drug use on
society as a whole. In making the privately optimal choice, Q0, the individual does not take these costs into account.
To obtain the total social costs of illicit drug use, we add the private costs
to the social costs and this is represented by the marginal social cost curve
(MSC). The allocation of resources resulting from the private choice of Q0
is Pareto-ineﬃcient from society’s point of view. If we make the conventional assumption that there are no benefits to the rest of society from individual drug use, then the eﬃcient level of consumption is Q1, where the
social costs of illicit drug use are equal to the benefits. Indeed, at any point
between Q1 and Q0, the total costs of drug use are greater than the benefits,
and eﬃciency can be improved by government intervention that reduces
consumption from Q0 to Q1 (note that if the external costs are suﬃciently
large, Q1 will correspond to the origin, that is, zero consumption or absolute
prohibition).
Culyer (1973) examines whether society has grounds for concern over
illicit drug use and proposes six principal propositions upon which prohibition arguments should be based. MacDonald (2004) summarizes them as
follows:
1.
2.
3.
4.
5.
6.

One individual’s use of drugs imposes costs on others in society, either
through anti-social behaviour or acquisitive crime.
Drug users impose an additional burden on a publicly provided health
service either through treatment or rehabilitation.
Society simply finds the use of drugs undesirable.
Drug users should be protected as they do not act in their own best
interests.
An individual’s choice to consume drugs may lead to an escalation in
society of an undesired activity.
Drug users are less productive members of society.

The above propositions relate directly to the externalities and merit goods
frameworks of welfare economics. Indeed, proposition (4) is the classic

Economics uncut

paternalistic case for drug prohibition, which conflicts with the traditional
emphasis on consumer sovereignty in the first-year economics textbook.
The main diﬃculty with the proposition, however, is the assumption that the
rest of society is well-informed about the dangers of illicit drug use whereas
the individual drug user is not. One could make a case with reference to the
rational addiction model discussed earlier that this is not always true; individuals often make their choices with at least some, if not all, knowledge of
the risks attached to the consumption of the good or activity. On the other
hand, one could appeal to the extensions of the model that incorporate the
possibility of the drug user experiencing ‘learning and regret’. As we noted,
such models suggest a role for policy-makers to inform individuals about
the potentially negative consequences of illicit drug use and addiction.
6 Conclusion
This opening chapter has surveyed the contribution of economists in
understanding perhaps one of the most notorious aspects of human
deviancy: drug addiction. We have revealed the significant progress made
in developing insightful theories of addictive behaviour. Also, despite
having to rely upon rather imperfect data, economists have provided
numerous estimates of how an illicit drug user responds to price changes.
Indeed, both the theoretical and empirical advances that we have discussed
have the potential to improve drug policy-making.
However, while we have demonstrated that economists have a good
understanding of the determinants of the demand for illicit drugs, little is
known about both the theoretical and empirical nature of the supply of
these substances. For instance, although we can gauge from Chapter 4 the
probable motivations of a drug dealer, we have only limited knowledge of
what the cost structure of a typical ‘firm’ in this industry looks like
and, importantly, how it responds to changes in market conditions and
environment (see Levitt and Venkatesh, 2000). In summary, there remain
rich areas of research into illicit drugs for economists to explore.
Notes
1.
2.
3.

Recent history of drug use is measured as the proportion of total activity allocated to
drug use.
The term discount is used here to refer to any reason to care less about a future consequence.
An alternative approach is to make the discount function dependent on the stock of
addictive good. Orphanides and Zervos (1998) extend the Becker–Murphy model of
rational addiction by assuming that individuals with a larger addictive stock (‘addicts’)
discount future utility more heavily. In the extreme, some individuals who choose to
become addicts end up being completely myopic.
Gruber and Mullainathan (2002) test for time-consistent versus time-inconsistent preferences by looking at the eﬀects of cigarette taxes on a smoker’s utility. With time-consistent

Economics of drug addiction

preferences, higher future cigarette taxes would unambiguously reduce a smoker’s
current utility; however, with time-inconsistent preferences, a smoker may prefer (and
thus have higher current utility) to face higher future cigarette taxes to give him or herself
a commitment device in the future when he or she knows that he or she will have the selfcontrol problems. Hyperbolic discounting is also consistent with observing a positive
demand for commitment devices in a variety of contexts (aids for quitting smoking,
checking oneself into treatment facilities, smokers generally in favour of workplace
smoking bans and so on).
Here we are assuming that net external costs are positive. Chapter 2 suggests that they
are small and might even be negative. For example, consider individuals substituting
alcohol with the, arguably, less dangerous marijuana.

References
Ainslie, G.W. (1975), ‘Specious Reward: A Behavioral Theory of Impulsiveness and Impulse
Control’, Psychological Bulletin, 82, 463–96.
Bardsley, P. and N. Olekalns (1999), ‘Cigarette and Tobacco Consumption: Have AntiSmoking Policies Made a Diﬀerence?’, The Economic Record, 75, 225–40.
Becker, G.S. and K.M. Murphy (1988), ‘A Theory of Rational Addiction’, Journal of Political
Economy, 94 (4), 675–700.
Becker, G.S., M. Grossman and K.M. Murphy (1991), ‘Rational Addiction and the Eﬀect of
Price on Consumption’, American Economic Review, 81 (2), 237–41.
Becker, G.S., M. Grossman and K.M. Murphy (1994), ‘An Empirical Analysis of Cigarette
Addiction’, American Economic Review, 84 (3), 396–418.
Becker, G.S., M. Grossman and K.M. Murphy (2004), ‘The Economic Theory of Illegal
Goods: The Case of Drugs’, National Bureau of Economic Research Working Paper,
No. W10976.
Betz, C., D. Mihalic, M.E. Pinto and R.B. Rafta (2000), ‘Could a Common Biochemical
Mechanism Underlie Addictions?’, Journal of Clinical Pharmacy and Therapeutics, 25,
11–20.
Blair, R.D. and R.J. Vogel (1973), ‘Heroin Addiction and Urban Crime’, Public Finance
Quarterly, 1 (14), 457–67.
Bretteville-Jensen, A. and M. Sutton (1996), ‘Gender, Heroin Consumption and Economic
Behaviour’, Centre for Health Economics, Discussion Paper 147, University of York.
Cameron, S. (1999), ‘Rational Addiction and the Demand for Cinema’, Applied Economics
Letters, 6, 617–20.
Caulkins, J.P. (1996), ‘Estimating Elasticities of Demand for Cocaine and Heroin with Data
from the Drug Use Forecasting System’, School of Public Policy & Management Working
Paper Series, No. 95–13, Carnegie Mellon University.
Chaloupka, F.J., M. Grossman and J.A. Tauras (1999), ‘The Demand for Cocaine and
Marijuana by Youth’, in Frank J. Chaloupka, Michael Grossman, Warren K. Bickel and
Henry Saﬀer (eds), The Economic Analysis of Substance Use and Abuse: An Integration of
Econometric and Behavioral Economic Research, University of Chicago Press, pp. 133–55.
Cheng, L.S., G.E. Swan and D. Carmelli (2000), ‘A Genetic Analysis of Smoking Behaviour
in Family Members of Older Adult Males’, Addiction, 95, 427–35.
Culyer, A. (1973), ‘Should Social Policy Concern Itself with Drug Abuse?’, Public Finance
Quarterly, 1, 449–56.
Cunningham, C.L., D.R. Niehus, D.H. Malott and L.K. Prather (1992), ‘Genetic Diﬀerences
in the Rewarding and Activating Eﬀects of Morphine and Ethanol’, Psychopharmacology,
107, 385–93.
Dave, D. (2004), ‘Illicit Drug Use Among Arrestees and Drug Prices’, National Bureau of
Economic Research Working Paper, No. W10648.
Daw Namoro, S. (2003), ‘Risk Perception and Addictive Behavior’, unpublished manuscript.
DeSimone, J. and M.C. Farrelly (2003), ‘Price and Enforcement Eﬀects on Cocaine and
Marijuana Demand’, Economic Inquiry, 41 (1), 98–115.

Economics uncut

DiNardo, J. (1993), ‘Law Enforcement, the Price of Cocaine and Cocaine Use’, Mathematical
and Computer Modelling, 17, 53–64.
European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) (2004), ‘The State
of the Drugs Problem in the European Union and Norway’, Annual Report.
Goel, R.K. and M.A. Nelson (2003), ‘Controlling Use of Undesirable Goods: Survey of
Eﬀectiveness of Tobacco Control Policies’, unpublished manuscript.
Grossman, M. and F.J. Chaloupka (1998), ‘The Demand for Cocaine by Young Adults:
A Rational Addiction Approach’, Journal of Health Economics, 17, 427–74.
Grossman, M., F.J. Chaloupka and I. Sirtalan (1998), ‘An Empirical Analysis of Alcohol
Addiction: Results from the Monitoring the Future Panels’, Economic Inquiry, 36, 39–48.
Gruber, J. and B. Köszegi (2001), ‘Is Addiction “Rational”? Theory and Evidence’, Quarterly
Journal of Economics, 116 (4), 1261–305.
Gruber, J. and S. Mullainathan (2002), ‘Do Cigarette Taxes Make Smokers Happier?’,
National Bureau of Economic Research Working Paper No. 8872.
Hammersley, R., F. Khan and J. Ditton (2001), Ecstasy and the Rise of the Chemical
Generation, New York: Routledge.
Herrnstein, R.J. and D. Prelec (1992), ‘A Theory of Addiction’, in George Loewenstein and
Jon Elster (eds), Choice over Time, New York: Russell Sage Foundation, pp. 331–60.
Herrnstein, R.J. and W. Vaughan, Jr. (1980), ‘Melioriation and Behavioral Allocation’, in John
E.R. Staddon (ed.), Limits to Action: The Allocation of Individual Behavior, New York:
Academic Press, pp. 143–76.
Kenkel, D., A.D. Mathios and R.L. Pacula (2001), ‘Economics of Youth Drug Abuse,
Addiction and Gateway Eﬀects’, Addiction, 96, 151–64.
Kleiman, Mark A.R. (1992), Against Excess: Drug Policy for Results, New York: Basic Books.
Laibson, D.I. (1997), ‘Golden Eggs and Hyperbolic Discounting’, Quarterly Journal of
Economics, 112 (2), 443–78.
Levitt, S.D. and S.A. Venkatesh (2000), ‘An Economic Analysis of a Drug-Selling Gang’s
Finances’, Quarterly Journal of Economics, 115 (3), 755–89.
Loewenstein, E. and J. Elster (eds) (1992), Choice Over Time, New York: Russell Sage
Foundation.
Macdonald, Z. (2004), ‘What Price Drug Use? The Contribution of Economics to an
Evidence-Based Drugs Policy’, Journal of Economic Surveys, 18 (2), 113–52.
Marshall, Alfred (1920) [1890], Principles of Economics, 8th edition, London: Macmillan
and Co.
Mobilia, P. (1993), ‘Gambling as a Rational Addiction’, Journal of Gambling Studies, 9 (3),
121–52.
Moore, M.H. (1973), ‘Policies to Achieve Discrimination on the Eﬀective Price of Heroin’,
American Economic Review, 63, 270–77.
Moore, M.H. (1990), ‘Supply Reduction and Drug Law Enforcement’, in Michael H. Tonry
and James Q. Wilson (eds), Drugs and Crime, A Review of Research in Crime and Justice,
Chicago: University of Chicago Press, pp. 109–57.
National Household Survey on Drug Abuse (NHSDA) (2003), ‘Comparison of Substance
Abuse in Australia and the United States’, NHSDA.
Nisbet, C.T. and F. Vakil (1972), ‘Some Estimates of Price and Expenditure Elasticities of
Demand for Marihuana Among UCLA Students’, Review of Economics and Statistics, 54,
473–75.
Olekalns, N. and P. Bardsley (1996), ‘Rational Addiction to Caﬀeine: An Analysis of Coﬀee
Consumption’, Journal of Political Economy, 104, 1100–104.
Orphanides, A. and D. Zervos (1995), ‘Rational Addiction with Learning and Regret’, Journal
of Political Economy, 103 (4), 739–58.
Orphanides, A. and D. Zervos (1998), ‘Myopia and Addictive Behaviour’, Economic Journal,
108 (1), 75–91.
Pacula, R.L., M. Grossman, F.J. Chaloupka, P.M. O’Malley, L.D. Johnston and M.C. Farrelly
(2001), ‘Marijuana and Youth’, in Jonathon Gruber (ed.), Risky Behavior Among Youths:
An Economic Analysis, University of Chicago Press, pp. 271–326.

Economics of drug addiction

Phlips, Louis (1983), Applied Consumption Analysis, Advanced Textbooks in Economics,
Vol. 5, Revised Edition, Amsterdam: North-Holland Publishing Company.
Prochaska, J.O. and C.C. DiClemente (1983), ‘Stages and Processes of Self-change of
Smoking: Toward an Integrated Model of Change’, Journal of Consulting and Clinical
Psychology, 51 (3), 390–95.
Rachlin, H. (1997), ‘Four Teleological Theories of Addiction’, Psychonomic Bulletin and
Review, 4 (4), 462–73.
Read, D. and B. Van Leeuwen (1998), ‘Predicting Hunger: The Eﬀects of Appetite and Delay
on Choice’, Organizational Behavior and Human Decision Processes, 76 (2), 189–205.
Reuter, P. and V. Greenfield (2001), ‘Measuring Global Drug Markets: How Good are the
Numbers and Why Should We Care About Them’, World Economics, 2 (4), 159–73.
Robinson, T.E. and K.C. Berridge (2001), ‘Incentive-sensitization and Addiction’, Addiction,
96, 103–14.
Saﬀer, H. and F.J. Chaloupka (1995), ‘The Demand for Illicit Drugs’, Economic Inquiry, 37 (3),
401–11.
Saﬀer, H. and F.J. Chaloupka (1999a), ‘The Demand for Illicit Drugs’, Economic Inquiry,
37 (3), 401–11.
Saﬀer, H. and F.J. Chaloupka (1999b), ‘Demographic Diﬀerentials in the Demand for Alcohol
and Illicit Drugs’, in F.J. Chaloupka, M. Grossman, W.K. Bichel and H. Saﬀer (eds), The
Economic Analysis of Substance Use and Abuse, University of Chicago Press, pp. 187–212.
Schelling, T.C. (1978), ‘Ergonomics, or the Art of Self-Management’, American Economic
Review, 68 (2), 290–94.
Shane, F., G. Loewenstein and T. O’Donoghue (2002), ‘Time Discounting and Time
Preference: A Critical Review’, Journal of Economic Literature, 40 (2), 351–401.
Silverman, L.P. and N.L. Spruill (1977), ‘Urban Crime and the Price of Heroin’, Journal of
Urban Economics, 76, 78–90.
Stigler, G.J. and G.S. Becker (1977), ‘De Gustibus non est Disputandum’, American Economic
Review, 67, 76–90.
Thaler, R.H. (1981), ‘Some Empirical Evidence on Dynamic Inconsistency’, Economics
Letters, 8, 201–7.
United Nations Oﬃce on Drugs and Crime (UNODC) (2005), ‘World Drug Report 2005’,
UNODC.
Van Ours, J.C. (1995), ‘The Price Elasticity of Hard Drugs: The Case of Opium in the Dutch
East Indies, 1923–1938’, Journal of Political Economy, 103, 261–79.
Wagstaﬀ, A. and A. Maynard (1988), ‘Economic Aspects of the Illicit Drug Market and Drug
Enforcement Policies in the United Kingdom’, Home Oﬃce Research Study 95, London:
Home Oﬃce.
Waters, T. and F. Sloan (1995), ‘Why Do People Drink? Tests of the Rational Addiction
Model’, Applied Economics, 27, 727–36.
West, R. (2001), ‘Theories of Addiction’, Addiction, 96 (1), 3–13.
White, M.D. and W.A. Luksetich (1983), ‘Heroin: Price Elasticity and Enforcement
Strategies’, Economic Inquiry, 21, 557–64.

Economics of drug prohibition
Jeﬀrey A. Miron

According to conventional wisdom, illicit drugs are responsible for a broad
range of social and personal ills, including crime, diminished health and
reduced productivity. Popular thinking attributes these ills mainly to the
characteristics of drugs themselves. For example, standard accounts suggest
that psychopharmacological eﬀects of drugs make users commit violent
and other crimes. Similarly, standard depictions suggest that mind-altering
and addictive properties of drugs cause users to suﬀer poor health or diminished productivity.
This chapter explains that the social and personal ills typically associated
with illicit drugs have little to do with drugs themselves; instead, they result
from the economic incentives created by drug prohibition. This does not
mean drug use is benign; drug use carries signiﬁcant risks in some situations, but the magnitude of these risks is not markedly diﬀerent from that
of many legal goods.
Sections 1 and 2 of the chapter analyse the eﬀect of prohibition on the
price and quantity of drugs consumed. The standard defence of prohibition assumes this policy eliminates or substantially reduces the market for
drugs; in this case, prohibition eliminates or substantially reduces any negative eﬀects of drug use. The analysis here, however, shows that while price
and quantity plausibly diﬀer between a legal market and a prohibited
market, both a priori reasoning and existing evidence suggest this
diﬀerence is far smaller than assumed in most accounts. Thus, even if drug
use itself causes the ills typically associated with drugs, prohibition has only
a modest eﬀect in reducing these ills.
Sections 3 and 4 then show that, while reducing drug use and associated
ills only moderately, prohibition itself causes many of the ills usually attributed to illicit drugs. In particular, prohibition causes most of the violence
in illicit drug markets, and prohibition encourages drug users to commit
income-generating crimes such as theft or prostitution. The analysis further
shows that prohibition causes or exacerbates the risks associated with drug
use. At the same time, much of the evidence that drug use causes crime,
poor health or diminished productivity is distorted, misleading or grossly
exaggerated.
This chapter is not a cost–beneﬁt evaluation of prohibition; such an
evaluation must address a large number of issues that are outside the scope
44

Economics of drug prohibition

of the analysis here. The point of what follows is to indicate that the
unusual features of drug markets result mainly from the legal status of
drugs, not from the properties of drugs themselves. Rather than requiring
specialized assumptions or treatment, an understanding of illicit drug
markets follows readily from application of standard economic principles.
1

Price and quantity of drugs under prohibition: theory

1.1 Prohibition and the demand for drugs
Prohibition potentially reduces the demand for drugs via several mechanisms.1 First, it can foster a social norm that drug use is wrong, thereby
discouraging use. This eﬀect is hard to quantify since norms are diﬃcult to
measure. On the other hand, prohibition might increase demand by glamorizing drugs or creating a forbidden fruit.
Prohibition can also reduce demand by those who exhibit ‘respect for the
law’, even if such persons do not believe drug use is wrong. There is little
direct evidence on this eﬀect, but some information can be gleaned from laws
that are weakly enforced, such as speeding laws, sodomy laws, and certain
tax laws. In each case, violation of the law is common and the degree of noncompliance suggests caution in assuming that respect for the law, per se, substantially reduces the demand for drugs.
A third mechanism by which prohibition can reduce the demand for
drugs, even by those who put little weight on respect for the law or on social
norms that denigrate drug use, is by sanctioning the purchase or possession
of drugs. These sanctions can include heavy ﬁnes, long jail terms and loss
of professional licences, amongst other consequences.
The impact of these sanctions on the demand for drugs is probably
modest, however. First, the actual penalties imposed for possession are typically far below the maximums allowable under law. Second, although there
are many arrests for drug possession, there are also many drug users, and
the number of purchases or ‘possessions’ is far larger than the number of
users. Third, many of the arrests that do occur are incidental to the commission of other crimes, such as prostitution, theft, shop-lifting, or
vagrancy. Thus, persons who are otherwise law-abiding face minimal
chances of arrest for drug possession.
1.2 Prohibition and the supply of drugs
Prohibition can also aﬀect the supply of drugs via several mechanisms.
It potentially reduces supply by imposing costs that would not be borne by
legal suppliers. Black market suppliers must produce, transport, distribute
and store their goods secretly, or bribe law enforcement authorities to look
the other way. These suppliers must compensate employees for the risk of

Economics uncut

arrest, injury and incarceration, for the absence of fringe beneﬁts, as well
as for any stigma associated with working in a black market. Suppliers
cannot use the oﬃcial dispute resolution system and must therefore employ
mechanisms like violence that arguably have higher costs.
Despite this tendency to raise costs, however, prohibition does not
necessarily produce dramatic increases in the costs of supplying drugs
(Miron, 2003a). Conditional on operating in secret, black market suppliers face low marginal costs of evading tax and regulatory policies that
ordinarily add costs for legal suppliers. These policies include income
taxes, excise taxes, environmental regulation, safety and health regulation,
child labour laws, minimum wage laws and others. Since suppliers in a legal
market can always choose to evade such laws (that is, act as though the
good is prohibited), the fact that black market suppliers evade tax and regulatory costs can never cause prices to be lower in a black market than in
a legal market. But the eﬀect of prohibition in raising prices can be arbitrarily small.
Several other considerations suggest prohibition has a modest eﬀect on
the costs of supplying drugs. To begin with, prohibition can involve
reduced expenditure for enforcement of the taxation and regulation policies that would otherwise exist. It might also have a modest eﬀect on costs
because the eﬃcacy of enforcement expenditure is likely greater for taxation
and regulation than for prohibition. Taxes and regulations create complainants who monitor compliance with these policies, including employees
(labour market regulation), customers (false or misleading advertising),
consumer watchdog groups (environmental regulation) and rival ﬁrms
(most cost-increasing policies, including taxation). Prohibition creates no
such complainants; indeed, prohibition prevents Pareto-improving
exchanges. Thus, enforcement must rely on informants, sting operations
and busts, all of which require expenditure. Moreover, the complainants visà-vis tax and regulatory policies are less likely to complain about violations
of these policies in a black market, since they might thereby risk legal sanctions themselves. This means that, for a given level of enforcement, the costs
imposed by a policy are likely higher in a legal market.
A third reason prohibition’s impact on costs might be moderate is that in
legal markets, advertising comprises a substantial fraction of the price of
many goods. In a prohibited market, ﬁrms face increased costs of advertising, since such activities can reveal their identity or location. In monopolistically competitive models the net impact of this is ambiguous:
advertising can make demand more or less elastic, so decreased advertising
has an ambiguous eﬀect on price. The plausible assumption for drugs is
that advertising would enhance product diﬀerentiation, as with alcohol,
cigarettes or soft drinks. In this case, advertising makes demand less elastic,

Economics of drug prohibition

so an increased price of advertising implies more elastic demand and lower
goods prices.
A ﬁnal factor that suggests a moderate cost eﬀect of prohibition is its
impact on market power. One view is that prohibition facilitates evasion of
anti-trust laws, thereby increasing market power, or lowers the marginal
costs of extreme punishments (violence), thereby enhancing the potential
for collusive agreements. At the same time, certain enforcement activities
appear to enhance competition. For example, eﬀorts to prevent Peruvian
coca paste (an intermediate product between coca leaf and cocaine) from
being transported to Colombia for processing caused Colombian processors to develop coca-growing capabilities in Colombia.
Alternatively, the arrest and incarceration of a dominant supplier can
encourage price wars among the remaining suppliers as they compete for
the arrested supplier’s market share. More formally, prohibition can make
prices noisy, which inhibits collusion in some classes of models (for
example, Green and Porter, 1984), or makes it proﬁtable for suppliers to
incur the ﬁxed costs of new distribution networks, which then compete with
pre-existing arrangements. Thus, the net eﬀect of prohibition on market
power is ambiguous and probably depends on both the level and kind of
enforcement.
In summary, this section suggests that prohibition has raised costs relative to what would occur in a legal market, but by far less than asserted in
many accounts. Miron (2003a) estimates that black market cocaine is currently 2–4 times, and heroin 6–19 times its price in a legalized market. In
contrast, prior research had suggested that cocaine is 10–40 times, and
heroin hundreds of times, its legal price. MacCoun and Reuter (1997) note
that the price of marijuana in the Netherlands, which has de facto legalized
marijuana, is little diﬀerent from the price in the USA.
2 Price and quantity of drugs under prohibition: evidence
Theoretical reasoning therefore implies that prohibition reduces drug consumption. This reasoning also suggests, however, that the reduction in drug
consumption is potentially modest. Evidence on this issue is incomplete due
to lack of good data, but existing research suggests some broad conclusions.
First, the amount of drug use that occurs under prohibition suggests by
itself that prohibition’s impact on drug consumption is moderate. As
shown in Figure 1.1 of Chapter 1, for example, roughly 40 per cent of the
US population admits to having tried marijuana, and this probably understates the true frequency because some people lie or forget about their past
drug use. Moreover, many of those who do not try marijuana presumably
do so for reasons other than prohibition; many persons do not drink or
smoke even though these activities are legal.

Economics uncut

A second piece of evidence comes from experience in the USA with consumption of opiates before 1914, the year the Harrison Narcotics Act criminalized opiates, cocaine and other drugs.2 Available data show that
per capita consumption increased on average until the mid-1890s and then
decreased steadily until 1914 (Miron, 2004). These ﬂuctuations prior to
prohibition indicate that many factors aﬀect opiate consumption, not just
its legal status. In particular, public health concerns raised by the medical
profession appear to have caused the substantial decline in opiate use
between the mid-1890s and 1914 (Musto, 1973).
More recent experience also fails to suggest that prohibition reduces
drug consumption. Over the past 30 years in the USA, enforcement of drug
prohibition has expanded dramatically, as we shall discuss further in
Chapter 3. For example, the real, per capita budget of the Drug Enforcement Administration (DEA) has increased by a factor of more than three,
and the drug arrest rate has increased by a factor of roughly two (see Figure
2.1).
Over the same period, however, the trends in drug production and consumption have been essentially ﬂat, and the real, purity-adjusted prices of
both cocaine and heroin in the USA have more than halved (see Figure 2.2).
Measures of drug consumption in the USA that include all use show a substantial decline during the 1980s followed by a partial rebound during the
1990s, to roughly 80 per cent of the 1975 level. Measures of heavy consumption, however, show if anything an increase over this period (see
Basov et al., 2001).
This evidence does not prove that increased enforcement had no eﬀect;
prices might otherwise have fallen further and drug consumption
increased. But absent an explanation for why this should have occurred,
this evidence constitutes a puzzle for the view that enforcement reduces
consumption and increases price. DiNardo (1993) ﬁnds no evidence that
enforcement, as measured by cocaine seizures, raised cocaine prices or
reduced cocaine use among high school seniors during the 1980s.
Similarly, Yuan and Caulkins (1998) ﬁnd that a greater number of drug
seizures is associated with lower black market prices of cocaine and
heroin.
A further approach to estimating the impact of prohibition is to combine
estimates of the eﬀect of prohibition on demand and on costs with an estimate of the elasticity of demand. This approach is problematic for several
reasons. First, it requires estimating the eﬀect of prohibition on the demand
curve itself; such estimates do not exist. Second, this approach requires estimating the eﬀect of prohibition on costs; such estimates exist but are subject
to substantial uncertainty. Third, this approach requires estimating the elasticity of the demand for drugs, which, as discussed in Chapter 1, is a diﬃcult

49
0

Real DEA budget and total drug arrests per 1000 population in USA (1975–2003)

Drug Enforcement Administration and Bureau of Justice Statistics.

Figure 2.1

Source:

Real DEA budget (1999 US$) per 1000 population
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
Total drug arrests per 1000 population

DEA budget
Drug arrests

1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000

Median prices per pure gram of cocaine and heroin in USA (1974–2000)

Basov et al. (2001).

100

200

300

400

500

600

700

Figure 2.2

Source:

1999 US$ (cocaine)

800

1000

2000

3000

4000

5000

6000

7000

1999 US$ (heroin)

Heroin

Cocaine

Economics of drug prohibition

exercise. Nevertheless, Kuziemko and Levitt (2004) use this approach to
estimate that increases in enforcement since 1985 in the USA increased
cocaine prices and reduced cocaine consumption by 10–15 per cent over the
1985–96 period. However, their conclusion relies on regressions of cocaine
prices on drug arrest rates, commitment rates for drug arrests and other
variables. They ﬁnd a positive and signiﬁcant eﬀect of drug enforcement,
but note that interpretation of these coeﬃcients is diﬃcult. Changes in drug
demand might cause changes in price that would be positively correlated
with drug arrest rates. Moreover, since enforcement increased substantially
over this period, this conclusion is consistent with the view that prohibition
has a moderate but not a dramatic eﬀect in reducing drug consumption.
In addition to information gleaned from drug prohibition, there is also
evidence from the prohibition of alcohol in the USA during the 1920–33
period (Dills and Miron, 2004). Data on alcohol consumption are not
readily available for this period, but the cirrhosis death rate is a reasonable
proxy. Cirrhosis was indeed lower during Prohibition than during periods
before or after, but cirrhosis had reached its low, Prohibition level by the
time Prohibition began, implying factors other than Prohibition caused the
low level of cirrhosis. Taking into account the possible roles of state prohibitions, pre-1920 federal anti-alcohol policies, the age composition of the
population and economic factors such as alcohol taxes or income, Dills and
Miron (2004) estimate that national prohibition reduced cirrhosis by about
10–20 per cent.
A diﬀerent kind of evidence comes from the experience of states or countries that have decriminalized marijuana. Decriminalization reduces the
penalties for possession from jail terms and heavy ﬁnes to small ﬁnes only.
The evidence provided by decriminalization is potentially weak; changes in
the law sometimes ratify ex post what has already taken place, and changes
in the law can reﬂect incidental factors rather than real changes in enforcement (see Pacula et al., 2003 for evidence consistent with these concerns).
Nevertheless, existing evidence provides little indication that decriminalization leads to increased marijuana use.
A ﬁnal piece of evidence comes from comparing drug use rates between
the USA and other rich countries such as those in Western Europe, Japan
or Australia. These countries have prohibition laws similar to those in the
USA, especially for harder drugs such as cocaine and heroin, but, as discussed in Chapter 3, they enforce prohibition to a substantially lesser
degree. Thus, if prohibition enforcement reduces drug use, these countries
should have higher use rates than the USA. As shown in Figures 1.1 and
1.2 of Chapter 1, however, the use rate is little diﬀerent on average in the
USA compared with other rich countries; indeed, many of these countries
have lower drug consumption. This fact is only suggestive, since it does not

Economics uncut

control for other determinants of drug use, but it fails to indicate that
prohibition reduces use.
In summary, both theory and evidence suggest that prohibition has only
a modest impact on the size of the illicit drug market. This means that, even
if the ills associated with illicit drugs result from drug use, prohibition
reduces them only mildly. It also means that a substantial drug market
remains, although this market operates underground. This latter fact has
important implications for understanding the various social and personal
problems typically attributed to drugs. The following two sections of this
chapter address these implications.
3

Drugs, prohibition and crime

3.1 The impact of prohibition on crime
Prohibition promotes violent crime for several reasons.3 Most importantly,
participants in illicit markets cannot resolve disputes via standard,
non-violent mechanisms. Drug suppliers cannot use the legal system to
adjudicate commercial disputes such as non-payment of debts. Drug
traﬃckers risk legal penalties themselves if they report their employees for
misuse of ‘company’ funds or property. Purchasers of illicit drugs cannot
sue drug sellers for product liability, nor can drug sellers use the courts to
enforce payment. Along a diﬀerent line, rival traﬃckers cannot compete via
advertising and are more likely to wage violent turf battles. Thus, participants in an illicit market for drugs resolve disagreements with violence,
although it should be noted that the extent to which prohibition promotes
violence depends on the degree of enforcement (see Miron, 1999, 2001b).
Prohibition may also encourage violence by increasing the proﬁtability
of drug traﬃcking. One model of what occurs under prohibition is that
suppliers enter the illicit drug market until the total return from this activity equals the total return from legal activity, taking into account the risks
of incarceration, injury or death and any stigma/glamour associated with
working in a black market. Assuming everyone has the same willingness
to accept these features of black market activity, prohibition does not
imply excess proﬁts in the prohibited sector. If there is heterogeneity in the
willingness to work in the black market, however, those willing to do so
select into this sector, earn ‘rents’ to this characteristic and are better oﬀ
under prohibition. Such persons have more to protect under prohibition
and therefore have an additional reason to use violence.
Despite these tendencies to increase violence, prohibition might reduce
violence on net if prohibition reduces drug use and drug use causes violence.
Existing evidence, however, suggests the net eﬀect is to increase violence. A
ﬁrst piece of evidence is the behaviour of the homicide rate over the past

Economics of drug prohibition

century in the USA (Friedman, 1991; Miron, 1999). As illustrated in Figure
2.3, the murder rate rose rapidly after 1910, when many states adopted drug
and alcohol prohibition laws. The rate also rose through World War I, when
alcohol and drugs were ﬁrst prohibited nationally, and it continued to rise
during the 1920s as eﬀorts to enforce alcohol prohibition increased. The
rate then fell dramatically after alcohol prohibition’s repeal in 1934 and
(except for wartime) remained at modest levels for several decades. In the
late 1960s, the rate increased dramatically again and stayed at historically
high levels through the 1970s and 1980s, coinciding with a drastic increase
in drug law enforcement. More formally, Miron (1999) shows that enforcement of drug and alcohol prohibition is strongly correlated with the homicide rate even after controlling for a broad range of other determinants.
Comparisons across cities and countries also indicate a positive relation
between prohibition enforcement and violence. Brumm and Cloninger
(1995) ﬁnd in a sample of cities in the USA that higher drug arrest rates
are associated with higher homicide rates, while Miron (2001b) and
Fajnzylber et al. (1998, 1999) ﬁnd a positive relation across countries
between prohibition enforcement and homicide. These results must be
interpreted with caution because they document correlations that might
not be causal, but they are nevertheless consistent with the view that
prohibition causes violence.
Micro-evidence suggests even more clearly that prohibition increases
violence. Goldstein et al. (1989, 1997) ﬁnd in a sample of New York City
precincts during 1988 that almost three-quarters of ‘drug-related’ homicides were due to disputes over drug territory, drug debts and other drug
trade-related issues, rather than to the psychopharmacological eﬀects of
drugs. In a substantial fraction of the cases where such eﬀects might have
been involved, the perpetrator had also consumed alcohol.
Finally, a body of evidence documents that violence is commonly used
in a range of prohibited industries, independent of the characteristics of the
good. Violence committed by pimps or johns against prostitutes is a welldocumented feature of prostitution markets, in which prostitutes cannot
easily report violence without risking legal sanctions themselves (see, for
example, Lowman and Fraser, 1996 and also Chapter 7 of this volume).
Similarly, violence was an important feature of the gambling industry
during its early years in the USA, when entry was limited; the incidence has
decreased as legal gambling has mushroomed. More broadly, violence is
commonly used to resolve disputes in countries like Russia where the
oﬃcial dispute resolution system is ineﬀective.
In addition to promoting the use of violence in drug markets, prohibition promotes income-generating crime such as theft or prostitution.
Prohibition raises drug prices compared with a legal market, as discussed

0.0

2.0

4.0

6.0

8.0

10.0

Murder rate per 100 000 population in USA (1900–2001)

National Center for Health Statistics, Vital Statistics.

2001 ﬁgure includes terrorism attacks in USA on 11 September.

Figure 2.3

Source:

Note:

Murder rate per 100 000 population

12.0

1900
1903
1906
1909
1912
1915
1918
1921
1924
1927
1930
1933
1936
1939
1942
1945
1948
1951
1954
1957
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999

Economics of drug prohibition

above, which means users require a higher real income to ﬁnance a given
quantity of consumption. In many cases, such users would engage in criminal activity even if drugs sold for legal prices (many persons engage in
income-generating crime to ﬁnance purchases of alcohol, TVs and the
like), but prohibition appears to increase the amount of such crime.
Greenberg and Adler (1974), for example, document that some individuals
already inclined to commit crimes also happen to use drugs. Brown and
Silverman (1974, 1980) and, as discussed in Chapter 1, Silverman and
Spruill (1977) provide evidence that higher drug prices are associated with
higher rates of income-generating crime.
A diﬀerent kind of crime encouraged by prohibition is corruption of
police, prosecutors, judges and politicians. This exists because lawsuits,
lobbying and campaign contributions do not exist in a prohibited industry.
In addition, drug traﬃckers have high proﬁts to protect and thus added
incentive to bribe or threaten those who might impede these proﬁts.
Evidence on the magnitude of corruption is diﬃcult to obtain, but anecdotal evidence for such an eﬀect is abundant (see, for example, US General
Accounting Oﬃce, 1998).
The eﬀects of prohibition in generating crime and corruption are especially prevalent in drug-producing countries (see, for example, Atkins, 1998;
Melo, 1998 and Toro, 1998, concerning Bolivia, Columbia and Mexico
respectively). Beyond the standard violence- and corruption-producing
eﬀects discussed above, prohibition promotes civil unrest by providing
income to rebel groups such as the Shining Path in Peru, the FARC in
Colombia, or, until recently, the Taliban in Afghanistan. Under prohibition, these groups sell protection services to drug traﬃckers. For instance,
McClintock (1988) discusses the fact that prohibition eﬀorts in the USA
cemented a relation between Peruvian drug traﬃckers and the Shining Path
during the 1980s.4
3.2 The eﬀect of drug use on crime
The conclusion that prohibition, rather than drug use, causes most crime
associated with drugs stands in contrast to the usual claim that drug use
itself causes crime. Reviews of the literature have consistently concluded,
however, that there is little evidence of a causal role for drugs, and to the
extent that such evidence exists, it is stronger for alcohol (see, for example,
Duke and Gross, 1993, pp. 37–42, 53–4, 64–6, 73–4 and US Department of
Justice, 1992, p. 5). Fagan (1993) concludes: ‘there is little evidence that
alcohol or drugs directly cause violence’ and that, ‘several reviewers have
concluded that alcohol is the substance most likely to lead to psychopharmacological violence’ although, ‘there is some evidence that cocaine,
barbiturates, amphetamines, phencyclidine (PCP), and steroids also have

Economics uncut

psychopharmacological properties that can motivate violence’. He also
notes that, ‘the most consistent and predictable relationship between substances and violence is a result of traﬃcking in illicit drugs’.
The evidence usually oﬀered as showing an eﬀect of drug use on crime
does not stand up to careful scrutiny. This evidence consists of statistics that
document a high frequency of drug use among arrestees. For example, the
National Institute of Justice (2000) documents that in the 34 cities monitored in 1999, at least 50 per cent of adult male arrestees tested positive for
one or more illicit drugs. These statistics document that many criminals use
drugs, but they do not demonstrate that drug use causes criminal behaviour.
First, the correlation between drug use and crime is contaminated by the
fact that many arrests are for drug possession; the sense in which this shows
drug use ‘causes’ crime is purely tautological. Second, these data provide
no evidence that drug users were under the inﬂuence of drugs when they
committed their crimes or any evidence that the inﬂuence was to make the
user criminogenic. Bennet and Wright (1984) report the results of interviews with burglary oﬀenders. Many report that they had consumed
alcohol before committing their oﬀence, and about a third said they committed their oﬀence under the inﬂuence of alcohol. Most saw no causal
relation, however. For example, these oﬀenders suggested that they sometimes planned their oﬀences while in drinking situations or that they just
drank a lot generally. Given that a substantial percentage of criminals had
consumed alcohol before commission of their crimes, at a minimum the
standard evidence does not indicate which substance is implicated.5
In addition, the fact that many criminals are also drug users shows
merely that drug use is correlated with criminal behaviour. The methodology used in these analyses would also demonstrate that consumption of
fast food or wearing blue jeans causes criminal behaviour. If people from
particular socio-economic groups engage in a range of illicit behaviour,
then one will ﬁnd this correlation independent of any eﬀect of drug use on
criminal behaviour. Thus, this kind of evidence is uninformative. Stated
diﬀerently, the set of arrestees is not a random sample of the population.
Data on the behaviour of arrestees do not indicate how many people
consumed drugs without engaging in criminal behaviour (other than drug
possession itself) and thus say nothing about the tendency of drug use to
cause such behaviour.
4

Drugs, prohibition and the welfare of drug users

4.1 The impact of prohibition on drug quality
In addition to causing violence and other crime, prohibition has substantial eﬀects on the welfare of drug users. As with crime, the usual view is that

Economics of drug prohibition

drug use itself is what damages the health or reduces the productivity of
users. As with crime, however, prohibition plays a far more important role.
This does not mean drug use is benign. But to the extent that drugs entail
greater risk than a host of legal goods, the reason is prohibition rather than
the properties of drugs themselves.6
The most obvious way in which prohibition may reduce the welfare of
drug users is by subjecting them to legal penalties for purchasing or possessing drugs. Under prohibition, users risk arrest, ﬁnes, incarceration, loss of
professional licences, public embarrassment and more, as the result of their
drug use. These sanctions can be far worse than any eﬀect of drug use per se.
A second way prohibition may harm drug users is by reducing quality
control in the market for drugs. As we noted earlier, purchasers of a prohibited good cannot easily sue a seller of defective goods, while sellers of a
quality product cannot easily attract business by advertising. Moreover,
government agencies in the USA that monitor quality, such as the Food
and Drug Administration or Federal Trade Commission, cannot easily
monitor suppliers in an illicit market. While there are some instances of
accidental overdoses and poisonings in a legal market and prohibition
probably reduces the total amount of drug use, the evidence presented in
this chapter suggests that prohibition’s impact on drug consumption is
modest and so accidental overdoses and poisonings are likely to be higher
per unit of drug consumption under prohibition.
A number of examples illustrate this point. During Prohibition in the
USA, deaths due to alcoholism rose relative to other proxies for alcohol consumption, presumably because consumption of adulterated alcohol
increased (Miron and Zwiebel, 1991). Indeed, federal regulation required
manufacturers of industrial alcohol to adulterate their product, knowing
that much would be diverted to illicit consumption (Merz, 1932). In one case,
an adulterant used by bootleggers to disguise alcohol as medicine turned out
to cause permanent paralysis, victimizing thousands (Morgan, 1982).
Similarly, the chemical paraquat, which the US government encouraged
Mexico to spray on marijuana ﬁelds, caused sickness in many consumers
(Duke and Gross, 1993).
Prohibition also creates an incentive for drugs to be marketed and consumed in the purest form possible. Suppliers prefer concentrated forms of
drugs because these are easier to conceal from authorities. Consumers can
dilute potent forms if they wish, but they typically lack reliable information
on the purity of the drugs they purchase. This exacerbates the tendency
toward accidental overdoses.
Another way that prohibition harms drug users is by raising drug prices.
The direct eﬀect is to require additional expenditure to ﬁnance a given
quantity of drug consumption, which implies fewer resources for food,

Economics uncut

clothing, shelter or health care. Beyond this direct eﬀect, however, the
elevated prices caused by prohibition encourage users to employ risky consumption methods, like injection, that give the biggest bang for the buck.
This harms users because of drug paraphernalia laws and prescription laws
that restrict purchase of clean needles and syringes, thereby encouraging
the sharing of needles contaminated with HIV and other blood-borne diseases (National Research Council, 1995).
A further way that prohibition harms drug users (and others) is by discouraging the medical use of drugs. Marijuana cannot be legally prescribed
in most countries, despite evidence that it provides relief from nausea, pain
and muscle spasms and alleviates symptoms of glaucoma, multiple sclerosis and migraine headaches, amongst other ailments (Grinspoon and
Bakalar, 1993). Most opiates can be legally prescribed, but under prohibition doctors worry about legal or regulatory penalties and thus potentially
underprescribe (The New York Times, 1994; Sullum, 1997).
Prohibition also harms drug users to the extent that it fosters drug-testing
of employees and others. Drug-testing would exist even if drugs were legal,
as argued in Chapter 3, especially for safety-sensitive or other highskill occupations. Much drug-testing, however, results from governmentmandated testing programmes that have emerged as a means of enforcing
prohibition. Under these programmes, employees who test positive face
sanctions like job loss even if they have never used drugs on the job.
4.2 Drug use and addiction revisited
The arguments above explain why prohibition causes many of the adverse
consequences of drug use. In addition, the evidence suggesting that drug
use itself causes reduced health, diminished productivity or other personal
ills, is biased, exaggerated or misleading. One standard claim is that drug
use is addictive.7 There is some truth in this view, but drugs are far less
addictive than commonly portrayed. One possible measure of addictiveness is the degree to which use continues after initial use. High continued
use rates do not necessarily suggest addiction in the classic sense; if those
who enjoy a particular good consume it frequently, the continued use rate
will be high even if there is no addictiveness. But addiction does imply a
high continued use rate, and this has been used frequently as a measure of
addiction.
The data presented in Chapter 1 indicate that across all categories of
drugs, at most a third of those who have ever used a drug say they have used
that drug in the past year. This does not mean drugs are never addictive, but
it fails to suggest a high degree of addictiveness. The fact that continued use
rates for marijuana, which is not regarded as physically addictive, are
similar to those for crack cocaine, which is regarded as highly addictive,

Economics of drug prohibition

also challenges the more extreme claims about addictiveness of drugs.8
Relatedly, the continued use rates for alcohol and tobacco are even higher
than those for illicit drugs, and casual observation suggests continued use
rates for other legal goods are even higher (for example, chocolate, caﬀeine).
A diﬀerent measure of addictiveness is the degree to which consumers
use a particular substance casually or irregularly. The stereotypical depictions of addiction suggest that experimentation progresses inevitably into
regular use and that irregular use occurs rarely. In fact, a sizeable percentage of heroin users consume only occasionally (Zinberg, 1979), and measurable withdrawal symptoms from opioids rarely occur until after several
weeks of regular administration (Jaﬀee, 1991). Bennett (1986) discusses a
literature search and interviews conducted during the 1980s with opiate
users in the UK. His analysis suggests a modest degree of addictiveness.
Few users are pushed into use by sellers; most start with friends or acquaintances. For most it takes months or years to become addicted. Many
addicts voluntarily abstain for weeks, months or years, and many mature
out of addiction. The most common reason given for using opiates is that
users ‘like them’ and feel their lives are better on opioids. Many use opiates
to self-medicate or to follow friends; few exhibited signs of compulsion.
Further evidence that addiction is less important than typical portrayals
comes from the experience of returning Vietnam veterans. Robins et al.
(1974) interviewed samples of veterans 8–12 months after their return from
Vietnam. They found that most addicted veterans gave up their narcotic use
voluntarily before departure or after a short, forced treatment period at
departure. In subsequent work, Robins et al. (1980) document that
although most veterans had access to cheap heroin in Vietnam, only about
35 per cent tried it and only about 19 per cent became addicted. They also
conclude that heroin use does not consistently lead to daily use and addiction, that addiction frequently ceases without treatment, that maintaining
recovery from heroin addiction does not require abstinence, and that the
reason for high levels of social disability among heroin users is likely attributable to characteristics of the users rather than to heroin per se.
As with addiction, the negative health consequences of drug use are
often overstated. All drugs carry some health risk, but the degree to which
illicit drugs are physically detrimental is far less than generally portrayed,
provided they are consumed under safe circumstances. The Merck Manual,
a standard reference book on diagnosis and treatment of diseases
(Merck & Co., Inc., 1992), states that, ‘people who have developed tolerance [to heroin] may show few signs of drug use and function normally in
their usual activities . . .. Many but not all complications of heroin addiction are related to unsanitary administration of the drug’. It also writes
that, ‘there is still little evidence of biologic damage [from marijuana] even

Economics uncut

among relatively heavy users’. Concerning cocaine, the manual does not
mention eﬀects of long-term use but emphasizes that all eﬀects, including
those that promote aggression, are short-lived. Many of the health risks
discussed for all drugs result from overdoses or adulterated doses, not moderate or even heavy levels of use.
Similar evaluations appear elsewhere. These evaluations in no way imply
that drug consumption is without risk; indeed, many highlight a range of
adverse reactions and eﬀects that can occur as a result of drug use. But they
consistently indicate that such adverse reactions are not the norm and that
even heavy, regular use can and does occur with modest or minimal negative consequences (see, in particular, Grinspoon and Bakalar, 1979 or
Morgan and Zimmer, 1997 on cocaine; Grinspoon and Bakalar, 1993, Hall
et al., 1994, British Medical Association, 1997 or O’Brien, 1996 on marijuana; Trebach, 1982 or Zinberg, 1979 on heroin).
A critical problem with standard depictions of the health consequences
of drug use is that they rely on data sources that are biased towards those
who suﬀer the worst consequences from drug use. One example is data from
clients of drug treatment programmes. Such persons are presumably those
who have had the worst experiences with drugs, but this does not demonstrate what the average or typical experience might be (just as data from
alcohol treatment facilities would give a misleading impression about the
typical consequences of alcohol consumption). Even a robust correlation
between drug use and poor health does not indicate the eﬀect of drug use
on health, since drug use is often associated with a range of behaviours and
characteristics that might be detrimental to health.
Similarly, many discussions fail to distinguish the eﬀects of casual use
from those of heavy use, or the eﬀects that can occur for some users as
opposed to those that usually occur for most users. Moderate doses of overthe-counter medications such as ibuprofen relieve pain eﬀectively for
millions of persons each day, but a large dose causes severe damage to
the liver. Peanuts are an excellent food source that enhances health for
most consumers, yet they can be deadly for persons who are allergic.
Standard antibiotics such as penicillin provide relief from moderate to lifethreatening diseases for most users but cause illness or death in persons who
are allergic. These examples merely illustrate the general principle that
many goods are beneﬁcial in moderation even though potentially harmful
in excess; likewise, many goods are beneﬁcial to most persons even though
they can be harmful for some persons.
4.3 Drug use and labour market outcomes
A diﬀerent alleged harm of drug use is reduced income or likelihood of
employment. According to this view, drug use inhibits concentration,

Economics of drug prohibition

coordination, motivation and other factors that contribute to successful
job market experience. For example, widely-cited estimates of the costs of
drug abuse suggest that drug use reduces productivity in the USA by tens
of billions of dollars each year (Harwood et al., 1998). (See Miron, 2003b,
for a critique of estimates of the costs of drug abuse.)
In fact, the existing evidence on the relation between drug use and labour
market outcomes faces severe methodological diﬃculties. These studies
utilize data on the wages and drug use of individuals, along with auxiliary
information on demographic and economic characteristics such as age,
education, occupation or experience. The basic analysis regresses wages on
drug use plus measured individual characteristics. The problem is that drug
use and wages are both plausibly correlated with unmeasured characteristics such as optimism, motivation, sociability, creativity or risk aversion.
Thus, a ﬁnding that drug use and wages are correlated can reﬂect the
inﬂuence of these omitted individual characteristics. Stated diﬀerently, the
standard methodology is not experimental; it does not come from observing the behaviour of individuals who have been randomly assigned to
diﬀerent levels of drug use. Equivalently, there are no obvious instruments
for drug use. Some analyses of this type employ two equation models, with
the second having drug use as the dependent variable, but the identifying
assumptions are implausible and are based mainly on non-linearities.
The results of existing studies of drug use and wages are therefore
diﬃcult to interpret at best. In addition, the results do not support the conclusion that drug use is associated with lower wages; a persistent puzzle in
this literature is that the estimated relation between drug use and wages is
often positive (see, for example, Kaestner, 1991, 1994a; Register and
Williams, 1992; Gill and Michaels, 1992; Kenkel and Ribar, 1994, who
obtain similar results for alcohol consumption).
The estimated relation between drug use and certain other labour market
outcomes, such as employment status or hours worked, is more frequently
negative, but even for these outcomes there are many ‘paradoxical’ results
in the literature (see, for example, Zarkin et al., 1992; Kaestner, 1994b;
Zarkin et al., 1998). Given the methodological problems that confront
examination of this question, the right conclusion is that there is no evidence in either direction.
4.4 Drug use and accidents
An additional harm often attributed to drugs is automobile accidents
caused by drivers under the inﬂuence. Such incidents are indeed a negative
consequence of drug use, but several controlled studies indicate that marijuana has a smaller detrimental eﬀect on driving performance than alcohol.
Crancer et al. (1969) examine the simulated driving performance of subjects

Economics uncut

under the inﬂuence of marijuana and alcohol in comparison with a control
group that consumed neither substance. They ﬁnd that the marijuana consumers had more speedometer errors than under control conditions, but no
signiﬁcant diﬀerence in accelerator, brake, signal, steering or total errors.
The alcohol group accumulated signiﬁcantly more accelerator brake,
signal, speedometer and total errors but no signiﬁcant diﬀerence in steering
errors (see also Smiley, 1986 or US Department of Transportation, 1993,
1999 for similar conclusions). The fact that marijuana appears to impair
driving less than alcohol is especially relevant due to evidence (presented in
Chapter 3) that marijuana and alcohol are substitutes in consumption.
A similar harm often linked to drug use is workplace accidents. The evidence that drugs cause a substantial number of accidents is at best mixed,
and alcohol is implicated at least as often as illicit drugs (National Research
Council, 1994). Kaestner and Grossman (1998) examine the relation
between drug use and workplace accidents using the 1988 and 1992
National Longitudinal Survey of Youth (NLSY). For males, they ﬁnd weak
evidence that drug users have a higher accident rate than non-users; for
females, they ﬁnd no evidence of such an eﬀect. Moreover, their evidence
shows only whether drug users have elevated accident rates; it does not
show in any way that drug use caused these accidents.
4.5 Drug use and the unborn fetus
A further problem potentially caused by drug use is harm to the unborn
fetus. A vast literature examines the correlation between illicit drug use and
adverse pregnancy outcomes such as low birth weight or shorter gestation
length. There is no doubt that excessive use of certain substances can have
negative eﬀects on pregnancy outcomes, but careful evaluation of the evidence suggests the adverse eﬀects of drugs have been drastically overstated.
Many pregnant women consume legal substances like alcohol or tobacco
that might also contribute to poor postnatal outcomes, but some studies do
not control for legal substance use in assessing the eﬀects of illicit drugs.
Relatedly, many studies control imperfectly for other factors that aﬀect
pregnancy outcomes, including access to pre-natal care, income, presence
of a father, nutrition, exercise or sleep habits. Further, as with studies of
drug use and crime or drug use and productivity, even studies that control
for a broad range of observed factors cannot address the unobserved
diﬀerences across women that might simultaneously explain both drug use
and poor pregnancy outcomes. In any event, the studies that exist do not
consistently suggest substantial or necessarily persistent eﬀects of drug use
on children, and the negative eﬀects that do occur are not obviously
diﬀerent from those of legal goods such as alcohol or cigarettes. (For
detailed discussion of these issues, see Behnke and Eyler, 1993; Richardson,

Economics of drug prohibition

et al., 1993; US Department of Health and Human Services, 1996; Inciardi
et al., 1997; LaGasse et al., 1999).
5 Conclusion
Standard accounts of drugs and drug markets often portray them as
diﬀerent from other goods, somehow defying the usual laws of supply and
demand or requiring special economic principles or insights to become
understandable. We have conﬁrmed the ﬁndings of Chapter 1 that current
drug markets do indeed exhibit characteristics that are rare or non-existent
in other markets, but the main purpose of this chapter was to emphasize
that these characteristics result from the legal status of drugs rather than
from the characteristics of drugs themselves. Moreover, we demonstrated
that a normative analysis of prohibition must distinguish between the
eﬀects of prohibition and the eﬀects of drugs themselves and an obvious
question raised by our analysis is whether prohibition is the best policy
toward illicit drugs. By assessing a range of alternatives to prohibition,
Chapter 3 addresses this issue.
Notes
1.
2.
3.
4.

6.
7.
8.

The discussion in this section draws heavily on Miron and Zwiebel (1995) and Miron
(1998, 1999, 2001a, 2001b, 2003a, 2004).
The Harrison Act did not prohibit marijuana; this resulted from the 1937 Marijuana
Tax Act.
The discussion in this section draws heavily on Miron and Zwiebel (1995) and Miron
(1998, 1999, 2001a, 2001b, 2004).
In addition to increasing ‘drug-related’ crime, prohibition can aﬀect non-drug-related
crime. Enforcement of prohibition potentially diverts resources from deterrence of other
crime (Benson and Rasmussen, 1991; Benson et al., 1992; Rasmussen et al., 1993 and
Sollars et al., 1994). Similarly, incarceration of drug oﬀenders can crowd out incarceration of other criminals. Alternatively, incarceration of drug oﬀenders might incapacitate
persons who commit both drug and non-drug crime. On the last two eﬀects, see
Kuziemko and Levitt (2004).
Greenfeld (1998) summarizes data from the National Crime Victimization Survey, the
Uniform Crime Reports, the National Incident-Based Reporting System, Bureau of
Justice Statistics (BJS) surveys of probationers, jail and prison inmates, BJS censuses of
prisons and jails, and the Fatal Accident Reporting System on the relation between
alcohol consumption, drug use and crime of various types. The focus is mainly on
alcohol; it contains less extensive data on drug use. With respect to alcohol, this review
provides the standard type of data indicating an association between alcohol use and
crime. For example, many victims of crime perceive that the oﬀender had been drinking
at the time of the oﬀence.
The discussion in this section draws heavily on Miron and Zwiebel (1995) and Miron
(1998, 2001a, 2004).
The discussion here takes as given that addiction per se is a negative consequence, but
this view is open to debate (see Chapter 1 for a discussion of this issue).
While these data are potentially biased by under-reporting, and the degree of underreporting with respect to recent use might be greater than with respect to lifetime use,
there might be more forgetting of lifetime use than of recent use.

Economics uncut

References
Atkins, A. (1998), ‘The Economic and Political Impact of the Drug Trade and Drug Control
Policies in Bolivia’, in Elizabeth Joyce and Carlos Malamud (eds), Latin America and the
Multinational Drug Trade, Great Britain: Macmillan Press Ltd.
Basov, S., M. Jacobson and J.A. Miron (2001), ‘Prohibition and the Market for Illegal Drugs:
An Overview of Recent History’, World Economics, 2 (4), 133–58.
Behnke, M. and F.D. Eyler (1993), ‘The Consequences of Prenatal Substance Use for the
Developing Fetus, Newborn, and Young Child’, The International Journal of Addictions,
28 (13), 1341–91.
Bennett, T.H. (1986), ‘A Decision-Making Approach to Opioid Addiction’, in Derek B. Cornish
and Ronald V.G. Clarke (eds), The Reasoning Criminal, New York: Springer-Verlag,
pp. 83–102.
Bennet, T. and R. Wright (1984), ‘The Relationship Between Alcohol Use and Burglary’,
British Journal of Addiction, 79, 431–7.
Benson, B.L. and D.W. Rasmussen (1991), ‘The Relationship Between Illicit Drug
Enforcement Policy and Property Crimes’, Contemporary Policy Issues, 9, 106–15.
Benson, B.L., I. Kim, D.W. Rasmussen and T.W. Zuehlke (1992), ‘Is Property Crime Caused
by Drug Use or by Drug Enforcement Policy?’, Applied Economics, 24, 679–92.
British Medical Association (1997), Therapeutic Uses of Cannabis, the Netherlands: Harwood
Academic Publishers.
Brown, G.F. Jr. and L.P. Silverman (1974), ‘The Retail Price of Heroin: Estimation and
Applications’, Journal of the American Statistical Association, 69 (347), 595–606.
Brown, G.F. Jr. and L.P. Silverman (1980), ‘The Retail Price of Heroin: Estimation and
Applications’, in Irving Leveson (ed.), Quantitative Explorations in Drug Abuse Policy,
New York: SP Medical and Scientiﬁc Books, pp. 25–53.
Brumm, H.J. and D.O. Cloninger (1995), ‘The Drug War and the Homicide Rate: A Direct
Correlation?’, Cato Journal, 14 (3), 509–17.
Chaloupka, F.J. and A. Laixuthai (1997) ‘Do Youths Substitute Alcohol and Marijuana?
Some Econometric Evidence’, Eastern Economic Journal, 23 (3), 253–76.
Crancer, A. Jr., J.M. Dille, J.C. Delay, J.E. Wallace and M.D. Haykin (1969), ‘Comparison
of the Eﬀects of Marihuana and Alcohol on Simulated Driving Performance’, Science,
164, 851–4.
Dills, A. and J.A. Miron (2004), ‘Alcohol Consumption and Alcohol Prohibition’, American
Law and Economics Review, 6 (2), 285–318.
DiNardo, J. (1993), ‘Law Enforcement, the Price of Cocaine, and Cocaine Use’, Mathematical
and Computer Modelling, 17 (2), 53–64.
Duke, S.B. and A.C. Gross (1993), America’s Longest War: Rethinking Our Tragic Crusade
Against Drugs, New York: G.P. Putnam’s Sons.
Fagan, J. (1993), ‘Interactions Among Drugs, Alcohol, and Violence’, Journal of Health
Aﬀairs, Winter, 65–79.
Fajnzylber, P., D. Lederman and N. Loazya (1998), ‘Determinants of Crime Rates in
Latin America and the World: An Empirical Assessment’, Washington, DC: World
Bank.
Fajnzylber, P., D. Lederman and N. Loazya (1999), ‘Inequality and Violent Crime’, World
Bank, manuscript.
Friedman, M. (1991), ‘The War We Are Losing’, in Melvyn B. Krauss and Edward P. Lazear
(eds), Searching for Alternatives: Drug-Control Policy in the United States, Stanford,
California: Hoover Institution Press, pp. 53–67.
Gill, A.M. and R.J. Michaels (1992), ‘Does Drug Use Lower Wages?’, Industrial and Labor
Relations Review, 45 (3), 419–34.
Goldstein, P.J., H.H. Brownstein, P.J. Ryan and P.A. Bellucci (1989), ‘Crack and Homicide in
New York City, 1988: A Conceptually Based Event Analysis’, Contemporary Drug
Problems, Winter, 651–87.
Goldstein, P.J., H.H. Brownstein, P.J. Ryan and P.A. Belluci (1997), ‘Crack and Homicide in
New York City: A Case Study in the Epidemiology of Violence’, in Craig Reinarman and

Economics of drug prohibition

Harry G. Levine (eds), Crack in America: Demon Drugs and Social Justice, Berkeley:
University of California Press, pp. 113–30.
Green, E. and R. Porter (1984), ‘Noncooperative Collusion under Imperfect Price
Information’, Econometrica, 52, 87–100.
Greenberg, S.W. and F. Adler (1974), ‘Crime and Addiction: An Empirical Analysis of the
Literature, 1920–1973’, Contemporary Drug Problems, 3, 221–69.
Greenfeld, L.A. (1998), ‘Alcohol and Crime: An Analysis of National Data on the Prevalence
of Alcohol Involvement in Crime’, Washington, DC: Oﬃce of Justice Programs, US
Department of Justice.
Grinspoon, L. and J.B. Bakalar (1979), ‘Cocaine’, in Robert L. Dupont, Avram Goldstein and
John O’Donnell (eds), Handbook on Drug Abuse, Washington, DC: NIDA.
Grinspoon, Lester and James B. Bakalar (1993), Marihuana: The Forbidden Medicine, New
Haven: Yale University Press.
Hall, W., N. Solowij and J. Lemon (1994), ‘The Health and Psychological Consequences of
Cannabis Use’, Canberra: Australian Government Publishing Service.
Harwood, H., D. Fountain and G. Livermore (1998), ‘The Economic Costs of Alcohol and
Drug Abuse in the United States, 1992’, Rockville, Maryland: US Department of Health
and Human Services.
Inciardi, J.A., H.L. Surratt and C.A. Saum (1997), Cocaine-Exposed Infants: Social, Legal,
and Public Health Issues, Thousand Oaks, CA: Sage Publications.
Jaﬀee, J.H. (1991), ‘Opiates’, in Ilana Glass (ed.), The International Handbook of Addictive
Behaviors, pp. 64–8.
Kaestner, R. (1991), ‘The Eﬀect of Illicit Drug Use on the Wages of Young Adults’, Journal
of Labor Economics, 9 (4), 381–412.
Kaestner, R. (1994a), ‘New Estimates of the Eﬀect of Marijuana and Cocaine Use on Wages’,
Industrial and Labor Relations Review, 47 (3), 454–70.
Kaestner, R. (1994b), ‘The Eﬀect of Illicit Drug Use on the Labor Supply of Young Adults’,
Journal of Human Resources, 29 (1), 126–55.
Kaestner, R. and M. Grossman (1998), ‘The Eﬀect of Drug Use on Workplace Accidents’,
Journal of Labor Economics, 5, 267–94.
Kenkel, D.S. and D.C. Ribar (1994), ‘Alcohol Consumption and Young Adults’
Socioeconomic Status’, Brookings Papers: Microeconomics, 119–75.
Kuziemko, I. and S.D. Levitt (2004), ‘An Empirical Analysis of Imprisoning Drug Oﬀenders’,
Journal of Public Economics, 88 (9–10), 2043–66.
LaGasse, L.L., R. Seifer and B.M. Lester (1999), ‘Interpreting Research on Prenatal
Substance Exposure in the Context of Multiple Confounding Factors’, Clinics in
Perinatology, 26 (1), 39–54.
Lerner, R. (1998), ‘The Drug Trade in Peru’, in Elizabeth Joyce and Carlos Malamud (eds),
Latin America and the Multinational Drug Trade, Great Britain: Macmillan Press Ltd.
Lowman, J. and L. Fraser (1996), ‘Violence Against Persons Who Prostitute: The Experience
in British Columbia’, Technical Report 1996–14e, Department of Justice Canada.
MacCoun, R.J. and P. Reuter (1997), ‘Interpreting Dutch Cannabis Policy: Reasoning by
Analogy in the Legalization Debate’, Science, 278, 47–52.
MacCoun, Robert J. and Peter Reuter (2001), Drug War Heresies: Learning from Other Vices,
Times, and Places, Cambridge: Cambridge University Press.
McClintock, C. (1988), ‘The War on Drugs: The Peruvian Case’, Journal of Interamerican
Studies and World Aﬀairs, 30 (Summer/Fall), 127–41.
Melo, J.O. (1998), ‘The Drug Trade, Politics, and the Economy: The Colombian Experience’,
in Elizabeth Joyce and Carlos Malamud (eds), Latin America and the Multinational Drug
Trade, Great Britain: Macmillan Press Ltd.
Merck & Co., Inc. (1992), The Merck Manual of Diagnosis and Therapy, 16th edition, Robert
Berkow (ed.), Rahway, NJ: Merck Research Laboratories.
Merz, Charles (1932), The Dry Decade, Garden City, NY: Doubleday, Doran and Co.
Miron, J.A. (1998), ‘Drug Prohibition’, in Peter Newman (ed.), The New Palgrave Dictionary
of Economics and the Law, London: The Macmillan Press, pp. 648–52.

Economics uncut

Miron, J.A. (1999), ‘Violence and the U.S. Prohibitions of Drugs and Alcohol’, American Law
and Economics Review, 1–2 (Fall), 78–114.
Miron, J.A. (2001a), ‘The Economics of Drug Prohibition and Drug Legalization’, Social
Research, 68 (3), 835–55.
Miron, J.A. (2001b), ‘Violence, Guns, and Drugs: A Cross-Country Analysis’, Journal of Law
and Economics, 2 (4), 615–33.
Miron, J.A. (2002a), ‘Review of Drug War Heresies: Learning from Other Vices, Times, &
Places’, by Robert J. MacCoun and Peter Reuter, The Independent Review, 7 (2), 297–300.
Miron, J.A. (2002b), ‘The Eﬀect of Marijuana Decriminalization on the Budgets of
Massachusetts Governments, With a Discussion of Decriminalization’s Eﬀect on
Marijuana Use’, Report to the Drug Policy Forum of Massachusetts, October, 2002.
Miron, J.A. (2003a), ‘Do Prohibitions Raise Prices? Evidence from the Markets for Cocaine
and Heroin’, Review of Economics and Statistics, 85 (3), 522–30.
Miron, J.A. (2003b), ‘A Critique of Estimates of the Economic Cost of Drug Abuse’, manuscript, Boston University, manuscript.
Miron, Jeﬀrey A. (2004), Drug War Crimes: The Consequences of Prohibition, Independent
Institute.
Miron, J.A. and J. Zwiebel (1991), ‘Alcohol Consumption During Prohibition’, American
Economic Review, 81, 242–7.
Miron, J.A. and J. Zwiebel (1995), ‘The Economic Case Against Drug Prohibition’, Journal of
Economic Perspectives, 9 (4), 175–92.
Morgan, J.P. (1982), ‘The Jamaica Ginger Paralysis’, Journal of the American Medical
Association, 245 (15), 1864–7.
Morgan, J.P. and L. Zimmer (1997), ‘The Social Pharmacology of Smokeable Cocaine: Not
All It’s Cracked up to Be’, in Craig Reinarman and Harry G. Levine (eds), Crack in
America: Demon Drugs and Social Justice, Berkeley: University of California Press,
pp. 131–70.
Musto, David F. (1973), The American Disease, New Haven: Yale University Press.
National Institute of Justice (2000), ‘1999 Annual Report on Drug Use Among Adult and
Juvenile Arrestees’, Washington: US Department of Justice, Oﬃce of Justice Programs.
National Research Council (1994), ‘Under the Inﬂuence? Drugs and the American Work
Force’, Jacques Normand, Richard O. Lempert and Charles P. O’Brien (eds), Washington,
DC: National Academy Press.
National Research Council (1995), ‘Preventing HIV Transmission: The Role of Sterile
Needles and Bleach’, Washington, DC: National Academy Press.
New York Times, The (1994), ‘It Pains a Nation of Stoics to Say “No” to Pain’, 3 April, p. E5.
O’Brien, C.P. (1996), ‘Drug Addiction and Drug Abuse’, in Goodman and Gilman’s The
Pharmacological Basis of Therapeutics, 9th edition, Joel G. Hardman, Lee E. Limbird,
Perry B. Molinoﬀ and Raymond W. Ruddon (eds), New York: McGraw-Hill.
Pacula, R.L., J.F. Chriqui and J. King (2003), ‘Marijuana Decriminalization: What Does
it Mean in the United States?’, National Bureau of Economic Research Working Paper
No. 9690.
Rasmussen, D.W., B.L. Benson and D.L. Sollars (1993), ‘Spatial Competition in Illicit Drug
Markets: The Consequences of Increased Drug Law Enforcement’, Review of Regional
Studies, 23 (3), 219–36.
Register, C.A. and D.R. Williams (1992), ‘Labor Market Eﬀects of Marijuana and Cocaine
Use Among Young Men’, Industrial and Labor Relations Review, 45 (3), 435–48.
Resignato, A.J. (2000), ‘Violent Crime: A Function of Drug Use or Drug Enforcement?’,
Applied Economics, 32, 681–8.
Reuter, P. (2001), ‘The Limits of Supply-Side Drug Control’, Milken Institute Review, 3 (1),
14–23.
Richardson, G.A., N.J. Day and P.J. McGauhey (1993), ‘The Impact of Prenatal Marijuana
and Cocaine Use on the Infant and Child’, Clinical Obstetrics and Gynecology, 36 (2),
302–18.
Robins, L.N., D.H. Davis and D.N. Nurco (1974), ‘How Permanent Was Vietnam Drug
Addiction?’, American Journal of Public Health, 64, 38–43.

Economics of drug prohibition

Robins, L.N., J.E. Helzer, M. Hesselbrock and E. Wish (1980), ‘Vietnam Veterans Three Years
After Vietnam: How Our Study Changed Our View of Heroin’, in Leon Brill and Charles
Winick (eds), The Yearbook of Substance Use and Abuse, vol. 2, New York: Human
Sciences.
Silverman, L.P. and N.L. Spruill (1977), ‘Urban Crime and the Price of Heroin’, Journal of
Urban Economics, 4, 80–103.
Smiley, A. (1986), ‘Marijuana: On-Road and Driving Simulator Studies’, Alcohol, Drugs, and
Driving, 2 (3–4), 121–34.
Sollars, D.L., B.L. Benson and D.W. Rasmussen (1994), ‘Drug Enforcement and the
Deterrence of Property Crime Among Local Jurisdictions’, Public Finance Quarterly, 22 (1),
22–45.
Sullum, J. (1997), ‘No Relief in Sight’, Reason, 28, 22–9.
Toro, M.C. (1998), ‘The Political Repercussions of Drug Traﬃcking in Mexico’, in Elizabeth
Joyce and Carlos Malamud (eds), Latin America and the Multinational Drug Trade,
Great Britain: Macmillan Press Ltd.
Trebach, Arnold S. (1982), The Heroin Solution, New Haven: Yale University Press.
US Department of Health and Human Services (1996), ‘Behavioral Studies of Drug-Exposed
Oﬀspring: Methodological Issues in Human and Animal Research’, Cora Lee Wetherington,
Vincent L. Smeriglio and Loretta P. Finnegan (eds), National Institute on Drug Abuse
Research Monograph 164, Rockville, MD: NIDA.
US Department of Justice (1992), ‘Drugs, Crime and the Justice System’, NCJ-133752,
Washington, DC: USGPO.
US Department of Transportation (1993), ‘Marijuana and Actual Driving Performance’,
Washington, DC.
US Department of Transportation (1999), ‘Marijuana, Alcohol, and Actual Driving
Performance’, Springﬁeld, VA: National Technical Information Service.
US General Accounting Oﬃce (1998), ‘Law Enforcement: Information on Drug-Related
Police Corruption’, Report to the Honorable Charles B. Rangel, House of Representatives,
Washington, DC: USGAO.
United National Oﬃce for Drug Control and Crime Prevention (2003), ‘Global Illicit Drug
Trends 2002’, ODCCP Studies on Drugs and Crime (www.undcp.org/ global_illicit_drug_
trends.html).
Yuan, Y. and J.P Caulkins (1998), ‘The Eﬀect of Variation in High-Level Domestic Drug
Enforcement on Variation in Drug Prices’, Socio-Economic Planning Sciences, 32 (4), 265–76.
Zarkin, G.A., M.T. French and J. Valley Rachal (1992), ‘The Relationship Between Illicit Drug
Use and Labor Supply’, Research Triangle Institute, manuscript.
Zarkin, G.A., T.A. Mroz, J.W. Bray and M.T. French (1998), ‘The Relationship Between Drug
Use and Labor Supply for Young Men’, Journal of Labor Economics, 5, 385–409.
Zinberg, N.E. (1979), ‘Non-Addictive Opiate Use’, in Robert L. Dupont, Avram Goldstein
and John O’Donnell (eds), Handbook on Drug Abuse, Washington, DC: National Institute
on Drug Abuse.

Economics of drug liberalization
Mark Thornton, Bruce L. Benson and
Simon W. Bowmaker

In Chapter 2, we drew attention to the many unintended consequences of
drug prohibition, including violent crime and corruption, and suggested
that prohibition causes or exacerbates the various risks associated with
illicit drug use. Tullock and McKenzie (1985) argue that economists have
always been anti-prohibition, dating back to the days of the ban on alcohol
in the USA:
In the early part of this century, many well-intentioned Americans objected
to the consumption of alcoholic beverages. They succeeded in getting the
Constitution amended to prohibit the sale of alcohol. By the 1930s most of them
had given up because they discovered how diﬃcult it was to enforce the law. If
they had consulted economists, I’m sure they would have been told that the
law would be very diﬃcult and expensive to enforce. With this advice they
might have decided not to undertake the program of moral elevation. The same
considerations should, of course, be taken into account now with respect to
other drugs.

There have been exceptions, however. Irving Fisher, one of the USA’s
greatest mathematical economists, was a leading proponent of alcohol prohibition. As late as 1927, Fisher claimed he could not find one economist to
speak out against prohibition at a meeting of the American Economic
Association, and in The Noble Experiment, published in 1930, Fisher clearly
remained a strong believer in the virtues of alcohol prohibition:
Summing up, it may be said that Prohibition has already accomplished incalculable good, hygienically, economically and socially. Real personal liberty, the
liberty to give and enjoy the full use of our faculties, is increased by Prohibition.
All that the wets can possibly accomplish is laxity of enforcement or nullification: in other words, enormously to increase the very disrespect for the law which
they profess to deplore. Hence the only satisfactory solution lies in fuller enforcement of the law.

More recently, Thornton (1995, 2004) reports that in relation to illicit
drugs, the majority of economists (in the USA at least) are relatively antiprohibition. He reports the findings of a survey he conducted in 1995 of
117 randomly selected professional economists based on membership of
the American Economic Association. Of those who oﬀered an opinion,
68

Economics of drug liberalization

58 per cent were in favour of drug policy in the USA being steered toward
decriminalization of drugs. Only 16 per cent favoured complete legalization, while 71 per cent of those who gave a response other than keeping
the status quo, favoured either legalization or decriminalization. Less than
2 per cent supported stronger prohibition other than longer prison sentences and increased enforcement budgets.
In a separate survey, conducted in 2004, Thornton solicited the views of
US economists who are actively engaged in drug policy research. Three
general conclusions emerged. First, most argued that the current policy of
prohibition in the USA is fairly ineﬀective, very ineﬀective or even harmful
(as suggested in Chapter 2). Second, most agreed that the current policy
stance ought to be shifted. Third, most believed that this shift ought to be
in the direction of liberalization. A source of disagreement, however,
centred on the degree of liberalization.
Therefore, the purpose of this chapter is to examine the results of an
economy shifting its drug policy away from prohibition toward some form
of legalization. We begin in Section 1 with a precise description of current
drug policies in both Europe and the USA. This is followed by an analysis
of the following policy approaches toward illicit drugs: government
monopoly (Section 2), regulation (Section 3), sin taxes (Section 4) and the
free market (Section 5). Examples are given from drug and non-drug
markets and the potential costs and benefits of each policy are evaluated in
light of theoretical analysis as well as historical and contemporary evidence. Section 6 concludes the chapter.
1 Drug policy in Europe and the USA: an overview
In recent years, Europe has continued to move away from stereotypical USstyled ‘war on drugs’ ideology. The shift has been characterized by recognition that illicit drug use is a health issue (an addiction, an illness) rather
than a crime issue. In view of this, alternative measures implemented and
currently under debate in Europe include legalization (permitting commercial sales), decriminalization (eliminating user sanctions) and harm
reduction (medical provision of heroin to addicts, for instance).
1.1 European demand-side policies
It should be noted that many European nations have had a drug policy of
de facto decriminalization for many years, but it is only recently that
governments have been amending their legislation to reduce or remove
criminal penalties for illicit drug use and possession. In 2001, for example,
Belgium, Finland, Greece, Luxembourg, Portugal and Switzerland drafted,
proposed or approved legislation for the decriminalization of minor drug
use and possession, in most cases for marijuana. In the UK in 2004, the

Economics uncut

declassification of marijuana from a Class B drug to a Class C drug
took eﬀect, representing the first major change to UK drug policy since the
introduction of the 1971 Misuse of Drugs Act. Under the new law, possession of marijuana remains illegal but will ordinarily not be an arrestable
oﬀence.
In the Netherlands, contrary to many perceptions, marijuana possession
also remains illegal, but is tolerated by the Dutch authorities. In 1976,
under the auspices of the separation of markets concept, the Dutch revised
their Opium Act to separate the ‘hard’ drug market from the ‘soft’ drug
market and prevent soft drug users from interacting with hard drugs. One
consequence of the change in laws was the emergence of ‘coﬀee shops’
throughout the nation, oﬀering marijuana products for sale. By 1997, there
were around 1200 such shops operating in the Netherlands but this number
had fallen to 782 by 2002 as Dutch authorities closed premises because of
nuisance complaints and code violations1 (Osborn, 2003).
Table 3.1 summarizes the position of a number of European countries
on the issue of decriminalization of illicit drug use. It shows that while not
every European country has either reduced or removed penalties for minor
drug oﬀences, all have taken steps to provide a variety of treatment and
harm reduction measures. For example, Sweden, seen as one of Europe’s
‘hard-liners’ on illicit drugs, oﬀers a suspension of prison sentence for
minor drugs oﬀences on condition that the oﬀender seeks treatment under
a ‘treatment contract’ (DEA, 2002).
In the past, drug substitution programmes typically consisted of treating
heroin addicts with declining doses of methadone, which was pioneered by
the Swiss. Today, several European nations, including Switzerland, now
oﬀer maintenance programmes, which, like other harm reduction measures, are intended to regulate the drug use of those who are unwilling to
seek traditional forms of treatment. Maintenance programmes usually
relate to an opiate such as morphine or heroin, but methadone is still most
common. A number of countries in Europe also oﬀer heroin prescription.
Heroin is prescribed in the UK, for instance, through general practitioners
to an estimated 500 clients (EMCDDA, 2003).
It was the rapid spread of the HIV virus among intravenous drug users
in the 1980s that induced many European nations to seek such measures
that would reduce the harmful eﬀects of illicit drug use for those who
refused treatment. A wide variety of these so-called harm reduction measures have developed throughout Europe. The most common are needleexchange programmes, pill testing and consumption rooms (sometimes
referred to as ‘shooting galleries’).
Needle-exchange programmes are used widely in France (in 87 out of
100 départements) and the French government subsidizes pill-testing

Source:

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Substitution
treatment
No
No
No
Yes
No
No
Yes
No
No
No
Yes
Yes

Heroin
prescription
Yes
Yes
Yes
Yes
No
No
Yes
No
Yes
No
Yes
Yes

Pill
testing
No
No
No
Yes
No
No
Yes
Yes
Yes
No
Yes
No

Consumption
rooms
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

Needle
exchange
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes

Decriminalization
of personal use

European approaches to reducing drug-related crime and harm (as at 2004)

The Senlis Council (2004).

Austria
Belgium
France
Germany
Greece
Italy
Netherlands
Portugal
Spain
Sweden
Switzerland
United
Kingdom

Table 3.1

All drugs
Marijuana

All drugs
All drugs
All drugs
All drugs

All drugs

All drugs
Marijuana

Decriminalization
of use

Economics uncut

Table 3.2

1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000

Quantity of illicit drug seizures in Europe (1985–2000)

Cannabis
(kg)

Ecstasy
(tablets)

LSD
(doses)

Heroin
(kg)

131 146
164 641
167 304
263 701
224 279
274 874
320 379
372 841
516 666
743 473
778 417
674 081
764 698
831 074
819 598
758 812

n.a.
n.a.
1 217
9 742
51 528
65 479
530 956
661 174
2 568 963
2 593 275
2 663 754
9 962 120
4 349 845
6 235 902
14 874 024
19 295 458

52 190
187 076
124 207
119 952
177 386
375 147
238 595
809 803
1 164 561
401 084
585 151
438 275
320 358
152 737
121 621
124 373

1 990
2 076
2 312
2 876
2 864
4 666
6 029
5 389
4 809
5 889
5 265
5 542
5 815
4 979
7 138
8 811

Cocaine Amphetamines
(kg)
(kg)
1 043
1 925
3 912
6 751
7 788
16 439
16 508
17 693
16 892
29 013
20 699
32 304
43 207
31 611
43 366
26 242

277
404
582
486
383
630
891
1 307
1 674
2 026
1 605
3 534
5 266
4 800
3 871
3 218

Notes: Countries under study are Belgium, Denmark, Germany, Greece, Spain, France,
Ireland, Italy, Luxembourg, Netherlands, Austria, Portugal, Finland, Sweden, UK, Norway.
Source: EMCDDA.

projects2 (EMCDDA, 2003). About five million syringes are distributed
or sold each year in Switzerland, about 25 per cent of them through pharmacies (EMCDDA, 2003). Drug consumption rooms are found in
39 European cities, where users can take drugs obtained elsewhere, under
supervision, in hygienic conditions, without fear of arrest (EMCDDA,
2004).
1.2 European supply-side policies
While many European nations are focusing on treating and reducing the
harm caused by illicit drug use, most are also continuing to employ substantial enforcement resources against the supply of illicit drugs. As
Table 3.2 shows, the quantity of illicit drugs in Europe seized over the 1985
to 2000 period has generally been increasing.
Stricter penalties on those organizations and individuals that traﬃc illicit
drugs in Europe are also being introduced. In countries such as Austria,
France, Greece and the UK, drug traﬃcking can result in sentences up to

Economics of drug liberalization

life imprisonment, while in Luxembourg, supplying illicit drugs to minors
carries a penalty of up to lifelong forced labour.
The UK in particular has stepped up its eﬀorts in recent years to dismantle and punish drug traﬃcking organizations. In 1995, the 1994 Drug
Traﬃcking Act was implemented, which allows the court to assume that all
current assets, including any owned by the oﬀender during the previous six
years, are the result of traﬃcking oﬀences. Unless the oﬀender can prove
otherwise, the court may seize these assets. Further, the 2003 Proceeds of
Crime Act established an Assets Recovery Agency, which grants police and
customs oﬃcers powers to seize and search for money generated from drug
traﬃcking.
The penal procedure and the drug classification system determine the
traﬃcking penalties in the UK. The 1971 Misuse of Drugs Act divides controlled substances into three classes: A, B and C. Class A (which includes
‘hard drugs’ such as cocaine and heroin) traﬃcking is punishable by up to
life imprisonment and, in 2000, the Powers of the Criminal Courts Act
established a minimum seven-year sentence for a third conviction of
Class A drug traﬃcking. Moreover, in 2001, the Criminal Justice and Police
Act empowered courts in the UK to strengthen controls on convicted
traﬃckers. For instance, a ban can be placed on all overseas travel of a convicted traﬃcker for up to four years in an attempt to reduce his or her
opportunity to re-engage in traﬃcking activities (DEA, 2002).
1.3 US demand-side policy
In Chapter 2, we noted that enforcement of drug prohibition in the USA
has increased dramatically over the past 30 years. However, the widespread
perception that the USA has a uniformly aggressive prohibitionist policy is
actually somewhat misleading. Policy implementation clearly varies over
time, across types of drugs and across enforcement jurisdictions. For
instance, consider policy toward marijuana consumption. In 2000, 46.5 per
cent of all drug arrests in the USA were for marijuana oﬀences (Federal
Bureau of Investigation, 2000), and possession charges accounted for
almost 88 per cent of these arrests. This suggests that prohibition remains
the dominant policy in the USA regarding marijuana.
Yet, there are important oﬀsetting trends in demand-side policy that
should be recognized. By some indicators at least, both the probability
and severity of punishment for marijuana possession oﬀences appear
to be declining in many states, for instance, often despite increasing
arrests. Also, consider the fact that during the 1970s 11 states ‘decriminalized’ marijuana possession and use by significantly reducing their
criminal sanctions against marijuana possession.3 While no additional
decriminalization occurred in the 1980s, Arizona decriminalized in 1996,

Economics uncut

Nevada did so in 2001 (indeed, Nevada had a referendum to legalize
marijuana in 2003, and while it failed to pass, about a third of the voters
supported it), and similar policies have received considerable political
attention in other states.4 A number of other states have sharply reduced
penalties for marijuana possession, particularly after 1989, even though
they have not fully decriminalized (Pacula et al., 2003). For instance, seven
additional states have specified first-time marijuana possession as a noncriminal oﬀence.5 Furthermore, medical uses of marijuana have been
legalized in several states during the last decade, despite strong federal
resistance.
Liberalization trends for heroin and cocaine can also be detected in many
states. A growing number of arrested drug users are being diverted out of
traditional criminal justice punishment options into programmes focusing
on treatment rather than incarceration. These treatment programmes are
still backed by threats of criminal justice system sanctions, however, unlike
voluntary treatment programmes or treatment programmes run through
public health agencies. For instance, the first ‘drug court’ was established
in Miami, Florida in 1989, and as of 8 January 2003 the model had been
adopted, with various modifications, in 1424 jurisdictions, with at least one
such programme in all 50 states, the District of Columbia, Guam, Puerto
Rico and at least two Federal Court Districts.
Drug courts are judge-supervised treatment programmes, often in lieu of
a criminal conviction and/or sentence, which require regular appearances
before the judge, mandatory participation in treatment, frequent drugtesting (typically once per week), the potential for sanctions such as short
jail terms for positive drug tests or other failures to meet the programme
requirements (for example, perhaps finding and keeping a job), and the ultimate threat of diversion into regular criminal justice sanctions if the participant is judged to be not living up to the mandates of the programme. In
other words, such coercive treatment programmes keep the criminal justice
system at the centre of drug policy even as drug abuse is implicitly or explicitly being recognized as a medical issue.
1.4 US supply-side policy
Again, let us start with marijuana policy. The federal government continues to treat marijuana as a ‘Schedule I’ substance (in the same category as
opiates, cocaine and other hard drugs), with significant statutory penalties
for use, but the major focus is on suppliers. To illustrate the supply-side
federal prohibitionist focus, and the relatively large and expanding role of
marijuana in this context, consider drug seizures as an indicator of drug
control eﬀort. Table 3.3 shows federal drug seizures by drug type over the
1990 to 2002 period.

Economics of drug liberalization

Table 3.3 Pounds of illicit drugs seized by US federal authorities by drug
type (1990–2002)
Fiscal year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002

Heroin

Cocaine

Marijuana

Hashish

1 704
3 067
2 552
3 516
2 898
2 569
3 737
3 121
3 499
2 733
6 640
4 392
6 900

235 885
246 325
303 289
244 315
309 710
234 105
253 297
252 329
266 029
284 631
248 827
239 957
225 122

483 353
499 097
783 477
772 086
1 041 445
1 308 171
1 429 786
1 488 362
1 777 434
2 282 313
2 614 746
2 674 826
2 412 365

17 062
178 211
4 048
26 080
1 625
32 020
32 096
53 051
596
1 678
23 987
433
193

Source: Federal-wide Drug Seizure System (FDSS).

By this measure at least, federal eﬀorts against marijuana have become
increasingly dominant over time, but because the federal focus is on supplyside enforcement (interdiction, pursuit of traﬃckers), not demand-side
enforcement (only a very small portion of the drug oﬀenders sentenced to
federal prisons are convicted for possession), they have not increased
marijuana-related arrests the way the states in aggregate have. Of course, it
must be recognized that marijuana is much more diﬃcult to hide than most
other drugs, so seizure eﬀorts should be more successful against marijuana.
It should also be noted that seizure eﬀorts are not just focused on drugs
themselves. Seizures of money and other assets allegedly associated with
drug market activity (for example, used in transportation or in transactions, purchased with proceeds from drug sales) are an important focus as
well.
The federal emphasis on the supply-side is also illustrated in Table 3.4,
which shows recent federal drug control spending by function. Note that
demand-side budgets for treatment and prevention (and for treatment and
prevention research) have been continually expanding, but supply-side
budgets for international policy initiatives, interdiction and domestic law
enforcement spending (money used for the criminal justice system,
research on criminal justice policy and intelligence, which certainly could
be for either demand- or supply-side enforcement, but is in fact mostly

1330.8

322.2

219.6

3667.7

1937.5

1406.5
464.0
3808

1106.9

309.6

206.5

3446.1

1836.3

1549.3
389.9
3775.5

Source:

2155.6
746.3
5002.5

2100.6

4028.4

249.9

373.5

1407.6

1997.4

19991

Oﬃce of National Drug Control Policy (ONDCP).

Notes:
1
Actual;
2
Final;
3
Enacted.

1795.2

1823.1

Drug
treatment
Drug
prevention
Prevention
research
Treatment
research
Demand-side
total
Domestic law
enforcement
Interdiction
International
Supply-side
total

19981

19971

1904.4
1619.2
5761.9

2238.3

4139.1

280.8

421.6

1445.8

1990.9

20002

1895.3
617.3
4975.4

2462.8

4443.1

326.8

489.0

1540.8

2086.5

20012

1913.7
1084.5
5792.9

2794.7

4781.0

367.4

547.8

1629.0

2236.8

20022

2147.5
1105.1
6206.7

2954.1

4812.4

382.9

611.4

1553.6

2264.6

20032

US federal drug control spending (US$ million) by function (1997–2005 fiscal years)

Function

Table 3.4

2534.1
1159.3
6876.3

3182.9

4991.1

412.4

607.2

1550.4

2421.1

20042

2662.9
1131.3
7083.4

3289.2

5079.2

423.1

615.4

1139.0

2494.3

20053

Economics of drug liberalization

focused on the supply-side) is always higher and increasing in aggregate,
except for the reduction in 2001. Furthermore, supply-side spending has
been rising at a faster overall rate than demand-side spending, indicating
that federal legislators and bureaucrats are in general agreement on
enforcement focus.
At the state level, the aggregate data regarding policies against drug
traﬃcking do not provide a clear picture of trends, for similar reasons to
those that create confusion about the overall direction of demand-side
policy changes. Arrests for traﬃcking have been falling in recent years (the
opposite of possession arrests), while convictions appear to be on an
upward trend (see Table 3.5). These two factors imply that the likelihood of
conviction given arrest is rising. Countering this is the apparent reduction
in the likelihood of imprisonment for felony traﬃcking convictions
(see Table 3.6). Similarly, traﬃcking sentences are declining, although
reduced sentences are largely oﬀset by increasing portions of sentences
served (see Tables 3.7 and 3.8 respectively).
In other words, in aggregate, courts appear to have a diﬀerent policy
agenda (reducing both the likelihood and length of prison sentences) than
legislators (building prisons so the portion of sentence served increases), as
with possession, while police appear to be increasing focus on possession
arrests. Countering this is the increasing level of seizure activity we noted
earlier. In fact, many observers of US drug policy trends suggest that the
dramatic increase in enforcement that occurred in the last half of the 1980s
is a result of changes in asset seizure laws, which mandated that policing
agencies could keep the proceeds from their seizure eﬀort (see Rasmussen
and Benson, 1994; Benson et al., 1995), rather than placing these revenues
into general fund accounts or other non-criminal justice accounts
Table 3.5

Felony drug convictions in US State Courts (1992–2002)

Year

Total drug
convictions

Possession
convictions

Traﬃcking
convictions

Marijuana
traﬃcking
convictions

1992
1994
1996
1998
2000
2002

280 232
274 245
347 774
314 626
319 700
340 330

109 426
108 815
135 270
119 443
116 300
127 530

170 806
165 430
212 504
195 183
203 400
212 810

16 376
15 931
20 618
22 975
25 300
21 340

Source: Bureau of Justice Statistics, ‘Felony Sentences in State Courts’, 1992, 1994, 1996,
1998, 2000, 2002 (published every other year).

Economics uncut

Table 3.6 Percentage of felony drug convictions in US State Courts
sentenced to prison, jail, and probation (1992–2002)
Year

1992
1994
1996
1998
2000
2002

Possession

Traﬃcking

Prison

Jail

Probation

Prison

Jail

Probation

33
34
29
36
33
34

29
32
41
29
31
28

38
34
30
35
36
38

48
48
39
45
41
42

27
23
33
26
28
26

25
29
27
29
31
32

Source: Bureau of Justice Statistics, ‘Felony Sentences in State Courts’, 1992, 1994, 1996,
1998, 2000, 2002 (published every other year).

Table 3.7 Mean and median sentences in months for felony drug
convictions sentenced to prison, jail and probation by US State
Courts (1992–2002)
Year

Possession

Traﬃcking

Prison
Jail
Probation
Prison
Jail
Probation
mean/med. mean/med. mean/med. mean/med. mean/med. mean/med.
1992
1994
1996
1998
2000
2002

55/36
50/36
41/24
35/24
34/24
33/24

4/3
4/3
5/5
4/3
5/3
5/3

45/36
37/24
37/36
36/25
33/24
32/36

72/48
66/48
55/36
54/36
52/36
51/36

8/6
7/6
7/6
6/4
7/6
8/6

51/36
40/36
45/36
40/36
39/36
39/36

Source: Bureau of Justice Statistics, ‘Felony Sentences in State Courts’, 1992, 1994, 1996,
1998, 2000, 2002 (published every other year).

(for example, education). The disposition of asset seizures clearly is a determinant of drug enforcement eﬀorts at any rate (Mast et al., 2000), and total
asset seizures have risen dramatically as drug enforcement has increased.
There is no centralized accounting of asset seizures by all policing agencies,
but examinations of available records from individual states consistently
show increasing seizures.
The apparently conflicting indicators become even more confusing when
we consider interstate diﬀerences. There are some fairly clear liberalization

Economics of drug liberalization

Table 3.8 Percentage of prison sentence served and average time served in
months, for drug convictions sentenced to prison by US State
Courts (1992–2002)
Year

1992
1994
1996
1998
2000
2002

Possession

Traﬃcking

% sentence
served

Mean time
served
(months)

% sentence
served

Mean time
served
(months)

27
34
40
40
49
40

15
17
16
14
17
14

34
32
42
41
49
45

24
21
23
22
26
24

Source: Bureau of Justice Statistics, ‘Felony Sentences in State Courts’, 1992, 1994, 1996,
1998, 2000, 2002 (published every other year).

actions being taken in some states, even on the supply-side. For instance,
some states are allowing production of marijuana for medical uses,
although the federal government continues to oppose these policies.
Further, open discussion of legalizing marijuana has actually been taking
place in a few states (for example, New Mexico where the past governor
openly advocated legalization and Nevada where a legalization referendum
garnered over 30 per cent of the votes cast). However, the open marketing
of marijuana for general consumption does not appear to be on the
horizon, even in states with a long history of liberalization, and liberalization regarding the sale of other drugs (for example, prescriptions for
heroin for addicts) is not even being considered.
2 Government monopoly
We now turn to an in-depth evaluation of the alternative policies available
in switching from prohibition to some form of legalization. These include
government monopoly, regulation, the sin tax and the free market. In turn,
each of these policies represents a spectrum of possible policy choices. For
example, the sin tax rate could be set at a low or high rate; government
regulation could be strict or lax, on the producer, consumer or both: government might regulate production, distribution, consumption or some
combination of all three. Government monopoly can be open, as in the case
of state-run liquor stores in the USA, or restricted as in the case of
methadone clinics. Many policy reformers advocate that policies be set

Economics uncut

with respect to specific drugs so that, for example, marijuana be left
completely free from government intervention while heroin be regulated by
a tightly-controlled government monopoly. (For a good discussion of the
policy options and establishing policies according to specific drugs, see
Nadelmann, 1992 and MacCoun et al., 1996.)
One policy that is often considered a replacement policy for prohibition
is to establish a government monopoly for the distribution of drugs, particularly narcotic drugs such as heroin. This approach would place the production and distribution of drugs in the hands of the state and thus
provides direct control over most aspects of the marketplace. Several US
states monopolize the distribution and sale of liquor, and a few also do so
for wine (see Benjamin and Anderson, 1996; Benson et al., 2003). Two
other examples of these privatized government monopolies are the market
for human organ transplants and state lotteries.
One consequence of this policy approach is that government can directly
control the product. It can establish rules for the production, distribution
and consumption of the product and therefore mandate the composition
of the product (for example, potency), price, quantity limits and hours of
operation. With human organ transplants the government prohibits the
sale of organs and determines who gets those available. With state lotteries
the government regulates what products can be sold, their price and the
method of sale. Some state liquor monopolies control the wholesale distribution, choosing products and setting wholesale prices, while others
monopolize retailing as well, thus determining the number, location, operating hours and practices of retail outlets, along with prices and products
to sell.
Government monopolies can also establish regulations concerning
who is allowed to purchase and consume the product. In the area of drugs,
government-run liquor stores restrict the sale of their products to adults,
although the liquor is often resold by adults to minors or obtained from
government stores by minors by theft or deception. Methadone clinics have
a monopoly on the distribution of narcotics, but they generally only
provide the drug to registered addicts. The methadone is often provided at
no charge, but the addicts are required to consume the product on the
premises in order to prevent resale.
The results of government monopoly vary depending on whether it is
contracted out or publicly run and whether it distributes its product at high
prices or gives the product away for free to pre-determined consumers.
State-run liquor monopolies and state lotteries generally provide diminished access, high prices, limited product selection and high levels of tax
revenue to the government.6 Revenues typically range from 30–50 per cent
of sales but the high prices tend to encourage smuggling. Most importantly,

Economics of drug liberalization

restricting access by this means does little to distinguish (and punish) bad
behaviour from good behaviour (Whitman, 2003).
Government-controlled human organ transplant organizations provide
‘free’ organs, but as we noted earlier, only in a limited supply that is administratively allocated, and so many patients suﬀer and die because of the
resulting shortage. Most drug maintenance programmes remain limited to
a small percentage of all addicts who do see a general improvement in terms
of consuming a safer product, better economic status and lower levels of
criminal activity.
For instance, in January 1994, the Swiss authorities opened the first
heroin maintenance clinics, and at the end of the three-year trial, more than
800 patients had received heroin on a regular basis. MacCoun and Reuter
(1999) report that the crime rate among all patients declined over the course
of treatment, use of non-prescribed heroin dropped sharply and the unemployment rate among the patients fell from 44 per cent to 20 per cent.
Overall, though, as currently constituted, such programmes have potential
but have not demonstrated their eﬀectiveness in reducing the many harms
of prohibition (MacCoun and Reuter, 2001). They are highly restrictive,
oﬀer little in the way of access to legal drugs, and in the case of methadone
clinics, diversions of the drug have been known to establish new markets
for illicit methadone.
3 Government regulation
Economic analysis of regulation often focuses on price regulations. While
price regulations might be applied in a legalized drug market, it is much
more likely that other aspects of the market would be regulated. In fact,
while price constraints are often a part of government-imposed regulations
in markets, regulations actually deal with many other aspects of most
markets (Benson, 2003). One need only examine the markets for prescription drugs, or various alcohol markets. (For example, see Sass and
Saurman, 1993 to get an idea of the wide array of controls that government
can and does impose on legal markets.)
Thus, there have been many reformers who have suggested that illicit
drugs be made legally available but only through a regulated process
whereby buyers and sellers meet certain government requirements. Kleiman
(1992) argues that alcohol drinkers and marijuana smokers pay a high tax,
have a revocable licence, and a limit on the amount they consume. Under
his scheme, cocaine users would be registered and could receive a limited
amount of cocaine from regulated distributors either at a high price or
under therapeutic supervision. Tobacco users would be registered, sellers
would be licensed, quantities would be limited and heavier taxes would be
imposed. Heroin prohibition would be rigidly enforced, but addicts would

Economics uncut

be registered and placed in maintenance and treatment programmes. The
cost of administering Kleiman’s approach would be extremely high, of
course, and violations would probably be rampant, but there are alternative
approaches as well (see Kleiman, 1992; Thornton, 1994 for a critique).
The prescription-licence approach has many variations both inside and
outside of drug markets. Within ‘prescription’drugs markets, consumers are
generally registered, licensed or given a prescription for narcotic drugs from
a medical doctor or drug treatment therapist. Drugs can then be purchased
from a licensed pharmacy or maintenance programme facility in limited
quantities. Permission to consume could be obtained along a spectrum that
runs from only for legitimate medical needs, to addiction maintenance and
treatment, to any adult who the doctor determines is knowledgeable and
healthy enough to consume such drugs. Drug prices can range from the
highly taxed to free at government-run maintenance programmes.
In the 1920s, the UK adopted a system by which doctors could prescribe
heroin to addicted patients for maintenance purposes. MacCoun and Reuter
(1999) note that this worked reasonably well until the mid-1960s when, ‘. . .
a handful of physicians began to prescribe irresponsibly and a few heroin
users began taking the drug purely for recreational purposes, recruiting
others like themselves’. This led to a large relative increase in heroin addiction, although MacCoun and Reuter stress that the problem was small in
absolute terms (about 1500 known addicts in 1967). In response, the
Dangerous Drugs Act of 1967 significantly reduced access to heroin maintenance, with long-term prescription being limited to a small number of specially licensed drug-treatment specialists. Today, as we noted in Section 1,
doctors in the UK prescribe heroin to only a very small number of addicts.
Benjamin and Anderson (1996) point out that the form of alcohol
control (taxation and regulation) among states in the USA clearly is a function of the cost of inducing compliance. Most states along the Canadian
border where smuggling is easy employ a very diﬀerent approach to alcohol
control than most interior or Southern States. Similarly, alcohol control
diﬀers in the traditional ‘moonshine’ states in the Appalachian Region
where social norms support illegal production, compared with other states.
One good example of where an illegal market was legalized and regulated
is the casino gambling industry in many US states. Here casinos are
licensed, regulated and taxed. Generally, the requirements and taxes are
considered normal rather than strict or lax and the results have been quite
positive (see Chapter 13).
4 Sin taxes
Another alternative to prohibition is to allow drugs to be sold in the
market, but to impose a special tax on the product above the normal sales

Economics of drug liberalization

tax, called an excise tax. This sin tax approach is common on alcohol and
tobacco products, but is also used in the case of gasoline and a variety of
other products. The excise tax has also been called a ‘luxury’ tax when
applied to luxury products, for instance, where the sin apparently is conspicuous consumption, but it is clear that such taxes have been used
throughout history in order to raise revenue for the government.
Many legalization advocates explicitly combine legalization with high sin
taxes, including Gary Becker (see interview in this book and Becker et al.,
2004). The expectation is that such taxes will reduce consumption,
although the magnitude of the reduction depends on the elasticity of
demand and the size of the tax. It will also raise revenues that presumably
can be used to enforce compliance. Indeed, this is Becker’s preferred policy,
although other advocates of a sin tax approach would direct revenues
toward prevention and treatment.
One potential advantage of a sin tax approach is that it is relatively politically attractive. It provides politicians with revenues and allows them to
continue to condemn the product by imposing a sin tax. Many excise taxes
are also earmarked for voter-preferred government programmes such as
law enforcement, road building and education. They can also be directed to
addiction treatment programmes or education campaigns against drug use,
as exemplified by the allocation of some of the money collected from
tobacco companies in the USA. Because most revenues are fungible,
however, the overall impact is to increase the purchasing power of government (Shughart, 1997).
This suggests one flaw in the sin tax approach. While many of its advocates contend that the resulting revenues can and should be earmarked for
enforcement, treatment or some other specific purpose, legislators will
probably reduce revenues directed at such purposes from other sources as
the earmarked sin tax revenues rise. Spending on the desired activity will
not rise in the way that advocates intend it to. Nonetheless, this approach
is more likely to be considered by political decision-makers than a pure free
market, which we discuss next, and if adopted, it may provide many of the
advantages of the free market approach and avoid many of the costs of prohibition (Caputo and Ostrum, 1994). It is not likely to deliver the level of
benefits that its advocates suggest, however, so voters may become dissatisfied over time.
Sin taxes are generally advocated, at least in public debate, not primarily
because they are revenue sources, but because they raise the price of the
taxed product and therefore reduce the quantity demanded. The political
nature of the taxing decision means that this consumption-reducing impact
may be mitigated, however. The level of the tax is likely to be high in a US
taxing jurisdiction where many citizens do not consume the product for

Economics uncut

social, moral and/or religious reasons, for instance, because these people
support eﬀorts to reduce consumption by others as well. On the other hand,
where the dominant norms of the community do not discourage consumption, most citizens may oppose a high excise tax because they consume the
product. Thus, the tax will be high where consumption is already curtailed
by social forces, and low where consumption is widespread.
Note in this regard that the negative correlation between excise taxes and
consumption does not necessarily mean that the high taxes are the primary
cause of the reduced consumption. Both tax rates and consumption levels
may be caused by other factors. Mast et al. (1999) illustrate this for beer
taxes across the USA: a strong negative correlation appears to hold in consumption regressions with few explanatory variables, but as more controls
are added, such as those reflecting the religious make-up of the population,
the coeﬃcient on excise taxes falls, becoming quite small and statistically
insignificant in a fully specified supply and demand model. This does not
mean that the law of demand does not hold, but it does suggest that even
relatively high excise taxes do not play nearly as an important role in
demand and supply determination as other market characteristics do.
Other shortcomings of the sin tax approach also require recognition. For
instance, technically, most excise taxes are paid by the seller. This creates
high compliance costs for sellers. In essence, enforcement costs are shifted,
at least to a degree, from the public sector under prohibition, to sellers.
Naturally, sellers have incentives to avoid such costs, so if they are high
(and/or if the tax is so high that it dramatically reduces their sales and revenues), illegal sales will continue, much as under prohibition. Even though
selling is not per se illegal, many sales will be illegal because taxes are not
collected. A substantial portion of the product sold in this black market
may be less potent and less dangerous than the products sold under prohibition because legally produced products will simply be sold illegally.
However, some illegal products may also be produced and sold (moonshiners still produce alcohol in many places in order to sell it in untaxed
underground markets, for instance).
The excise tax is also a regressive tax in that it takes a larger percentage
of income from low-income households who purchase the taxed product
than from high-income households. In the political arena this may be seen
as a benefit of the sin tax, in that it presumably reduces consumption more
eﬀectively among low-income groups. On the other hand, such taxes create
relatively strong incentives for buyers, and particularly low-income individuals who want to consume the good, to turn to black market sources.
Most sin tax advocates assume that raising taxes simply raises price and
results in reduced consumption. They fail to see that as prices rise consumers have incentives to look for substitutes and producers have incentives

Economics of drug liberalization

to supply them. One substitute for highly taxed goods is the same good sold
in an illegal market. Thus, the incentives of suppliers to avoid compliance
costs, and the incentives of consumers to avoid paying high taxes create the
inevitable result of an ongoing black market, with low-income consumers
in particular being pushed into these ‘illegal’ activities. The relative sizes of
the legal and illegal market is a function of the level of taxes and the
expected punishment for buying and selling in the illegal market.
Consider an example, Quebec’s experiences with taxes on cigarettes. An
9 August 1993 article in Maclean’s (a leading news magazine in Canada)
noted that, ‘tax levels in excess of 60 percent on cigarettes have convinced
many smokers that they are justified in breaking the law’ (quoted in Benson
and Rasmussen, 1996). At the time, roughly half of the cigarettes in Quebec
(and some other Canadian provinces with similar taxes)7 were being sold in
illegal markets in order to avoid excise taxes. Tax evasion is not the only
‘crime’ that accompanies high excise taxes, however. Canada’s illegal cigarettes were being smuggled across the border. For instance, Canada
exported about 7.6 billion cigarettes to the USA in 1992, and police estimated that about 80 per cent were smuggled back into Canada. Cigarette
smuggling became so lucrative that organized crime became involved.
Loads of cigarettes were crossing the St Lawrence from the USA to
Canada, in large boats, generally stolen for this purpose, painted black for
night-time crossings, and armed with machine guns for protection against
police and other criminal organizations. Rival gangs exchanged gunfire as
they competed for shares of the illegal market.
As smuggling and violence increased, demands for already-pressed policing, courts, prisons and other law enforcement services increased. Canadian
police were forced to become involved in the same kind of interdiction and
law enforcement eﬀorts against cigarette markets that law enforcement in
the USA is involved in with drug prohibition. Revenues for enforcement did
not increase suﬃciently to counter the increasing criminal activities, and the
flow of illegal cigarettes went unabated, undermining the ability of the
provincial government to collect the excise taxes. The eﬀort to crack down
on this sin tax-induced crime meant that either additional taxes from other
sources would have to be raised or a larger portion of Canada’s criminal
justice resources would have to be shifted away from eﬀorts to control nonsin tax-induced property and violent crime. Quebec citizens and/or policymakers apparently recognized the tradeoﬀs, as a massive reduction in
cigarette taxes was announced by the Quebec government in 1994.
This suggests that a major problem with the sin tax approach is the
diﬃculty in setting the tax rate. Low tax rates would have little eﬀect on consumption, while high tax rates can spur black markets to develop in order to
avoid the taxes, so the underground production, smuggling, crime and

Economics uncut

corruption associated with prohibition also occur with significant sin taxes
(see Benson and Rasmussen, 1996; Thornton, forthcoming). Becker et al.
(2004) are among the minority of sin tax advocates in that they recognize that
illegal markets will persist under this approach, thus necessitating continued
spending on enforcement. They suggest that setting the optimal level of
expected punishment for black market activities will eliminate that market.
If their model and predictions are correct, it must be that the Canadian
government did not set the optimal levels of expected punishment, but the
same implication arises for alcohol taxes (Benjamin and Anderson, 1996)
and most other highly taxed goods (see Thornton, forthcoming).
For instance, as European governments attempted to establish control
over maritime trade in order to tax it, and granted franchises for numerous
trading monopolies between 1500 and 1800, the ‘average merchant and
seaman’ responded with piracy and smuggling, and a substantial part of
maritime commerce was carried out in violation of the laws of some
nation-state (Rosenberg and Birdzell, 1986). Furthermore, the middle and
even the upper classes willingly wore, drank and ate smuggled goods
(Rosenberg and Birdzell, 1986). In 1776, Adam Smith described the implications of such a policy, beginning with a characterization of the typical
smuggler as:
a person who, though no doubt highly blameable for violating the laws of his
country, is frequently incapable of violating those of natural justice, and would
have been, in every respect an excellent citizen, had not the laws of his country
made that a crime which nature never meant to be so. In those corrupted governments where there is at least a general suspicion of much unnecessary
expense, and great misapplication of the public revenue, the laws which guard it
are little respected. Not many people are scrupulous about smuggling, when,
without perjury, they can find any easy and safe opportunity of doing so. To
pretend to have any scruple about buying smuggled goods, though a manifest
encouragement to the violation of the revenue laws, and to the perjury which
almost always attends it, would in most countries be regarded as one of these
pedantic pieces of hypocrisy which, instead of gaining credit with any body,
serves only to expose the person who aﬀects to practice them, to the suspicion
of being a greater knave than most of his neighbours. By this indulgence of the
public, the smuggler is often encouraged to continue a trade which he is thus
taught to consider as in some measure innocent; and when the severity of the
revenue laws is ready to fall upon him, he is frequently disposed to defend with
violence, what he has been accustomed to regard as just property.

The expectation that such incentives can be overcome by devoting excise
tax revenues to enforcement eﬀorts is simply not borne out by history.
Yet another drawback of a sin tax approach that corresponds to a
prohibition approach is that, like prohibition, taxes are targeted against
consumption in general, not the external harms that some consumption

Economics of drug liberalization

may produce. For example, the tax on red wine in the USA does have the
eﬀect of reducing the consumption of red wine over the entire economy,
but this reduces the health and other benefits of red wine, and yet does little
to target specifically the potential harms of wine consumption, such as
automobile accidents (Mast et al., 1999).
5 Free market
Very few economists advocate a completely free market in drugs. Most
favour replacing one form of government intervention (prohibition) with a
diﬀerent form(s) of government intervention. However, the purpose of this
section is to provide a rigorous examination of how such a free market
might operate.
First, of course, the supply of and demand for a drug would be determined
solely on the basis of market forces. Competitive conditions would result in
relatively low prices and diversified oﬀerings of competitive products, while
consumer sovereignty would dictate that the products that best satisfied consumers in terms of price, quality and so on, would dominate the market.
We would also expect a large number of suppliers to enter the market and
for most suppliers to leave the underground economy. When a new drug is
introduced or legalized there is usually a period of relative chaos and misinformation. Over time, however, we would observe more stability and
better information as the more successful firms (generally those that establish reputations for oﬀering uniformly good quality at a competitive price)
would grow and capture a larger market share, while less successful firms
would fail and exit the industry. Surviving firms would develop mass production and distribution techniques and employ advertising to promote
sales and develop ‘good will’ amongst its customers.
It is impossible to project what a mature market would look like in terms
of the precise number of firms and products, but it is safe to say that most
consumption would be served by commercial production, rather than by
home production. Looking at other mature industries such as soft drinks,
cigarettes, toothpaste, beer and over-the-counter drugs, we find that a small
number of firms supply the majority of the products sold in the marketplace.
These firms each oﬀer a variety of diﬀerentiated products that are massproduced, heavily advertised and sold through well-developed distribution
networks. The existence of a small number of firms does not mean a lack of
competition or the presence of a monopoly. Indeed, competition can be very
intense, as most sources of reduced competition arise through state action
to limit the free market (for example, price controls, entry restrictions
through licensing and other such measures, and assigned markets).
In the absence of such artificial constraints on markets, smaller firms,
regional firms and firms that cater to special tastes are likely to exist

Economics uncut

alongside the large firms. For instance, consider the proliferation of small
local and regional micro-breweries in the USA, which are successfully competing against the large producers of national brands. There might also be
boutique firms, like brewpubs in the beer industry or vacation resorts that
make and sell their own wines, although these would likely constitute a very
small share of the overall market. Likewise, home production is possible
(especially with marijuana), but this form would be negligible in a low-tax
environment.
The underground or black market for these products would be small, but
perhaps not negligible, if the market faces regulation of product distribution (for example, minimum consumption ages). The existence of a wide
variety of producers, many of whom may be small, helps ensure that a wide
diversity of consumers’ tastes are satisfied and provides a source of product
innovation and competition for the large producers, preventing them from
‘dominating’ the market in the sense of limiting price/quality competitions,
even though they may ‘dominate’ in terms of market share.
The advocates of legalization are often quick to emphasize all of the
costs of prohibition but often downplay or dismiss the increased consumption expected under legalization. The supporters of prohibition rest
much of their case on the increase in consumption that would be experienced with legalization, but neglect to consider that legalization would
remove the costs of prohibition. Indeed, some prohibitionists believe that
all the problems related to drugs and drug prohibition, including crime and
corruption, will simply get worse as a function of the (greatly) increased
consumption of drugs, despite considerable evidence to the contrary.
Prohibitionists suggest that once the prohibition is lifted, consumption
of drugs will skyrocket because of a lack of legal restriction, a significant
decrease in price and the use of commercial advertising promoting their
use. They feel that lower prices would increase consumption among current
consumers but more importantly legalization would increase the number of
consumers who currently abstain only because of the legal threats or the
perceived morality of the law.
They point to the experience of the Netherlands. From the mid-1980s,
Dutch drug policy developed from the simple decriminalization of marijuana to the active commercialization and promotion of it. Although legal
restrictions remained as we noted in Section 1, the result was a rapid
increase in the number of marijuana users. For example, in 1984, 15 per cent
of 18–20-year-olds in the Netherlands reported having used marijuana at
some point in their lifetime. By 1992, that figure had more than doubled to
33 per cent. Of course, this does not control for additional factors at work,
but MacCoun and Reuter (1999) note that it is consistent with evidence
from other markets (alcohol, tobacco and legal gambling) that commercial

Economics of drug liberalization

promotion of such activities increases consumption. However, they also
emphasize that the growth in marijuana use, ‘. . . has not led to a worsening of the Dutch heroin problem, . . . cocaine use is not especially high by
European standards, and a smaller fraction of marijuana users go on to use
cocaine or heroin in the Netherlands than in the United States’.
Overall, in a free market, many abstainers would not consume the drugs,
particularly hard ones such as heroin and cocaine even if they were legal
because they consider the consumption of those drugs to be immoral, dangerous or repugnant. Those who only refrain because of the legal threat
would probably consume the drugs responsibly for fear of running afoul of
other legal threats, such as Driving Under Influence (DUI) laws or the loss
of their job or reputation.
In fact, legalization reformers sometimes suggest that there will be little
or no increase in consumption because illegal drugs are readily available
and competitively priced against their legal counterparts. With heroin and
cocaine selling for less than $5 and often available inside government
prisons, a case can be made that prohibition does not really restrict the availability of illegal drugs to any significant degree (Miron, 2004). However, we
cannot completely discount the impact that purchasing illegal drugs
involves more than just price and availability, namely going to prison, losing
your job or overdosing on drugs, and that these threats do diminish the
actual number of drug consumers. Indeed, we explained in Chapter 1 that
it is a drug’s ‘eﬀective’ price that is relevant, not necessarily its market price.
We also noted in Chapter 2 that some individuals may be aﬀected by the
‘forbidden fruit’ eﬀect, which actually increases their demand for illegal
products. By making a good illegal you draw attention to it and encourage
its use as a way of rebelling against society or unjust laws. If prohibition
creates this forbidden fruit eﬀect and increases sales in a black market, then
demand could decrease as a result of legalization. There is some evidence
for the existence of this eﬀect, but nothing that would suggest its magnitude in the case of drugs. While it is considered to be a small component in
the overall demand for illicit drugs, the primary group that this is said to
impact is teenagers and young male adults. This group is a major concern
in the debate, but it may be less of a positive consideration if those same
risk-taking individuals simply seek out alternative hazardous behaviours,
especially those more dangerous than drugs.
Moving from prohibition to a completely free market would likely lead
to an increased consumption of drugs. First, current consumers would face
a lower price and increased quality. Second, new consumers would enter
the market due to the lifting of criminal sanctions and the improved safety
of the products. Third, there would be a substitution from drugs that are
currently legal and highly taxed, such as alcohol and tobacco, into newly

Economics uncut

legalized drugs, which are not. Fourth, legal products would tend to be sold
in lower potency forms so that the quantity of product as measured by
weight or volume sold would increase. Fifth, there would be a surge in
demand for the legitimate medicinal uses of marijuana, cocaine and heroin
that are currently prohibited or restricted. Indeed, marijuana also has
many potential commercial uses as a fibre substitute and for the oil from
its seeds. It can be successfully grown in many climates without the use of
large amounts of fertilizers and insecticides.
Estimates of the projected increase in consumption of these drugs are
highly speculative because they are based on inaccurate historical information and questionable assumptions about how legalization would actually
aﬀect an individual’s demand. Such estimates are often constructed for political purposes or to espouse a particular policy viewpoint and are not trustworthy (see Michaels, 1987). Furthermore, increased consumption is often
viewed by all sides as a negative, but scientifically it is a positive result if no
‘negative externalities’ are associated with that consumption (see Miron,
1991).
Normally price and quantity are overriding economic considerations, but
in the movement from prohibition to legalization everything about the
product and its market changes dramatically, relegating price and quantity
to minor roles. What are often neglected in the drug policy debate are the
real characteristics of free markets governed by the rules of private property, contract and tort, and a ‘free-for-all’ market without any enforceable
rules of the game. The fact is that rules against fraud, duress, imposition of
intentional or accidental harms, and trespass or other involuntary takings,
do exist in all modern economies. Furthermore, even when these rules are
not instituted by the state, various kinds of ‘privately-imposed’ regulations
would virtually guarantee that the production, distribution and consumption of drugs such as marijuana, cocaine and heroin would be significantly
diﬀerent compared with black market conditions. With legalization,
market behaviour will look more like Budweiser, Marlboro and Coca Cola,
and less like Al Capone, Miami Vice and The Sopranos.
Even in a free market, suppliers operate in an environment of both competition and a legal framework that provides certain restrictions, guidelines to behaviour and penalties for violations, mistakes and even bad luck.
This legal framework is apart from the government regulatory apparatus
discussed earlier, which could add various requirements regarding product
characteristics (potency, consistency), distribution (age restrictions, characteristics of actual retail outlets such as location or operating hours),
marketing strategies (advertising restrictions) and so on, if the competitive
market process itself does not control such factors in a way that satisfies
those with political decision-making power or influence. In short, the free

Economics of drug liberalization

market for drugs would be neither a ramped-upped version of the illegal
market or a free-for-all orgy of intoxication.
Forces such as competition for reputation and goodwill, and the risks to
a firm’s capital due to contract and tort liabilities provide much in the way
of what is commonly referred to as ‘consumer protection’, which is almost
entirely absent in black markets. Therefore, we will describe some of the
major aspects of free markets that are either absent or greatly curtailed with
prohibition of drugs and that will either be established or expanded with
legalization and how they influence and moderate behaviour. Indeed, while
‘regulation of markets’ can be provided by government, it also can be provided by other organizations, both formal and informal (see Holcombe and
Holcombe, 1986; Thornton, 1996; Benson, 2001).
We begin with the informal constraints on drug use and abuse that are
considered to have a deterrence eﬀect. While neither commercial nor corporate in nature, informal social sanctions such as embarrassment,
expressions of disapproval, resentment, ridicule, rejection, ritual and
shaming can have a powerful eﬀect on behaviour and can even be more
eﬀective than legal sanctions, especially under the right circumstances (see
Zinberg, 1987). The eﬀectiveness of these sanctions is based on their
immediacy and personal nature. Sanctions imposed by peers, mentors
and guardians are particularly powerful in moderating bad behaviour.
Social sanctioning can correct and prevent behaviours before they
become problematic. In contrast, legal sanctions are only ‘probabilistic’,
and when eventually enforced (that is, going to prison), can actually lead
to long-term association with drugs and crime. Social sanctions would be
more widely practised under legalization, because drug use is necessarily
a concealed behaviour, not an open social behaviour (see MacCoun and
Reuter, 2001).
There are also several constraints on the behaviour of suppliers. For
example, in terms of advertising, suppliers would want to promote their
products with the most cost-eﬀective advertising, but would be prevented
from making false and fraudulent claims for their products. If a supplier
makes fraudulent claims, consumers can sue. This legal constraint works
imperfectly, but tolerably well, in the market for other products. Even in the
absence of actual lawsuits, false and misleading claims work against suppliers of products where purchases are made on an ongoing basis or where
the corporation produces a wide variety of products that could all be
‘tinged’ by the misleading promotions. Building a reputation for reliability
can be a very valuable investment in a competitive market (Benson, 2001).
Much of the serious health consequences of illicit drugs are due to a lack
of information and the presence of misinformation about illicit drugs.
For example, the majority of heroin deaths are not due to heroin per se,

Economics uncut

but consuming heroin with other drugs, which unknowingly causes ‘overdose’. Legalization, competitive advertising and labelling would certainly
address the most egregious aspects of this problem (see Ekelund and
Saurman, 1988).
In terms of potency, purity and product quality, even an unregulated free
market would work much diﬀerently from an illegal market. In an illegal
market, there is an incentive and marked tendency toward higher-potency
products, where purity is variable and uncertain and product quality is low.
With high and variable potency, customers face high risks and many die or
end up in the emergency room of the hospital due to overdoses. As we
emphasized in Chapter 2, suppliers of deadly drugs face little loss or
reprisals for selling dangerous products (see also Thornton, 1991, 1998).
In the market, such behaviour is controlled through the forces of competition and the increased flow of information. Any supplier who sold dangerous or deadly products is not only quickly eschewed by customers; they
also are subject to multimillion dollar lawsuits for negligence. Through
either means, the entire value of the company could be lost. In the market
for alcohol, caﬀeine and tobacco we have seen trends towards lower
potency products, drug-free alternatives and safer products. Other firms
have introduced a large number and wide variety of products and services
to help consumers break their habits. In terms of product safety, consumers
will experience far less danger from legal products compared with illegal
products. The two popular narcotic drug products before narcotics prohibition were both low potency and controlled dosage: Bayer’s heroin pills and
Coca-Cola, which contained cocaine (see Courtwright, 2001). Indeed, in a
free market, there would be an incentive for firms to develop new drugs that
provided the ‘high’ with minimal adverse eﬀects.
As we noted earlier, a free market policy does not imply a free-for-all in
drug consumption. Stores currently post signs stating such rules as ‘no
shoes, no service’, ‘no smoking permitted’, and ‘no alcohol beyond this
point’. Would you shop at stores that allowed marijuana smoking by its
customers or employees? Some people would, but others would not. We
would therefore have high expectations that marijuana smoking would be
largely restricted to specialized marijuana bars and the consumer’s private
property even in the absence of any government regulation. There are a
variety of market mechanisms and alternatives to government regulation
that have been demonstrated to be more eﬀective at protecting the interests
of buyers (see Poole, 1982; Benson, 2001).
Likewise, we fully expect to see employer sanctions on the use of certain
drugs while working on the job. Airlines and trucking companies that
restrict their pilots and drivers from consuming alcohol before or during
work also restrict them from smoking marijuana and a variety of other

Economics of drug liberalization

legal and illegal drugs. Drug restrictions on workers, especially those with
dangerous and sensitive jobs, would continue and possibly be strengthened
if drugs were legalized (see Hartwell et al., 1998). It turns out that one of
the most eﬀective forms of drug treatment is simply having a job (see
Silverman and Robles-Sotelo, 1998; Feinauer, 1990).
Indeed, businesses used drug-testing and behaviour restrictions on
employees long before they were ever required by government (see
McGuire and Ruhm, 1993). It is also worth noting that while much remains
to be established in the relationship between drug use and economic outcomes (employment, income and wealth), it is clear that labour markets
provide one of the most eﬀective checks on and punishment for drug abuse,
whether legal or illegal. Businesses do this not because they necessarily
oppose using drugs, but because they have an economic incentive to do so.
For comparison, while the cumulative aggregate eﬀect of law enforcement
over time is large, the probability of getting caught by law enforcement in
a single act of selling or consuming illicit drugs is quite small compared
with the much higher probability of being caught and swiftly punished
if your alcohol or drug use threatens your employer. As we suggest in
Chapter 4, the certainty and swiftness (or ‘celerity’) of punishment appears
to have a greater deterrent eﬀect compared with the severity of punishment
(see also Norman et al., 1994).
What about the case of children? In most cases, businesses would not sell
dangerous drugs to children. Businesses have reputations to uphold and the
bigger the business, the more important that reputation is. It should also be
noted that, under both common and civil law, businesses cannot make valid
legal contracts with minors. Pharmacies generally do not sell dangerous
drugs to children, auto dealers do not sell them cars, and all sorts of other
companies do not do business with children, either out of concern for the
children, their own reputations or their financial viability and survival.
With regard to the latter, companies that sell drugs to children would face
severe legal consequences if harm comes to the children. Therefore, legal
liability and negligence laws would greatly discourage suppliers from selling
most drugs to minors. The incentives to restrict sales of dangerous and
potentially poisonous drugs to children would be particularly strong in a
free market, and much stronger than they are in an illegal market. Some
adults may buy drugs and sell them to children, of course, as some do with
alcohol and tobacco products, and some children would obtain false identification or other means of purchasing drugs, but even then, the drug
product is likely to be less potent, and safer due to reduced variability in
quality.
Clearly, some problems arise or persist with a free market approach for
drugs. As just suggested, drugs would still be available to children, although

Economics uncut

the products generally would be in safer forms. If adults and young adults
can freely purchase the drugs there would be access to minors as there is
now with minimum age requirements. Indeed, there probably would be
marginal sellers who might specialize in selling to children.
A second potential problem is that drug abuse can lead to negative health
consequences even if the products are legally produced. We would expect
the number of deaths and emergency room visits due to drug overdose to
significantly decline despite overall increases in consumption, but there
would still be people who harm their health via drug abuse, just as with
alcohol and tobacco products today (although probably on a much smaller
scale). Something that greatly exacerbates this problem is that the bulk of
medical services is paid via government and insurance. This creates a moral
hazard in that individuals can place their health at greater risk (with drug
abuse and other health risks) given that the government will pay for their
medical and hospital expenses in the event of an overdose, accident or
other health problem. In other words, if individuals could not depend on
the government to provide free medical services, most would be more riskaverse in their drug use. Private insurers are less susceptible to this problem
because they provide incentives (for example, lower rates for non-smokers)
and reduce adverse selection problems by providing group insurance
policies that include healthy and unhealthy individuals.
In addition, there is a moral hazard associated with government unemployment insurance and welfare benefits. Severe drug abuse (which is what
we are most concerned about) often leads to economic consequences such
as the loss of employment, the inability to find a job and, worse still,
poverty (Kaestner, 1998). This form of market punishment is less threatening if individuals have state-provided unemployment insurance in the
short term and welfare benefits in the longer term. Removing this safety
net for health and unemployment problems associated with drug abuse
would reduce, although not eliminate, risk-taking behaviour associated
with drug abuse.
There are also problems associated with drug use and driving automobiles, but these problems are not unique to drugs that are currently illegal.
As we revealed in Chapter 2, the eﬀects of alcohol consumption on driving
are apparently more significant than those associated with most illegal
drugs. Even though marijuana, cocaine and heroin may be less intoxicating
and pose less of a threat than alcohol, they still could impose negative
externalities on other drivers if the overall level of intoxicated driving were
to increase. This need not occur, however. If marijuana consumption
increases in a free market, for instance, with no change in alcohol consumption, then the level of danger and number of accidents could increase,
but if marijuana is a substitute for alcohol, the negative consequences of

Economics of drug liberalization

intoxicated driving could decline even with increased marijuana consumption. In this regard, studies by DiNardo and Lemieux (2001), Chaloupka
and Laixuthai (1997) and Model (1993) suggest that marijuana and alcohol
are substitutes. Indeed, Model (1993) found that the decriminalization of
marijuana in some states in the USA is associated with lower numbers of
emergency room episodes related to alcohol (and other illegal drugs), while
Chaloupka and Laixuthai (1997) report that lower marijuana prices lead to
a significant drop in the probability of automobile accidents.
Yet another potential problem is that in a free market, existing drugs and
particularly new drugs could have unknown side-eﬀects from long-term
consumption. This was a problem with cigarettes because the long-term
results such as lung cancer were only determined after cigarettes were
popular for a long time and people started living long enough for the longterm health consequences to be readily apparent. Indeed, this can be a
problem with all drugs, whether sold in free or black markets, and it is not
a problem that can be easily dealt with under any policy. Government
testing for a drug’s eﬀectiveness and side-eﬀects does not screen for longterm consequences and actually worsens the potential problem by providing products with the government’s seal of approval (see Higgs, 1995).
Importantly, however, as we have already suggested, this problem is likely
to be more significant under prohibition. Tort liability creates relatively
strong incentives for legal drug producers to consider and attempt to mitigate potential long-term eﬀects, after all, but illegal producers do not face
such liabilities. Furthermore, since such consequences can never be fully
avoided, tort law does create the potential for those harmed by products
purchased in legal markets to at least recover some compensation, something that is virtually impossible when the harm arises from a drug purchased from an illegal firm under prohibition.
6 Conclusion
This chapter spelled out the major forms of drug legalization that are available to policy-makers throughout the world. Government monopoly, regulation, sin taxes and the free market were all described with the help of
well-developed economic models that allowed us to explore their costs and
benefits. As each of these policies has been used in a variety of diﬀerent
industries, including drugs and other illicit markets, we also had a wealth
of historical experience to draw upon in order to further understand their
implications. While Chapter 2 established that drug prohibition is harmful
in numerous respects, this chapter’s analysis showed that many of the alternatives are certainly not perfect either. Each has its own drawbacks and
costs. Given that the reader will likely face the prospects of changing drug
policy in their lifetime, this should be valuable information.

Economics uncut

Notes
1.

2.
3.

5.
6.
7.

Coﬀee shops are restricted under the Opium Act from advertising, selling hard
drugs, admitting or selling to those under 18 years of age, selling more than 5 grams to
a person at any one time and causing a nuisance. Moreover, the availability of coﬀee
shops depends upon policy in any particular Dutch municipality, that may choose to
allow them, allow them with certain restrictions, or not allow them at all (see Martineau
and Gomart, 2001, for a discussion of the evolution of Dutch coﬀee shops). At the
time of writing (May 2005), the very existence of coﬀee shops is being threatened by
new measures including a possible ban on foreigners from buying drugs in coﬀee shops
and an anti-smoking law stipulating that employees should not be exposed to tobacco
smoke.
Pill testing refers to the concept of providing illicit drug users with quality assessment of
their chosen drug.
Oregon in 1973; Colorado, Alaska and Ohio in 1975; California, Maine and Minnesota
in 1976; Mississippi, New York and North Carolina in 1977 and Nebraska in 1978.
Oregon recriminalized in 1996 but in 1998 voters rejected recriminalization and reestablished decriminalization. Alaska also recriminalized in 1990, although it is not clear
that they did so in practice, but recently decriminalized again.
Note that the actual characteristics of each of these 13 states’ decriminalization acts
diﬀer quite substantially, and many other states have adopted at least some of the legal
changes that these 13 states have (see Pacula et al., 2003 for details about what appear to
be the relevant aspects of liberalization in state laws).
Connecticut, Louisiana, Massachusetts, New Jersey, Vermont, Wisconsin and West
Virginia.
But see Benson et al. (2003) who find that some state liquor monopolies clearly collect
less revenue than they could through a sin tax approach, perhaps due to the bureaucratic
ineﬃciencies and/or consumer-unfriendly locations and operating practices.
In fact, it was estimated that about one of every nine cigarettes consumed in Canada as
a whole were illegally purchased.

References
Becker, G.S., M. Grossman and K.M. Murphy (2004), ‘The Economic Theory of Illegal
Goods: The Case of Drugs’, National Bureau of Economic Research Working Paper,
No. W10976.
Benjamin, D.K. and T.L. Anderson (1996), ‘Taxation, Enforcement Costs, and the Incentives
to Privatize’, in T.L. Anderson and P.J. Hill (eds), The Privatization Process: a Worldwide
Perspective, Lanham, MD: Rowman & Littlefield, pp. 39–54.
Benson, B.L. (2001), ‘Knowledge, Trust, and Recourse: Imperfect Substitutes as Sources of
Assurance in Emerging Economies’, Economic Aﬀairs, 21, 12–17.
Benson, B.L. (2003), ‘Regulatory Disequilibrium and Ineﬃciency: The Case of Interstate
Trucking’, Review of Austrian Economics, 15, 229–55.
Benson, B.L. and D.W. Rasmussen (1996), ‘Predatory Public Finance and the Origins of the
War on Drugs: 1984–1989’, The Independent Review: A Journal of Political Economy, 1,
163–89.
Benson, B.L., D.W. Rasmussen and D.L. Sollars (1995), ‘Police Bureaucrats, Their Incentives,
and the War on Drugs’, Public Choice, 83, 21–45.
Benson, B.L., D.W. Rasmussen and P.R. Zimmerman (2003), ‘Implicit Taxes Collected by
State Liquor Monopolies’, Public Choice, 115, 313–31.
Bureau of Justice Statistics (2003), Sourcebook of Criminal Justice Statistics 2003,
Washington, DC: United States Department of Justice.
Caputo, M.R. and B.J. Ostrum (1994), ‘Potential Tax Revenue from a Regulated Marijuana
Market: A Meaningful Revenue Source’, American Journal of Sociology, 53, 473–90.
Chaloupka, F.J. and A. Laixuthai (1997) ‘Do Youths Substitute Alcohol and Marijuana?
Some Econometric Evidence’, Eastern Economic Journal, 23 (3), 253–76.

Economics of drug liberalization

Courtwright, David T. (2001), Forces of Habit: Drugs and the Making of the Modern World,
Cambridge, MA: Harvard University Press.
DiNardo, J. and T. Lemieux (2001), ‘Alcohol, Marijuana, and American Youth: The
Unintended Consequences of Government Regulation’, Journal of Health Economics, 20,
991–1010.
Drug and Alcohol Services Information System (2001), ‘Coerced Treatment Among Youths:
1993 to 1998’, The DASIS Report (21 September), 1–3.
Drug Enforcement Agency (DEA) (2002), ‘Drug Intelligence Brief – The Changing Face of
European Drug Policy’, April.
Ekelund, R.B and D.S. Saurman (1988), Advertising and the Market Process: A Modern
Economic View, San Francisco, CA: Pacific Research Institute for Public Policy.
European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) (2003), ‘The State
of the Drugs Problem in the European Union and Norway’, Annual Report.
European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) (2004), ‘The State
of the Drugs Problem in the European Union and Norway’, Annual Report.
Executive Oﬃce of the President, Oﬃce of National Drug Control Policy (ONDCP) (2001),
‘Drug Treatment in the Criminal Justice System’, ONDCP Drug Policy Information
Clearing House Fact Sheet (March), 1–6.
Executive Oﬃce of the President, Oﬃce of National Drug Control Policy (ONDCP) (2003),
‘Drug Data Summary’, ONDCP Drug Policy Information Clearing House Fact Sheet
(March), 1–8.
Federal Bureau of Investigation (various years), Uniform Crime Report.
Feinauer, D.M. (1990), ‘The Relationship Between Workplace Accident Rates and Drug and
Alcohol Abuse: The Unproven Hypothesis’, Labor Studies Journal, 15 (4), 3–15.
Fisher, Irving (1930), The Noble Experiment, New York: Alcohol Information Committee.
Grinspoon, Lester and James B. Bakalar (1993), Marihuana: The Forbidden Medicine, New
Haven, CT: Yale University Press.
Hartwell, T., P.D. Steele and N. Rodman (1998), ‘Workplace Alcohol-testing Programs:
Prevalence and Trends’, Monthly Labor Review, 121 (6), 27–34.
Higgs, Robert (ed.) (1995), Hazardous to our Health: FDA Regulation of Health Care Products,
Oakland, CA: Independent Institute.
Holcombe, R.G. and L.P. Holcombe (1986), ‘The Market for Regulation’, Journal of
Institutional and Theoretical Economics, 142, 684–96.
Kaestner, R. (1998), ‘Does Drug Use Cause Poverty?’, Cambridge, MA: National Bureau of
Economic Research Working Paper No. 6406.
Kleiman, Mark A.R. (1992), Against Excess: Drug Policy for Results, Basic Books, New York.
MacCoun, R.J. and P. Reuter (1999), ‘Does Europe Do It Better? Lessons from Holland,
Britain and Switzerland’, The Nation, 20 September.
MacCoun, Robert J. and Peter Reuter (2001), Drug War Heresies: Learning from Other Vices,
Times, & Places, Cambridge University Press.
MacCoun, R.J. P. Reuter and T. Schelling (1996), ‘Assessing Alternative Drug Control
Regimes’, Journal of Policy Analysis and Management, 15 (3), 330–52.
Martineau, H. and E. Gomart (2001), ‘Coﬀee Shops in the Netherlands: The Dutch Tolerance
in Practice’, OFDT (Observatoire François des Drogues et des Toxicomanies),
www.ofdt.fr/BDD/publications/docs/tend11gb.pdf.
Mast, B.D., B.L. Benson and D.W. Rasmussen (1999), ‘Beer Taxation and Alcohol-Related
Traﬃc Fatalities’, Southern Economic Journal, 66 (2), 214–49.
Mast, B.D., B.L. Benson and D.W. Rasmussen (2000), ‘Entrepreneurial Police and Drug
Enforcement Policy’, Public Choice, 104, 285–308.
McGuire, T.G. and C.J. Ruhm (1993), ‘Workplace Drug Abuse Policy’, Journal of Health
Economics, 12 (1), 19–38.
Michaels, R.J. (1987), ‘The Market for Heroin Before and After Legalization’, in R. Hamowy
(ed.), Dealing with Drugs: Consequences of Government Control, San Francisco: Pacific
Research Institute for Public Policy, pp. 289–326.
Miron, J.A. (1991), ‘Drug Legalization and the Consumption of Drugs: An Economist’s
Perspective’, in Melvyn B. Krauss and Edward P. Lazear (eds), Searching For Alternatives:

Economics uncut

Drug-control Policy in the United States, Stanford, CA: Hoover Institution Press,
pp. 68–76.
Miron, Jeﬀrey A. (2004), Drug War Crimes: The Consequences of Prohibition, Oakland, CA:
Independent Institute.
Model, K.E. (1993), ‘The Eﬀect of Marijuana Decriminalization on Hospital Emergency
Room Drug Episodes: 1976–1978’, Journal of the American Statistical Association, 88,
737–47.
Nadelmann, E.A. (1992), ‘Thinking Seriously about Alternatives to Drug Prohibition’,
Daedalus, 121 (3), 85–132.
Norman, J., R.O. Lempert and C.P. O’Brien (eds) (1994), Under the Influence? Drugs and the
American Workforce, Washington, DC: National Academy Press.
Osborn, A. (2003), ‘Joint Operation’, The Guardian, 24 November.
Pacula, R.L., J.F. Chriqui and J. King (2003), ‘Marijuana Decriminalization: What Does
it Mean in the United States?’, National Bureau of Economic Research Working Paper
No. 9690.
Poole, R.W. (ed.) (1982), Instead of Regulation: Alternatives to Federal Regulatory Agencies,
Lexington, MA: D.C. Heath and Company.
Rasmussen, David W. and Bruce L. Benson (1994), The Economic Anatomy of a Drug War:
Criminal Justice in the Commons, Lanham, MD: Rowman and Littlefield.
Rosenberg, N. and L.E. Birdzell Jr. (1986), How the West Grew Rich: The Economic
Transformation of the Industrial World, Basic Books, New York.
Sass, T.R. and D.S. Saurman (1993), ‘Mandated Exclusive Territories and Economic
Eﬃciency: An Empirical Analysis of the Malt-Beverage Industry’, Journal of Law and
Economics, 36, 153–77.
Senlis Council (2004), Document Centre – Drug Policies Table (http://www.senliscouncil.net/
modules/document_centre/drug_policy_in_europe).
Shughart, William F. (ed.) (1997), Taxing Choice: The Predatory Politics of Fiscal
Discrimination, New Brunswick, NJ: Transaction Publishers.
Silverman, K and E. Robles-Sotelo (1998), ‘Employment as a Drug Abuse Treatment
Intervention: A Behavioral Economic Analysis’, Cambridge, MA: National Bureau of
Economic Research Working Paper No. 6402.
Smith, Adam (1937) [1776], An Inquiry into the Nature and Causes of the Wealth of Nations,
E. Cannan (ed.), New York: Modern Library.
Thornton, Mark (1991), The Economics of Prohibition, Salt Lake City, UT: University of
Utah Press.
Thornton, M. (1994), ‘Review of Against Excess: Drug Policy for Results’, Review of Austrian
Economics, 7 (1), 147–50.
Thornton, M. (1995), ‘Economists on Illegal Drugs’, Atlantic Economic Journal, 23 (2), 73.
Thornton, M. (1996), ‘The Market for Safety’, Business Quest: A Journal of Applied Topics in
Business and Economics (http://www.westga.edu/~bquest/1996/uwlab.html).
Thornton, M. (1998), ‘The Potency of Illegal Drugs’, Journal of Drug Issues, 28 (3), 725–40.
Thornton, M. (2004), ‘Prohibition vs. Legalization: Do Economists Reach a Conclusion on
Drug Policy?’, Econ Journal Watch, 1 (1), April.
Thornton, M. (forthcoming), ‘Harm Reduction and Sin Taxes: Why Gary Becker is Wrong’,
Advances in Austrian Economics.
Tullock, Gordon and Richard B. McKenzie (1985), The New World of Economics:
Explorations into the Human Experience, 4th edition, Homewood IL: Richard D. Irwin, Inc.
United Nations Oﬃce on Drugs and Crime (UNODC) (2005), ‘World Drug Report 2005’,
UNODC.
Whitman, Douglas G. (2003), Strange Brew: Alcohol and Government Monopoly, Oakland,
CA: The Independent Institute.
Zinberg, N.E. (1987), ‘The Use and Misuse of Intoxicants: Factors in the Development of
Controlled Use’, in R. Hamowy (ed.), Dealing with Drugs: Consequences of Government
Control, San Francisco: Pacific Research Institute for Public Policy, pp. 247–79.

PART II
GUNS AND ROSES

Economics of crime
Bruce L. Benson and Simon W. Bowmaker

In this chapter we turn to crime. As early as 1763 in his Lectures on
Jurisprudence, Adam Smith observed that, ‘. . . the disorders in any
country are more or less according to the number of retainers and dependents in it’. The Scotsman was referring to the ‘idle and luxurious life’
enjoyed by servants, which, ‘renders them altogether depraved both in mind
and body’, and hence unwilling and unable to survive when discarded by
their masters other than by ‘crimes and vices’. Moving toward a commercial economy was the solution according to Smith. Not only would this
‘give the poorer sort better wages’, but also an independence, which is the
most eﬀective means of protection from crime.
In the year following Smith’s Lectures, Cesare Beccaria wrote Essays on
Crime and Punishment. The Milanese jurist, criminologist and economist
was a fervent opponent of the severities and abuses of criminal law, particularly capital punishment and torture. He argued that for punishment to be
eﬀective, it should ‘never be an act of violence committed by one or many
against a private citizen’, and that the harshest penalties must be imposed
in moderation while being both ‘proportionate to the crime, and established
by law’.
Beccaria’s writings stimulated and provided a guide for reforms in the
penal codes of many European nations and his pioneering ideas of presumption of innocence and the protection of civil liberties later impacted
upon the US Constitution and especially the Bill of Rights. He also had a
great influence upon Jeremy Bentham, the founding father of utilitarianism. Bentham argued that the rational choice theory proposed by Beccaria
assumed that individuals commit crime because the benefit outweighs the
cost. The Englishman called this thought process the ‘hedonic calculus’ and
concluded, like Beccaria, that punishment should be designed to persuade
individuals that criminal activity was not worth the price to be paid. In his
review of the Hard Labour Bill in 1778, Bentham argued for a ‘general
plan of punishment . . . in which solitary confinement might be combined
with labour’.
In 1785, Bentham’s compatriot, William Paley, suggested in The
Principles of Moral and Political Philosophy that solitary incarceration
did provide the greatest opportunity of ‘reform’, although the length of
confinement ought to be measured by the ‘quantity of work’ achieved
101

102

Economics uncut

by the prisoner, which would ‘excite industry, and . . . render it more
voluntary’.
Unlike Bentham and Beccaria, however, Paley defended the use of capital
punishment. In particular, he embraced the Georgian ‘Bloody Code’, which
incorporated the idea that this form of punishment was a more powerful
tool if many diﬀerent types of crime could be swept ‘into the net’ of capital
oﬀences. ‘Paley’s Net’, as it came to be known, would be executed only in
exceptional cases and so we would find that ‘few actually suﬀer death, whilst
the dread and danger of it hang over the crimes of many’.
Edwin Chadwick, another nineteenth-century economist and utilitarian
reformer, analysed the allocation of criminal justice resources (circa
1829–41) as an evolved publicly provided good with open-access common
pool characteristics (see Ekelund and Dorton, 2003). Chadwick’s primary
goal was crime prevention: ‘A good police would be one well-organized
body of men acting upon a system of precautions, to prevent crimes and
public calamities; to preserve public peace and order’. Given this goal he
identified the evolved criminal justice system as a common pool and
presented a rationale for restructuring the institutions of the entire system
to deal with common pool problems. He believed that the system misplaced
incentives for crime prevention, creating dissipations due to the common
access to policing and other criminal justice services, so he recommended a
number of interrelated incentive alterations for both enforcement and
penal systems.
Surprisingly, economists’ interest in crime and criminal justice declined
in the late nineteenth and early twentieth centuries. It is only during the past
four decades, largely in consequence of Gary Becker’s seminal 1968 contribution1 that their interest in crime has been revived. Like his classical
school predecessors, Becker argued that potential criminals respond
rationally and consistently to incentives. Crime is viewed as a problem of
constrained utility maximization with individuals maximizing expected
utility by choosing between legal and illegal activities after considering the
expected gains and costs associated with each of the alternatives. It follows,
according to Becker, that potential criminals will be deterred from
oﬀending by, first, increases in the probability of being caught and punished and, second, increases in the amount of punishment if caught.
This approach to the issue was initially opposed by criminologists, before
it eventually became accepted that economics could make a significant contribution to the design of legal penalties and enforcement schemes. Indeed,
in the last few years, economists have directly impacted upon government
penal policy, particularly in the USA. For instance, Isaac Ehrlich, whose
work is discussed later in this chapter, played a leading role in the
re-introduction of the death penalty option post-1976 and, more recently,

Economics of crime

103

the empirical work of economists such as John Lott and Steven Levitt has
stimulated considerable public debate and policy change in areas ranging
from gun control to punishment policy (see Lott, 1998; Levitt, 1998).
This chapter proceeds as follows. After a brief overview of the world’s
crime problem, Section 2 provides an illustration of the economic theory
of the supply of criminal oﬀences. In order to test the hypotheses generated
by this theory, it must be recognized that the supply of oﬀences may have
to be examined simultaneously with demand for crime prevention and law
enforcement services. This demand is considered in Section 3 in the context
of an examination of the empirical research relating to the so-called ‘deterrence hypothesis’, that is, whether crime rates rise with a reduction in the
opportunity costs of crime and a decline in the expected cost of crime.
Then, in Section 4, we investigate how resources are allocated within the
public-sector criminal justice system, why they are allocated the way they
are, and what the consequences are of this allocation process. Section 5
examines possible reasons for the declining crime rates in the USA during
the 1990s in light of the theoretical and empirical findings outlined in the
previous sections. Finally, Section 6 provides some concluding thoughts on
the economics of crime and ideas for future research in this field.
1 A criminal world
Direct comparisons of crime rates between countries are fraught with
diﬃculties; laws are diﬀerent across international boundaries, and methods
of collecting crime data vary enormously, with some countries relying on
administrative data collected by the police and others relying more on
surveys. Further, there are problems in comparing recorded crime rates
over time within countries because of changes in levels of reporting to the
police and recording by them. With these caveats in mind, we now show
that over the past few decades, crime has for the most part been a growth
industry in many countries.
Figure 4.1, for example, reveals that in England and Wales the number
of recorded crimes per 100 000 population in 1951 was 1197. By 2003/04,
this figure had increased almost ten-fold to 11 309. A similar picture
emerges when we observe the experience of the USA. Table 4.1 shows that
the All Index Crimes rate (which includes the seven categories listed)
jumped by over 115 per cent between the mid-1960s and the mid-1990s.
However, the USA did enjoy an unexpected fall in crime during the 1990s.
As Table 4.1 reveals, the All Index Crimes rate dropped by 30 per cent
between 1990 and 2003, with the Murder and Non-negligent manslaughter
rate declining by 40 per cent. Yet, despite this sharp fall, Figure 4.2
shows that over the 2000 to 2002 period, the USA still had one of the
highest murder rates in the world at 5.6 per 100 000 population. The two

104

1901

248

Figure 4.1

1911

269
1921

272
1931

398
1941

859

1951

1 197

1961

1 747

1971

3 349

1981

5 971

Recorded crime rate per 100 000 population in England and Wales (1901–2003/04)

Home Oﬃce, UK.

2000

4000

6000

8000

10000

12000

Source:

Recorded crime rate per 100 000 population

1991

10 325

2003/04

11 309

105

31.8
33.1
32.1

9.6
12.1
18.7
26.3
36.8
37.1
41.2
37.1
32.0

Forcible
rape

Federal Bureau of Investigation.

5.6
5.6
5.7

2001
2002
2003

Source:

5.1
5.1
7.9
9.6
10.2
8.0
9.4
8.2
5.5

Murder and
Non-negligent
manslaughter

148.5
146.1
142.2

60.1
71.7
172.1
220.8
251.1
208.5
257.0
220.9
144.9

Robbery

318.6
309.5
295.0

86.1
111.3
164.8
231.1
298.5
302.9
424.1
418.3
323.6

Aggravated
assault

741.8
747.0
740.5

508.6
662.7
1084.9
1532.1
1684.1
1287.3
1235.9
987.1
728.4

Burglary

Recorded crime rate per 100 000 population in USA (1960–2003)

1960
1965
1970
1975
1980
1985
1990
1995
2000

Table 4.1

2485.7
2450.7
2414.5

1034.7
1329.3
2079.3
2804.8
3167.0
2901.2
3194.8
3043.8
2475.3

Larceny –
theft

430.5
432.9
433.4

183.0
256.8
456.8
473.7
502.2
462.0
657.8
560.4
414.2

Auto
theft

4162.6
4124.9
4063.4

1887.2
2449.0
3984.5
5298.5
5950.0
5207.1
5820.3
5275.9
4124.0

All index
crimes

Murder rate per 100 000 population by country (2000–02 average)

Home Oﬃce (UK).

Figure 4.2

Source:

2.9 3.6
1.7 1.8 1.8 1.8 1.8 2.0 2.2 2.2 2.4 2.5 2.5 2.7 2.8
0.5 0.9 1.0 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.2 1.2 1.2 1.3 1.4 1.4 1.5

Data for Austria, Italy, Romania, Russia, Sweden (1999–2001).

0.0

10.0

20.0

30.0

40.0

Note:

Murder rate per 100 000 population
50.0

Cyprus
Norway
Japan

Spain
Malta
Austria
Italy

Australia
France

Slovakia
Turkey

60.0

Luxembourg
Denmark
Sweden
Switzerland
Germany
Portugal
Slovenia
Greece
Netherlands
Ireland (Eire)
New Zealand
England & Wales
Canada
Poland
Hungary
Scotland
Romania
Czech Republic
Finland
Northern Ireland
Bulgaria

106
5.6

USA
Estonia
Russia

48.8

22.2

10.4
10.3
7.1

Belgium
Lithuania
South Africa

Economics of crime

107

countries with the highest murder rates per 100 000 population were South
Africa (48.8) and Russia (22.2). Figure 4.3 shows murder rates over the
2000 to 2002 period for selected world cities with the highest rates being
found in the US cities of Washington DC (42.8), Dallas (18.8), San
Francisco (8.3) and New York (8.1) and the Eastern European cities of
Tallinn (Estonia, 9.4) and Vilnius (Lithuania, 8.9).
Figure 4.4, meanwhile, reveals that the US imprisonment rate of 702 per
100 000 population was the highest in the world in 2002. The latest data
show that as of 30 June 2004, the number of inmates held in US state and
federal prisons and jails as well as local jails totalled 2 131 180, an increase
of 85 per cent from 1 148 702 in 1990 (Harrison and Karberg, 2005).
Russia, the former world leader in imprisonment rates, had reduced its
imprisonment rate to 608 per 100 000 population by 2002, with this figure
expected to fall further following the approval of a prisoner amnesty by the
Russian parliament in 2000. England and Wales (136), Portugal (132) and
Scotland (127) had fairly high rates of imprisonment compared with others
in Western Europe, which partly reflects the longer sentences imposed in
these countries. Higher rates of imprisonment were reported in South
Africa (425) and in some Eastern European countries, including Estonia
(351), Lithuania (333) and Poland (212).
One reason for the diﬀerences in crime statistics across countries is that
politics is central to the criminal justice system, even in terms of defining
what action is criminal. For example, Anglo-American common law has
always made a distinction between customary law crimes (such as murder,
robbery, theft and rape) and so-called positive-law crimes, which have
become crimes exclusively because some political interest group won or lost
a political battle. Of course, objectives of criminalization are often unclear
because of multiple demands of special interest groups and the misleading
nature of political rhetoric, but changes in criminal law are driven by political forces (see Benson, 1990).
In other words, criminal deviancy is not simply a philosophical or a moral
issue; it is a political question. Which acts are defined as criminal depend on
the interest of the persons with suﬃcient power and influence to manage to
have their views prevail. Once certain acts are designated as criminal, how
the laws are implemented will also reflect the political power of the various
aﬀected groups. This suggests that viewing criminals as somehow diﬀerent
from other people, in the sense that their behaviour is abnormal or irrational, is questionable. An individual may be on the wrong side of the law
simply because he or she is not in the winning political coalition. Thus, the
decision to engage in crime may be quite rational. This leads us to Gary
Becker’s economic theory of crime, which approaches criminal activity in a
manner that diﬀers significantly from other social science paradigms.

Murder rate per 100 000 population by city (2000–02 average)

Home Oﬃce (UK).

Figure 4.3

Source:

3.8 3.9
3.0 3.2 3.2
1.9 2.1 2.1 2.2 2.6 2.8 2.8
1.1 1.1 1.5 1.5 1.5 1.6 1.6 1.7 1.8 1.8 1.9
0.2 0.4 0.5 0.5 0.7

Data for Vienna (1999–2001).

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Note:

Murder rate per 100 000 population
40.0

Berne

Tokyo
Lisbon

Geneva
Paris
Oslo

Berlin

Vienna

Warsaw

45.0

Lefkosia
Athens & Pireus
Canberra
Ottawa
Ljubljana
Wellington
Sydney
Copenhagen
Edinburgh
Madrid
Helsinki
Budapest
Dublin
London
Ankara
Brussels
Amsterdam
Prague
Bratislava

108
6.2

Belfast

9.4
8.1 8.3 8.9

New York
San Francisco
Vilnius
Tallinn

42.8

18.8

Dallas
Washington DC

109
Finland
Malta

Belgium

France

Canada

120 121 127

Spain

158
132 136 143 146

Prison population rate per 100 000 national population by country (2002)

Home Oﬃce (UK).

101 102 114
83 86 87 87 88 94
68 70 73 76 77
67
63
61
57
48 53

Figure 4.4

100

200

300

400

500

600

700

800

Source:

Prison population rate per 100 000 national population

Cyprus
Japan
Slovenia
Northern Ireland
Denmark
Switzerland
Sweden
Greece
Ireland (Eire)
Germany
Luxembourg
Turkey
Austria
Netherlands
Australia
Bulgaria
Scotland
Portugal
England & Wales
Slovakia
New Zealand
Czech Republic

177

Hungary

212

Poland

333

Lithuania

351

Estonia

425

South Africa

608

Russia

702

USA

110

Economics uncut

2 The supply of criminal offences
Becker does not try to specify the ultimate causes of crime (that is, the ‘preferences’ that individuals have, whether ‘normal’ or ‘abnormal’). Instead, he
attempts to identify variables that structure the costs of and returns to
criminal activity in order to predict the incidence of crime. That is, he
explains observed criminal behaviour directly through the examination of
social and economic variables rather than indirectly as a result of the psychological makeup of potential criminals.
As we noted earlier, Becker assumes that criminals respond rationally
and consistently to incentives. Therefore, crime arises from the optimizing
decisions of rational economic agents who maximize expected utility functions that take as arguments the expected return from criminal participation (both monetary and psychic rewards), the probability of being
apprehended for an oﬀence, and the ‘monetary equivalent’ of the severity
of punishment. The main implications of the model are its unambiguous
prediction that the ‘supply of oﬀences’ is an increasing function of the
return to criminal activity and a decreasing function in the level of deterrence factors. However, whether crime is reduced more by increases in the
probability of apprehension or in the expected severity of punishment
(given apprehension) depends on the distribution of risk aversion across
individuals in the economy.
Ehrlich (1973) extends Becker’s model by casting participation in the
light of the theory of occupational choice. Individuals derive income from
time spent in either legitimate or illegitimate (that is, criminal) activities.
The diﬀerence between the two is that income generated from participation
in illegitimate activities is subject to uncertainty due to deterrence factors.
Ehrlich assumes a utility function of the form:
US U(XS, C)

(4.1)

In this utility function, C is the amount of time devoted to consumption
or non-market activity and X is the stock of a composite market good.
He then assumes that there are two states in S: arrest (a) or success (b),
and that:
Xa WI WT(T)WL(L)

(4.2)

Xb WI WT(T)F(T)WL(L)

(4.3)

The level of the initial asset is WI, while WT, F, and WL are earnings from
illegal activities, the monetary equivalent of punishment and earnings from

Economics of crime

111

legal activities respectively. T is time spent in illegal activities, and L is time
in legal earning activities. Expected utility is given by:
E[U(XS, C)](1P) U(Xa, C)P U(Xb, C)

(4.4)

Note that P is the probability of a successful crime. Under these circumstances, the objective is to maximize expected utility over T subject to a time
constraint:
Max E[U(XS, C)] subject to kTLC

(4.5)

The total hours available to the individual are given by k. The Kuhn–Tucker
first-order optimality conditions are:
(EU/T) 0

(4.6)

[(EU/T) ]T0

(4.7)

(4.8)

The term represents the marginal utility of time spent in consumption.
If C is fixed then the optimal allocation of working time between T and
L, given an interior solution (that is, the individual does not devote all of
his or her time to either illegal or legal activities), is represented by:
(WTWL)/(WTFWL)PU(Xa)/[(1 P)U(Xb)]

(4.9)

Derivatives are indicated by the prime () symbols. The left-hand side of
equation (4.9) is the slope of an opportunity boundary or the production
possibility curve of the composite X between the two states of the world
[(dXb/dT)/(dXa/dt)], as illustrated in Figure 4.5, while the right-hand side
is the slope of an indiﬀerence curve defined along dU* 0. Given that T, L
and C 0, the production possibility curve is defined only between A and
B in Figure 4.5. That is, A represents a point where the decision-maker
specializes in illegal activities while B is a point where no illegal activities
are pursued at all. C is the optimal allocation between T and L given the
indiﬀerence curve shown.
Like Becker’s model, Ehrlich’s framework does not produce unambiguous comparative statics results with respect to the relative eﬀectiveness of
the certainty versus severity of punishment. Consideration of the secondorder conditions by totally diﬀerentiating equation (4.9) demonstrates that
the equilibrium position of a risk-preferrer must be to the left of that of a
risk-neutral, and even further to the left of a risk-avoider, all else equal.

112

Economics uncut
Xa

W I+WT (kC)

Indifference curve

Certainty line

W I+WL (kC)

B
Production possibility
curve

45°
W I+WT

(kC)F (kC)

W I+WL (kC )

Source: Ehrlich (1973).

Figure 4.5

The allocation of time between legal and illegal activities

Furthermore, whether in equilibrium crime pays or does not pay in terms
of the expected marginal returns is simply a reflection of an oﬀender’s attitude toward risk, since in equilibrium the expected marginal returns from
crime would exceed, be equal to, or fall short of, the marginal returns from
legitimate activity, depending on whether the oﬀender is a risk-avoider,
risk-neutral or risk-preferrer, respectively. The deterrent eﬀect of an
increase in severity of punishment also exceeds or falls short of the deterrent eﬀect of a similar increase in the probability of punishment if the
oﬀender is a risk-avoider or a risk-preferrer, respectively.
The theoretical ambiguities arising from the Becker and Ehrlich models
motivated the development of several more theoretical models. For
instance, Block and Heineke (1975) show that if time spent in legitimate
and illegitimate income-generating activities is allowed to enter directly
into the utility function as arguments, then ambiguous comparative statics
result with respect to deterrence factors, despite assumptions on the degree
of risk preference of the agents. Thus, key policy implications cannot be
predicted by formal theoretical models and must be examined empirically
before their consequences can be anticipated. They also demonstrated that
the ambiguity associated with the relative eﬀectiveness of the certainty
versus severity of punishment in Becker’s model arises from the particular
notion of there existing a monetary equivalent of punishment, something

Economics of crime

113

these authors argue may not actually exist. As such, the distribution of risk
across potential oﬀenders does not allow one to make any inference about
the relative eﬀectiveness of deterrence factors.
Critics of this economic approach to crime contend that while the model
might make sense for economic crimes like robbery and burglary, the
so-called ‘crimes of passion’ such as murder, rape and assault do not generate monetary rewards and, therefore, cannot be analysed from an economic perspective. However, this objection arises from a misunderstanding
of the economist’s position, which is not simply that people attempt to maximize monetary wealth. The economic model can be applied to any situation where an individual is seeking to achieve his objectives (in addition to
the goods and services that can be obtained with monetary wealth, individuals may gain utility from power, security, prestige among peers, risky activities as risk-preferrers and so on), so the fact that crimes of passion often do
not involve transfers of money does not mean that they cannot be explained
by the economic model. Whether the model actually predicts the incidence
of crimes of passion can only be determined by empirical analysis.
3 Supply of offences and demand for crime prevention
A large empirical literature explores various determinants of crime. The
first empirical studies after Becker reintroduced crime to economists were
undertaken by Ehrlich. In his 1973 paper discussed earlier, he examined the
deterrent impact of the likelihood (and severity) of punishment on all seven
of the FBI index crimes and found that these crime rates vary inversely with
the probability of apprehension and punishment by imprisonment. Indeed,
the so-called crimes of passion (murder, rape, assault) responded just as
strongly to the expected costs of punishment as did property crimes.
3.1 Capital punishment
In this context, Ehrlich’s work (1975, 1977 and Ehrlich and Gibbons, 1977)
on the deterrent eﬀect of capital punishment is probably the most wellknown (and, therefore, most controversial) in the literature. In his 1975
paper, Ehrlich examined US murder and execution figures for the 1933 to
1969 period, together with measures of social factors such as unemployment and per capita income. His model revealed a statistically significant
negative relationship between the murder rate and the execution rate and
he noted, ‘in light of these observations, one cannot reject the hypothesis
that punishment, in general, and execution in particular, exert a unique
deterrent eﬀect on potential murderers’. In fact, Ehrlich suggested a
‘tradeoﬀ between executions and murders’, and estimated that over the
period studied, ‘an additional execution per year . . . may have resulted, on
average, in 7 or 8 fewer murders’.

114

Economics uncut

Ehrlich’s findings on capital punishment generated a storm of controversy concerning both his data and his econometric method (see Passell and
Taylor, 1977; Baldus and Cole, 1975; Bowers and Pierce, 1975; Peck, 1976;
Passell, 1975; Bailey, 1982; McGahey, 1980). However, as Yunker (1982)
reported after examining the empirical techniques and a priori hypotheses
tested in the deterrence literature, the ‘hypotheses relied upon by the
defenders of capital punishment possess greater inherent plausibility than
those relied upon by its opponents’.
Similarly, Rubin (1978) surveyed the literature and explained that for
research concluding that there was no deterrent eﬀect:
many of these studies were anecdotal in nature . . . But this kind of evidence
cannot prove anything. We need some sort of statistical study to determine the
true relationship. Many of the earlier sociological studies were statistical, but the
statistics were not very sophisticated . . . The kind of study done by Ehrlich . . .
is able to compensate for most diﬀerences that are thought to be significant. The
results of Ehrlich’s studies are very strong in indicating a deterrent eﬀect of
capital punishment.

Thus, the evidence provided by the critics of Ehrlich generally was more
questionable than Ehrlich’s techniques and data. Furthermore, economists
have continued to explore the issues using increasingly sophisticated statistical techniques. The general result is that additional empirical support for
most of Ehrlich’s findings has been produced. For example, Dezhbakhsh
et al. (2003), using US county-level post-moratorium panel data and a
system of simultaneous equations, found that each execution results, on
average, in 18 fewer murders, with a margin of error of plus or minus ten.
Mocan and Gittings (2003) merge a state-level dataset, which includes
crime and deterrence measures and state characteristics with information
on all death sentences imposed in the USA between 1977 and 1997 and
estimate that six murders are deterred per execution (see also Wolpin, 1978;
Horney and Ineke, 1992; Sollars et al., 1994; Lott, 1998; Benson and Mast,
2001, for other studies of the deterrence hypothesis).
However, Levitt (2004) suggests caution in the interpretation of such
figures for two reasons. First, while Figure 4.6 shows that the number of
executions has generally been increasing in the USA since the restoration
of the death penalty in 1976, this form of punishment is used only sparingly and the necessary legal process is often prolonged. According to
Levitt, therefore, the rational criminal presented in the previous section is
unlikely to be deterred by the threat of execution. Second, he argues that
even if we accept the validity of these empirical estimates of the deterrent
eﬀect of the death penalty, ‘the observed increase in the death penalty
from 14 executions in 1991 to 66 in 2001 would eliminate between 300 and

115
Number of executions in USA (1930–2004)

100

150

200

250

Bureau of Justice Statistics.

Figure 4.6

Source:

Number of executions
1930
1932
1934
1936
1938
1940
1942
1944
1946
1948
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004

116

Economics uncut

400 homicides, for a reduction of 1.5 percent in the homicide rate, or less
than one-twenty-fifth of the observed decline in the homicide rate over this
time period’. Perhaps it should be noted, however, that the modern data
involves very small probabilities of execution, so such extrapolations to
probability values outside the range of the data may be inaccurate.
While Ehrlich and other economists have found significant support for
the deterrence hypothesis, it does not follow that these economists advocate
capital punishment or severe criminal penalties. The deterrence hypothesis
is a positive implication of economic theory that empirical tests have
generally supported. Whether capital punishment is desirable or not is a
normative issue and the suggestion that it deters is not the same as suggesting it should be used. Figure 4.7 shows public opinion in the USA over
the post-1976 period for the death penalty has been overwhelmingly in
favour of its use. However, while those economists who support the death
penalty do so largely because they believe it has a deterrent eﬀect, Table 4.2
reveals that only a small percentage of US death penalty supporters (at
least for individuals convicted of murder) cite this reason for their stance.
According to a 2001 Gallup poll, almost half of those in favour of the
death penalty put forward the ‘eye for an eye’ argument.
Ehrlich himself has argued that the best way to reduce crime, including
murder, appears to be through the creation of economic opportunities.
Indeed, this probably continues to be the strongest implication of the
literature.
3.2 Extensions to the standard empirical model of crime
One criticism Ehrlich made of early statistical deterrence research was the
use of single-equation statistical models.2 This criticism is predicated on the
idea that crime rates aﬀect the demand for crime control and criminal justice.
For instance, voters are expected to demand and be willing to pay for more
criminal justice resources (police, prisons and so on) when crime rates are
high than when they are low, all else equal. The investment in criminal justice
resources in turn aﬀects crime rates via the probability of arrest and conviction and severity of punishment. Therefore, the supply of oﬀences and the
demand for criminal justice resources are determined simultaneously, and
failure to control for this will produce biased estimates of deterrence relationships. If high crime rates cause relatively high investments in policing, for
example, and relatively high investments in policing reduce crime rates, then
the coeﬃcient on police in a single-equation model explaining crime rates will
be biased downward. Similarly, individuals and organizations are more likely
to employ private security measures (alarms, guards, crime watch arrangements and so on) in the face of rising crime rates, all else equal, and such
private investments also may simultaneously reduce at least some crime rates.

117

Nov-72

Mar-72

Jan-69

Oct-71

Jun-67

May-66

Mar-60

Jan-65

Aug-57

Nov-53

Mar-56

Dec-37

Gallup, various years (www.gallup.com).

Aug-00

Feb-01

Feb-00

Jun-00

May-95

Feb-99

Sep-94

Sep-88
Jun-91

Jan-86

Nov-85

Jan-85

Mar-78

Jan-81

Apr-76

Figure 4.7 US public opinion on death penalty (selected years, 1937–2003)
(Are you in favour of the death penalty for a person convicted of murder?)

Source:

May-03

May-02

Oct-02

Oct-01

May-01

For
Against
No opinion

118

Economics uncut

Table 4.2 US death penalty supporters’ reasons for favouring policy
(1991–2001) (Why do you favour the death penalty for persons
convicted of murder?)

Eye for an eye/punishment fits crime
Save taxpayers money/prison costs
Acts as deterrent/sets an example
They deserve it
Support/believe in death penalty
Depends on type of crime committed
To keep them from repeating crime
Biblical reasons
Relieves prison overcrowding
If no doubt they committed crime
Life sentences don’t always mean life
They cannot be rehabilitated
To serve justice
Fair punishment
It benefits families of victims
Other
No opinion

Feb.
2001 %

Feb.
2000 %

Jun.
1991 %

48
20
10
6
6
6
6
3
2
2
2
2
1
1
1
3
1

40
12
8
5
0
6
4
3
0
0
0
1
3
6
0
10
3

40
12
8
5
0
0
4
3
0
0
0
1
2
6
0
10
3

Source: Gallup, various years (www.gallup.com).

In response to the above, the standard empirical model has evolved over
time to consist of a set of simultaneous equations explaining the crime
rate(s), the probability(ies) of arrest and a measure(s) of deterrence resources
(for example, police employment or budget) as dependent variables
(for reviews, see Cameron, 1988; Benson, et al., 1994). The generic empirical
model of crime that developed in the literature assumes that the crime rate
aﬀects the resources available to the police (that is, voters’ willingness to pay
for criminal justice resources such as policing and prisons is a function of the
level of crime as well as other factors), which in turn aﬀects the crime rate via
the deterrence eﬀects of the probability of arrest (and perhaps the severity of
punishment). The standard equations of this model are:
CR f(PA, SEV, SOCEC, COM1)

(4.10)

PAg(POL, COM2)

(4.11)

POLh(CR, COM3)

(4.12)

Economics of crime

119

where CR is the jurisdiction-level crime rate being studied during a particular time period (some models have considered more than one crime rate
simultaneously), PA is the probability of arrest in the jurisdiction in that time
period (or perhaps in the previous time period, assuming that potential criminals base their expectations on the probability of arrest on past data), POL
represents the jurisdiction’s police resources in the relevant time period, and
SEV is the severity of the sanction for the crime during the time period, which
can include the probability of conviction if arrested as well as the average
sentence (SEV may also be a function of crime rates, of course, since expenditures on punishment resources are also determined in the political arena;
thus, additional equations may be added to account for the endogeneity of
SEV). SOCEC is a vector of variables that reflect the socio-economic characteristics of the jurisdiction during the time period being studied. Factors
such as income, unemployment, educational attainment and racial characteristics of the population are control variables or other sources of incentives
and constraints for the ‘rational’ decision-makers who are assumed to
respond to deterrents. COM1, COM2 and COM3 are vectors of other community characteristics that may influence crime, the probability of arrest and
the demand for policing (for example, tax capacity, income and demographic
factors). Some recent work suggests that an equation or equations to control
for the demand for private security may also be important to include since
private security measures may be either substitutes for or complements to
public crime control eﬀorts (see Benson, 1998; Benson and Mast, 2001).
While simultaneity bias can be an issue in some crime studies, both
empirical and theoretical findings suggest that it may not be appropriate to
model some individual crime rates in simultaneous equation models. After
all, while the aggregate crime rate might influence the demand for policing
(as well as punishment and/or private security) resources, individual crime
categories may not, particularly if they constitute only a small part of
overall crime activity. In this context, for instance, there is little empirical
evidence of a connection between violent crime rates and police resources
as implied by equation (4.12). More property crime apparently increases
the demand for police resources, but violent crime is not a significant determinant of police resources.
Benson et al. (1992) provide direct evidence of this in a study of the
determinants of police non-capital expenditures in Florida. They report
that these outlays are significantly and positively correlated with the property crime rate (with an estimated elasticity ranging from 0.46 to 0.61).
The eﬀect of the violent crime rate on police expenditures, in contrast, is
generally not significant and has an estimated elasticity between 0.04
and 0.003 (see also Avio and Clark, 1976; Hakim et al., 1979; Sollars
et al., 1994; Benson et al., 1994).

120

Economics uncut

Second, a common finding in the literature is that a higher probability of
arrest has a deterrent value but that the marginal increments in police
resources do not produce a higher probability of arrest as implied by equation (4.11).3 This result is not too surprising given that the property and
violent crimes that are studied in these models often account for only a
small portion of total police activity. Increases in police resources need not
increase deterrence of any specific crime because those resources can be
allocated to other activities. Thus, two of the three links required for a
simultaneous model of at least some crime rates and the probability of punishment are very weak or non-existent: police apparently do not receive
greater budgets with higher violent crime rates, although they do with
higher property crime rates, and police discretion in the allocation of
resources reduces the likelihood that marginal increments to police
resources will alter the probability of being arrested for at least some
crimes. Whether simultaneous equations should be used or not is, therefore,
an issue that has to be considered in the specification of empirical models,
using endogeneity tests.
The deterrence relationship is generally found by using a reported
oﬀence rate as a dependent variable, and the arrest/reported oﬀence ratio
serves as a proxy for the probability of arrest. Therefore, reported oﬀences
are explained by arrests multiplied by the inverse of reported oﬀences, and
critics of the literature note that the negative relationship could simply be
a product of spurious correlation resulting from this measurement error.
For instance, Brier and Fienberg’s (1980) influential review (reinforced by
others), concludes that there is, ‘no reliable empirical support in the existing econometrics literature either for or against the deterrence hypothesis’.
In light of such criticisms, Levitt (1998) developed a model for determining the extent of measurement error. Tests of his model for the seven
major Index I crime categories conclude that measurement error biases are
likely to be relevant in only one of them: auto theft. Thus, Levitt’s results
suggest that a study focusing on the violent crime categories (murder, sexual
oﬀences, assault, robbery) or on most property crimes (burglary, larceny)
will not suﬀer from such problems. In this context, a fixed-eﬀects model,
which is designed to deal with unobserved heterogeneity among observations in time-series cross-section pools of data, is also suggested by Levitt’s
findings, and this empirical device has been employed in most of the recent
econometric studies of crime. By including a dummy variable for each jurisdiction, unobservable factors that do not change over time can be controlled for. After all, such unobserved heterogeneity is characteristic of
jurisdictions since communities with very similar socio-economic and
demographic characteristics have very diﬀerent crime rates. Further,
importantly, if crime reporting by victims in a jurisdiction is relatively

Economics of crime

121

constant over time, the use of fixed-eﬀects models also attenuates another
measurement error problem that has plagued this literature: the widely recognized inaccuracies in the number of oﬀences known to the police arising
due to the uneven reporting of victims and police departments. Thus,
empirical testing of the hypotheses generated by the economic theory of
crime is becoming increasingly sophisticated in order to deal with data limitations, but as the empirical literature expands and becomes more sophisticated, the robustness of certain key results suggested by Ehrlich’s studies
is becoming increasingly clear.
It can be concluded with considerable confidence that the opportunity
cost of committing crime matters. In a strong economy with substantial
opportunities for employment and high returns to labour, crime rates are
relatively low, all else equal. It is also clear that the crime rates are negatively related to the probability of punishment. As the probability of punishment increases, crime rates fall, ceteris paribus. On the other hand,
empirical inferences of the deterrence eﬀect of the severity of punishment
are less robust. Some studies, such as Ehrlich’s capital punishment research
and others cited above, find that relatively severe punishment deters crimes,
while other studies using diﬀerent data or specifications find that this is not
the case. This should not be surprising, nor should it lead to the rejection
of the deterrence hypothesis for at least three reasons.
First, the severity of punishment is a subjective concept. Some people
may find that capital punishment is more severe than life in prison without
parole, for instance, while others may feel that the opposite holds.
Second, non-measurable social sanctions (for example, loss of reputation, social ostracism, the diﬃculty that accused or convicted criminals
have in getting a legal job or finding contractual partners) may be much
more ‘severe’ for some types of crimes than the relatively more objectively
measurable sanctions imposed by the criminal justice system. This is likely
to be particularly true for individuals with significant opportunity costs in
the form of high income-earning possibilities that would be sacrificed if the
person is arrested (see Lott, 1987). These social sanctions may be important deterrents but because they cannot be measured, the deterrence eﬀect
cannot be observed. However, they may aﬀect interpretation of empirical
results using measurable punishment. For instance, if severity is measured
by length of sentence, and unobserved social sanctions are roughly constant, then real punishment is the length of sentence plus a constant.
Doubling the length of sentence will not double real punishment under
these circumstances, so empirical results could suggest that criminals are
risk-preferrers even if they are not.
Third, sanctions do not matter if the probability of punishment is
extremely low. In fact, there is no reason to even learn about what a

122

Economics uncut

punishment may be if one believes that the chances of being punished are
trivial. A low probability of punishment might be anticipated because an
individual is a very good criminal, because of misinformation, because
victims choose not to report most of the crimes of a particular type (for
example, small thefts), or because very few scarce policing resources are
allocated to control a particular crime.
4 The allocation of criminal justice resources
Most resources are scarce, including those allocated by the public sector.
Thus, many public sector decisions, including those involved with the criminal justice system, essentially are resource allocation problems. Legislators
(national or local, depending on where a budget decision is made) must
decide how much funding to allocate to purchase resources for law enforcement (for example, police and policing capital like patrol cars; courts and
prosecutors; prisons, parole systems) given alternative uses of tax revenues
(for example, education, highways). Budgets for the purchase of law
enforcement resources must then be allocated among competing uses. For
example, police resources can be allocated to the control of drug markets,
prostitution markets, illegal gambling; solution of robberies and burglaries; solution of rapes, assaults and murders; traﬃc control; crime prevention eﬀorts and various community service and information services.
If there is political demand to increase eﬀorts against one particular type
of crime, there are two possible responses. First, a legislature may allocate
more total expenditures to law enforcement so that more can be allocated
to the control of that crime without sacrificing the current police eﬀorts
against other crimes. This either requires increased taxes or reduced
expenditures for other publicly provided goods and services. Therefore, it is
more likely that competing political demands will have to be taken into
account and the resulting compromises will mean that at least part of the
response will take a second form: the shifting of some law enforcement
resources away from other uses. To see the consequences, let us first examine
the underlying mechanisms that drive the allocation of existing law enforcement resources.
4.1 The tragedy of the criminal commons
When resources are scarce in the sense that the competing demands for
their use far exceed their supply, they must be rationed among these competing uses. Many rationing techniques are possible. For instance, in a free
market where resources are privately owned, those resources (as well as
goods and services) are rationed by price: anyone willing to pay the marketdetermined price has access to the desired resources. Public law enforcement resources are not rationed by price, of course. Indeed, some would

Economics of crime

123

suggest that they are not rationed at all since everyone supposedly has
access to them free of charge. But the fact that no money price is charged
for common access resources simply means that some other rationing criteria must arise (see Chadwick, 1929; Shoup, 1964).
The implications of common access, or common pool resources, have
been examined extensively in relation to publicly or commonly owned
resources such as grazing land and fishing grounds, but the similarity
between the problems that arise in grazing or fisheries commons and in law
enforcement are striking. In the classic commons problem, for instance,
when many individuals have free access to graze cattle on the same land,
each has incentives to use up as much grass as possible before others with
access do the same.
Free access public law enforcement resources are similarly crowded and
ineﬃciently employed. For example, prosecutors and judges have common
access to prison space. They have at best only weak incentives to limit the
number of prisoners they herd into the commons when making their
sentencing recommendations and decisions, or to consider alternatives to
imprisonment such as fines, restitution, community service, lesser degrees
of constraint like community control, treatment programmes and work
release. Many such alternatives are well known to prosecutors and judges,
and some are frequently employed, but the incentives to consider them
are relatively weak because these decisions are made in a common pool
environment.
Indeed, in as much as prosecutors have incentives to demonstrate to their
local constituencies that they are tough on crime, imprisonment is a relatively attractive punishment. Even if they recognize that their actions add
to the crowding problem, their private benefits (the political support they
get from their tough image) may exceed their private costs (perhaps the
anxiety associated with the recognition that they are crowding prisons and
raising costs to society at large). Thus, imprisonment is chosen relatively
frequently and the eﬀect is that prosecutors and judges as a group crowd
the common access prisons much as cattle owners crowd common access
grazing land.
State legislators in the USA can also crowd the criminal justice commons
since, like judges and prosecutors, they have weak incentives to conserve
scarce prison resources. When passing laws that increase the penalties for
certain crimes, such as mandatory minimum sentences, the legislature’s
action is equivalent to increasing the number of people who can use the
classic grazing commons. Of course, the legislature also has the power to
oﬀset the resulting overcrowding since it can increase the number of prison
beds, the equivalent of increasing the common land. But the temptation to
overcrowd the commons is likely to prevail, since individual legislators are

124

Economics uncut

unlikely to suﬀer any costs from this course of action, particularly in the
short run. Legislators can reap political benefits by passing longer sentences for crimes, so they appear to be tough on criminals, while ignoring
the fact that the law can undermine other aspects of the criminal justice
system. Since expanding criminal justice resources involves the politically
unpopular task of either cutting other government functions or raising
taxes, the politically astute course of action is to crowd the commons.
Crowding occurs at all levels of the criminal justice system. In fact, one
reason for judges’ and prosecutors’ failure to consider alternatives to
imprisonment is that the courts themselves are severely crowded, implying
that judges may not have time to explore alternative punishments at length
(Neely, 1982; Benson, 1990).
Judges must also consider other crowded conditions in their sentencing
decisions as well. Probation is the most commonly used alternative sanction to imprisonment, for example, and the typical sanction for first-time
oﬀenders. In fact, roughly three times as many convicted oﬀenders in the
USA are placed on probation each year as are sentenced to prisons and jails
combined. But as Byrne (1988) noted, ‘Although prison crowding draws
national attention and increased resources, “probation crowding” poses a
more immediate threat to the criminal justice process and to community
protection’. The probation population in the USA has been increasing at
roughly the same rate as the prison population, severely taxing the ability
of the system’s limited resources to monitor and supervise probationers.
All alternatives to imprisonment, other than forgiveness and release,
require the use of some scarce resources for monitoring, so the potential for
crowding arises even if judges do not herd criminals into the common pool
prisons. Scarce police resources also see many reported crimes crowded out
of the system.
But crowding is only one aspect of a commons problem. Indeed, even
though crowding may be the most visible consequence of common access,
it is not the most significant consequence. When grazing land is over-used,
its rapid deterioration in quality means that the output of the production
process that employs that land similarly deteriorates in quality. The same
problem results with common access law enforcement resources. Prisons,
for instance, are intended to serve several functions (produce several
outputs), including punishment of convicted criminals, deterrence of
potential criminals, incapacitation of dangerous criminals, and rehabilitation of criminals who can be reformed. All of these outputs of the prison
system decline in quality because of the commons problem and crowding
(crowding could make prisons less pleasant, thereby increasing punishment
and deterrence, but in the USA, for instance, prison systems have been
forced by federal court order to alleviate crowding, and this is generally

Economics of crime

125

done by releasing prisoners early, thereby holding the unpleasantness of
prison roughly constant while reducing time served). Early release programmes mean that criminals are not punished to the degree that judges,
victims and others in society feel justice requires. Potential criminals recognize that prison sentences are rarely fully served and that the portion of
sentences served declines with increased admissions after capacity is
reached, so deterrence may be diminished. Some non-rehabilitated and
insuﬃciently deterred criminals are released early, despite the need for continued incapacitation, and commit more crimes. Finally, when prison
budgets are consumed by eﬀorts to accommodate overflowing prison populations, resources may not be available for rehabilitation programmes; and
if they are, early release may mean that criminals are not enrolled in rehabilitation programmes long enough to benefit from them.
Individual legislators, prosecutors and judges who make the decisions
that produce crowded prisons are not liable for the costs of their decisions.
Instead, costs are imposed on taxpayers who face the ever-increasing cost
of building more prison space; on victims who are not satisfied with the
level of punishment oﬀenders actually incur; on citizens who are victimized
by prematurely released criminals and other criminals who are not
suﬃciently deterred; on criminals, who may have reduced opportunities for
rehabilitation, face the increased violence and abuse that may accompany
prison crowding, and endure additional social sanctions upon release
because people do not feel that they have been suﬃciently punished; and on
corrections oﬃcials who must focus their eﬀorts and resources on coping
with crowded conditions rather than on other functions, such as rehabilitation and treatment.
Common access really means that many public law enforcement
resources are rationed on a first-come-first-served basis. The first case of a
certain type filed gets the first available slot on the court docket, for
instance, and the first prisoner of a particular type sentenced gets the first
open space of that type in the prison system. But rationing by first-comefirst-served generally means rationing by waiting in queues, as backlogs of
unmet demands build up. Thus, as noted above, there is a backlog of court
cases waiting to be heard and county jails in the USA fill with prisoners
waiting to be tried or to be placed in state prisons. In fact, 52 per cent of
the prisoners held in local jails during 1987 were queued up, awaiting
arraignment or trials, and a substantial portion of the remaining 48 per
cent, who were convicted inmates, were being held until they could be transferred to another authority (US Department of Justice, 1988). Similarly,
backlogs of thousands of supposedly open police cases receive little or no
attention, waiting forever to be solved. In Tallahassee, Florida, for instance,
there were approximately 15 900 crimes reported to police in 2001 that

126

Economics uncut

would require investigation to be solved. Of those, roughly 7000 had no
obvious leads for investigators to follow, so no investigator was even
assigned to the cases. Furthermore, another 4000 cases that actually had
leads that could have been followed were not assigned to an investigator,
probably because they were not considered to be important enough by the
police to warrant attention. Only 5900 of the reported cases were assigned
to investigative personnel.
When rationing by waiting becomes prevalent and the time cost of
waiting grows, many potential demanders opt out of the queue and choose
not to be served. Consequently, prosecutors and judges must expedite
waiting cases by plea bargaining. Furthermore, many crime victims choose
not to report crimes, either because of the low probability of satisfaction
(since police facing excess demands for their services will likely be unable
to solve the crime; plus the likelihood that even if the police are successful,
prosecutors will drop the case or plea bargain it down to a level of punishment that is unsatisfactory from the victim’s perspective) or because of the
time cost associated with cooperating in prosecution. Annual Victimization
Surveys conducted by the US Department of Justice suggest that over
60 per cent of the crimes against persons and property are never reported
to the police.
Other rationing mechanisms also apply when there is excess demand. For
one thing, suppliers generally have considerable discretion when demand
exceeds supply. They can discriminate among demanders, choosing to serve
some but not others. In light of the numbers reported above, for instance,
the Tallahassee Police chose which crimes to investigate and which to
ignore. These choices may reflect many factors, such as police oﬃcials’ individual subjective evaluations of the relative merit of diﬀerent crimes, political pressures (neighbourhoods with large numbers of wealthy and
politically influential residents may get more patrol cars per capita and
more rapid response time than those with large numbers of poor and politically powerless residents), and the self-interest motives of police. The allocation of police resources across the various crime categories reflects such
decisions and essentially divides the larger common pool into smaller ones.
The decision of a patrol oﬃcer to give one driver a warning while giving
another a traﬃc ticket for the same kind of infraction (for example, because
one driver was more contrite, friendlier, more attractive, politically influential, a friend) is another example.
4.2 Allocation of resources and probability and severity of punishment
The allocation processes in the criminal justice system determine the probability and severity of punishment for diﬀerent crimes, of course.
Continuing with the supply and demand analogy, these allocation decisions

Economics of crime

127

result in a set of ‘expected prices’ of crimes in the form of expected punishments that the criminal justice system sets for various criminal acts.
According to the economic theory of crime outlined in Section 2, potential
criminals compare these expected prices to the benefits they anticipate from
criminal acts as they make their decisions to commit crimes.
The expected price of crime is not the actual sentence given to convicted
criminals. Rather, the expected price is determined by the probability of
punishment times the actual or average punishment. The probability of
punishment involves a series of uncertain events. First, the crime or its consequences must be observed. When a crime has a victim this is likely to
occur, of course, but the ‘quality’ of observations can also vary, thus influencing the probability of successful prosecution (for example, whether the
criminal is likely to be observed in the act, producing an eye witness, or
whether the crime will only be discovered after the fact, thus requiring
stronger physical evidence for conviction). With consensual crimes like
drug exchanges, gambling and prostitution, such observation is much less
likely, of course, as it requires either that police see the transaction or that
some potential reporting witness does. Not surprisingly, these crimes are
often much more diﬃcult and costly to enforce.
If the criminal act or its consequences are observed, the observation must
then be reported to police, unless the observer is someone in a position to
act on the observation (for example, police or private security oﬃcer). As
indicated above, victims actually observe many crimes that they do not
report. With consensual crimes, the problem is obviously even more significant, as witnesses almost never report them unless they are ‘snitches’ who
exchange ‘evidence’ for money or some other favour, such as freedom
from prosecution for crimes they have committed. Third, the criminal who
committed the reported crime must be arrested. Fourth, the oﬀender must
be charged and prosecuted after being arrested. Fifth, the prosecution
must be successful before a sanction is imposed. Given a conviction, the
criminal can be sanctioned in one of a number of alternative ways, as suggested above. Possible punishments include, in the USA, the death penalty
(a highly unlikely outcome except for the most heinous violent oﬀences),
prison terms, supervised release including probation, immediate release by
being sentenced to time already served in jail, or a fine. Socially imposed
sanctions are real alternatives that should also influence criminal decisions,
but here the focus is on the ‘price’ of (expected punishment for) crime
imposed by the public sector criminal justice system.
There is no way to determine what portion of crimes that are committed
are actually observed, so the first probability listed above cannot be estimated. There is evidence regarding the other steps in the process, however.
According to the US Department of Justice, Bureau of Justice Statistics’

128

Economics uncut

Table 4.3

Expected prison time for Florida criminals in 1992

Crime

(PVR)

(PA)

(PP
PC
PPS)

1.0
0.53
0.70
0.60
0.36
0.92

0.68
0.56
0.25
0.15
0.18
0.15

0.52
0.52
0.25
0.30
0.14
0.14

Homicide
Rape
Robbery
Burglary
Larceny
Auto theft

(APS)
(years)
21.8
12.9
8.6
5.5
4.1
4.1

(PSS)
(%)

Expected
Prison
Time

38.1
45.8
44.9
37.3
33.6
33.6

2.99 years
338 days
61.3 days
13.2 days
4.5 days
9.8 days

Victimization Survey for 1992 (this year is chosen in order to use it with
other data from that year, as explained below), only about 39 per cent of all
Index I crimes were reported in the USA, although reporting varies by
crime type from 92 per cent for auto theft to 15 per cent for crimes of
larceny resulting in losses of less than $50. Using Florida data, reported
crimes cleared by arrest ranged from a high of 68 per cent for murder to a
low of 15 per cent for motor vehicle theft and burglary in 1992.4
In Table 4.3, PVR is the probability of victim reporting from the
Department of Justice’s Annual Victimization Survey. Victims cannot report
that they are victims of homicides, of course, so it is simply assumed that the
probability that homicides are reported/discovered is 1.0. PA represents the
probability of arrest given that a crime is reported, which is the clearance rate
reported by the Florida Department of Law Enforcement (FDLE). PP is the
probability of prosecution given arrest, PC denotes the probability of conviction given prosecution, and PPS is the probability of a prison sentence
given conviction. These probabilities are combined into a single probability
of imprisonment given arrest because the arrest data come from the FDLE
while the prosecution and conviction data are from the courts and the imprisonment data are from the Florida Department of Corrections (FDOC), and
the courts’reporting categories do not match with the FDLE and FDOC categories except for Robbery and Burglary. Therefore, the individual probabilities can only be directly estimated for those crimes, while the combined
probabilities can be estimated for the others. APS represents the average
prison sentence given imprisonment from the FDOC. PSS denotes the
average portion of a sentence that is served, also from the FDOC.
The numbers in Table 4.3 might be thought of as approximations of
the expected price (that is, probability-weighted prison term) the criminal
justice system is charging for committing these crimes, assuming that
imprisonment is the punishment that criminals consider to be most severe.

Economics of crime

129

Violent crimes against persons have relatively high expected penalties, compared with property oﬀences, but all of the probability-weighted prison
terms for most crimes committed in Florida were very low, and not surprisingly, Florida had the highest crime rates of all states in the USA at the time.
Clearly, the allocation of criminal justice resources matters, as it influences the decisions to commit various kinds of crimes. As noted in
Section 1, crime rates have been falling in the USA in the 1990s, and one
explanation is that the rapid increase in the allocation of criminal justice
resources devoted to drug enforcement ended in the early 1990s. Other
explanations, to which we now turn, are consistent with the economic
theory of crime and can also be cited.
5 Explanations for the 1990s’ fall in US crime rates
As Table 1.1 shows, the FBI All Index Crime Rate was 5820.3 per 100 000
population in 1990. The crime rate dropped continuously through the
1990s, reaching 4124.0 in 2000. Why? Politicians have ready answers as they
claim credit through their support of prison construction, longer mandated
sentences, greater police funding and so on. They also claim credit for the
strong economy and low levels of unemployment, which reduce the incentives to commit property crimes. Criminologists cite many of these same
causes as well, along with the changing age distribution of the population,
as the size of the crime-prone cohort of young males has been shrinking
due to ageing and reduced birth rates. In fact, as discussed in Chapter 11,
Donohue and Levitt (2001) oﬀer evidence that legalized abortion is an
important factor in explaining these reductions in crime.
5.1 The impact of a growing economy
The growing economy that the USA enjoyed through the 1990s had two
important impacts. First, as Ehrlich and many other economists have
stressed, crime rates fall as opportunities for legal employment improve and
wages rise. This may well be the most important factor explaining the
decline in crime during the period, although Levitt (2004) points out that
the 1960s was also a decade of strong economic growth but large increases
in crime were experienced. (Gary Becker also does not believe the strong
economy to be an important factor as his interview in this book reveals.)
Second, the rapid growth in income and wealth of the 1990s meant that
governments saw tax revenues increasing substantially without tax rate
increases. Federal, state and local legislators could increase spending on
many things, including law enforcement, and they did so. The resulting
prison construction (along with the growing funding and use of diversionary programmes for drug oﬀenders) meant that prisoners now serve
longer sentences and/or more criminals are sentenced to prison relative to

130

Economics uncut

probation and other alternatives, all else equal. Thus, deterrence may be
more eﬀective due to the increased severity of punishment. Greater police
funding means police can make more arrests. Of course, whether this deters
Index I crimes or not depends on how police resources are allocated.
As noted earlier, drug enforcement declined in the early 1990s, releasing
policing (and prison) resources for other uses. With the addition of more
resources in the mid-to-late 1990s, state, local and federal government
increased spending on policing, so deterrence eﬀects became stronger. In
fact, these growing law enforcement budgets ultimately allowed the police
to simultaneously increase drug enforcement and enforcement eﬀorts
against other crimes. For example, drug arrests in Florida reached 126 087
in 2002, up from the 1997 level of 81 846. This increase in the late 1990s and
early 2000s may help explain the fact that crime rates appear to be levelling
out. The 2002 drug arrest rate per 10 000 population (76) in Florida actually
exceeded the 1989 rate (66.83), for instance.
5.2 The rise in private protection
There is another potential explanation for falling crime rates that has
gone largely unnoticed. As early as 1968, Gary Becker (1968) observed,
‘A variety of private as well as public actions also attempt to reduce the
number and incidence of crimes: guards, doormen, and accountants are
employed, locks and alarms installed, insurance coverage extended, parks
and neighbors avoided, taxis used in place of walking or subways, and so
on’. Three decades later, private citizens in the USA have responded to the
fear of crime by investing increasing amounts of their own time and money
in crime prevention, including crime watch and other types of neighbourhood or building watching, patrolling and escort arrangements, installation
of alarms and other detection devices, improved locks and lighting, investments in self-protection such as martial arts training and guns, and employment of private security personnel.
Information on these activities is relatively scarce, but a number of
studies have been conducted over the years that indicate dramatic increases
in the employment of most of these resources and services (Benson, 1998).
Private security employment was roughly equal to public police personnel
in 1970, for instance, but it is estimated that there are approximately three
times as many private security employees today than there are public police.
Spending on private security is estimated to be more than three times the
level of spending on the entire public sector criminal justice system
(Philipson and Posner, 1996) and this does not include the opportunity
costs for individuals who alter their behaviour (for example, give up opportunities such as travel, social activities) to avoid becoming a crime target,
costs that are ‘likely to be substantial’ (Ayres and Levitt, 1998).

Economics of crime

131

It is not just numbers and expenditures that are rising. Both increasingly
sophisticated labour and capital are being combined to produce ever higher
levels of security where it is demanded. Indeed, some technological innovations that probably were not even motivated by fear of crime, such as the
widespread ownership and use of cell phones, have probably had dramatic
impacts on the incentives to commit crimes, as they have lowered the cost
and increased the speed of reporting crimes. Many innovations have been
directly motivated by a demand for crime control, however. The increasing
level of technological sophistication for relatively low cost home alarm
systems is really quite amazing, for instance, as technologies that were on
the cutting edge only a short time ago are obsolete. Such technologies also
lower the cost of observing and reporting crimes. Training for security personnel has had to improve dramatically in order to take advantage of the
new technologies. Security personnel still do include minimum-wage night
watchmen that are the stereotype of security from the 1950s and 1960s, of
course, but they also include highly trained and skilled electronic-security
experts and security design consultants, as growing demand for detection
and deterrent equipment and technological advances that are lowering
the price of such equipment stimulate demand for more educated, skilled
and specialized private security labour while reducing the demand for
unskilled watchmen.
Studies of the consequences of these private sector crime control activities are also scarce. However, several informative studies do exist. For
example, Ayres and Levitt’s (1998) empirical analysis of the impacts of
Lojack, a hidden radio transmitter installed in cars, which can be remotely
activated if the car is stolen, concluded that a 1 percentage point increase
in installations of Lojack in a market is associated with a 20 per cent decline
in auto thefts within large cities and a 5 per cent reduction in the rest of the
state. These devices greatly reduce the expected loss for car owners who use
them, since 95 per cent of the cars equipped with Lojack are recovered compared with 60 per cent for non-Lojack equipped cars, but this direct benefit
for the individual using the device clearly is only part of the total benefit
arising from its availability and use because a potential car thief does not
know whether a potential target vehicle is protected by Lojack or not. This
uncertainty acts as a general deterrent against auto thefts, and, ‘Lojack
appears to be one of the most cost-eﬀective crime reduction approaches
documented in the literature, providing a greater return than increased
police, prisons, job programs, or early education interventions’ (Ayers and
Levitt, 1998). Lott’s (1998) controversial work on concealed handguns
reaches similar conclusions, finding that violent crimes, including murder,
rape and robbery, are significantly deterred when potential criminals know
that citizens are allowed to carry concealed handguns.

132

Economics uncut

Similarly, Zedlewski’ s (1992) comprehensive statistical study used data
from 124 Standard Metropolitan Statistical Areas in an eﬀort to analyse
the eﬀect of both public police and private security on the overall safety
environment of communities and on oﬀender decision-making. Greater
levels of security personnel were associated with reduced levels of community crime, suggesting that investments in private security produce spillover
benefits to the community at large. More recently, Benson and Mast (2001)
employed a large cross-section time-series pool of data to consider the
impact of private security on crime. They found robustly significant and
negative relationships between rapes and private security, as well as some
but less robust support for general deterrence of robbery.5 Thus, there is
growing evidence that the growth in private security helps explain the
decline in crime rates through the 1990s.
Even though there have been many changes during the 1990s, the economic theory of crime probably can explain the dropping crime rates
enjoyed in the USA during that decade. They generally are factors that
increased legal income-earning opportunities and/or increased the probability or severity of punishment. Looking ahead, Levitt (2004) identifies
the coming of age of so-called ‘crack babies’ in the USA and ‘those who
have spent their childhood years in families and neighborhoods ravaged by
crack’ as factors that could potentially contribute to increased crime rates
over the next decade due to the ‘criminogenic eﬀect’, although he suggests
that the overall downward trend in crime rates experienced in the 1990s
should continue.
6 Conclusion
This chapter has shown that economists have long harboured criminal
interests. Our principal focus has been the examination of two specific
contributions: first, the theory and empirical evidence relating to the ‘deterrence hypothesis’, and second, the analysis of resource allocation within
the public sector criminal justice system. One topic for economists to
further investigate is the impact of technology improvements on crime. In
Section 5, we suggested that advances in alarms and other detection devices
(together with their increased use) may also have contributed to falls in
crime in the USA. Yet, it may also be true that the recent introduction of
other technology such as the Internet has led to crime creation. This is particularly relevant to two topics discussed later in this book: pornography
and music. In the former case, there are concerns that the Internet may have
driven a rise in child pornography crimes (see Chapter 6 for an analysis of
the impact of the Internet on porn consumption in general and how police
authorities can enforce laws in this area). Meanwhile, Chapter 14 examines
the extent to which the Internet has stimulated music piracy.

Economics of crime

133

Overall, one would hope that public policy on crime can be influenced by
research carried out by economists. Indeed, it is almost ten years since
DiIulio (1996) appealed to the economics profession for ‘help’ with matters
of criminal justice, while simultaneously lamenting the fact that:
many leading experts and policymakers doubt that economic perspectives on
crime have much to oﬀer, see zero value in attempts to model the conditions
under which given public policies can cut crime, and disdain eﬀorts to measure
(or, heaven forbid, quantify) the social costs and benefits of competing crime
policy options.

Of course, our examination of the impact of Isaac Ehrlich’s work in the
1970s on capital punishment suggests this statement may not be universally
true and certainly there are signs over the past several years that the work
of economists such as Steven Levitt at Chicago is stimulating policy debate.
Armed with increasingly powerful tools to understand individual behaviour, the aggregation of individual choices and actions, and to establish
causality, we conclude that economists are comfortably placed to make significant contributions well into the twenty-first century.
Notes
1.
2.
3.
4.

Becker (1968).
Ehrlich made this point in his 1973 article, and employed simultaneous equation
techniques, but some economists continued to employ single-equation models (see
Sjoquist, 1973).
This is emphasized in Cameron’s (1988) critical review of the economics of crime and
the deterrence hypothesis, but for an alternative interpretation that focuses on the allocation of policing resources, see Benson et al. (1994) and Benson et al. (1998).
1992 data were gathered by one of the authors for a report prepared for the Florida
Chamber of Commerce, so these data are employed here. Compiling the data is costly
because much of it is not published and it must be obtained from several diﬀerent sources
(for example, the Florida Department of Law Enforcement, the US Department of
Justice, the Florida Department of Corrections and the Florida Supreme Court).
Furthermore, the data are not categorized in the same way by each agency so data must
be reclassified or used in various estimation techniques in order to perform the calculations presented in the table.
These results are preliminary since the purpose of the analysis is to test the robustness
of the Lott model, not to present a fully developed model of the influence of private
security on crime rates.

References
Avio, K.L. and C.S. Clark (1976), Property Crime in Canada: An Econometric Study, Toronto:
University of Toronto Press.
Ayres, I. and S.D. Levitt (1998), ‘Measuring Positive Externalities from Unobservable
Victim Precaution: An Empirical Analysis of Lojack’, Quarterly Journal of Economics,
113 (1), 43–77.
Bailey, W. (1982), ‘Capital Punishment and Lethal Assault Against Police’, Criminology,
19, 608–25.

134

Economics uncut

Baldus, D. and J. Cole (1975), ‘A Comparison of the Work of Thorstein Sellin and Isaac
Ehrlich on the Deterrent Eﬀect of Capital Punishment’, Yale Law Review, 85, 170–86.
Barclay, G. and C. Tavares (2003), ‘International Comparisons of Criminal Justice Statistics’,
Home Oﬃce, Issue 12/03.
Becker, G.S. (1968), ‘Crime and Punishment: An Economic Approach’, Journal of Political
Economy, 76 (2), 169–217.
Benson, Bruce L. (1990), The Enterprise of Law: Justice Without the State, San Francisco:
Pacific Research Institute.
Benson, B.L. (1995), ‘Understanding Bureaucratic Behavior: Implications from the Public
Choice Literature’, Journal of Public Finance and Public Choice, 8 (2–3), 89–117.
Benson, Bruce L. (1998), To Serve and Protect: Privatization and Community in Criminal
Justice, New York: New York University Press.
Benson, B.L. and B.D. Mast (2001), ‘Privately Produced General Deterrence’, Journal of Law
and Economics, 44 (2), 725–46.
Benson, B.L. and D.W. Rasmussen (1996), ‘Predatory Public Finance and the Origins of the
War on Drugs: 1984–1989’, The Independent Review: A Journal of Political Economy, 1 (2),
163–89.
Benson, B.L., I. Kim and D.W. Rasmussen (1994), ‘Estimating Deterrence Eﬀects:
A Public Choice Perspective on the Economics of Crime Literature’, Southern Economic
Journal, 61, 161–8.
Benson, B.L., I. Kim and D.W. Rasmussen (1998), ‘Deterrence and Public Policy: Tradeoﬀs
in the Allocation of Police Resources’, International Review of Law and Economics, 18 (1),
77–100.
Benson, B.L., I.S. Laburn and D.W. Rasmussen (2001), ‘The Impact of Drug Enforcement on
Crime: An Investigation of the Opportunity Cost of Police Resources’, Journal of Drug
Issues, 31 (4), 989–1006.
Benson, B.L., D.L. Sollars and D.W. Rasmussen (1995), ‘Police Bureaucrats, Their Incentives,
and the War on Drugs’, Public Choice, 83 (1–2), 21–45.
Benson, B.L., I. Kim, D.W. Rasmussen and T.W. Zuehlke (1992), ‘Is Property Crime Caused
by Drug Use or Drug Enforcement Policy?’, Applied Economics, 24, 679–92.
Bentham, J. (1962) [1778], ‘A View of the Hard Labour Bill’, in John Bowring (ed.), The Works
of Jeremy Bentham, Vol. IV, New York: Russell and Russell, pp. 3, 6–7, 32.
Block, M.K. and J.M. Heineke (1975), ‘A Labor Theoretic Analysis of the Criminal Choice’,
American Economic Review, 65, 314–25.
Bowers, W. and G. Pierce (1975), ‘The Illusion of Deterrence in Isaac Ehrlich’s Research on
Capital Punishment’, Yale Law Journal, 85, 187–208.
Brier, S.S. and S.E. Fienberg (1980), ‘Recent Econometric Modelling of Crime and
Punishment: Support for the Deterrence Hypothesis?’, Evaluation Review, 4, 147–91.
Buchan, B. (2002), ‘Punishing the Poor: Early “Liberal” Arguments for Penal and Police
Reform’, International Journal of Sociology of Law, 30, 201–18.
Byrne, J.M. (1988), ‘Probation’, Washington, DC: National Institute of Justice, GPO:1.
Cameron, S. (1988), ‘The Economics of Crime Deterrence: A Survey of Theory and
Evidence’, Kyklos, 41, 301–23.
Chadwick, E. (1929), ‘Preventive Police’, London Review, 1, 252.
Dezhbakhsh, H., P.H. Rubin and J.M. Shepherd (2003), ‘Does Capital Punishment Have a
Deterrent Eﬀect? New Evidence from Postmoratorium Panel Data’, American Law and
Economics Review, 5 (2), 344–76.
DiIulio, J.J. (1996), ‘Help Wanted: Economists, Crime and Public Policy’, Journal of Economic
Perspectives, 10 (1), 3–24.
Dodd, T., S. Nicholas, D. Povey and A.Walker (2004), ‘Crime in England and Wales, 2004’,
Home Oﬃce Statistical Bulletin, 10/04, July.
Donohue, J.J. III and S.D. Levitt (2001), ‘The Impact of Legalized Abortion on Crime’,
Quarterly Journal of Economics, 116, 379–420.
Ehrlich, I. (1973), ‘Participation in Illegitimate Activities: A Theoretical and Empirical
Investigation’, Journal of Political Economy, 81, 521–65.

Economics of crime

135

Ehrlich, I. (1975), ‘The Deterrent Eﬀect of Capital Punishment: A Question of Life or Death’,
American Economic Review, 397–417.
Ehrlich, I. (1977), ‘Capital Punishment and Deterrence: Some Further Thoughts and
Evidence’, Journal of Political Economy, 741–88.
Ehrlich, I. and J.C. Gibbons (1977), ‘On the Measurement of the Deterrent Eﬀect of Capital
Punishment and the Theory of Deterrence’, Journal of Legal Studies, 6, 35–50.
Ekelund, R.B. and C. Dorton (2003), ‘Criminal Justice Institutions as a Common Pool: The
Nineteenth Century Analysis of Edwin Chadwick’, Journal of Economic Behavior and
Organization, 50 (3), 271–94.
Hakim, S., A. Ovadia and J. Weinblatt (1979), ‘Interjurisdictional Spillover of Crime and
Police Expenditures’, Land Economics, 55, 200–212.
Harrison, P.M. and J.C. Karberg (2005), ‘Prison and Jail Inmates at Midyear 2004’, Bureau
of Justice Statistics Bulletin, April.
Horney, J. and M. Ineke (1992), ‘Risk Perceptions Among Serious Oﬀenders: The Role of
Crime and Punishment’, Criminology, 30, 575–93.
Levitt, S.D. (1998), ‘Why Do Increased Arrest Rates Appear to Reduce Crime: Deterrence,
Incapacitation or Measurement Error?’, Economic Inquiry, 36 (3), 353–72.
Levitt, S.D. (2004), ‘Understanding Why Crime Fell in the 1990s: Four Factors that Explain
the Decline and Six that Do Not’, Journal of Economic Perspectives, 18 (1), 163–90.
Lott, J.R. (1987), ‘Should the Wealthy Be Able to “Buy Justice”?’, Journal of Political
Economy, 95, 1–68.
Lott, John R. (1998), More Guns, Less Crime: Understanding Crime and Gun-Control Laws,
Chicago: University of Chicago Press.
Manzoni, Alessandro (1964), The Column of Infamy, Prefaced by Cesare Beccaria’s ‘Of Crimes
and Punishments’, Translated from the Italian by Kenelm Foster OP and Jane Grigson, with
an Introduction by A.P. d’Entrêves, London: Oxford University Press.
Mast, B.D., B.L. Benson and D.W. Rasmussen (2000), ‘Entrepreneurial Police and Drug
Enforcement Policy’, Public Choice, 104 (3/4), 285–308.
McGahey, R. (1980), ‘Dr. Ehrlich’s Magic Bullet: Econometric Theory, Econometrics, and the
Death Penalty’, Crime and Delinquency, 26, 485–502.
Mocan, H.N. and K.J. Gittings (2003), ‘Getting Oﬀ Death Row: Commuted Sentences and
the Deterrent Eﬀect of Capital Punishment’, Journal of Law and Economics, 46 (2), 453–78.
Mocan, H.N., S.C. Billups and J. Overland (2000), ‘A Dynamic Model of Diﬀerential Human
Capital and Criminal Activity’, National Bureau of Economic Research Working
Paper No. 7584.
Myers, S.L. (1983), ‘Estimating the Economic Model of Crime: Employment Versus
Punishment Eﬀects’, Quarterly Journal of Economics, 98, 157–66.
Neely, Richard (1982), Why Courts Don’t Work, New York: McGraw-Hill.
Paley, William (1785), The Principles of Moral and Political Philosophy, London: Faulder.
Passell, P. (1975), ‘The Deterrent Eﬀect of the Death Penalty: A Statistical Test’, Stanford Law
Review, 28, 61–80.
Passell, P. and J.B. Taylor (1977), ‘The Deterrent Eﬀect of Capital Punishment: Another
View’, American Economic Review, 67 (3), 445–51.
Peck, J. (1976), ‘The Deterrence Eﬀect of Capital Punishment: Ehrlich and His Critics’, Yale
Law Journal, 85, 359–67.
Philipson, T. and R.A. Posner (1996), ‘The Economic Epidemiology of Crime’, Journal of
Law and Economics, 39 (2), 405–33.
Rasmussen, David W. and Bruce L. Benson (1994), The Economic Anatomy of a Drug War:
Criminal Justice in the Commons, Lanham, MD.
Rasmussen, D.W., B.L. Benson and D.L. Sollars (1993), ‘Spatial Competition in Illicit Drug
Markets: The Consequences of Increased Drug Enforcement’, Review of Regional Studies,
23, 219–36.
Reaves, B.A. (1992), ‘State and Local Police Departments’, Bureau of Justice Statistics
Bulletin (February), 4.
Resignato, A.J. (2000), ‘Violent Crime: A Function of Drug Use or Drug Enforcement?’,
Applied Economics, 32 (6), 681–8.

136

Economics uncut

Rubin, P.H. (1978), ‘The Economics of Crime’, Atlantic Economic Review, July–August,
13–26.
Shoup, C. (1964), ‘Standards for Distributing a Free Government Service: Crime Prevention’,
Public Finance, 19, 383–92.
Sjoquist, D. (1973), ‘Property Crime and Economic Behavior: Some Empirical Results’,
American Economic Review, 83, 439–46.
Smith, Adam (1982) [1763], Lectures on Jurisprudence, R.L. Meek, D.D. Raphael and P.G.
Stein (eds), Indianapolis: Liberty Fund.
Sollars, D.L., B.L. Benson and D.W. Rasmussen (1994), ‘Drug Enforcement and Deterrence
of Property Crime Among Local Jurisdictions’, Public Finance Quarterly, 22, 22–45.
US Department of Justice (1988), ‘Jail Inmates 1987’, Bureau of Justice Statistics Bulletin, 1.
US Department of Justice (1989), ‘Prisoners in 1988’, Bureau of Justice Statistics Bulletin, 5.
Wolpin, K.L. (1978), ‘Capital Punishment and Homicide in England’, American Economic
Review, 68, 422–7.
Yunker, J. (1982), ‘The Relevance of the Identification Problem to Statistical Research on
Capital Punishment: A Comment on McGahey’, Crime and Delinquency, 28, 96–123.
Zedlewski, E. (1992), ‘Private Security and Controlling Crime’, in G.W. Bowman, S. Hakim
and P. Seidenstat (eds), Privatizing the United States Justice System: Police Adjudication,
and Corrections Services from the Private Sector, Jeﬀerson, NC: McFarland and Company.

Economics of marriage and divorce
Leora Friedberg and Steven N. Stern

Why do people get married? Love, sex, children, money. Why do they get
divorced? Probably for the same reasons. While most people would, at
most, want to hear what economists have to say about the financial aspects
of marriage, economists (who, surprising as it may seem, marry and divorce
like everyone else) have taken it on themselves to oﬀer useful hypotheses
about the rest of it too.
For example, they have studied whether divorce laws aﬀect the divorce
rate. Before 1970, many states in the USA required both spouses to agree
before a divorce could take place; now, most states allow one spouse to
initiate a divorce unilaterally. Under certain assumptions that economists
have spelled out, the type of divorce law would not actually aﬀect the
number of divorces, even while it does aﬀect the way that divorced couples
share resources. Economists also have shown that the type of divorce law
aﬀects how married couples share resources, even if they do not divorce,
and, theoretically, it might even influence how often they have sex.
In order to understand these ideas, we have to develop a model of both
marriage and divorce (or, in other words, a model that allows for changes
over time in the utility from marriage). In this chapter, we will set the stage
for this analysis by discussing trends in marriage and divorce in Section 1.
The major changes in matrimonial patterns in North America and Europe
that we will highlight serve as a backdrop for the analysis we will introduce
in the rest of the chapter. Then, to keep things simple, we will discuss separately the gains from marriage versus living together (Section 2), the
reasons why people marry (Section 3), the nature of decision-making
within marriage (Section 4), and the nature of the decisions to marry and
to divorce (Section 5).1
1 Trends in marriage and divorce
Figures 5.1 to 5.5 illustrate trends in marriage and divorce in several industrialized countries. Figure 5.1 shows long-term trends in marriage and
divorce rates in the USA, while Figures 5.2 to 5.5 compare recent trends
in other industrialized countries. The figures show marriage and divorce
rates by year – the number of marriages and divorces for each 1000 people
in the population. It would also be interesting to know the number relative
to the eligible population, that is, those available to marry (the unmarried)
137

138

1974

1972

1970

1968

1966

1964

1962

1960

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

Marriage and divorce rates per 1000 population in USA (1960–2003)

United Nations Center for Health Statistics.

Figure 5.1

Source:

Marriage and divorce rates per 1000 population

2002

2000

1998

1996

Marriage rate
Divorce rate

139

1966

1964

1962

1960

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

Marriage rates per 1000 population in France and Canada (1960–2003)

Statistics Canada (Canada) (www.statcan.ca) and Council of Europe (France) (www.coe.int).

Incomplete data for Canada.

Figure 5.2

Source:

Note:

Marriage rate per 1000 population

2002

2000

1998

1996

Canada
France

140

1968

1966

1964

1962

1960

1970

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

Marriage rates per 1000 population in Spain, Sweden and UK (1960–2003)

Council of Europe (www.coe.int).

Incomplete data for Sweden.

Figure 5.3

Source:

Note:

Marriage rate per 1000 population

2002

2000

Spain
Sweden
UK

141

0.5

1.5

2.5

1966

1964

1962

1960

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

1968

Divorce rates per 1000 population in France and Canada (1960–2002)

Statistics Canada (Canada) (www.statcan.ca) and Council of Europe (France) (www.coe.int).

Incomplete data for Canada.

Figure 5.4

Source:

Note:

Divorce rate per 1000 population

3.5

2002

2000

1998

1996

France

Canada

142

1964

1962

1960

Figure 5.5

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

Divorce rates per 1000 population in Spain, Sweden and UK (1960–2003)

Council of Europe (www.coe.int).

1966

Source:

1968

Incomplete data for Spain and UK.

0.5

1.5

2.5

1970

Note:

Divorce rate per 1000 population

3.5

2002

2000

1998

Spain
Sweden
UK

Economics of marriage and divorce

143

and divorce (the married). We are not able to report those statistics,
however, because they are more diﬃcult to come by and are not measured
as accurately.
It is apparent from Figure 5.1 that we need to distinguish between shortand long-run patterns in the data. Over the long run, the divorce rate in the
USA has steadily increased while marriage rates have bounced up and
down. The recent trends have been starker, though. The divorce rate more
than doubled in the recent past, rising from 2.5 per 1000 people in 1965 to
5.2 in 1980. It levelled oﬀ in the 1980s and then dropped back to 3.8 by
2003, but it remains the case that almost half of marriages end in divorce
(Cherlin, 1992). More recently, marriage rates have dropped oﬀ, falling by
25 per cent from 10.5 in 1980 to 7.5 in 2003.
Looking at Figure 5.1, it is not entirely clear that these recent developments represent major breaks from past behaviour. For example, the
divorce rate in the USA rose steadily in the early part of the century and
then plateaued in the 1950s, making the later increase look especially sharp.
In fact, if it had risen at the same annual rate as occurred between 1920 and
1950, the divorce rate would have hit 5.0 in 1990, slightly above the actual
rate of 4.7. Similarly, the marriage rate remains above the low of 8.4 that it
hit in 1958, which followed a period of very high marriage rates. What
distinguishes the recent period is the same patterns – which, if anything,
have been starker – occurring in other countries, along with a number of
contemporaneous trends that are entirely new, including an increased frequency of cohabitation, an increase in the age at first marriage and a
decline in the birth rate.
Figures 5.2 to 5.5 compare trends in marriage and divorce rates since
1960 in a selection of representative industrialized countries. Several features of the data are worth discussing.
First, the USA has had consistently higher marriage and divorce rates
than other industrialized countries have. Interestingly, the number of marriages relative to divorces started out widely divergent across countries and
very low in the USA but has converged to around two marriages per divorce
each year in most of these countries.
Second, divorce rates in industrialized countries rose universally between
1960 and 1990. The increase started a little earlier in the USA and also flattened out sooner, around 1980. While the absolute increase was also largest
in the USA, the percentage increases were higher in almost all other industrialized countries. The divorce rate more than tripled in France and more
than quintupled in Canada and the UK. Moreover, the USA is somewhat
unusual in the 25 per cent decline that has taken place since the divorce rate
peaked. In most other countries the rate of increase has either stopped or
slowed, but not turned negative.

144

Economics uncut

Third, marriage rates in industrialized countries began to fall universally
at some point between 1965 and 1985 and continue to drop today. These
declines began after divorce rates started to rise and in many cases were preceded by a short-lived increase in marriage, perhaps reflecting remarriages
among the newly divorced. The decline in marriage rates in the USA has
been smaller in relative terms, at about 25 per cent from its peak, than the
decline in other countries, which generally exceed 40 per cent.
2 Marriage versus cohabiting
What is the diﬀerence between shacking up and getting hitched? Marriage
is a contractual arrangement, with rules determined by church or state.
For example, Jewish weddings are not complete until the bride and groom
sign the ketubah, a marriage contract which, ‘ . . . spells out the husband’s
obligations to the wife during marriage, conditions of inheritance upon
his death, and obligations regarding the support of children of the marriage. It also provides for the wife’s support in the event of divorce’
(Anonymous, 2004a). Brides in many cultures brought (or still bring)
dowries to their husbands, with religious or secular law determining
the disposition of the dowry in the event of death or divorce. In other
cultures, husbands must provide a bride price to the bride or her father.
With some exceptions, non-marital relationships do not entail the same
contractual obligations.2 Several aspects of marriage as a contract are
important.
First, marriage as a transaction may be costly in terms of time, eﬀort
and/or money to enter into and to leave. It implies that the utility from marriage must exceed the utility from being apart as well as the costs involved
in getting married, and, possibly, divorced (although we do not emphasize
this in the model of getting married, which we develop later).
Second, divorce as the dissolution of a contract entails financial obligations between the spouses. For example, one spouse may be required to
pay income support to the other, and property acquired during the marriage (or even before) must be split according to some rule, perhaps
depending on behaviour during the marriage. There are a few motives for
these financial obligations associated with divorce. One is to provide
support for children issuing from the marriage, although these rules aﬀect
childless couples as well. Another motive is to punish certain types of
behaviour, which are viewed as violating the contractual obligations of
marriage. An additional motive is to compensate spouses for some types
of investments in the marriage, which are undertaken with the belief that
the marriage will last. We will elaborate on the motives for these ‘sharing
rules’ as we outline the reasons why people marry and subsequently
divorce.

Economics of marriage and divorce

145

Third, the nature of the marriage contract has varied greatly across
religions, countries and time periods. The Old Testament, for example,
established a right of husbands to unilaterally divorce their wives, for any
reason or no reason, although not without absolving all financial obligations. Subsequent Talmudic law (developed in oral form before Christ and
in writing after) gave Jewish wives the right to unilateral divorce when the
husband, ‘ . . . is physically repulsive because of some medical condition or
other characteristic, when he violates or neglects his marital obligations
(food, clothing and sexual intercourse), or, according to some views, when
there is sexual incompatibility’ (Anonymous, 2004b). Jesus, in various
gospels, recognized (though disapproved of) divorce, while the Roman
Catholic Church forbids it. Occasionally, legal and/or religious divorce law
has changed in response to demands of influential society members (for
example, Henry VIII). We will discuss some reasons why the legal regime
governing marriage and divorce has varied as the circumstances surrounding marriage have shifted.
3 What are the gains from marriage?
At the outset, we listed several reasons – love, sex, children and money –
why people might get married. Let us consider in turn how each of them
aﬀects an individual’s utility, and moreover why the impact may depend on
whether the couple marries instead of just living together.
3.1 Love and sex
While love may be a many-splendoured thing, we will model it as a simple
gain in utility from sharing your life with another. Love may matter
because you enjoy being with the one you love, and it may matter because
it causes you to care about the well-being of the one you love. In other
words, love may be an argument of your utility function and it may
change your utility function by placing weight on the utility of another
person. Next, there is sex. But is sex any diﬀerent from love from a modelling point of view? They both require another, and they both oﬀer
utility, given the right partner. Sex may complement love, if emotional
intimacy enhances sexual satisfaction. However, there are other important
characteristics of sex. One that we will ignore in this chapter is that there
is a relatively well-functioning market for sex (via prostitution, discussed
in Chapter 7), but not one for love, since sex is an act but love is an
emotion, which is diﬃcult to transact. Another characteristic of sex is that
it involves certain risks. It may be for the latter reason that sex is often
associated with marriage.
One risk of sex is disease. Monogamy reduces the risk of disease, but
how does one insure monogamy? First, both marriage and cohabitation

146

Economics uncut

increase physical proximity and hence the ability to monitor the partner.
Second, marriage, as a legal contract, often specifies penalties that raise the
cost of adultery. For instance, it may provide grounds for the other partner
to end the marriage or aﬀect the distribution of income and property after
divorce.
Another risk of sex is pregnancy. The risk of pregnancy explains why sex
within marriage may be preferred to sex outside of marriage – because the
welfare of children, and so the welfare of parents who care about their children, is enhanced by making marriage diﬃcult to end and by the financial
obligations imposed by marriage (although these obligations may be
imposed for children born out of wedlock as well).
Seen in this light, one can understand why the advent of eﬀective contraception in recent decades – by reducing the risk of pregnancy resulting
from sex – has reduced the frequency of marriage and increased the frequency of living together. While condoms have been around for millennia
(Youssef, 1993), the introduction of the birth control pill has had a drastic
eﬀect, especially after legal decisions in the late 1960s allowed it to be distributed widely among unmarried women. That is, because it allows women
to control their fertility, and women bear more of the costs of childbearing
than men do.3
3.2 Children
Sex naturally brings us to talk about children. There are a few reasons why
people have kids besides the possibility that they are an accidental byproduct of sex. One is that they feel a biological imperative. Another is that
they enjoy kids. Those two reasons look the same in a simple economic
model – they are things that raise an individual’s utility.
What makes kids interesting for our purposes is that one kid may raise
the utility of both parents at the same time. Thus, kids are public, not
private, goods. In other words, one kid provides utility to two parents, and
the utility one parent gets from spending time with the kid is not diminished
by the utility the other gets from spending time with the kid – as long as the
parents live together. If not, then kids are more like other private goods, for
example spaghetti – if one person eats the spaghetti, the other cannot. Since
kids cost about the same whether they live with one parent or both, it is
eﬃcient, from this perspective, for parents to live together. We will not have
a chance to say anything about the decision of whether to have kids,
although it may have some of the same features as other types of marital
decisions, which we will discuss below. Other interesting decisions, which
we will not discuss, are how many kids to have, and the trade-oﬀ between
investing in child quantity (by having more kids) versus child quality (by
spending more time or money on each kid).

Economics of marriage and divorce

147

3.3 Public goods
We already pointed out that kids are public goods. In a variety of other
ways as well, it is likely that two can live as cheaply as one. Kitchens can be
used to provide two meals about as easily as one, and larger sized bed linens
are cheaper than two sets of smaller sized bed linens. One spouse’s enjoyment of heat in the winter and air conditioning in the summer is little
aﬀected by the other’s as long as their preferences over the ambient temperature are not too diﬀerent. The same applies to home decor (rugs and
paintings) and to home entertainment (television sets, cable service and
stereos), once again provided that tastes are relatively similar – that is, that
both spouses like to watch the same TV programme, or that one does not
mind taping the football game to watch at a later time. Moreover, if being
in love also means that your tastes are more similar, then that enhances the
‘public good’ aspect of many household commodities.
Suppose that G stands for quantity of public goods associated with
living together. We also will introduce X to stand for a private good, or
bundle of private goods, which individuals consume. Consider a couple
with similar incomes deciding whether to live together. If they live apart,
then each faces a budget constraint like AB in Figure 5.6. With similar

X
(private B
goods)

E
D

G (public goods)

Figure 5.6

Budget constraints and indifference curves in marriage

148

Economics uncut

indiﬀerence curves, the best each alone can do is reach point D where the
‘best’ indiﬀerence curve is tangent to the budget constraint. However, if
they live together, then the cost per person of G is cut in half (because they
can share it). Thus, the new budget constraint is CB, and they can pick a
point like E (ignoring, for now, the possibility that they may have diﬀerent
preferences for G relative to X).
Kids provide a strong argument for marriage as opposed to cohabitation.
Since kids require costly investments that take a long time to reach fruition,
the contractual nature of marriage (the fact that it may be diﬃcult to leave
and that leaving it imposes continued financial obligations explicitly related
to kids) increases one’s willingness to invest in kids. The same can be said of
other public goods – it is worth investing in furniture together if we are reasonably sure that the relationship will last – though with somewhat less force.
3.4 Time
There is another type of public good that is important to talk about separately. It originates with the idea that a person’s time can be used to
produce something that is a public good. For example, a clean house is a
public good because both you and your spouse enjoy it, and the utility you
get from it does not diminish the utility your spouse gets from it. Similarly,
if you spend time teaching the kids to say ‘please’ and ‘thank you’, then
both you and your spouse enjoy the outcome of polite kids.
What makes these examples diﬀerent from the public goods discussed
above is that time, rather than money, is used to produce them. However,
money is being used indirectly, because using an hour of your time to clean
the house has an opportunity cost – you could be working and earning a
wage instead. That is why economists refer to activities of this sort as nonmarket or ‘household production’ – it highlights the notion that you are
using your time for something that is productive (it raises your utility) but
is diﬀerent from regular market production (by which you earn a wage to
buy market goods that raise your utility).
The idea of household production gets even more interesting when we
consider the following question: how should the burden of household production be divided between you and your spouse? One option is that you
could split the work – you could do the household work in the morning and
then go to work while your spouse takes over, or you could do it one day
and then go to work the next day, or you could do it for one year and then
go to work the next year. Another option is that you could do all the household work while your spouse works in the labour market full-time. A third
option is that both you and your spouse could work and pay someone else
to do your household work; this is possible for almost anything except
actually giving birth to a child.

Economics of marriage and divorce

149

The option that is least common is the first. There are two reasons why
it is rare for both spouses to split household work and market work. One is
that part-time jobs often pay much less per hour than full-time jobs. Thus,
one spouse working in a full-time job will often earn considerably more
than both spouses working in a half-time job. The other reason is that
market work often involves a career, rather than simply a job. By a career,
we mean a job that requires a long-term relationship with an employer or
fellow workers, or a long-term investment in skills, since many skills (for
example, carpentry, surgery, computer programming) take time to acquire
and deteriorate or grow obsolete when they go unused. Thus, ‘career’ jobs
typically pay more than short-term jobs.
All of this leads us to the notion of specialization – it often makes sense
for one spouse to specialize in market work and the other to specialize in
household work. Again, the reason is that part-time work is not highly
remunerative and interrupts careers. Individuals on their own cannot
specialize, but individuals in a relationship can, if one spouse provides the
income (which also goes further when it is used to purchase public goods
for two people) while the other spouse provides the time for household production. Moreover, marriage may be the preferred arrangement to foster
specialization precisely because of the concern that, when one spouse
specializes in household skills, they are giving up not just the current wage
but a career and hence higher future wages. You would be more willing to
do this if you receive some sort of commitment of compensation for the
loss of your career in case the relationship ends. Alimony and propertysharing arrangements associated with divorce provide that commitment.
Consider a simplified model where utility depends upon consumption of
potatoes as measured by X and the cleanliness of the house as measured by
C, where CCm Cf Cm and Cf represent the proportion of time spent
cleaning the house by the male and female, and 1 Cm and 1 Cf represent
the proportion of time spent working. Let wm be the daily wage of the male,
wf be the daily wage of the female, and p be the price of potatoes. Further,
assume that the couple maximizes U(C,Xm) U(C, Xf), subject to the
budget constraint,

wi(1 Ci) p(Xm Xf)

im, f

when married, while each member of the couple maximizes U(Ci, Xi)
subject to
wi (1Ci)pXi
for i m, f when single.

150

Economics uncut

Define (X*m , C*m) and (X*f , C*f ) as the optimal choices of the male and the
female when single. Note that the couple, when married, can choose
(X*
m, X*
f , C*
m, C*
f ) (the same bundle) and get utility, U(C*
m C*
f , Xm)
U(C*m C*f , Xf), which is more than the sum of utilities when single. This
occurs because they each benefit from the other’s cleaning time as well as
their own. Alternatively, the spouse with the higher wage can reduce his or
her cleaning time and spend more time in the labour market allowing the
couple to buy more potatoes. For example, the male could reduce his cleaning time to C**
m max(C*
m C*
f , 0), which is either the diﬀerence between
how much time he would spend if single (if C*m C*f ), or else zero.
Consequently, both spouses would receive at least as much household production as when they were single, and the family would be able to buy more
potatoes.
A last question arises when we consider housework. Why is it that women
have historically specialized in household production and men in market
production? A key reason is because women bear the children, so they have
to withdraw from market work for a length of time. Another possible
reason is that women are better at or care more about raising children,
though this is far from being proven. In terms of the model, we could
capture this idea by letting CaCf Cm where a now captures the productivity of women in household production relative to men. The statement
that women are better at household production is equivalent to the statement that a 1.
A third possibility is that men are more productive in market work as
reflected in relative wage rates. In fact, the model suggests that all that really
matters is the ratio of wages relative to the ratio of marginal products in
the household. As wf /wm decreases relative to a, men will increasingly
specialize in market activity and women will specialize in household production. While men’s wages have historically greatly exceeded women’s, the
gap has narrowed in recent years. In 2002, women’s average pay was 76 per
cent of men’s, up from 59 per cent over a 40-year period (Anonymous,
2004c). At this point, it is unclear whether women continue to be paid less
for equal work (and hence choose to specialize in household production
more often), or whether the wage gaps persist because women continue to
specialize in household production for other reasons.
4 How do married couples make decisions?
The model we developed above showed the potential gains from getting
married. Those gains are not only emotional but result from a more
eﬃcient use of time and resources enabled by two people living together.
However, we have not yet said anything about how those resources are actually used. There are several sources of possible conflict. Some money will

Economics of marriage and divorce

151

be spent on marital public goods, but some will also be spent on private
goods consumed by one spouse and not the other. Spouses must also
choose how to spend time on market production (work), household
production (housework) and leisure. Consider, for example, three types of
decisions:
1.

Given a certain amount of money that is available to spend, how will
couples decide whether to spend money on golf clubs or on Manolo
Blahnik sandals? And what if another way to spend the money is on
dinner at a nice restaurant? Or on a tricycle for the baby?
How will labour supply decisions by each spouse aﬀect how much is
spent on golf clubs, sandals or a tricycle? And how will spending possibilities aﬀect each spouse’s labour supply decisions?
How will the couple decide whether to spend money on golf clubs,
sandals, a tricycle, a nice dinner – or else save for the future?

These issues involve the allocation of money between private goods (golf
clubs, sandals and so on, designated X in our model above) and public
goods (dinner together, a tricycle, saving for future expenses, and so on designated G). They also involve the allocation of time between working in the
market (1 C in our model above, which earns a wage, w) and in household production (and producing marital public goods C, like a clean
house), and taking leisure.
The available evidence, as well as introspection, tells us that individuals
are not completely self-eﬀacing or altruistic in marriage, and therefore
spouses bargain over how to split shared resources. However, even as we
make hard-hearted assumptions about the nature of bargaining, we can
allow for the possibility that spouses get some utility from the well-being of
their spouse.
4.1 How do families make decisions?
We assume that individuals maximize the utility gained from such decisions, subject to a budget constraint. Does a family act as a single decisionmaker, maximizing a family utility function? If so, the family utility
function depends on the total amount that the family spends on diﬀerent
goods regardless of who consumes them or on the utility of each family
member. In these cases, we do not need to know how family members interact when these decisions are made. Interestingly, a model that yields the
same outcome is one in which one member of the family is a dictator who
makes all the decisions to maximize his or her own utility.
However, empirical analysis has rejected a key implication of these
‘unitary’ models of family decision-making: the allocation of consumption

152

Economics uncut

within the family should not depend on who earns what within the family.
In other words, unitary models imply that we would not have to know who
brings home the bacon and who fries it up in a pan in order to predict
whether high-fashion sandals, high-tech golf clubs or a tricycle is purchased. However, two studies found that the share of income spent on
men’s versus women’s and children’s clothing varied in Canadian families
(Browning et al., 1994) and in British families (Lundberg et al., 1997),
depending on the proportion of family income earned by husbands versus
wives. Another study showed that, as the share of cash income accruing to
women in Côte d’Ivoire rose, the share of spending on food rose and the
share on alcohol and tobacco fell (Hoddinott and Haddad, 1995).
4.2 Bargaining among family members
If families do not make decisions as a single unit, then we are back to the
model of individual decision-making with which students of microeconomics are familiar. Nevertheless, being in a family still changes the way
that decisions are made because family members share control over
resources and are aﬀected by actions of other family members. A natural
way to model these types of interactions is using game theory.
Economists have focused on a particular type of game theory to explain
family decision-making – games with cooperative bargaining between
agents (or spouses, in this case). Define the value to the male of being in the
relationship as Rm and the value of being separate as Sm. Similarly, assume
that the value to the female of participating in the relationship is Rf and the
value of not participating is Sf . R includes the benefits of relationships that
we discussed earlier, and S includes the value of extra privacy and independence and the cost of loneliness along with any monetary eﬀects. As
long as spouses can agree on how to split the surplus, Rm Rf Sm Sf ,
from being together, then they will remain together. If they cannot cooperate, then the best they can do is get utility Sm and Sf from being apart – so
these are the threat points in the cooperative bargaining game because each
individual will prefer to get S rather than to stay in the relationship and get
RS. Their bargaining determines how the surplus will be split, and thus
the actual values of Rm and Rf . Of course, they will both walk away if there
is no surplus from marriage (Rm Rf Sm Sf 0), which is the same as
saying that the threat points exceed the gains from marriage for both
spouses (Rm Sm and Rf Sf ). Manser and Brown (1980) and McElroy and
Horney (1981) defined the Nash cooperative bargaining model to marriage
and defined the threat points as the utility from separating.
John Nash demonstrated some important characteristics of the cooperative bargaining problem (Nash, 1950). He showed that, under relatively
general conditions, cooperative bargaining will yield an allocation that

Economics of marriage and divorce

153

maximizes the product of each spouse’s surplus, (Rm Sm)*(Rf Sf ),
subject to the household budget constraint. This solution is Pareto-eﬃcient
– neither spouse could be made better oﬀ without making the other worse
oﬀ. As an example, suppose that the utility of each spouse when separated is, Sm Sf 0, and the utility when they are together is Rm U(Xm) and
Rf U(Xf), where U(.) is a standard utility function with decreasing
marginal utility in X. Moreover, suppose that family income Y is going to
get split so that, YXm Xf . The solution that maximizes
(Rm Sm)*(Rf Sf )Rm *Rf
U(Xm)*U(Xf)
U(Xm)*U(YXm )
is Xm Xf Y/2 – each spouse gets half of Y. If we changed the problem
so that Sm Sf 1, the solution would not change, as long as Y/2 1; if
Y/2 1, then the spouses would separate. On the other hand, if we changed
the problem so that Sm 0 and Sf 1, then the solution is Xm Y/2 a and
Xf Y/2a, where the value taken by a is positive (so f gets more stuﬀ
than m) and depends on the actual form of the utility function U(.). It is
important to note that the same results apply when U(.) depends not only
on private consumption X but also on marital public goods and even on the
other spouse’s utility, so that spouses display some altruism (that is, they
are willing to give up some of their own consumption in order to increase
the utility of the other spouse).
The key result, then, is that the allocation of resources to each spouse
increases with their threat points. Knowing that, we can begin to answer the
questions posed above. Along the way, we will also consider whether separation is the appropriate threat point in all situations of marital dispute over
resources.
4.3 Decisions involving spending: golf clubs versus sandals
Suppose that all the money that the couple has is going to be spent immediately, and entirely on private goods. The choice of who gets more of the
private goods (golf clubs versus sandals) depends on each spouse’s bargaining power – that is, their threat points. If Sm Sf , then spouse m gets
more stuﬀ.
What determines the threat points? There are a few factors involved in
determining each individual’s utility from being separate. One factor is how
unhappy, emotionally, each partner would be when separated instead of
together. If partner m would be much lonelier, while partner f has a lot of
hobbies, then that reduces m’s bargaining power relative to f ’s and hence
m’s ability to lay claim to private goods Xm while married. A similar

154

Economics uncut

outcome would be observed if f is much more likely to attract a new romantic partner than m.
A second factor aﬀecting utility from being separate is each partner’s
financial resources outside of marriage. That in turn depends on their
earning power and on the legal regime that governs the separation of
marital assets following divorce. For example, if one partner (the wife, say)
gives up a career to specialize in household production, that reduces her
bargaining power over the distribution of resources within marriage. Why
would she agree to this future loss in bargaining power? Perhaps she does
it after extracting a large enough ‘payment’ up front in terms of family
resources (perhaps a diamond?) that it makes up for the loss in bargaining
power later on.
Another possibility is that she can find a way to bind her husband to an
initial agreement not to exercise his bargaining power later on. For
example, the legal regime in many states in the USA is now set up to favour
the partner who specializes in household production. On the other hand,
many states used to distribute property after divorce according to who had
legal title to it, which at the time was generally the husband, and some states
even forbade women from owning property – which gave the husband most
of the bargaining power within marriage.
A third factor aﬀecting utility from being separate involves child custody.
Custody laws that favour mothers will increase their bargaining power
within marriage. However, it is becoming more common to assign joint
custody following divorce.
4.4 Decisions involving spending: private goods versus dinner together
When we incorporate the possibility of marital public goods, we have to
reconsider whether spouses make decisions in the same way. There are a
couple of possibilities. If they bargain cooperatively, then they will choose
the Pareto-eﬃcient level of marital public goods. The eﬃcient level is dictated by adding together each partner’s marginal valuation of the public
good, relative to private goods (whereas the eﬃcient level of private goods
is dictated by equating the marginal valuations).
However, public goods are subject to free-rider problems when agents act
non-cooperatively. The idea is that, if I know that you will provide some of
the public good (by taking care of the kids, for example), then I will not
bother supplying as much myself. Moreover, if you follow the same reasoning, then too little of the public good (child care) is supplied overall.
Free-riding has been observed in many situations involving the private provision of public goods in non-marital settings.
Which model is correct? In the case of child care, it appears that married
parents tend to act cooperatively (the first model) while divorced parents

Economics of marriage and divorce

155

tend to act non-cooperatively (the second model) because it is diﬃcult to
monitor the other partner’s contributions (Weiss and Willis, 1985). It may
even be the case that the non-cooperative bargaining outcome (underprovision of the public good) is the operative threat point in a cooperative
bargaining setting – it may be more realistic than considering divorce
during every conflict (Bergstrom et al., 1986; Woolley, 1988; Lundberg and
Pollak, 1993).
There is also evidence that mothers care more about child welfare – and
hence may contribute more to the public good, either in a cooperative or
free-riding setting – than fathers. For example, unearned income given to
Brazilian mothers is associated not only with increased fertility but also
with greater improvements in child health compared with unearned income
given to fathers (Thomas, 1990).4
4.5 Buying stuff today versus saving
Another option is not to spend – to save resources for the future. For
example, an individual will save in case of a financial emergency or to
finance retirement. It will be worth doing this if the value of consumption
in the future is worth more than the value today.
However, a couple may make a diﬀerent decision. This may happen in a
situation where some of the future uses are public (benefitting both) and
some are private (benefitting only one), yet both spouses may have control
over financial assets (as is common today). Therefore, there is a risk that
they decide to save for the future, but then one chooses to buy private goods
later on. If both fear that the other will do this, then the free-riding result
from above will hold here as well – the couple will undertake too little saving
because they are not confident that the other will refrain from buying
private goods for themselves later. A related concern is that financial assets
will be divided in the event of divorce, in which case they cease to be a
public good; thus, the increase in divorce rates in the USA may be a cause
of the decline in private savings rates.
5 Transitions into and out of marriage
We have discussed the gains from marriage (and cohabitation) and the
nature of decision-making by couples. At this point, we will talk about how
couples decide whether to enter into or leave a relationship.
5.1 Getting together
Suppose a couple is considering forming a relationship of some type. Later,
we will discuss diﬀerent types of relationships (cohabitation, marriage) and
how couples might choose among them. For now, we can abstract away
from the type and just call it a relationship.

156

Economics uncut

Again, define the value to the male and female of being in the relationship as Rm and Rf and the value to them of being separate as Sm and Sf .
Further, for now, assume that both the male and the female know the
values of Rm, Sm, Rf, and Sf. These issues are discussed in more detail
below.
Then, Becker’s (1981) model of relationship formation says that a relationship will occur if and only if
(Rm Rf)(Sm Sf)RS

(5.1)

In other words, the relationship will occur if the total value to the couple
of forming a relationship is greater than the total value to the couple of
being single. Define the left-hand side of equation (5.1) as the total net
value of the relationship.
The condition defined in (5.1) is uncontroversial when both the male
and female prefer being together (Rm Sm and Rf Sf). Imagine, however,
that
Rm Sm

Rf Sf ;

that is, the male wants to have the relationship and the female does not. If
the total net value of the relationship is positive, then there is some side
payment or transfer p that the male can pay to the female such that
Rm Sm p 0,
Rf Sf p 0;
that is, both would want to form a relationship. The side payment need only
satisfy the conditions that
Rm Sm

Sf Rf .

If the couple sets pRm Sm, then the male is willing to be in the relationship because Rm Sm p0, and the female is willing to be in the relationship because
Rf Sf p(Rf Sf)(Rm Sm)

(because the total net value, as defined in (5.1), is positive). Similarly, if the
couple sets pSf Rf , then the female is willing to be in the relationship

Economics of marriage and divorce

157

because Rf Sf p0, and the male is willing to be in the relationship
because
Rm Sm p(Rm Sm)(Sf Rf)

(again, because the total net value is positive). On the other hand, if the
total net value is negative, then there is no side payment that will induce
both partners to want to form a relationship; the payment required by one
of the partners with a negative value is greater than the net benefit of the
relationship of the other.
In summary, any value of p between Sf Rf and Rm Sm satisfies the conditions for a relationship to occur. What determines the actual value of p?
It may be the outcome of the Nash bargaining game we described earlier,
so that the side payment p increases with Sf and decreases with Sm. In other
words, as the wife’s value of being separated increases relative to the
husband’s, then the payment made from the husband to the wife (if p 0)
grows, or the payment made from the wife to the husband (if p0) shrinks
closer to zero. Even if some other form of bargaining determines p, as long
as it is eﬃcient, it will not aﬀect whether a relationship begins.
Throughout this section, we have assumed that the couple can transfer
utility from one member to the other with p. However, ‘transferable utility’
is a tricky concept. Bergstrom (1997) shows that, when there is no public
good in the family, assuming transferable utility is not restrictive. Further,
he derives a necessary and suﬃcient condition on utility functions for transferable utility to exist when there is a public good. We note the existence of
the issue and proceed.
How does one partner make a side payment to the other? It may be in the
form of allocation of household chores or relationship assets, as we discussed earlier, or it may be something less tangible like the payer being
‘extra nice’ to the recipient of the side payment. One might argue that, in
the real world, we do not see such side payments. However, sometimes the
side payments may be diﬃcult to observe (for example, being ‘extra nice’).
Also, we pointed out earlier that many empirical studies have found
evidence of unequal allocations of resources within families, depending
upon Rm Sm relative to Sf Rf . A good example of such a side payment
is that Stern was so anxious to work with Friedberg on this chapter that he
agreed to let her have all royalties associated with the chapter.
5.2 Breaking up
Now consider a couple in a relationship – which implies that, at the time
they chose to form the relationship, the total net value of the relationship
defined in (5.1) was positive. What if, for some reason, the total net value

158

Economics uncut

changes later on? If the total net value of the relationship is still positive,
then the couple may renegotiate the side payment p so that both continue
to prefer remaining in the relationship (even if one would not without the
side payment). If the total net value of the relationship is now negative,
then there is no side payment that will make both members of the couple
better oﬀ in the relationship than apart, and the relationship ends.
The key aspects of this decision are the following. First, the couple will
not break up as long as nothing about the relationship changes, because the
total net value remains positive. Thus, we have to think about what may
change in the relationship in order to explain why divorces occur. Second,
the model we have described so far assumes that all transitions into and out
of relationships are efficient. A couple forms a relationship if and only if
the total net value is positive, and they dissolve the relationship if and only
if the total net value is negative. This is a key result about bargaining that
was first proposed by Coase (1960). His work made it clear that several
assumptions are necessary for these transitions to occur eﬃciently. Later
on, we will discuss the validity of these assumptions, some of which – most
notably, perfect information and costless transitions – may not be reasonable when applied to marriage and divorce. Another implication of his
work is that factors aﬀecting the distribution of resources (and hence the
utility of one partner relative to the other) will not influence whether a
divorce occurs, as long as divorce occurs eﬃciently. We will also discuss specific applications of this result later on.
In the meantime, we will ask what factors may ultimately precipitate a
divorce? Among economists, there have been three approaches to modelling the source of changes in the total net value of the relationship. One is
to model a change in the characteristics of the relationship that determine
R and S over time. A second approach is to model how partners may learn
more about the true values of R and S over time. A third is to consider the
possibility that actions taken by partners during marriage directly aﬀect
subsequent values of R and S. We will present evidence suggesting that, in
fact, all of these probably occur.
5.3 Changes
What is it about a relationship that may change over time? We have to
consider factors that aﬀect at least one of the values Rm, Sm, Rf , and Sf .
Some changes might arise because the characteristics or circumstances of
one of the partners changes unforeseeably (due to, perhaps, job loss, or an
inheritance, or the introduction of Viagra or Prozac) or some characteristic of the relationship itself changes (because, say, the TV breaks down,
so the couple no longer spends time together watching their favourite
programme).

Economics of marriage and divorce

159

Consider an improvement in labour market opportunities for women.
This probably raises Rm, Rf , and Sf . Both Rf and Sf rise because the female
has new opportunities that previously were not available, so she is better oﬀ
whether she is married or not. Rm would also rise as long as the husband
gets some of the benefit from his wife’s good fortune, perhaps because she
purchases more marital public goods. Even though Rm, Rf, and Sf all
increase, however, it is likely that Sf jumps the most (because now the
female is less likely to have to rely on a male for those expensive sandals);
it may be the case that Sf increases by more than the sum of Rm Rf (the
value of marriage). If that is the case, there will be some couples in a relationship who will now separate because the total net value of the relationship becomes negative. There will also be some couples who would have
formed a relationship in the past who are no longer willing to do so, again
because the total net value of the relationship is now negative. Thus, the
model suggests that the observed improvements in market opportunities
for women should reduce the marriage rate and, at least in the short run,
increase the divorce rate, as has happened.
Consider, instead, a new divorce law that requires husbands to pay
alimony to wives (or, say, the partner with more income outside of marriage
to pay the partner with less income). This will raise Sf while reducing Sm by
the same amount, but, under one important assumption, which we will
discuss in a moment, it would not alter the total net value of the relationship (Rm Rf)(Sm Sf). Consequently, if divorces occur eﬃciently, then
it would not precipitate a divorce. Even though it makes separation more
attractive to females, it makes it less attractive to males, so males will raise
their side payments p to avoid divorce, which will satisfy females. Therefore,
even though it would not cause a divorce, it would alter the distribution of
resources within marriage.
It is important to point out the key assumption on which this conclusion
rests, since it helps illustrate the main point. We must assume that the
factors determining alimony payments do not induce either ex-partner to
alter their income – for example, by choosing to earn less in order to either
win more or pay less alimony. If this assumption is violated, then the new
divorce law will alter the value of divorce (by reducing it) and hence the
incentive for couples to divorce.
All of these examples show that, if divorces occur eﬃciently, we have to
draw a distinction between those changes that alter the total net value of
the relationship RS, and those that shift resources within the relationship without altering the total net value. As we mentioned earlier, we
will discuss the assumptions necessary for divorces to occur eﬃciently,
and we will analyse the consequences of alimony laws if divorces occur
ineﬃciently.

160

Economics uncut

5.4 Learning
What happens if the couple learns about Rm and Rf over time? Consider a
model where a couple meets and receives imperfect information – a
signal – about RRm Rf. For example, they spend time with each other
and start to learn how much they like each other. Based on the signal, the
couple updates their beliefs about the value of R and then chooses
whether to form a relationship. If they choose to form a relationship, then
they enjoy each other’s company and next period receive another signal
about the relationship. Again they update their beliefs about the value of
R and make a new, possibly diﬀerent decision. Such a model can help
explain why we observe both cohabitation and marriage. The advantages
of marriage relative to cohabitation include some legal, religious or societal benefits, along with, perhaps, a greater level of commitment and intimacy; the disadvantage of marriage relative to cohabitation is the large
cost associated with dissolving a marriage. Such a model suggests that
couples who choose to marry without cohabiting received very good
signals about the value of their match and are willing to risk a relatively
small probability of future divorce to gain the advantages of marriage.
Couples who choose to cohabit received a signal good enough to form a
relationship but not good enough to commit to marriage. Cohabiting
couples are using cohabitation as an ‘option’ to commit in the future if,
with better information, the couple discovers that it has a very good
match.
This model of learning may help explain certain facts about cohabitation and marriage. Many empirical researchers have found that cohabitation leads to higher divorce rates (Brien et al., 2006). More precisely, if
we compare two apparently identical married couples, one of whom
cohabitated prior to marriage, then the couple who initially cohabited is
subsequently more likely to divorce. This fact has led some conservative
social commentators to argue that cohabitation is destroying the institution of marriage, which perhaps it does by somehow reducing the taste
for marriage. The theory above suggests another story: those who marry
right away received very good initial signals and, on average, have very
good matches; while those who cohabit received initial weaker signals
and, on average, have worse matches. Even though the ones with
worse initial signals receive later signals good enough to induce them to
marry, on average, they are not as good matches as those who received
good enough signals to marry right away. An implication of this model is
that, if everyone were forced to cohabit prior to marriage, then divorce
rates would decline because the cohabitation period would have no
direct eﬀect on the quality of the subsequent marriage but it would serve
as a learning period. During the learning period, some couples would

Economics of marriage and divorce

161

discover that they are not a good match and would not marry in the
first place.
5.5 Commitment
Another potentially informative fact is that, the longer a couple is in a relationship, the less likely they are to end it. There are two popular hypotheses to explain this phenomenon. One hypothesis is that it is essentially a
statistical phenomenon related to selection. Consider the learning model
outlined above, in which couples learn over time about the true value of the
relationship R. Each period the couples who realize they have a negative
total net value separate. Early on, all of the couples with negative match
values separate. Over time, additional couples learn that they have a negative match value, but fewer do so over time because the remaining relationships have a higher average value of R. As this process continues, the
separation rate for couples married at a certain point in time falls because
the couples with lower values of R are selected out through divorce, even
though there is no direct eﬀect of duration on relationship quality.
The second hypothesis is that the relationship itself causes the value of
the relationship to increase over time. In other words, the couple invests in
their relationship as it develops. They invest by learning how to interact
with each other, maybe by buying assets (such as a house) together, having
children together, or in other ways intensifying their commitment to each
other. Each time they invest in their relationship, they build a larger wedge
between the total value of the relationship R and the total value of separating, S, thus decreasing the probability of separation.
We can take the model one step further and note that these investments
are sunk in the match itself, and if the relationship dissolves, then most, if
not all, of the value of the investments is lost. For example, it may be very
costly to children for a married couple to divorce, or there may be large
transactions costs associated with dividing up the value of a house upon
separation. An implication of this point is that couples with very good
matches will be more likely to invest in their relationship than those with
less good matches. The couples with mediocre matches will hesitate to
invest because they know there is a significant probability of dissolution in
the future. Moreover, married couples will invest more than cohabiting
couples because, a) they have better matches (see the discussion above) and,
b) separation costs are higher for marriage than for cohabitation, so separation is less likely. The phenomenon that the level of investment depends
upon the quality of the match is called endogenous investment. It has the
feature that investment increases the variability of match quality, making
the best relationships even better and not having much of an eﬀect on
mediocre relationships.

162

Economics uncut

5.6 Breaking up is hard (or easy) to do
In the simple model above, we showed that all separations are eﬃcient –
they occur if and only if the total net value of the relationship (Rm Rf)
(Sm Sf)RS turns negative. Most non-economists are not comfortable
with this result, so it is useful to examine the underlying assumptions on
which the result depends.
In this case, one oﬀending assumption is that partners have perfect information: this requires that the male knows as much about Rf and Sf as the
female, and vice versa. Consider, instead, a model where each member of a
relationship has some information about his or her own preferences not
observable by the other member. The couple has to bargain about the size
of any side payment p from one to the other. If each had perfect information about the other’s preferences, they would know the range of side payments (Sf Rf and Rm Sm) necessary to satisfy the condition that both
would be better oﬀ in the relationship than not. However, if each has some
unobservable information, then he or she has an incentive to lie about it to
get a better deal. Knowing that, partners will still bargain, but now they will
consider the trade-oﬀ between getting a bigger side payment (by exaggerating about the value of one’s outside option, S) against the loss associated
with a possible separation if the other partner refuses a large side payment.
In such a situation, some couples break up even though the total net value
of the relationship is positive (so breaking up is too easy).
Another controversial assumption is that transitions are costless.
Suppose instead that the government (or perhaps a religious authority)
imposes a significant monetary or non-monetary cost D on divorcing
couples. The immediate eﬀect of such a cost is to decrease Sm and Sf For
example, if the divorcing couple split the cost D equally, then the new values
of being single are Sm 12D and Sf 12D. Since Sm and Sf decline and there
is no change in Rm and Rf , the total net value of the relationship increases.
Thus, the eﬀect of the divorce cost is, unsurprisingly, to decrease the
number of couples who divorce (so breaking up is too hard).
It is important to note, however, that the allocation of the divorce costs
between the male and female has no eﬀect on whether a divorce occurs,
since the decision to divorce depends only on the total net value (including
divorce costs). A law that required the cost to be shared equally would not
aﬀect the total net value diﬀerently than a law that requires the husband,
say, to pay the entire cost. All that would change would be the relative size
of Sf and Sm. The husband would be willing to provide a larger side
payment to stay married and avoid Sm D, but the wife would only be
willing to accept it if the total net value was positive. This is analogous to
the result, presented above, that laws mandating alimony payments after
divorce do not alter the incidence of divorce.

Economics of marriage and divorce

163

In the last few examples, we considered how changes in divorce laws
aﬀect the frequency of divorces (imposing a divorce cost reduces the incidence of divorce, but the allocation of the cost between partners does not).
But how do they aﬀect the marriage rate? Consider a couple deeply in love –
they are considering marriage and not worrying about the future.
Following in the footsteps of other romantic couples before them, they
compute the total net value of the relationship, and, if it is positive, negotiate side payments and marry. Divorce costs and alimony rules play no
role in their decision of whether to marry because they do not consider possible changes in the value of the relationship. If, in contrast, they recognize
that things may change and divorces do occur, then imposing future divorce
costs essentially increases the cost of marrying today, thus decreasing the
total net value. Therefore, the marriage rate would decrease, but also those
marriages that do occur would be stronger (in the sense that the total net
value of the marriage would be higher).
5.7 Policy interventions
As we have demonstrated, if divorces occur eﬃciently, then divorce laws
aﬀecting the distribution of resources will not aﬀect whether a divorce
occurs. Even divorce laws that do not directly determine the distribution of
resources can be interpreted in the same way. Before 1970, most states in
the USA required divorcing couples to demonstrate some fault – adultery,
abuse, abandonment and so on – in order to grant a divorce. This was
implicitly a mutual consent divorce regime; if a couple wanted to divorce,
they could lie about fault and obtain a divorce. Led by California in 1970,
most states changed their law to allow both no-fault and unilateral
divorce – so that one couple could instigate a divorce unilaterally and
without alleging fault. This raises two questions: should this aﬀect the
divorce rate, and did it aﬀect the divorce rate?
If divorces occur eﬃciently, then changing the law from mutual to unilateral divorce should not aﬀect the divorce rate. Instead, it acts to redistribute resources from the spouse who wants to stay in the relationship
(because Rm Sm) to the spouse who wants to leave (because Rf Sf ).
Under a mutual consent divorce law, the spouse who wants to leave will be
willing to make a side payment that the spouse who wants to stay is willing
to accept to agree to a divorce, as long as Sm Rm Rf Sf , so the gain to
spouse m from leaving the relationship exceeds the gain to spouse f from
staying in the relationship. Rearranging terms, this implies that (Rm Rf )
(Sm Sf ) 0, or the total net value of the relationship is negative.
On the other hand, if the total net value is positive, then the payment that
f is willing to make to obtain a divorce is smaller than the payment that m
is willing to take in order to grant a divorce. Now, suppose that the divorce

164

Economics uncut

law changes to allow unilateral divorce – would f now go ahead and get that
divorce? Because the total net value of the relationship is positive, m would
now be willing to pay f to stay in the marriage, and the payment would be
big enough to make f willing to stay. Thus, even though the law now allows
f to leave unilaterally, f would not go unless the total net value of the relationship is negative – so the divorce rate would be unchanged, even though
it looks like the law change has made divorce easier. What changes would
be the direction of payments, from the spouse who wants to leave the marriage to the spouse who wants to stay.
However, empirical evidence suggests that the divorce rate rose in states
in the USA after they changed their divorce laws (Friedberg, 1998). This
suggests that divorces do not occur eﬃciently. This belief is widespread
among policy-makers, who have recently reacted by discussing policies that
would increase the cost of divorce. A number of states in the USA have proposed giving people a choice of two marriage types: a regular marriage or
a ‘covenant marriage’. A covenant marriage has extra costs associated with
divorce. The Catholic Church has suggested it will approve only covenant
marriages in states that provide them.
The analysis in our perfect information model suggests that such government interference will not have the intended eﬀect because all marital separations are eﬃcient. In this subsection, we consider various possible
reasons why it might be welfare improving for the government to discourage divorce.
The most obvious possibility is the eﬀects of divorce on children. A large
empirical literature shows that children suﬀer in many ways when their
parents divorce. If the divorcing parents do not take into account the cost
of divorce to their children, then it might be appropriate for the government
to increase divorce costs to simulate the costs to the children. While this
argument has some merit, it also has problems. First, it is diﬃcult to
measure the eﬀect of divorce on children. The relevant counterfactual is the
welfare of the children with unhappy parents who would like to be
divorced. In other words, discouraging the divorce of an unhappy couple
with children does not magically create a happy couple. Second, parents do
care about their children, and maybe they give equal weight to their own
happiness and their children’s happiness in making divorce decisions. We
have no measures of how much they internalize such costs. Finally, divorce
costs are imposed on couples without children when such an argument is
irrelevant. In fact, if anything, such an argument should lead the government to encourage divorce among couples without children so that they do
not divorce once children arrive.
Another possibility that has been discussed in the literature is that, when
a couple divorces, it changes the size and the distribution of people in the

Economics of marriage and divorce

165

marriage market. The increase in the number of single people provides a
benefit for those already single, and this benefit is not internalized by the
divorcing couple. The eﬀect on the quality distribution of single people is
harder to sign. In any case, because there is an externality, the government
may play a role. The problem with this argument is that the externality goes
the wrong way! Since divorce provides a benefit to others not internalized
by the divorcing couple, the government should provide a subsidy to divorcing couples.
The final possibility is that increased divorce costs reduce divorce probabilities, causing couples to invest more and improve their relationships.
Consider a case where a new husband and wife bargain over sending the
wife to medical school while the husband supports her. He is more likely to
agree to such a plan if he knows that she will not be able to divorce him
later after she has her medical degree. Without high divorce costs, it may be
diﬃcult for her to credibly commit to not divorcing later. We can tell a
similar story about the couple investing in eﬃcient division of labour. As
discussed above, it may be eﬃcient for the wife to specialize in household
production and the husband in market production. Without a commitment
not to divorce, the wife may not be willing to specialize because she knows
that, if they later divorce, she will need to develop market skills. The
problem with this argument is that the best solution here is not an imposed
divorce cost; rather it is an eﬃcient contracting mechanism. In the first
example, the husband and wife should sign a binding contract determining
how the proceeds from a medical education will be shared between them if
a divorce occurs. In the second example, the husband and wife should agree
upon divorce-contingent alimony at the time they are bargaining about
specialization. An eﬃcient contract solves the commitment problem
without causing bad marriages to remain intact.
6 Conclusion
In this chapter, we have aimed to demonstrate that economic tools can help
us understand how individuals make decisions related to marriage and
divorce. Economists have not only developed models to explain but also have
found evidence to support several hard-hearted theories. One is that the gains
to marriage are not simply from sharing love and aﬀection but from a more
eﬃcient use of resources (both money and time). Another is that the distribution of those resources within a marriage depends at least in part on the bargaining power of each spouse. A third is that divorces may occur eﬃciently,
so that religious or social laws governing divorce do not aﬀect whether
divorces occur but do aﬀect the distribution of resources within marriage.
As we mentioned early on, even economists have not yet come up with
definitive explanations for the recent trend away from marriage. One reason

166

Economics uncut

is that, as we have argued, many diﬀerent factors aﬀect decisions to marry
and divorce, and these factors can interact in complicated ways. A second
reason is that it is diﬃcult to control for all of these relevant, and changing, factors when attempting to test competing theories. For example, we
can write down a model showing that married women are working more
because the increased probability of divorce raises the gains to investing in
their careers, and another model showing that an increase in married
women’s labour supply raised divorce rates by augmenting friction within
marriages. To test whether either, or both, explanations are valid, we need
to find exogenous reasons why women are working more (to see whether
that led to more divorce) and why divorce rates have risen (to see whether
that led to increased labour supply). Did recent increases in women’s wages
exogenously cause women to work more? Perhaps, or else an increase in
married women’s labour supply, which boosted their investments in education and careers and their incentive and power to fight gender discrimination, raised women’s wages. Did the shift in divorce laws in the early 1970s
from requiring mutual agreement by spouses to divorce to allowing one
spouse to leave unilaterally exogenously cause more divorce? Perhaps, or
else laws were changed in response to pressure from people who increasingly wanted to exit their marriages.
In summary, challenges remain for economists who are in turn challenging researchers in other disciplines to provide coherent explanations and
empirical tests of marriage and divorce behaviour.
Notes
1.
2.
3.
4.

In most cases, our theories apply to state-sanctioned marriage, whether between two
people of diﬀerent or of the same genders. Without intending any bias against homosexuals, we will sometimes use gender-specific terminology.
Most famously, ‘palimony’ (alimony for a pal) has been granted by courts under relatively stringent conditions – the cohabiting relationship had to involve, ‘an express or
implied contract’ (http://www.palimony.com/7.html).
Goldin and Katz (2002) documented the consequences in terms of later marriage and
increased investment by women in education and careers that followed the dissemination
of the pill.
The indicators of child health were calorific intake, height, weight and survival probabilities. Similar outcomes were found in Côte d’Ivoire (Hoddinott and Haddad, 1995)
and in Thailand (Schultz, 1990).

References
Allen, D.W. (1998), ‘No-Fault Divorce In Canada: Its Cause and Eﬀect’, Journal of Economic
Behavior and Organization, 37 (2), 129–49.
Anonymous (2004a), ‘Marriage’, Judaism 101, http://www.jewfaq.org/marriage.htm.
Anonymous (2004b), ‘Divorce’, Judaism 101, http://www.jewfaq.org/divorce.htm.
Anonymous (2004c), ‘Fact Sheet: Equal Pay for Working Women’, www.alfcio.org.
Becker, G.S. (1974), ‘A Theory of the Allocation of Time’, Economic Journal, 75, 493–517.
Becker, G.S. (1981), A Treatise on the Family, Cambridge, MA: Harvard University Press.

Economics of marriage and divorce

167

Bergstrom, T. (1997), ‘A Survey of Theories of the Family’, in Mark Rozenzweig and Oded
Stark (eds), Handbook of Population Economics, New York: North Holland, pp. 21–74.
Bergstrom, T., L. Blume and H. Varian (1986), ‘On the Private Provision of Public Goods’,
Journal of Public Economics, 29 (1), 25–49.
Binner, J. and A. Dnes (2001), ‘Marriage, Divorce, and Legal Change’, Economic Inquiry, 39,
298–306.
Brien, M., L. Lillard and S. Stern (2006), ‘Cohabitation, Marriage, and Divorce in a Model of
Match Quality’, International Economic Review, forthcoming.
Browning, M., F. Bourguignon, P. Chiappori and V. Lechene (1994), ‘Incomes and Outcomes:
A Structural Model of Intra-Household Allocation’, Journal of Political Economy, 102 (6),
1067–96.
Cherlin, A. (1992), Marriage, Divorce, and Remarriage, 2nd edn, Cambridge, MA: Harvard
University Press.
Coase, R. (1960), ‘The Problem of Social Cost’, Journal of Law and Economics, 3 (1), 1–44.
Coelho, C. and N. Garoupa (2004), ‘Did Divorce Law Reforms Raise Divorce Rates? Evidence
from Portugal’, Draft, Universidade Nova de Lisboa.
Friedberg, L. (1998), ‘Did Unilateral Divorce Raise Divorce Rates? Evidence from Panel
Data’, American Economic Review, 88 (3), 608–27.
Goldin, C. and L. Katz (2002), ‘The Power of the Pill: Oral Contraceptives and Women’s
Career and Marriage Decisions’, Journal of Political Economy, 110 (4), 730–70.
Hoddinott, J. and L. Haddad (1995), ‘Does Female Income Share Influence Household
Expenditures? Evidence from Côte d’Ivoire’, Oxford Bulletin of Economics and Statistics,
57 (1), 77–96.
Lundberg, S. and R. Pollak (1993), ‘Separate Spheres Bargaining and the Marriage Market’,
Journal of Political Economy, 101 (6), 988–1010.
Lundberg, S., R. Pollak and T. Wales (1997), ‘Do Husbands and Wives Pool their Resources?
Evidence from the UK Child Benefit’, Journal of Human Resources, 32 (3), 463–80.
Manser, M. and M. Brown (1980), ‘Marriage and Household Decision Making: A Bargaining
Analysis’, International Economic Review, 21, 31–44.
McElroy, M. and M. Horney (1981), ‘Nash-Bargained Decisions: Toward a Generalization of
the Theory of Demand’, International Economic Review, 22 (2), 333–49.
Nash, J. (1950), ‘The Bargaining Problem’, Econometrica, 18 (1), 155–62.
Schultz, P.T. (1990), ‘Testing the Neoclassical Model of Family Labor Supply and Fertility’,
Journal of Human Resources, 25 (4), 599–634.
Thomas, D. (1990), ‘Intra-Household Resource Allocation: An Inferential Approach’, Journal
of Human Resources, 25 (4), 635–64.
United Nations (various years), Demographic Yearbook, New York: Department of Economic
and Social Aﬀairs, Statistical Oﬃce, United Nations.
United States National Center for Health Statistics (various years), Vital Statistics of the
United States, Hyattsville, MD: National Center for Health Statistics, United States Center
for Disease Control.
Weiss, Y. and R. Willis (1985), ‘Children as Collective Goods in Divorce Settlements’, Journal
of Labor Economics, 3 (3), 268–92.
Wolfers, J. (2003), ‘Did Unilateral Divorce Raise Divorce Rates? A Reconciliation and New
Results’, National Bureau of Economic Research Working Paper No. 10014.
Woolley, F. (1988), ‘A Non-Cooperative Model of Family Decision Making’, LSE Working
Paper, TIDI/125.
Youssef, H. (1993), ‘The History of the Condom’, Journal of the Royal Society of Medicine,
86 (4), 226–8.

PART III
BODY AND SOUL

Economics of pornography
Samuel Cameron

In this chapter, we look at the contribution economics can make to
analysing the behaviour of markets for pornography and to the formulation of public policy in this area. For instance, we show that economics can,
through standard supply and demand models, shed light on the impact of
regulatory policies on pornography and provide a framework that isolates
diﬀerent elements in the eﬀect of technological change on the long-run
equilibrium price and output levels of each type of pornography.
It should be noted from the outset, however, that as a general rule most
academic debate and teaching on the subject of pornography is to be found
within the realms of feminism, philosophy and legal theory. Economists
have made scant contribution with the odd exception. Judge Richard
A. Posner, for example, has repeatedly applied orthodox free market
Chicago thinking to porn in several of his books (see Economic Analysis of
Law, 1973; Sex and Reason, 1994; Frontiers of Legal Theory, 2001).
Heterodox economist David George (2001, pp. 122–4) brieﬂy discusses
porn as an instance of what he calls ‘preference pollution’. In this scenario,
free markets impose weakness of will upon us leading to the ‘wrong choice’
of consumption basket. We end up on a lower indiﬀerence curve than might
otherwise have been obtained.
Yet, there are perhaps two simple areas in which debate on porn is
focused, which could be summed up in these propositions:
1.

There is a huge and growing amount of porn, which threatens to undermine the stability of civilization as we know it. This is a position to
which the book we have just mentioned by George seems to subscribe.
Pornography generates extremely high proﬁts for its entrepreneurs.
This idea is to be found all over the campaigning literature by Christian
groups in the USA, despite the lack of reliable statistics on the porn
industry. It also inspired the witty, punning title of a book by attorney
Frederick Lane III, namely, Obscene Proﬁts: The Entrepreneurs of
Pornography in the Cyber Age, published in 2000.

This chapter cannot hope to comprehensively answer these charges as they
involve two highly contentious things: the reliability and accuracy of statistics and the rights and wrongs of the prevalence of porn in the world. The
171

172

Economics uncut

latter inevitably involves value judgements and thus any economist who
claims to be providing a clear-cut proof of the matter is likely to be involved
in some act of disingenuity. What we can do is provide a framework for
addressing these issues, which will illustrate novel uses of economic concepts for those already used to them and, for those unused to these concepts,
it will help to introduce them to their more widespread applicability.
The structure of the chapter is as follows. In Section 1, we look brieﬂy at
research outcomes on the eﬀects of pornography. Next, in Section 2, we
move on to the nature of porn as an industry just like any other industry,
which requires us to consider the issue of statistical evidence. Section 3 examines Amartya Sen’s argument about Paretian liberals as applied to pornography. Section 4 looks at the nature of the demand for pornography with
reference to addiction and variety seeking. Section 5 outlines government
approaches to regulation of porn while Section 6 uses some supply and
demand analysis to illuminate the issues of the eﬀect of the Internet on
pornography and, in particular, its impact on the eﬃciency of enforcement
eﬀorts by the regulatory authorities. Section 7 considers the existence of possible positive externalities arising directly from both the acts of consumption
and production of porn, as well as discussing whether porn should be viewed
as a complementary or substitute good. Section 8 concludes the chapter.
1 Social outcomes of pornography: negative externalities
Economic research on this subject is thin. As far as I am aware, there have
been no empirical studies of pornography in the major economics journals.
Further, I have searched the main economics working papers (that is, a
record of works in progress) websites and have not found any papers on
pornography.
Evans and Winick (1996) appear to provide the only econometric study
of the eﬀects of porn. They studied the non-operation of anti-pornography
statutes in four states in the USA during periods between 1973 and 1986.
Since non-enforcement of pornography statutes is associated with an
increase in the supply of pornographic materials, this provided an opportunity to examine the impact of statutes and pornography availability on
sexual oﬀences. They found that there were no statistically signiﬁcant
changes in rates for rape, prostitution or sex oﬀences during the statute suspension. Therefore, we could perhaps conclude from this one isolated study
that there is little evidence of any substantial externality eﬀect, in terms of
recorded deviant behaviour, from pornography.
Although there is a dearth of econometric studies, there have been a
number of comparable statistical studies by researchers in other ﬁelds in the
sense that aggregate statistics on sex crimes are correlated with some kind
of measure of porn output or changes in the regulatory regime of a region

Economics of pornography

173

(see, for example, Bryant, 2001; Gentry, 1991; Donnerstein et al., 1987;
Diamond and Uchiyama, 1999; Scott 1985; Fisher and Barak, 2001). The
obvious problem in such studies is eliminating factors, other than access to
porn, which may be responsible for spatial or temporal ﬂuctuations in the
rates of rape and related activities.
For what it is worth, the overwhelming evidence of the above studies is that
we cannot ﬁnd any clear correlation between porn quantity, style/content or
its regulation and the rate of sex oﬀending. Some commentators have also
pointed out that in the USA the era of increased porn supply (through VHS
tapes, cable and now DVDs and the Internet) has witnessed sharp falls in
rape rates. For example, the estimated total of 89 107 forcible rapes reported
to law enforcement during 1999 was the lowest total since 1985 and the rate
in 1999, 64 out of every 100 000 females, was 5 per cent lower than the previous year’s. The latest data show that in 2003 an estimated 93 433 rapes
were reported in the USA (all statistics are from FBI sources), 1.9 per cent
lower than estimates for 2002. Indeed, the overall level is now extremely low
in comparison with the early 1980s.
As we have already implied, one has to be cautious interpreting such
statistics as there are changes in sentencing, criminal justice expenditure,
police policy and user behaviour going on at the same time. Nevertheless,
it cannot be avoided that the expanded prevalence of porn does not seem
to be producing aggregate evidence of a negative externality eﬀect in the
form of sex crimes.
There is, of course, experimental research on the eﬀects of porn such as
the survey by Zillmann (1989), which seems to report some negative attitudinal outcomes of exposure to porn. However, mainstream positivist psychological laboratory research does face certain ethical and epistemological
problems in this area. That is, researchers can only deal with willing subjects for the experiment and they cannot reasonably be allowed to test
whether observed attitudinal changes would translate into actual behavioural changes of the type that has not been found in the aggregate data
analysis reviewed above. Experiments also suﬀer from the artifactual
feature of imposing a consumption style and environment on the subjects
that is diﬀerent from that which they would voluntarily choose in the real
world. This may induce some attitudinal changes that would not otherwise
occur when one is consuming in the desired way.
2

Characteristics of the industry

2.1 Production and consumption
Let us look at the scope for classification of porn as part of the economy.
Although it has no such thing as an SIC (Standard Industrial Classiﬁcation)

174

Economics uncut

coding in censuses of industrial production, it clearly makes sense to talk of
there being a ‘sex economy’, comprising explicitly marketed legal and illegal
porn along with other activities such as prostitution (the subject of Chapter
7), lap dancing and telephone sex chat lines. Some of the legal activity, or
that which takes place under the umbrella of legal production, will be partly
represented in production and consumption statistics.1
In terms of employment, the pornography industry sees many people in
categories other than simply being the actors or models. Additional
employment is generated in writing scripts and literature, photography,
marketing, website design, sales, legal advice and all the other ancillary
services found in other forms of production.
Moreover, economics suggests that the sector will also generate standard
Keynesian multiplier eﬀects of creating additional employment in other
sectors. This could be vital to local economies where there is specialization of the sector due to agglomeration and networking economies (for
example, the ‘sex tourism’ to places such as Amsterdam and Thailand).
A local specialization in porn production can also be a valuable source of
export earnings (albeit untaxed) due to heavy demand in other countries,
which in turn reﬂects major national diﬀerences in legal regulation of the
deﬁnition and availability of porn.
The most reliable statistics on porn are those produced by the porn ﬁlm
industry, which has a publication, Adult Video News, which is comparable
to the music industry’s Billboard. According to a newspaper report, ‘Wall
Street Meets Pornography’, by Timothy Egan in The New York Times in
2000, Americans buy or rent more than $4 billion a year worth of graphic
sex videos from retail outlets and spend an additional $800 million on less
explicit sexual ﬁlms.
There are no trustworthy statistics on the consumption and production
of porn on the Internet, although that does not stop people producing
many contradictory and unreliable estimates, mainly from surveys of poor
methodological content.2 Apart from the issue of people telling lies in
surveys, there are other problems with data in this area. Some ‘net porn’ is
free and accurate statistics would require logs of the traﬃc to providers and
this is not generally available. In addition, the same surveys are all that are
available on paid net porn and we do not have access to statistics on credit
card payments to these sites.
Even if one wanted to take a chance on using survey estimates on time
spent visiting porn sites, there is a further problem on what this time ﬁgure
really means. Everyone knows that ‘pagejacking’ in the form of pop-up
windows attempting to take you somewhere else is a regular feature of
surﬁng, especially for ‘free’ sites as the funding is likely to come from porn
pay sites. Thus, time spent surﬁng the net for porn sites may contain a large

Economics of pornography

175

and hard-to-measure fraction that is consumption search rather than
actual consumption.
The size of the illegal porn sector (bearing in mind that what is illegal in
some countries is not in others) is extremely hard to proxy. The only way to
do this would be if there were a regular and consistent seizure policy by the
regulatory bodies and it were fair to assume that this seized a fairly constant
portion of the total supply. This requirement is not met in reality, however.
For example, on the issue of consistency, recent UK policy on child porn
has been to put huge resources into targeting high proﬁle (consumer)
suspects, which means small amounts detected per case compared with a
strategy of spreading investigations across the sector.
2.2 Size distribution of ﬁrms
The porn sector is a very diverse one in terms of the size distribution of
ﬁrms. Due to the development of technology, there are now one-person
ﬁrms in the form of people who may provide still pictures, ﬁlm clips or
even interactive sexual performances via webcams. At the latter point, the
deﬁnition is crossing over to prostitution. There are also small ﬁrms distributing what one might call almost ‘home-made’ porn without trained
actors or scriptwriters. Traditionally, this might have been called ‘amateur’
porn (see O’Toole, 1998 for a discussion of amateur porn). However, the
use of the word amateur to describe a product being supplied by a full-time
producer and sold for a price is a little diﬃcult to handle in terms of economic analysis. Eﬀectively, amateur now really means one of two things:
1.

A genre or style of ﬁlm-making that might be seen as more ‘real’ than
the heavily stylized products found in commercial DVDs and on ‘pay’
television stations.
Models or webcam performers who do not have any kind of agent to
represent them. In some cases these may be genuine college students
brieﬂy subsidizing their education by a foray into the sex industry. In
Chapter 7, we provide evidence of this phenomenon in the UK in relation to prostitution.

At the other end of the scale, we have entities such as the Playboy corporation, which is a multi-product ﬁrm supplying magazines, DVDs, its own
TV channel and a website. For the traditional full-length feature-ﬁlm, there
are ‘majors’ in the same way that there are in regular movie production and
they do have an annual Oscar-style award ceremony.
However, due to the restricted area of consumption these ﬁrms cannot
approach the size of proﬁts to be made in the mainstream ﬁlm sector. In
theory, they will nevertheless receive an implicit subsidy as some workers in

176

Economics uncut

the porn sector enter it to acquire relevant human capital and contacts for
later use in mainstream ﬁlms. This means they should be willing to accept
lower wages than would otherwise be the case.
Perhaps the principal economic characteristic of the sector is that there
are few entry barriers. The main obstacle to new producers of pornography
is ﬁnding consumers. In the case of ﬁlms, this problem has decreased due
to the emergence of DVDs and pay-TV channels, which ﬁll the gap in
outlets due to the fact that in many parts of the world porn ﬁlms are not
allowed in conventional movie theatres and specialist porn cinemas tend to
be heavily restricted. Overall, barriers to entry in pornography may in fact
be lower than in some other niches of the sex industry such as prostitution,
as the emphasis on depiction and suggestion of sexual acts means that
‘looks’ per se are less important.
Within the self-employed Internet-based sector we observe the interesting phenomenon of cooperation between rivals in a sector approximating
perfect competition. Websites of individual models feature links to other
suppliers of the same product. This seems most unusual; we would be very
surprised to see plumbers, electricians or interior designers advertising the
services of their rivals. The explanation for the free advertising of site links
in porn could be that people are not suﬃciently impressed by a website
without many URL links and so all that is happening here is cooperation
due to desperation. A more convincing analysis, however, requires attention
to some non-standard features of consumption, which create network
externality gains for those who advertise (seemingly) rival suppliers. These
features in the shape of ‘variety demand’ and addiction are discussed in
Section 4.
2.3 Proﬁts
Let us brieﬂy address the issue of proﬁts that was raised in the introduction.
From an economic point of view, there are a whole host of reasons why we
would expect some individuals and some websites and corporations that
supply hotel pay-per-view porn movies to make large proﬁts (whether or not
we regard these as super-normal proﬁts in the textbook sense is a diﬀerent
matter). These are all points that apply to the sex industry in general:
1.
2.

In any industry some operators will make huge proﬁts, whilst many
other less well-known cases struggle and/or go out of business.
Porn is supplied in a dynamic technologically progressive industry and
therefore there will be standard ‘ﬁrst-mover advantages’ to producers
who establish a niche market or perfect a delivery method.
Compensating diﬀerentials – whilst the pornographer is still the recipient of social opprobrium, standard labour market theory dictates that

Economics of pornography

177

they should be receiving higher wages to compensate for the psychic
costs of being disliked.
Lack of non-proﬁt motivation – whilst there may be some ‘evangelical’
entrepreneurs who see the supply of porn as a crusade to help
humankind, it is likely that on balance the porn entrepreneur lacks
some of the non-proﬁt motives found in some other industries.

Given the above it is thus something very far from remarkable that some
porn suppliers make large proﬁts. Indeed, condemning this seems like some
kind of ‘moral double-counting’ from those who consider porn sinful, as
the act of production and consumption has already been condemned. One
is left wondering if those who make the proﬁt critique would be happier if
porn proﬁts were more evenly spread amongst the pornsters.
3

Defining pornography

3.1 The legal and philosophical perspective
We now address the thorny issue of the deﬁnition of pornography more
directly, as this is an area in which everyone thinks they know what the
word means, yet legal deﬁnitions have been diﬃcult to agree upon. The
original meaning of the word was as a description of the life and activities
of prostitutes. This derives from the Greek, pornographos (porne prostitutes and graphein to write). It acquired its present meaning in the nineteenth century and seems to have two key elements, one being the obscenity
of depiction of sexual acts (our principal focus for the remainder of this
section) and the other, the exploitation of workers in the production of the
material and/or the consumers of the material.
Therefore, analysis of pornography falls within the topic of censorship,
that is, state decisions on what citizens should and should not be allowed
to see. The obscenity charge includes such things as swearing and incitements to violence in cultural products like cartoons and rap and rock music
records. Much legislation in the past that has dealt with pornography has
also tended to be under the rubric of ‘obscenity’ but it has been very
diﬃcult to deﬁne pornography as a criminal oﬀence. Instead, it has been
frequently treated in legal matters as material ‘with a tendency to deprave
or oﬀend’ with the decision on how depraving or oﬀensive the matter being
left to a judge and/or jury.
American Supreme Court decisions have provided ample illustration of
the diﬃculties of arriving at an obscenity deﬁnition (see Sears, 1989 for
some history on this). According to Silver (1994), Justice Potter Stewart
pronounced that he would know pornography ‘when he saw it’. This
implies either that his preferences are representative of those that should

178

Economics uncut

dominate social decision-making or that they represent majority social
preferences. Two other judges quoted in Silver (1994), Justices White and
Brennan, were particularly keen to use erect penises as a dividing line
between porn and ‘erotica’ although they diﬀered on how much deviation
was allowed from traditional heterosexual intercourse. Justice Brennan’s
so-called ‘limp dick’ test has remained today as a dividing line in movie censorship. This is recognized by porn ﬁlm-makers who insist not only on erections but ejaculation as being essential to the production of a genuine (as
opposed to ‘soft’) porn movie (see O’Toole, 1998).
As alluded to above, lurking in the Supreme Court decisions has been the
notion of an ‘average person’ standard, that is, the judges are agents representing the median preferences of the population. This has been a motif
also of UK legal debate. For example, in the trial of D.H. Lawrence’s novel,
Lady Chatterley’s Lover, as an obscene book, the judge invited the jurors to
consider whether they would allow their servants to read it.
However, average person standards run into diﬃculties when we come to
child pornography, especially when it is intertwined with claims that a work
is ‘erotic art’. That is, can the ‘average person’ know that something is a
genuine work of artistic merit, particularly so if they have not invested in
the ‘art appreciation capital’ that emerges from Beckerian models of rational addiction? This seems then to require delegation of judgement to allow
some role for expert witnesses. A good example of this type of situation was
reported by BBC News on 27 March 2002 when Canadian author
J.R. Sharpe was found ‘not guilty’ of child pornography because his work
had some artistic merit. Several university professors compared his work
with Charles Dickens and James Joyce thereby causing the charge to fail.
This is only one of numerous cases all over the world where material
charged with being pornographic is defended as art.
Artistic merit is also a problematic element to bring into the deﬁnition
because we still seem to be driven back to the issue of the intent of the producer rather than the nature of the product. In economic terms, we could argue
that a particular book/ﬁlm may have artistic merit, which is one type of output,
but the porn content is a joint product. Thus, the purchaser gets two goods. Is
it then acceptable to supply porn just because it is jointly supplied with another
good that is deemed to be perhaps a ‘merit good’ by paternalistic policy advisors and governments? Even if one accepts this, it is diﬃcult to prevent the consumer skipping to the passage of a book or a ﬁlm that is deemed to be porn.
The logical response is to appoint censors to either cut out the oﬀending
segment or require it to be edited diﬀerently. If the joint outputs are neither
readily separable nor divisible, then this is not so easy. In such cases, it is
diﬃcult to resolve the argument that the alleged porn component cannot
be removed without eﬀectively destroying the whole product.

Economics of pornography

179

3.2 The economic perspective
But, why from an economist’s point of view should we be spending
time trying to deﬁne porn in the ﬁrst place? If individuals are rational
utility maximizers, then why do they need to be barred from pornography? Furthermore, if they do, why does an exception need to be made
for ‘art’?
Problems arise over the optimality of individual liberty in consumption
as discussed by Sen (1970) once we start to permit people’s disgust at other
people’s consumption to be classiﬁed as a negative (consumption-onconsumption) externality. Sen posed the dilemma for the Paretian liberal in
terms of permission to read Lady Chatterley’s Lover. In itself, this indicates
how the moral climate has shifted in English-speaking countries as there is
now little concern about such books (Henry Miller’s Tropics novels are
another example) being published in continental Europe.
Given this, we will reformulate Sen’s example in terms of the publication
of a book of graphic bondage photographs, which we will assume to be
with the consent of the models pictured therein and causing them no pain.
However, if the people concerned were rational masochists then any physical pain they suﬀer may, in fact, be a source of utility to them, which would
mean they will accept a lower equilibrium wage for the shoot than the nonmasochist would accept.
So, let us assume there are two people, Mr 1 who wishes that no one
should read the book and Ms 2 who wishes to read what we will call
bob (book of bondage). There are three possible outcomes as shown in
Figure 6.1: no one can be permitted to read the book or one can be made
to read it and not the other. Mr 1 is a paternalistic prude who, as just noted,
prefers that no one should read the book and ranks second the case where
only he should read bob as Ms 2 is at risk of welfare loss from being
exposed to bob (the ‘demerit good’ argument).
Ms 2, who might be described as an evangelical liberationist (that is, she
seeks to expose others to challenges to their sexual preconceptions), has

Mr 1

Preference ordering

Figure 6.1

Ms 2

No one reads book

Mr 1 reads book

Ms 2 reads book

No one reads book

Sen and the impossibility of a Paretian liberal

180

Economics uncut

as her least favoured option that neither should read bob. Her preferred
position is that Mr 1 should have to read the book due to a utility eﬀect
from the pleasure of shocking the prude. Her second best option is for only
her to read the book. According to Sen, ‘liberal’ society would nominate
that the ‘no bob’ situation is the optimal choice as the other situations
involve Mr 1 being forced to read the book.
This is just a simple example of a voting paradox, which would vanish if
all people had the same preferences. Yet, if all people had identical preferences there would be no need for a debate about censoring pornography,
but it is important to note the forced consumption element of the example
given. We can always argue that people have a free choice not to consume
porn. This seems to lie behind the greater tolerance of porn/obscenity in
some art forms than others. For example, cinematic movie releases and live
theatre tend to be less heavily censored than television provision.
The rationale for this would seem to be the diﬀerences in comparative
risk of those whose preferences may need protection (for example, children,
the prudish and easily shocked). Thus, although such people may not be
forced to watch sado-masochism, which is not essential to a dramatic plot
or an advertisement, there is clearly a cumulative risk that they may experience accidental forcing from inertia or channel hopping. The same argument applies to Internet ‘pagejacking’ by pop-up windows.
Some would argue (David Friedman, for example) that Sen’s argument
is a misreading of what ‘liberal’ means. He assumes the liberal wants each
person to have the controlling decision between some two states of the
world – where a state of the world describes everything that happens to
everyone. We would more normally assume that what a liberal wants is to
control certain choices relevant to her/himself only. For example, if I am a
liberal I will want to have control over what book I read and I would wish
that others have the same choice so long as no third party suﬀers.3
Further, there may be no conﬂict between liberal and Paretian outcomes
in a world where there are no obstacles to eﬃcient contracts. That is, the two
individuals in the above example could come to an agreement that the prude
should not read the book but the liberal can read the book. This does seem
to face the problem that a prude, by deﬁnition, would not want the liberal
to be allowed to read the book. It takes us into a regress of saying that it is
impossible to be a Paretian liberal in a world where not everyone is a liberal,
which I would assume to be the point Sen was making.
Indeed, we should pause to consider that it is questionable whether the
rather restrictive concept of Pareto optimality is ever likely to be of great
usefulness in analysing policy issues that are so heavily interwoven with disparate value judgements. Again, I would be inclined to think that this is the
point Sen was trying to make (following on from Arrow’s Impossibility

Economics of pornography

181

Theorem4) in order to provoke liberals to come up with a more convincing
welfare economics.
4 Demand analysis of pornography
We now consider the demand analysis of porn within the framework of the
routine assumptions of the undergraduate economics textbook, beginning
with a discussion of what constitutes the ‘price’ of porn to the consumer.
Following this, we look at some modiﬁcations to the textbook assumptions,
which may provide greater insights into the economics of porn.
4.1 The price of pornography
The price of porn to the consumer could be assumed simply as the list purchase price of such a title. Generally, there is little problem with this if the
porn is 100 per cent legal. However, if the porn is illegal then the market
price is no longer a true reﬂection of the marginal cost of acquiring a unit
of porn. Illegality adds some new costs to the list price:
1.
2.
3.
4.

Search costs of ﬁnding titles as illegality will impose transaction costs
due to lack of advertising and secrecy.
Additional costs following from (1) in the form of restricted opportunities to consume from the need to avoid detection.
Explicit costs of punishment by the legal system if caught such as ﬁnes
and prison sentences.
Punishment costs from beyond the legal system, for example, from a
partner or parent or through loss of a job or restricted promotion
opportunities due to consumption at work. Workers can be sacked or
disciplined for wasting work time surﬁng pornographic websites.
‘Rip oﬀ’ risks from unscrupulous suppliers. In the heyday of purchasing VHS videotapes from covert retail outlets in the UK, the situation
arose where an individual asked for a particular title and was asked to
wait. The vendor then went downstairs and wrote the desired title on
the box of an entirely diﬀerent ﬁlm (see O’Toole, 1998). Due to the
illegality of the product, the buyer does not have recourse to consumer
protection legislation.

Hence, as in the case of illicit drugs discussed in Chapter 1, the eﬀective or
‘full’ price of the product is higher than the price at the point of sale and
so illegality would lower consumption in some proportion to the extent that
it adds to the list price. This does, of course, illustrate the fact that even a
zero price does not mean a free good. A rational utility-maximizing worker
surﬁng a free site (or using a ‘hack’ into a pay site) on his or her work computer is not receiving a totally ‘free’ good.

182

Economics uncut

4.2 Addiction
An additional worry about porn, aside from any ‘social pollution’ of negative externalities, is that it may have addictive properties in consumption.
The psychological position on this is presented in Cooper et al. (1999).
From an economist’s perspective, addiction to porn may make any such
social pollution worse and it might also be used as a case for intervention
to protect the porn consumer from himself or herself.
If porn is an addiction (assuming for the moment this is the case), it is
not delivered, unlike illicit drugs, in direct chemical form. Does this make
any diﬀerence? According to the rational addiction model discussed in
Chapter 1, the answer is no. The important thing is that some mental
process creates a link between the marginal utility of consumption of units
of the good in diﬀerent periods. It does not matter whether chemicals
directly taken into the body or produced in the mind by external stimulation of the senses forge this link.
Within this framework, fears about porn are in terms of it being a
harmful addiction. This has been expressed by strict religions over all sexual
activities outside of those designed to produce children within marriage (see
Cameron, 2002). Such fears have, at times, given rise to claims that mental
and physical deterioration will befall the sexually indulgent individual.
The addicted porn user will have low levels of social interaction in the traditional sense of face-to-face contact. Some porn consumers are in a trading
network of social interaction, notably on the Internet due to its illegality, but
this is not the major feature of their consumption, which has low sociability.
Rather, it is the possible growth of a fantasy element where an individual
seeks to control an imaginary world in which he or she is more powerful and
successful than in the real one. We can illustrate this in terms of the
Becker–Lancaster goods characteristics/time allocation model of household
production. The aggregate fantasy production function can be written as:
Fantasy output f (time spent fantasizing, fantasy-related goods,
fantasy energy supplied, imagination, random error)
Individuals may substitute between diﬀerent fantasies according to the
relevant prices and the shadow prices of time. In the speciﬁc case of
pornography, the goods are magazines, videos or visits to websites, all of
which may be fairly time-intensive if fantasy output is to be created. One
can be more than one sort of fantasist. The obsessive stalker or fan of one
person demands low levels of variety as his or her utility is increased by
having more items from the same category (for example, more pictures of
the same person) because obsession by its nature has to be with a small set
of items or people. However, it should be noted that charged obsessives are

Economics of pornography

183

not typically porn collectors, as they may form their fantasies from collections of images that would not be deemed ‘obscene’ by the ‘normal’ person.
For instance, many cases of obsessive stalking concern ordinary people,
plus actors, singers and newsreaders, of whom there are no pornographic
images available.
4.3 Variety-seeking
The evidence suggests that the major porn markets seem to be characterized by high levels of variety-seeking (O’Toole, 1998). As already noted,
there are many overground Internet news groups in which porn consumers
trade, discuss and catalogue their collections. This is comparable to what
goes on in the general range of cultural products such as music and ﬁlms.
Our earlier observation that fantasy goods such as porn may be very timeintensive, relative to other goods, seems to imply low levels of consumption
unless a person has ‘idle’ time available to them or, in the opposite case, high
levels of income.
Fantasy may have the unusual (for a consumer choice model) property
of substantial zones of increasing marginal utility, and better still (or worse
still depending on your point of view) increasing returns to some of the
fantasy production function inputs. This strange eﬀect has been previously
remarked upon in a piece on the phenomenon of collecting by Troilo
(1999). The collector may have an increasing marginal utility for items in
the collection if the sequence of acquisition is appropriately structured.
5 Regulation of porn: methods and problems
Pornography attracts attention because it is perceived as a problem requiring government action. In terms of welfare economics, the simplest case for
controlling it is that it may be viewed, as mentioned earlier, as a form of
‘social pollution’. Therefore, policy towards porn can be divided into the
following distinct categories:
1.
2.
3.

Price controls.
Outright quantity controls such as banning.
Quantity control through licensing and regulation, in particular,
through zoning by age. A notable instance of this is to be found in the
1995 Child Protection and Obscenity Enforcement Act, passed in the
USA following the 1986 report of the Meese Commission. This
required porn ﬁlm-makers to keep detailed records proving that no
under-age performers were used. In theory, this should have shifted up
the cost curves of legitimate adult porn ﬁlm-makers and thereby
reduced equilibrium output of adult ﬁlms even though that was not the
ultimate aim of the policy.

184
4.
5.

Economics uncut
Moral suasion possibly by encouraging the consumer to feel guilty or
fearful of adverse mental and physical health consequences.
Product intervention via censorship. For mainstream ﬁlms, producers
may be required to make cuts in order to get an age certiﬁcate such as
an ‘18’ or a ‘12’. Without these cuts the ﬁlms would be condemned to
the limited circulation of the porn circuit.
Regulation of distribution channels. For example, a porn magazine
may be required to be shrink-wrapped in an opaque cover and/or
located in a position in a shop where it is not easily seen, by accident,
by children or others likely to take oﬀence. This is the origin of the
phrase ‘top shelf magazine’. In the case of broadcast porn, the title
may be restricted to channels that are not ‘free-to-air’ and can therefore be blocked to inappropriate consumers.

These options diﬀer greatly in terms of their costs of enforcement. Most of
them are considerably cheaper than the outright ban. Option (1), price controls, is seldom discussed. Sex industry products in general have, unlike
chemically addictive goods, not been treated with additional ‘sin’ taxation
in the modern era. They have either been made illegal or given normal tax
treatment, that is, workers pay conventional income taxes and products are
taxed at standard rates.
A porn tax might seem attractive (analogous to a Pigouvian pollution
tax) from an older welfare economics perspective, but in terms of more
modern welfare economics, the option of selling permits, or licences, to
porn merchants would seem a better price-based solution. While the sale
value of the permits could be determined at a level that is social welfare
maximizing, both approaches face the diﬃculty of being seen as dangerous
due to making porn seem like a legitimate business activity rather than an
‘evil’ that has to be somehow contained. In view of this, porn policy in
general terms veers towards the persuasion, prevention and punishment
approaches, the latter two being discussed in Section 6 in relation to the
impact of the Internet on pornography.
Yet, any policy faces problems of adaptive behaviour by producers and
consumers. One option is smuggling of products from more tolerant
nations. This ﬂared up in the 1970s with movements of material from the
Scandinavian nations to the UK and the USA, particularly after the 1969
Danish legalization of pornography, which included very liberal treatment
of child porn. Physical movements of goods across borders can be prevented by systematic surveillance, particularly as videos and books are
quite hard to disguise. Eventually, the liberality of legal Scandinavian child
porn was curbed, however, thus reducing the burden on customs agencies
abroad.

Economics of pornography
6

185

Supply and demand analysis of the effect of the Internet

6.1 Dirty postcards
The arrival of widespread access to pornography on the Internet can be
fairly said to have sparked moral panic about the threat posed to the stability of society. It is probably worth remembering that this has happened
before, albeit with a much more humble technological advance – the picture
postcard. The ability to mass-market postcard images cheaply only arrived
in the 1890s. Before this, porn was largely manifested in the written word
and expensive non-reproducible images. It was the preserve of the wealthier classes and managed to largely escape regulation despite the pressure of
organizations such as the National Vigilance Association and women’s
purity groups in the UK (Sigel, 2002). Even until recently it was common
to refer to ‘dirty postcards’.
A number of prosecutions in the UK in the 1890s brought to light the
distinction between porn and erotica in terms of the ﬁtness of the consumer. That is, the chief legal principle seemed to be that nakedness and so
on were not art when consumed by those of lower income and status. They
were deemed to be too easily corrupted and excited to be safe receiving such
material. This takes us back to modern anxieties over digital television and
now the Internet where concern is that porn is falling too readily into the
wrong hands.
Of course, the global miracle of the Internet and the allied development
of widespread digital copying have transformed all problems associated
with international trade in cultural goods. The most obvious important
innovation is the fact that the goods no longer need to be literally moved in
real time in a cumbersome format across national boundaries. Rather, they
can be e-mailed in compressed formats or downloaded from websites.
Given a high-speed connection, an individual can rapidly download a huge
collection of still pornographic images, although moving-image compression of acceptable picture quality is still much slower.
One consequence of all this is that the transaction costs of obtaining a
product are greatly reduced. In addition, it is now virtually impossible to
prevent the product moving from a territory where the content is legal to a
consumer into one where it is illegal. This is especially germane to the issue
of child porn where the age limit is 18 in the USA but much lower in many
other places. This creates incentives for suppliers to locate on servers in the
most tolerant countries. Even if only one person in the prohibiting nation
downloads a new collection it is now fairly costless for this collection to be
proliferated if the individual belongs to a covert-sharing group. The digital
dimension and its accompanying cost falls have added fuel to the kind of
moral panic felt in the heyday of ‘dirty postcard’ prosecutions.

186

Economics uncut

Of course, the Internet is of considerable assistance to suppliers of illegal
‘hard’ porn, although the same resources are available to the policing
authorities. It follows that laws can be passed and strategies formulated to
regulate pornography on the Internet. This consists of two main aspects:
detection and punishment and, on the other hand, prevention (for legal
discussion, see Akdeniz, 1997).
6.2 Detection and punishment
Police surveillance of Internet activity includes ‘cottaging’ their way into an
illegal porn ﬁle-sharing newsgroup and hacking into people’s supposedly
‘private’ e-mails. This is often backed up by high-proﬁle ‘naming and
shaming’ and stiﬀer prison sentences for supplying the material. One would
expect that if the consumers were rational utility maximizers that they
would reduce their consumption levels as every act of transaction now
carries a higher potential cost. Indeed, the important variable for deciding
the optimal level of enforcement is the elasticity of porn consumption with
respect to the punishment. Unfortunately, we know nothing about this.
However, we can be sure that the consumer of forbidden porn increases
his or her risk of punishment greatly if he or she has an archive of porn
that is not easily hidden. Once magazines have been purchased or images
downloaded they have to be stored for repeat consumption (assuming that
the marginal utility of gazing does not decline so rapidly that they become
almost instantly valueless upon storage). The larger the collection, the
greater is the risk of detection by signiﬁcant disapproving others (parents,
partners). Moreover, the size of the collection may also inﬂuence the size
of guilt/shame punishments inﬂicted by these signiﬁcant others as it may
be easier to escape with a small punishment if a convincing case can be
made that it was idle curiosity or a temporary weakness.
While a digital collection of images is much more easily hidden than a
magazine, book or videocassette, cases are common of people being
charged due to porn collections on the hard disk of their home or work
computers. Indeed, the rapid fall in price and increase in capacity of ﬂash
memory cards, such as those used in digital cameras, provides great opportunities for evading detection as these are easily concealed. They can be
read through a USB card reader attached to the machine when consumption is desired or, in an emergency, can probably be ﬂushed down the toliet.
6.3 Prevention
Prevention of access to Internet porn requires that the would-be consumer
is unable to access the site. This can be accomplished by speciﬁc software
that refuses connection to a site. Such software (for example, Cyber Patrol
and Net Nanny) is aimed at parents wishing to prevent children seeing porn.

Economics of pornography

187

Alternatively, many pay porn sites now require the purchase of an adult
‘ID key’ in order to prove that the consumer is legally entitled to view the
material. ISP providers can also lock certain sites as unsuitable for their
client base and government regulation requires that they comply with the
pornography statutes in their own countries. This is not particularly
diﬃcult to overcome, however. Indeed, all prevention measures can be circumvented at some level of eﬀort/cost by those whom they were meant to
stop. The level of such costs depends essentially on the computer literacy
of the person being prevented. In fact, more information technology training for schoolchildren may have the unintended by-product of making
them more proﬁcient at ‘hacking’ through Internet security.
6.4 Marginal productivity of porn preventers and consumers
The precise eﬀect of the advent of the Internet depends on the relative
amount by which it shifts the marginal productivity schedules of ‘preventers’ versus those of consumers. This is illustrated on a supply and demandstyle diagram in Figure 6.2. The axes show the quantity of illegal porn and
the risk of being caught and punished. The demand-style schedule (D) is
downward-sloping on the assumption that the consumer treats risk as
exogenous and is more willing to consume porn when risk is lower, ceteris

EC2

Risk
in
consumption

EC1

EC3

R1
R4
D2
D1
Q3

Figure 6.2

Regulation of pornography on the Internet

Q4 Quantity of porn

188

Economics uncut

paribus. This curve can move due to changes in the rate of consumption by
existing consumers and changes in the participation rate (that is, the fraction of people consuming). The enforcement curve (EC) shows the risk that
enforcement agencies are able to produce. Other things being equal, they
should be able to increase the number of porn arrests when there is more
pornography in existence as there are more additional trades to detect. If
the law becomes easier to enforce, this curve will shift upwards.
Let us suppose that the schedules D1 and EC1 represent the pre-Internet
world in a fairly anti-porn country where there is a settled legal framework.
Let us also assume that there is a fairly elastic supply of porn from the rest
of the world. The model has an equilibrium risk (R1) and volume of porn
consumed (Q1), which should be stable because a fall below R1 will lead to
excess demand, which drives the risk up to R1 and a rise above R1 will lead
to excess supply, which drives the risk down.
Now imagine the Internet suddenly arriving in this scenario by shifting the
D function to the right to D2 but leaving the EC function unchanged. Part of
this shift is due to increased participation as many who would not have overcome the obstacles to obtaining porn via other methods now have access to
the Internet. This leads to a higher equilibrium consumption rate at Q2 but
also a higher equilibrium risk at R2 and is a characterization of the very
short-run impact where the authorities have not responded to the new situation.
If they can develop strategies that mean they now ﬁnd arrests easier to
produce then, ceteris paribus, the EC1 schedule shifts to EC2 causing a
further rise in equilibrium risk (R3) and a decrease in the volume of porn
traded (Q3). On the other hand, the productivity of punishment eﬀort may
decrease. This would be particularly likely in the medium term when legislation has not been adjusted to accommodate changed circumstances and
is shown as a fall in EC to EC3 giving a lower risk of detection (R4) than
the original post-Internet equilibrium and an increase in the volume of
porn traded (Q4).
7

A reconsideration of porn as an economic good

7.1 Social beneﬁts
So far, we have been treating all pornography as a negative externality along
the lines of pollution. If this were the case, then we simply face a trade-oﬀ
between the marginal social costs of deterrence versus the marginal social
beneﬁts of reduction. But the picture would be radically changed if it were
to be argued that there are, in fact, social beneﬁts of pornography. Clearly,
the optimal amount of resources devoted to its prevention would decline
and, in the extreme case, we could ﬁnd an optimum of legality of all porn.

Economics of pornography

189

Denmark was the only country where the possible positive eﬀects of the
availability of child pornography played a role in the political debate; it was
suggested that child pornography could serve as an outlet for tensions that
might otherwise lead to child abuse. This argument was proposed by the
criminologist Berl Kutchinsky and has been reformulated in various works
(Kutchinsky, 1970, 1985; Kutchinsky and Snare, 1999). He produced some
evidence of correlations between availability of hard pornography and a
decrease in reported cases of voyeurism and sex oﬀences against children.
Aside from outlet beneﬁts, it has even been argued that the demand for
porn has induced the suppliers of porn to stimulate technological advances,
which have brought considerable beneﬁts to non-porn consumers. This was
pointed out in an article by John Arlidge in the 3 March 2002 edition of
The Observer (UK) newspaper entitled, ‘What’s Porn Done For Us?’, where
he lists these spin-oﬀs:
1.
2.

Camcorder and VHS video machines were pioneered by porn producers who wanted to mass-market their goods.
Take-up of DVD players was accelerated by pornographers and their
customers because the technology enabled users to skip to and from
their favourite scenes.
Pay-per-view cable or satellite TV movies entered the market for
general movies only after porn ﬁrms introduced ‘premium’ services in
hotels and on digital networks.
Interactive television, common on digital sport channels, was developed by pornographers to allow users to focus on favourite actors and
actresses.
The Internet is used for many things other than porn and the proﬁts
and scale eﬀects generated by porn provide beneﬁts to non-porn users.

A further point not mentioned by Arlidge is that it is likely that the continuing collapse of region code protection in the DVD market is due to the
widespread international trade in porn DVDs. This beneﬁts non-porn consumers by expanding freedom of choice and it helps remove a barrier to
international price competition. In addition, it may be argued that widely
available porn can potentially improve the quality of other art/entertainment goods (such as mainstream Hollywood movies) by reducing the incentive to supply ‘watered down’ porn in an attempt to attract porn-rationed
consumers.
7.2 Consuming porn: substitute or complement?
Returning to economic theories of consumption, we might usefully restate
the case as to whether porn is most appropriately regarded as a complement

190

Economics uncut

or substitute to sex industry products, or is it rather an input to a production process. Some users may regard it as complementary to their normal
consumption of sexual satisfaction. Some makers (female, in particular) of
softer porn movies claim that couples may use them to revive their relationships. In this situation, the negative externality seems much smaller if
we rule out the non-liberal ‘disapproval’ factor by other people. The
substitution case is where pornography provides someone with an alternative to sexual outlets involving active interaction with other people (see
Cameron and Collins, 2000). This may be a positive temporary (therapeutic even) usage before they re-enter conventional relationships. Clearly, here
there is the ‘safety valve’ beneﬁt that can also be claimed for other sex
industry products.
8 Conclusion
Pornography is a subject with wide economic ramiﬁcations. Goods that
might be regarded as porn generate considerable amounts of employment,
earnings and output. To the extent that this is illegal, there will be an understatement of national income. There will also be a loss of potential tax revenues from making production illegal. Where porn is legal, those who object
to it are in the position of regarding it as a ‘bad’ (social pollution), the inclusion of which overstates rather than understates the true real national
income. The amount of any such overstatement would be even greater if it
were decided to adjust real national income to take account of utility losses
from crimes, oppression and stress that might be thought to follow from the
presence of porn in the economy. However, there are two problems with this
that have been highlighted by the discussion in this chapter:
1.

To date there is no convincing scientiﬁc research that shows that
pornography has harmful eﬀects on its employees or consumers. Even
if disturbed personality traits can be found among consumers and
workers, this may simply represent a ‘self-selection’ bias whereby those
with such problems are attracted to these products but their problems
are not originated or worsened by porn output.
There is immense diﬃculty in agreeing what constitutes ‘porn’. To the
person who regards anything they personally object to, of a sexual
nature, as pornography, the distinction may appear obvious. However,
legal action requires clear-cut criteria that can be applied. The more
unclear these criteria are, the more deadweight loss is inflicted upon the
economy in terms of resources used up in complex legal cases.
Economics does not oﬀer a way of resolving disputes about deﬁnition
within the traditional welfare economics framework of sovereign preferences. Rather, Sen’s 1970 paper on the impossibility of a Paretian

Economics of pornography

191

liberal shows this approach does not enable us to say much about the
direction policy should take. One can attempt to break through this
straitjacket by delving into the possibility of more complex preference
functions. Endogenous preferences where the porn seller is creating
additional wants in the buyer is one possibility and irrational addiction
is another. Even so, we still face the problem that it is very hard to
produce empirical evidence to support these factors as grounds for
intervention.
Notes
1.

2.
3.

Some porn production and consumption will be mixed up in the statistics for other production and consumption unless explicit attempts were made to identify it. Government
household surveys may ask about spending on magazines, ﬁlms and so on but not
whether they were pornographic. A certain amount of the output of the leather industry, for instance, may be from inputs into the production of bondage equipment, which
ends up in pornographic ﬁlms or still pictures or webcam interaction. Yet, there is no way
of knowing what proportion this is.
A good guide to all of this can be found at http://www.caslon.com.au/xcontentproﬁle.
htm.
Certainly Sen’s deﬁnition of liberalism under footnote 1 of his paper is one that seems
to interpret liberalism as a form of dictatorship under certain pairs of alternatives. He
expands on the footnote of this stating that, ‘liberalism is elusive and open to alternative interpretations. Some uses of the term may not embrace the condition described
here, while many uses will. I do not wish to engage in a debate on the right use of the
term’.
Kenneth Arrow’s (1951) ‘impossibility theorem’ suggests that no voting system can generate consistent social preferences from conflicting views about how society ought to be
organized.

References
Akdeniz, Y. (1997), ‘Governance of Pornography and Child Pornography on the Global
Internet: A Multi-Layered Approach’, in L. Edwards and C. Waelde (eds), Law and the
Internet: Regulating Cyberspace, Hart Publishing, pp. 223–41.
Arlidge, J. (2002), ‘What’s Porn Done For Us? You May Disapprove of Pornography But the
Spin-oﬀs of the Industry Are All Around’, The Observer, 3 March.
Arrow, Kenneth J. (1951), Social Choice and Individual Values, London: Chapman and Hall.
Becker, G.S. and K.M. Murphy (1988), ‘A Theory of Rational Addiction’, Journal of Political
Economy, 96 (4), 675–700.
Bryant, P. (2001), ‘Using Crime Mapping to Measure the Negative Secondary Eﬀects of Adult
Businesses in Fort Wayne, Indiana: A Quasi-Experimental Methodology’, paper presented
to National Institute of Justice’s Crime Mapping Research Center’s 2001 International
Crime Mapping Research Conference: Dallas, Texas.
Cameron, Samuel (2002), The Economics of Sin: Rational Choice or No Choice At All?,
Cheltenham, UK and Northampton, MA., US: Edward Elgar.
Cameron, Samuel and Alan Collins (2000), Playing the Love Market: Dating, Romance and
the Real World, London: Free Association Books.
Cooper, A., C.A. Scherer, S.C. Boies and B.L. Gordon (1999), ‘Sexuality on the Internet: From
Sexual Exploration to Pathological Expression’, Professional Psychology: Research and
Practice, 30 (2), 154–64.
Diamond, M. and A. Uchhiyama (1999), ‘Pornography, Rape, and Sex Crimes In Japan’,
International Journal of Law and Psychiatry, 22 (1), 1–22.

192

Economics uncut

Donnerstein, E., D. Linz and S. Penrod (eds) (1987), The Question of Pornography: Research
Findings and Policy Implications, The Free Press – A Division of Macmillan, Inc.
Edwards, D.M. (1992), ‘Politics and Pornography: A Comparison of the Findings of the
President’s Commission and the Meese Commission and the Resulting Response’
(http://home.earthlink.net/~durangodave/html/writing/Censorship.htm).
Egan, T. (2000), ‘Wall Street Meets Pornography’, The New York Times, 23 October.
Evans, J.T. and C. Winick (1996), ‘The Relationship Between Non-enforcement of State
Pornography Laws and Rates of Sex Crime Arrests’, Archives of Sexual Behavior, 25 (5),
439–53.
Fisher, W.A. and A. Barak (2001), ‘Internet Pornography: A Social Psychological Perspective
On Internet Sexuality’, Journal of Sex Research, 38 (4), 312–23.
Gentry, C. (1991), ‘Pornography and Rape: An Empirical Analysis’, Deviant Behavior:
An Interdisciplinary Journal, 12, 277–88.
George, David (2001), Preference Pollution: How Markets Create the Desires We Dislike,
Ann Arbor: The University of Michigan Press.
Hardy, S. (1998), The Reader, The Author, His Woman and Her Lover: Soft-core Pornography
and the Heterosexual Man, London: Cassel.
Kutchinsky, B. (1970), ‘Studies on Pornography and Sex Crimes in Denmark: A Report to the
US Presidential Commission on Obscenity and Pornography’, Copenhagen: New Social
Science Monographs, London.
Kutchinsky, B. (1985), ‘Pornography and its Eﬀects in Denmark and the United States:
A Rejoinder and Beyond’, in R.F. Thomasson (ed.), Comparative Social Research,
Greenwich, CT: JAI Press, vol. 8, pp. 301–30.
Kutchinsky, Berl and Annika Snare (eds) (1999), ‘Law, Pornography and Crime – The Danish
Experience’, Scandinavian Studies in Criminology, 16, Norway: Pax Forlag.
Lane, Frederick S. III (2000), Obscene Proﬁts: The Entrepreneurs of Pornography in the Cyber
Age, New York: Routledge.
O’Toole, Laurence (1998), Pornocopia: Porn, Sex, Technology, and Desire, 2nd edition,
London: Serpent’s Tail.
Posner, Richard A. (1973), Economic Analysis of Law, Little Brown.
Posner, Richard A. (1994), Sex and Reason, Cambridge, MA: Harvard University Press.
Posner, Richard A. (2001), Frontiers of Economic Theory, Cambridge, MA: Harvard
University Press.
Scott, J.E. (1985), ‘Violence and Erotic Material: The Relationship Between Adult
Entertainment and Rape’, paper presented at the annual meeting for the American
Association for the Advancement of Science, Los Angeles, California.
Sears, A.E. (1989), ‘The Legal Case for Restricting Pornography’, in Dolf Zillmann and
Jennings B. Bryant (eds), Pornography: Research Advances and Policy Considerations,
Lawrence Erlbaum Associates.
Sen, A.K. (1970), ‘The Impossibility of a Paretian Liberal’, Journal of Political Economy, 78,
152–7.
Sigel, L.Z. (2002), Governing Pleasures: Pornography and Social Change in England, 1815–1914,
New Brunswick, New Jersey and London: Rutgers University Press.
Silver, J. (1994), ‘Movie Day at the Supreme Court or “I Know It When I See It”: A History
of the Deﬁnition of Obscenity’ (http://library.lp.ﬁnd law.com).
Tomkinson, M. (1982), The Pornbrokers, London: Virgin Books.
Troilo, G. (1999), ‘Collecting’, in P.E. Earl and S. Kemp (eds), The Elgar Companion to
Consumer Research And Economic Psychology, Cheltenham, UK and Northampton, MA,
US: Edward Elgar, pp. 88–92.
Zillmann, D. (1989), ‘Eﬀects of Prolonged Consumption of Pornography’, in Dolf Zillmann
and Jennings B. Bryant (eds), Pornography: Research Advances and Policy Considerations,
Lawrence Erlbaum Associates.

Economics of prostitution
Peter G. Moﬀatt

This chapter applies basic principles of economics to the market for prostitution services. Section 1 attempts to identify the position taken and the
role played by this market within the commercial world, paying close attention to its interactions with other markets, and to the ways in which it is
evolving over time. Section 2 draws some international comparisons of
legal status and industry size, and then in Section 3, the chapter examines
some theoretical models of prostitution, seeking among other things to
explain how equilibrium earnings are so high in a profession with only rudimentary skill and capital requirements. The distinction between the
‘primary’ market of marriage and the ‘secondary’ market of prostitution
plays a central role throughout the economic analysis. Within this framework, particular attention is paid to the concepts of opportunity cost, compensating wage diﬀerentials, price discrimination, asymmetric information
and welfare considerations. In Section 4, the chapter moves on to examine
the empirical evidence relating to prostitution, focusing on an econometric
model of the price of sexual services. Several economic concepts emerge in
this empirical section, including age–earnings proﬁles, strategic behaviour,
consumer surplus and sunk costs. Section 5 examines the issue from a
policy perspective, describing and evaluating various models of legalization. Section 6 concludes the chapter.
1 Introduction
The prostitution industry is huge and operates in some form, either legally
or illegally, in every corner of the world. The sheer size of this largely
untaxed market makes it of considerable interest to economists. Although,
as we shall see in Section 2, data sources are notoriously imperfect and it is
diﬀicult, if not impossible, to obtain accurate estimates of levels of activity in this market, a rough idea is provided by The Economist (Anonymous,
1998), which estimates total global income from prostitution to be in excess
of US $20 billion per annum.
It is necessary to deﬁne prostitution at the outset. A dictionary deﬁnition
might be, ‘the undertaking of sexual actions for payment’. Edlund and
Korn (2002) cite references that raise a logical problem with this type of
deﬁnition: namely that some married women satisfy the deﬁnition routinely
within marriage. Edlund and Korn therefore prefer a deﬁnition in terms of
193

194

Economics uncut

payment for sex that is non-reproductive from the purchaser’s point of
view. Such a deﬁnition is adopted here.
By far the largest part of this industry is the sale of sex by female prostitutes to male clients, and it is on this market that we shall focus in the
chapter. Provision by males to homosexual male clients is a small market
by comparison, as is provision by males to female clients. Provision by
females to female clients appears to be virtually non-existent.
From an economist’s point of view, a key feature of the prostitution
market is that it is a ‘secondary’ market. In using this terminology, we are
following the contribution of Doeringer and Piore (1971) who apply the
idea to the labour market: the primary segment of the labour market is
characterized by high wages, good working conditions and chances of
advancement, and employment stability; the secondary segment is characterized by low pay, poor working conditions and lack of stability and
advancement prospects. Here, as suggested by Cameron (2002), we are
drawing an analogy in the market for sex: the ‘primary’ market is the
market for marriage (considered in Chapter 5 of this volume), where both
parties derive long-term beneﬁts, analogous to career advancement, from a
lasting relationship. In comparison, the beneﬁts from either consuming or
supplying commercial sex are strictly short term and highly superﬁcial. By
distinguishing the secondary market from the primary in this way, it is possible to develop an economic model that solves the wage diﬀerential puzzle;
that is, a model that is capable of explaining the signiﬁcant excess of the
wage earned in the prostitution industry over that earned, with similar skill
levels, in other industries. This economic model provides the focal point of
this chapter, and is analysed in detail in Section 3.
Another reason why the distinction between the primary and secondary
market is useful is that it allows clear explanation of the ways in which the
existence of the prostitution market may be seen to enhance societal welfare,
acting as the overﬂow mechanism necessary for the non-commercial market
to function in equilibrium. These welfare considerations are discussed in
Section 3, and again in Section 5, where we discuss the implications of
toleration and legalization from a policy perspective.
Apart from its intimate link to the marriage market, there is no doubt
that the prostitution market also has strong interdependences with other
markets, for example pornography (Chapter 6), drugs (Chapters 1–3) and
crime (Chapter 4). The link to pornography arises because pornographic
material may be seen as a substitute for prostitution services, and the
recent explosion in the availability of Internet porn may well have brought
about a fall in the demand for prostitution. Indeed, a German brothel
manager is quoted in The Economist (Anonymous, 1998) as explaining
his drop in business over the previous ten years with, ‘too much sex on

Economics of prostitution

195

television; why should people pay here when they can get it for free just
sitting at home’.
The strong interdependences between the prostitution market and the
drugs and crime markets are described in down-to-earth terms in a UK
Home Oﬃce Report (May et al., 1999). This report demonstrates using case
studies how each market feeds and perpetuates the others. A more recent
report (Home Oﬃce, 2004) contains a survey of eight studies that investigate the relationship between prostitution and drugs: in all eight studies,
more than half of the prostitutes sampled were users of Class A drugs. The
conjugate existence of such markets is clearly socially undesirable, since it
leads to vulnerable young people becoming trapped on a highly dangerous
treadmill, often at the mercy of unsavoury and unscrupulous individuals
such as drug-dealers, pimps and violent clients. Providers are also at a high
risk of contracting sexually transmitted diseases (STDs) and serious diseases associated with substance abuse, and passing them to their clients. In
such an environment, clients are also at risk of oﬀences such as ‘clipping’,1
which place the providers at the further risk of reprisal by infuriated clients.
In recent years, the link between prostitution and migration has taken on
new levels of importance. For example, of the 500 000 prostitutes estimated
to operate in Western Europe, around 250 000 are estimated to be immigrants (EUROPAP regional reports).2 This is an especially important
policy issue if it is believed that a signiﬁcant proportion of such immigration is ‘forced’. Although there is some recent evidence of the evil of human
traﬃcking,3 there is also a view, expounded in a Spectator article in April
2003, ‘Happy hookers of Eastern Europe’, by Phelim McAleer, that the
problem is not nearly as severe as some campaigners claim.
Having drawn attention to the unsavoury side of the industry, a balanced
treatment also requires consideration of the ‘respectable’ side. There is an
enormous variation in the types of environment in which sexual services are
provided and in the prices paid. At the bottom end of the spectrum, we see
street-girls who normally deliver outdoors or in clients’ cars, and might
charge between US $5 and US $100 depending on geographical location.
At the top end, we see high-class ‘escorts’ who might charge over US $1000
for a lengthy session in luxurious surroundings.
It is natural to ask why the variation in price is so much higher than that
observed for other types of service such as hairdressing or restaurant meals.
The answer has two parts. First, there is, as with other types of service, a
good deal of variation in the quality of provision, both in terms of the
physical and social attributes of the provider, and in terms of the dedication with which the service is provided. Second, high prices are used by
both parties to the transaction as a ‘screening’ device. On the supply-side,
high-class providers use their price signals to attract customers of the

196

Economics uncut

safest, healthiest, cleanest and most intellectually reﬁned variety. On the
demand-side, wealthy clients are attracted by high prices because they perceive them as a signal of quality. Recall the terms under which the Richard
Gere character secured the services of the Julia Roberts character in the
1990 movie Pretty Woman. The idea here is similar to that underlying
Rosen’s (1981) ‘Economics of superstars’, where it is recognized that, for
wealthy consumers, money is no object when quality of service is paramount (see Chapters 12 and 14, which explore similar ideas in the more
conventional context of sporting and musical superstars). The simultaneous use of screening on both sides of the market is one of the reasons why
equilibrium in the market can be maintained in the face of such enormous
price variation.
One of the reasons why the prostitution industry might be looked upon
diﬀerently from others is the social stigma associated with selling or purchasing sex, and the generally poor reputation of the profession. However,
there is a large volume of anecdotal evidence that the ‘taboo’ of prostitution is being gradually lifted. A Spectator article in November 2002,
‘Hookers at sports day’, by John Gibb, describes cases of middle-class
married females in the UK who have recently taken up prostitution in
response to economic misfortune or in order to ﬁnance luxuries such as
children’s private schooling. There is also recent evidence in the UK of
signiﬁcant numbers of students turning to prostitution in order to ﬁnance
their studies following changes in the student grant system: allegedly more
than half of all sex workers in Leeds are full-time students (Chapman,
2001). Isaacs (1993, p. 204) quotes a London University philosophy
undergraduate who claims proudly to earn £5000 per month in an escort
agency, more than enough to support her studies. That well-educated
people are prepared to enter this industry and to talk openly about their
experiences in it is perhaps evidence of a softening of attitudes towards
commercial sex.
An interesting issue is how high-class providers are able to attract clients
in environments in which commercial sex is illegal, in ways that do not
tarnish their ‘clean’ reputations, or make them liable to prosecution. One
method is to include in their advertisement a disclaimer to the following
eﬀect (variants of which appear in many UK websites):
NB: I am a professional independent escort and any money paid to me is for my
time and companionship ONLY. Anything else that occurs between us is a
matter of choice between consenting adults.

Of course, such façades do not always fool the authorities, as discovered
recently by high-class ‘madam’ Margaret MacDonald. British-born
MacDonald ran an international business with more than 400 male and

Economics of prostitution

197

female escorts, each charging around €1000 per hour, of which she took
40 per cent. In court, she admitted running the escort service, but vehemently denied being involved in prostitution in any way, claiming that her
role was simply to introduce clients to escorts. The French court was
unsympathetic and jailed her for four years in 2003 for, ‘aggravated procuring for the purpose of prostitution’.4
An important development of recent years is that the supply-side has
become much better organized, with sex workers’ unions emerging in many
countries. One union, which has been in existence since 2000, is the
International Union of Sex Workers5 (IUSW) whose terms of reference
include promotion of decriminalization, access to health clinics, the promotion of re-training programmes for sex workers wishing to leave the profession and legal assistance for those wishing to sue those who exploit their
labour. It also campaigns for an end to social attitudes that stigmatize those
who are or have been sex workers. In March 2002, the IUSW became
aﬃliated to one of Britain’s largest unions, the GMB, perhaps a further sign
that the industry is moving towards the commercial mainstream.
In Section 4, we pay attention to the demand-side of the market, principally using empirical analysis. In the past, the econometric modelling of the
demand for prostitution services has been made diﬃcult by the obstacles to
data collection.6 The problem is that both providers and clients are from
‘hidden’populations and it is unrealistic to expect either group to supply reliable information on their activities when traditional data collection methods
are used. This situation changed in the 1990s with the emergence of the
Internet. A large number of websites have emerged as information exchanges
for clients, and such sites have, albeit unintentionally, provided a rich data
source for statistical analysis. One of the most famous is the Punternet site,
data from which is used in our empirical analysis of Section 4.
2 Overview of world prostitution
The question of whether prostitution is legal in a given country rarely has
a simple answer. As we shall see in Section 5, there are many diﬀerent
approaches to decriminalization and legalization, and so a wide spectrum
of degrees of legalization is observed. According to information extracted
from Table 10.1 of Cameron (2002), countries with the most lax policies
include Austria, the Dominican Republic, Singapore and Switzerland. In
these countries, not only is prostitution fully legal, but the operation of
third party gainers (for example, ‘pimps’ or ‘madams’) is also permitted.
Next in the spectrum we have countries where prostitution is legal, but
third party gainers are not permitted. These include Australia, Germany,
Honduras, Hungary and the Netherlands. Many other countries appear
to have ambiguous systems, for example, China, Japan and the United

198

Economics uncut

Arab Emirates, where prostitution is illegal but tolerated, and the USA,
where it is legal in certain counties within the state of Nevada, but illegal
everywhere else. A further complication is that countries often change
their legal status, the most recent example being New Zealand, whose parliament voted narrowly for legalization in May 2003. At the harsh end of
the spectrum, we have a small number of countries in which prostitution
is illegal and punishable by death. To provide one recent example, in
February 2001, more than 1000 people watched while two women convicted of prostitution were hanged in Afghanistan’s Kandahar sports
stadium.7 Similarly, harsh penalties are evident in Iran and, until the
recent regime change, in Iraq.
Whatever the legal status of prostitution, it seems to occur in some form
everywhere in the world. Obtaining estimates of levels of activity in
diﬀerent locations is an awkward task due to the notorious unreliability of
most data sources. It is particularly hard to obtain reliable data from developing countries, and although it is widely believed that they have higher
levels of prostitution than the developed world (Boonchalaksi and Guest,
1998), it is hard to conﬁrm this claim empirically. The claim is consistent
with the commonly assumed positive association between poverty and
Table 7.1

Number of prostitutes in Western European countries

Country

Austria
Belgium
Denmark
France
Finland
Germany
Greece
Holland
Italy
Luxembourg
Norway
Sweden
UK

Population
(millions)

Number of
prostitutes
(thousands)

Population
per prostitute

8
10
5
59
6
80
11
16
58
0.45
4.5
8.5
60

6
12
6
40
4
150
12
25
60
0.30
3
2.5
80

1333
833
833
1475
1500
533
880
640
967
1500
1500
3400
750

Source: EUROPAP regional reports; collated by the Department of Politics at the
University of Exeter (http://www.ex.ac.uk/politics/pol_data/undergrad/aac/scale.htm).
The ﬁgure for Germany is extracted from Morell (1998). All ﬁgures relate to the situation
pertaining at some time during the late 1990s.

Economics of prostitution

199

prostitution. In this section of the chapter, we shall analyse some data that
are available for developed countries.
Table 7.1 shows estimates of the number of prostitutes in diﬀerent
Western European countries, alongside ﬁgures for total population. The
number of prostitutes is plotted against population size in Figure 7.1. Here,
as expected, we see that larger countries have higher levels of prostitution.
Similar data for states in the USA are presented in Table 7.2. Here, the
best available data are the number of arrests for prostitution and related
oﬀences in each state in a given year. This may be used as a rough proxy for
the number of prostitutes operating in that state, although it is acknowledged that this measure is also inﬂuenced by such factors as the attitude of
police towards prostitution. Figure 7.2 shows a plot of this variable against
state populations. Again, we see a clear positive relationship.
What is most interesting about the plots shown in Figures 7.1 and 7.2 is
that in both cases the number of prostitutes appears to rise more than in
proportion with population size. Results from applying double logarithmic

150000

Number of prostitutes

120000

90000

60000

30000

0
0

20 000 000

40 000 000
60 000 000
Population

80 000 000

Figure 7.1 Number of prostitutes against population in European
countries

200

Economics uncut

Table 7.2 Number of arrests in given year for prostitution and related
oﬀences in USA
State

Year

Population

Number of
arrests

Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Hawaii
Idaho
Iowa
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Nebraska
Nevada
New Jersey
New York
North Carolina
North Dakota
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
Tennessee
Texas
Utah
Virginia
West Virginia
Wisconsin
Wyoming

1999
2000
2000
2000
2000
1999
2000
2000
2000
2000
2000
1998
2000
2000
1999
2000
2000
2000
2000
2000
1998
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
2000
1998
1997
2000

4 370 000
626 900
5 130 600
2 673 400
33 871 600
4 056 000
3 405 600
783 600
1 211 500
1 294 000
2 926 300
4 369 000
1 274 900
5 296 500
6 175 000
9 938 400
4 919 500
1 711 300
1 998 300
8 414 400
18 175 000
8 049 300
642 200
3 450 700
3 421 400
12 281 100
1 048 300
4 012 000
5 690 100
20 851 800
2 233 200
7 078 500
1 811 000
5 170 000
493 800

187
117
2 542
326
12 401
1 366
694
142
469
6
266
637
23
989
1 818
1 474
2 027
469
4 177
1 956
10 774
1 090
0
284
824
2 338
382
977
1 305
6 321
443
726
142
1 907
4

Source: FBI Arrest statistics.

Economics of prostitution

201

12500

Number of prostitutes

10000

7 500

5 000

2 500

0
0

Figure 7.2

10 000000

20000000
Population

30000000

40000000

Number of prostitutes against population in US states

regressions of prostitution on population for each data set are as follows
(with standard errors in parentheses):
Europe:
log( prostitution)9.371.15 log( population)
(se)
(0.09)
USA:
log(prostitution)15.51.45 log(population)
(se)
(0.19)
The slope coeﬃcients in both of these regressions are signiﬁcantly greater
than one, conﬁrming that a rise in the population of a country or state is
expected to give rise to a more than proportionate increase in the number of
prostitutes in that country or state. This empirical observation is loosely
consistent with the reasoning of MacCoun and Reuter (2001) that since

202

Economics uncut

social controls are tighter in rural settings, and since a single supplier
requires a reasonable number of clients for adequate support, suppliers
tend to be drawn disproportionately to areas of higher population density.
The European data of Table 7.1 also seems to indicate that there is
roughly one prostitute for every thousand of population. Making the
courageous assumption that this proportion prevails throughout the world,
and using current estimates of world population, we may deduce that there
are approximately 6 million prostitutes in the world today.
Information on the demand-side is available in the form of the proportion
of males who are demanders of prostitution services. Sullivan and Simon
(1998) ﬁnd that 17.7 per cent of American males have paid for sex. Cameron
and Collins (2003) ﬁnd that 4.9 per cent of UK males have paid for sex, but
only 1.3 per cent have done so in the past ﬁve years. Carael et al. (1991) ﬁnd
that 3.3 per cent of Frenchmen have participated in the past ﬁve years. There
is some evidence that participation rates are higher in less developed countries: the Global Program on AIDS/World Health Organization estimated
the proportion of men using prostitutes in a single year to be around 10 per
cent in many African countries. These participation rates suggest that
demanders usually constitute a minority of the male population.
3

Economic analysis of prostitution

3.1 Overview of economic theories
As hinted in Section 1, the central question to be addressed is why earnings
per hour are so high in a profession with very basic skill and capital requirements, low ﬁxed costs, no formal training requirements and no barriers to
entry such as professional organizations. Moﬀatt and Peters (2004) recently
used UK data to estimate that the weekly earnings of a typical prostitute
are roughly two times that of non-manual workers, and three times that of
manual workers. There are many approaches to solving this wage
diﬀerential puzzle. One approach is to focus on the demand-side of the
market: Della Giusta et al. (2004) develop a theory in which the demandside is characterized by the assumption that males are willing to pay a
premium to avoid the ‘hidden’ costs of obtaining so-called ‘unpaid’ sex
such as wining and dining, gifts and the longer-term consequences of
attachment. Alternatively, one could focus on the supply-side, and attribute
the earnings premium to the compensation necessary for the risk of violence, risk of contracting STDs and for the generally unpleasant physical
experiences that prostitution entails.
A third approach, probably the most interesting to economists, is to
focus on the opportunity cost of prostitution in terms of foregone marriage
market opportunities. The underlying assumption here is that marriage and

Economics of prostitution

203

prostitution are incompatible for a female; one cannot be a wife and a prostitute simultaneously. While this assumption may not be universally true,
there is a good deal of evidence on the diﬃculties faced by women attempting to do both. The assumption is particularly tenable when considering the
sex industry in Victorian Britain (Fisher, 1997). After all, it is well known
that in the past, even loss of virginity was considered a serious penalty in
the marriage market (and still is in some societies).
One model that follows this approach is due to Edlund and Korn (2002).
We shall focus on this particular model in detail later in this section, for two
main reasons. First, this is the ﬁrst theory of prostitution to be published
by a major economics journal. Second, the model is useful because it is
capable of predicting certain known features of the sex industry which
other models are not.
3.2 Entry
Formal entry barriers such as required qualiﬁcations and professional
organizations do not appear to exist in the prostitution market, although
the International Union of Sex Workers (IUSW) is currently promoting
access to training as noted earlier. One entry barrier that may be encountered is the male managers of massage parlours and escort agencies ‘trying
out’ new employees. However, it usually transpires that this is a cheap ploy.8
Those entering the lower end of the market may be subjected to informal
barriers such as the threat of physical violence from incumbents or their
associates. Under such circumstances, we might expect to see entry
restricted to those who have been introduced by incumbents.
A point emphasized by Cameron (2002) is that the entry decision is completely detached from the supply decision. The supply decision may be
determined in a standard neoclassical labour supply framework, depending
in the usual way on non-wage earnings and the wage earned per unit supplied. The entry decision, in contrast, is heavily dependent on non-economic
considerations. Many individuals in society are restricted by ‘negative
capital’, for example, feelings of sin or social stigma, from entering this profession. Once the decision has been made to enter, however, preferences may
change dramatically, and the provider may ﬁnd that they experience the
usual beneﬁts from raising supply to an optimum. This two-stage decisionmaking process is referred to by Cameron as a ‘threshold crossing’ problem,
and if data on number of hours supplied were available, it could be modelled econometrically using the ‘double-hurdle’ approach (Cragg, 1971).
3.3 Market structure
The most commonly occurring market structure in prostitution is a collection of sole traders acting independently and competitively. In fact, in

204

Economics uncut

many UK cities, locations from which more than two providers operate are
unlawful and closed down if detected. Such incentives create the ideal conditions for a competitive market. There is, however, some anecdotal evidence of collusive behaviour with all providers operating in a particular
area committing to the same prices, with (possibly violent) punishment
strategies in place for deviators.
As soon as the enterprise takes on a third party gainer (‘pimp’) the enterprise changes status from sole trader to small ﬁrm. According to survey
evidence presented in the Home Oﬃce report by May et al. (1999), the
actual number of providers in the UK who are controlled by pimps is much
smaller than commonly believed; the great majority claim to operate independently. This suggests that, for most prostitutes, the services provided by
a pimp cost more than they are worth.
The cost structure varies according to the style of provision. At the most
basic level, providers who operate on the street conduct their business outdoors or in clients’ cars, and therefore do not require business premises. The
only obvious ﬁxed costs incurred are for seductive clothing, while the only
variable costs are for condoms and lubricants. Providers operating in
premises face the rental costs of those premises, and might well take advantage of the ﬁxed location by investing in capital equipment such as whips,
chains, handcuﬀs, strap-ons and uniforms of various types, in order to raise
the standard of service. Providers operating in massage parlours pay ﬁxed
costs to the proprietor, detailed by O’Connell Davidson (1998).
While, for the reasons suggested in the opening paragraph of this subsection, the market in a particular locality might be described as competitive, the best description of the structure of the market as a whole is one of
monopolistic competition: there is a reasonable amount of product
diﬀerentiation, and this is one of the reasons why prices vary so much
through the wider market. Another reason for the signiﬁcant price
diﬀerences is the bilateral price screening described in Section 1.
There are few large ﬁrms to talk of in the prostitution industry. However,
an interesting recent development, in May 2003, was when Australia’s
largest brothel, Daily Planet, became the world’s ﬁrst brothel to ﬂoat on the
stock exchange. In the longer term, Daily Planet aims to expand internationally. However, it remains to be seen how successful large organizations
like this one become. After all, such mega-commercialization surely has the
potential to remove much of the attraction of this particular product.
The potential for large organizations to exploit monopoly power
depends as always on the availability of close substitutes. One obvious substitute to commercial sex is non-commercial, non-marital sex (for example,
extra-marital aﬀairs, one-night stands, ‘swingers’ clubs9), which has
undoubtedly grown in importance over the last half-century. Less obvious

Economics of prostitution

205

substitutes are magazines, videos, TV and Internet porn and ‘telephone
sex’, which have also shown a recent surge. The continuing availability of
these may be expected to act as a check on monopoly power, whether or not
regulatory agencies ever choose to become involved.
3.4 Explaining the high compensating wage diﬀerential
Edlund and Korn’s (2002) model considers two separate but closely
related markets: the market for marriage, to which we have been referring
as the ‘primary’ market, and the market for sex, the ‘secondary’ market.
Use of this terminology was justiﬁed in Section 1. Market clearing conditions are derived for each market, in terms of the price of commercial
sex and the number of prostitutes operating. Then, in a manner similar
to IS–LM analysis in macroeconomics, a graph is used to identify the
price and ‘quantity’ that are consistent with equilibrium in both markets
simultaneously.
Here, we shall present a basic version of this model. Assume that society
contains N males and N females. Of the N females, n are prostitutes, and
N n are married. Note that all unmarried women are assumed to be prostitutes. This is because the only downside to marriage (in this very simple
model) is foregone prostitution opportunities. Since N n is the number of
married women, this is also the number of married men, implying that
there are n unmarried men, equal to the number of prostitutes. However, as
we shall see, the unmarried men do not match one-to-one with prostitutes
for the simple reason that married men also demand the services of prostitutes; each prostitute is therefore likely to be shared between married and
unmarried men.
Each female bears one child. If a female is married, both partners obtain
utility from the child; if she is unmarried, only she beneﬁts.
Everyone supplies one unit of labour. Married women are employed in
some activity other than prostitution, and earn w. Married women also
receive pm from their husbands, as payment for marriage. Prostitutes earn
p, the price of a unit of commercial sex. Note that one unit of commercial
sex does not amount to the servicing of one client. As mentioned previously, each prostitute services more than one client in each time period, so
a ‘unit’ is some number (not necessarily a whole number) greater than one,
of encounters with clients.
Assume that each unmarried man demands d ( p) units of commercial
sex. Note that quantity demanded depends on the price per unit, p, and by
the law of demand, we expect this dependence to be negative. Further,
assume that each married man demands d ( p) units, where 0 1. In
words, married men demand a ﬁxed fraction of the amount of commercial sex enjoyed by unmarried men. It is a key requirement of the model

206

Economics uncut
pays
nλp

pm +
n + (N n) λ
has child

receives pm + w
has child
married

married

FEMALE

MALE

single

prostitute

pays
np

n + (N n) λ
no child

receives p
has child

Figure 7.3 Male and female decisions, money and parenting outcomes, and
the matching process
that
0, that is, married men demand at least some commercial sex; if
they did not, the model would break down. is therefore an important
parameter and we shall refer to it as the ‘inﬁdelity parameter’.10
Each person is free to choose whether to marry. The consequences of
each choice for each gender are shown in Figure 7.3, which summarizes all
of the model’s assumptions. Double-ended arrows represent matching.
With reference to Figure 7.3, we next derive equilibrium conditions.
First, consider the female decision, and note that the return from marriage
is pm w while the return from prostitution is p. A child is enjoyed in both.
For equilibrium, return from the two activities must equalize, so:
p* p*m w

(7.1)

where stars denote equilibrium values. To understand (7.1), imagine a situation in which p pm w, so the return from marriage is greater than that
from prostitution. In such a situation, there would be a transfer of women
from prostitution to marriage, which would cause a rise in the price of commercial sex, p, and this process would continue until the equilibrium condition (7.1) is met. (7.1) may be interpreted as the condition for equilibrium
in the marriage market.
Equation (7.1), although very simple, is very useful in that it provides the
answer to the wage diﬀerential puzzle identiﬁed earlier. It shows that wages
earned by prostitutes are always higher than those earned by women in

Economics of prostitution

207

alternative employment, and that the diﬀerential is given by pm, the ‘value’
of marriage. Thus, equation (7.1) makes it very clear that the solution to
the puzzle is in terms of the opportunity cost of prostitution, that is, the
prostitute’s foregone beneﬁts from marriage. As discussed in Chapter 5,
there is no doubt that the beneﬁts of marriage are substantial. Indeed,
Clark and Oswald (2002) use data on 10 000 people, together with an economic model of ‘happiness’, to estimate the value of marriage at £70 000
per annum.
It was mentioned earlier that the ‘inﬁdelity parameter’ must be positive
for the model to be sensible. Using Figure 7.3 and equation (7.1), we can
see why this is. Imagine 0, so married men demand no commercial sex.
The cost to a man of being married would then be only pm, which, by equation (7.1), is always less than the cost of being single, p. Even ignoring the
additional advantage of fatherhood, every man would prefer marriage.
Therefore, if 0, everyone would be married, and there would be no prostitution at all.
Next, we consider the market for commercial sex. Recall that each of the
n unmarried men demands d ( p) units of commercial sex, while each of the
N n married men demands d( p) units, where , introduced earlier, is
the inﬁdelity parameter. Market demand for commercial sex is then:
n d(p)(Nn) d(p)[n (N n)] d(p)

(7.2)

Market supply of commercial sex is simply:
n

(7.3)

This follows from the assumption that each of the n prostitutes supplies
exactly one ‘unit’ of commercial sex.
We combine equations (7.2) and (7.3) in order to derive the equilibrium
condition for the sex market:
n[n (Nn)]d(p)

(7.4)

Equation (7.4) implicitly deﬁnes the market clearing price of commercial
sex, p*, as a function of the number of prostitutes, n. To make this function
explicit, we need to specify a functional form for d(p). For good measure,
we shall do this. We assume:
d (p)2p

(7.5)

208

Economics uncut

which clearly satisﬁes the law of demand. Substituting (7.5) into (7.4) and
rearranging gives:
p* 2

n
n (N n)

(7.6)

Equation (7.6) is the equilibrium condition in the commercial sex market.
Notice that the relationship between n and p* is negative. To verify this, set
n0 in equation (7.6): if there are no prostitutes, the market clearing price
is 2; then set nN: if all women are prostitutes, the market clearing price
is only 1.
The marriage market equilibrium condition (equation (7.1)) and the sex
market equilibrium condition (equation (7.6)) are combined in Figure 7.4.
The point of intersection of the two lines identiﬁes the combination of
price (p*) and number of prostitutes (n*), which clears both markets simultaneously. Based on the observation made from real data in Section 2, we
may predict that the actual value of n* in the context of Europe is approximately N/500.
The key predictions from this very basic model can be seen directly from
Figure 7.4. An increase in the wage earned in alternative employment w, or
alternatively an increase in the price paid to a woman for marriage pm, both
raise the position of the horizontal line in the diagram, causing both an
increase in the equilibrium price of prostitution and a decrease in the
number of prostitutes in the market.
These predictions might be seen as fairly obvious, but the purpose of this
sub-section has been nothing more than to introduce the reader to the most
basic version of the model. More useful predictions emerge when the model
is generalized. To understand the technical aspects of the more general
p

marriage market
equilibrium

p* = pm + w

sex market
equilibrium

Figure 7.4

Marriage and sex market equilibrium

Economics of prostitution

209

models, the interested reader is referred to the paper by Edlund and Korn
(2002). In the sub-section that follows, we brieﬂy outline some predictions
that emerge from the more general models, avoiding technical detail.
3.5 Extensions
First, in addition to predicting that prostitution decreases in the alternative
female wage w, the model can be generalized in such a way that prostitution
also decreases in male income. One of the additional assumptions required
for this result is that child quality depends on marital resources allocated to
the child. The prediction that prostitution decreases in both male and
female earnings is useful because it provides an explanation for the higher
rates of prostitution in poorer countries, and for the long-term decline in
prostitution seen in developed countries. Empirical evidence for these two
features of the industry is documented by Edlund and Korn (2002).
Second, the assumptions that all men face identical wages, and all women
in ‘alternative’ employment face identical wages, are relaxed. The prediction then is that higher earners of both genders are selected into marriage,
while women with poor labour market possibilities are drawn towards prostitution. This too is an accurate representation of reality.
Third, the possibility of a tax on prostitution is introduced. This has the
eﬀect of reducing the number of prostitutes, since, in equilibrium, it is the
prostitute’s price net of tax that is now equated to the wife’s earnings.
Fourth, the assumption of a closed population with equal numbers of
males and females is relaxed. It is important to do this to allow for the
growing phenomena of sex tourism and sexual migration. Some holiday
destinations, particularly in the Far East, are popular to Western men
seeking commercial sex. It is reasonable to assume that the number of
males in the relevant market is greater than the number of available females.
Moreover, it is reasonable to assume that the male tourists are high demanders of commercial sex (perhaps, a third category of males should be introduced to the model, with a value greater than 1). So, there are two reasons
why the proﬁtability of prostitution rises relative to marriage, and why we
expect the number of prostitutes to be higher under such circumstances.
Sexual migration is the phenomenon of large numbers of young females
migrating from developing countries to richer ones to work in the sex industry. As cited in Section 1, around one half of Europe’s prostitutes are immigrants. This must have the eﬀect of raising the number of females in the
commercial sex market, and one might expect the equilibrium price to fall
as a result, obviously causing great irritation to home-based providers.
French prostitutes, in particular, have been heard complaining bitterly in
recent years about the impact on ‘their’ markets of the heavy inﬂux of
Eastern European providers into French cities (see Anonymous, 2004).

210

Economics uncut

3.6 The eﬀects of asymmetric information
The Edlund and Korn model explains why the market price is so high. The
presence of asymmetric information in the market means that the actual
price paid may be even higher. This issue is highly relevant in markets with
‘transient’ consumers, for example, holidaymakers purchasing sex, since
such clients may have no local knowledge of the market price. It is often the
case that this market price is considerably lower than that prevailing in the
client’s home country, and this may clearly be exploited by local providers.
Providers can maximize the consumer surplus they extract from locally
inexperienced clients by engaging in ﬁrst-degree price discrimination.
They simply name an excessive price and then lower it until the client is in
agreement.
Another sort of asymmetric information arises because advertisements
appearing in newspapers, magazines, phone boxes, or on the Internet, are
generally posted by the providers themselves. There is therefore a tendency
for these advertisements to give a misleading description of the provider’s
characteristics and of the quality of the services they oﬀer. The industry is,
therefore, prone to the problems caused by informational asymmetries, as
ﬁrst described by Akerlof (1970). One of the negative outcomes is a version
of so-called Gresham’s Law: high-quality providers tend to be driven out
of the market by those of lower quality.
We might at this point acknowledge the lasting eﬀect that the emergence
of the Internet may have on the prostitution industry. Obviously, providers
advertise on the Internet, but equally importantly, a number of websites
have appeared with the purpose of information exchanges for clients,
including the Punternet site, data from which is analysed econometrically
in Section 4. Since information supplied on these sites is provided by clients,
and the web-managers appear vigilant in detecting and deleting false
reports, there is no reason to expect it to be biased in any way. As more and
more information on providers becomes available through this new
medium, and as more and more clients adopt it as a routine source of information, the informational asymmetries are therefore expected to diminish
substantially. In consequence, high-quality providers are likely to see their
businesses ﬂourish, while low-quality providers are likely to be driven out
of the market early.
3.7 Economic welfare considerations
Cameron (2002), with the support of a number of entertaining anecdotes,
presents a number of reasons why the prostitution market, in its capacity
as a ‘secondary’ market, has the potential to raise levels of economic
welfare. These welfare issues form the basis of many arguments in favour
of legalization (see Section 5). Some of these ideas are summarized here.

Economics of prostitution

211

One rather extreme perspective on the sexual climate in human societies
is that male lust permeates through the community in a continuous raging
torrent, with which the primary (marriage) market cannot cope. There is
acute excess demand by males at all times. The secondary (prostitution)
market plays the vital role of an overﬂow mechanism, without which
equilibrium in the primary market could never be attained. The secondary market can also be seen as an externality abatement technology,
since the negative externalities generated by excess male demand such as
sexual harassment, cajolement and assault are likely to be reduced by its
existence.
When we consider an individual with very speciﬁc, non-conformist,
sexual needs and desires, it may well be the case that these demands are not
able to be satisﬁed in the staidness of the primary market. In the secondary
market, every service is available at a price, and can be obtained through
specialist advertising channels. Whenever the deviant demander is matched
with a cooperative supplier from the secondary market, a Pareto improvement may be said to have occurred. Indeed, we ﬁnd evidence in Section 4
that the provision of ‘kinky extras’ not only raises the price of the service,
but also signiﬁcantly raises the level of satisfaction derived by the client.
Another aspect of equilibrium in the primary market relates to expectations regarding partners’ sexual pasts. In some settings, there may be an
expectation for the bride to be a virgin, but at the same time for the bridegroom to be highly sexually experienced. For these incongruent expectations to be realized in equilibrium, a secondary prostitution market is
essential, being the only means for males to acquire the necessary experience prior to marriage without sacriﬁcing the virginity of the pool of
potential brides.
More convincingly, there may simply be an expectation on young males
to have at least some experience before entering the primary market. The
secondary market would again be essential in providing males with the
minimal experience necessary to shed their embarrassment and enter
the primary market with conﬁdence. Perhaps the most famous ﬁctional
example is Holden Caulﬁeld in JD Salinger’s The Catcher in the Rye who,
as a 16-year-old contemplating the loss of his virginity to a call girl, reﬂects:
‘I ﬁgured if she was a prostitute and all, I could get in some practice on her,
in case I ever get married or anything. I worry about that stuﬀ sometimes.’
4 Empirical studies of prostitution
In this section, we move from the theoretical to the real, and consider
empirical studies of the prostitution industry, which use data extracted
from real markets. We also report in detail on the analysis of one particular dataset, analysed previously by Moﬀatt and Peters (2004).

212

Economics uncut

4.1 Previous empirical work
Relatively little work has been done on the empirical analysis of the prostitution industry. One possible reason for this lies in the obvious diﬃculties
associated with data collection. Prostitutes are unlikely to reveal information about their work since they cannot be sure that the researcher is not
connected to the police or the tax department. Clients are unlikely directly
to supply information on encounters with prostitutes since they may, for
reasons of stigma and feelings of sexual inadequacy, prefer to keep the
activity a secret unto themselves. Although data on clients have been
obtained in the past, for example from self-reported surveys (Faugier and
Sargeant, 1997) or from police records (Benson and Matthews, 1995), the
samples are small, and the information obtained is mainly concerned with
civic and health status of the individuals, rather than the economic features
of the market that are of interest here.
In spite of these problems, some empirical work has been done in the
past. Cameron et al. (1999) obtain data from escort advertisements
extracted from gay publications in London. They then perform econometric estimation, which identiﬁes the characteristics of the escort that make
them more likely to oﬀer particular services. They ﬁnd, for example, that
older prostitutes are more likely to supply deviant sexual services such as
bondage and ﬂagellation, than their younger colleagues.
Cameron and Collins (2003) focus on the demand-side, using a large UK
lifestyle dataset to identify the characteristics of males that determine
demand for prostitution services. The dependent variable in their model is
binary: one if the respondent has purchased sex; zero otherwise. As noted
in Section 2, the proportion in their sample who have done so is 4.9 per cent.
They ﬁnd that the most important determinants of participation are religion and risk disposition. Interestingly, age does not appear to have a
signiﬁcant eﬀect on the probability of recent participation. Sullivan and
Simon (1998) obtain similar results using a sample of American males,
although the proportion in this sample who have purchased sex is noticeably higher, at 17.7 per cent.
Rao et al. (2003) investigate the determinants of price paid for sex in
India. In particular, they focus on the impact of condom use, which they
ﬁnd to have a strong negative eﬀect on price, inevitably amounting to a disincentive from the provider’s point of view. This ﬁnding has far-reaching
societal implications for health and education policy, not least because
unprotected sex with prostitutes is one of the primary causes of the spread
of HIV/AIDS in developing countries.
Moﬀatt and Peters (2004) carried out an empirical study of the prostitution market in the UK, using data extracted from the Punternet
website (http://www.punternet.com).11 The main purpose of this site is as

Economics of prostitution

213

an information exchange for clients. The site restricts itself to heterosexual
encounters between male clients and female providers. Clients are invited
to submit ‘ﬁeld reports’ to the site on prostitutes whom they have recently
encountered. The report is submitted pro-forma, and contains the location
and duration of the encounter, the working name and contact details of the
provider, some information about her physical attributes and personality, a
description of the services rendered, and the price paid. The site is heavily
used, with over 55 000 reports at August 2005.
The Punternet data were used by Moﬀatt and Peters (2004) to investigate
the way in which the prices of prostitutes’ services are determined.
Therefore, they focus on the demand-side of the market, estimating the
value perceived by clients of each attribute of the service, and assuming
supply to be ﬁxed. Models of this type are known as hedonic pricing
models. Their most usual application is to data on property prices (for
example, Hughes, 1996), but as seen in Chapter 10, they can also be applied
to assisted reproduction markets.
An additional piece of information supplied in each report is whether or
not the client would return to the same provider in the future. This yields a
binary variable, which may be perceived as an indicator of client satisfaction. Binary data analysis of this variable may be used to identify features
of the service provider and of the encounter that increase client satisfaction. It is then interesting to check whether the factors that increase client
satisfaction are the same factors that attract a price premium according to
the hedonic pricing analysis. After all, in a smoothly functioning market, it
should be the case that any product attribute that increases customer satisfaction also increases the price.
The Punternet dataset shall be analysed here, although not at the level of
sophistication of Moﬀatt and Peters (2004). First, the determinants of
price, then the determinants of client satisfaction are identiﬁed, and a comparison is made between the two sets of results. Before this, some
exploratory data analysis is reported.
4.2 Exploratory analysis of the Punternet dataset
The selected sample consists of 982 complete reports that were submitted
to the site between January 1999 and July 2000. Since the focus of attention
is the per-encounter price paid in pounds sterling, a histogram of this variable is shown in Figure 7.5. The price distribution has a clear mode at
around £60, and shows a strong positive skew.
Figure 7.6 shows a scatter plot of price against duration of encounter in
minutes, with a non-parametric regression12 (smooth) super-imposed.
From the smooth, a clear positive relationship is identiﬁable and, interestingly, a small degree of convexity is detected: it appears that the marginal

214

Economics uncut

200

Frequency

150

100

0
50

Figure 7.5

100

150
200
Price in pounds

250

300

A histogram of price of prostitute’s services (£)

eﬀect of duration on price becomes higher at higher values of duration.
This suggests the possibility of market segmentation, with longer durations
typically being associated with a higher class of provider, namely the escort.
Figure 7.7 shows a scatter plot and smooth of price against age of
provider in years. Non-linearity is again evident. The relationship is fairly
level at low ages, but beyond the age of around 35, price appears to decline
monotonically. This pattern calls for the use of age and age-squared, at
least, as explanatory variables in the pricing model.
4.2.1 Determinants of price An ordinary least squares regression was
performed using the sample of 982 observations described earlier, with

Economics of prostitution

215

600

Price in pounds

500

400

300

200

100

0
0

60
80
Duration in minutes

100

120

140

Figure 7.6 Scatter of price of prostitute’s services (£) against duration in
minutes, with smooth

price of encounter in pounds sterling as the dependent variable and a set of
explanatory variables including duration of encounter, provider characteristics, arrangement dummies, type of service and regional dummies. The set
of explanatory variables included in the model was decided using a generalto-speciﬁc model selection procedure. Results are reported in Table 7.3.
The R2 of 0.66 reported at the end of Table 7.3 indicates that two-thirds
of the variation in price is explained by this regression, which is actually
quite impressive for cross-section data.
The ﬁrst slope coeﬃcient reported is for duration of the encounter in
minutes. The coeﬃcient of 1.28 indicates that an additional minute added
to the duration of the encounter adds £1.28 to the price, ceteris paribus.
Note also from the low standard error that this eﬀect is very precisely estimated.
The next set of estimates relates to a selection of provider characteristics.
To deal with the inverted U-shaped relationship between price and age of
provider, we have included age-squared in addition to age. The signs of the

216

Economics uncut
300

Price in pounds

200

100

0
15

Age of provider

Figure 7.7 Scatter of price of prostitute’s services (£) against age of
provider in years, with smooth
two coeﬃcients conﬁrm that the relationship is inverted U-shaped. The
turning point can be computed from the two coeﬃcients as:
1.81
23.8
2 0.038
This implies that the highest prices are commanded by providers aged just
under 24 years and tells us something about the age–earnings proﬁle of
prostitutes. However, it must be remembered that the payment for a single
encounter is being modelled here, not total annual earnings. In order to
discover, for example, the age at which total annual earnings are maximized, information would be required on the ﬂow of clients for diﬀerent
ages of provider. This information is not readily available, although see
Matthews (1997).
Next, a dummy variable is present indicating whether the provider is
‘Very thin’, according to the description supplied by the client. Note that
this variable has a signiﬁcantly positive coeﬃcient of 6.92, which

Economics of prostitution

217

Table 7.3 OLS regression to identify the determinants of price of
prostitute’s services
Variable
Constant
Duration in minutes
Provider characteristics:
Age
Age-squared
Very thin
Attractive
Arrangement (base: walk-up):
Girl’s home
Parlour
Client’s home
Hotel
Service:
Full sex
Kinky extras
Region (base: rest of UK):
Soho
Central London
Greater London
Midlands
Greater Manchester
North East
Merseyside
Scotland
Sample size
R2

Coeﬀ. (standard error)
17.03 (15.32)
1.28 (0.045)**
1.81 (0.93)*
0.038 (0.015)**
6.92 (2.12)**
7.60 (5.39)
17.75 (3.08)**
11.74 (3.02)**
49.13 (6.79)**
59.65 (5.91)**
2.42 (3.54)
4.20 (1.90)*
9.80 (3.57)**
29.24 (3.15)**
7.19 (2.66)**
3.10 (3.82)
7.35 (3.60)*
8.89 (4.83)*
27.70 (15.55)*
11.00 (4.58)**
982
0.66

Notes:
Dependent variable: price of encounter in pounds.
Standard errors are given in parentheses.
* Indicates that a variable is signiﬁcant at the 5% level, ** indicates signiﬁcance at the 1%
level.

represents the premium attracted by providers having this characteristic.
The ﬁnal provider characteristic is ‘Attractive’, another dummy variable
based on subjective information supplied by the client. Note that the
coeﬃcient is negative (although insigniﬁcant), and this is counter-intuitive;
it implies that attractive providers command lower prices. An explanation
for this apparently strange result shall be oﬀered later in the section.

218

Economics uncut

The next set of variables are dummy variables for arrangement. The base
case (that is, the excluded dummy) here is what is known as a ‘walk-up’.
A walk-up is a ﬂat, usually in an apartment block, with a sign outside and
a doorbell answered by a ‘maid’, from which sex can be purchased on a
drop-in basis. Because the coeﬃcients of the included arrangement
dummies are all positive and signiﬁcant, it appears that walk-up is by far
the cheapest arrangement, ceteris paribus. This is explained by the high
eﬃciency of such establishments and the associated low quality of the
service. The most expensive arrangements are client’s home and hotel, the
latter commanding almost £60 more than a similar encounter in a walk-up.
This is not surprising either: hotel visits are usually associated with ‘escort’
providers, widely acknowledged to be of the highest quality.
Next in the list are a pair of variables representing the type of service delivered. Many such variables were available in the dataset; only one appeared
to be signiﬁcant. The signiﬁcant variable is a dummy indicating whether
‘Kinky extras’ are supplied. Such extras appear to add a few pounds to the
price of the encounter. Another dummy is included if ‘Full sex’ was supplied, but this does not appear signiﬁcant. The message here is that although
price depends strongly on the duration of the encounter, the actual service
supplied has very little eﬀect. It seems that providers are broadly indiﬀerent
to what services they supply as long as they are paid for the time spent.
Finally, there is a set of regional dummy variables, the base case being
‘Rest of UK’. Here, a clear pattern emerges: prices are by far the highest in
Central London (excluding Soho, where there is a thriving walk-up
market), and lowest in the North of England and Scotland. It therefore
appears that distance from the capital is a key determinant of price. This is
interesting because it ties in with one of Edlund and Korn’s (2002) key theoretical predictions (see Section 3), namely that the price paid for prostitution services increases with the wage earned in alternative employment; it
is well known that average wages are considerably higher in London than
in the North of England and Scotland. In a sense this regional pattern is
not at all surprising, since a similar story could be advanced in the context
of any profession. However, since even the most obvious predictions are
sometimes not supported by data, it is always reassuring when they are.
4.2.2 Determinants of satisfaction As already noted, one of the pieces of
information supplied in each report is whether or not the client would wish
to return to the same provider in future. This is a useful piece of information because it provides an indication of whether or not the client is satisﬁed
with the service received. While most clients appear to be satisﬁed, by no
means all are: of the 982 reports used in this study, 202 clients reported that
they would not return.

Economics of prostitution

219

It is interesting to investigate the determinants of this satisfaction variable. Since it is a binary variable, ordinary least squares regression cannot be
used. Logistic regression is used instead. Once again, a general-to-speciﬁc
model selection procedure is used to decide on a set of variables. The results
are presented in Table 7.4.
The coeﬃcient of duration appears to have the expected sign: the longer
the encounter lasts, the more likely the client is to be satisﬁed. The
coeﬃcient of price also has the expected sign: the more a client pays, ceteris
paribus, the less likely he is to return to the same provider.
As in the price equation, there is a set of variables here that represent
provider characteristics. First, an inverted U-shaped age eﬀect is again
apparent. The turning point can again be computed:
0.126
31.5
2 0.002
This implies that providers aged 31.5 supply the most satisfaction. This
appears to contrast with the results from the price equation analysed
earlier. There, we found that providers aged 23.8 charge the highest price.
This contrast may partly be explained by the fact that older providers have
Table 7.4 Results from binary logit analysis of client satisfaction
variable
Variable

coeﬀ. (asy. se)

Constant

4.25 (1.43)

Duration
Price

0.074 (0.009)**
0.017 (0.003)**

Characteristics
Age
Age-squared
Attractive
Afro-Caribbean

0.126 (0.086)
0.002 (0.001)*
1.77 (0.51)**
1.82 (0.63)**

Service
Full sex
Kinky extras
Sample size

0.75 (0.32)*
0.84 (0.28)**
982

Notes:
Dependent variable: would you return? yes 1; no 0.
Asymptotic standard errors are given in parentheses.
* indicates that a variable is signiﬁcant at the 5% level, ** indicates signiﬁcance at the 1%
level.

220

Economics uncut

had more opportunity to establish a regular clientele than younger
providers, and regular clients are, by deﬁnition, more likely to indicate
readiness to return. Another explanation may be in terms of the greater
value older providers place on repeat business. Albert (2001) provides a case
study of a Nevada brothel, including (pp. 156–7) an interview of two older
providers who criticize the casual attitude of their younger colleagues
towards clients, adding that, ‘us working girls from the old school understand the value of repeat business’. The message here is that older providers
are nurturing their clients with a view to regular custom, while younger
colleagues are more conﬁdent of an unending stream of ﬁrst-time clients
and therefore pay less attention to quality of service.
The dummy variable ‘Attractive’ is included in the logistic regression. The
strongly signiﬁcant positive coeﬃcient is as expected, indicating that attractive providers are more likely to attract a return visit. However, again we see
an inconsistency with the results of the price equation; there, we found that
attractive providers charge a lower price. This inconsistency is explained in
terms of strategic behaviour by the provider. Unattractive providers charge
a high price precisely because they do not expect a return visit, and they are
extracting maximum consumer surplus from the client on the only opportunity. An obvious question that arises here is why a rational customer is
willing to pay this higher price for a less desirable service. One answer might
lie in the ‘sunk costs’ of search and travel: having already incurred these
costs, and in doing so built up an expectation of sexual satisfaction, it is
rational to pay once but never return.
The dummy variable ‘Afro-Caribbean’ also has a strongly signiﬁcant positive coeﬃcient. Apparently, providers of this ethnicity generate the most
satisfaction. Ethnicity had no eﬀect in the price equation, and was therefore not included there. So, once again we are seeing an inconsistency
between the two equations. If the market were perfectly functioning, we
might expect the price charged by Afro-Caribbean providers to rise, to
reﬂect the higher levels of client satisfaction that they generate.
The last two variables in the list are service indicators. If a client has had
full sexual intercourse, he is more likely to return to the same provider. This
is quite obvious, but the signiﬁcant eﬀect here may partly be due to endogeneity of this explanatory variable: more enthusiastic clients are clearly
more likely to opt for full sex during the current encounter.
Finally, the variable ‘Kinky extras’ has a strong positive eﬀect on the
likelihood of a return visit. This may be explained in terms of such services
providing satisfaction. However, there is another explanation: the specialist services indicated by this variable may only be available from a small
number of providers, so clients requiring such services are inevitably drawn
to return to the same specialist provider.

Economics of prostitution

221

5 Policy
Policy debate concerning prostitution usually revolves around the issue of
legalization. This issue is currently of particular topical interest in the UK,
with a four-month nationwide consultation on approaches to legalization
launched by the Home Oﬃce in July 2004 (Home Oﬃce, 2004). This is the
ﬁrst thorough review of prostitution law in the UK for 50 years. Here, we
state the standard cases for and against legalization, and then outline and
evaluate various alternative models.
The principal argument against legalization, stated forcefully by Hughes
and Roche (1999), is that it is giving societal approval to an activity that is
sordid, unhealthy, dangerous, debasing to sexual relationships, associated
closely with STDs, substance abuse, violence and many other socially
undesirable phenomena. A related argument is that legalization is likely to
encourage more people to become prostitutes. Most people agree that the
activity is degrading to the individual as well as dangerous and unpleasant,
and policies should be designed with the objective of minimizing the
number of people who are subject to such degradation. However, it is not
clear that legal restrictions meet this objective. MacCoun and Reuter (2001)
have pointed out that no academic studies have ever established a link
between legal availability and the extent of prostitution.
A separate set of arguments is concerned with the direct impact of the
activity on the surrounding community. Communities most likely to reject
prostitution are small towns or middle-class suburbs, predominantly populated by infrequently-moving homeowners and families. Their opposition
to prostitution is more likely to arise from an objection to the ‘visual pollution’ (for example, kerb-crawlers, used condoms on the pavement, sexually explicit business cards in phone boxes) and other distasteful spillovers
from the activity, than from any inherent moral objections. In the UK, the
famous Wolfenden Committee Report (Wolfenden, 1957) articulated this
pragmatic standpoint, asserting that the law should only be concerned
with, ‘the manner in which the activities of prostitutes and those associated
with them oﬀend against public order and decency, expose the ordinary
citizen to what is oﬀensive and injurious, or involve the exploitation of
others’.
Most policies outlawing prostitution are directed towards providers
themselves, with ﬁnes or prison sentences imposed on violators. An interesting alternative model is administered in Sweden, where, in January 1999,
it became illegal to purchase sexual services, with violators facing up to six
months in prison (Grundberg, 2003). The eﬀect on the market was dramatic, with street prostitution falling by more than two-thirds. Moreover,
it has become evident that the activity has been displaced, with the Internet
now providing the principal medium for attracting clients. The Swedish

222

Economics uncut

policy approach, of curbing demand rather than supply, has therefore
proved successful in reducing the undesirable spillovers listed in the preceding paragraph.
Some arguments in favour of legalization were presented earlier in this
chapter, although those arguments all had a strictly theoretical tone. From
a more pragmatic standpoint, the principal argument in favour of legalization is that the social problems associated with prostitution are likely to be
greatly alleviated if the activity is overseen by the authorities. For example,
free medical checks would limit the spread of STDs; keeping central registers of providers would alleviate the recent serious problems of human
traﬃcking; support and advice available through drop-in centres may act
as a check on the type of violence from pimps and clients to which many
providers are otherwise subjected.
Running in parallel with arguments promoting drug legalization
(see Chapter 3), some predict that legalization of prostitution would reduce
the revenue, and therefore the attractiveness, of organized crime. As a
further beneﬁt to society, a portion of the revenue currently raised by criminals would instead become tax revenue for government. The number of
undesirable characters on the scene, including ‘pimps’, would undoubtedly
decrease, as would associated levels of street violence. Also, the time and
money spent by the police ﬁghting the losing battle of enforcing prostitution law could be diverted to more productive activities.
Another moral argument in favour of legalization is that it removes a
restriction on freedom. At a certain level it seems strange that a society that,
for example, allows a woman to end another’s life through abortion, denies
a woman the choice to sell her body in order to make a living.
Having made a decision on whether to support legalization in principle,
we are then faced with the choice between a number of models of legalization, to which we now turn. These models have been described in detail by
Reynolds (1986).
First, the ‘laissez-faire’ or ‘tolerance’ model is one in which government
and police intervention is kept to a minimum, despite legislation prohibiting prostitution. Such inaction by the authorities may be the result of pressure from the courts, pressure from the public or some form of bribery. The
model often occurs in large cities in which police are over-burdened with
crimes of a more serious nature, or in locations whose economy relies
heavily on adult tourism. A good example of the latter is Thailand, where
the sex industry as a whole is thought to account for 14 per cent of GDP.13
In such settings, police action would be called upon by the public only when
limits of acceptable behaviour have been breached.
Second, the ‘regulation’model is characterized by a licensing system. Such
a scheme has recently been applied by the US government to prostitutes

Economics of prostitution

223

operating near the USA–Mexico border (Marizco, 2003). The licences are
not intended to promote prostitution, but rather provide a means of exercising control over what is considered a socially undesirable activity.
Licensing is useful for maintaining health standards, for detecting illegal
traﬃcking of women and generally for keeping track of providers. Also,
crimes against clients are likely to be reduced since oﬀending providers
become easily identiﬁable. However, an obvious downside is that society may
well frown upon the idea of local government monitoring the provision of
commercial sex.
Finally, the ‘zoning’model describes a situation in which a minimum physical distance is set between residential areas and the locations where prostitution is permitted. Such regulations normally result in a concentration of
activity in one location, known as a ‘tolerance zone’ or ‘working area’, and
often located in an industrial area. Many European cities apply such a
model, most notably in the Netherlands, and some UK cities look likely to
adopt similar schemes soon (Humphries, 2003). Although such plans tend
to provoke emotive negative responses from the Press, the schemes are likely
to beneﬁt all parties. The police are fully aware of the sort of activities taking
place in the zone and ﬁnd it easy to control the conduct of both providers
and clients. Concentration of businesses can also beneﬁt the clients and
providers themselves. For clients, choice is maximized and search costs minimized. For providers, the complete market can be reached without the need
for advertising, and publicly provided facilities such as night-shelters and
CCTV surveillance have the potential to make life easier. Last but not least,
the general public beneﬁts from the spectacle of prostitution being eliminated from their doorsteps. (See Krugman, 1991 for a full economic theoretic explanation of the beneﬁts of so-called ‘market clustering’ in a more
general context.)
A question that arises is, given the apparent Pareto-optimality of the
zoning model, why such schemes do not emerge spontaneously. The answer
is that it is only the preferences of market participants that are reflected in
the prevailing scheme. When left to the free market, no account is taken of
the preferences of local residents and other members of the public whose
lives are aﬀected. In economic terms, the spillover eﬀects of prostitution are
negative externalities. The government’s role is to actively enforce a zoning
model in order to eliminate these externalities.
One thing that is clear from the descriptions of these various models of
legalization is that it is appropriate to apply such schemes at a geographically local level. Therefore, it may be seen as more sensible for prostitution
policy to be delegated to local government, where local knowledge may be
exploited in designing an optimal scheme, than for national governments
blindly to enforce blanket schemes.

224

Economics uncut

It is well-known that sex workers themselves are generally in favour of
legalization, or at least decriminalization (see, for example, the IUSW
website). Some third party gainers are also in favour. Thai brothel owners,
for instance, claim that it would relieve them of paying bribes to the
police.14 However, it is also interesting that wherever legalization has been
adopted, for example in Nevada in the USA, sex workers claim to feel vulnerable to state control, being requested to work long hours for low pay.
Many also complain of the soulnessness of legal brothels as workplaces. It
is, therefore, not surprising that many opt to work outside of the legal
system. As seen in Table 7.2, there were 4177 arrests in Nevada in the year
2000 for illegal prostitution.
When legalization is introduced, we also need to consider the destiny of
those workers who are excluded from the legal market due, for example, to
failing the obligatory health check, or to being an illegal immigrant. These
rejects would naturally turn to the illegal sector, which would now consist
of a high-risk ‘underclass’, but would nevertheless continue to attract
clients since the lower costs associated with the illegal market, for example
from avoiding tax, would allow undercutting. It must also be acknowledged
that the most socially repugnant sectors of the market such as child prostitution will never emerge from underground, since no one is likely to
propose legalization of these sectors.
Thus, we are led to the conclusion that any decision by a government to
legalize a market would not overnight solve all of the social problems associated with prostitution; it would be impossible to have a legal market
without having an illegal one operating alongside it.
6 Conclusion
The economic model introduced in this chapter makes an important distinction between two markets: marriage (primary) and commercial sex
(secondary). One aspect of the importance of this distinction is that the
eﬀects of prostitution on economic welfare are seen clearly in this context:
the presence of the secondary market is essential for the socially cherished
marriage market to function in equilibrium.
The model is also capable of explaining a number of known features of
the market for prostitution, for example, that the number of prostitutes
falls as earnings from alternative employment rise. The key quantity
appearing in the model is the opportunity cost of prostitution, being simply
the ﬁnancial beneﬁts of marriage. The high price paid to prostitutes as compared with other workers with similar skill levels is explained in terms of
the very high level of this opportunity cost.
A criticism of the economic model might be that it relies exclusively on
economic factors. A complete model of the decision to enter the market for

Economics of prostitution

225

sexual services, either as a client or a provider, would surely incorporate the
impact on self-esteem, the social stigma, the risks and so on. Also, the
assumption that there are no possibilities other than marriage and prostitution was never fully tenable, least of all today, given the signiﬁcant rise in
non-marital, non-commercial sex observed in the second half of the twentieth century. Nevertheless, as a highly simpliﬁed version of reality, the
model serves a useful purpose. Another reason for outlining the model in
this chapter is, as previously mentioned, that this is the ﬁrst model of
prostitution to be published by a leading economics journal.
The demand-side has been analysed in some detail in Section 4. This
sort of analysis has been made possible by the emergence of the Internet,
ﬁrst in inducing market participants to provide information, albeit
through pseudonyms, of a sensitive and personal nature, and second in
providing an easily accessible source of large amounts of such information. The analysis has been useful in the uncovering of certain interesting
features of the market. Also, since it is a study based on actual prices paid,
the results could be used to estimate the total ﬂow of income generated
by the industry (see Moﬀatt and Peters, 2004), which would, in turn, be
useful in estimating the tax revenue that would be generated under legalization. However, it should also be remembered that the introduction of a
tax would have the eﬀect of reducing the number of providers in the
industry.
The key ﬁndings in the empirical analysis are that price is driven mainly
by factors such as duration of the encounter, type of arrangement and location, but not, as some might expect, by the actual services provided. This
suggests that providers are indiﬀerent to what services they perform, as
long as they are paid for their time. Another interesting set of ﬁndings
concern the discrepancies observed between the ‘price equation’ and the
‘satisfaction equation’. A point made there is that, in a smoothly functioning market, any feature of a service that increases client satisfaction should
attract a price premium. We noticed, however, that in many cases such a
correspondence was not observed. The ﬁnding that ‘unattractive’ providers
generate less satisfaction while charging higher prices is one example, and
this was explained in terms of strategic behaviour on the part of the
provider: she is exploiting the sunk cost vulnerability of ﬁrst-time customers. This and other such ﬁndings suggest that market signals are not
functioning perfectly and there exist unrealized proﬁt opportunities.
Section 5 was mainly concerned with the legalization debate. This is one
debate in which most people have very strong views on one side or the other.
This may be because both sides of the argument are quite powerful, but it
is more likely to be because people are insuﬃciently informed. This section
was a modest attempt at a balanced treatment of the key issues in the

226

Economics uncut

debate. The key conclusions were that, whatever the perceived beneﬁts from
legalization, it is only appropriate to apply schemes at a geographically
local level, and certain types of legalization have the potential to alleviate
but not eliminate social problems. Another key point is that there is no evidence that the legal status of prostitution has ever actually inﬂuenced levels
of activity in the market. Therefore, legalization can be used to alter the
shape but not the size of the market.
An important point made was that societal rejection of prostitution has
its roots mainly in objections to the undesirable spillovers generated by the
industry, rather than in any moral objection. If these spillovers are reduced
or eliminated, society is likely to become much more tolerant. The Internet
is one means of reducing the visual pollution, since information on the
Internet is normally seen only by those actively searching for it. There is no
doubt that the use of the Internet as an advertising medium by sex workers
has taken oﬀ spectacularly: typing ‘escort service’ into a search engine gives
more than 100 000 hits! This must be seen as a beneﬁt to society if it means
that non-participants in the market are no longer exposed to it. Furthermore, if the Internet becomes the universal medium for advertising and
information exchange, evils such as child prostitution become more easily
detectable, since the Internet is accessible by everyone including the law
enforcement authorities.
Various models of legalization and toleration were also described in
Section 5. Each model clearly has attractions and limitations. On balance,
the zoning model appears to strike a sensible compromise, and it is not surprising that schemes of this type are springing up in many locations in the
Western world. However, one point that was not made is that legalization
may change the nature of the product in the eyes of the client, and reduce
the demand. Perhaps, there are some clients who are drawn to commercial
sex precisely because it is illegal, sordid, slightly risky, even dangerous.
Indeed, this perverse desire on the part of human males may be seen as a
further explanation of the high compensating wage diﬀerential that has
been a recurring theme in this chapter.
Notes
1.
2.
3.

‘Clipping’ is the practice of taking payment from a client and immediately departing
without delivery of the agreed service.
Original source: EUROPAP regional reports; data collated by Department of Politics
at the University of Exeter, http://www.ex.ac.uk/politics/pol_data/undergrad/aac/
scale.htm.
For example, two Thai women were jailed for ﬁve years in June 2003 for running a multimillion pound prostitution racket in the UK, in which hundreds of young Thai girls
were brought over to work as prostitutes and could not earn ‘freedom’ until they had
serviced between 500 and 1000 clients. For full report, see: http://news.bbc.co.uk/1/hi/
england/london/2960328.stm. Also in the UK, an Albanian asylum seeker was jailed for

Economics of prostitution

4.
5.
6.
7.
8.
9.

10.
11.

12.
13.
14.

227

ten years in December 2003 for similar oﬀences. See: http://newsvote.bbc.co.uk/1/hi/
uk/3340921.stm.
For full report see http://news.bbc.co.uk/1/hi/world/europe/3209541.stm.
http://www.iusw.org/start/.
One possible data source, which has been available for many years from sex shops in the
UK, is McCoy’s British Massage Parlour Guide.
See http://www.angelﬁre.com/stars/dorina/dpasiamiddleeast.html.
See http://www.punternet.com/saunaguide.html#The%20Dark%20Side.
‘Swinging’ is a pastime enjoyed by consenting adults of both genders, in which couples
congregate usually at a party at a private house, for the opportunity of temporarily swapping partners and engaging in sexual activities. The pastime is popular among couples
wishing to spice up their sex lives without compromising their marriages.
Interesting empirical studies of inﬁdelity appear in the microeconometrics literature,
dating back to Fair (1978). Results from these studies could conceivably be used to
obtain estimates of .
We are grateful to the web-manager ‘Galahad’ for permission to use the site for the
purpose of the empirical analysis that follows, and to Sheena Louks for painstaking
extraction of data from the site. Other sites exist that fulﬁl similar functions, for example,
the international site, World Sex Guide, http://www.worldsexguide.org.
The procedure used to obtain the smooth is known as lowess, which is originally due to
Cleveland (1979), and is now available in recent versions of SPSS.
See http://news.bbc.co.uk/1/hi/world/europe/3068827.stm.
See ‘Thais mull legalising sex trade’, http://newsvote.bbc.co.uk/1/hi/world/asia-paciﬁc/
3240824.stm.

References
Akerlof, G. (1970), ‘The market for lemons: quality uncertainty and the market mechanism’,
Quarterly Journal of Economics, 84, 488–500.
Albert, A. (2001), BROTHEL: Mustang Ranch and Its Women, USA: Random House.
Anonymous (1998), ‘The sex industry’, The Economist, 14 February.
Anonymous (2004), ‘It’s a foreigner’s game’, The Economist, 4 September.
Benson, C. and R. Matthews (1995), ‘Street prostitution: ten facts in search of a policy’,
International Journal of the Sociology of Law, 23, 395–415.
Boonchalaksi, W. and P. Guest (1998), ‘Prostitution in Thailand’, in L.L. Lim (ed.), The Sex
Sector: The Economic and Social Bases of Prostitution in Southeast Asia, Geneva:
International Labour Oﬃce.
Cameron, Samuel (2002), Economics of Sin: Rational Choice or No Choice At All?,
Cheltenham, UK and Northampton, MA, US: Edward Elgar.
Cameron, S. and A. Collins (2003), ‘Estimates of a model of male participation in the market
for female heterosexual prostitution services’, European Journal of Law and Economics, 16,
271–88.
Cameron, S., A. Collins and H. Thew (1999), ‘Prostitution services: an exploratory analysis’,
Applied Economics, 31, 1532–29.
Carael, M., J. Cleland and L. Adeokun (1991), ‘Overview and selected ﬁndings of sexual
behaviour surveys’, AIDS, 5, S65–S74.
Chapman, M. (2001), ‘Hard up students turn to vice’, BBC Internet report,
http://news.bbc.co.uk/1/hi/education/1303782.stm.
Clark, A. and A. Oswald (2002), ‘A simple statistical model for measuring how life events
aﬀect happiness’, International Journal of Epidemiology, 31, 1139–44.
Cleveland, W.S. (1979), ‘Robust locally weighted regression and smoothing scatterplots’,
Journal of the American Statistical Association, 74, 829–36.
Cragg, J.G. (1971), ‘Some statistical models for limited dependent variables with application
to the demand for durable goods’, Econometrica, 39, 829–44.
Della Giusta, M., M.L. di Tomasso and S. Strom (2004), ‘Another theory of prostitution’,
University of Reading, Economics and Management Discussion Papers, 013-2004.

228

Economics uncut

Doeringer, P. and M.J. Piore (1971), Internal Labour Market and Manpower Analysis,
Lexington, MA.
Edlund, L. and E. Korn (2002), ‘A theory of prostitution’, Journal of Political Economy, 110,
181–214.
Fair, R. (1978), ‘A theory of extramarital aﬀairs’, Journal of Political Economy, 86, 45–61.
Faugier, J. and M. Sargeant (1997), ‘Boyfriends, “pimps” and clients’, in G. Scrambler and
A. Scrambler (eds), Rethinking Prostitution, London: Routledge.
Fisher, T. (1997), Prostitution and the Victorians, New York: St Martin’s Press.
Gibb, J. (2002), ‘Hookers at sports day’, The Spectator, 23 November, 34–5.
Grundberg, K. (2003), ‘Sweden’s prostitutes ply their trade on the internet’, Agence France
Presse, 13 January.
Home Oﬃce (2004), ‘Paying the price: a consultation paper on prostitution’, http://news.bbc.
co.uk/l/shared/bsp/hi/pdfs/16_07_04_paying.pdf.
Hughes, W.T. (1996), ‘Uncertain neighbourhood eﬀects and restrictive covenants’, Journal of
Urban Economics, 39, 160–72.
Hughes, D.M. and C.M. Roche (1999), Making the Harm Visible, published by The Coalition
Against Traﬃcking in Women.
Humphries, P. (2003), ‘Green light districts’, The Guardian, 14 March.
Isaacs, A. (1993), Cassel Dictionary of Sex Quotations, London: Cassell.
Krugman, Paul (1991), Geography and Trade, MIT Press.
MacCoun, Robert J. and Peter Reuter (2001), Drug War Heresies: Learning From Other Vices,
Times and Places, Cambridge University Press.
Marizco, M. (2003), ‘Eﬀorts to control border prostitution ineﬀective, experts say’, Associated
Press, 30 April.
Matthews, R. (1997), Prostitution in London: An Audit, Middlesex University.
May, T., M. Edmunds, M. Hough and C. Harvey (1999), Street Business: The Links Between
Sex and Drugs Markets, Police Research Series, Paper 118, Home Oﬃce.
McAleer, P. (2003), ‘Happy hookers of Eastern Europe’, The Spectator, 5 April, p. 25.
Moﬀatt, P.G. and S.A. Peters (2004), ‘Pricing personal services: an empirical study of earnings in the UK prostitution industry’, Scottish Journal of Political Economy, 51, 675–90.
Morell, S. (1998), Aktionsmappe gegen Frauenhandel – Frauenhandel ist Frauen-verachtung,
Tubingen, Germany: Terre des Femmes.
O’Connell Davidson, J. (1998), Prostitution, Power and Freedom, UK: Polity Press.
Rao, V., I. Gupta, M. Lokshin and S. Jana (2003), ‘Sex workers and the cost of safe sex: the
compensating diﬀerential for condom use among Calcutta prostitutes’, Journal of
Development Economics, 71, 585–603.
Reynolds, Helen (1986), The Economics of Prostitution, Illinois, USA: Thomas.
Rosen, S. (1981), ‘The economics of superstars’, American Economic Review, 71, 845–58.
Sullivan, E. and W. Simon (1998), ‘The client: a social, psychological, and behavioural look at
the unseen patron of prostitution’, in J.E. Elias and G. Brewer (ed.), Prostitution, New York:
Prometheus Books.
Wolfenden, J. (1957), Report of the Committee on Homosexual Oﬀences and Prostitution,
London: Her Majesty’s Stationary Oﬃce.

Economics of suicide
Samuel Cameron

Résumé
Razors pain you;
Rivers are damp;
Acids stain you;
And drugs cause cramp;
Guns aren’t lawful;
Nooses give;
Gas smells awful;
You might as well live.
Dorothy Parker (1863–1967)

Suicide is not a regular part of the diet of the trainee economist. It is not
found in the typical microeconomics textbook so many professional economists remain unaware of the topic, ‘economics of suicide’. This chapter
sets out to show that it does indeed exist and why we should be interested.
There are three major factors in the field that might stimulate an economist to participate. First, microeconomic theory is founded on a rational
choice model that might find the phenomenon of suicide to be an interesting challenge. Second, the marked patterns to be found in the statistics of
suicide present a topic for the application of the formidable arsenal of
econometric techniques. For example, sociological pioneers in the subject
had noticed some degree of correlation of suicide rates with the business
cycle. Third, the tools of standard microeconomic theory may also be informative in looking at policy measures toward suicide.
The chapter proceeds as follows. Section 1 begins with a brief historical
overview of the study of suicide, with particular focus upon Emile
Durkheim’s socio-economic contribution. Next, Section 2 presents some
international statistics on suicide. We show the variation in rates of suicide
between countries, among diﬀerent age groups and gender, as well as how
these trends have changed over time. We also make reference to the
diﬃculties involved in interpreting such statistics.
Sections 3 and 4 examine economic models of suicide. We begin with the
simple notion of a demand for and supply of suicide and a discussion of
the landmark economic contribution to the subject by Hamermesh and
Soss in 1974. Section 4 extends the analysis by introducing more complex
approaches to suicide, including a game-theoretic framework.
229

230

Economics uncut

The empirical work undertaken by economists on suicide is presented in
Section 5. We examine the results of various studies, which are mostly based
at the aggregate level and include the testing of theories relating to gun availability and suicide. Section 6 investigates the potential economic costs of
suicide, including ‘celebrity’ suicides. The death of rock star, Kurt Cobain,
and its accompanying externalities, is a particular focus. Given the potential economic costs of suicide, we examine the economics of anti-suicide
policy in Section 7 with an analysis of how an anti-suicide bureau might
operate. Section 8 concludes the chapter.
1

An overview of the study of suicide

1.1 The Durkheim approach
Suicide is a subject that has attracted tremendous interest for thousands of
years, particularly from its religious treatment and also due to the fact that
its measured incidence displays very marked patterns by gender, age,
nation, ethnicity, state of the economy and time period. The most famous
contribution to theorizing on the subject was made in 1897 by economic
sociologist, Emile Durkheim, in his book, Le Suicide.
Durkheim put suicide on the agenda as a subject for social scientific debate
in contrast to the mainly moral and religious discussion of earlier times. He
identified four categories of suicide, which are a mixture of theorizing and
generalization, from his inspection of European suicide rate data:
1.1.1 Anomic suicide Anomic derives from the Greek word, anomie,
meaning lawlessness. Nomos means usage, custom or law and nemein means
to distribute. Therefore, anomy is social instability resulting from breakdown of standards and values. Anomic suicide, according to Durkheim, is
related to too low a degree of regulation or external constraint on individuals. This can occur when the ‘normal form’ of life is disrupted through either
a dramatic change in society or in the personal situation of an individual.
Durkheim divides anomic suicide into economic and domestic anomie.
The former can be generated during periods associated with economic
depression or over-rapid economic expansion. Individuals experience a
sense of disorientation and role confusion because of a large change in
their relationship with society. Similarly, in domestic anomie, a person may
not be able to cope with the change represented by the death of his or her
husband or wife. In both cases, the strain of the situation may force an
individual to commit suicide.
1.1.2 Fatalistic suicide In contrast to anomic, fatalistic suicide may
occur when regulation is too strong. Durkheim (1897) suggests that,

Economics of suicide

231

‘persons with futures pitilessly blocked and passions violently choked by
oppressive discipline’ may commit suicide. The duress of prison life is often
invoked as a situation where this might occur.
Having said that, we can observe that in situations of political confinement individuals react in the opposite way by struggling to construct
reasons to continue living. It may be that such attempts are an instrumental device to self-motivate individuals who hold on to the belief that one
day they will be freed from the situation.
1.1.3 Egoistic suicide Egoistic suicide refers to the case where the individual lacks suﬃcient integration within society, resulting in a feeling of not
belonging to any meaningful social group: ‘However individualized a man
may be, there is always something collective remaining – the very depression and melancholy resulting from this same exaggerated individualism’
(Durkheim, 1897).
According to Durkheim, individuals who are strongly integrated into a
family structure, a religious group or some other type of integrative group
are less likely to experience these feelings and commit suicide. Indeed, his
particular interest in this respect was the suicide rate amongst unmarried
males. This is to some extent reflected in the focus on divorce in empirical
work on suicide and does raise some questions about sexuality. For
example, an unmarried gay male who lives in a solid and supportive community should, on Durkheim’s premise, be less likely to commit suicide
than a comparable unmarried or divorced heterosexual male.
1.1.4 Altruistic suicide Integration is too great in the case of altruistic
suicide and the individual sacrifices themselves for the benefit of others.
One extreme of this would be suicide bombing to promote what the bomber
perceives as the collective good of a nation or religious group. More generally, in societies where life is diﬃcult and likely to be short it is not
unknown for older persons to take their own lives so that they will not be
a burden to their relatives. This is eﬀectively an intergenerational transfer
of utility, which would not be made if, at the margin, the would-be suicide
found that his or her marginal productivity was suﬃciently high to justify
continuing to live.
One can find other examples of this type. For example, a mother might
sacrifice her own life if this were essential to provide some type of body part
to save the life of her child. Or, reversing the situation to ‘assisted suicide’,
a person may find himself or herself in a situation of faculty loss such that
they have to transfer agency to someone else to do the altruistic deed. This
dilemma has been discussed extensively by economist Thomas Schelling
(1978, 1983) who couches the issue in terms of a divergence of preference

232

Economics uncut

sets. That is, the agents of assisted suicide face a problem in deciding
whether the person would now really want the termination of life he or she
specified when in a state not deemed fit to trigger it.
As is clear from the above, Durkheim located the four types of suicide
within a grid of social regulation. The core reason for any type arising is
that the individual is unable to integrate his or her own needs adequately
with those of society. Most of Durkheim’s taxonomical enterprise revolves
around the notion that people who do not feel they fit into society as
‘normal’ are at risk of suicide. At the time in which he wrote, this seemed
to lead to much exaltation of marriage as a state, that would stave oﬀ
suicide. Indeed, this family-oriented approach still seems to underlie much
contemporary research on suicide despite the increasingly overt diversity in
sexuality in developed economies. Outside the mainstream of suicide
research there is a body of literature on the higher suicide rates amongst
gay teen males.1 What this phenomenon suggests is that suicide is as much
a function of the hostility of society toward any apparent deviation as it is
of the individual to adapt to the needs of society.
1.2 By their own (invisible) hand
Subsequent to Durkheim, there has been a great deal of writing on the
subject by philosophers, psychiatrists and sociologists. There are even
‘suicidologists’2 who devote themselves entirely to the study of suicide and
have regular conferences on the subject. A small but significant contribution to the field of suicide has been made by economists. As noted earlier,
the pioneering work is by Hamermesh and Soss in 1974 and is discussed in
Section 3. Further, Durkheim’s approach has been assailed with a myriad
of criticisms and there are some features he neglected, which recent developments in microeconomics seek to incorporate. One is the increasing
tendency to look at conflict in preferences along the lines pursued by
Schelling, and the other is the use of signalling games to explore the use of
attempted suicide as a bargaining tactic to induce changes in behaviour by
other people. Both developments are discussed in more detail in Section 4.
Prior to the twentieth century, however, economists had shown little
interest in the subject. In 1759, Adam Smith devoted several pages to
suicide in The Theory of Moral Sentiments. The bulk of this discussion was
about ancient philosophical positions (especially Stoicism) on the matter of
whether it was morally wrong to kill oneself. His own position seems to be
that suicide is not a rational act but rather an outcome of some neurotic
disorder.
Seventeen years earlier, Smith’s fellow Scottish enlightenment pioneer of
economics, David Hume in Of Suicide, discussed largely the same issues.
Although also coming to the conclusion that suicide was not a crime to be

Economics of suicide

233

punished, he goes on to see suicide as a rational choice, saying, ‘I believe no
man ever threw away life while it was worth keeping’. In fact, Hume draws
attention to age and misfortune as causes of choosing to kill oneself, which
turn out to be the main drivers of the decision in the modern economic
theory of suicide. It is probably fair to say, however, that economists
working on the empirical study of the subject, rather than the pure theory,
have been somewhat more eclectic, even tending to incorporate a fair degree
of Durkheim’s original ideas in their reasoning (see, for example, Cameron,
2001; Lester and Yang, 1997; Brainerd, 2001).
2 Statistics of suicide
One guaranteed by-product of suicide is a large number of statistics routinely collected by national governments around the world. Figure 8.1
shows the wide variation in suicide rates across the USA and Europe over
selected time periods. For example, overall suicide rates range from 21.5 per
100 000 population in Finland in 2000 to 3.26 in Greece in 1999. Sweden
has experienced a 42.5 per cent fall in suicide rates between 1970 and 1999,
while Spain has experienced an increase in suicide rates of 52.6 per cent
between 1970 and 1998. It is worth nothing that despite the worldwide perception of the USA as being plagued by murder, in 2000 it had a suicide
rate of 10.7 per 100 000 population, which is greater than its murder rate
(see Figure 4.1 in Chapter 4).
Figures 8.2 and 8.3, meanwhile, clearly demonstrate the excess of male
over female suicide rates in all featured nations at all levels of the suicide
rate. Within the male and female population, again suicide rates vary
across countries and time. For example, Finland had particularly high
male and female suicide rates of 32.9 and 10.5 respectively in 2000, while
Greece had very low male and female suicide rates of 6 and 1.4 respectively. Spain experienced large increases in both male and female suicide
rates of 54.5 per cent and 35.6 per cent respectively over the 1970 to 1998
period.
According to McIntosh (2002), men in the USA are four times more
likely to die by suicide, while women are three times more likely to attempt
suicide. He also points out that women’s attempts are more frequent but less
violent and vice versa for men. (Two out of three male suicide victims in the
USA die by firearm compared with one out of three for women. The most
common method for female victims is overdose/poisoning.)
Table 8.1 shows the distribution of suicide rates by gender and age. The
suicide rate appears to increase with age across most countries. The highest
suicide rates are found among males in the 65 and over age group, while the
highest female suicide rates are reported in the 45–54 age group. In terms
of youth suicide, male suicide rates are greater than those among females

234
18.3
17.5
16.4
13.2

22.4
23.0
21.5

15.9 16.1

3.4 3.3
6.1 6.6
9.1 9.0 8.9 9.0 9.6

Norway (1970)

4.4 4.7

Suicide rates per 100 000 population in USA and Europe (1970–2000)

24.8

7.2

Data from Rodriguez Andres (2003) and National Center for Health Statistics (www.cdc.gov/nchs).

0.0

5.0

10.0

15.0

20.0

25.0

Greece (1970)

30.0

Figure 8.1

Source:

Suicide rate per 100 000 population

Austria (1970)
Austria (2000)
Belgium (1970)
Belgium (1996)
Denmark (1970)
Denmark (1998)
Finland (1970)
Finland (2000)
Germany (1990)
Germany (1999)

Greece (1999)
Italy (1970)
Italy (1999)
Netherlands (1970)
Netherlands (1999)

Norway (1999)
Portugal (1971)
Portugal (2000)
Spain (1970)
Spain (1998)

22.1

Sweden (1970)

17.5

19.5

12.7

Sweden (1999)
Switzerland (1970)
Switzerland (1998)
7.9

UK (1970)

7.0

UK (2000)

10.7

13.1

US (1970)
US (2000)

235
Figure 8.2

32.9

18.6

23.6

5.0
6.0
9.6

Italy (1970)

12.412.7
10.811.7

Norway (1970)

19.4
17.7

7.9 7.6
31.4

11.8

Male suicide rates per 100 000 population in USA and Europe (1970–2000)

19.6

23.0

27.9

23.1

27.7

Finland (1970)

38.6

Data from Rodriguez Andres (2003) and National Center for Health Statistics (www.cdc.gov/nchs).

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0 39.0

45.0

Source:

Suicide rate per 100 000 population

Austria (1970)
Austria (2000)
Belgium (1970)
Belgium (1996)
Denmark (1970)
Denmark (1998)

Finland (2000)
Germany (1990)
Germany (1999)
Greece (1970)
Greece (1999)

Italy (1999)
Netherlands (1970)
Netherlands (1999)

Norway (1999)
Portugal (1971)
Portugal (2000)
Spain (1970)
Spain (1998)
Sweden (1970)

29.9
27.3

18.3

Sweden (1999)
Switzerland (1970)
Switzerland (1998)
9.9

UK (1970)

19.8

11.1

UK (2000)
US (1970)

17.5

US (2000)

236
9.7
7.2
10.5
9.7

Finland (1970)
8.6

5.9

1.9
1.4
3.6

Italy (1970)

2.9
6.8
5.9
5.3
6.7

3.5
1.6
2.4
3.2
13.2

Female suicide rates per 100 000 population in USA and Europe (1970–2000)

8.8
10.8
16.0

Data from Rodriguez Andres (2003) and National Center for Health Statistics (www.cdc.gov/nchs).

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0 13.4

16.0

18.0

Figure 8.3

Source:

Suicide rate per 100 000 population

Austria (1970)
Austria (2000)
Belgium (1970)
Belgium (1996)
Denmark (1970)
Denmark (1998)

Finland (2000)
Germany (1990)
Germany (1999)
Greece (1970)
Greece (1999)

Italy (1999)
Netherlands (1970)
Netherlands (1999)
Norway (1970)
Norway (1999)
Portugal (1971)
Portugal (2000)
Spain (1970)
Spain (1998)
Sweden (1970)

7.4

Sweden (1999)

10.3

Switzerland (1970)
8.6

Switzerland (1998)
6.3

UK (1970)
3.1

UK (2000)

7.4

US (1970)
4.1

US (2000)

Economics of suicide

237

in the 15–24 age group. Moreover, male suicide rates for this age group have
increased over time in the USA and in almost all European countries.
McIntosh (2002) notes that in 2000, suicide was the third leading cause
of death of young people in the USA, whereas it was the eleventh leading
cause of death overall. Cutler et al. (2001) point out that the rise in suicide in
the USA among the 15–24 age group between 1950 and 1999 has been
centred mainly in the country’s rural states. At the time of Durkheim’s
writing, suicide was primarily an urban phenomenon and in fact he used
this as evidence that traditional agrarian (and particularly Catholic) society
played an important role in fostering a well-functioning social environment.
What the figures presented in this section do not show, however, is the
massive surge in suicide rates in the Eastern European economies in transition from communism, which is perhaps a form of anomic suicide. In
2000, Estonia, Ukraine, Latvia, Belarus and Russia all had significantly
higher male suicide rates per 100 000 population than any of the countries
featured in Figures 8.1 to 8.3 at 45.8, 52.1, 56.6, 63.6 and 70.6 respectively.
Moreover, as revealed in Chapter 4, a number of these countries, particularly Russia, have also developed a concurrent surge in the murder rate.
2.1 Problems with suicide statistics
Unfortunately, suicide data are even more subject to diﬃculties of interpretation and problems of error than routine economic statistics. The collection of suicide figures falls into the category of population statistics.
They are classified as a cause of death along with the other causes of
death and this obviously makes no attempt to harmonize diﬀerences in
the recording practices across countries. For example, some countries
record classified suicides only, whilst others have additional figures for
death by injury where it is not possible to determine whether the death
was on purpose or by accident. A classic case of this is the use of guns at
home and overdosing on medication. It is not uncommon to see it
reported that someone shot himself or herself by accident whilst cleaning
a gun or that death was due to an overdose of prescribed medication
taken by accident. Clearly, as we cannot retrospectively read the mind of
the dead person it is impossible to say what the true intentions were at the
time of the act.
Attempted suicides pose even greater problems of consistent data
comparisons. They will not be routinely recorded in national demographic
statistics. In the UK, the Samaritans quote figures for ‘self-harm’, which
are estimates taken from the Centre for Suicide Research at the Department
of Psychiatry, Warnerford Hospital, Oxford. Workers in psychiatric
services will tend to ask ‘at risk’ individuals whether they have felt like
‘harming themselves’ rather than whether they are thinking of suicide.

238

Year

Age

0.5
0.4
0.3
0
0.4
0.3
0
0.78
0.3
0.4
0.4
0.3
0
0
0.3
0.2
0.5
0.4

21.8
26.9
21.4
7.9
10.4
10.7
31.2
22.4
12.5
9.4
12.8
13.6
3.8
1.7
7.2
3.5
8.5
5.7

23.3
31.8
30.4
13.3
20.2
25.6
46.1
41.3
26.1
17.1
17.7
22.3
7.3
4.1
11.5
5.9
15
6.4

31.9
46.7
42.1
20.8
27
39.5
50.4
55.5
35.8
25.5
23.4
25.2
5.8
6.1
10.4
7.4
16.6
11.2

32.8
61.8
34.7
31.8
33.3
56
44.3
59.6
34.3
37.0
24.9
33.7
7.2
6.9
12.2
11.6
18.7
15.8

44.1
65.3
38.2
46.1
25.3
48.9
37.3
62.4
31.2
50.9
25.9
35.7
8.2
10.6
15.3
18.2
14.8
24

48.6
72.7
34.7
59.8
25.3
45.7
34.1
66.6
39.3
55.4
30.9
39.0
7.0
10.5
22.2
24
18.8
25.4

0.6
0
0.3
0
0
0
0.4
0.2
0.2
0
0.1
0.1
0
0.1
0
0.1
0.3
0.1

5.8
5.7
5.7
2.6
2.8
5.5
8.1
6.8
3.4
4.4
3
4
0.9
1.4
1.9
2.3
4.4
2

5.3
8.2
10.1
6.6
4
10.8
12.2
9.8
7.6
7.5
4.4
6.9
2
1.8
2.5
2.7
6.4
5.5

11.5
16.4
15.2
12
7.3
22.7
13.6
11.3
11.3
8
6.8
7.9
1.2
1.0
3.6
3.5
7.7
7.5

11.5
23
15.6
17.1
13.5
30.3
17.4
21
14.9
11.7
8.7
12.9
2.2
3.3
3.5
5.4
9.7
11.2

16.3
25.6
13.4
23.2
18.1
34.2
18.1
16.0
14.3
16.2
10.5
14.3
1.5
4.3
4.9
7
9.2
12.6

17.5
27.5
15.3
27.3
10.3
26.8
13.3
12.1
14.8
17.8
10.9
19.4
3.2
1.7
6.2
7.5
8.1
15

0–14 15–24 25–34 35–44 45–54 55–64 65–74 0–14 15–24 25–34 35–44 45–54 55–64 65–74

Female

Male

International suicide rates by age and gender per 100 000 population

2000
1970
Belgium
1996
1970
Denmark
1998
1970
Finland
2000
1970
France
1999
1970
Germany
1999
1970
Greece
1999
1970
Italy
1999
1970
Netherlands 1999
1970

Austria

Table 8.1

239

Source:

USA

1999
1970
2000
1971
1998
1970
1999
1970
1998
1970
2000
1970
2000
1970

0.2
0.2
0
0.3
0.1
0.1
0.3
0.4
0.5
0.5
0.1
0
1.2
0.5

28.1
5.3
3.6
4.2
7.8
2
14.7
17.8
21.8
22.6
10.6
5.8
17.1
13.5

26.1
15.2
4.9
9.2
13.5
4.1
16.1
28.4
28.5
26.6
17
8.9
19.6
19.8

19.2
17.3
8.7
14
12.2
7.1
24.7
44.7
27.6
28.7
16.8
11.7
22.8
22.1

23.2
19.3
6.4
22.4
13.1
9.6
26.3
52.5
35.3
47.7
15.8
13.9
22.4
27.9

25.4
24.7
10.8
35.7
15.9
15.6
27.6
54.7
44.3
44.6
11.7
17.9
19.4
32.7

31.2
19.6
20.7
50.2
23.2
22
30.4
46.3
42.7
50.5
10.4
19.7
22.7
36.0

0.2
0
0
0.3
0.1
0
0
0.4
0.3
0
0.1
0
0.3
0.2

9.1
2
0.5
1.4
1.5
0.9
6.3
7.7
5.6
5
2.9
2.4
3.0
4.2

8
4.3
1.2
3.1
3.1
1.4
7.4
15.4
5.4
9.9
3.9
4.6
4.3
8.6

Data from Rodriguez Andres (2003) and National Center for Health Statistics (www.cdc.gov/nchs).

Switzerland

Sweden

Spain

Portugal

Norway

9.9
10
1.1
3
3.8
1.6
11.7
19.2
10.5
8.2
4.4
7.4
6.4
11.9

11.1
10.4
2.1
6.3
4
3.8
11.9
26.1
14.5
19.3
4.9
10.7
6.7
12.6

5.4
9
3
5.9
4.8
5.4
10
18.2
13.6
19.5
3.8
12.9
54
11.4

9.4
6.9
3.8
6.3
7.8
5.7
9.6
15.6
16.7
22.2
3.6
14.1
4.0
9.0

240

Economics uncut

Notwithstanding this, a certain volume of self-harm (such as someone
cutting himself or herself to get attention) may not be attempted suicide.
To the non-economist, the main worry is simple accuracy in counting of
the totals. However, for the economist who is likely to draw heavily on the
coeﬃcient estimates from multiple regression equations there is the more
serious problem of an inflated error variance leading to over-acceptance of
null hypotheses about the relationships with the supposed determinants of
suicide rates. As discussed further in Section 3, the economist will tend to
model suicide as a supply function, which should show the total intended
oﬀer of suicides at the existing ‘prices’ of suicide. They are, in eﬀect, allocation of time functions where the individual allocates time to suicide
attempts, not all of which will be successful. So, any estimated supply curve
based on recorded suicides will be distorted by two types of omission:
1.

Some deaths that were due to suicide are included in the figures under
‘other causes of death’. The misallocation of suicides to other sources
of death is often deliberate due to the culture of a particular region. In
the extreme case we find a zero number of suicides. This is the case for
a country such as Egypt as it is forbidden by national law.
The non-inclusion of the failed attempts. The number of these massively exceeds the number of completed suicides. For example, in the
USA, for every teen that commits suicide (one-hundredth of 1 per cent
each year), 400 teens report attempting suicide (4 per cent per year),
100 report requiring medical attention for a suicide attempt (1 per cent
per year), and 30 are hospitalized for a suicide attempt (0.3 per cent
year) (Cutler et al., 2001).

Researchers can make adjustments by moving the estimated number of suicides in other death categories to the suicide category and they can add the
failed attempts to the completed suicides. However, evidence suggests that
the underlying correlates of these are very diﬀerent. Despite all this, one can
at least be confident that the total number of suicides presented in this
section for all countries are an underestimate of the true number.
3 The simple microeconomics of suicide
The previous section presents a very diverse and confusing picture, which
defies simple explanation in terms of social or cultural factors. What now
follows is an examination of how suicidology could perhaps benefit from
more of an economic focus. For instance, we outline the Hamermesh and
Soss (1974) economic theory of suicide, which predicts that suicide rises
with age and falls with income. But first we ask whether we can look at the
issue using a simple demand and supply framework.

Economics of suicide

241

3.1 Demand for and supply of suicide
Lester and Yang (1997) claim that we can indeed have a supply and
demand model for suicide. However, this seems rather hard to imagine
without a great deal of strain. Admittedly, many economic situations in
which we are accustomed to seeing textbooks apply supply and demand
models are not literally applicable. Clearly, suicide is not a product that is
sold in a market at a price of so much per unit and, unless it were to
become mainly a spectator sport, it is hard to see how there are ‘consumers’ who demand it. There may be a derived demand for inputs to
suicide production such as drugs for peaceful termination of life, but there
is no demand for suicides as such. However, we can make a case, as we do
in Section 4, that there is an implicit or derived demand for celebrity suicides.
The supply of suicide may be viewed as a ‘price’ response in the sense of
there being a net profit or benefit from suicide in terms of the utility
diﬀerence between death and life. The costs of suicide include transaction
costs in the form of the money costs and time costs of eﬀecting departure
from the world. For example, as we noted in Section 2, men and women
tend to choose radically diﬀerent methods for committing suicide. Indeed,
a suicidal individual may want to research this and might have to use up
time seeking out his or her most preferred choice. On the other hand, one
economic factor conducive to suicide is that low-cost methods are always
available and, if he or she is at liberty and no one suspects his or her intentions, little eﬀort will be needed to make the attempt undetected.
We can find some casual evidence to support the argument that suicides
have a rational component in the shape of price responsiveness. Gassing
oneself in the domestic oven was once a long-established method of suicide
in England and Wales. In 1963, for instance, suicide by domestic gas
accounted for more than 40 per cent of suicides in England and Wales but
when the facility to do this was reduced by the falling carbon monoxide
content of the gas supply (see Kreitman, 1976) there was a marked drop in
the total number of suicides as shown in Figure 8.4.
Specifically, between 1963 and 1975, the annual number of suicides in
England and Wales fell from 5714 to 3693 at a time when suicide continued to rise in most other European countries. Few of those prevented from
using gas appear to have found some other method of killing themselves.
This tale is consistent with the notion of a ‘marginal suicide’ who would
drop out of the supply of suicide attempts if the net rate of return falls
suﬃciently due to a cost rise. However, the suicide rate did return to its
‘typical’ level after a few years, suggesting that there is a diﬀerence between
the long-run and short-run price elasticities of suicide due to a lagged
adjustment to equilibrium taking place.

242

1960

1959

1958

1971

1970

1969

1968

1967

1966

1965

1964

1963

1962

1961

The gas story (suicides in England and Wales, 1958–77)

National Statistics, UK.

1000

2000

3000

4000

5000

Figure 8.4

Source:

Total number of suicides

6000

Total suicides

Suicides by domestic gas

1977

1976

1975

1974

1973

1972

Economics of suicide

243

Attempted and completed suicide may have explicit and implicit ‘punishment’ costs, for example, in religious doctrine. These may be the basis of
the common treatment of suicide as a criminal oﬀence. Suicide was only
decriminalized in the UK in 1961 and in the Republic of Ireland in 1993
and it has been forbidden in many religions with the suggestion that it condemns the individual to Hell or Purgatory and can involve the loss of
certain burial rights. Having said that, in some minor religions suicide has
been a way of ‘life’ for individuals and in some cults, group suicide of
members has been carried out.
Yet, standard economic models of suicide do not fully consider the role
of religion in the suicide decision even though the Subjective Expected
Utility model has its origin in worries about the existence of God. This is
in the form of ‘Pascal’s Wager’ wherein seventeenth-century French mathematician Blaise Pascal explored the wisdom of deciding on the existence
of God by way of calling heads or tails on a coin toss, namely, a 50–50 bet
(see Chapter 9). Extrapolating this to the notion that a suicide may land one
in Hell, it follows that belief in God may retard suicides and, ceteris paribus,
will do to a greater extent the more risk-averse one is. Other explicit and
implicit ‘punishment’ costs of attempted and completed suicide arise
through the eﬀects on others, which are explored in Section 4.
3.2 The Hamermesh and Soss model
So, how can we fit suicide into an economic model? Is it not rather outside
the boundaries of such an approach, as economics is so heavily premised
on rational choice whereas the voluntary termination of one’s life seems to
be far from rational? Looking at the pioneering contribution by
Hamermesh and Soss (1974), the answer to these questions seems to be
‘easily’ and ‘no’.
Hamermesh and Soss (1974) continued the trend of ‘economic imperialism’ inspired by Gary Becker whereby traditional topics and problems of
other disciplines were penetrated by the insights of economic models
without much incorporation of the approach of the other disciplines. This
strictly neoclassical economics approach was continued in a textbook on
investment decisions by Dixit and Pindyck (1994) and is carried on in a
paper by Mirer (1998).
The core economic model of suicide revolves around consumption and
the basic premise is an individual maximizing total discounted lifetime
utility. Utility is a function of consumption, which in turn is a function of
age and income, and Hamermesh and Soss further assume that the individual has a given discount rate, which is known over his or her lifetime. In
order for suicide to be a welfare improvement, the individual must have
reached the point where any further existence brings a negative discounted

244

Economics uncut

utility, although it is important to note that this does not necessarily mean
that all future periods have negative net utility.
Specifically, the present value of lifetime expected utility, Z, at age a, is
given by:

Z(a,Y ) a

er (ma)UmP(m)d m

(8.1)

where a* denotes life expectancy or maximum attainable age, Y is permanent income, r is the discount rate, P(m) is the probability of living at age m
given that one is alive at age a, and Um is utility for an individual, which, as
already noted, is a function of consumption, which in turn depends upon
age and income.
From equation (8.1), therefore, suicide is a straightforward investment
value decision. The term on the left represents the value of the life an individual may have left. If his or her permanent income rises then, ceteris
paribus, the value of life rises and continuing to live is preferable to committing suicide. Age enters into the lifetime utility function as a negative
component, however. This is because as an individual gets older, his or her
most valuable asset (life) has decreased in duration and thus must have a
lower total discounted value. Formally, er (m a) decreases because (m a)
increases with age. Moreover, the lower the probability of survival, P(m),
the higher the probability the individual will commit suicide.
In this model, life itself does not have an independent value. Rather it is
just a collection of sums of utility and the person does calculations on the
time pattern of these to see whether he or she should switch oﬀ the utility
generating machine that is his or her body before its current expected termination date. Dixit and Pindyck (1994) suggest that one should circumvent this limitation by allowing for ‘option value’ of life. Option value gives
rise to ‘option demand’. In the traditional cost–benefit example, we might
value the pure existence of a wildlife park even though we never intend to
go to see it. In the suicide example, we attach value to having a life even
though we might think that it will not be worth living again in terms of its
net utility benefit from the components of existence. This is consistent with
certain traditions in religious teaching.
A modified version of this model could have implications for an individual’s
health stock and euthanasia. If life expectancy has decreased then the incentive to invest in health has decreased suggesting that one is voluntarily further
reducing expected life length even though the decision is not an explicit one
to perform a suicidal act. In fact, euthanasia (assisted suicide) might appear
as an optimal choice in cases of decreased life expectancy.

Economics of suicide
4

245

Not so simple?

4.1 Preferences and suicide
The microeconomic model of suicide is perhaps complex in its mathematical
development but is undoubtedly simple in terms of the underlying assumptions.
The would-be suicide is treated as a rational individual in the sense of having a
stable unified intertemporal utility function with a constant discount rate and
risk perceptions based on a Von Neumann–Morgenstern utility function.
All these assumptions have been increasingly questioned in fields such as
insurance and risk-taking in general. In the case of such an important decision as deciding to kill oneself, it is surely even more imperative that key
assumptions be scrutinized. I will concentrate on the most philosophical of
these assumptions, rather than the more technical ones, that is, there is a set
of completely known, unified and stable preferences.
The standard model can be modified to allow for ‘mistakes’ in the sense
of deciding to kill oneself due to faulty discounting or over-estimation of
future periods of negative utility. Indeed, this suggests a ‘public goods’ role
for providing useful information about such things as depression and
addiction. Such mistakes are, therefore, due to market failure or individual
cognitive limitations.
If we take the last insight somewhat further, namely, that the individual
has a limited ability to process information in order to arrive at a rational
choice, then we move into the field of ‘expanded preference’ models. The
simplest way to expand the preference ordering approach of economics is
to suppose that a person has two distinct sets of preferences. This raises
two questions: (i) Which one of these will they try to maximize? (ii) Which
one of these should they try to maximize? The first question covers the
area of ‘positive economics’ whilst the second covers the area of ‘normative economics’, including the whole formal apparatus of welfare economics. A number of authors have written on such ‘dual preference’ models,
which do have a long pedigree in philosophical literature. Various distinctions can be made as to the status of each set of preferences:
(i) Long-run versus short-run preferences
This could involve the notion of correcting ‘mistakes’ if individuals are
involved in dynamic learning. Admittedly, a completed suicide allows no
scope for such learning but, even so, someone on the margins of attempted
suicide may learn from observing the conduct of others.
(ii) Higher versus lower preferences
The individual may wish to follow the higher preference set but succumbs
to temptation and gives in to the lower. In the case of suicide, this could go

246

Economics uncut

either way. An individual may regret the acting out of suicidal urges as a
lower order preference that requires controls to be placed on him or her.
Conversely, an individual may believe that suicide corresponds to the
optimum dictated by his or her higher order preferences but regrets the possible cowardliness or recalcitrance in giving in to the temptation to stay
alive. In a dual preference set-up of this type, the individual would optimize
by choosing a suicide method that is harder to withdraw from part way
through the attempt, provided it did not have a level of money and transactions costs that oﬀset the gain.
(iii) Temporary versus permanent preferences
This approach was implied earlier when we discussed Schelling’s papers
on assisted suicide. Modern interest in it derives from the work of Jon
Elster (1998, p. 70) on the neglected role of the emotions in economics. It
should not be confused with category (i), long-run versus short-run preferences. In that case, the short-run preferences will turn into the long-run
preferences when the individual reaches full equilibrium. Any variable
change that alters the cost of suicide would promote this. Elster’s temporary preference notion concerns the case of preferences that flare up in
response to stimuli. In the suicide case, bad or depressing news may precipitate the temporary preference set to the level where suicide seems
optimal in the current frame of mind. However, if the act is delayed long
enough (which could happen due to random variation in such things as
delays in the time taken to procure any necessary suicide equipment or
pills) the permanent preferences will take over and greatly decrease the
likelihood of completed action. This is the way the Samaritans in the UK
seek to operate by providing a phone-in service for those on the margins
of suicide.
A notable case of temporary preferences for suicide is ‘romantic suicides’. This does not refer to the narrow ‘Romeo and Juliet’ scenario of
lovers who face insurmountable odds to be together as this falls into the
standard model. The romantic suicide, as termed here, is one where an individual wants to commit suicide because he or she feels that the current
period has just delivered such a sublime experience that everything afterwards is doomed to be a comparative disappointment even though the flow
of utility may be positive at all times. This has been expressed in various art
forms over the years and achieved a notable popular manifestation in the
song, ‘There Is A Light That Never Goes Out’, by The Smiths, released in
1986. In this the narrator wishes to die because he is in a state of bliss due
to finally being beside the object of his unrequited love.

Economics of suicide

247

4.2 Para-suicide signalling games
Specialists use the term ‘para-suicide’ to denote a case of deliberate selfharm not intended to result in death. This raises questions about how this
could be seen to bring any utility at all to the perpetrator in a strictly
economic model. One form of reconciliation would be to allow a source
of direct utility from self-harm. A wider explanation of para-suicide,
however, would seem to require us to view it as an extreme strategy in a
game-theoretic model of relationship management first developed formally
by Rosenthal (1993).
Let us assume that there are two parties, the would-be para-suicide who
we term AP (aggrieved party) and the person he or she sees as responsible
for his or her condition, SO (significant other). The strategy of para-suicide
is then a form of investment risk taken in an attempt to raise the discounted
utility of staying alive by inducing changes in the behaviour of others.
The strategy is needed because the AP sees the SO as causing them negative utility and is unable to find a superior means of conveying the satisfactory degree of angst. If the SO only feels weak external eﬀects of the
negative externalities they inflict on the AP then a strategy of exaggeration
is needed. In other words, if lesser strategies of exaggeration are subject to
diminishing returns (due to a lessening impact on the SO’s utility function),
then the probable optimality of para-suicide will increase.
The showbusiness and entertainment world provides illuminating tales of
the para-suicide persuasion strategy. Well-known musician Peter Gabriel
experienced attempted suicides from women who wanted children he was
not willing to supply. Fortunately, these people survived. On the other hand,
the actress Sarah Miles details in her autobiography the repeated emotional
blackmailing of suicide threats by a woman who had become dependent on
her. Finally, when thrown out of Miles’s house the person jumped oﬀ a
building. This case resonates with us due to the implicit cost of guilt we
would feel if a serial para-suicide or threatener thereof killed themselves.
4.2.1 Suicidal children Cutler et al. (2001) use a game-theoretic framework to explain the rise in attempted suicides amongst teens in the USA,
which we noted in Section 1. They suggest that a child’s suicide signal may
convince their parents that they are genuinely miserable so that the latter
choose to allocate more resources to the child such as time and money. This
is referred to as the direct value of a suicide signal. In other cases, the
parents may in principle be unwilling to distribute these resources to the
child, but give in due to the internal or external cost (embarrassment, for
example) at having a child attempt suicide.
To illustrate this formally, Cutler et al. consider a child with a utility
function V(T, Z), where T is the amount of time or money the parents

248

Economics uncut

transfer to the child and Z is a vector of other factors that aﬀect the child’s
happiness. The parents obtain utility both from their own consumption
[U(Y T)] and the happiness of their child, [aV(T, Z)]. The child’s happiness is assumed to be known to the child but not to the parents.
Unhappy children may want more parental input and the value of
parental resources is greater when the child is exogenously less happy (that
is, d 2V/dTdZ0). As child utility is not observable to the parent, suicide
attempts are a credible threat to signal this unhappiness if there is some
probability that they succeed and if the utility loss from death is smaller for
unhappy children. Cutler et al. show that if parents cannot observe the
child’s happiness, equilibrium is where children with ZZ* attempt suicide
and children with Z Z* do not.
In summary, these results indicate that a child communicates his or her
unhappiness to their parents so that more resources such as time and
money are received by the child. Suicide is, therefore, one potential signal
because a child who is less happy values future life less than one who is
happier. However, while a suicide attempt may convince parents of their
child’s unhappiness, occasionally, for the signal to be transmitted, the child
must die.
5

Empirical work

5.1 Alcohol, business cycles and suicide
In Section 2 we described some international statistics on suicide. Much of
the literature on suicide is empirically based at an aggregate level in contrast
to the highly individualistic formulation of the Hamermesh–Soss model,
for instance. In their original 1974 paper, Hamermesh and Soss estimate
equations using aggregate cross-section and time-series US data, which
found that suicide rates tend to increase with age and decrease with income.
It is extremely unlikely, however, that the ideal database for testing the
economic model of suicide could ever be obtained. This would require a
large survey-based analysis using a random sample of many people over a
large number of years, which included faithful recording of suicide
attempts. The nearest data to this is the University of Southern Illinois’
Core Institute, which conducts annual surveys of college students in the
USA, focusing on drinking, drug use and outcomes associated with drinking and drug use. Markowitz et al. (2002) use the Core Institute’s 1991
survey and report a causal relationship from alcohol and illicit drug consumption to suicide thoughts and attempts.
Other surveys from the USA containing information on suicide include
the Youth Risk Behavior Survey (YRBS) and the National Comorbidity
Survey (NCS). The YRBS samples diﬀerent groups of teenagers (between

Economics of suicide

249

10 000 and 15 000) in grade 9–12 in the USA every two years between 1991
and 1999, while the survey respondents in the NCS are 8098 individuals
aged 15 to 54 between 1990 and 1992, with the majority of interviews being
conducted in 1991. Markowitz et al. (2003) combine the two surveys in a
study that reports that a causal relationship between ‘binge’ drinking and
suicide attempts among youth in the USA is very unlikely, but a causal relationship does exist between clinically defined alcohol use disorders and
suicide attempts among girls.
Marcotte (2003) uses the NCS conducted in 1991 and 1992 and employs
the indirect approach of estimating an earnings function with the log of
annual earnings as the dependent variable and suicide-related dummies
appearing as ‘explanatory’ variables for earnings. This provides little evidence that actual past attempts influence earnings as the relationship
may well reflect a number of other factors that are associated with suicide,
but it does show that thinking about suicide is correlated with lower later
earnings.
The more typical econometric study uses aggregate statistics following in
the path of Hamermesh and Soss. A paper by Rodriguez Andres (2003), for
instance, investigates the determinants of suicide rates in 15 European
countries, using age-standardized suicide rates for both men and women
over the period 1970 to 1997 as the fluctuations to be explained. The results
show that divorce, per capita income, unemployment, income inequality
and alcohol consumption have significant eﬀects on suicide rates but these
results diﬀer across age groups and gender. For instance, female suicide
rates seem to be more sensitive to changes in per capita income than male
suicide rates. Unemployment has a positive influence on suicide rates for
both sexes. Income inequality exhibits a positive and significant eﬀect on
male suicide rates but a negative and significant eﬀect on female suicide
rates. Divorce appears to have a positive and significant eﬀect on suicide
rates and its impact declines with age. A negative and significant eﬀect was
found of alcohol consumption on suicide rates for both sexes.
The Rodriguez Andres paper focuses on a fairly stable set of economies.
In contrast, the work of Brainerd (2001) looked at 22 transition economies
of Eastern Europe, which indicated that male suicide rates are highly sensitive to the state of the macroeconomy. This suggests that the steep and
prolonged declines in GDP in the western countries of the former Soviet
Union may have been partly to blame for the suicide epidemic that we highlighted in Section 1. The evidence also indicates that the general adult male
mortality crisis in the region had an epidemic or cascading quality in that
the loss of a spouse or friend (or simply declining life expectancy) contributed to rising suicide rates. Female suicide rates were not statistically
significantly related to the state of the macroeconomy and were more

250

Economics uncut

strongly related to alcohol consumption. A lesser sensitivity of female
suicide rates to income movements was also found in a study of Taiwanese
data by Cheung and Huang (1997) and in work reported in Lester and
Yang (1997).
One response to the problem of scarce and deficient suicide data is to
apply simulations to a more precisely formulated theoretical model. In
these simulations the researcher traces out the probability of an event
occurring in response to diﬀerent combinations of the parameters and
causal variables. For instance, Mirer (1998) uses a life-cycle model of retirement planning in which rational choice regarding the length of life can lead
to suicide. This model is similar to the Hamermesh–Soss model, although
it is couched in discrete rather than continuous terms and contains an
explicit term for risk-aversion. Similarly, it ignores the bequest motive,
which may appear to be relevant in actual cases of suicide.
Mirer’s model is concerned with the allocation of a given level of initial
wealth over an uncertain lifetime, and the critical assumption is that an
individual regards the utility of living with some particular low level of
consumption, C*, as equivalent to the utility of death. In the simulations,
various fractions of the sustainable level of consumption are used to
approximate C* and Mirer calculates the likelihood of suicide in some
plausible economic circumstances. In essence, the model attributes suicide
to a fear of poverty in old age for someone in employment, implying
unemployment may also be a cause of suicide. If unemployment is perceived as long term by an individual, who does not have adequate sources
of non-earned income, then it can be seen as a form of forced retirement.
Similarly, sudden arrivals of a large debt problem from causes such as business collapse or gambling addiction would precipitate suicide on a rational
choice basis.
However, we should not automatically jump to the conclusion that statistically significant relationships of suicide with income validate the microeconomic model of suicide, which is based essentially on substitution at the
margin. Income and unemployment at the aggregate level may be indicators
of stress (along with other stressors like social tension and divorce) bearing
down upon individuals as per the various formulations of Durkheimian
models (see Lester and Yang, 1997). For example, male–female diﬀerences
in aggregate income sensitivity may represent males feeling more distress at
the loss of their symbolic role as the wage earner. Falling or negative growth
rates and rising unemployment may also create a social milieu in which the
support networks available to help potential suicides are curtailed by lack
of resources. That is, there may be restricted budgets for mental and other
health service provision, but also more reluctance from family and friends
to allocate time inputs to those in need.

Economics of suicide

251

It should also be apparent from the above discussion that the studies by
economists include many variables that do not follow obviously from the
economic model; that is, alcoholism and divorce are treated largely as
proxies for stress. Studies by non-economists such as those surveyed in Stack
(2000) and Platt and Hawton (2000) also include these variables but interpret other variables such as income and unemployment in a diﬀerent way
from economists. Usually these interpretations derive from some version of
Durkheim’s ideas.
5.2 Gun availability and suicide
Attempts to estimate a loosely formulated supply equation for suicide have
tended to focus on precipitating factors (labour market conditions, alcohol,
divorce and so on) rather than opportunity factors, notwithstanding the
well-known UK ‘gas supply eﬀect’ that we noted in Section 3. Leaving aside
all other statistical problems, this could seriously bias the results found in the
research studies. The most obvious opportunity factor that is likely to vary
substantially across countries, and over time, is access to guns. Table 8.2
provides some evidence on this, and in doing so reflects the unfortunate fact
that gun availability is not readily measured.
Surprisingly, there is no obvious simple correlation between gun availability and suicide. The small literature on the relationship is surveyed by
Kleck (1997) who comes to the conclusion that guns simply create a substitution eﬀect, in the mode of death, rather than adding to the total supply
of suicides. He points out that in the USA from 1972 to 1995, the estimated
stock of guns increased by 50 per cent but the aggregate suicide rate fluctuated within a fairly narrow range.
In the absence of satisfactory information on gun ownership rates, one
could resort to arguments based on diﬀerences in gun legislation across
countries, although Stolinsky (2000) notes that this does not help support the
idea of a gun–suicide relationship as many countries with liberal gun control
regimes have fairly low gun suicide rates. Ludwig and Cook (2000) examined
the impact of the Brady Law in the USA on murder and suicide rates. In states
where criminal background checks were not already required for gun purchases, the Brady Law mandated a background check and a five-day waiting
period. When the laws were introduced in March 1994, 18 states already had
laws as strict as or stricter than the Brady Law, and so Ludwig and Cook were
able to categorize these as the ‘control group’. The other 32 states were the
‘experimental’ group. They found that changes in rates of homicide and
suicide for treatment and control states were not significantly diﬀerent. The
only category in which they found a statistically significant change caused by
the Brady Law was a reduction in firearms suicides among persons aged 55
years or over. However, while firearms suicides fell significantly among this

252

Economics uncut

Table 8.2 International suicide rates and firearm suicides per 100 000
population
Country

Year

Total
suicide

Firearm
suicide

% Households
with guns

Estonia
Hungary
Slovenia
Finland
Brazil
Denmark
Austria
Switzerland
France
Mexico
Belgium
Portugal
United States
Japan
Sweden
Germany
Taiwan
Singapore
Canada
Mauritius
Argentina
Norway
N. Ireland
Australia
New Zealand
Scotland
Hong Kong
Netherlands
South Korea
Ireland
Italy
England/Wales
Israel
Spain
Greece
Kuwait

1994
1994
1994
1994
1993
1993
1994
1994
1994
1994
1990
1994
1993
1994
1993
1994
1996
1994
1992
1993
1994
1993
1994
1994
1993
1994
1993
1994
1994
1991
1992
1992
1993
1993
1994
1995

40.95
35.38
31.16
27.26
3.46
22.13
22.12
21.28
20.79
2.89
19.04
14.83
12.06
16.72
15.75
15.64
6.88
14.06
13.19
12.98
6.71
13.64
8.41
12.65
12.81
12.16
10.29
10.10
9.48
9.81
8.00
7.68
7.05
7.77
3.40
1.66

3.13
0.88
2.51
5.78
0.73
2.25
4.06
5.61
5.14
0.91
2.56
1.28
7.35
0.04
2.09
1.17
0.12
0.17
3.72
0.09
3.05
3.95
1.34
2.35
2.14
0.31
0.07
0.31
0.02
0.94
1.11
0.33
1.84
0.43
0.84
0.06

n/a
n/a
n/a
23.2
n/a
n/a
n/a
27.2
22.6
n/a
16.6
n/a
39.0
n/a
15.1
8.9
n/a
n/a
29.1
n/a
n/a
32.0
8.4
19.4
22.3
4.7
n/a
1.9
n/a
n/a
16.0
4.7
n/a
13.1
n/a
n/a

Source: Adapted from http://www.guncite.com/gun_control_gcgvintl.html, which gives the
original sources.

Economics of suicide

253

group, the overall suicide rate did not fall by a statistically significant amount
because non-firearm suicides rose at the same time.
6

Economic costs of suicide for society

6.1 Loss of human capital stock
Given the statistics documented in Section 2, it appears that the economic
costs of suicide are potentially very large. Suicide might be likened to a war
that carries oﬀ a certain portion of the stock of human capital on a regular
basis. Human capital can also be eroded by damage due to failed or parasuicides. If suicide is disproportionately concentrated amongst the old and
variously infirm and poor/unemployed it could be argued that the economic loss is greatly diminished because of reductions in health care
expenditures and social security payments.
Looking at the UK, there seems to be a selective removal of some parts
of the human capital stock. National Statistics provides data on the relative statistical risk of suicides by occupation in the form of the PMR
(Proportional Mortality Ratio). This shows the proportion of deaths from
suicide in a group compared with the national average rate, which is set at
100. Therefore, a PMR of 200 indicates twice the average risk, while a PMR
of 50 indicates half the average risk. Over the period 1991 to 1996 for men
aged 20 to 64 and women aged 16 to 59, Tables 8.3 and 8.4 show a profound
skew towards loss of those in caring and information professions.
American PMR figures are discussed in detail by Stack (2000) using the
1990 Population Census, but only for 21 states. This shows similarities to the
UK pattern with dentists, doctors, mathematicians and scientists, artists
and social workers scoring above the expected rate of suicide. His figures
show only farmers not to be a high-risk group although there is a slight risk
factor for farm workers. Stack’s figures are not broken down by gender.
Researchers attempting to assess occupational risk of suicide face the
problem of whether or not persons in so-called high-risk occupations have
the said risk because of stress, economic or otherwise, associated with the
occupation or because of its demographic composition. Second, there is the
‘problem’ of tiny samples. For instance, although dentists appear to be at
high risk from suicide according to Tables 8.3 and 8.4, they only represent
a small fraction of the total population, only a small fraction of them die
in a given year, and only a small fraction of those deaths are by suicide.
However, it does seem likely that one important factor influencing the risk
in a number of the occupations featured in the tables is access to potentially
lethal means of suicide (for example, shotguns in the case of farmers,
and pharmaceuticals in the case of those in the caring professions).
Indeed, 36 per cent of male farming-related suicides in the UK involve

254

Economics uncut

Table 8.3

UK men (aged 20–64) suicide by occupation (1991–96)

Occupation
Veterinarians
Dental practitioners
Pharmacists
Garage proprietors
Sales representatives –
property and services
Medical practitioners
Farmers, horticulturists,
farm managers
Publicans
Other motor drivers
Cleaners, window cleaners,
road sweepers
Painters and decorators,
French polishers
Builders
Shop salesmen and assistants
Gardeners, groundsmen
Carpenters and joiners

PMR

Suicides

324
249
171
155
151

9
25
25
43
97

147
144

71
190

128
124
122

129
221
204

119

389

119
118
117
115

332
296
234
384

Source: Kelly and Bunting (1998).

Table 8.4

UK women (aged 20–59) suicide by occupation (1991–96)

Occupation
Veterinarians
Medical practitioners
Domestic housekeepers
Waitresses
Professional and related in
education, welfare and health
Students
Cleaners, window cleaners,
road sweepers
Nurse administrators, nurses
Hospital ward orderlies
Source: Kelly and Bunting (1998).

PMR

Deaths

500
285
247
187
183

4
25
16
37
26

139
138

132
95

137
130

240
139

Economics of suicide

255

firearms compared with only 5 per cent of all male suicides (Kelly and
Bunting, 1998).
6.2 Celebrity suicides: the Kurt Cobain case
Now we discuss the consequences of celebrity suicide. Take, for example,
the suicide in 1994 of Nirvana rock star, Kurt Cobain. This has two components: the suicide itself and the impact of his suicide on the value of his
product (music).
Artists and celebrities can be classified into two groups. Some artists
produce works that are relevant to their choice of suicide, while others do
not. For example, Cobain’s music and lyrics reflect the misery, despair, hopelessness and helplessness that he experienced in his own life. Thus, his death
from suicide validated his artistic products (including lyrics) in the eyes of
his audience. Listening to his music may result in sensations, reflections and
inspirations that provide insights into one’s own life (as well as that of
Cobain). The products acquire iconic and increasing value for those who
have already purchased the products.3 This can be characterized as a positive externality, but diﬀerent from a monopoly restriction of output, which
results from the supply of works decreasing after an artist dies.
In the realm of literature, the poets Sylvia Plath and Anne Sexton wrote
poems about their mental illness and their suicidality, after which they
killed themselves. The writings of Ernest Hemingway, although somewhat
autobiographical in nature, do not provide clues to the despair and suicidality of his old age. It could be hypothesized that the positive externality
will be greater in the first group than in the latter group.
Let us examine the economic impact of Kurt Cobain’s suicide. To do so,
we need to compare the social benefit from his suicide with the social cost
of his lost life.
6.2.1 Positive externalities of Cobain’s death Among the economic gains
are those mentioned above that accrue to Cobain’s fans, but there are also
positive externalities to the record companies that produced his compact
discs and accompanying merchandise, as well as the economic gains to
his widow and heirs. The headline in the weekly trade journal Billboard on
23 April 1994, read ‘Cobain Death Spurs Rush At Retail’ (Rosen and
Morris, 1994). Nirvana had four albums already released, and all four made
gains in the week following Cobain’s suicide:
●
●

In Utero (1993) jumped from number 72 to 27 in the Billboard 200,
with an increase in sales from 18 000 to 40 000 units.
Nevermind (1991) jumped from number 167 to 56, with an increase in
sales from 7000 to 20 000 units.

256
●
●

Economics uncut
Incesticide (1992) re-entered the Billboard 200 at number 135, with
an increase in sales from 2000 to 8000.
Bleach (1989), Nirvana’s debut album, entered the Top Pop Catalogue
chart for the first time at number 6, with an increase in sales from 2000
to 9000 units.

Sales of posters and collectibles (such as t-shirts) also rose. However,
suicide as a cause of death may have several eﬀects on sales that may be
diﬃcult to separate:
1.

Owners of intellectual property associated with the deceased artist will
step up marketing once the suicide occurs, such as releasing compilations of the artist’s songs. In the case of Kurt Cobain, his publishers
also considered releasing unpublished tracks, and the group’s album,
Unplugged In New York, was released in 1994, after Cobain’s suicide.
The suicide is its own advertisement, reminding people that the artist
existed and, by its newsworthiness may alert hitherto non-consumers
of the product to its existence.
The suicide may have a ‘salience’ eﬀect on the consumer, which could
include a ‘Van Gogh romantic suﬀering eﬀect’ (as perpetuated in the
pop song ‘Vincent’ recorded by Don McLean), which makes the artist
somehow seem more important.
Associates may capitalize on the suicide (intentionally or unintentionally). For example, people who worked with the artist may enhance
their own career, as perhaps New Order did when they emerged from
the remains of Joy Division after lead singer Ian Curtis hanged himself.
Interestingly, Cobain’s suicide occurred just four days before his wife’s
debut album was released (Live Through This, by Hole, led by Courtney
Love). Cobain’s suicide most certainly had an impact on sales of this
album too.
Other artists may benefit by making ‘covers’ of the original music by
the deceased artist(s), as might the writers of the songs recorded by the
deceased artist.

6.2.2 Negative externalities of Cobain’s death The social cost of a
suicide such as Kurt Cobain’s typically consists of the potential value of his
future productivity and the associated externalities. However, the life-cycle
of artistic creativity and productivity must also be taken into account.
Artists do not generally stay creative and productive throughout their lives.
In Cobain’s case, he died at an early age (27). Thus, he might have reached
his peak in a few more years and then begun a downhill slide. It is generally the case in ‘popular’ music that artists produce their best-selling works

Economics of suicide

257

early on as compared with visual and literary artists. Therefore, the potential productivity of Cobain’s future artistic output may be much less than
was generated by his suicide, even without taking into account his maladaptive lifestyle and his substance abuse. Indeed, it is possible that future
mediocre works might have blighted a legacy, leading to negative reappraisals and possible lower sales of the peak-period work.
However, we may still face a problem that emulation eﬀects are stimulated, leading to the emergence of some non-beneficial suicides. In other
words, there may be negative externalities associated with Cobain’s death.
If these emulated suicides are in suﬃcient number, then this may oﬀset the
gains from the suicide of the artist. Emulation may come about due to the
loss of an emotional support system for some consumers. In the case of an
artist whose work is based on misery and depression, other depressed individuals may lose (after the suicide) the reassurance that here is someone
‘like them’ (a kind of imaginary friend) who is able to cope by expressing
how he or she feels. This may reduce their reasons for continuing to live.
Does the suicide of an icon persuade or dissuade others to commit
suicide? It is diﬃcult to find examples of dissuasion. The death of Jesus
Christ can be viewed as a victim-precipitated homicide, which may
have been motivated in part by suicidal desires, probably unconscious
(Wolfgang, 1959). In a way, Jesus Christ’s death perhaps dissuaded some in
subsequent years not to commit suicide. However, imitation is definitely the
norm. Phillips (1974) examined the impact of newspaper stories about suicides from 1947 to 1968 and found that, in general, there was an increase in
suicides in the month after the story. The more publicity devoted by the
Press to the suicide, the greater the increase in the number of suicides.
Stack (1990a) found that both celebrity suicides (such as Marilyn Monroe’s
in 1962) and suicides by high profile non-celebrities (such as Victor
Kravchenko, a Russian defector to the USA, in 1966) increased the number
of suicides in the following month, but the impact of celebrity suicides was
much greater than the impact of non-celebrity suicides.
For example, in the month after the suicide of Marilyn Monroe in 1962
(reported on page 1 of The New York Times), there were 197 more suicides
in the USA than expected (1838 versus 1641, which in eﬀect captures the
size of the negative externality). The suicide of James Forrestal in 1949
(a former US Secretary of Defense) was followed by an excess of 56 suicides. The self-immolation of Norma Morrison in 1965 to protest against
the Vietnam War resulted in an excess of 58 suicides. However, not all suicides reported on page 1 of The New York Times were followed by a rise in
the observed number of suicides over the number expected. For instance,
the suicide of novelist Ross Lockridge in 1948 was followed by a decrease
of 12 suicides in the following month over the number expected.

258

Economics uncut

Stack (1990b) commented on the impact of ‘star suicides’, noting that
stars promote a high level of mass identification among the audience since
he or she represents the ordinary and the unusual at the same time. The
audience identifies with the star on two levels: in the ordinary roles that the
star plays (father, lover and so on) on the screen, and in the luxurious life
that the star leads in real life. There is also evidence that the imitation (or
suggestion) eﬀect of celebrity suicides is greater among those who resemble the celebrity in age and sex (Schmidtke and Hafner, 1988).
There have been two preliminary studies of the impact of Kurt Cobain’s
suicide on suicidal behaviour. Jobes et al. (1996) noted that Cobain’s suicide
was met with immense public grieving:
Radio stations played Nirvana’s music around the clock; television, particularly
MTV, showed concert footage, interviews, and music videos. Grieving fans gathered at Cobain’s residence, leaving notes and flowers. News of his death was the
lead story in the nation’s press and made the front cover of major magazines
such as Newsweek and People magazines; Rolling Stone devoted an entire issue
to Cobain.

In Seattle, there was one copycat suicide (a 28-year-old male attended the
candlelight vigil, went home and shot himself in the head), but no increase
in suicides in the next four weeks. However, the Seattle Crisis Clinic experienced an increase in calls in the weeks following Cobain’s suicide (Jobes
et al., 1996). In Australia, Martin and Koo (1997) compared the change in
the suicide rate for those aged 15 to 24 in the four weeks after Cobain’s
suicide as compared with the same period in neighbouring years. No
increase was noted. In fact, there were fewer suicides than in the neighbouring years. Thus, in these two preliminary and small studies, no suggestion eﬀect was observed.
However, across the USA as a whole, the average number of suicides per
day in April 1994 (Cobain committed suicide on 8 April 1994) was 90.8. In
the two previous years, the average numbers per day in April were 83.8 and
87.8; in the two subsequent years, the average numbers per day in April
were 88.3 and 86.6. Thus, there was a peak in suicide in the April of the year
in which Cobain committed suicide.4
7 Economics of anti-suicide policy
As we have seen in the previous section, suicide may involve direct loss of
output from destruction of part of the human capital stock and thus could
be an important economic, as well as social, problem. However, unlike
murder and drug addiction, it has not attracted government appointment
of a specific agency, ‘tzar’ or guru to deal with the issue. Instead, the major
direct attention to the suicide problem comes from voluntary public goods

Economics of suicide

259

provision. In the UK, as already noted, this comes in the shape of the
Samaritans. In this final section we provide some basic analysis of how a
government ‘anti-suicide bureau’ would operate.
If suicide and para-suicide are reflective of deeper underlying social problems that might be represented in the term ‘decrease in social capital’ (see
Sobel, 2002, for a discussion of the social capital concept from an economist’s viewpoint), then more benefits might accrue from directing policies
towards these problems. As far as I am aware the only paper by an economist
on the eﬃciency of suicide prevention is that by Medoﬀ (1986), but there is
a literature on suicide prevention centres and policies particularly focused on
universities where an at-risk group is exposed to additional stress factors
(see, for example, the report of University of British Columbia, 1999).
Medoﬀ’s paper presents some very interesting results, which suggest
there might be some mileage in the type of analysis presented in this
section. His results consist of two-stage least squares estimates of a suicide
supply equation using American state data for 1979 of the usual type with
an added variable for the density of suicide prevention centres per population. These are centres of the ‘Samaritan’ type designed to change the tastes
of those contemplating suicide. The estimates suggest that one more suicide
centre would have reduced suicide rates by 3.7 white males per 100 000 population and by one white female per 100 000 population.
Using some crude estimates of net worth of life saved based on earnings,
Medoﬀ concludes that the benefit–cost ratios showed a return per dollar of
between 4.96 and 7.10 for suicide prevention centres. This suggests that
more centres might be a highly eﬃcient policy but one should have some
caution over such an isolated study. It is notable that he obtains some
unusual results for the variables that are traditionally included in such
models, casting some doubt on the validity of the estimates.
Supposing that we can use public spending to reduce suicide rates then
we may still face standard public finance problems of an equity–eﬃciency
trade-oﬀ illustrated in the crudely simplified diagram of Figure 8.5.
This is a strictly equilibrium analysis as we abstract from eﬀects on
national output from the expenditure on suicide prevention. We also
abstract from all demographic dimensions (wealth, residence, gender, race
and so on) of suicide variability other than age and further divide the population of potential suicides into ‘old’ and ‘young’. We assume that the
total number of young is the same as the total number of old. The figure
has two schedules: a social welfare function giving society’s preferences
over the possible combinations of old–young suicides and a transformation
curve showing the suicide levels that can be obtained by allocating a fixed
budget for a ‘suicide prevention bureau’ diﬀerently with respect to expenditure that deters the old and the young.

260

Economics uncut

Old
suicides
prevented

Of
f
e

SWF2
SWF1

Young suicides
prevented

Figure 8.5 Optimal suicide reduction with choice over ‘old’ and ‘young’
suicides
The social welfare function (SWF) is negatively sloped as it reflects the
trade-oﬀ between suicide of young and old. It is also very steep (that is, the
marginal rate of substitution is very large) because the assumption that
saving a young person is more valuable than saving an old person implies
that if we give up saving a young person, we must save several more old
people to leave us equally well oﬀ. The transformation curve is assumed to
be downward-sloping on the grounds that there is some trade-oﬀ in deterring ‘old’ versus ‘young’ suicides and vice versa. It is drawn as concave on
the assumption that the marginal rates of reduction of suicide via spending do not decrease at the same rate.
The figure shows the standard welfare optimum where the marginal rate
of substitution in (social) consumption is equal to the marginal rate of
transformation in production. This is point e, which has a very high ratio
of old-to-young suicides due to the assumed shape of the two schedules.
However, suppose someone objected that this is unfair. The ‘fair’ point in
terms of equal risk is shown at point f, which involves a welfare loss to

Economics of suicide

261

society equal to the utility value of the gap between SWF1 and SWF2. Any
move to equality might be said to involve ‘sacrificing’ the number of young
equal to Ye Yf in order to ‘save’ the number of old equal to Of Oe.
8 Conclusion
It may have come as something of a surprise to learn that economists have
made serious contributions to the study of suicide. Notwithstanding, their
approach to the subject is far from surprising, in that variations in suicide
rates are explained from a basis of individualistic rational choice economics. The pure theory papers on the subject have been more in the nature of
spin-oﬀs of other areas like the human capital model, retirement planning
and health capital models.
One thing economics brings to any topic is a degree of precision and
clarity. Economic models clearly predict that rises in income and age will,
respectively, decrease and increase the rate of suicide. Although this has
been largely borne out in econometric work, the major features of suicide in
many countries present something of a challenge. For example, suicide rates
have shown a tendency to rise in some places in the face of rising average
incomes and particularly within the group of younger males. If these trends
persist we could delegate the explanation to other subject areas or simply
write them oﬀ as mechanical statistical aberrations (namely, younger males
might be getting better at completing attempted suicides and/or the reluctance to record suicides as being so might be declining suddenly). The alternative is to incorporate more external factors in the rational choice model in
the form of influences of social groups and society.
Notes
1.
2.
3.

See http://www.youth-suicide.com/gay-bisexual/ for references and abstracts of papers.
See, for example, the website of the American Association of Suicidologists at http://
www.suicidology.org.
Some artists may also sing or write of misery and depression while living a life of luxury,
such as James Taylor who often based his songs on his own mental ill-health, or Johnny
Marr and Patrick Morrissey of The Smiths. Their deaths would probably not have such
impact.
I would like to thank Dr John McIntosh for supplying these data.

References
Bernstein, P.L. (1996), Against the Gods: The Remarkable Story of Risk, New York: John
Wiley and Sons.
Brainerd, E. (2001), ‘Economic Reform and Mortality in the Former Soviet Union: A Study
of the Suicide Epidemic in the 1990s’, IZA Discussion Papers No. 243, January, Bonn.
Cameron, S. (1997), ‘Economics of Preference Change: The Case of Therapy’, Journal of
Economic Psychology, 18, 453–63.
Cameron, S. (2001), ‘Self-execution, Capital Punishment and the Economics of Murder’,
American Journal of Economics and Sociology, 60 (4), 881–90.

262

Economics uncut

Cheung, H. and W. Huang (1997), ‘Economic Correlates of Regional Suicide Rates: A Pooled
Cross-section and Time Series Analysis’, Journal of Socio-Economics, 26, 277–89.
Cutler, D., E.L. Glaeser and K.E. Norberg (2001), ‘Explaining the Rise in Youth Suicide’, in
Jonathan Gruber (ed.), Risky Behavior Among Youths: An Economic Analysis, Chicago:
University of Chicago Press, pp. 219–69.
Czapinski, J. (1995), ‘Money Isn’t Everything: On the Various Social Costs of Transformation’,
Polish Sociological Review, 112, 289–302.
Dixit, Avinash K. and Robert S. Pindyck (1994), Investment Under Uncertainty, Princeton
University Press.
Durkheim, Emile (1897), Le Suicide, Paris: Felix Alcan.
Elster, J. (1998), ‘Emotions and Economic Theory’, Journal of Economic Literature, 36 (1),
47–74.
Hamermesh D. and N. Soss (1974), ‘An Economic Theory of Suicide’, Journal of Political
Economy, 82, 83–98.
Hume, D. (1904) [1741–42], ‘Of Suicide’, in Essays: Moral, Political and Literary, London,
Edinburgh, Glasgow, New York and Toronto: Henry Frowde.
Jobes, D.A., A.L. Berman, P.W. O’Carroll, S. Eastguard and S. Knickmeyer (1996), ‘The Kurt
Cobain Suicide Crisis: Perspectives from Research, Public Health and the News Media’,
Suicide and Life-Threatening Behavior, 26, 260–71.
Kelly, S. and J. Bunting (1998), ‘Trends in Suicide in England and Wales, 1982–1996’,
Population Trends, No. 92, Summer.
Kleck, Gary (1997), Targeting Guns: Firearms and Their Control, New York: de Gruyter.
Kreitman, N. (1976), ‘The Coal Gas Story: United Kingdom Suicide Rates 1960–1971’,
British Journal of Preventive Social Medicine, 30, 86–93.
Lester, David and Bijou Yang (1997), The Economy and Suicide: Economic Perspectives on
Suicide, Commack, NY: Nova Science.
Ludwig, J. and P.J. Cook (2000), ‘Homicide and Suicide Rates Associated with Implementation
of the Brady “Handgun Violence Prevention Act” ’, Journal of the American Medical
Association, 284 (5), 585–91.
Marcotte, D. (2003), ‘The Economics of Suicide, Revisited’, Southern Economic Journal,
69 (3), 628–43.
Markowitz, S., P. Chatterji, R. Kaestner and D. Dave (2002), ‘Substance Use and Suicidal
Behaviors Among Young Adults’, National Bureau of Economic Research Working Paper
No. W8810.
Markowitz, S., P. Chatterji, R. Kaestner and D. Dave (2003), ‘Alcohol Abuse and Suicide
Attempts Among Youth – Correlation or Causation?’, National Bureau of Economic
Research Working Paper No. W9638.
Martin, G. and L. Koo (1997), ‘Celebrity Suicide: Did the Death of Kurt Cobain Influence
Young Suicides in Australia?’, Archives of Suicide Research, 3 (3), 187–98.
McIntosh, J. (2002), ‘USA 2000 Oﬃcial Data’, prepared for American Association of
Suicidology.
Medoﬀ, M.H. (1986), ‘An Evaluation of the Eﬀectiveness of Suicide Prevention Centers’,
Journal of Behavioral Economics, 15, 43–55.
Mirer, T. (1998), ‘Rational Suicide and the Optimal Length of Life’, State University of New
York, Albany, unpublished manuscript.
Parker, Dorothy (1926), Enough Rope: Poems by Dorothy Parker, New York: Boni and
Liveright.
Phillips, D.P. (1974), ‘The Influence of Suggestion on Suicide: Substantive and Theoretical
Implications of the Werther Eﬀect’, American Sociological Review, 39, 340–54.
Platt, S. and K. Hawton (2000), ‘Suicidal Behaviour and the Labour Market’, in Keith Hawton
and Kees van Heeringen, The International Handbook of Suicide and Attempted Suicide,
Chichester, England: John Wiley and Sons, pp. 303–78.
Rodriguez Andres, A. (2003), ‘Suicide Rates in European Countries: An Empirical
Investigation’, unpublished manuscript.
Rosen, C. and C. Morris (1994), ‘Cobain Death Spurs Sales at Retail’, Billboard, 23 April.

Economics of suicide

263

Rosenthal, R.W. (1993), ‘Suicide Attempts and Signaling Games’, Mathematical Social
Sciences, 26, 25–33.
Schelling, T. (1978), ‘Strategic Relationships in Dying’, in E. McMullin, Death and Decision,
Boulder, Colorado: Westview Press.
Schelling, T. (1983), ‘Ethics, Law and the Exercise of Self-command’, in Sterling
M. McMurrin (ed.), The Tanner Lectures on Human Value IV, Salt Lake City: University of
Utah Press.
Schmitdke, A. and H. Hafner (1988), ‘The Werner Eﬀect After Television Films: New
Evidence for an Old Hypothesis’, Psychological Medicine, 18, 665–76.
Smith, Adam (1976) [1759], The Theory of Moral Sentiments, Indianapolis: Liberty Classics.
Sobel, J. (2002), ‘Can We Trust Social Capital?’, Journal of Economic Literature, 40 (1),
139–54.
Stack, S. (1990a), ‘A Reanalysis of the Impact of Non-celebrity Suicides’, Social Psychiatry
and Psychiatric Epidemiology, 25, 269–73.
Stack, S. (1990b), ‘Audience Receptiveness, the Media, and Aged Suicide’, Journal of Aging
Studies, 4 (2), 195–209.
Stack, S. (2000), ‘Work and the Economy’, Chapter 8 in Ronald W. Maris, Alan L. Berman
and Morton M. Silverman (eds), Comprehensive Textbook of Suicidology, New York:
Guilford Publications.
Stolinsky, D. (2000), ‘America: The Most Violent Nation?’, The Medical Sentinel, 5 (6),
199–201.
University of British Columbia (1999), ‘Report of the Suicide Prevention Planning
Committee’.
Wolfgang, M.E. (1959), ‘Suicide by Means of Victim-precipitated Homicide’, Journal of
Clinical and Experimental Psychopathology, 20, 335–49.

Economics of religion
Robert J. Stonebraker

Economists hold nothing too sacred to invade with supply and demand
curves, not even matters of faith. Religion may not be bought and sold like
fish and chips, but there are parallels. It is a diﬀerentiated commodity that
is produced and consumed in a competitive marketplace. Consumers make
choices about which brands of religion to consume and how much religion
to consume. Some of us choose to ‘believe’, and others not. Some of us
choose to be Catholic, others Baptist, others Muslim. What factors drive
those choices? Is it the hand of God, or is it the hand of economics?
Religion is supplied and marketed by firms called churches, or maybe synagogues or mosques. Each must choose production technologies, set prices
and deal with free-riders. How are these choices made? They must compete
among themselves as well as with secular firms for the time, attention and
financial support of potential adherents. How do they compete and what
determines which will prosper and which will not?
Section 1 of this chapter considers the demand for religion and how it
can be analysed in terms of such traditional economic factors as scarcity,
opportunity cost, prices of related goods and capital stocks. Section 2
analyses supply-side issues with a particular emphasis on how providers of
religious goods and services cope with endemic free-rider issues. Section 3
explores the problems of risk and uncertainty in religious choice and
Section 4 concludes by looking at how competition and government regulation shape the structure and conduct of religious markets.
1 The demand for religion
Religion took quite a bashing from nineteenth-century intellectuals.
Nietzsche proclaimed that ‘God is dead’. Karl Marx and Sigmund Freud
predicted that religion would be crushed by the relentless march of science
and rationality, and sociologists began theorizing about the dawning age of
secularization. Similar claims mark much modern discourse.
Curiously, the data refuse to submit. Religious adherence and church
attendance as a percentage of population have risen over time in the USA.
In recent surveys more than 60 per cent of Americans claim membership in
a specific religious group compared with a paltry 17 per cent in 1776
(Iannaccone et al., 1997). While religious declines have accompanied modernization in a handful of countries, recent studies indicate that religion
264

Economics of religion

265

remains remarkably robust in most areas of the world. In the Gallup
International Millennium Survey almost two-thirds of respondents claimed
that God is very important in their lives (see Figure 9.1) and over 80 per cent
considered themselves to be a part of some religion (see Figure 9.2).
1.1 Scarcity and the benefits of religion
This continuing demand for religion should be no surprise. Economists
assume that scarcity drives behaviour. We invest heavily to find new and
better ways of producing more and more goods. While these eﬀorts certainly matter, human technologies have inherent limits; limits that can be
transcended by gods. Religion oﬀers the chance that supernatural beings
might supply us with rewards, both current and eternal, well beyond what
we could supply ourselves. One important lure of religion is the potential
increase in our collective production possibility frontiers. In this sense, religion becomes an additional technology that rational humans might use to
combat scarcity. Using the terminology of sociologists Rodney Stark and
William Bainbridge (1987), religion provides compensators that promise
people rewards they could not otherwise obtain in the physical world of
the ‘here and now’. According to economist Laurence Iannaccone, ‘the
demand for supernaturalism is as basic and irrepressible as the wants it
seeks to satisfy’ (Iannaccone, 2002b).
Moreover, the advance of science and ‘rational’ thought will not necessarily undermine religion. In the world of supernaturalism Iannaccone
(2002b) diﬀerentiates magic, which involves the belief in impersonal supernatural forces, from religion, which involves belief in supernatural beings.
Magic can objectively be studied. For example, claims that tarot cards accurately forecast the future can be subjected to scientific analysis. Controlled
experiments can be used to verify or reject such claims. However, supernatural beings elude such simple tests. It is true that we cannot prove that
supernatural beings exist; but neither can we prove that they do not.
Confronted with such uncertainty, rational humans might well choose to
cut their risks and jump onto the religious bandwagon. As the philosopher
Pascal put it long ago, ‘Let us weigh the gain and loss in wagering that God
is . . . If you gain, you gain all; if you lose, you lose nothing. Wager then,
without hesitation that He is’ (Pascal, 1660). To Pascal, investments in religious activity are perfectly rational reactions to eternal risk.
Because religion conflicts neither with science nor rational thought in
this context, there is no logical reason to expect it to wither away in modern
society. Magical aspects of the supernatural that can be disproved will fall
prey to science, but religion will not. Indeed, among college professors,
those in the ‘most’ scientific fields such as mathematics, chemistry, physics
and biology profess the most religious faith (Iannaccone et al., 1998).

266

100

Total

Western Europe

Eastern Europe

West Africa

Respondents claiming to be part of a religion

Southeast Asia

Gallup International Millennium Survey, 1999, http://www.gallup-international.com/.

Figure 9.1

Source:

North America

South America

267

Total

Western Europe Eastern Europe

West Africa

Southeast Asia

Respondents claiming God is very important in their lives

Gallup International Millennium Survey, 1999, http://www.gallup-international.com/.

Figure 9.2

Source:

100

North America

South America

268

Economics uncut

Many consumers emphasize the potential benefits religion might provide
them in the future. Indeed, an initial economic study posited that the
demand for religion springs from the demand for afterlife consumption
(Azzi and Ehrenberg, 1975). However, religious activity can also oﬀer concrete benefits in the here and now. It is one avenue, among others, that might
provide consumers with a sense of aﬃrmation and a sense of meaning or
purpose. As such, it can create a sense of happiness or serenity in the
present. Religion can also provide a wide assortment of other social and
personal benefits. Religious activity can create improvements in both physical and emotional well-being through a variety of channels (Ellison, 1998).
For example, church attendance can build stronger social networks, provide
both formal and informal support systems and improve skills for coping
with stress. In addition, by discouraging tobacco and alcohol, religious
beliefs can promote healthier life-styles.
1.2 Determinants of demand
To model the demand for religious activity most economic studies begin
with a variant of Gary Becker’s (Becker, 1965) household production
model. They assume that households and individuals will allocate their
scarce time between secular and religious goods to maximize utility.
Although initial studies emphasized the ‘afterlife’ value of religious activity, recent ones pay increased attention to the current benefits derived
from participation (Sullivan, 1985; Sawkins et al., 1997; Heineck, 2001). In
these models individuals maximize an intertemporal utility function that
depends upon the consumption of secular and religious goods and services
in each time period. Specifically they assume:
U U(S1, S2, . . . , St,. . . , Sn, R1, R2, ,. . . , Rt,. . . , Rn)

(9.1)

The St represent an individual’s consumption bundles of secular goods and
services in each time period t, and the Rt are consumption bundles of religious goods and services. The values of St depend upon the quantities of a
composite secular good purchased in period t (QSt) and the amount of
leisure time in period t (TSt). Each Rt, in turn, depends upon the quantity
of a composite religious good purchased (QRt) and the amount of time
spent engaging in religious activities (TRt). Or:
St S(QSt, TSt) and

(9.2)

Rt R(QRt, TRt)

(9.3)

Economics of religion

269

In these models, time is devoted either to consuming leisure (TSt), to consuming religious activities (TRt) or to work that provides the income needed
to purchase the composite secular and religious goods.
Religious activity responds to conventional microeconomic variables in
these models. For example, price should negatively aﬀect the consumption
of religious goods just as it negatively aﬀects our willingness to buy secular
items. Prices could be explicit prices for purchased religious goods or
implicit prices that reflect the opportunity cost of time spent in religious
activities. As a simple illustration, when inclement weather, three-day holidays or championship football games raise the opportunity cost of time,
attendance at religious services falls. Statistical studies typically have examined the impacts of such traditional variables as wage rates, the price and
availability of complementary and substitute goods and consumer tastes
and preferences.
Wage rates can have several eﬀects. Since higher wage rates increase the
opportunity cost of time spent in religious activities, consumers might substitute toward work and away from religious activities (TRt falls). On the
other hand, higher wages can also raise income and release individuals
from having to work as many hours. With increased income and fewer
hours on the job, consumers might also decide they can better aﬀord to
invest money and time into religious involvement. In this scenario, both QRt
and TRt can rise. Thus, the theoretical impact on the quantity of religion
demanded is ambiguous and statistical studies do show mixed results
(Sawkins et al., 1997).
More interestingly, the models predict that wages should aﬀect the type
of religious involvement demanded. When wages and income rise, consumers should shift from time-intensive to money-intensive forms of religion. Using the terminology above, QRt should rise relative to TRt. And it
does. For example, studies find that consumers with higher incomes do not
necessarily spend more time in religious pursuits, but they do throw more
money in the oﬀering plates. Financial contributions relative to attendance
rises with income. Wealthy consumers tend to join congregations that hire
professional staﬀ to preach, to teach, to sing, to cook and to clean. Religious
groups that rely upon and expect members to shoulder those responsibilities often attract members with lower wages (Iannaccone, 1998).
Complementary and substitute goods also matter. Just as a cup of tea can
be enhanced by judicious amounts of lemon and sugar, the benefit of religious activity can be enhanced if it is shared by other family members.
Indeed, families with mixed religious preferences experience significantly
higher rates of divorce (Lehrer and Chiswick, 1993). As we would expect
given these religious complementarities, people are inclined to marry within
their religion or to convert to their spouse’s religion in a mixed marriage.

270

Economics uncut

We also find higher levels of religious activity among spouses who share the
same faith (Iannaccone, 1998) and among families with children (Sawkins
et al., 1997). As for substitutes, the most likely suspect is employment. Hours
spent on the job are hours not available for consuming most religious activities. Employment does have the predicted eﬀect. Historically, men have
been more likely to enter religious communes such as the Shakers during
periods of high unemployment (Murray, 1995) and any modern cleric can
cite examples of how the influx of women into the labour market decreased
the number of volunteer hours that women oﬀered their local church.
Tastes and preferences or perceived value is a final consideration. If the
benefits of religion stem from its ability to provide compensators and
social support, then its perceived value might be moulded by perceived
needs for these factors. For example, there is some evidence that religiosity
is aﬀected by illness and the recent death of a loved one (Miller, 1995) and,
although the evidence is mixed, the tendency for adults’ religious activity to
increase with age might partly stem from increased concern with afterlife
consumption.
1.3 Religious capital
Extending this basic model to include human capital sheds additional light
on patterns of religious behaviour (Iannaccone, 1992). In the secular world,
individuals routinely invest in education, training and practice to build
human capital. The sacred realm is no diﬀerent. Just as an athlete practises
to build athletic skills and a neurosurgeon studies to build medical skills,
those who study and participate in religious activities create ‘religious
capital’ that enhances their religious skills and enjoyment.
The appreciation and enjoyment of religion, like that of music and art
and literature, is a skilled consumption activity; it requires investments in
human capital. Those who have never struggled to produce music themselves will not appreciate the full power and beauty of a Bach requiem. To
be pleasurable, a stimulus must contain the right mix of redundancy and
novelty. Experiences too familiar or redundant are boring; we need an
element of surprise or novelty to hold our attention or interest. But experiences can be too new as well. Those we perceive as totally foreign can
bewilder, even frighten. A film such as Alien might delight a teenager accustomed to cinematic horror, yet terrorize a younger child with no prior exposure to movie mayhem.
Movies are not alone. According to the late economist Tibor Scitovsky,
art cannot be fully appreciated unless it is in a reasonably familiar style, and
successful literature must portray characters and dilemmas to which we can
relate. Similarly, conversation, even gossip, must be about people or places
we know if it is to hold our interest (Scitovsky, 1992). In this sense, cultural

Economics of religion

271

capital provides the background we need to appreciate additional information; it provides the redundancy needed to appreciate additional experiences. Just as knowledge of art can increase the demand for and the benefits
from additional experiences in art, religious capital can raise the demand for
and benefits from additional religious activity. Those who know the liturgy
and theological nuances find more meaning in church services and sermons.
Those already familiar with the melodies find hymns more uplifting. Those
already plugged into congregational networks find more social support.
Religious capital creates a feedback loop. Participation in religious activities creates religious capital that, in turn, raises the benefit from and
demand for further participation. The fact that this capital grows over time
provides an alternate explanation as to why religious participation tends to
increase over an adult’s lifespan. Similar to models of addiction discussed
earlier in this book, increased consumption in one period raises the demand
for consumption in the next. Mathematically:
Rt R(QRt, TRt, RCt)

(9.4)

where RCt is the amount of religious capital ‘owned’ by a consumer. Over
time, RCt grows with the consumer’s religious consumption. Specifically:
RCt RC(QRt1, TRt1, RCt1)

(9.5)

Much religious capital cannot easily be transferred across denominations or even congregations. Thirty years of faithful participation in the
local Methodist congregation does nothing to prepare one for life as a
Hindu. Because religious capital is faith-specific, brand loyalty should predominate. A sale on Pepsi will readily cause consumers to switch from
Coke, but switching from Methodism to Hinduism means the costly loss of
religious capital. As a result, most children follow in the religious footsteps
of their parents, relatively few people switch faiths, and those who do tend
to switch at earlier ages before the amount of religious capital at risk
becomes too large (Iannaccone, 1992).
2 The supply of religion
While some religion can be self-supplied, most is produced and marketed
through non-profit firms such as churches, synagogues or mosques. The
non-profit form largely results from information imperfections. In markets
for which information is easily and cheaply obtained, consumers can make
eﬃcient and informed choices. For search goods (those for which quality
can be determined easily by inspection), the process is trivial. Selecting the
correct-sized mailing envelope or the dress shirt of appropriate colour

272

Economics uncut

requires little stress. However, experience goods cannot be easily judged
until after they have been consumed. For these we seek advice from others.
We consult consumer reports and friends to assess which autos are likely to
be most durable and which wrinkle-free slacks are truly wrinkle-free.
Information problems are most severe for credence goods, that is, goods
whose quality we cannot objectively measure even after use. For example,
suppose we take a pill our doctor prescribes and then we recover. The recovery might have been caused by the pill, but we might have recovered anyway.
We cannot be sure. Crediting recovery to the pill requires a subjective leap
of faith. Religion is the consummate credence good. We may never know,
at least in this life, if we picked a winner. Credence goods expose consumers
to increased risk and vulnerability. With no eﬀective method of judging
product quality for themselves, consumers must take some supplier claims
on faith.
Can suppliers be trusted? For-profit firms might be especially prone to
exploiting their information advantage with bogus claims of product
quality. Non-profit firms have less of a financial stake in our choices and,
therefore, might be less likely to deceive or to skimp on quality. As a result,
non-profit firms are more believable and easier to trust. In addition, managers might self-select. Those primarily interested in monetary rewards
might locate in proprietary firms, while those with more altruistic bents
plausibly could gravitate toward non-profit employment. Given that nonprofit firms often open their accounting ledgers to the public and even allow
consumers a voice in how their organizations are managed, it is not surprising that they dominate markets for credence goods such as education,
health care and religion (Rose-Ackerman, 1996).
Despite their non-profit nature, religious organizations and leaders are
certainly not immune from financial incentives. Numerous religious entrepreneurs have parlayed evangelistic eﬀorts into substantial personal fortunes and recent economic studies suggest that prior to the competitive
entry during the Protestant Reformation, Catholic oﬃcials routinely ‘controlled and manipulated doctrine and rules in order to increase its revenues’
(Ekelund et al., 1996).
Although money does matter, many modern studies model religious
organizations as clubs that produce quasi-public goods that are accessible
to members, but not to outsiders or non-members. Their products include
worship services, religious education opportunities, counselling and a wide
array of social activities. Many of these are produced within the organizations by participating members, but others are produced by agents such as
clergy and other professional staﬀ. In club models, clergy are treated as
agents of the members much as corporate managers are agents of the
stockholders.

Economics of religion

273

2.1 External eﬀects and free-riding
Along with traditional principal–agent issues, clubs and religious organizations are plagued by a variety of external eﬀect and free-rider problems.
Because church members produce and consume many products jointly, their
utilities are interdependent. Participation, both in terms of volunteer labour
and contributions, creates positive spillover or external eﬀects. The more
one member contributes, the more others have available to them. The benefit
received by one member depends on the quantity of participation by others.
To modify our earlier model, a consumer’s utility in period t (Ut) is given by:
Ut U(St, Rt, Qt)

(9.6)

where Qt represents the quantity and quality of religious production by
other members of the group in period t.
If individual members ignore their impact on others, ineﬃciencies result.
As with other club goods, most religious goods are non-excludable. Once
produced, they are available to all members at no additional cost. Because
religious goods produced by one member are available to all, the marginal
benefit of religious output to the group exceeds that received by the individual producing member (see Figure 9.3).
The marginal benefits (MB) to the group cover marginal cost (MC) all
the way to output R1t. But members who care only about themselves will
rationally balance their personal benefits with cost and produce only R0t,
Costs/
Benefits

MB to group
MB to individual
members
R0t

Figure 9.3

R1t

Quantity of religious
production in period t

Spillover benefits and eﬃciency in religion

274

Economics uncut

an ineﬃciently small amount from the perspective of the group. Of course,
selfless individuals who care only about the greater good would consider
the spillover benefits and produce larger quantities. Many members do voluntarily contribute generously in terms of both time and money. But many
do not. Faced with the opportunity to enjoy the benefits of religious output
whether or not they pay, many choose not to pay. They choose instead to
free-ride on the contributions of others.
Free-riders create at least two problems for a religious congregation or
group. First, free-riders can be expensive; they raise costs for the more
committed members. Free-riders might avoid helping to supply religious
outputs, but they are seldom shy about demanding them and often complain loudly when such goods and services are slow in coming. Clerics of
every stripe can identify free-riders who rarely attend services or contribute
funds, yet place disproportionate demands on their time and energy.
Second, free-riders can demoralize other members of the congregation.
Worship services are chapters in a congregation’s continuing and collective
journey of faith. Occasional and uncommitted participants do not appreciate where the congregation has been or where it is going. They do not sing
with enthusiasm because the melodies and responses are unfamiliar. They
do not pray with conviction because their own commitment is marginal.
They do not seek out and greet new visitors because they cannot identify
which people are visitors. Committed believers add excitement to worship;
free-riders often reduce it.
Of course, members can free-ride only on religious benefits provided in
the ‘here and now’. As David Friedman has commented privately, when it
comes to matters of salvation, ‘God can monitor the free-riders’.
2.2 Sacrifice and stigma
Secular clubs alleviate free-riding through membership screening and
pricing. Committees scrutinize applicants and deny membership to those
deemed unlikely to act in the best interest of the group. Those passing this
initial test are then charged explicit fees. Just as governments levy compulsory taxes that force citizens to pay for eﬃcient quantities of public goods,
secular clubs levy initiation fees and annual dues to finance their programmes. Potential free-riders who are unwilling to pay and contribute are
excluded and lose access. In the face of scarcity, we cannot allow free access
to all goods and services. We must ration some potential consumers out of
the market. Most firms ration access to their products by pricing the products themselves. Secular clubs ration access to their products by pricing
membership in the club.
Although some religious groups do restrict entry in various ways, most
consider screening committees and annual dues theologically oﬀensive.

Economics of religion

275

They need more creative ways to weed out uncommitted members likely to
free-ride. Unwilling to screen potential members through explicit prices,
they use implicit prices instead. These implicit prices often are charged in
terms of sacrifice and stigma (Iannaccone, 1992).
Religious membership may not require a monetary sacrifice, but it often
requires significant non-monetary ones. Hare Krishnas are expected to
shave their heads and don saﬀron robes. Jehovah’s Witnesses are expected
to engage in door-to-door evangelism and forego blood transfusions. The
Amish must do without cars and electricity. When religious groups impose
prices this high, free-riders vanish quickly.
Sacrifice and stigma serve a second function as well. Not only do they
keep free-riders out, they raise the commitment of those who remain. For
example, Amish men find it diﬃcult to interact easily in the secular world.
Their beards, hats and drab black clothing set them apart and make them
curiosities subject to the bemused stares of others. As a result, they
avoid such interactions when possible. They prefer to remain in their own
communities among those of similar beliefs. In other words, secular and
sacred activities are substitute goods. By imposing heavy demands on
their members, religious groups eﬀectively raise the price to their members
of interacting with people outside their own faith. By increasing the price
of secular substitutes, they raise the demand for religious activities.
Since the equilibrium output in a free-riding world is ineﬃciently low (see
Figure 9.3), this increased demand for religious activity drives output closer
to the optimal level for the group. As Iannaccone puts it: ‘Distinctive diet,
dress, grooming and social customs constrain and often stigmatize
members, making participation in alternative activities more costly.
Potential members are forced to choose: participate fully or not at all’
(Iannaccone, 1992).
In this context, imposing harsh requirements that isolate members from
other societal groups can be a rational method to discourage free-riding
and build a cohesive community. High-cost or high-tension religious cults
and sects are often accused of brainwashing their members. But there is
considerable evidence that people are well aware of the costs and benefits
in joining these groups and that their memberships were not coerced
(Richardson, 1991; Iannaccone et al., 1998).
The high-cost strategy can work. Such groups enjoy significantly higher
levels of giving and participation per member than do other religious
groups. Iannaccone (1992) divided denominations into four categories:
Most Churchlike, Churchlike, Sectlike and Sect. Those denominations
categorized as Most Churchlike impose the lowest demands on members
while Sects demand the most. For a sample of Northern Californian
church members, Iannaccone found that, compared with members of the

276

Economics uncut

Table 9.1 Denominational characteristics: 1963 Northern Californian
data

Size of congregation
(No. of members)
Household income
($ per year)
Respondent education
(years)
Sunday attendance
(services/year)
Evening attendance
(meetings/year)
Church contributions
(% of yearly income)
Church friends (No. out
of 5 closest friends)
Non-church memberships
(number claimed)

Most
Churchlike

Churchlike

Sectlike

Sects

939

787

335

173

$10 140

$9 435

$8 399

$6 944

14.5

13.9

12.3

12.5

33.8

38.4

44.2

49.1

34.5

44.9

69.0

96.8

2.64

3.36

4.64

8.43

1.32

1.51

1.80

3.15

3.59

3.07

2.13

1.72

Note: Using two-tailed tests, diﬀerences between the mean values for Most Churchlike and
Sects are statistically significant for all reported variables.
Source: Raw data are from 1963 Northern Californian Church Member Study. Table is
adapted from Iannaccone (1992).

Most Churchlike denominations, members of Sects attended more Sunday
services per year (49.5 compared with 34.5) and more evening events per
year (96.8 compared with 34.5). They also contributed significantly larger
percentages of their incomes (8.43 per cent compared with 2.64 per cent)
and were less likely to belong to other organizations (1.72 compared with
3.59). More complete results are listed in Table 9.1. Similar diﬀerences can
be found using national data and these results persist in regressions that
control for diﬀerences in such background characteristics as age, sex, education, income, marital status and geography.
Iannaccone’s sacrifice and stigma thesis also explains an intriguing
paradox of church growth. Conventional wisdom would predict that
making membership easy would increase a group’s rate of growth.
However, those faiths that demand the most sacrifice and stigma are
among the fastest growing. Mormons have enjoyed rapid growth, yet place
heavy demands upon members in terms of expected monetary contributions and evangelistic service. On the other hand, mainstream Protestant

Economics of religion

277

denominations place minimal demands upon members and have been
losing members and market share in recent decades. Membership is cheap,
but so is free-riding. According to Berman (2003), a similar model can be
used to explain the success of radical and high-cost religious militias such
as the Hamas and Taliban.
Sacrifice and stigma are not foolproof strategies for enhancing church
growth. Not all potential members will be persuaded, and not all current
members will stay. As economists should expect, the strictest and most
demanding religious groups do tend to attract those with the least attractive secular options: the less educated, the less well-to-do, the young,
women and minorities (see Table 9.1). These groups also experience cyclical swings. When the economy drops into a recession and unemployment
rates rise, membership in strict religious groups also tends to rise. When the
economy recovers and job opportunities improve, membership in these
same groups falls (Iannaccone, 1992).
2.3 Optimal firm size
Large congregations are especially vulnerable to free-rider woes. The larger
the group, the easier it is to hide. Shirking one’s proportionate responsibilities is tough in a group of two, but relatively easy in a group of 1000.
A congregation of 1000 members might be lucky to find 300 at worship on
a particular Sunday. On the other hand, being large does bring benefits.
Multiple staﬀ members support a diversified menu of programmes and
interest groups that smaller congregations can only imagine. There are cost
advantages as well. Attendance at liturgies, church school classes, youth
groups and Bible studies rarely approaches physical capacity. Stuﬃng in
extra participants almost always lowers the cost per person. Even when
capacity is strained, constructing larger facilities often generates economies
of scale. A building large enough to handle 1000 parishioners is not likely
to be twice as expensive as one built to accommodate 500. Statistical studies
document both eﬀects. Within many Christian religious traditions, larger
congregations attract fewer contributions per member, exactly what we
would expect if proportionately more free-riders are present (Stonebraker,
1993; Zaleski and Zech, 1994). However, larger congregations also enjoy
scale economies that lower operating costs per member.
For some mainline Protestant churches the size advantage dominates. As
congregational size rises, average operating costs fall faster than average revenues. In these cases, the larger congregations are able to allocate disproportionately more funds for benevolence and mission projects that aid
ministries beyond the congregation itself (Stonebraker, 1993). However,
patterns in stricter denominations might diﬀer. High-cost religious groups
and sects that rely on sacrifice and stigma are likely to find smaller sizes more

278

Economics uncut

eﬃcient. The larger the group, the more diﬃcult and expensive it becomes
to monitor and enforce behavioural standards. Theory predicts that mainline groups that impose few standards and have no significant monitoring
costs will find larger sizes more eﬃcient while high-cost sects will find
smaller sizes more eﬃcient. Interestingly, these are precisely the patterns
we find. High-cost religious groups tend to be smaller than low-cost groups
(see Table 9.1).
3 Religious risk
Religious suppliers must also confront risk. To Pascal, choosing to believe
in God was elementary logic. If we choose not to believe, we risk eternal
damnation. If we choose to believe, we risk nothing. If God does exist,
believers will be rewarded. If God does not exist, believers end up in the
same boat as non-believers.
But things are not so simple. Which God? Which religion? The prerequisites for salvation vary markedly among religions. Diﬀerent faiths
demand diﬀerent and often mutually exclusive beliefs and actions. The
oﬃcial dogma of the Shiite Muslim and the Southern Baptist each condemns the other to an eternal hell. There is no safe belief.
Suppliers try to convince potential adherents that theirs is the one true
faith. Consumers are often sceptical but, since religion is a credence good,
claims cannot easily be disproved, at least in this life. Cautious buyers will
seek strategies that might cut the likelihood of falling prey to religious
fraud. For example, they prefer testimonials from trusted friends rather
than from strangers, and testimonials from those with no financial interest
at stake. Personal invitations from trusted friends and lay testimonials have
proven to be successful evangelistic tools and the fact that religious professionals often earn considerably less than others with similar education and
training probably contributes to their credibility as well. Nonetheless, risk
remains.
In financial markets we combat risk through diversification; we spread
investments over a wide variety of corporate stocks and bonds. Might religious diversification bring similar rewards? Religious diversification is not
a new idea. Ancient Rome oﬀered many potential and competing deities.
Venus, the goddess of love, handled matters of the heart while Ceres, the
goddess of agriculture, took care of crops and Mars, the god of war,
handled military matters:
All these existed, side by side, in an atmosphere of mutual toleration. Adherence
to one did not preclude involvement in another. Citizens could, and many
did, participate in multiple cults – sacrificing to several gods, worshipping
the Emperor, and undertaking a variety of diﬀerent rites and initiations.
(Iannaccone, 1995)

Economics of religion

279

More modern examples exist as well. Indians might worship a variety of
Hindu gods, while Japanese families might claim both Buddhism and
Shintoism, and yet rent a Catholic church for their daughter’s wedding
(Ono, 1998). New Age proponents mix and match tarot readings, meditation, astrological charts, trips to a mystical energy vortex, crystals, pyramids and witches’ covens.
Despite its potential, religious diversification remains the exception in
Western countries. Most adherents choose a specific brand of faith and
consume it exclusively. Of course, no firm wants its customers to diversify.
Ford prefers that its customers not diversify with General Motors and
Pepsi prefers we not diversify with Coke. But secular firms rarely command
the fierce customer devotion enjoyed by churches. What enables a particular religious faith to command the sole allegiance of its worshippers?
3.1 Strategies for exclusivity
Secular firms wanting exclusive consumer allegiance first must diversify
themselves. They become full-line merchandisers and market themselves
for one-stop shopping. Ford produces a full line of vehicles and Pepsi a full
line of soft drinks. So it is in the sacred world. Because gods in the Roman
panoply specialized, none could command exclusive devotion. But the
Judeo-Christian God is an all-encompassing deity. Churches and synagogues in this tradition provide a full range of cradle-to-grave services; services that include a comprehensive theological system and a broad network
of fellow members to meet both social and emotional needs.
Comprehensive services are necessary, but not suﬃcient. Consumers will
grant exclusive allegiance only if firms make it eﬃcient to do so. Developing
exclusive technology is one common corporate tactic. Enthusiasts lured by
Sony’s PlayStation games had better buy the PlayStation game platform as
well; Nintendo’s are not compatible. Similar ploys succeed in religion.
While the Greco-Roman gods tolerated multiple allegiances, the JudeoChristian tradition demands exclusivity. The Ten Commandments are
quite explicit. ‘You shall have no other gods before me’, and ‘You shall not
make yourself a graven image . . . you shall not bow down to them or serve
them; for I the Lord your God am a jealous God’ (Exodus 20:3–4).
Polytheism and Christianity are not compatible.
The first two commandments forbid Christians to shop for alternative
gods, but they place no restrictions on sampling diﬀerent traditions within
Christianity. They explain why Christians do not diversify into Hinduism,
but not why Presbyterians fail to diversify into Catholicism. Perhaps such
diversification is costly. Strategies that reward allegiance and/or penalize
defection might build loyalty. Each November many supermarkets in the
USA oﬀer a free Thanksgiving turkey to those who spend at least $250 in

280

Economics uncut

the preceding month. Consumers that patronize one store exclusively can
usually make the $250 level. But those that diversify and split their purchases among several stores may not. Airlines grab allegiance in the same
way with frequent-flier programmes. Consumers that patronize a single
carrier exclusively more easily can run up enough miles to win free trips.
Competing Christian groups pursue similar strategies. Christianity is
largely a communal faith. Worship, sacred rites and other services are produced and consumed collectively within a congregation of believers.
Consumers not spliced into the intricate social networks within these congregations derive few benefits. But these network connections seldom arise
naturally; they must be carefully cultivated over time. And those who split
their shopping among several congregations are unlikely to be successful.
To drive home the point, many Christian faiths forcibly evict members who
are intent on comparison shopping. The Catholic who strays into
Mormonism risks ex-communication and separation from the family of
faith. Mormons who regularly seek absolution from Catholic priests risk
analogous penalties.
4 Religious markets
The success of a religious group might be aﬀected by the types of demands
it places upon members and by its ability to command the exclusive loyalty
of those members. However, just as in the secular world, success will also
hinge upon the quality of the product delivered. While religious groups
certainly provide a variety of social products, their main products are
doctrines. These doctrines vary significantly across religious traditions,
yet most share a common logical structure. They contain a series of assertions, typically improvable, about what specific outcomes will result from
what specific religious actions. While economists cannot prove that one
doctrine is ‘better’ than another, we can make predictions about the types
of doctrines that are most likely to succeed in a competitive marketplace
(Iannaccone, 2002b).
4.1 Religious products and market competition
First, only those doctrines that oﬀer positive expected net benefits can
expect to flourish. Consumers will quickly reject religions that promise negative payoﬀs. Second, successful doctrines will reserve the greatest net benefits for those who are the most faithful. If they oﬀer the same rewards to
all, the marginal benefit will be zero. Consumers will expend little or no
time and money pursuing religious activities if they promise no additional
gain. This creates problems for many liberal Protestant denominations.
According to Iannaccone, these groups often lose members to more conservative congregations ‘not because they oﬀer too much or too little, but

Economics of religion

281

because they cannot bring themselves to promise members substantially
more than they oﬀer non-members’ (Iannaccone, 2002b). Similarly, he
argues, ‘with little to oﬀer and less to withhold’, it is not surprising that
atheists have made little headway in religious markets (Iannaccone, 2002b).
Successful doctrines must be able to attract sellers just as they attract
buyers. We cannot buy what no one is willing to sell. The fact that priests
traditionally would claim a share of any sacrificial oﬀering certainly
might have aﬀected how vigorously they touted such oﬀerings, and the burgeoning New Age market for tarot readings, meditation lessons and channelling sessions has been helped by the variety of people willing to sell such
services.
While not all doctrines are created equal, those that persist show remarkable diversity. Religious markets in many countries, including those of the
USA, are characterized by many firms, diﬀerentiated products and easy
entry. Like restaurants and clothing manufacturers, religious groups
exhibit monopolistic competition.
Because consumers have diﬀerentiated food tastes, restaurants race to fill
every available spot in their product space. Some specialize in ethnic dishes,
others in seafood. Some oﬀer drive-through service while others oﬀer sixcourse, three-hour extravaganzas. Religious groups oﬀer a similar diversity.
We can find groups that oﬀer short worship services or long ones, high liturgies or none at all, conservative theologies or liberal ones. New entrants will
quickly fill unexploited market niches in religious product space. If
Pentacostalists are under-served in a local market, storefront suppliers
rapidly will emerge.
Although diversity and competition dominate many religious markets,
such pluralism generates very diﬀerent reactions. Some insist that pluralism erodes the overall demand for religiosity. They maintain that competing religious claims undercut the credibility of all. According to sociologist
Steve Bruce (1992), ‘Pluralism threatens the plausibility of religious belief
systems by exposing their human origin . . . A chosen religion is weaker
than a religion of fate because we are aware that we choose the gods rather
than the gods choosing us’. In economic terms, religious pluralism raises
consumer costs by increasing uncertainty and lowering the expected utility
of commitment (Hull and Bold, 1998). In this view, a single monolithic
faith will maximize religious faith and participation.
However, many economists disagree and suggest that pluralism might
increase religiosity. Pluralism allows religious groups to specialize and
diﬀerentiate their products to more accurately match up with diverse consumer tastes and preferences. Just as an ice cream parlour that oﬀers only
a single flavour will lose sales, a country that oﬀers only one brand of faith
will lose religiosity.

282

Economics uncut

a: Monopoly church
liberal theology

conservative theology

X
A1

X
A

b: Competitive entry
liberal theology

conservative theology
2X

2X
B1

Figure 9.4

Theological product space

Economists Pedro Pita Barros and Nuno Garoupa (2002) have recently
used a spatial location model to analyse the eﬀect of religious competition.
Using a simplified variant of their approach, imagine a ‘theological product
space’ in which consumer tastes are uniformly distributed along a spectrum
from conservative to liberal views. Next, assume that consumers will join a
church only if the church’s theology is reasonably consistent with the consumers’ own tastes. For example, suppose consumers will join only if the
church is located within X units of the consumers’ own positions in theological product space. In Figure 9.4a, if a conservative monopoly church
locates at point A, it will attract members located between A1 and A2 (A
X). If the church wants to attract members with more liberal tastes it can
do so by moving to the theological left. But this will cause its more conservative members to drop out. Anything a single, monopoly faith does to
please one market segment will alienate another.
If new churches enter the mix, they can locate in areas not served by the
incumbent. For example, if a new church enters at point B (see Figure 9.4b),
consumers located between B1 and B2 (BX) join the fold. As new faiths
emerge through time, they capture the minds and hearts of the heretofore
heathen. Pluralism boosts religion participation.
Competition might also force churches to produce more eﬃciently.
Clergy in a monopoly religion have less need to recruit members, less need
to preach and teach eﬀectively, and less need to respond and tailor their
product to meet needs of their members. A lack of competitive pressure
means a lack of evangelistic zeal. Monopoly power is likely to sap the

Economics of religion

283

evangelical fervour of the dominant faith over time. Monopolies in the
sacred realm are just as prone to sloth as those in the secular world.
Pluralism proponents bolster their case by citing the example of the USA
where both religious participation and religious competition are unusually
high by international standards. Many statistical studies also support the
positive impact of pluralism. Using appropriate control variables, they find
significant positive correlations between measures of pluralism and religiosity (Iannaccone, 1991). However, many of the studies are plagued by
data and statistical shortcomings and some do reach opposite conclusions
(Olson and Hardaway, 1999; Voas et al., 2002).
The impact of market concentration might also depend upon its cause.
For example, religious participation in Ireland, Poland and the Canadian
province of Quebec has been remarkably high even though all are heavily
dominated by Catholicism. However, in these areas, the church often served
to galvanize opposition to outside political repression. Attending church
was viewed as much an indication of nationalism as of religiosity.
Interestingly, with the demise of the Soviet hegemony in Poland, religious
participation has begun to wane. The same decrease in religiosity has been
seen in Quebec as the French Canadians have increased their local political
power (Stark and Iannaccone, 1996).
4.2 Regulated religion
While the overall impact of pluralism remains an issue, economists generally
agree that artificial restrictions on religious pluralism, especially those
imposed by governments, do restrict participation (Chaves and Cann, 1992).
Religious suppliers are no keener on competition than their secular
counterparts. Adam Smith recognized this more than 200 years ago. He
knew that if clergy are confronted by competitive pressure, ‘their exertion,
their zeal and industry’ must rise (Smith, 1776). He knew that churches, like
other producers, would fight to insulate themselves from the rigours of such
competitive pressure; that, barring natural economic entry barriers, they
would seek refuge in protective government regulation. They have.
Religions provided divine endorsements and legitimacy to kings who, in
return, granted state-endorsed monopoly privileges. The religion of the
king became the national, established religion, and alternative faiths were
repressed through both legislation and violence. As Iannaccone puts it,
‘From Old Testament Israel to contemporary Iran, religious uniformity has
arrived on the edge of the sword, and only the sword has suﬃced to maintain it’ (Iannaccone, 1991).
Although the American colonies were settled by refugees seeking to avoid
monopoly church repression, they were intolerant themselves. They quickly
set up government-backed religions of their own and used state power to

284

Economics uncut

erect entry barriers against new, upstart faiths. Congregational churches
received revenues from colonial New England taxes, and itinerant ministers
were banned from preaching without the prior approval of local clergy.
Nonetheless, the constitutional separation of church and state has allowed
the USA to escape the formation of a federally established church. This laid
the groundwork for religious diversity that has been responsible for one
of the most competitive religious markets in the world.
In many European countries state-established and state-protected
churches endure. The German government still subsidizes Lutheran
congregations through state-imposed membership taxes and several
Scandinavian governments continue to pay Lutheran pastors as civil servants. Despite woeful attendance, the established churches of Europe
display considerably less evangelistic zeal than their American counterparts:
Europe’s monopoly churches are, like its monopoly utilities, directed towards the
needs of producers rather than consumers. Producer-run businesses favour technical excellence over cost-eﬀectiveness. Compare European cathedrals with the
no-frills premises of the new fundamentalists . . . Unconstrained by market
forces, European churches devote energy to doctrinal niceties [that] do not
interest most users. The seminaries of Southern Baptists are concerned with
practical skills, just as Wal-Mart’s training is directed towards minding the store,
not the financial engineering of a Harvard MBA. (Kay, 2003)

The message seems clear: government-imposed entry barriers restrict competition and output.
Deregulation set the stage for several major religious shifts in the USA.
The first, following the 1791 constitutional prohibition on enacting federal
laws that establish an oﬃcial religion, unleashed a torrent of entry in the
early nineteenth century that increased consumption of religious services
significantly (Olds, 1994). Changes in immigration policies helped lead to
changes in the population that spurred the growth of specific religious
groups. Immigrants from Ireland and Italy strengthened the competitive
presence of Catholics. The relaxation of controls on Asian immigrants in
the latter half of the twentieth century brought increased numbers of
Buddhists and Hindus into the country. Perhaps more importantly, the new
laws also enabled the entry of guru-entrepreneurs who quickly showed the
ability to market their religious practices successfully among native-born
Americans as well as among Asian immigrants (Iannaccone et al., 1997).
Changes in Federal Communication Commission (FCC) broadcast
regulations also proved to be important. Prior to the 1960s, FCC policies pushed broadcasters to grant free airtime to religious programming.
Well-heeled station owners, in turn, parcelled out the free time to the
denominations to which they themselves belonged: liberal, mainline

Economics of religion

285

Protestant denominations. However, when new policies no longer gave
preferential treatment to broadcasters oﬀering free airtime to religious
groups, the major stations and networks quickly began shifting to paid religious programming instead. The mainline denominations hesitated at
paying for access to a heretofore free resource, but conservative faiths did
not. Having been denied free airtime, more conservative evangelists had
been forced to buy airtime on local stations for years. They jumped quickly
into the broadcasting breach. They knew the market and they knew how to
cajole their audience into funding their broadcasts. Within a short period
of time fundamental Christians began to dominate the American airwaves
(Iannaccone et al., 1997).
Deregulation has spawned similar changes in other countries. After the
Allied Occupation disestablished the Shinto religion and repealed religious
regulations, Japan experienced a dramatic religious revival (Iannaccone
et al., 1997). Deregulation of Catholic state-monopolies has led to the rapid
growth of Protestant churches in Latin America, and the demise of Soviet
control created comparable increases in Hungary and other Eastern
European nations (Froese, 2001).
4.3 Competition and doctrine
In addition to its impact on religious participation, competitive pressure
can also influence the nature of the product or doctrine being marketed.
Barros and Garoupa (2002) consider a case in which consumers that were
originally forced to join a monopoly church are suddenly given the option
of not joining. Those most inclined to leave the church are likely to be those
most opposed to religious strictness. As a result, an established church concerned about its power is likely to liberalize to stem the potential outflow
of members. If tastes and preferences change in a way that makes the population less religious, we should observe the same liberalizing eﬀect as the
established church tries to hold onto its market.
If new faiths are allowed to enter the market, the established church
should move in the direction of its new competition. Entry by conservative
sects is likely to pull the established church to the right to cut oﬀ support
for the new threat. Similarly, competition from more liberal faiths should
pull the established church to the left. Barros and Garoupa use this theory
to explain why the dominant churches in the UK and Scandinavia (where
the main competitors are fairly liberal) are less strict than are the dominant
faiths in Southern Europe and Brazil where the competition comes primarily from the right.
Competitive pressures might also explain the Catholic Schism of 1054.
Barros and Garoupa explain that the Catholic Church enjoyed far more
monopoly power in the West than in the East. Catholic rulers controlled

286

Economics uncut

Western Europe and acknowledged the spiritual leadership of Rome. With
no major challengers, the Western Catholic Church could aﬀord to market
a conservative doctrine. However, the Eastern Catholic Church faced significant competitive pressure from Muslims and local sects, and found itself
being pulled in very diﬀerent theological directions; directions that ultimately caused the split.
Competition might also explain why church oﬃcials have increasingly
soft-pedalled the concept of Hell. With a near monopoly, the Catholic
Church could aﬀord to deliver a fire and brimstone message. Toe the religious line, or burn forever. According to Barros and Garoupa, increased
competition, both from Protestants after the Reformation and from the
‘non-church’ during the Enlightenment, forced oﬃcial doctrine to become
more user-friendly. In addition, a doctrine that condemns non-Catholics
to Hell is more viable in a world populated almost entirely by Catholics.
The doctrine becomes harder to stomach in a world of religious pluralism
in which many of one’s best friends and relatives are non-Catholics
(Iannaccone, 2002b).
5 Conclusion
Where do we stand? As we have seen throughout this book, economics is a
science of choice in which people rationally consider the expected costs and
benefits of their actions. It assumes that if behaviour diﬀers across individuals or across time, it must be because perceived costs and benefits diﬀer
across individuals or across time. Such notions generate little controversy
in secular enterprises and microeconomic theories consistently predict realworld behaviour with uncanny accuracy. The theory works.
Does it work equally well in the sacred world? Can we model religious
behaviour with the same tools we use to analyse General Motors and CocaCola? For many years, social scientists considered religious choice antithetical to rationality. Religious behaviour was deemed illogical and
‘scientific’ analyses were often couched in terms of social and psychological pathologies.
Recent scholarship suggests otherwise. Decisions about how much and
what types of religious activity to demand and supply appear remarkably
consistent with traditional economic maximization behaviour. Religious
choices closely mirror those made in more secular pursuits. Negativelysloped consumer demand curves for religion shift in response to factors
such as changes in tastes or price and availability of substitute and complementary goods. Suppliers of religion compete for members, search for
strategies to lock in customer loyalty, and welcome government regulations
that restrict competition. Religious markets evolve much like secular ones.
Prices, outputs and product mixes respond to shifts in consumer tastes,

Economics of religion

287

technologies and competitive pressures. In the sacred world, just as in the
secular world, the theory works.
References
Azzi, C. and R.G. Ehrenberg (1975), ‘Household Allocation of Time and Church
Attendance’, Journal of Political Economy, 83 (1), 27–56.
Barros, P.P. and N. Garoupa (2002), ‘An Economic Theory of Church Strictness’, Economic
Journal, 112 (3), 559–76.
Becker, G.S. (1965), ‘A Theory of the Allocation of Time’, Economic Journal, 75, 493–517.
Berman, E. (2000), ‘Sect, Subsidy and Sacrifice: An Economist’s View of Ultra-Orthodox
Jews’, Quarterly Journal of Economics, 115 (3), 905–53.
Berman, E. (2003), ‘Hamas, Taliban and the Jewish Underground: An Economist’s View
of Radical Religious Militias’, National Bureau of Economic Research Working Paper
No. w10004.
Bruce, S. (1992), ‘Pluralism and Religious Vitality’, in Steve Bruce (ed.), Religion and
Modernization, Oxford: Clarendon Press, pp. 170–94.
Chaves, M. and D.E. Cann (1992), ‘Regulation, Pluralism and Religious Market Structure’,
Rationality and Society, 4, 272–90.
Ekelund, Robert B., Robert F. Hebert and Robert D. Tollison (1996), Sacred Trust: The
Medieval Church as an Economic Firm, New York: Oxford University Press.
Ellison, C.G. (1998), ‘Religion, Health and Well-Being’, Journal for the Scientific Study of
Religion, 37 (4), 692–4.
Finke, Roger and Rodney Starke (1992), The Churching of America 1776–1990: Winners and
Losers in our Religious Economy, New Brunswick: Rutgers University Press.
Froese, P. (2001), ‘Hungary for Religion: A Supply-Side Interpretation of the Hungarian
Religious Revival’, Journal for the Scientific Study of Religion, 40 (2), 251–68.
Gallup International Millennium Survey (1999), (http://www.gallup-international.com/).
Heineck, G. (2001), ‘The Determinants of Church Attendance and Religious Human Capital
in Germany: Evidence From Panel Data’, Discussion Papers of DIW Berlin, paper no. 263,
DIW Berlin, German Institute for Economic Research.
Hull, B.B. and F. Bold (1998), ‘Product Variety in Religious Markets’, Review of Social
Economy, 56 (1), 1–19.
Iannaccone, L.R. (1990), ‘Religious Practice: A Human Capital Approach’, Journal for the
Scientific Study of Religion, 29 (3), 297–314.
Iannaccone, L.R. (1991), ‘The Consequences of Religious Market Regulation’, Rationality
and Society, 3 (2), 156–77.
Iannaccone, L.R. (1992), ‘Sacrifice and Stigma: Reducing Free-Riding in Cults, Communes,
and Other Collectives’, Journal of Political Economy, 100 (2), 271–97.
Iannaccone, L.R. (1995), ‘Risk, Rationality, and Religious Portfolios’, Economic Inquiry,
33 (2), 285–96.
Iannaccone, L.R. (1998), ‘Introduction to the Economics of Religion’, Journal of Economic
Literature, 36 (3), 1465–96.
Iannaccone, L.R. (2002a), ‘Looking Backward: A Cross-National Study of Religious Trends’,
paper presented at November 2002 Meetings of the Society for the Scientific Study of
Religion.
Iannaccone, L.R. (2002b), ‘From Scarcity to Spirituality: The Consequences of Religious
Entrepreneurship’, (http://icg.harvard.edu/ec2390d/Papers/Papers_Fall_2002/October_
8_2002_Laurence_Iannaccone_Religious_Entrepreneurship.pdf).
Iannaccone, L.R., R. Finke and R. Stark (1997), ‘Deregulating Religion: The Economics of
Church and State’, Economic Inquiry, 35, 350–64.
Iannaccone, L.R., R. Stark and R. Finke (1998), ‘Rationality and the “Religious Mind” ’,
Economic Inquiry, 36, 373–89.
Kay, J. (2003), ‘It Pays Churches to have Faith in Markets’, Financial Times, 10 April,
p. 13.

288

Economics uncut

Lehrer, E.L. and C.U. Chiswick (1993), ‘Religion as a Determinant of Marital Stability’,
Demography, 30 (3), 385–404.
Miller, A.S. (1995), ‘A Rational Choice Model of Religious Behavior in Japan’, Journal for the
Scientific Study of Religion, 34 (2), 234–44.
Murray, J.E. (1995), ‘Determinants of Membership Levels and Duration in a Shaker
Commune, 1780–1880’, Journal for the Scientific Study of Religion, 34 (1), 35–49.
Olds, K. (1994), ‘Privatizing the Church: Disestablishment in Connecticut and
Massachusetts’, Journal of Political Economy, 102 (2), 277–97.
Olson, D.V.A. and C.K. Hardaway (1999), ‘Religious Pluralism and Aﬃliation Among
Canadian Counties and Cities’, Journal for the Scientific Study of Religion, 38 (4), 490–508.
Ono, Y. (1998), ‘Here Comes the Bride All Dressed in a Kimono’, Wall Street Journal, 9 June
A25.
Pascal, Blaise (1660), Pensees.
Richardson, J.T. (1991), ‘Cult/Brainwashing Cases and Freedom of Religion’, Journal of
Church and State, 33 (1), 55–74.
Rose-Ackerman, S. (1996), ‘Altruism, Nonprofits, and Economic Theory’, Journal of
Economic Literature, 34 (2), 701–28.
Sawkins, J.W., P.T. Seaman and H.C.S. Williams (1997), ‘Church Attendance in Great Britain:
An Ordered Logit Approach’, Applied Economics, 29 (2), 125–35.
Scitovsky, Tibor (1992), The Joyless Economy (revised edition), New York: Oxford University
Press.
Smith, Adam (1776), An Inquiry into the Nature and Causes of the Wealth of Nations, reprinted
by Modern Library, New York, 1965.
Stark, Rodney and William S. Bainbridge (1987), A Theory of Religion, Bern: Peter Lang.
Stark, R. and L.R. Iannaccone (1994), ‘A Supply-Side Reinterpretation of the
“Secularization” of Europe’, Journal for the Scientific Study of Religion, 33 (3), 230–52.
Stark, R. and L.R. Iannaccone (1996), ‘Response to Lechner: Recent Religious Declines in
Quebec, Poland, and the Netherlands: A Theory Vindicated’, Journal for the Scientific
Study of Religion, 35 (3), 265–71.
Stonebraker, R.J. (1993), ‘Optimal Church Size: The Bigger the Better?’, Journal for the
Scientific Study of Religion, 32 (3), 231–41.
Sullivan, D.H. (1985), ‘Simultaneous Determination of Church Contributions and Church
Attendance’, Economic Inquiry, 23 (2), 309–20.
Voas, D., D.V.A. Olson and A. Crockett (2002), ‘Religious Pluralism and Participation: Why
Previous Research is Wrong’, American Sociological Review, 67, 212–30.
Zaleski, P.A. and C.E. Zech (1994), ‘Determinants of Religious Giving in Urban Presbyterian
Congregations’, Review of Religious Research, 36 (2), 197–206.

PART IV
CONCEPTION AND
REJECTION

10 Economics of assisted reproduction
Sherrie A. Kossoudji

Typically, when two people decide to become parents, they procreate by
copulation and produce a child. What do people do if, for some reason,
they cannot produce their own children but want to be parents? Until
recently, if a couple wanted to have children, but was unable to do so
because of fertility problems, there were limited options. They could adopt
a child to become the legal parents of a child who had no legal parents, have
a child who was the progeny of only one of them, or could become foster
parents, the temporary (and paid) caregivers of a child who still had legal
parents. Such activities are historically common and were often market
exchanges. Zelizer (1985), noting the importance of the work that foster
children provided for their families, claims that, ‘the legitimacy of child
labor was essential to early nineteenth-century substitute care arrangements’. In fact, child auctions (for fostering) were not prohibited in Sweden
until 1918 (Lundberg, 2000). Zelizer also argues that the changing social
value of children rendered them as, ‘exclusively emotional and moral
assets’ who were transformed from economic assets into ‘priceless’ children
in the early twentieth century. Children may have become recommodified:
at the beginning of the twenty-first century, excess demand for children to
adopt in the United States has led to an international adoption market that
has greatly increased the supply of adoptable children.
A child is created when an ovum is fertilized by a sperm and together
they become an embryo. The embryo must reside in a womb for approximately nine months to develop properly into a live baby. Recent developments in biotechnology have expanded the number of ways in which any
of these steps can happen. Newer markets are now extensive as the ability
to extract, process, freeze, transport and implant eggs, sperm and embryos
becomes commonplace. Simple sperm or egg ‘donation’ allows a couple
with one fertile partner to acquire the raw genetic material from someone
of the opposite sex.1 Artificial insemination by donor (DI) and fertilization of the ovum in vitro (IVF) are a collection of non-invasive to moderately invasive techniques to fertilize a woman’s ovum with a man’s sperm.2
There are relatively few places where these techniques are illegal, although
‘donation’ of sperm or eggs for money is illegal in a number of places.
With or without these techniques, the embryo must be able to develop in
a womb.
291

292

Economics uncut

The market in surrogate motherhood consists of a contractual arrangement where a woman agrees to carry a baby to term for a couple and to give
up any parental rights she may have. The baby may or may not have any
genetic components of either member of the commissioning couple, or of
the surrogate mother.3 Surrogate motherhood, although it still exists and is
legal in some parts of the world, presented both legal and moral dilemmas
that, as we discuss later, nearly destroyed (and significantly reduced) its
market.
A final market possibility exists. Reproductive cloning for humans skips
the process of uniting egg and sperm and, while still an untested possibility,
has raised questions from the religious to the economic. Two sets of economics authors have considered the implications of cloning. Posner and
Posner (1999) suggest that, among other impacts, cloning may accelerate the
breakdown of traditional roles of men and women as marriage incentives
change and as women find fewer incentives to invest in skills that are complementary to men’s. Saint-Paul (2002) concludes that a Handmaid’s Tale
society will emerge under cloning, where there is, ‘a reproductive class at the
bottom of the distribution of ability, a productive class at intermediate
levels; and a replicated class at the very top of the ability distribution’.
All of these markets, with the exception of the market for clones, exist in
a number of countries. There is no consensus, though, on the legality of
procedures, the enforceability of contracts, or the acceptance of commercialization. Surrogacy is legal in Spain; accepted in the UK, but not its commercialization; and its procurement is illegal in Denmark. In the United
States most assisted reproduction falls under state rather than federal law,
meaning that there can be 50 diﬀerent legal environments for each kind of
assisted reproduction. For instance, commercial surrogacy is illegal in Utah
(where the law is currently being contested) but legal in neighbouring
California. Table 10.1 gives a brief breakdown of current law in a number
of European countries and in the United States.4
The economics of assisted reproduction is not well understood by economists because it is still a new and rarely studied phenomenon. For this
reason, this chapter will not present a standard literature review that summarizes what economists think about the subject. Instead, we will heuristically proceed with a series of economic explorations related to assisted
reproduction to gauge the scope of the issues and how economists might
contribute to them. In Section 1, we show that economists do discuss
the economics of the decision to have children by considering the demand
for children. The traditional literature relating to this work is discussed.
We briefly mention the issue of supply of children, which is not typically
modelled in economics because children are assumed to be self-produced.
Next, in Section 2, is an exploration of the development of markets in

Economics of assisted reproduction

293

Table 10.1 Assisted reproduction services in Europe and USA
(as at 2004)
Country

IVF

DI (sperm)

DI (egg)

DI (embryo)

Surrogacy

Austria
Germany
Denmark
France
Norway
Sweden
Netherlands
Spain
United Kingdom
Greece
Hungary
Italy
Poland
United States

Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yesd

Noa
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yesd

Noa
No
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yesd

Noa
No
No
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yesd

No
No
Nob
No
No
No
Yes
Yes
Yesc
Yes
Yes
Yes
Yes
Yesd

Notes:
a
All forms of assisted conception with donated gametes are prohibited with the exception of
semen donation in IVF.
b
Surrogacy in itself is not forbidden, only its procurement. Also, surrogacy contracts are not
forbidden but considered legally invalid.
c
Surrogacy is accepted but specific regulation prohibits commercialization.
d
Legal in at least some US states.
Source: http://www.gov.gg/childrenboard/pdf/AssistedReproduction.pdf.

assisted reproduction. Even though there is a large market in assisted reproduction services (which uses the genetic components of the couple), there
will be no analysis of this medical service market in the chapter.5 After a brief
explanation of the size of the market in assisted reproduction in Section 3,
the discussion in Section 4 centres on economic issues in three markets that
currently exist: the markets for eggs, sperm and surrogacy. We draw out the
issues of supply and demand in these markets, discuss incentives, excess
demand, supply price, reservation price, risk and return, property rights and
altruism. Section 5 provides a discussion of the ethics of markets in assisted
reproduction. Section 6 contains some concluding thoughts.
1

Microeconomics of demand for and supply of children (before
reproductive markets)
In An Essay on the Principle of Population, Thomas Malthus (1798) was
concerned about population change but often used individual incentives to

294

Economics uncut

explain the demand for children: ‘The labourer who earns eighteen pence a
day and lives with some degree of comfort as a single man, will hesitate a
little before he divides that pittance among four or five, which seems to be
but just suﬃcient for one.’ Economists have since explored in a more
nuanced way how people make decisions about having children and the
role of economics in such decisions. In Chapter 5, we suggested two reasons
why people have children: the biological imperative and the enjoyment children provide. We noted that both factors look the same in a standard economic model because they increase an individual’s utility. Let us elaborate
on this idea.
In the basic microeconomic theory of the family, the demand for children (how many children and the ‘quality’ of children a couple desires) typically depends on the preferences of the parents, the relative price of
children and income. In the simplest model, a couple will maximize a utility
function, U f (n, q, Z), where n is the quantity of children, q is the quality
of children (typically inferred by the time and monetary resources spent on
each child for education, health care and so on), and Z is a composite of all
other goods, subject to a budget constraint, which might be specified as
pc qn pzZ I, where pc is the price of each child, pz is the price of all
other goods and I is income.6 The higher the price of children, the lower the
demand, and the higher the income, the higher the demand for children
(children are normal goods).7
In general, economists show that couples will trade oﬀ quality and quantity of children. Today, most couples choose fewer but higher quality children. Other considerations also arise in these models. For example, higher
wage rates for women are negatively related to the demand for children
because of the substitution eﬀect between time raising children and time in
the labour market as discussed in Chapter 5 (see also Rosenzweig and
Schultz, 1985; Becker, 1981; Applebaum and Katz, 1991).
Until recently, understanding the supply of children was simple and not
modelled by economists. Most men produce sperm on demand and can
potentially produce millions of oﬀspring. Women are born with all the eggs
they will ever produce (up to one to two million), and release one at each
ovulatory cycle. Fecundity (the probability of conceiving in any menstrual
cycle) is exogenous and random. If a couple wanted to have children, then,
with some probability, unprotected sexual intercourse would lead to the fertilization of a woman’s egg by a man’s sperm.
Rosenzweig and Schultz (1985) rue the absence of an integrated framework to analyse both the demand for and supply of children and note that:
‘Economists have recognized the joint relevance of biological and behavioral factors in determining fertility . . . but this perception has not been
suitably incorporated into the empirical study of fertility.’

Economics of assisted reproduction

295

2 The evolution of reproductive markets
Today, a prospective parent can go to the Internet, drop a vial of semen
from a donor with specific selected characteristics into a ‘shopping cart’
and have that semen delivered in twenty-four hours. Similarly, one can sift
through the profiles and pictures of women who are egg donors and select
eggs from women with desired characteristics and arrange an egg delivery.
These markets are two segments that loosely fall under the rubric of
Assisted Reproduction Technologies (ART), which is a shorthand term for
the numerous procedures aided by technology used to produce a baby.8
Now the cost of producing children can radically diﬀer among people of
similar incomes and values because some prospective parents pay to gain
rights to the genetic components that build the child while others do not.
How did we get here?
The rise of assisted reproduction is the story of need and technology
changing together. The very first artificial insemination (AI) took place in
1790 in Scotland. In that case, a woman was inseminated with her
husband’s sperm. Until recently, male infertility was a taboo subject and
donor insemination (inseminating a woman with sperm from a man who
was not her husband) was a well-kept secret. The first known case of donor
insemination (DI) occurred in 1884 but was not publicly admitted until
1909. According to David Plotz (2001) the delay was because the case was
so shocking:
Dr. William Pancoast, a professor at Jeﬀerson Medical College in Philadelphia,
found that a woman in his care couldn’t get pregnant because of her husband’s
infertility. At the urging of some of his medical students and with the permission of the husband – but not the wife – Pancoast anesthetized her and impregnated her with semen taken from the ‘best-looking’ student in the class. She never
was told what was done to her or that her husband was not her child’s father.
This kind of subterfuge, made-up-on-the-spot, loosey-goosey standards and
reliance on overwilling medical students would become the dismal defining qualities of donor insemination (DI).

Although it was known to be practised throughout the early to mid1900s, insemination with live sperm was conducted with discretion, or outright secrecy. Women, or couples, made arrangements privately with a
doctor. Who was the father of the child? Various laws gave the babies thus
produced tenuous legal standing or made them outright illegitimate in most
countries. Religious leaders around the world denounced DI in the 1950s.
Indeed, little may have changed if it were not for rapid changes in technology in the 1960s and 1970s: the ability to successfully freeze and use sperm
later to develop an embryo and the technology to fertilize an egg outside of
the body and successfully implant the embryo in a womb. These changes

296

Economics uncut

liberated baby production from a specific time and a specific body. Joining
these technical abilities was a critical change in law in 1973 in the USA,
which stated that the husband of an inseminated woman was the legal
father of a DI baby as long as he knew of the treatment.9
These, and other rapid changes, resolved a number of issues that had
relegated DI to the private arena where there was a fragmented ability to
procure the components that produce a baby (we might call this an incomplete market); sperm banks could provide women and their doctors with
needed sperm on demand. The public market in sperm was born along with
the first sperm bank in the early 1970s. California Cryobank, the longest
continuously open public sperm bank in the USA, opened in 1977. Soon
after, Baby Louise, the first baby known to have been successfully born after
fertilization outside of the body, ‘in-vitro’, was born in England in 1978.
Assisted reproductive technologies, and the markets that accompanied
them, took oﬀ.
Several historical facts may have sped the development of assisted reproductive technologies. The first is that the baby boom generation, the largest
generation in history, reached childbearing age in the 1960s and 1970s; the
second is that increasing labour force participation among women led to a
delay in childbearing among baby boomers until the ages where infertility
was more common; the third is that the resulting soaring demand for adoption, combined with a shortage of children available to be adopted (possibly as a result of the legality of abortion, the availability of birth control,
and the illegality of market transactions for babies), left many potential
parents without children. Medical procedures to eliminate, reduce or ‘get
around’ fertility problems were guaranteed a large market if ethical and
legal quandaries could be eliminated. As we have already noted, such problems were publicly illuminated in surrogacy. Early on, the surrogate’s egg
was implanted with a commissioning father’s sperm. Fears about surrogate
mothers who might fight for custody, and a number of court cases where
such custody battles took place (raising such legal and ethical questions as,
‘Can a baby be owned and, if so, who owns the baby?’ and, ‘Is surrogacy
the same thing as selling a baby?’) nearly destroyed the market for surrogacy. But the first known baby resulting from embryo transfer (where the
egg was not from the woman in whose womb the embryo was implanted)
was born in 1984 and gestational surrogacy, where the surrogate carries a
baby that is not genetically hers, became popular (Mathews, 1984). This
method made the parentage of the baby clearer, was contractually easier to
defend and reduced the prevalence of lawsuits over custody. Indeed, gestational surrogacy, often called ‘rent a womb’, is now the norm.
After hundreds of court cases relating to various procedures of assisted
reproduction, both the law and market regulation (which had lagged behind

Economics of assisted reproduction

297

the market development) have begun to catch up. The Uniform Parentage
Act of 2000 in the USA asserts in Article 7, Section 702, that, ‘a donor is
not a parent of a child conceived by means of assisted reproduction’.
3 Assessing the size of markets in assisted reproduction
How big is the potential market for assisted reproduction (AR)? In the year
2001, there were at least 421 AR laboratories in the USA and surrounding
territories.10 The medical characteristics of fecundity along with the
medical, demographic and behavioural characteristics of fertility are
appropriate preconditions for a potentially large demand for assisted reproduction. Medical characteristics of fertility include a woman being allergic
to her husband’s sperm. She is fecund, he is fecund, they just are not
fertile together. Demographic characteristics include the decline of fecundity with age. Behavioural characteristics include labour market participation that may be associated with a delay in childbearing until ages when
fecundity is low.
According to the Centers for Disease Control (CDC), of the 60 million
women of childbearing age in 1995 in the USA, about 13 per cent received
an infertility service. If each of these women were to demand some form of
assisted reproduction, then, in the USA alone, there are at least 8 million
potential customers of assisted reproduction. Including others who have
not received infertility treatments but who may desire children (say, women
without partners or gay and lesbian couples) the potential demand for
assisted reproduction is even greater in the USA.
While the potential demand is large, the potential supply is even larger.
Each man and woman carries millions of eggs and sperm, but only
demands a few to create children for himself or herself. If each man and
woman produced only enough sperm or eggs to satisfy their own demand
for children, then there would be very few people willing to enter the market
in assisted reproduction as suppliers. But most people have a substantial
oversupply of self-produced sperm and eggs. Thus, our biological characteristics set the preconditions for large markets for eggs and sperm to exist.
Further, because problems with infertility can originate with either partner
or the combination of people in the couple, the size of each market (sperm,
egg, surrogacy) is independently potentially large.11
We do know that in the USA in 2001, 107 587 procedures were performed
by AR as defined by the CDC, resulting in 29 344 live births and 40 687
infants. The AR market is clearly growing. Just between 1996 and 2001,
procedures increased by 66 per cent and live births by 101 per cent (CDC,
2003b). It is diﬃcult to assess actual market size from these CDC statistics,
however. They do not include pregnancies resulting from sperm donation
when the egg is not handled in the laboratory (underestimating the size of

298

Economics uncut

the assisted reproduction market), although they do include the medical
procedure of assisted reproduction, using the egg and sperm of the existing couple (overestimating the size of the market). Meanwhile, the CDC
reports that there were 571 gestational carriers of fresh eggs in 2001 (CDC,
2003b). Unfortunately, there are no statistics on the size of the market in
surrogate motherhood.
4 Reproductive markets: thinking about the economic issues
As with any economic analysis, we must consider the determinants of
supply and demand in each market. The diﬀerence here is that there is not
a long history of economic analysis of these markets. What are the main
issues that have to be addressed and which economic tools may help us do
so? Reproductive markets are characterized by both profit and altruistic
motives of suppliers, varying levels of risk associated with both supply and
demand, and prices that depend on the characteristics of suppliers and
demanders. To investigate these markets, we will use the rare existent
article on the subject and the fundamental tools of economics to examine
the components of supply and demand. In the following discussion, it
will be assumed that contracts (for eggs, sperm or surrogacy services) are
traded.
4.1 Supply
Why do people supply their own eggs, sperm or wombs so that others may
have a baby? What are the incentives to supply AR contracts? Many
authors have written about the multiple incentives of surrogate mothers,
egg and sperm donors. Although a host of reasons lie behind any person’s
decision, when Klock et al. (2003) evaluated nine possible motivations by
querying egg donors about the motivation behind their participation, they,
like other authors, found that altruism and compensation together were the
primary factors.
Because of these dual motives, the markets for eggs, sperm and surrogacy
share the common feature that the supply of contracts can be realized
through purchase (where legal) or donation. What does the supply curve
look like? For the moment, set aside ethical issues in buying and selling
such contracts. If their purchase is illegal, and the ‘market’ supply is
restricted by a price equal to zero, then only altruistic suppliers will enter
the market (see Figure 10.1).12 The supply curve (S1) is a straight horizontal line at P0 and ends at the quantity of altruistic suppliers of contracts,
Qs. If the demand is no larger than the supply at P 0 (D1), then a price
will never arise, and this ‘market’ is in equilibrium. Typically, however,
demand is at D2, the market is in disequilibrium, and there is excess demand
for contracts.

Economics of assisted reproduction

299

Price

D2
D1

S1
Qs
Figure 10.1

Quantity

A ‘market’ for AR contracts with donation only

Whether or not they agree that a market is the solution, most authors
agree that there is excess demand for assisted reproduction, blood
supply and organs (see Ethics Committee of the American Society of
Reproductive Medicine, 2000; Harris and Alcorn, 2001; Thorne, 1998).
Guerin (1998) writing about sperm donation in France, where, at the time,
sperm could only be donated, documents statistics showing that the
number of couple candidates for AR with donor semen was six times
greater than the number of sperm donors proposing a donation, and up to
ten times greater than the number of donors accepted.13
There is a growing number of authors who argue that allowing suppliers
to be compensated will increase supply, and take care of excess demand. But
there is a longstanding controversy over supply behaviour in such markets.
If a market emerges in an arena where altruism is an important incentive,
then altruists may feel some revulsion (or derive negative utility) when the
market emerges and they may refuse to participate. Titmuss (1997) derived
supply change predictions when a market for human blood joins a system of
donation.14 It is possible that a market with a positive price could end up
with fewer equilibrium contracts traded than one with no price at all! To see
this, examine Figure 10.2. Once payment is available, some altruistic donors

300

Economics uncut

Price

D2
D1
P2

P1
S1

Qm1

Qm2

Quantity

Figure 10.2 A market for assisted reproduction contracts with a legal
positive price
(Qs – Qr) may withdraw from the market. Others will enter the market once
the price is positive, resulting in the supply curve S1, that starts at P0, and
looks like a typical supply curve after the supply of altruistic donors is
exhausted. If the demand curve is D1, then the equilibrium quantity of contracts is Qm1 and the equilibrium price is P1. But the equilibrium quantity
traded is less than if the price were zero. We believe, however, that D2 is more
representative of demand, and there is a positive equilibrium price, P2, and
a traded quantity of contracts, Qm2, that is greater than Qs.
Hewitson (1997), in her model of surrogate motherhood, derives a supply
curve with these features. In each market (sperm, egg, surrogacy), both altruism and the profit motive are present. What other factors go into the supply
function? Ignoring risk (discussed below) each market requires varying level
of medical, emotional and physical time commitment on the part of donors:
1.

Surrogate mothers must subject themselves to initial medical and psychological exams, and to significant monitoring and restriction of their

Economics of assisted reproduction

301

behaviour during the pregnancy. They may experience disutility from
the monitoring required in many surrogate motherhood contracts and
from the elimination of behaviours that they may enjoy, but are banned
through the contract period.15 Activities that increase utility will lower
the supply at each price while activities that reduce utility increase the
supply at each price. Further, surrogate mothers may derive utility or
disutility from gestating a fetus, birthing a child and surrendering the
baby. Finally, carrying a baby to term is a time-intensive activity and,
just as in the demand for children, surrogates’ value of time will influence supply at every price.16
Typically, egg donors undergo a several weeks’ regime of daily selfinjections (hormones to bring more than one egg out during ovulation), a final injection of a diﬀerent drug to release the eggs, and then
an ultrasound and local anesthetic for the egg retrieval. Each of these
procedures may reduce utility (and, hence, supply), although the
change may be minor.
There seems to be little outside of altruism and compensation that
determines the supply of sperm. If you think about it, this makes
perfect sense. While there are some requirements to be met for sperm
donors at the major clinics (for example, they must have an initial physical exam, blood draws to test for diseases, and they may have to
commit to a period of participation), donating sperm is a non-invasive
activity of minimal time commitment, with essentially no medical risk.

4.2 Demand
Why do people demand sperm, ova or surrogacy contracts and what determines the demand curve? Surprisingly, almost no work has been done on
this issue. To the author’s knowledge, as of January 2004, there are no published articles in the economics literature on assisted reproduction outside
of the Hewitson (1997) article. Looking back to the older literature on the
demand for children, however, we can speculate on how the couple’s maximization process might be diﬀerent. Remember that the utility function
contains, as its arguments, the number of children, the quality of children,
and other goods. In a model of the demand for children via assisted reproduction, the same features will be present with at least two additions. First,
the ‘quality’ argument (money spent on existing children) may now have to
be split into post-birth and pre-birth ‘quality’ considerations. Second,
people derive utility from genetic continuity (see Saint-Paul, 2002; Posner
and Posner, 1999) and so we might see an additional argument that is some
function of the genetic makeup of the couple. Of course, genetic continuity was implicitly present before, but since children were typically selfproduced, it was automatic. What that function looks like is not yet known.

302

Economics uncut

However, with assisted reproduction, the level of genetic continuity (genes
from both man and woman, from the woman only, from the man only, or
from neither) varies as does the price of such continuity. The genetics of the
mother and the father may enter the utility function separately. Further, a
couple that must include the genetic makeup of another person may want
to find a donor who is physically like them (that is, they may want someone
of the same race, with the same hair and eye colour, and so on as one
member of the couple). In other words, there may be an argument for ‘look
and feel’ genetics. Finally, there are characteristics that are valued by
society (intelligence, beauty and so on) that may enter into the utility function. This was explicit, but not exclusive to, the ‘genius’ sperm bank, which
raised many questions about eugenics and designer babies:17
Over the past decade, sperm banks have appeared around the world catering to
women clients who care mostly about the physical appearance of the sperm
donor. . . . Fortunately . . . a few sperm banks select sperm donors mainly based
on achievements and intelligence. Their sperm donor catalogs contain extensive
information about sperm donors’ scientific discoveries and inventions, published
papers and patents, school records, music and artistic abilities, athletic abilities,
as well as the usual race and appearance information . . . Genius sperm banks
select sperm donors based mainly on achievements and genetic quality
rather than based solely on sperm donor appearance, race and sperm quality.
They cater to clients who want to improve the intelligence of their child by
selecting a sperm donor of superior intelligence and outstanding achievements.
(http://www.geniusspermbank.com/)

The demand for children changes markedly under a regime of assisted
reproduction when we consider the income constraint. Now, there is a price
attached to the birth of each child that was not present before (the additional price of sperm, or egg, or womb and the price of the medical
procedures) and there is a price variation attached to the ‘quality’ of the
genetic components (see the discussion on price below). Ceteris paribus,
children born from assisted reproduction are more expensive than children
born the old-fashioned way.
Let us consider just a couple of the implications of these changes in the
model. What used to be automatic is no longer so. Couples may now trade
oﬀ genetic continuity for social value depending on the price. Also, because
price can depend on the characteristics of the genetic material, both the
price of having a child and the price to produce a certain quality of child
at birth can be higher for AR couples (than in the old-fashioned case).
Recall that quality in the old-fashioned utility function is produced by
money spent on the child after birth. One potential implication of these
extra costs to produce a baby is that it is possible that AR couples will have
fewer children of lower post-birth quality than other couples with the same

Economics of assisted reproduction

303

income and tastes (ignoring the fact that AR births are much more likely
to produce multiple children than old-fashioned births).
4.3 Risk in demand and supply
Risk is a common attribute of reproduction. The probability that a baby
will be conceived in any given reproductive cycle (called the ‘take home baby
rate’ in assisted reproduction clinics) is only 20 per cent for couples with past
proven fertility and is relatively low in all circumstances of assisted reproduction. The probabilities vary, and are higher for fresh eggs than for frozen
(see Table 10.2). Even though the probability of producing a baby is low for
fertile couples, in the typical case there is no cost for a couple to attempt to
produce a baby in a given cycle through copulation. But every attempt to
produce a child costs if using AR. Typical AR procedures include multiple
stages and planned monetary payments at each stage. The CDC classifies
the AR process leading to a child into five stages: cycles started, retrievals,
transfers, pregnancy, childbirth. The total cost varies, but can be many
thousands of dollars (and the couple may not end up with a child).
Suppose a couple requires a donor egg to conceive a child. Fertility
Alternatives is one egg donor agency that posts fees (the fee structures are
remarkably similar across the various agencies). Among other charges (it
notes that prices will vary widely), Fertility Alternatives requires a profile
fee (US $25), agency fee (US $2500 or 50 per cent of egg donor’s fee,
whichever is higher), egg donor’s fee (US $5000–US $15 000), facilitator’s
fee (US $250), mandatory health insurance fee (US $500), medical screening fee (US $800), attorneys’ fees (US $750), and so on. Other expenses
Table 10.2

Probability of a live birth in any cycle

ART procedures

Fresh eggs

Frozen eggs

Non-donor eggs
Donor eggsb

25.4%
43.3%

19.5%a
23.5%

Gestational carriers
Non-donor eggs
Donor eggs

36.8%
46.1%

22.6%
32.8%

Notes:
The pregnancy rate per cycle of couples with proven past fertility is approximately 20 per
cent (Society for Assisted Reproductive Technology [SART], A Patient’s Guide to Assisted
Reproductive Technologies, p. 7, http://www.sart.org/TextPatients.htm).
a
Per thaw.
b
Per transfer.
Source: ART Statistics from CDC (2003a).

304

Economics uncut

include egg donor travel expenses, lost wages for the donor, all medical
expenses incurred by the donor and all costs of the IVF procedure (see
http://www.fertilityalternatives.com/ipfeesed.html). Studies that examined
various economic costs associated with a successful delivery with IVF
around the world were collected in Garceau et al. (2002). Estimated costs
(typically payment of fees by patients) ranged from £10 372 in Finland, to
£25 499 in Canada, to £52 566 in the United States.
These fees are typically non-refundable if a baby is not produced by the
contract. As a question of the economics of childbearing, this risk is critical. First, the cost to ‘produce’ a child is much higher for a couple that
requires assisted reproduction. Second, the budget constraint is likely to
have a probabilistic function of childbirth in it (a couple could pay for AR,
but not end up with a child, making other payments moot). In the end, it
may mean that diﬀerent prospective parents with the same income and the
same desired quantity and quality of children (who will face radically
diﬀerent costs to produce children depending on their AR needs and on the
probability of child conception) will have diﬀerent demands for children.
Third, in the USA, all these costs are borne by the prospective parents and
none by society. Health insurance still does not cover most AR procedures,
and no egg or sperm purchases. In contrast, in France, IVF is fully reimbursed by the social security system; in Belgium, Denmark and Norway, the
state bears almost all costs; in England and Wales, some IVF procedures
are funded by the National Health System (see Garceau et al., 2002). As a
result, AR contracts in the USA will only be demanded by potential
parents who have substantial disposable income and who can aﬀord to deal
with the probabilistic outcome. Fourth, it is clear that high-income couples
will make up the majority of demanders of AR. Since income varies by race
and ethnicity, demand will subsequently reflect the race, ethnicity and the
tastes of high-income people.
However, as in many markets where the outcome is risky, an insurance
market is emerging. Some donor agencies in the USA now oﬀer insurance
policies, as do many IVF clinics. They are often called ‘shared-risk’ programmes.18 At the moment, these insurance markets carry interesting economic and ethical questions because the insurance market is not
independent of the services market. Typically, a client pays a higher fee for
IVF by purchasing a policy through the agency. If the procedure is successful, then the agency keeps all fees. If not, a high proportion of the
agency’s fees (but not, typically, drug costs and other medical costs) is
refunded. This results in a funny kind of moral hazard because the agencies themselves provide insurance: ‘such programs have a built-in conflict of interest which is likely to skew clinical decision-making toward
achieving pregnancy regardless of the impact on the patient in order to

Economics of assisted reproduction

305

avoid paying a refund’ (Ethics Committee of the American Society for
Reproductive Medicine [ASRM], 2000).
Ethically, many believe that shared-risk programmes violate Opinion
6.01 of the American Medical Association’s Code of Medical Ethics, which
states that a physician’s fee cannot be made contingent on the outcome
(ASRM, 2000).
There are many other forms of risk in these markets, which must be considered in any economic analysis. Both surrogate mothers and egg donors
face medical risks associated with the egg retrieval, implantation or pregnancy. Often, surrogate mothers lose their compensation under conditions
of spontaneous abortion (taking place without any action on the part of
the woman carrying the child). At this time, there is no insurance market
for surrogate mothers who have gone through arduous medical and psychological testing and procedures, but fail to produce a viable baby.
Some forms of risk in the AR market can be classified as property rights
risks. Who has legal rights over the baby that is conceived as a result of a
contract?19 The ways in which the property rights risk manifests itself
depends on the specific market. As discussed earlier, in the past, legal disputes typically arose in surrogacy because both the commissioning couple
and the surrogate mother desired custody. Sometimes, however, it is
because there has been some problem with the baby’s physical or mental
health and no one wants custody of the baby.
Of course, property rights can get complicated both legally and ethically.
In surrogacy, there are potentially five people involved in the creation of the
baby: the commissioning father, his wife or partner (who must legally adopt
the baby), a sperm donor, an egg donor and the surrogate. The surrogacy
market, in particular, is sensitive to issues of risk and agencies spend considerable eﬀort on risk reduction. First, agencies require the commissioning parents to pay for an independent lawyer who looks out for the interests
of the surrogate mother. They also advertise whether any specific surrogate
has already had a successful surrogate pregnancy (signalling a lower risk of
legal disputes later). Finally, they require the commissioning father to pay
for counselling sessions, either with a psychologist or in a support group of
surrogate mothers. These sessions appear designed to help the surrogate see
herself as a temporary caregiver rather than a mother, and this reduces the
risk of a custody dispute (Kolczykiewicz, 2003). There has been no such
issue for sperm donors (the Uniform Parentage Act of 2000 specifically
states that the sperm donor is not the father).20
4.4 Putting supply and demand together
How does one think about price in AR markets? Some people may find it
uncomfortable to consider prices for men’s sperm and women’s eggs.

306

Economics uncut

However, as discussed in Chapter 7, economists often work with a concept
called hedonic pricing, where an attempt is made to decompose the total
price of a product into implicit prices for diﬀerent traits of that product. So,
what are the traits that might be valued in AR markets? Adoption markets
(where formal prices are typically illegal) have always been associated with
high demand for more desirable characteristics, in particular, race, looks and
health (see Kossoudji, 1990). Current advertising still reflects such attitudes:
The total expense for an adoption depends on such factors as the child’s age and
nationality/ethnicity, whether the child has special needs, the specific agency you
work with, how long you are willing to wait, what complications arise, and (for
foreign adoptions) what country you adopt from. You’ll usually find that the costs
involved in adopting a healthy Caucasian newborn/infant are higher than
in adopting a child of other races, children with special needs, or children from
other countries. (http://adoptionservices.org/adopting_families_Adoption_Cost.
htm)

Assisted Reproduction markets are slightly more complicated, however.
The way that firms advertise their product tells us something about what
they believe customers desire (and buy). At least one egg donor agency,
Tiny Treasures, advertises ‘Ivy League egg donors’. Both sperm banks and
egg donor agencies on the Internet oﬀer extensive information on their
donors.21 Although each agency is unique in the specific information provided in the profiles, a host of characteristics can be found on each individual donor, related to racial or cultural characteristics (self-reported race,
ethnicity, father’s and mother’s race or ethnicity, religion), physical traits
(height, weight, eye colour, hair colour, hair texture, skin tone), health traits
(blood type, personality test results, personal medical history, family
medical history), and ability traits (level of education, grades, SAT scores,
sometimes which university attended, occupation or major at college). Are
some of these characteristics worth more than others?
Pricing operates diﬀerently in egg and sperm markets. Sperm banks do
not typically price an individual donor’s sperm. Instead, sperm banks may
specialize in donors with particular characteristics. A review of over 30
sperm banks showed a minimum price of US $100 per vial, a maximum
price of US $425 per vial, and a mean price across all sperm banks of
approximately US $200 per vial.22 In this sense, sperm banks may operate
like firms in a monopolistically competitive industry. There are sperm
banks based on religion, ethnicity and achievements. One sperm bank specializes in sperm from gay men and, as noted earlier, one now defunct
sperm bank specialized in the sperm of Nobel Prize winners.
There is at least one sperm bank that prices broadly by education (many
sperm banks select donors on the basis of education. In fact, of the 1509

Economics of assisted reproduction

307

donors discussed below, not a single one had not been to college). Fairfax
Cryobank charges US $280 for sperm from men with (or earning) doctorates, US $215 for sperm from men without doctorates, but from whom all
of the relevant personal information has been collected, and US $135, a discounted rate, for sperm from men who became donors before extensive
information was gathered.23 Given the range of prices by agency, there are
two ways that men could have higher sperm prices. First, by having the
broad characteristics that bring a higher price within an agency or, second,
to sell sperm to an agency whose higher prices reflect a higher overall
quality of donor. Without individual variation in prices, it can be diﬃcult
to tell which characteristics are desirable. So we use an alternative way to
determine the desirability of a donor’s sperm. Many sperm agencies publicize whether a man’s sperm has already been sold (sometimes listed as
resulting in a pregnancy). We will use this selection as a proxy for desirability and assess which characteristics make it more likely that a man’s
sperm was chosen.
It is far too early to arrive at a definitive decision about the determinants
of price in these markets. However, we can present an empirical analysis
that is illuminative and absolutely preliminary. As already mentioned, the
author collected information on 1509 sperm donors from over 30 sperm
banks that advertised on the Internet. From these, 490 had information on
a previous pregnancy or selection. Not too surprisingly (given that sperm
banks do not individually price sperm) there is almost no characteristic that
is associated with a higher or lower sperm price (see Table 10.3). The regression on prices shows that only a graduate education (US $19.83 more than
mere college) and height (US $1.91 per inch) are associated with increased
prices. This is not too surprising since one sperm bank specifically prices by
education.
Surprisingly (assuming selection is a good proxy for desirability), however,
only race influences selection. No characteristic explains diﬀerences in selection probabilities except for being Asian or being of multiracial heritage
(both negative at 5 per cent level of significance).24 What does this tell us?
Racial aspects of the market are unclear. White, African American, Latinos
and Native American men are not selected at diﬀerential rates. Earlier it was
noted that the income distribution may play a large role in determining the
tastes of demanders. It may well be that there is racial matching for sperm
(people choose sperm from individuals of their own identified race) and that
there are fewer demanders of Asian or multiracial heritage (note that this
could be an income eﬀect, or a genetic eﬀect – a lower proportion of Asian
men, for example, may be infertile). It also raises the possibility, however,
that the characteristics of the male do not influence desirability (or that the
filters of the various agencies are strong enough to weed out any obviously

308

Economics uncut

Table 10.3 Sperm donor compensation and selection eﬀect of individual
characteristics

Height
Weight
Graduate education
Eye colour
brown/black
green
Hair colour
brown/black
red
Race
Latino
Asian
Black
Mixed
N

Compensation (US$)

Selection

1.91* (0.82)
0.09* (0.07)
19.83* (3.74)

0.985
0.999
1.283

3.48* (4.95)
4.37* (4.87)

1.631
1.538

2.45* (5.42)
1.61 (13.16)

0.780
1.000

2.94* (11.65)
2.89* (6.14)
2.44* (8.78)
16.07* (8.75)
879

1.268
0.474*
1.416
0.358*
438

Notes:
Standard errors in parentheses. Regression includes a constant: the omitted characteristics
are a white, blue-eyed, blond, with a high school diploma. The compensation regression is
OLS, and the selection regression is a binomial logit (the reported coeﬃcients are the change
in the log odds associated with that variable – a value of 1 means no change, greater (less)
than 1 means increased (decreased) probability). Standard errors are not reported for these
constructed numbers.
*Significant at 5% level.

undesirable characteristics, and eye and hair colour, weight, height and so on
are secondary in decision-making).
Women set their own compensation request at egg donor agencies. This
is known as an ‘oﬀer’ price. We cannot tell how far the final price diﬀers from
the oﬀer price because this information is not made public by agencies. In
this dataset of over 2000 egg donors, the average compensation request is
US $5145, the minimum is US $1800 and the maximum is US $20 000,
although we cannot say for sure whether a woman will receive her compensation request. Therefore, the same technique is used for egg donors as with
sperm donors (having been already selected is a proxy for desirability).
Table 10.4 shows the same regression results for women as for men. But
the results for women are remarkably diﬀerent. First, education plays the
critical role in the compensation request and in the probability of being currently matched. A college education is worth US $448 more than a high
school education and a graduate education is worth US $2365 more than a

Economics of assisted reproduction

309

Table 10.4 Egg donor compensation and selection eﬀect of individual
characteristics

Height
Weight
Education
college
graduate school
Eye colour
brown/black
green
Hair colour
dark
Race N/A
N

Compensation

Selection

88.43* (34.01)
14.69* (5.08)

1.131*
0.967*

448.26* (223.35)
2384.87* (322.68)

1.499
4.740*

263.28 (225.08)
285.76 (222.70)

0.669
0.931

217.39 (212.22)

1.170

693

296

Notes:
Standard errors in parentheses. Regression includes a constant: the omitted characteristics
are a white, blue-eyed, blonde, with a high school diploma. The compensation regression is
OLS, and the selection regression is a binomial logit (the reported numbers are the change
in the log odds associated with that variable – a value of 1 means no change, greater (less)
than 1 means increased (decreased) probability). Standard errors are not reported for these
constructed numbers.
*Significant at 5% level.

high school education (about 15 per cent of egg donors only have a high
school education). Further, physical characteristics matter for women.
Every extra inch of height brings an additional US $88 while every extra
pound reduces compensation by US $15. Selection characteristics are
similar: a graduate education increases the odds of being currently
matched, as does height. Weight reduces the probability of a match.
Why do a woman’s characteristics appear to relate so strongly to price
while men’s do not? A good part of this result is an artifact of the way that
prices are set in the two diﬀerent markets. But why is there no individual
pricing for men, but women set their own price? One reason surely is that
removing eggs is a more costly and risky procedure than removing sperm.
At the same time, both egg and sperm are just a set of genetic components.
Outside of a diﬀerential for risk, there is no obvious reason why women set
their own price but there is typically undiﬀerentiated pricing for sperm.
While it may be empowering for women to set their own compensation
price, it begs the question of the determinants they use to calculate that
price. Almost certainly education plays a dual role in egg pricing. It
enhances the desirability of a woman’s eggs because it proxies intelligence.

310

Economics uncut

It also is correlated with the value of a donor’s time (her opportunity cost).
It is curious that for men, race appears to influence desirability, but for
women, physical characteristics outside of race play a role in the price that
a woman sets. Women may be more culturally sensitive to the value of
looks. Selection, as a proxy for desirability, shows almost exactly the same
results as the price regressions, suggesting that there are some fundamentally sound cultural reasons for setting such price diﬀerences.
5 Ethics in assisted reproduction
The previous sections of the chapter function in the realm of positive economics (‘How does it work?’), but when we look at ethics, we move into the
realm of normative economics (‘How should it work?’). Ethical arguments
about assisted reproduction are rich and varied. There are hundreds, if not
thousands, of articles on the big and minute ethical questions raised by AR.
Most arguments have to do with the proper scope of the market. Where
should the market stop?
First and foremost, people argue that there should be no price put on a
human being’s body. The argument is against commodification of the person
because this way we lose our value as humans. The Human Fertilization and
Embryology Authority (HFEA, 1998) claimed just that when recommending that egg donation should be a gift and that payments should be eliminated. Radin (1996), like many, argues that we ought to be inalienable to the
market and that commodification is harmful to personhood. Thus, there are
‘contested commodities’. Even the language of commodification as part of
the discourse is itself harmful. Surrogacy has provided fodder for a number
of discussions. Anderson (1990) argues that surrogacy is women’s labour,
that is, the labour of carrying a baby to term, and if treated as a commodity
is degraded. Further, children are reduced to commodities.
Is assisted reproduction tantamount to selling babies? Hirschman (1991)
states that AR markets are, ‘centered around the production and acquisition of babies – babies in the form of component sperm and eggs, babies
in the form of fresh or frozen embryos, . . . and babies as full term living
infants’. She argues that, in the end, although such markets are ‘morally
troublesome’, they can be, ‘ethically construed as positive if they serve the
larger sacred goal of creating families’. Ethical questions often confront
religious questions. What makes a family? Who can make a family? What
makes a mother or a father?
6 Conclusion
In the end, these ethical questions drive us to a better understanding of ourselves. This debate is crucial to help us determine what to do about market
regulations, laws about the permissibility of some markets, and the modes

Economics of assisted reproduction

311

of market organization. At the same time, while we argue ethics, the
markets go on, fed by technology, need, altruism and greed. We have outlined here some of the critical features of markets in assisted reproduction.
For instance, we know that the potential size of the markets is enormous
and that the actual size is growing rapidly. We also know that there are
diﬀerent pricing structures in egg, surrogacy, and sperm markets and that
individual prices in the egg market vary markedly by the characteristics of
the donor. Further, various forms of insurance have arisen to reduce the
impact of risk in the markets. As relatively new markets, we can argue that
they follow the basic constructs of economic theory.
What we do not know is why men do not set their own price for sperm,
and why physical characteristics are important for egg donor’s desirability
but not sperm donor’s (in either case, 50 per cent of the DNA is contributed
by the donor). We also do not have a sense of the changes in social welfare
that have emerged from diﬀerent market organizations. In summary, we do
not yet know many of the details of how these markets function, nor what
stable equilibria look like in the markets for sperm, eggs and surrogacy.
Understanding the positive and normative issues of markets of today
can help us figure out the social and economic consequences of the markets
that are coming and the ethical arguments that may support or fall against
the development of those markets. Genetic engineering and ‘designer
babies’ are only a step away. As Francis Fukuyama (2002) states:
even if genetic engineering on a species level remains twenty-five, fifty, or one
hundred years away, it is by far the most consequential of all future developments in biotechnology. The reason is that human nature is fundamental to our
notions of justice, morality, and the good life, and all of these will undergo
change if this technology becomes widespread.

Notes
1.
2.
3.
4.

People without partners and gay and lesbian couples are also customers in the market
for assisted reproduction.
In in-vitro (glass) fertilization, the egg and sperm are combined outside of the body, typically, in fact, in a plastic petri dish. The first baby born from in-vitro fertilization was
called a ‘test tube’ baby.
The surrogate contract itself is actually between the commissioning father and the
surrogate mother. The man’s partner then enters into an adoption contract after the
baby’s birth.
Much of the legal and market information in this chapter focuses on the United States.
Laws and regulations are changing every day in this arena and may be diﬀerent from
those portrayed here. Also, laws and regulations continue to diﬀer significantly across
countries as does prevalence of use of existing markets.
A typical example is when the egg of a woman with blocked fallopian tubes is fertilized
with her partner’s sperm by IVF, then inserted into her uterus. This market is a market
for the medical procedure itself. Hereafter, except where noted, assisted reproduction will
refer to the market where some genetic or physical component is purchased rather than
the medical procedure itself.

312
6.

9.
10.
11.

12.
13.
14.

15.

16.

17.

Economics uncut
There are many microeconomic models of the demand for children, some of them much
more complex than this one. For a more in-depth look, A Treatise on the Family, by Gary
Becker (1981) is a good place to start. Models often diﬀer by whether the couple lives in
a more developed or less developed economy.
A conundrum arose when wealthier societies began having fewer children and the
relationship between income and the number of children appeared to turn negative. This
problem was resolved by asserting that a couple cared not just about the number of children in the family but also about the ‘quality of children’ in the family.
The assisted reproduction literature is rife with acronyms and many of the terms do not
yet have a clearly determined definition. The Centers for Disease Control (CDC) in the
United States may be unique in that AR only includes techniques where both the egg and
sperm are handled in the laboratory. CDC bases its definition on the 1992 law that
requires it to produce an annual report on success rates at AR clinics. This means that
statistics on the size of the market based on CDC success rates omit all DI or AI procedures where only the sperm is handled by the clinic. Unfortunately, CDC is the repository of clinic statistics. Statistics on sperm donation alone do not appear to exist. In this
chapter, all procedures are enveloped by AR as specified above, except where specific
CDC statistics are cited.
The Uniform Parentage Act (1973) attempted to make the laws in all states the same
regarding parenthood. There were, and are today, some other conditions on determining the father of a DI baby.
384 reporting to the CDC (Centers for Disease Control and Prevention) as required by
law and 37 clinics known to be operating, but not reporting to CDC (see CDC, 2003b,
Appendix C). This does not include sperm banks.
Statistics on infertility tend to be similar across nations. According to one French study,
fertility problems originated with the woman in 38 cases out of 100, with the man in 20
cases, and with some combination of the two in 38 cases. In eight cases, infertility had
an unknown source (de la Rochebrochard, 2001).
Notice that we ignore the cost of medical services, lawyers and so on. Here we concentrate on the compensation to the supplier of contracts.
His argument, however, is that donation could help eliminate excess demand if recruitment were improved.
The 1997 volume is an expanded edition of Titmuss’s classic 1970 volume, The
Gift Relationship. Market withdrawal was only one component of the argument.
Changes in health risks, among other factors, also play an important role in the debate.
Arrow and Titmuss argued about these outcomes. For a summary of the issues, see
Thorne (1998).
Contracts may specify regular alcohol or drug tests, repeated psychological exams to
ensure that the surrogate will willingly give up the baby, and regular medical exams.
Surrogate mothers may be banned from extreme physical activity, risky vacations and
smoking. A simple sample contract may be viewed at surromomsonline.com. This
sample contract includes language such as: ‘2) The Surrogate agrees not to participate in
dangerous sports or hazardous activities, and promises not to knowingly allow herself
to be exposed to radiation, toxic chemicals or communicable diseases. 3) The Surrogate
further agrees not to smoke any type of cigarettes, drink alcoholic beverages, or use any
illegal drugs, prescription or non-prescription drugs without consent from her obstetrician or midwife.’
Putting much of this together, Hewitson’s supply utility function is Ui ( Mq*)
( R) (A) A p where the terms in the first parentheses are parameters of
the (dis)utility of gestating a fetus, birthing and surrendering a child, and the monitoring
costs of the optimal child quality, the terms in the second parentheses refer to the surrogate’s value of privacy and her propensity to engage in risky behaviours. All these are multiplied by either a profit ( 1) or altruistic (A 1) contract. Both and are altruism
parameters, and p is the surrogate’s compensation.
The ‘genius’ sperm bank was actually called the Repository for Germinal Choice and was
founded by Robert Graham. Only men who had won the Nobel Prize could donate

Economics of assisted reproduction

18.

19.
20.
21.

22.
23.
24.

313

sperm. More than 200 babies were born from sperm issued by this sperm bank before it
went under (see Plotz, 2001).
One egg donor agency oﬀers an optional shared-risk policy. In the basic plan prospective parents receive a 50 per cent discount on the agency fee for donor matching and
coordination of a subsequent egg donation cycle (US $375). In the premium plan,
prospective parents will pay no agency fee for donor matching and coordination of a
subsequent egg donation cycle (US $750) (see http://www.tinytreasuresagency.com).
In fact, one of the big ethical questions is whether rights should be thought of as property rights (which turns the baby into property – something we ought not to do).
My thanks to Mariusz Kolczykiewicz, a former student, for permission to use this
surrogacy risk information from a paper he wrote for my class.
Some information is free and public on the web, and some must be purchased. The following quote is from one agency: ‘California Cryobank (CCB) provides many types of
information to help clients select a donor. Long donor profiles, baby photos, audio interviews, Keirsey Temperament Sorter Reports and Facial Feature Reports are available for
clients to purchase. Short donor profiles and Staﬀ Impression Reports are available for
free’ (http://www.cryobank.com).
Author’s dataset of approximately 30 sperm banks and all of their donors.
These prices are for IUI specimens and are correct as of January 2004.
This dataset may contain biases that have yet to be understood. The range of information varied by agency so the selection of agencies may influence the results. Age may be
a positive factor in selection, as may grades and college scores. Many firms did not advertise the same information. The chosen regression was the one with the largest sample size
for men and women with the most comparable variables so that characteristics could be
compared. There were not enough egg donors with information about race to construct
a comparable regression. When the sperm donors’ regressions were run without race,
nothing was significant, indicating that race did not substitute for other characteristics.
Since sperm prices do not vary by individual, the regression should be considered illustrative only. Finally, there are not reported standard errors in the selection equation
because the marginals (changes in the log odds) are reported.

References
Anderson, E. (1990), ‘Is Women’s Labor a Commodity?’, Philosophy and Public Aﬀairs, 19 (1),
71–92.
Applebaum, E. and E. Katz (1991), ‘The Demand for Children in the Absence of Capital and
Risk Markets: A Portfolio Approach’, Oxford Economic Papers, 43 (2), 292–304.
Becker, Gary S. (1981), A Treatise on the Family, Cambridge, MA: Harvard University Press.
British Broadcasting Company (1985), ‘Inquiry over “Baby for Cash” Deal’, 4 January,
http://news.bbc.co.uk/onthisday/hi/dates/stories/january/4/newsid_2495000/2495857.stm.
Cannell, Fenella (1990), ‘Concepts of Parenthood: The Warnock Report, the Gillick Debate
and Modern Myths’, American Ethnologist, 17 (4), 667–88.
Centers for Disease Control and Prevention (CDC) (2002), ‘Assisted Reproductive
Technology Success Rates: National Summary and Fertility Clinic Reports’, Atlanta,
Georgia: US Department of Health and Human Services.
Centers for Disease Control and Prevention (CDC) (2003a), ‘Surveillance Summaries’,
MMWR, 200 (3), 52 (No. SS-9).
Centers for Disease Control and Prevention (CDC) (2003b), ‘Assisted Reproductive
Technology Success Rates: National Summary and Fertility Clinic Reports’, Atlanta,
Georgia: US Department of Health and Human Services.
de la Rochebrochard, E. (2001), ‘Sterility, Fecundity: What About the Men?’, Population and
Societies, Sep, 371, 1–3.
Ethics Committee of the American Society of Reproductive Medicine (ASRM) (2000),
‘Financial Incentives in Recruitment of Oocyte Donors’, Fertility and Sterility, Aug, 74 (2),
216–20.
Fader, Sonia (1993), ‘Sperm Banking: A Reproductive Resource’, California Cyrobank, Inc.

314

Economics uncut

Field, Martha A. (1988), Surrogate Motherhood: The Legal and Human Issues, Cambridge,
MA: Harvard University Press.
Fukuyama, Francis (2002), Our Posthuman Future: Consequences of the Biotechnology
Revolution, New York: Picador.
Garceau, L., J. Henderson, L.J. Davis, S. Petrou, L.R. Henderson, E. McVeigh, D.H. Barlow
and L.L. Davidson (2002), ‘Economic Implications of Assisted Reproductive Techniques:
A Systematic Review’, Human Reproduction, 17 (12), 3090–109.
Guerin, J.G. (1998), ‘The Donation of Gametes is Possible Without Paying Donors:
Experience of the French CECOS Federation’, Human Reproduction, 13 (5), 1129–32.
Harris, C.E. and S.P. Alcorn (2001), ‘To Solve a Deadly Shortage: Economic Incentives for
Human Organ Donation’, Issues in Law and Medicine, Spring, 16 (3), 213–33.
Hewitson, G. (1997), ‘The Market for Surrogacy Contracts’, Economic Record, 73 (222),
212–24.
Hirschman, E.C. (1991), ‘Babies for Sale: Market Ethics and the New Reproductive
Technologies’, The Journal of Consumer Aﬀairs, 25 (2), 358–90.
Human Fertilization and Embryology Authority (HFEA) (1998), ‘Seventh Annual Report
and Accounts’, www. hfea.gov.uk.
Klock, S.C., J.E. Stout and M. Davidson (2003), ‘Psychological Characteristics and Factors
Related to Willingness to Donate Again Among Anonymous Oocyte Donors’, Fertility and
Sterility, 79 (6), 1312–16.
Kolczykiewicz, M. (2003), ‘Risk and Compensation in Surrogacy Contracts’, mimeo.
Kossoudji, S.A. (1990), ‘Pride and Prejudice: Culture in the Labor Market and the Home’, in
Steven Shulman and William Darity (eds), The Question of Discrimination: Racial
Inequality in the U.S. Labor Market, Wesleyan University Press, pp. 293–314.
Lundberg, S. (2000), ‘Child Auctions in Nineteenth Century Sweden: An Analysis of Price
Diﬀerences’, Journal of Human Resources, 35 (2), 279–98.
Malthus, Thomas (1798), An Essay on the Principle of Population, London: J. Johnson in
St Paul’s Churchyard.
Mathews, Jay (1984), ‘Boy’s Birth is First from Embryo Transfer’, Washington Post,
4 February, A14.
New Jersey Supreme Court (1988), ‘In the Matter of Baby M, A Pseudonym for an Actual
Person’ (No. A-39. 109 N.J. 396; 537 A. 2d. 1227; 1988 N.J. Lexis 1; 77 A. L. R. 4th 1),
14 September, 1987, Argued, 3 February, 1988, Decided.
Plotz, D. (2001), ‘The Rise of the Smart Sperm Shopper’, 20 April, Slate website.
Posner, E.A. and R.A. Posner (1999), ‘The Demand for Human Cloning’, The Hofstra Law
Review, 27 (3), 579–608.
Radin, Margaret Jane (1996), Contested Commodities, Cambridge, MA: Harvard University
Press.
Rosenzweig, M.R. and T.P. Schultz (1985), ‘The Demand and Supply of Births: Fertility and
its Life Cycle Consequences’, American Economic Review, 75 (5), 992–1015.
Saint-Paul, G. (2002), ‘Economic Aspects of Human Cloning and Reprogenetics’, IZA
Discussion Paper No. 608.
Thorne, E.D. (1998), ‘When Private Parts are Made Public Goods: the Economics of Marketinalienability’, Yale Journal on Regulation, 15 (1), 149–75.
Tindall, Harry L. et al. (2000), ‘The Uniform Parentage Act 2000’, drafted by the National
Conference of Commissioners on Uniform State Laws, approved and recommended for
enactment by all states.
Titmuss, Richard M. (1997), ‘The Gift Relationship: From Human Blood to Social Policy’,
expanded edition, Ann Oakley and John Ashton (eds), New Press.
Zelizer, Viviana (1985), Pricing the Priceless Child: The Changing Social Value of Children,
New York: Basic Books.

11 Economics of abortion
Leo H. Kahane

Abortion is an old practice dating back (at least) several thousand years to
ancient Egyptian, Greek and Roman civilizations. The procedure continues to be carried out in virtually all societies across the globe and in modern
times has become a point of divisive, and at times explosive, social debate.
The moral and emotional dimensions of abortion tend to dominate the discussion surrounding the procedure, but over the last several decades economists have carried out research on such topics as the supply of and
demand for abortion, the eﬀects of the availability of abortion on crime
rates, how abortion has aﬀected the incidence of so-called ‘shotgun’ marriages, as well as others. The purpose of this chapter is to describe how the
methodology employed by economists has been put to work in analysing
the topic of abortion in a number of interesting ways.
The layout of this chapter is as follows. Section 1 contains a brief discussion of the procedure, providing some terminology and recent statistics
on trends in abortion rates throughout the world. This section is intended
to give the reader a sense of the size of the ‘market’ for abortion services as
well as prepare them for the discussion that comes later. The rest of the
chapter is divided up into several sections, each devoted to a particular type
of economic analysis. Section 2 focuses on the demand for and supply of
abortion services and demonstrates how these fundamental tools of economic analysis can be used to understand the workings of the market.
Section 3 follows with a discussion of how the availability of abortion services and their legality has had ramiﬁcations for other related issues including teen pregnancy rates, the propensity for pre-marital sexual relations,
ﬂuctuations in crime rates across time, as well as others. The chapter ends
with Section 4, which contains some concluding remarks.
1 Terminology and demographics of abortion
Induced abortion can be carried out at virtually any stage of pregnancy. The
earlier the procedure, the greater the variety (and simplicity) of methods
available to end the pregnancy. Pregnancies that are less than nine weeks can
be terminated with the relatively new drug, mifepristone, which was developed and made available in France in 1988 and became known as the
‘RU-486’ abortion pill. This drug was legalized in Great Britain in 1991, in
Sweden in 1992 and in the United States in 2000. More invasive is the method
315

316

Economics uncut

of uterine evacuation using a vacuum (typically for pregnancies up to
14 weeks in duration). Perhaps the most controversial method of abortion
is that of intact dilation and extraction, often referred to as ‘partial-birth’
abortion in the USA, which involves dilation of the cervix, destruction and
removal of the fetus (typically for pregnancies that are 13 weeks or older).
1.1 Abortion usage
As for trends, the use of abortion services varies greatly across countries.
Table 11.1 presents data published by the Alan Guttmacher Institute (AGI)
Table 11.1 Worldwide estimated abortions and abortion rates (legal and
illegal) (1995)
Abortions
(millions)

Abortion rate (per 1000
women aged 15–44)

45.5
10.0
35.5

35
39
34

5.0
1.9
0.6
0.6
0.2
1.6

33
41
35
17
19
37

26.8
12.5
8.4
4.7
1.2

33
36
28
40
32

Europe
Eastern Europe
Northern Europe
Southern Europe
Western Europe

7.7
6.2
0.4
0.8
0.4

48
90
18
24
11

Latin America & Caribbean
Caribbean
Central America
South America

4.2
0.4
0.9
3.0

37
50
30
39

United States & Canada

1.5

Oceania

0.1

World
Developed regions
Developing regions
Africa
East Africa
Middle Africa
North Africa
Southern Africa
West Africa
Asia
East Asia
South Central Asia
Southest Asia
West Asia

Source:

The Alan Guttmacher Institute (1999, Appendix Table 3, p. 53).

Economics of abortion

317

for the estimated number of abortions and the abortion rate in 1995 by
region.
As the table reveals, worldwide about 45.5 million abortions were performed in 1995, which translates into an abortion rate of 35 abortions per
1000 women of childbearing age (aged 15–44). Abortion rates are typically
highest among women in their early 20s and lowest for women in their 40s.
However, deﬁning the abortion ratio as the number of abortions per 100
pregnancies, the data suggest that the relationship between this measure
and a woman’s age is ‘U’-shaped. That is, women who are relatively young
(for example, less than 20 years old) or relatively old (for example, more
than 35 years old) tend to have a greater abortion ratio than those in
between (AGI, 1999).
As for geographical diﬀerences, the developed regions’ abortion rate of
39 is notably higher than the rate of 34 for developing regions. The
African continent and Asia have the lowest abortion rates (33) whereas
Europe has the highest (48), the latter being noticeably aﬀected by the
high Eastern European rate of 90 abortions per thousand women of childbearing age.1
Abortion trends across time diﬀer substantially across countries.
Throughout much of the world there appears to be a trend towards a lower
abortion rate from 1990 to 1996 (AGI, 1999). Table 11.2 reports annual
ﬁgures for abortions and the abortion rate for England and Wales and the
United States and Figure 11.1 plots these abortion rates across time (and
includes simple estimated quadratic trends).
As the table and associated ﬁgure show, in all locations the numbers and
rates increase substantially shortly after court decisions making abortion
legal nationally (the Abortion Act 1967 for England and Wales, which came
into eﬀect on 27 April 1968 and the Roe v. Wade US Supreme Court decision in 1973). In the USA, the rate peaks in 1980–81 (with a rate of 29.3)
and then declines fairly steadily throughout the following years. The time
path for England and Wales is quite diﬀerent from that of the USA as the
abortion rate, following noticeable dips from 1973 to 1977 and again from
1990 to 1995, has shown a rather consistent upward, slowly diminishing
trend. It is interesting to note that the gap between the two rates appears to
be narrowing.
1.2 Availability of abortion services
Regarding the availability of legal abortion services, this also varies
greatly across countries and time.2 According to Rahman et al. (1998),
about 61 per cent of the world’s population live in countries that
have legalized abortion. Approximately 25 per cent live in countries that
prohibit abortions under most circumstances. Concerning numbers of

318

Economics uncut

Table 11.2

Year

1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000

Abortions and abortion rates: England and Wales and USA
(1969–2000)
Number of abortions
(thousands)

Abortion rate (abortions per
1000 women aged 15–44)

England &
Wales*

United
States**

England &
Wales*

United
States**

49.8
76.0
94.6
108.6
110.6
109.4
106.2
101.9
102.7
111.9
120.6
128.9
128.6
128.6
127.4
136.4
141.1
147.6
156.2
168.3
170.5
173.9
167.4
160.5
157.8
156.5
154.3
167.9
170.1
177.9
173.7
175.5

–
–
–
–
744.6
898.6
1034.2
1179.3
1316.7
1409.6
1497.7
1553.9
1577.3
1573.9
1575
1577.2
1588.6
1574
1559.1
1590.8
1567
1609
1556.5
1528.9
1495
1423
1359.4
1360.2
1335
1319
1314.8
1313.0

5.3
8.1
10.1
11.5
11.7
11.5
11.1
10.5
10.5
11.3
12.0
12.6
12.4
12.3
12.1
12.8
13.1
13.5
14.2
15.3
15.5
15.8
15.2
14.8
14.7
14.6
14.4
15.6
15.8
16.5
16.1
16.1

–
–
–
–
16.3
19.3
21.7
24.2
26.4
27.7
28.8
29.3
29.3
28.8
28.5
28.1
28.0
27.4
26.9
27.3
26.8
27.4
26.3
25.7
25.0
23.7
22.5
22.4
21.9
21.5
21.4
21.3

Sources:
*Oﬃce for National Statistics: Abortion Statistics 2001, England & Wales, series AB no. 28.
**The Alan Guttmacher Institute (2003, Table 1, p. 8). Values in italics are estimated by
interpolation.

Economics of abortion

319

Abortion rate

25
20
15
10
5
0
1969

1974

1979

1984

1989

1994

1999

Year
United States

England & Wales

Quadratic trend
(United States)

Quadratic trend
(England & Wales)

Note: Abortion rate is deﬁned as the estimated number of abortions per 1000 women aged
15–44 years. Estimated equations for the quadratic trends are:
Abortion in US 194 977 196.4(Year) 0.0494(Year)2, R2 0.765.
Abortion in England & Wales23 228 23.175(Year) 0.0058(Year)2, R2 0.861.

Figure 11.1 Abortion rates for England and Wales and USA (1969–2000)
providers in countries, data are scarce. However, for the USA, available
data show that there has been a marked downward trend. Data for the
period 1992 to 1996 show a decline in the number of providers by approximately 14 per cent, nationwide, and an 11 per cent decline from 1996 to
2000. Compared with 1982, when the number of providers in the USA
was at its peak, the ﬁgure for 2000 shows a decline of approximately 37 per
cent during the entire period (Finer and Henshaw, 2003a). As for the
reasons for such a decline, there may be several, including greater awareness of contraception and the decreased demand and discouragement of
supply due to the activities of those who oppose abortion. We discuss the

320

Economics uncut

eﬀects of anti-abortion activities on the supply and demand of abortion
in Section 2.
Concerning the kinds of providers, there are three basic types in the
United States: hospitals, clinics and physicians’ oﬃces. According to Finer
and Henshaw (2003b), a survey of 1819 abortion providers in 2000 yielded
the following observations:
1.
2.
3.

Approximately 95 per cent of all abortions are performed in nonhospital facilities.
The mean price for an abortion (over all types of facilities) at a gestation of ten weeks is US $468 (the median being US $370).3
The price of an abortion rises as the gestation increases. For example,
the mean price for an abortion (over all types of facilities) at ten weeks’
gestation is US $468, at 16 weeks’ gestation is US $774 and at 20 weeks’
is US $1179. The principal reason for this positive relationship between
gestation and price is that for a greater gestation, the abortion procedure becomes more complicated, time-consuming and requires greater
skill. Thus, these increased costs drive up the price for the procedure.
As the gestation increases, fewer facilities oﬀer abortions. For example,
90 per cent of all surveyed providers oﬀer abortion services for pregnancies of eight to ten weeks in gestation. However, only 33 per cent
will oﬀer abortion services for a gestation of 20 weeks. This inverse
relationship is mostly due to the fact that as the gestation increases and
the procedure then becomes more complicated (see the last point),
fewer providers have the expertise and facilities capable of performing
these procedures.
In 2001, providers with caseloads of less than 30 abortions per year had
a mean charge of US $787 for an abortion whereas providers with a
case load between 400 to 990 abortions per year had a mean of US
$368 per abortion. The diﬀerence implies some economies to scale.
These economies, however, apparently diminish rapidly as providers of
5000 or more abortions had a mean price of US $356 per abortion.
Approximately 8 per cent of women having abortions in non-hospital
facilities had to travel more than 100 miles to the service provider.
About 16 per cent travelled between 50 to 100 miles. Travel distance can
be a signiﬁcant barrier to the service as it adds to the overall cost of an
abortion. The cost component becomes even more important if the
client must make more than one trip. For example, as of 2001, four
states in the USA (Louisiana, Mississippi, Utah and Wisconsin) had
legislation requiring counselling between the client and the attending
or referring physician at least 24 hours prior to the abortion (Finer and
Henshaw, 2003b).

Economics of abortion

321

Abortion as an outcome: demand, supply and the market for abortion
services
The powerful tools of economic theory and empirical methods (most
notably econometrics) have been brought to bear on the topic of the incidence of abortion in essentially two ways: one with abortion on the lefthand side of the equals sign in an equation, and one on the right-hand side.
In the ﬁrst case, researchers have constructed theoretical models of abortion demand (and, to a lesser degree, supply). In these models, abortion
demand is the dependent variable to be explained and is hypothesized to
be a function of various explanatory variables. In the research where abortion appears on the right-hand side of the equals sign (that is, as an
explanatory variable), researchers consider how the incidence of abortion
may be a determining factor of something else. This section deals with the
former, having the incidence of abortion as the dependent variable in
research. The study of abortion rates as an explanatory measure is the
topic of Section 3.
2.1 The basic model of demand (and supply)
The demand for and, to a lesser degree, the supply of abortion services has
been researched by many authors (see, for example, Brown and Jewell,
1996; Garbacz, 1990; Haas-Wilson, 1996; Kahane, 2000; Medoﬀ, 1988,
2000). In the case of demand, the antecedent theory draws mainly from
Becker (1960, 1965), Mincer (1962, 1963) and the later work of Michael
(1973). In these works, fertility control is governed by the expected net
beneﬁt of the birth of an additional child. If the net beneﬁt of an additional
child is positive, then a fecund couple may seek to increase the number of
children they currently have. If the net beneﬁt of an additional child is perceived to be negative, then a couple may seek to prevent the birth of a child
using a number of birth control methods, including abortion.4 Thus, when
modelling the demand for abortion, both direct and implicit costs (or
opportunity costs) are considered. The typical demand function for abortion services looks something like the following:
Abortion Demand f (Price, Income, Education, Moral
or Cultural Diﬀerences, Marital Status, Other Measures)

(11.1)

Given the lack of data on individual consumers of the service, the bulk
of the empirical research considers aggregate demand, usually across time
and/or regions. The most common approach is to use state-level data
(a single cross-section or panel) for the USA to determine the factors
explaining the diﬀerences in abortion demand across the states. Indeed, it
appears that virtually all of the empirical research estimating demand has

322

Economics uncut

been done for the markets in the USA. An exception is one of the earliest
empirical studies by Coelen and McIntyre (1978) who model abortion
demand in Hungary as part of a simultaneous equations model, which
includes the demand for births. Their primary conclusion regarding abortion demand is that pro-natalist policies enacted in Hungary in the late
1950s to late 1970s reduced the demand for abortion, while reductions in
abortion costs increased the abortion rate.
Like the Coelen and McIntyre study, most research by economists on
abortion does recognize the simultaneous determination of equilibrium
price and quantity and thus treats price as endogenous and an instrumental variable method is used to estimate supply and demand simultaneously.5
The typical equation for supply looks something like the following:
Abortion Supplyf (Price, No. of Physicians Relative to the
Population of Women of Child-Bearing Age, Average Hospital
Costs, Average Wages of Employees in Physicians’ Oﬃces) (11.2)
In equation (11.2), supply is assumed to be positively related to the price of
an abortion and the availability of those capable of performing an abortion, but negatively related to input costs proxied by the average cost of a
day in the hospital and the wages of employees working in physicians’
oﬃces. In some cases, average travel distance to a provider or the geographic concentration of providers is used to capture supply.
The most common measure of abortion demand and supply is the aforementioned abortion rate, equal to the number of abortions per 1000
women of childbearing age (15–44 years old). A second measure frequently
used, referred to as the abortion ratio, is the number of abortions per 1000
pregnancies of women of childbearing age. There are slightly diﬀerent
interpretations between these two measures. In the case of the former,
changes in the abortion rate may be due to a variety of factors, including
greater knowledge and usage of contraception. In the latter, however, contraception is not an issue since in this case conception has already occurred.
In any case, the bulk of the empirical research ﬁnds qualitatively similar
results for the two measures of abortion demand.
There are several points of consistency for much of the research on abortion demand; Table 11.3 provides a summary of some selected results. All
of the research ﬁnds that the law of demand (that is, there is an inverse relationship between the price of an abortion and the quantity of abortions
performed) is upheld.6 Most research ﬁnds the estimated price elasticity of
demand to be less than one (in absolute terms) which, as microeconomics
teaches us, is not usually consistent with proﬁt maximization.7 Given the
nature of abortion and the kinds of institutions that provide the service, the

Economics of abortion

323

Table 11.3 Summary of estimated price and income elasticities of
abortion demand for selected research
Article
Garbacz (1990)
Gohmann & Ohsfeldt (1993)
Kahane (2000)
Medoﬀ (1988)
Medoﬀ (2000)

Price elasticity

Income elasticity

0.68
0.91*
0.75
0.81
0.62

0.84
0.76*
1.96
0.79
1.25

Notes:
Demand is measured using the abortion rate, that is, number of abortions per 1000 women,
aged 15–44.
*Gohmann and Ohsfeldt (1993) find a range of price elasticities, some greater than 1,
for various models estimated. Their findings indicate that the price elasticity appears to
be sensitive to model specification. The value reported in this table is for their pooled,
two-stage least-squares model.

less than a proﬁt-maximizing price is likely intended to make the service of
abortion available to a wider consumer base. It should be noted, however,
that a sizeable portion of the industry is not-for-proﬁt, which includes
Planned Parenthood, one of the largest groups of providers in the USA.
Income is present in most of the studies and the results are remarkably
consistent: abortion is a ‘normal good’ (or service) as income repeatedly
turns out to be positive and statistically signiﬁcant. This result is consistent
with the hypothesis coming from earlier theory on fertility (noted above)
that women earning greater income and who become pregnant may ﬁnd that
the opportunity cost in terms of lost income (to them or their family) from
having an additional child is too great and as such may choose to terminate
a pregnancy in order to avoid those costs. Some studies (for example, HaasWilson, 1996; Kahane, 2000; Medoﬀ, 1988) include women’s labour force
participation rates as another proxy for opportunity costs due to childbirth.
Education is also present in many studies, but in this case opposing theories exist. Education, typically measured as the percentage of women of
childbearing age with a high-school degree (or perhaps college education),
may serve as another proxy of opportunity costs to childbirth. Women with
greater education may have more satisfying careers that may have to be put
on hold in the event of childbirth. This opportunity cost may encourage
women with greater education to terminate their pregnancies so as to avoid
any stoppage in their career, thus we would expect the demand for abortion
to be positively related to education level. On the other hand, women who
are better educated are more likely to be better aware of methods of

324

Economics uncut

contraception and would experience fewer unwanted pregnancies, thus
reducing the abortion rate. The ultimate, net eﬀect is an empirical question.
The research by Kahane (2000) and Medoﬀ (2000) come to diﬀerent conclusions with the former ﬁnding a positive yet insigniﬁcant relationship and
the latter ﬁnding a negative and weakly signiﬁcant relationship.8
In order to consider how diﬀerences in moral or cultural views of abortion
aﬀect the demand for abortion, several proxies have been employed. Two
controls that are commonly used are the racial composition (for example,
the percentage of a state’s population that is non-white) and relative membership in certain religions that openly oppose abortion (for example, the
numbers of Catholic, Southern Baptists, Evangelists and Mormons, relative
to state population). The former proxy is motivated by the fact that,
controlling for income diﬀerences, non-white women have historically
greater abortion rates, implying some cultural diﬀerences regarding the issue
of abortion. This control for race typically returns a positive and signiﬁcant
coeﬃcient in abortion demand regressions. As for the control for religion,
one would expect a negative coeﬃcient. The reasoning behind this expectation has to do with the idea that individuals who are adherents to a religious
faith that opposes abortion will more likely themselves have a moral aversion
to, and hence a lower demand for, abortion services, all else equal. Empirical
results, however, typically show insigniﬁcant eﬀects of religious membership
on abortion demand.
Last on the list of common controls is the measure of marital status. It
is argued that single women are more likely to terminate a pregnancy since
it is expected that single mothers would likely bear greater costs (direct and
indirect) to having and caring for a child than married mothers, all else
equal. This control, however, has found only weak empirical support.9
2.2 Public subsidies, consent laws and issues of access
One of the common variables to appear in the category of ‘other measures’
has to do with public subsidies for abortion. Following the 1973 Roe v.
Wade US Supreme Court case, which made abortion legal in the United
States, the US Congress passed the Hyde Amendment in 1976, which prohibited the use of federal funds (through the federal medical insurance
programme, Medicaid) to pay for abortions under most circumstances.10
A collection of states, however, decided to provide state funds for the procedure under most circumstances.11 Given this ‘natural experiment’, a
number of researchers have investigated the eﬀects of how diﬀering access
to public funds aﬀects abortion rates across states. The predicted eﬀect
is straightforward: public funds serve as a subsidy for the service and, ceteris
paribus, this would increase the demand for the service. The increase would
arise due to substitution eﬀects as well as perhaps income eﬀects. That is,

Economics of abortion

325

the eﬀect of the subsidy would be the possible switching of women from
other contraceptive measures (including abstinence) to using abortion as a
means of preventing unwanted pregnancies. In addition, the subsidy would
lower costs, thus making the procedure more aﬀordable and possibly cause
an increase in demand. These predictions are borne out, albeit weakly, by
the research. The results of most of the statistical analyses that control for
availability of state funding of abortions ﬁnd a direct relationship between
states that do provide funds and the abortion rate. The statistical robustness
of this result, however, is somewhat mixed: Levine et al. (1995), Medoﬀ
(1988, 2000) and Haas-Wilson (1996), for example, ﬁnd a statistically
signiﬁcant, positive relationship, whereas Haas-Wilson (1996) and Kahane
(2000) ﬁnd a positive, yet statistically insigniﬁcant relationship. Given the
consistency across studies of positive relationships, however, one may
presume that public funds tend to increase abortion rates, all else equal.
Another issue that lands itself into the ‘other measures’ category has to
do with legal impediments to abortion access for minors that have been put
in place by state legislatures. While the 1973 Roe v. Wade decision interpreted the constitutional right to privacy to include a woman’s right to
terminate a pregnancy, the Supreme Court did permit states to enact various
laws restricting or regulating a woman’s access to abortion. The 1989
Webster v. Reproductive Health Services (492, US 490) case and the 1992
Planned Parenthood of Southern Pennsylvania v. Casey (112, S. Ct. 2791)
case allowed states to enact laws requiring parental consent for minors and
mandatory 24-hour waiting periods for women seeking abortions when
these laws do not pose an ‘undue burden’ on women. As of 1993, 23 states
had laws in place (NARAL, 1993).12 The expected impact on abortion
demand is obvious: the laws are intended to reduce the number of abortions
by increasing the opportunity cost of an abortion both emotionally (by
involving the minor’s parents in the decision) and legally (in the case where
a judicial bypass of the law is sought).
The empirical question presented by these access laws is whether they
have been successful in signiﬁcantly reducing the number of abortions performed. The results of various research papers on this issue are, for the
most part, in agreement: parental consent laws do indeed reduce abortion
rates. As an example, Ohsfeldt and Gohmann (1994) write, ‘Parental
involvement laws appear to reduce the adolescent abortion rate relative to
the abortion rate of older teens or adults not subject to the laws. “Best
Model” estimates imply that such laws reduce the adolescent abortion rate
by about 18 percent, other things equal’. Haas-Wilson (1996) concurs, estimating a reduction between 13 to 25 per cent.
Another of the ‘other measures’ considered has to do with the eﬀects of
anti-abortion activities. Several of the papers cited earlier (for example,

326

Economics uncut

Haas-Wilson, 1996; Gohmann and Ohsfeldt, 1993) estimate ﬁxed-eﬀect
models in an attempt to control for ‘unobservable eﬀects’ such as antiabortion sentiment that may diﬀer across states. Kahane (2000), however,
attempts to capture these eﬀects more directly. Using cross-section data for
the 50 states in the USA in 1992, he estimates both supply and demand
functions using a two-stage least-squares procedure. An independent
variable included in both functions is the percentage of clinics that had
experienced picketing with physical blocking or contact between clients
and protestors or service providers and protestors.13
The hypothesized eﬀect of anti-abortion activities is rather straightforward: other things equal, states where clinics experience greater
900
800

700

Price

600
500
S

400
300

200

100
0
20 20.4 20.8 21.2 21.6 22 22.4 22.8 23.2 23.6 24 24.4 24.8 25.2 25.6
Abortion rate
Note: The S and D are the location of the curves when the eﬀects of anti-abortion
activities are included in the regression. The label e is the new equilibrium after the shifts,
with e being the equilibrium before the shifts.
Source:

Kahane (2000).

Figure 11.2 The estimated eﬀects of anti-abortion activities on the supply
and demand curves for abortion services

Economics of abortion

327

anti-abortion activities (as measured) would tend to have a reduced
demand and supply. Under this assumption, both the demand and supply
curve shifting left would work to reduce the equilibrium quantity of abortions performed, with the equilibrium price being indeterminate and
depending on the relative size of the shifts in the supply and demand curves
as well as their slopes. The estimated equations suggest that anti-abortion
activities did signiﬁcantly reduce both supply and demand leading to a
reduction in the equilibrium quantity by about 19 per cent and raising price
by about 4.3 per cent. The estimated demand and supply curves, with and
without the anti-abortion eﬀects, are shown in Figure 11.2.14
2.3 Other related issues of abortion demand
There are obviously a myriad of factors that govern the demand for abortion. Some of these reasons extend beyond the modelling of supply and
demand considered above. For example, China in the early 1980s attempted
to control population growth by mandating a one child per family policy
coupled with a forced abortion policy for any woman who had an unauthorized pregnancy and sterilization for couples with two or more children (see Graham et al., 1998). Further, in some countries (for example,
China, Korea, Bangladesh) male children are apparently more desirable
than female children. The reasoning behind this preference is that some
parents view their children as a source of support as the parents age,
coupled with the notion that male children are more likely to better support
their parents later in life than are female children. Furthermore, male children will carry on a family name. There is some evidence, in fact, that for
these reasons abortion has been used as a means for achieving such sex preferences of children (see, for example, Bairagi, 2001). Future research that
considers abortion demand across countries would need to incorporate
such cultural diﬀerences into the model.
3

Abortion as a cause: the effects of abortion availability on fertility,
premarital sex, shotgun weddings and crime
Another fascinating application of economics modelling using abortion is
where measures of abortion (usage or availability) enter as an explanatory
variable in models explaining other issues. That is, abortion appears on the
right-hand side of the equals sign. This section summarizes a few of the
more interesting research papers using this approach.
Key to analysis of this kind is the requirement that there be suﬃcient
variation of abortion access (and, hence, cost) in order to examine the
eﬀects of such changes in access on various dependent variables of interest.
Fortunately, from the standpoint of the researcher looking for data, a
variety of policy changes aﬀecting abortion access have occurred in the

328

Economics uncut

USA over the past four decades. Furthermore, there were signiﬁcant
changes in abortion policy in other parts of the world, especially in Eastern
Europe during the 1950s. These events have created the kind of ‘natural
experiment’ necessary for the testing of several interesting hypotheses
described below.
3.1 Abortion access and fertility
The availability of abortion services increases the options women face
regarding pregnancy resolution. Faced with the reality of being pregnant,
a woman who chooses to abort the pregnancy may have instead given birth
had abortion services not been legally available (with illegal abortion
remaining a possibility). Thus, the seemingly obvious result is that fertility
rates would tend to fall in places where abortions have become legal, other
things equal. Furthermore, a second consequence of legalized abortion
may be that the use of contraception may decline as abortion becomes a
substitute for contraception. Both of these issues were, in fact, studied in
an early research paper by Frejka published in 1983.
Following the Soviet Union, where abortion was legalized in 1955,
many of the countries in Eastern Europe made abortion legal a few years
later. Using data on abortion rates in Eastern Europe from the mid-1950s
to the late 1970s, Frejka analyses the eﬀects of abortion liberalization on
fertility rates and use of contraception. He concludes that the greater
availability of abortions led to an increase in abortion use, a signiﬁcant
drop in fertility and a reduced usage of contraception. In one of his more
striking examples, Frejka notes that for Romania the number of annual
abortions increased from 578 000 to 1 115 000 between the years 1959 to
1965 (abortion was made available on demand in Romania in 1957).
During the same period, the number of annual births declined by approximately 90 000.15
In the United States, beginning with the legalization of abortion in the
state of Colorado in 1967, a total of 19 other states had followed suit by
1972 (Klerman, 1999). The 1973 Roe v. Wade US Supreme Court decision
made abortion legal in all states, thus increasing overall access. The subsequent passage of the Hyde Amendment in 1976 (see Section 2 above),
however, eﬀectively allowed the decision of whether public funds will be
used to subsidize the cost of an abortion to be made at the state level. This
has resulted in diﬀering restrictions on the public ﬁnancing of abortions
across states. Finally, the enactment of teen parental consent laws in some
states, which were put in place and enforced beginning in the late 1980s to
early 1990s, has led to variation in abortion access. Jacob Klerman, in his
1999 article, uses these policy changes to explore the eﬀects of abortion
access on fertility of women in the USA.

Economics of abortion

329

Employing individual-level birth certiﬁcate data, Klerman considers
how the eﬀects of abortion legalization and variation in Medicaid funding
may have aﬀected fertility rates across diﬀerent racial and age groups.
Using a ‘diﬀerence-of-diﬀerence’16 regression approach he ﬁnds some interesting results, including the following: in the case of legalization, legal
access led to a moderate reduction in fertility among white women of about
2 per cent (with a greater reduction for women in their 20s as compared
with women in their 30s). The data for black women revealed no statistically signiﬁcant pattern resulting from abortion legalization.
With regards to the variation in Medicaid funding for abortion, Klerman
ﬁnds that the changes in the availability occurring in the period 1982–9217
led to the following eﬀects: availability of Medicaid funding had a large,
negative eﬀect on the fertility of black women (approximately a 10 per cent
reduction in births) with an even larger eﬀect on higher-order births
to women in their 20s (approximately a 15 per cent reduction). The pattern
is similar for white women, yet smaller in size (for example, an estimated
3 per cent reduction for higher-order births to women in their 20s).
Finally, using his estimated model, and disaggregated data for 1992 (the
last year’s data in his dataset) Klerman estimates the eﬀects of elimination
of legalization and Medicaid funding on the total fertility rate (TFR).
Starting with a base case of legalization and full funding he ﬁnds that for
white women, ending Medicaid funding would increase the TFR by about
2 per cent and making abortion illegal would increase the TFR by an additional 3 per cent. For blacks, the numbers are considerably larger: ending
Medicaid funding would lead to a predicted 10 per cent increase in the TFR
and making abortion illegal would lead to an additional 5 per cent increase
in the TFR.
3.2 Endogenous pregnancy
One of the implications of Frejka’s 1983 work is that the variation in the
availability of abortion services may lead to changes in the sexual behaviour of women, such as the usage of contraception. This idea can be
extended to the very issue of pregnancy itself. That is, contrary to the
perhaps more common view where pregnancy is treated as exogenous and
the pregnancy resolution is then considered, pregnancy may in fact be
endogenous. Thomas Kane and Douglas Staiger explored this issue in the
1996 research paper on teenage motherhood and abortion.
The hypotheses developed in Kane and Staiger (1996) are, in many ways,
intuitive and are nicely summarized in the following passage from their
paper:
Our simple model assumes that women get information during the early months
of pregnancy, and abort the pregnancy if the birth turns out to be unwanted

330

Economics uncut

based on this new information. Given the majority of teen pregnancies are conceived out of wedlock, the father’s willingness to marry is an obvious example
of such information. Contraception and abstinence decisions are made only on
the basis of information available before pregnancy occurs. In contrast, the abortion decision is made with the beneﬁt of new information. Abortion (unlike contraception or abstinence) works as an insurance policy to limit the downside risk
when that information is negative. Increasing the cost of abortion increases the
cost of this insurance policy and discourages women from becoming pregnant.
Some of these pregnancies would have resulted in births, so the model implies
that an increase in the cost of abortion results in a decline of wanted births. Of
course, there is a second more conventional eﬀect as the increased cost of an
abortion discourages some women from aborting unwanted pregnancies. Thus,
the net eﬀect of any restriction of abortion access on birthrates is ambiguous
and, in the end, an empirical question.

The dynamics of the above hypothesis are represented diagrammatically
in Figure 11.3. The ﬁgure shows a simple, sequential decision-making
scheme, which begins with the question of whether a woman is involved in
sexual relations. Given the woman is having sex, the possibility of pregnancy arises. If a woman having sex has become pregnant, then a decision
must be made of carrying the pregnancy to full-term and giving birth or
aborting the pregnancy, if abortion is easily available (the dashed arrow represents this latter conditionality). It is at this point in the scheme, where a
decision to proceed to birth or abort the pregnancy is needed, that the
woman may gain information that could aﬀect her decision. According to
Kane and Staiger, easy access to abortion gives women a low-cost ‘insurance policy’, which reduces the risk associated with having sex, that risk
being birth from an unwanted pregnancy. This reduced risk then in fact may
increase the frequency of sexual relations, represented by the dashed,
curved line ﬂowing back to the top of the diagram. Greater sexual relations
would likely lead to greater unplanned pregnancies, of which some may
result in birth as women, at the decision juncture for birth vs. abortion,
receive new information that leads them to choose birth. Thus, we have an
example of an ‘insurance policy’ leading to moral hazard. At the same time,
if the option of abortion becomes more costly (or even eliminated), the
‘encouragement’ factor of abortion (that is, the dashed curve ﬂowing back
to the sex decision) is lessened (or eliminated), thus reducing sexual activity, reducing the occurrence of unplanned pregnancies and this may work
to reduce the birthrate. At the same time, those unplanned pregnancies that
do occur are now more likely to end up in birth. Thus, these two factors work
in opposite directions and the end result on the birthrate is ambiguous.
In order to test the above hypotheses, Kane and Staiger use US countylevel data for 14 years (1973–88) and consider the eﬀects on the above
dynamics of changes in the geographic availability of abortion (as proxied

Economics of abortion

331

Sex?

Pregnancy?

Abortion?

Birth?

Figure 11.3 Sequential decision-making of pregnancy and pregnancy
resolution

by travel distance to the nearest provider), the availability of Medicaid
funding and the imposition of parental consent laws for minors on the
birthrates of teenage mothers.18 They ﬁnd that restricting access to abortion leads to a small, but statistically signiﬁcant decline in the teen birthrate,
with the bulk of the decline occurring for in-wedlock mothers. As an
example of their results, they ﬁnd that by increasing the distance to the
nearest provider by 25 miles leads to an estimated 1 per cent reduction in
the teen birthrate among in-wedlock mothers.19
3.3 Abortion and shotgun weddings
The discussion in the previous section touched on the possibility that the
availability of abortion may result in the change in behaviour of women
with regard to their sexual activity. This issue is also the topic of a very
interesting (and often cited) 1996 paper by Akerlof et al. The motivation

332

Economics uncut

for their paper can be understood by considering a few numbers. Between
the periods 1965–69 to 1980–84, the percentage of out-of-wedlock births
increased by 154 per cent for whites and by 64 per cent for blacks. Over the
same periods, the incidence of ‘shotgun weddings’ (that is, marriage occurring after pregnancy begins, but before the birth of the child) decreased by
25 per cent for white women and by 48 per cent for blacks.20 Figure 11.4
shows these trends over the entire period, at ﬁve-year increments. The questions that emerge are whether there is a linkage between these trends, and
what might be driving these changes.
The response by Akerlof et al. to these questions is that the availability
of abortion may be (at least partly) responsible. During the same periods,
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
1965–69

1970–74 1975–79
Years

1980–84

% Out-of-wedlock (white)
% Out-of-wedlock (black)
% Shotgun (white)
% Shotgun (black)
Source: Tables I and II from Akerlof et al. (1996).

Figure 11.4 Out-of-wedlock births and shotgun weddings, by race
(1965–69 to 1980–84)

Economics of abortion

333

the number of abortions among unmarried women aged 15 to 44 increased
from 88 000 in 1965–69 to 1.27 million in the period 1980–84. How may
these trends be related? The relationship can be understood by viewing
Figure 11.5 (which is a more elaborate version of Figure 11.3). The sequential decision-making process begins with the woman’s decision of whether
to engage in premarital sex or not. An aﬃrmative answer leads us down the
tree to the potential outcome where a woman has become pregnant and
then must consider carrying the pregnancy to term or aborting it, if abortion is easily available. According to Akerlof et al., the legalization of abortion led to a change in the ability of a woman to withhold premarital sex
from men. This is due to several forces. First, women who are willing to
choose abortion in light of pregnancy are perhaps more willing to engage
in premarital sex (here, abortion acts as the ‘insurance policy’ referred to by
Kane and Staiger, 1996). Second, these women may be less likely to demand
a promise to marry, in the event of pregnancy, as a precondition for premarital sex. Both of these possibilities then put pressure on other women
because of competition from women who do not require a promise to
marry. As Akerlof et al. put it, ‘those women who are not willing to use contraception or obtain an abortion will also engage in [premarital] sexual
activity, since they correctly fear that if they abstain their partners would
seek satisfaction elsewhere’.
Adding to this, Akerlof et al. point to the possibility that a man, with the
availability of abortion, may feel less compelled to marry a woman who
has become pregnant. That is, according to Akerlof et al., ‘the man reasons:
“If she is not willing to obtain an abortion or use contraception, why
should I sacriﬁce myself to get married?” ’ Thus, their model is consistent
with the observed decrease in shotgun weddings.
According to Akerlof et al., the ‘technology shock’ of the availability of
contraceptives and particularly of abortion (shown by the dashed arrow in
Figure 11.5) may have had the eﬀect of increasing the incidence of premarital sex (shown by the dashed curve in Figure 11.5), increasing out-ofwedlock births and reducing shotgun weddings. Their model’s predictions
are consistent with the historical data. Furthermore, they are somewhat
supported from survey results conducted with students.
3.4 Abortion and crime
As discussed in Chapter 4, the USA witnessed a marked drop in crime
during the early 1990s. Tracking the incidence of crime from 1973 to 1991,
Donohue and Levitt (2001) ﬁnd that violent crime increased by about
80 per cent, property crime increased by about 40 per cent with the murder
rate being essentially unchanged. Following 1991, however, these rates were
dramatically reduced by about 30 per cent in the ﬁrst two categories, and

334
Yes

Yes

No birth

Pregnant?

A similar version of this ﬁgure appears in Lundberg and Plotnick (1990).

No birth

Yes

Premarital
birth

Abortion?

Marriage?

Figure 11.5 Sequential decision-making of premarital sex, abortion, marriage and childbirth

Note:

No birth

Premarital sex?

Postmarital
birth

Yes

Economics of abortion

335

by approximately 40 per cent for murder. In search for a reason for this
dramatic turnaround in crime rates, Donohue and Levitt (2001) postulate
that a signiﬁcant portion of the decrease may be due to the legalization of
abortion. The mechanism by which abortion may reduce crime has two
components: increased abortion may reduce the size of future cohorts of
teenagers (who are more likely to commit such crimes), a sort of ‘volume’
eﬀect. Second, it may reduce the birthrate of children who may be born into
an environment that produces teenagers who have a greater propensity to
commit crime, a sort of ‘quality’ eﬀect. Their basic hypothesis on the latter
eﬀect is summarized in the following quotation from the Donohue and
Levitt paper:
children born after abortion legalization may on average have lower subsequent
rates of criminality for either of two reasons. First, women who have abortions
are those most at risk to give birth to children who would engage in criminal
activity. Teenagers, unmarried women, and economically disadvantaged are all
substantially more likely to seek abortions. Recent studies have found children
born to these mothers to be at high risk for committing crime in adolescence . . .
Second, women may use abortion to optimize the timing of childbearing . . .
[with the result being that] children are born into better [home] environments
and future criminality is likely to be reduced.

To say the least, this hypothesis has met with considerable controversy.
In support of their position that in the absence of abortion, children that
would have been born would have faced more diﬃculties, Donohue and
Levitt appeal to the research of others, notably the work of Gruber et al.
(1999). These authors ﬁnd that the marginal child not born because of
abortion would have grown up in a considerably adverse environment, in
comparison with the average child. They estimate that the marginal child
would have been 60 per cent more likely to be brought up in a single-parent
household, would be 50 per cent more likely to be born into poverty, would
be 45 per cent more likely to live in a household receiving welfare and have
a 40 per cent greater chance of dying before the age of one (Gruber et al.,
1999). A recent related paper by Bitler and Zavodny (2002) ﬁnds a negative
relationship between legalization of abortion and incidence of child abuse.
Levitt and Donohue test their hypothesis empirically regressing statelevel crime on the ‘eﬀective legalized abortion rate’ (appropriately lagged)
and a variety of other control variables using data for the period of 1985 to
1997.21 This period is chosen so as to allow children born after legalization
of abortion to have time to become teenagers, an age where the commencement of criminal activity is most common. Under their theory then,
children born in 1985 or later would have faced the possibility of abortion
when they were conceived. Because these pregnancies were not aborted,
however, means that these children were more likely to be ‘wanted’ and were

336

Economics uncut

more likely raised in a better home environment, with the end result being
lower criminal behaviour. The ‘eﬀective legalized abortion rate’ is computed
as a weighted average of the abortion rate across previous years with the
weights being computed as the proportion of total arrests in a given year
(for a given crime category) attributed to the age group of the cohort in
question. Donohue and Levitt reason that such a measure is appropriate
since the eﬀects of the abortion rate on criminal activity (if any) would be
gradually felt as the cohort ages and enters into the age where criminal
activity typically begins.22
Their results are the following: in states where the abortion rate was
higher in the 1970s and 1980s, the drop in crime was greater than in states
where the abortion rate was lower. Further, in the ﬁve states (Alaska,
California, Hawaii, New York and Washington) that made abortion legal
in advance of the Roe v. Wade 1973 decision, the drop in crime commenced
earlier. These results are consistent with the theory that the increase in the
availability of abortion led to a drop in the crime rate. Donohue and Levitt
estimate that an increase in the eﬀective abortion rate by approximately one
standard deviation reduces violent crime by about 13 per cent, reduces
property crime by about 9 per cent and reduces the murder rate by about
12 per cent. Further, they note:
Extrapolating our results out of sample to a counterfactual in which abortion
remained illegal and the number of illegal abortions performed remained steady
at the 1960s level, we estimate that . . . crime was almost 15–25 percent lower in
1997 than it would have been absent legalized abortion.

An impressive result that seemingly provides an answer to the question of
why crime rates fell dramatically in the early 1990s.
The paper by Donohue and Levitt has come under scrutiny, however.
Several authors have responded with their own research that seems to controvert the work of Donohue and Levitt.23 In particular, the research by
Lott and Whitley (2001) not only challenges the ﬁndings of Donohue and
Levitt, but in fact ﬁnds in their own analysis that legalizing abortion has
had no signiﬁcant eﬀect on crime with the exception of murder, which they
estimate may have increased following legalization of abortion.
The centrepiece to the theoretical counter-argument in the paper by Lott
and Whitley is simply an appeal to the previously discussed research of
Akerlof et al. (1996). Recalling one of the theses of that paper, increased
availability of abortion may have the eﬀect of increasing the incidence of
out-of-wedlock births, the opposite prediction to that proposed by
Donohue and Levitt, which may thus have the eﬀect of an increase in criminal activity. With competing theories, each of which may be at work at the
same time, the net result is an empirical issue.

Economics of abortion

337

Lott and Whitley are also critical of the empirical execution of
Donohue and Levitt. Speciﬁcally, they argue that Donohue and Levitt’s
use of aggregated data along with their ‘eﬀective abortion rate’ measure,
inadequately links abortion and crime to speciﬁc cohorts of individuals
across time (see Note 22). Furthermore, Lott and Whitley criticize the
approach of Donohue and Levitt, who implicitly assume that no abortions
were being performed in states other than the early legalizers in the preRoe v. Wade period; Lott and Whitley provide evidence to the contrary. In
their own empirical analysis, which they argue more closely links cohorts
to crime and abortion24 and which includes abortion rates of states other
than the ﬁve early legalizers, they come to the following conclusion:
There are many factors that reduce murder rates, but the legalization of abortion
is not one of them. Of the over six thousand regressions that we estimated . . .
only one regression implied even a small reduction in murder rates. All the other
estimates implied signiﬁcant if very small to modest increases in murder rates:
legalizing abortion would increase murder rates by around 0.5 to 7 percent.

The work of Lott and Whitley does seem to cast doubt on the ﬁndings
of Donohue and Levitt. To date, Donohue and Levitt have not issued an
oﬃcial written response. They have apparently, however, managed to replicate the results of Lott and Whitley. They have also discovered that the Lott
and Whitley results are highly sensitive to the chosen regression model
employed (Lott and Whitley estimate Poisson regressions). Using other,
less restrictive regression models (for example, the negative binomial) and
not lumping everyone over 30 into one group (as Lott and Whitley do)
Donohue and Levitt ﬁnd that they get back to their original result, that
abortion had signiﬁcantly reduced crime.25 In summary, the evidence seems
to weigh in favour of the Donohue and Levitt thesis, but clearly more work
on this topic is warranted.
4 Conclusion
There are perhaps few social-political-economic issues more divisive than
that of abortion. On 5 November 2003, President George W. Bush signed
into law the Partial Birth Abortion Ban Act of 2003. The ban, which is
aimed at ending a controversial type of late-term abortion described in
Section 1, had earlier passed in the US House of Representatives and the
Senate in lopsided votes (282 to 139 in the House and 64 to 33 in the Senate).
Prior to signing the ban, President Bush noted that the bill represented, ‘an
important step toward building a culture of life in America’. Proponents of
abortion fear that the bill moves the nation one step closer to eliminating a
woman’s right to seek abortion. The ban has already met with a legal challenge as the day after the signing of the bill a federal judge in New York

338

Economics uncut

granted a temporary restraining order that will likely delay prosecution of
any cases; the bill will most likely end up in the US Supreme Court. This
emotionally charged debate over abortion rights, which at times has led to
drastic behaviour,26 is likely continue on so long as there are women who
seek abortion services and those who believe abortion is morally wrong.
With this backdrop, economists have applied the tools of economic
analysis to the subject of abortion in much the same way they are applied
to other issues in society. As this chapter has hopefully demonstrated,
objective application of economics modelling and econometric methods
has been successfully employed to increase the understanding of the market
for abortion itself, and the aﬀects of abortion access on other important
social and economic phenomena. Given the millennia-old practice of abortion is not likely to disappear anytime soon (regardless of legal impediments), the potential work of economists on this topic will continue as well.
Notes
1.
2.
3.
4.
5.

8.
9.

10.

See AGI (1999), Appendix Table 4, p. 54 for a more detailed, country-by-country breakdown of estimated abortion rates. Countries with particularly high estimated abortion
rates include (for 1996): Vietnam (83.3), Romania (78.0) and Cuba (77.7).
Lack of data on the number of illegal abortions performed and the number of illegal
providers precludes any reliable discussion of these services.
Gestation is calculated as the number of weeks since a woman’s last menstrual period.
Note that the decision to increase the number of children born to a mother may not
necessarily be a joint decision.
Medoﬀ (1988, 2000), Gohmann and Ohsfeldt (1993) and Kahane (2000) use the 2SLS
approach. Garbacz (1990) does not treat price as endogenous, which leaves his results
somewhat in question given that Hausman tests performed by others (for example,
Gohmann and Ohsfeldt, 1993) reject the null hypothesis that price is exogenous.
Haas-Wilson (1996) does not include price as an explanatory variable. This omission is
somewhat puzzling and may be due to the unavailability of data. In any event, she estimates a ﬁxed-eﬀects model in which state-speciﬁc eﬀects may, in part, control for price
diﬀerentials.
An example where a ﬁrm (or group of ﬁrms) may ﬁnd it proﬁt maximizing to not price
their good or service on the elastic portion of the demand curve occurs when ancillary
goods or services are also sold by the same ﬁrm. For example, it has been found in professional sports that most ticket prices for sporting events are price inelastic. The reason
why owners of teams may underprice tickets is that they want to increase attendance,
which in turn boosts sales of ancillary goods such as concessions and parking with a ﬁnal
result of greater proﬁts overall.
The discrepancy may be in part due to the fact that Kahane includes both education and
women’s labour force participation rates in the same regression, which perhaps introduces problems of high multicollinearity.
Mothers co-habitating with fathers (married or not) or other supportive partners could
also be used to compare with single mothers. Data on such living arrangements, however,
are not typically available and thus simple marital status data are used to proxy the case
of single vs. non-single motherhood.
The Hyde Amendment was challenged in the courts. Following a series of legal decisions
it ultimately went into eﬀect in 1980 (see Levine et al., 1995, for more details). The
Amendment allows for the use of federal funds in the case of rape, incest or if the
mother’s life is in danger.

Economics of abortion

339

11.

According to the National Abortion Rights Action League, or NARAL, 12 states provided funds for abortions in most circumstances as of 1993: Alaska, California,
Connecticut, Hawaii, Massachusetts, New Jersey, New York, North Carolina, Oregon,
Vermont, Washington and West Virginia (NARAL, 1993, p. 145).
12. See Ohsfeldt and Gohmann (1994) for a more detailed discussion of the state laws
restricting abortion. It is important to note that most states provide for a judicial bypass
for parental consent when such consent may lead to an undue burden on the minor.
13. The measure was constructed with the help of Stanley Henshaw of the Alan Guttmacher
Institute who conducted a survey of clinics inquiring about the kinds of anti-abortion
activities the clinics experienced, if any.
14. A number of anti-abortion activities were considered, including simple picketing
(without contact or blocking) to more drastic measures such as bomb threats. Neither
of these turned out to be statistically signiﬁcant. In the case of simple picketing, nearly
all clinics (about 85 per cent) experienced this activity and as such there was too
little variation in this measure across states. As for bomb threats, few clinics (about
25 per cent) experienced this activity, making the eﬀect diﬃcult to pick up in the regressions. Picketing with contact was somewhere in between (about 52 per cent) and thus
had enough variation across states to be found statistically signiﬁcant in the regressions.
15. The abortion ﬁgures for Romania, however, do include some spontaneous abortions
along with induced abortions, thus inﬂating the numbers. The dramatic relative increase
in abortions pre- and post-legalization, however, should be largely unaﬀected by this
inclusion.
16. A ‘diﬀerence-of-diﬀerence’ approach considers the diﬀerence in a measure (for example,
fertility rates) between two groups (for example, whites and blacks) before an event
(for example, legalization of abortion) and after. Then the change in these diﬀerences
before and after the event (or the ‘diﬀerence of diﬀerences’) is used as a dependent
variable in a regression.
17. Following the 1976 Hyde Amendment a group of states stopped funding abortions
immediately, however, two court cases compelled many of them to resume funding until
1980. Later, in 1982 through 1992, an additional ﬁve states changed their policy on
Medicaid funding of abortions. Klerman (1999) analyses the impact of these funding
changes over both periods (1977–81 and 1982–92). He ﬁnds no signiﬁcant eﬀects on fertility from funding policy changes for the earlier period.
18. A variety of other controls are implemented including controls for county economic
measures, year and county-level ﬁxed eﬀects.
19. Similarly, negative results are found for the imposition of parental consent laws and
Medicaid funding. Kane and Staiger note, however, that in these cases it may be that the
results may simply reﬂect the already occurring downward trend in teen birthrates in
states that have enacted such laws.
20. For those unfamiliar with the phrase ‘shotgun wedding’ it refers to the case when a
woman becomes pregnant and the man responsible for the pregnancy is compelled (consider the image of the woman’s father holding a shotgun as an ‘incentive’ for the male)
to marry his pregnant partner.
21. These include the number of prisoners and police per capita, various economic measures, welfare payments, handgun control laws and per capita beer consumption. A panel
data regression with ﬁxed eﬀects is employed.
22. Their measure is computed in the following way:
Effective Abortion

Abortionta*(Arrestsa Arreststotal)
a

23.

where a indexes the age of the cohort and t indexes years, with Abortion being the number
of abortions per live birth.
Joyce (2003) challenges the results of Donohue and Levitt. Donohue and Levitt (2004),
however, respond, criticizing Joyce’s work, particularly for his lack of control of the

340

24.
25.
26.

Economics uncut
crack epidemic, which was nearing its peak during the years 1985 through 1990, the
period Joyce uses in his analysis. Donohue and Levitt demonstrate that if Joyce’s
approach is used, but with a lengthened dataset, their hypothesis that abortion reduced
crime is once again supported.
They employ the Supplemental Homicide Reports, which disaggregates the number of
murders by age of the perpetrator and state.
Steve Levitt communicated these results to the author.
For example, the 1993 murder of Dr David Gunn in Pensacola, Florida and John Salvi’s
1994 shooting rampage that left two dead and ﬁve wounded at a clinic in Brookline,
Massachusetts.

References
Akerlof, G.A., J.L. Yellen and M.L. Katz (1996), ‘An Analysis of Out-Of-Wedlock
Childbearing in the United States’, Quarterly Journal of Economics, 111 (2), 277–317.
Alan Guttmacher Institute (AGI) (1999), Sharing Responsibility: Women, Society & Abortion
Worldwide, New York: The Alan Guttmacher Institute.
Bairagi, R. (2001), ‘Eﬀects of Sex Preference on Contraceptive Use, Abortion and Fertility in
Matlab, Bangladesh’, International Family Planning Perspective, 27 (3), 137–43.
Becker, G.S. (1960), ‘An Economic Analysis of Fertility’, in Demographic and Economic
Change in Developed Countries, Universities-National Bureau Conference Series 11.
Princeton, NJ: Princeton University Press.
Becker, G.S. (1965), ‘A Theory of the Allocation of Time’, Economic Journal, (September),
493–517.
Bitler, M. and M. Zavodny (2002), ‘Child Abuse and Abortion Availability’, American
Economic Review Papers and Proceedings, 92 (2), 363–7.
Brown, R.W. and R.T. Jewell (1996), ‘The Impact of Provider Availability on Abortion
Demand’, Contemporary Economic Policy, 14 (2), 95–106.
Coelen, S.P. and R.J. McIntyre (1978), ‘An Econometric Model of Pronatalist and Abortion
Policies’, Journal of Political Economy, 86 (6), 1077–101.
Donohue, J.J. III and S.D. Levitt (2001), ‘The Impact of Legalized Abortion on Crime’,
Quarterly Journal of Economics, 116 (2), 379–420.
Donohue, J.J. III and S.D. Levitt (2004), ‘Further Evidence that Legalized Abortion Lowered
Crime: A Response to Joyce’, Journal of Human Resources, 39 (1), 29–49.
Finer, L.B. and S.K. Henshaw (2003a), ‘Abortion Incidence and Services in the United States
2000’, Perspectives on Sexual and Reproductive Health, 35 (1), 6–15.
Finer, L.B. and S.K. Henshaw (2003b), ‘The Accessibility of Abortion Services in the United
States, 2001’, Perspectives on Sexual and Reproductive Health, 35 (1), 16–24.
Frejka, T. (1983), ‘Induced Abortion and Fertility: A Quarter Century of Experience in
Eastern Europe’, Population and Development Review, 9 (3), 494–520.
Garbacz, C. (1990), ‘Abortion Demand’, Population Research and Policy Review, 9, 151–60.
Gohmann, S.F. and R.L. Ohsfeldt (1993), ‘Eﬀects of Price and Availability on Abortion
Demand’, Contemporary Policy Issues, 11 (4), 42–55.
Graham, M.J., U. Larsen and X. Xu (1998), ‘Son Preference in Anhui Province, China’,
International Family Planning Perspective, 24 (2), 72–7.
Gruber, J., P. Levine and D. Staiger (1999), ‘Abortion Legalization and Child Living
Circumstances: Who Is the “Marginal Child”?’, Quarterly Journal of Economics, 114 (1),
263–91.
Haas-Wilson, D. (1996), ‘The Impact of State Abortion Restrictions on Minors’ Demand for
Abortions’, Journal of Human Resources, 31 (1), 140–58.
Joyce, T. (2003), ‘Did Legalized Abortion Lower Crime?’, Journal of Human Resources, 38 (1),
1–37.
Kahane, L.H. (2000), ‘Anti-abortion Activities and the Market for Abortion Services: Protest
as a Disincentive’, American Journal of Economics and Sociology, 59 (3), 463–85.
Kane, T.J. and D. Staiger (1996), ‘Teen Motherhood and Abortion Access’, Quarterly Journal
of Economics, 111 (2), 467–506.

Economics of abortion

341

Klerman, J.A. (1999), ‘U.S. Abortion Policy and Fertility’, American Economic Review, 89 (2),
261–4.
Levine, P.B., A.B. Trainor and D.J. Zimmerman (1995), ‘The Eﬀect of Medicaid Abortion
Funding Restrictions on Abortions, Pregnancies and Births’, National Bureau of
Economic Research Working Paper No. 5066.
Levine, P.B., D. Staiger, T.J. Kane and D.J. Zimmerman (1996), ‘Roe v. Wade and American
Fertility’, National Bureau of Economic Research Working Paper No. 5616.
Lott, J.R. and J.E. Whitley (2001), ‘Abortion and Crime: Unwanted Children and Out-ofWedlock Births’, Yale Law and Economics Research Paper, No. 254.
Lundberg, S. and R.D. Plotnick (1990), ‘Eﬀects of State Welfare, Abortion and Family
Planning Policies on Premarital Childbearing Among White Adolescents’, Family Planning
Perspectives, 22 (6), 246–51, 275.
Medoﬀ, M.H. (1988), ‘An Economic Analysis of the Demand for Abortions’, Economic
Inquiry, 26 (2), 353–9.
Medoﬀ, M.H. (2000), ‘Black Abortion Demand’, Review of Black Political Economy, 28 (1),
29–36.
Michael, R.T. (1973), ‘Education and the Derived Demand for Children’, Journal of Political
Economy, 81 (2), part 2, S128–S164.
Mincer, J. (1962), ‘Labor Force Participation Rate of Married Women’, in H. Gregg Lewis
(ed.), Aspects of Labor Economics, Universities-National Bureau Conference Series 14,
Princeton, NJ: Princeton University Press.
Mincer, J. (1963), ‘Market Prices, Opportunity Costs and Income Eﬀects’, in C. Christ et al.,
Measurement in Economics: Studies in Mathematical Economics and Econometrics in
Memory of Yehuda Grunfeld, Stanford, California: Stanford University Press.
National Abortion Rights Action League (NARAL) (1993), Who Decides? A State-By-State
Review of Abortion Rights, 4th edition, The Naral Foundation/NARAL.
Ohsfeldt, R.L. and S.F. Gohmann (1994), ‘Do Parental Involvement Laws Reduce Adolescent
Abortion Rates?’, Contemporary Economic Policy, 12 (2), 65–76.
Rahman, A., L. Katzive and S.K. Henshaw (1998), ‘A Global Review of Laws on Induced
Abortion, 1985–1997’, International Family Planning Perspectives, 24 (2), 56–64.

PART V
FUN AND GAMES

12 Economics of sport
John Goddard and Peter J. Sloane

The economics of sport has proved to be an area of fascination for economists for a number of reasons, but the focus has been on the economics of
professional team sports to the relative neglect of non-team sports such as
golf, boxing, athletics, swimming, horse-riding and auto sports. Though
these can be and have been organized on a team basis this is usually not an
essential element and the teams are not generally organized into leagues
that are ranked in order of playing success over the season. The same
neglect is true of amateur or participant, as opposed to spectator sports,
though the numbers involved in these activities are markedly greater than
those who choose to watch professionals as live spectators rather than as
television viewers.
There are two main reasons why economists should be interested in team
sports. First, professional team sports can substitute for a laboratory for
economists, since we can observe operations normally out of the public eye
and also measure directly the productivity of individual employees because
of the richness of the available data. Second, leagues have indulged in
various forms of behaviour such as cartel behaviour, which would certainly
be regarded as anti-competitive in a conventional industry and thus raise
important issues for competition policy.
Such issues have come to the fore both in North America and Europe, but
the means of dealing with them have diﬀered to such a degree that reference
has been made to the ‘North American model’ and the ‘European model’
of sport. In the former case, Barros et al. (2002) note that: ‘Teams are organised into hermetically sealed leagues: both the number of competitors and
the identity of those competitors is ﬁxed by the members themselves, so that
entry at the level of the individual team is impossible “without consent” ’.
This gives rise to a number of features that are generally absent in Europe
such as the entry of new franchises with ﬁnancial compensation to the
incumbents at a price that is agreeable to them and franchise relocation to
exploit subsidies provided by city authorities. In the labour market there
exist player drafts (in which the teams negotiate contracts with leading
young players in reverse order to their ﬁnishing position in the league in the
previous season), roster limits, salary caps and restrictions on player trading.
In product markets, economic competition is limited by gate revenuesharing of the receipts of individual games, joint merchandising and the
345

346

Economics uncut

collective sale of broadcasting rights. In the USA, the major team sports to
which such arrangements apply are baseball, American rules football,
basketball and ice-hockey.
In Europe, football (or soccer) dominates the other team sports, which
include cricket, rugby union, rugby league and speedway, though none of
these is signiﬁcant in all European countries. While US sports are not
entirely absent in Europe, their popularity (with one or two exceptions such
as basketball in Southern Europe) is limited. Revenue-sharing is less
common in Europe, while player drafts and roster limits are generally
unknown and salary caps have only recently been introduced. Player
trading for cash, unusual in North America, is the norm in Europe, at least
in football. Furthermore, some teams are found in cities or towns with
small populations, which would not be countenanced in North America.
As a result there are greater imbalances in income in Europe and a greater
tendency for one or two teams to dominate in their domestic leagues. There
is no North American equivalent to qualiﬁcation for European-wide competitions, though success in the league results in entry into the play-oﬀs in
North America. However, according to the European Commission, the key
feature of the European model of sport is the system of promotion and
relegation, which means the turnover of teams is at least potentially much
greater than applies in North America.
US and European economists have also diﬀered on the objectives
pursued by team owners, with the proﬁt maximization paradigm dominant
in North America and non-proﬁt objectives (such as maximization of
games won subject to a proﬁt constraint) generally considered to be more
appropriate in Europe. It is possible, however, that the greatly increased
income from broadcasting and the tendency for sports teams to become
listed on the stock market may have led to a more commercial approach in
Europe that has narrowed such diﬀerences.
In this chapter we consider the objectives followed by team and league
owners (Section 1), the nature of product demand and, in particular,
whether this is positively related to the uncertainty of outcome (Section 2).
This is followed by an examination of the sports league’s role as a cartel,
including the need to impose strong anti-competitive controls to maintain
competitive balance (Section 3). Finally, we examine contentious issues such
as whether revenue-sharing improves competitive balance, whether salary
caps are necessary and eﬃcient and whether collective selling of broadcasting rights should be permitted (Section 4). Section 5 concludes the chapter.
1 Team and league objectives
The nature of the objectives of team owners and the leagues that control
them is a key question in the economics of sport, especially given the

Economics of sport

347

monopoly nature of these enterprises. The question of proﬁt maximization
versus utility maximization can be viewed at two levels: whether or not
team proﬁts are maximized and whether or not leagues attempt to maximize joint league proﬁts. The North American literature has tended to
de-emphasize utility maximization on the grounds that there is no evidence
that owners, whether sports fans or publicly-traded corporations, have
received less than a market rate of return on their investment. Few teams
in the USA make losses in any season. In contrast, in Europe, few football,
rugby or cricket teams make consistent proﬁts, relying on donations from
wealthy directors and owners or donations from supporters’ clubs to continue in existence.
In principle, there are a number of ways in which one might test whether
utility or proﬁt maximization is being pursued. First, accepting long-run
losses is inconsistent with proﬁt maximizing behaviour, since such ﬁrms
should only remain in business as long as short-run marginal revenue
covers average variable costs and makes some contribution towards average
ﬁxed costs. Second, paying players more than the value of their marginal
revenue product is inconsistent with proﬁt maximization. Third, hiring too
large a squad, deﬁned as paying a higher salary to the last player signed
than that player’s expected contribution to future team revenue infringes
the proﬁt maximization rule. Fourth, building a stadium larger than needed
(that is, where the marginal revenue product of capital is less than the cost
of capital) is likewise inconsistent with this hypothesis. Fifth, charging
prices for tickets below the marginal cost of providing the product infringes
the rule. Having a waiting list for season tickets for many years would be
strong evidence of behaviour inconsistent with proﬁt maximization, as
would setting prices on the inelastic segment of the demand curve, unless
this generated suﬃcient ancillary sales to cover the implied reduction in
total revenue.
There are, however, constraints on utility maximization. Many professional sports teams in North America are publicly held and holding companies will face pressures to ensure that losses are not made. Also, the fact
that there are many million-plus population cities without major league
teams means that the ability to relocate provides for possibilities of earning
more revenue through subsidies. In general in Europe, there is an overprovision of teams, so such opportunities for revenue generation do not
exist. The tendency for some European teams to become listed on the stock
exchange has, however, decreased their ability to indulge in non-proﬁtmaximizing behaviour.
While teams have individual interests, the league as a whole has a collective interest in overall proﬁt maximization. The league as an entity has no
interest in the total number of wins for each team, as each win within the

348

Economics uncut

league is also a loss for one of its teams. League-wide proﬁt maximization
requires, however, competitive balance. Thus, the universal problem in professional team sports leagues is to set up and police a structure that maximizes joint proﬁts, despite the interest of individual owners in either
individual proﬁt maximization or in utility maximization. Indeed, it is not
clear that a league that is made up entirely of teams whose owners were
proﬁt maximizers would behave any diﬀerently from a league that was
made up entirely of utility maximizers.
2 The nature of the product
One early study (Neale, 1964) referred to the peculiar economics of professional team sports, since it is not possible to produce any output without
the assistance of other producers. The essence of such sports leagues is
mutual interdependence. That is, individual producers (teams) have a
vested interest in the economic viability of other teams in order to maintain the interest of their fans and in turn the revenues generated from the
sale of the product (the joint game). The nature of the product thus creates
a requirement for uncertainty of outcome and this is the bedrock upon
which a myriad of restrictions is justiﬁed.
Competitive balance can have a number of meanings. First, there is event
or game uncertainty, which increases as the probability of either side
winning approaches 0.5. Second, there is seasonal uncertainty over which
team will eventually win the championship. Third, there is the absence of
long-run domination by one or two teams, in the sense that the same or one
or two teams win the league championship over a number of seasons. It is
argued that more equal revenue-sharing can assist in making leagues more
equal, though as we shall see this has been disputed by some economists.
However, evidence that dominance has detrimental eﬀects on attendance in
terms of total league revenue is more mixed. In theory, league proﬁts should
be maximized if teams won 50 per cent of their games and cities were all of
the same size. However, to the extent that the drawing power of member
teams varies, more wins by teams in larger areas will likely maximize league
revenues, so that there is an inherent conﬂict of interest between individual
team owners and the interests of the league as a whole.
We can portray the relationship between an owner’s preference map and
the uncertainty of outcome constraint as in Figure 12.1. The proportion or
probability of wins, p, is on the x-axis. The variance or uncertainty of
outcome, p (1p), is on the y-axis. The theory of the distribution of a
binary random variable with constant probability of success establishes a
ﬁxed relationship between the probability of winning p and p (1 p), given
by the inverted U shape labelled UU. We assume a conventionally shaped
indiﬀerence map with I1, I2 and I3 representing increasing levels of utility

Economics of sport

349

p (1 p)

I3
I2
I1

U
0
Figure 12.1

U
0.5

Uncertainty of outcome and owner utility

for the owner (and likely the fans too). Then, it can be seen that maximum
uncertainty of outcome, where p0.5, is not the level that maximizes the
utility of the owner. This occurs at p*, where the owner is prepared to trade
oﬀ some uncertainty of outcome in order to have his or her own team
winning more frequently and thus have a higher probability of winning the
league championship. Only if there is a monotonic relationship between
owner and fan utility and attendance, will p* be the level of winning that
will maximize attendance. Figure 12.1 describes a representative team, as it
is not possible for all franchises to win more than 50 per cent of their games.
The above gives rise to the possibility that the sports league market is
inherently unstable. Given that it is impossible for all teams to be successful in winning their games and given a positive relationship between playing
success and proﬁtability it may not be possible for all teams to be proﬁtable
at any one time. In fact, in North America even less successful teams make
proﬁts, but this does not seem descriptive of experience in Europe. Such a
relationship between uncertainty of outcome, instability and mutual interdependence is illustrated in Figure 12.2.

350

Economics uncut
L

Returns and
costs of
winning

S
A

TRS

WS WL

TRL

W S W L
Percentage wins

Source: Sloane (2002).

Figure 12.2

Uncertainty of outcome and mutual interdependence

We assume a two-team league consisting of a large city franchise, L, and
a small city franchise, S. Further, assume that the costs (TC) of producing
a winning team rise linearly and are identical for both teams, with the
proviso that when one team increases its investment in acquiring a winning
team, the other does not react, and that the returns to winning rise initially
at an increasing rate, but then at a decreasing rate, since enthusiasm may
wane if teams win too often. Let the total returns to the winning schedule
of the large-market team (TRL) be above that of the small-market team
(TRS) on the assumption that a given winning percentage will attract more
spectators in the former case. Thus, there will be a unique proﬁt maximization solution for each team. If the TRL schedule exceeds that of the
TRS schedule by a ﬁxed percentage, so that the slope of the former is steeper
than that of the latter over the relevant range, that of the large city team
will lie to the right of that of the small-market team. Instability is an
inevitable consequence of the zero-sum nature of the league. Thus, if the
large city team wins more often as denoted by L, this will result in the
small-market team winning less often, as denoted by S. The very success

Economics of sport

351

of one team drives the other into a loss-making winning percentage as
shown by the shaded areas. Furthermore, the small-market team by virtue
of its size is much more vulnerable to loss-making than the large-market
team as it has a smaller range of winning percentages, BC, over which it is
proﬁtable, than the large-market team (AD). Finally, the pursuit of playing
success subject to a break-even constraint will increase the likelihood of
losses relative to the proﬁt maximization objective, since the latter lies to
the right of the former (WL for the large-market team and WS for the
small-market team). Again, the zero-sum nature of the league implies
greater losses for the small-market team (as well as for the large-market
team) as it loses more games. In the diagram, the small-market team losses
and large-market team losses are represented by the vertical-lined and
hatched areas respectively.
Mutual interdependence is also inﬂuenced by the relative sizes of the
markets of the two teams. If the population of team L was greater, resulting in an upward shift of the TR curve to L, this would likely shift down
the TR curve of team S, as the fans of the latter realized that winning the
championship was even less likely than in the previous example. This
emphasizes the need for leagues to keep population size diﬀerences between
constituent teams as small as possible, a strategy much more successfully
followed in North America than in Europe. It is also clear that far from
commercialism destroying sports leagues it actually assists in maintaining
competitive balance. It is the pursuit of sporting success, as opposed to
proﬁt, that imbalances leagues.
3 Cartel behaviour
According to Neale (1964), professional team sports are natural monopolies and we may consider teams simply as individual plants in a common
enterprise. Others regard leagues as cartels, which impose constraints on
the activities of individual teams by the imposition of various rules
required to organize the league.1 These include sporting rules to determine
how games are conducted, including the determination of the ﬁxture list,
rules about the selection and employment of players by teams, as well as
the number of teams in membership. Other rules are required to determine
shares of revenue from games, television and other activities. Furthermore,
there may be rules limiting the ability of individuals to have ﬁnancial interests in more than one team in order to maintain conﬁdence in the integrity
of the competition. Such rules are unlikely to fall foul of competition policy
at least in their simpler form. However, other rules are designed to inﬂuence
market structure and conduct and may well raise issues for competition
policy. Examples are controls on league size, on the distribution of league
franchises and various forms of labour market controls.

352

Economics uncut

The standard defence of these restrictions is that without such controls
the wealthier teams would acquire so much player talent that they would
become too dominant and kill spectator interest in the league as a whole.
Some economists such as Rottenberg (1956) have used the invariance principle to argue that such restrictions will have no eﬀect on competitive
outcomes. This principle consists of two propositions. First, in the absence
of transactions costs the parties to any contract will settle on the most
eﬃcient outcome. Second, the same outcome will be achieved regardless of
the distribution of property rights (the invariance thesis). The implication
of this thesis is that the distribution of talent would be identical regardless
of whether leagues imposed limitations on the free mobility of labour. The
best players would still end up in the wealthiest teams. However, this takes
no account of externalities, the fact that a team signing up the best players
will not likely take into account the negative eﬀect of such a signing on
other teams. This has been referred to in the literature as the Yankee
Paradox (Vrooman, 1996). A further problem is that a player’s talent may
be team-speciﬁc (or what Vrooman refers to as the irreversibility proposition). Thus, it may be reasonable to assume that a team for whom a player
has performed over a number of seasons will be better informed about the
player’s potential than any other team interested in signing him (the lemons
phenomenon). The possibility of a winner’s curse operating seems much
more likely in a situation in which the maximization of playing success
dominates over proﬁt maximization considerations.
4 Current policy issues in professional team sports
According to the ‘cartel’ interpretation, one of the main functions of the
leagues in their capacity as organizing bodies is to implement rules aimed
at furthering the collective interest of their member teams. Historically,
this function has in general been accepted by courts in the USA, which
have permitted restrictions on the production of sporting events that
would normally be regarded as anti-competitive in other ﬁelds of business
and commerce. For example, in 1922 the Supreme Court granted major
league baseball (MLB) exemption from the Sherman antitrust laws, and
in 1961 Congress passed the Sports Broadcasting Act, entitling the
leagues to sell broadcasting rights collectively on behalf of their member
teams.
The North American leagues therefore impose restrictions on the operations of the teams, which seek to prevent any individual team from
achieving a level of dominance that would be damaging to the interest
that all teams share in maintaining some degree of competitive balance.
In the early history of North American professional team sports the most
important restriction of this kind was the reserve clause. Having signed a

Economics of sport

353

contract as a professional, the player’s team retained the option to renew
his contract, eﬀectively binding the player to his present employer. The
reserve clause was justiﬁed on the grounds that it was necessary to prevent
the richest teams from outbidding the rest for the services of the top
players, but one of its main eﬀects was to depress players’ compensation
(Scully, 1974). Other restrictions on the free play of market forces in
North American sports include the draft system, salary caps and revenuesharing. The reverse-order-of-ﬁnish draft system allows the weakest
teams from the previous season the ﬁrst pick of rookie players moving
from college (or school in the case of baseball) to professional level for
the ﬁrst time. Salary caps and revenue-sharing are considered in greater
detail below.
4.1 Free agency and salary caps
Before discussing mechanisms for limiting the remuneration of sports professionals, it is perhaps timely to consider the level of salary payments that
are found in many sports. In Table 12.1, the top 25 sports stars in terms of
earnings in 2004–05 are listed, covering ten diﬀerent sports. Followers of
professional sports are often uneasy at the perceived large incomes that star
players receive, as displayed in the table, though other groups such as ﬁlm
stars and musicians do not seem to attract such opprobrium. References to
players’ contracts being similar to those that operated under the slave trade
are also diﬃcult to swallow.
Why then do sports stars earn so much? The answer lies in the concept
of a rent of ability. Real stars are few in number and possess skills that
cannot be easily replicated by training the average player more intensively.
Rosen (1981) referred to this as the superstar phenomenon, and illustrated how human capital interacted with production technology to
magnify small diﬀerences in talent, resulting in large diﬀerences in earnings. Technology enables the product to be reproduced at low cost, since
large stadia enable large numbers of spectators to consume the product
at the same time, and broadcasting widens the market still further.
Characteristically, wages are highly convex with respect to star quality,
resulting in a highly skewed distribution of earnings with a long upper-tail
to the distribution. There is good reason for this. Hausman and Leonard
(1997) showed that the presence of star players can have a substantial eﬀect
on the number of television viewers. In the case of basketball, they found
that superstars such as Michael Jordan, Magic Johnson and Larry Bird
raised television ratings by around 30 per cent, in addition to their eﬀect on
paid attendance.
A recent study by Lucifora and Simmons (2003) tests for superstar eﬀects
on the salaries of Italian footballers. In line with Rosen’s model, this study

354

Economics uncut

Table 12.1

Top 25 sports star earnings in US$ (2004–05)

1 Tiger Woods (Golf) – $87.0m
2 Michael Schumacher (Motor Racing) – $60.0m
3 Oscar De La Hoya (Boxing) – $38.0m
4 Michael Vick (American Football) – $37.5m
5 Shaquille O’Neal (Basketball) – $33.4m
6 Michael Jordan (Basketball) – $33.0m
7 David Beckham (Soccer) – $32.5m
8 Kobe Bryant (Basketball) – $28.8m
9 Lance Armstrong (Cycling) – $28.0m
9 Valentino Rossi (Motor Cycling) – $28.0m
11 Alex Rodriguez (Baseball) – $27.5m
12 Phil Mickelson (Golf) – $26.8m
13 Andre Agassi (Tennis) – $26.2m
14 Derek Jeter (Baseball) – $25.5m
15 Manny Ramirez (Baseball) – $24.2m
16 Jeﬀ Gordon (Motor Racing) – $23.4m
17 Walter Jones (American Football) – $23.2m
18 Ronaldo (Soccer) – $23.0m
19 LeBron James (Basketball) – $22.9m
20 Matt Hasselbeck (American Football) – $22.8m
21 Maria Sharapova (Tennis) – $18.2m
22 Serena Williams (Tennis) – $12.7m
23 Annika Sorenstam (Golf) – $7.3m
24 Venus Williams (Tennis) – $6.5m
25 Lindsay Davenport (Tennis) – $6.0m
Note: Earnings estimates are for June 2004 to June 2005. They include salaries, bonuses,
prize money, endorsements and appearance fees. The earnings for NBA players (basketball)
are net of the 10 per cent escrow tax that is withheld from their salaries as part of the
league’s collective bargaining agreement with the player’s union.
Source: Forbes (www.forbes.com).

ﬁnds that earnings were highly convex in two goalscoring and performance
measures, after controlling for personal characteristics and team eﬀects.
The implied earnings ratio for the median superstar, with a scoring record
of more than 0.2 goals per game, was over four times that of the median
ordinary player, with a goal scoring rate below this ﬁgure. A player located
at the 99th percentile of the earnings distribution earned over ten times
more than players located at the median, and 45 times more than those
players located at the 10th percentile of the distribution. Only nine players
in Serie A, the major Italian League, scored at a rate of 0.5 goals per game
or better. The rewards for goal scoring exceeded those for supplying the

Economics of sport

355

ﬁnal pass that created the goal. Goal scoring is clearly the most glamorous
part of the sport in the eyes of the consumer.
An alternative way to represent the above is to compare the supply and
demand equilibrium of sports stars with that of another occupation, for
which there is a much higher demand, but at the same time a very elastic
supply schedule. Fort (2003) uses the example of teachers. As constructed
in Figure 12.3, the demand for teachers lies well to the right of that for
sports stars. However, the supply of sports stars is much lower than that of
teachers and much more inelastic, with the consequence that the earnings
of sports stars, Es, greatly exceed those of teachers at Et, though many more
teachers are employed, Lt, than is the case for sports stars, Ls.
Until the ﬁnal quarter of the twentieth century, the reserve clause played
a major role in restraining salary levels in North American professional
sports. Baseball’s reserve clause, eﬀective since 1880, was eventually
removed through collective bargaining in 1976. Subsequently, all MLB
players became free agents after completing a minimum number of years
of major league service (currently six years). In a settlement ﬁnalized
in 1983, the National Basketball Association’s (NBA) reserve clause was
abolished in exchange for the introduction of a salary cap, which sought to
oﬀset the potentially harmful eﬀects of free agency for competitive balance.
The salary cap limits each team’s expenditure on players’ compensation to

Wage

Sports star supply

Teacher demand
Es
Sports star
demand

Teacher supply

Employment

Source: Adapted from Fort (2003).

Figure 12.3

Relative supply and demand for teachers and sports stars

356

Economics uncut

53 per cent of gross league revenues divided by the number of teams in the
league. Several exemptions apply: for example, a team can exceed the cap
in order to match an oﬀer made to one of its ﬁrst team players by a rival
team. As part of its 1993 free agency settlement, the National Football
League (NFL) adopted a similar arrangement, with wages limited to 67 per
cent of designated revenues. An attempt to impose a salary cap in MLB
resulted in a prolonged players’ strike in 1994–95. The eventual settlement
included a luxury tax on expenditures on wages and salaries over a certain
limit, with the proceeds redistributed among the poorer teams: an arrangement that may be interpreted as a ‘softer’ version of a salary cap. A new
version of the luxury tax was implemented in 2002.
By equalizing expenditure on players’ wages and salaries throughout the
league, the NBA salary cap is intended to enable the small-market teams to
compete on equal terms with the large-market teams. In practice, however,
the exemption clause is invoked routinely by the large-market teams and the
cap has not succeeded in equalizing expenditures. Consequently total
wages regularly exceed 60 per cent of revenues, and the high-spending
teams’ wage bills are typically more than twice those of the low-spending
teams. If anything, competition in the NBA may have become slightly less
balanced since the cap was introduced (Quirk and Fort, 1999).
In theory, if all teams are required to spend exactly the same on wages,
and if a reverse-order-of-ﬁnish draft allocates the best new players to the
lowest-ﬁnishing teams, the long-term equilibrium allocation of playing
talent implies perfect competitive balance. The large-market teams would
like to spend more on players and the small-market teams would like to
spend less, but both are prevented from doing so by the rules of the salary
cap. The equilibrium is also ineﬃcient since total league revenues are not
maximized (Fort and Quirk, 1995).
This type of analysis assumes that the purpose of the salary cap is to
achieve greater competitive balance. Vrooman (1995), however, considers
an alternative scenario in which the real purpose of the cap is to assist the
league (in its role as a cartel) to maximize the combined revenues or proﬁts
of all its member teams. Suppose that in the absence of the cap, largemarket teams incur a higher marginal cost in hiring playing talent than
small-market teams. A player of any given level of ability tends to demand
more for signing for a large-market team than a small-market team,
perhaps because living costs tend to be higher in big cities, or because
players take account of the team’s ability to pay when formulating their
wage demands. In the absence of the salary cap, the large-market teams’
revenue advantage is partially oﬀset by a cost disadvantage, and their
tendency to dominate competition is checked to some extent. Under the
rules of the salary cap, however, all teams spend the same amount on

Economics of sport

357

players, so the marginal cost of talent does not aﬀect the decision to hire
talent. The objective of league revenue or proﬁt maximization requires each
player to be assigned to the team for which his marginal revenue product is
the greatest. In other words, revenue functions are the only determinant of
the allocation of talent; cost functions are irrelevant. For this reason, the
large-market teams dominate competition to a greater extent with the
salary cap than in its absence.
High rates of inﬂation in players’ wages in European football have been
a serious cause for concern in recent years. In England, for example, the
annual rate of wage inﬂation was running close to 30 per cent throughout
most of the 1990s. By 2001, the average ratio of wages and salaries expenditure to revenue in the Premier and Nationwide leagues was approaching
the 70 per cent benchmark regarded by accountants Deloitte and Touche
(2004) as the maximum likely to be sustainable by any team in the long
term. In 2002–03, the ratio for the average Premiership club was below this
benchmark at 61 per cent, but for the average Division One, Division Two
and Division Three2 clubs stood at 89 per cent, 85 per cent and 68 per cent
respectively. Elsewhere in Europe, many French, Italian and Spanish football teams have recently experienced similar diﬃculties. In Italy, for
example, all 18 Serie A teams returned an operating loss in 2001.
Rosen and Sanderson (2001) have drawn an analogy between rising
expenditure on players’ wages in professional sports and an arms race
between nations. For an individual team (or nation) unable to inﬂuence the
momentum of the spending process at the aggregate level, increasing its
own expenditure may be the only rational action, given that abstention
would mean losing ground. But in this zero-sum game, any expenditure by
one team that achieves a relative improvement in performance imposes a
negative externality on others. Teams tend to spend as much on players as
their revenues (or the willingness of their owners to subsidize losses) will
allow, but the beneﬁt of the extra expenditure is largely cancelled out by
others acting in the same manner. Against this background, the need for
some form of collective restraint on wage expenditure in European football
has been widely debated.
Given the openness of the post-Bosman footballers’ labour market, it
seems likely that to be eﬀective a North American-style salary cap would
have to operate Europe-wide. However, the practical and legal obstacles
may well be insurmountable, especially in view of the probable scepticism
of the European Commission on grounds of competition policy. Some
form of voluntary restraint might be possible, however. Starting from the
2004–05 season, European football’s governing body UEFA introduced a
licensing scheme which imposes conditions on the financial management
of football clubs and aims to promote good governance. The licensing

358

Economics uncut

scheme requires clubs to meet five criteria, including national financial
standards, in order to be permitted to compete in the Champions League
and the UEFA Cup. Meanwhile, the G14 group, comprising the leading
teams from several European countries, agreed in November 2002 that
each team’s wage expenditure should be limited to 70 per cent of annual
turnover from the 2005–06 season and are urging other clubs to follow
their guidelines.
4.2 Cheating in professional team sports
A major diﬃculty associated with any attempt to restrain expenditure on
players’ salaries is that it leaves open the possibility of cheating on the part
of individual teams, in the hope that they can gain an advantage over
others who adhere to the rules. This can be illustrated using the prisoner’s
dilemma game-theoretic model. In a two-team model, assume that the
league rules have capped the total size of each team’s wage bill. Suppose a
team that adheres to the rules of the cap earns an operating profit of
200. A team that ‘cheats’ and breaks the rules of the cap by overspending earns an operating proﬁt of zero. However, there is an additional prize
of 1000 for the team that wins the league. If both teams adhere to the
rules of the cap or if both teams cheat, each has a 50 per cent chance of
winning the prize. However, if one team cheats while the other adheres to
the rules of the cap, the odds change so that the cheating team has an 80
per cent chance of winning the prize. Figure 12.4 shows the expected
payoﬀs to each team under this structure, conditional on the actions of
both teams.
If both teams adhere to the rules of the cap, combined expected proceeds
are maximized (both teams earn 700). Will this joint maximizing
outcome actually be attained? Consider ﬁrst team A’s decision. If B does
not cheat, A’s expected proceeds are 700 if A does not cheat, and 800 if
A cheats. If B cheats, A’s expected proceeds are 400 if A does not cheat,
and 500 if A cheats. Whatever B does, it is therefore best for A to cheat.

Team B
Cheat

Adhere

Cheat

(500, 500)

(800, 400)

Adhere

(400, 800)

(700, 700)

Team A

Figure 12.4

The sporting prisoner’s dilemma

Economics of sport

359

Because the model is symmetric, B applies similar reasoning and reaches
the same decision. Therefore, a sub-optimal outcome is achieved in which
both parties cheat and combined expected proceeds are reduced (both
teams earn 500). Both parties would be better oﬀ if they could agree to
obey the rules, but neither is confident that the other party would stick to
its side of the bargain. There is plenty of evidence to indicate that cheating
behaviour is common when leagues impose restrictions of the kind
described above.
Cricket provides another example in which the form of corruption has
involved individual players accepting bribes from bookmakers or gamblers
for ﬁxing matches. The former South African captain Hansie Cronje admitted his involvement, but similar allegations have been levelled against
players in Australia, England, India, New Zealand, Sri Lanka and the West
Indies. It has been suggested that the cause of the problem is the low pay
received by international cricketers relative to other sports, as a consequence of cricket’s organizational structure, which uses the revenue from
test matches and international one-day games to subsidize domestic competition. Preston et al. (2002) suggest that cheating in sports falls into two
categories. Where the rewards for winning substantially exceed the rewards
for ﬁnishing second or third (in terms of money or prestige), this ‘winnertakes-all’ structure may lead athletes to take performance-enhancing drugs
or adopt other measures to gain an unfair advantage. A second form of
cheating may occur where the rewards to athletes are insuﬃciently
diﬀerentiated, as in cricket, so that the potential gains to cheating are much
greater than the penalties for being caught. In this case, raising pay and
increasing the penalties for being caught should help reduce the likelihood
of cheating or corruption.
4.3 Revenue-sharing and the collective sale of broadcasting rights
In North American professional team sports, revenue-sharing plays an
important part in oﬀsetting inequalities in drawing power between teams.
Practices concerning the distribution of gate revenues vary between sports.
NFL home and away teams, historically, shared gate revenues on a 60:40
split. In January 2001, NFL owners approved a new revenue-sharing policy
to begin in 2002 in which the league pools visiting teams’ shares of gate
receipts for all pre-season and regular season games and divides the pool
equally among all 32 teams. This is much more equalizing than arrangements elsewhere. For major league baseball, the split is 95:5 in the National
League and 80:20 in the American League with some modest sharing of local
revenues. In the NBA and NHL, there is no sharing of gate revenues. Broadly
speaking, national television revenues are shared equally between all league
member teams, while the home teams keep local television revenues.

360

Economics uncut

One of the best known ﬁndings of the early economics of professional
team sports literature is that the existence or nature of arrangements for
revenue-sharing among the members of a sports league should have no
eﬀect on competitive balance. Intuition might suggest otherwise: by narrowing the gap between the richer and poorer members of the league,
revenue-sharing might be expected to allow the poorer teams to compete
more eﬀectively in the market for playing talent. In turn, this might be
expected to improve competitive balance. Indeed, the aim of improving
competitive balance might well be one of the main reasons for having a
revenue-sharing arrangement. According to the theory, however, this intuition is misleading. The preceding argument fails to recognize the damaging eﬀect of revenue-sharing on the incentives for teams to compete in
order to attract playing talent.
Consider the two-team model introduced in Section 2. Suppose initially
there is no revenue-sharing, and each team keeps all of its own revenues. At
any particular allocation of playing talent between teams L and S, the marginal beneﬁt to team L of a small increase in its win ratio brought about by
hiring some incremental talent is MRL and the marginal beneﬁt to team S
is MRS. In accordance with the invariance principle, the proﬁt-maximizing
allocation of talent must be such that MRL MRS. If there were any divergence between MRL and MRS, a trade in talent could be devised that would
leave both teams better oﬀ. In this model there is an assumption of symmetry: the marginal eﬀect on team L’s revenue of an incremental increase
in team L’s playing talent is the same as the marginal eﬀect of an incremental reduction in team S’s talent. There is also an assumption that the
total stock of playing talent is ﬁxed, and that this fact is taken into account
by both teams when making their hiring decisions. The implications of
relaxing these assumptions are considered below.
Suppose now a revenue-sharing arrangement is introduced, whereby
each team keeps only a proportion, say 90 per cent, of its own revenues,
while the remaining 10 per cent is awarded to the other team. The marginal
beneﬁt to team L of hiring incremental talent is now 0.9 times the marginal
eﬀect on its own revenue resulting from the increase in its win ratio, plus 0.1
times the (negative) marginal eﬀect on team S’s revenue resulting from the
reduction in its win ratio, or 0.9MRL 0.1MRS. Similarly, the marginal
beneﬁt to team S of hiring incremental talent is 0.1MRL 0.9MRS.
Comparing the situation before and after the introduction of the
revenue-sharing arrangement, the marginal beneﬁt from hiring incremental talent falls by the same amount for both teams: 0.1MRL 0.1MRS.
Because the eﬀect on both teams’ marginal revenue functions is identical,
the location of the proﬁt-maximizing distribution of playing talent, and
therefore the degree of competitive balance, does not alter. However, there

Economics of sport

361

is a reduction in the equilibrium wage: the teams do not compete as ﬁercely
to attract talent if part of the additional revenue generated by additional
talent is shared. According to this model, revenue-sharing tends to depress
players’ compensation, but it does not aﬀect competitive balance.
This straightforward conclusion, that the eﬀect of revenue-sharing on
competitive balance is neutral, has been challenged by a number of economists who have developed alternative models, in which revenue sharing does
aﬀect the equilibrium distribution of playing talent (Atkinson et al., 1988;
Rascher, 1997; Kesenne, 2000):
1.

2.
3.

If the total supply of talent is not ﬁxed, so it is possible for one team
to increase its stock of talent without automatically depleting its competitors’ stock of talent, revenue sharing may aﬀect competitive
balance.
If teams pursue utility-maximizing rather than proﬁt-maximizing
objectives, revenue-sharing may aﬀect competitive balance.
If revenue depends not only on the relative values of each team’s allocations of playing talent (which determine their win ratios) but also on
the absolute values, on the grounds that spectator demand is sensitive
to the overall quality of talent on view, revenue-sharing may aﬀect competitive balance under both proﬁt- and utility-maximizing assumptions.

Unfortunately, no consensus has so far emerged as to the direction of these
eﬀects: in some models revenue-sharing can be shown to improve competitive balance, while in others the eﬀect is the opposite. For example, Fort
and Quirk (1995) consider the eﬀect of sharing television revenues derived
from contracts between teams and local television stations. Television audience levels are assumed to be sensitive to the absolute level of talent on
display, so the symmetry assumption is abandoned: the marginal eﬀect on
team L’s television revenue of an incremental increase in its own playing
talent is greater than the marginal eﬀect of an incremental reduction in
team S’s talent, because in the ﬁrst case the absolute level of talent on
display increases, while in the second it declines. In this formulation the
introduction of a revenue-sharing arrangement has a greater (negative)
eﬀect on the marginal revenue function of the large-market team than the
small-market team. Revenue-sharing, therefore, tends to improve competitive balance.
In contrast, Szymanski and Kesenne (2004) have suggested that the conclusion that revenue-sharing has a neutral eﬀect on competitive balance
depends on a (previously unrecognized) implicit assumption about the
nature of conjectural variation: the assumptions each team makes about
the reactions of its competitor in response to its own actions. If team L is

362

Economics uncut

considering the eﬀect on its own revenue of an increase in its own expenditure on talent, for example, in the model described above, team L assumes
that team S will passively acquiesce in an equivalent reduction in its talent.
This is a consequence of the model’s assumption that the total supply of
talent is ﬁxed. Whether the teams would recognize their interdependence in
this way in practice is open to debate, however. An assumption about the
nature of conjectural variation more widely used in the oligopoly and game
theory literatures, known as Nash conjectures, requires each team to base
its own decisions on the assumption that its competitor will always stick to
its present position. Team L would not take into account the ‘total ﬁxed
supply of talent’ constraint when evaluating the eﬀect of an increase in its
own stock of talent, and would not assume that team S’s stock of talent
would fall by an equivalent amount. In this case, the introduction of a
revenue-sharing arrangement still reduces the incentive for both teams to
spend on talent as before. But because team S has more to gain from a share
of team L’s revenue than vice versa, team S reduces its expenditure by more
than team L. Revenue-sharing, therefore, causes competitive balance to
deteriorate in this formulation.
In European football, practices concerning revenue-sharing have varied
signiﬁcantly, both over time and between countries. In England, for
example, until 1983, 20 per cent of the notional receipts from every league
match were paid to the visiting team. Subsequently, home teams have
retained their own gate receipts. Under another scheme, 4 per cent of all
notional receipts were paid into a pool, the proceeds from which were
distributed evenly among all league teams. The 4 per cent levy was reduced
to 3 per cent in 1986, prior to the complete withdrawal from this scheme of
the leading (Premier League) teams in 1992.
Although the general trend has been towards the erosion of arrangements for direct sharing of gate revenues, sharing of television revenues
through the collective sale of broadcasting rights is still widespread,
though not universal, throughout Europe. Among the big ﬁve European
leagues, rights are sold collectively (by the leagues and not the teams)
and exclusively (to a single broadcaster) in England and Germany. In
each case, the distribution of the proceeds between the teams is determined by formula. In France, the rights are sold collectively but there is no
exclusivity: currently the rights are shared between two broadcasters. In
Italy and Spain, the teams sell their own rights individually and retain the
proceeds.
The legality of arrangements for the collective sale of broadcasting
rights has recently been subject to scrutiny from the competition authorities at both national and European levels, and the future structure of
broadcasting rights deals is likely to be shaped in large measure by the

Economics of sport

363

courts. In the UK in 1996, the Oﬃce of Fair Trading (OFT) initiated proceedings under the Restrictive Trade Practices Act, 1976 at the Restrictive
Practices Court (RPC), in which the OFT argued that the Premier League
acted unlawfully as a cartel in negotiating the sale of broadcasting rights
on behalf of its member teams. According to the OFT, the exclusivity of
the Premier League’s contract with the satellite broadcaster British Sky
Broadcasting (BSkyB) restricted the number of live matches screened on
television, and inﬂated the price paid by BSkyB subscribers.
In July 1999, the RPC reached a judgement that appears to have surprised
many commentators. In rejecting the OFT’s case, the RPC took the view
that on balance these arrangements did not operate against the public interest. The OFT’s case rested partly upon the question as to whether access to
exclusive live Premier League coverage conferred monopoly power upon
BSkyB as sole suppliers to a distinct market for a speciﬁc programme type,
or whether it merely enabled BSkyB to compete more eﬀectively with other
broadcasters for audiences as part of a wider market for television programmes in general (Cave, 2000). The RPC accepted the latter view, which
implied that exclusivity could be beneﬁcial in helping to promote competition between broadcasters through improved programme quality. From
football’s point of view, collectivity enabled the Premier League to market
its championship as a single entity, and produced greater equality between
the television revenues earned by richer and poorer teams than would be the
case if the teams sold their broadcasting rights individually.
Although the outcome of the OFT’s challenge settled the issue temporarily, it seems unlikely to represent the ﬁnal word on this matter. In
December 2002, the European Commission issued a statement of objections to the Premier League concerning the collective sale of broadcasting
rights, reiterating a number of arguments aired previously at the RPC.
Collective selling, especially when combined with exclusivity, excludes
other broadcasters from access to live televised football, resulting in higher
prices and limited choice for consumers. While the Commission accepts
that some element of collectivity could be justiﬁed as a means of redistributing television revenues within football, it does not accept that the present
arrangements are the only mechanism through which redistribution could
be achieved. Citing arrangements for the future sale of European
Champions League broadcasting rights by UEFA as an exemplar, the
Commission suggested that the Premier League’s broadcasting rights could
be sold in smaller bundles. Some bundles might still be sold collectively,
while others might be sold separately by individual teams. In July 2003, the
Premier League did split its oﬀer into three separate packages, but because
BSkyB’s bid was accepted for each of these packages the Commission was
still not satisﬁed that the process was suﬃcient to maintain competition.

364

Economics uncut

However, perhaps not all recent events suggest that the market for broadcasting rights is easily capable of diversifying in quite the way the
Commission envisages. Pay-per-view transmission of live football has so
far achieved only a limited foothold in the UK market. The liquidation of
the terrestrial digital broadcaster ITV Digital in 2002, and the continuing
ﬁnancial diﬃculties of the two cable operators NTL and Telewest, have
strengthened BSkyB’s position as the most credible, and perhaps the only
major bidder for pay-television rights in the foreseeable future. Meanwhile,
recent turmoil in Italy, where the start of the 2002–03 season was delayed
while several of the smaller Serie A teams attempted to achieve satisfactory
terms for the sale of their broadcasting rights, serves ample warning as to
the potentially chaotic and divisive eﬀect of the abandonment of the collectivity principle.
5 Conclusion
The economics of sport represents an area of continuing fascination for
economists. Professional sports provide fertile territory for empirical economists, who can measure the productivity of individual employees thanks
to the richness of the available datasets, and for theoreticians who can
observe various forms of behaviour, such as decision-making through
cartels, which would either be hidden or regarded as anti-competitive in
most other areas of business and commerce. Such issues have been prominent throughout the histories of professional sports in both North America
and Europe, and remain so today, although the methods for dealing with
them have diﬀered, partly because of diﬀerences in the underlying structural characteristics of the North American and European models for the
organization of professional team sports. North American and European
economists have also diﬀered concerning the objectives pursued by team
owners, with proﬁt maximization the dominant assumption in North
America, but non-proﬁt objectives (such as maximization of games won
subject to a proﬁt constraint) widely considered to be more realistic in a
European setting.
Much of the discussion concerning the nature of professional sport’s
special status, which may or may not justify allowing sports leagues and
teams certain exemptions from the full rigours of competition law, concerns the interdependences that exist between the interests of individual
teams, and the collective interest that all share in the preservation of some
degree of competitive balance (in the sporting sense) among the member
teams of any league. Historically, US courts have tended to accept the argument that such considerations may indeed justify certain exemptions from
competition law. More recently, the European courts, while acknowledging
the principles of interdependence and collective interest at a general level,

Economics of sport

365

have demonstrated a certain degree of scepticism over particular issues, and
have regularly questioned whether stated objectives, such as the crosssubsidization of poorer teams by their richer counterparts, could not be
achieved by means other than certain speciﬁc forms of anti-competitive
practice. Issues such as whether revenue-sharing improves competitive
balance, whether salary caps are an eﬀective way of restraining expenditure
on players’ wages and salaries, and whether the collective selling of broadcasting rights should continue, appear likely to occupy much of the policy
debate for many more years to come.
Notes
1.

However, a more useful approach may be to treat the league as a form of joint venture
(Flynn and Gilbert, 2001). This is particularly so when there is a simple entity ownership
structure as in Major League Soccer in the USA in which the league as opposed to the
individual member club owns all player contracts.
Division 1, Division 2 and Division 3 were re-named The Championship, League 1 and
League 2 respectively for the start of the 2004–05 season.

References
Atkinson, S., L. Stanley and J. Tschirhart (1988), ‘Revenue Sharing as an Incentive in an
Agency Problem: An Example from the National Football League’, Rand Journal of
Economics, 19 (1), 27–43.
Barros, C.P., M. Ibrahimo and S. Szymanski (eds) (2002), Transatlantic Sport: The
Comparative Economics of North American and European Sports, Cheltenham, UK and
Northampton, MA, US: Edward Elgar.
Cave, M. (2000), ‘Football Rights and Competition in Broadcasting’, in S. Hamil, J. Michie,
C. Oughton and S. Warby (eds), Football in the Digital Age: Whose Game is it Anyway?,
Edinburgh: Mainstream Publishing, pp. 180–90.
Deloitte and Touche (2004), Annual Review of Football Finance, Manchester: Deloitte and
Touche.
Flynn, M.A. and R.J. Gilbert (2001), ‘The Analysis of Sports Leagues as Joint Ventures’,
Economic Journal, 111 (469), F27–F46.
Fort, Rodney (2003), Sports Economics, New Jersey: Prentice Hall.
Fort, R. and J. Quirk (1995), ‘Cross-subsidisation, Incentives and Outcomes in Professional
Team Sports’, Journal of Economic Literature, 33 (3), 1265–99.
Hausman, J. and G. Leonard (1997), ‘Superstars in the National Basketball Association:
Economic Value and Policy’, Journal of Labor Economics, 15 (4), 586–624.
Kesenne, S. (2000), ‘Revenue Sharing and Competitive Balance in Professional Team Sports’,
Journal of Sports Economics, 1 (1), 56–65.
Lucifora, C. and R. Simmons (2003), ‘Superstar Eﬀects in Sport: Evidence From Italian
Soccer’, Journal of Sports Economics, 4 (1), 35–55.
Neale, W.C. (1964), ‘The Peculiar Economics of Professional Sports’, Quarterly Journal of
Economics, 78 (1), 1–14.
Preston, I., S.F. Ross and S. Szymanski (2002), ‘Seizing the Moment: A Blueprint for the
Reform of World Cricket’, unpublished manuscript.
Quirk, James and Rodney Fort (1999), Hard Ball: The Abuse of Power in Pro Team Sports,
New Jersey: Princeton University Press.
Rascher, D.A. (1997), ‘A Model of a Professional Sports League’, in W. Hendricks (ed.),
Advances in the Economics of Sport, vol. 2, Greenwich, CT: JAI Press.
Rosen, S. (1981), ‘The Economics of Superstars’, American Economic Review, 71 (5), 167–83.
Rosen, S. and A. Sanderson (2001), ‘Labour Markets in Professional Sports’, Economic
Journal, 111 (469), F47–F68.

366

Economics uncut

Rottenberg, S. (1956), ‘The Baseball Player’s Labor Market’, Journal of Political Economy, 64
(3), 242–58.
Scully, G.W. (1974), ‘Pay and Performance in Major League Baseball’, American Economic
Review, 64 (6), 915–30.
Sloane, P.J. (2002), ‘The Regulation of Professional Team Sports’, in C.P. Barros, M. Ibrahimo
and S. Szymanski (eds), Transatlantic Sport: The Comparative Economics of North
American and European Sports, Cheltenham, UK and Northampton, MA, US: Edward
Elgar, pp. 50–68.
Szymanski, S. and S. Kesenne (2004), ‘Competitive Balance and Gate Revenue Sharing in
Team Sports’, Journal of Industrial Economics, 52 (1), 165–77.
Vrooman, J. (1995), ‘A General Theory of Professional Sports Leagues’, Southern Economic
Journal, 61 (4), 971–90.
Vrooman, J. (1996), ‘The Baseball Player’s Labour Market Reconsidered’, Southern Economic
Journal, 63 (2), 339–60.

13 Economics of gambling
Robert Simmons

This chapter applies economic analysis to the growing gambling sector.
Over the past 50 years, gambling has become legally recognized in all
industrialized economies with an increasing variety of gambling products
now on oﬀer. Technological change, primarily via the Internet, is likely to
sustain this growth path. The chapter begins with a brief summary of
the scope and composition of gambling in some major jurisdictions. In
Section 2, we then pose the fundamental economic question of why people
gamble at all, particularly in lottery games where the probability of winning
the jackpot prize is of the order of one in 14 million. Next, in Section 3 we
consider economic reasons why states and governments may opt to legalize
forms of gambling. A public choice approach is highlighted. In Section 4
we examine some contemporary policy issues involved in regulation and
deregulation of gambling, including the social benefits and costs of gambling, the use of state and government revenues from gambling for particular purposes and the threat to government tax revenues and government
regulation posed by the growth of Internet gambling. Section 5 concludes
the chapter.
1 The scope of gambling worldwide
Across the world, gambling covers many products and takes both legal and
illegal forms. In the USA there are presently only three states where legal
gambling does not take place within their borders (Hawaii, Tennessee and
Utah). The total amount wagered in 2001, termed handle within the industry, was estimated at US $859 billion. Deductions of winnings gives gross
revenues and these were US $73 billion in 2003, more than the combined
revenues from movie tickets, recorded music, theme parks, spectator sports
and video games (Christiansen Capital Advisors, 2003). Market-based
opportunities for gambling have grown enormously over the last decade as
the Internet has made it possible for gambling transactions to be carried
out quickly and easily from a home PC. Hence, Internet gaming companies
advertise a range of products including casino games (‘blackjack without
the black tie’) and betting on a wide range of sporting and political outcomes. But, although Internet gambling is a strong growth sector, it should
be stressed that this represents a small proportion of gross gambling revenues (8 per cent in the USA in 2003, up from 5 per cent in 2001) as many
367

368

Economics uncut

Table 13.1 Growth of gross wagering (handle) in Australia and USA
Year

Australia

USA

1998
1999
2000
2001
2002
2003

100.0
108.8
120.4
125.0

100.0
112.5
121.8
125.8
132.3
140.5

Note: Index numbers with base year 1998.
Source: Tasmanian Gaming Commission (Australia) and Christiansen Capital Advisors
(USA).

Table 13.2 Percentage composition of handle in Australia, UK and USA
(1998–2003)
Australia

Casinos & gaming machines
National & state lotteries
Betting

USA

1998

2001

1998

2001

2003

83.1
3.1
12.6

84.9
2.8
10.8

59.0
12.8
20.9

61.4
27.0
6.2

62.4
27.4
5.4

Sources: Tasmanian Gaming Commission (Australia), Christiansen Capital Advisors
(USA), and Department for Culture, Media, and Sport (UK).

or most citizens in the USA and UK still gamble by more traditional means
such as visiting casinos or buying lottery tickets in stores.
Worldwide, the most prevalent form of gambling is the purchase of
lottery tickets, either as part of a draw or as ‘instant’ scratchcards, which
reveal immediate win or loss. In the USA in 2003, 38 states plus the District
of Columbia operated lotteries. Seventy-two per cent of Americans admitted to playing a lottery game at least once in the year 2000 (Gtech, 2000).
In 2002, state referenda supported progress towards legislation to adopt
lotteries in North Dakota and Tennessee. All Canadian provinces have
lotteries, as have all European countries.
Tables 13.1 and 13.2 display the growth and composition of gross
annual wagering in Australia (1998 and 2001) and the USA (2001 and
2003). Table 13.2 also shows the composition of gambling in the UK for
1998, the only year for which useful figures could be obtained. These tables

Economics of gambling

369

reveal steady growth of wagering, together with sustained dominance of
wagering in casinos and on gaming machines in both Australia and the
USA. The latter has been sustained recently by the expansion of casinos
in Indian reservations. In contrast, the UK has greater proportions of
wagering accounted for by betting on racing and sports and by lottery
wagering, with a smaller weight taken up by casino and slot machine
gaming.
2 Why gamble?
A gamble is a state-contingent claim: an individual places a stake with the
prospect of either losing this amount or having the stake returned together
with some winnings. Winnings are calculated in two main ways. In parimutuel betting, associated with lotteries worldwide and horse racing in the
USA and Australia, bettors’ wagers are placed in a pool and winnings are
shared amongst successful bettors after taxes and commissions have been
deducted. The terms of the bet are not known until the lottery is drawn or
the race is started.
In fixed-odds betting, found in Las Vegas sports betting and UK bookmaker betting, a particular value of winnings for a successful stake is
announced to potential bettors. In Europe most sports bettors gamble on
who will win a game and the odds on the game outcome tend to be fixed.
US sports betting is notable for its use of a betting line or point spread. The
sports gambler bets on the result of a game compared with a predicted
margin of victory for one team. If Chicago Bears are predicted to beat
Detroit Lions by six points, a punter betting on Chicago will win if
Chicago’s margin of victory exceeds six. Conversely, a punter betting on
Detroit (against Chicago) will win if Chicago’s margin of victory is less
than six or if Detroit wins the game. Bookmakers adjust the point spread
so as to equalize amounts bet on each side of the line and hence generate a
risk-free profit at the fixed odds of 10 to 11. A winning wager of $10 will
return $19.09 (the $10 stake plus $10 times 10/11). The bookmaker’s commission, or transactions cost from the bettor’s point of view, is five cents in
the dollar.
Analysis of gambling in economics usually begins with the notion of
expected utility. Denoting utility by U, wealth by W, winnings by G, stake
by S and probability of winning a wager by p, expected utility E(U) is
given by:
E(U)pU(WG)(1p)U(WS)
Economic theory of consumer behaviour typically begins with the
assumption that individuals are rational and are risk-averse. In Figure 13.1,

370

Economics uncut
Utility
U (wealth)
U(18)
U(12)

0.5U(18) + 0.5U(6)

U(6)

Figure 13.1

Wealth

Risk aversion and gambling

risk aversion is exhibited by utility drawn as a concave relationship against
wealth, with diminishing marginal utility of money. If oﬀered a choice
between the gamble in Figure 13.1, oﬀering expected utility E(U), and a
‘sure thing’ that pays W with certainty, where monetary returns are
equal on average, a risk-averse person will always reject the gamble since
U(W) E(U). A risk-averse person prefers a constant level of wealth, W,
to a set of gambling opportunities in which W is the average level of wealth.
For example, if a coin is flipped and you receive 18 cents if the coin comes
up heads and six cents if it turns up tails the expected outcome is (0.5*18)
(0.5*6) 12 and this will be rejected in favour of receiving 12 cents with
certainty.
Risk aversion is equivalent to diminishing marginal utility of income.
Essentially, a risk-averse person gains less extra satisfaction from a $100
gain compared with the drop in happiness from a $100 loss. Even a ‘fair’
bet, as in the example above, where the individual’s expected utility is held
constant over a series of wagers and there is no net loss, is rejected by the
risk-averse individual. Of course, the existence of bookmaker commissions
and other transactions costs (such as taxes) will usually mean that ‘fair’ bets
are not observed in practice.
Expected utility theory is well-established in economics. Although it has
been challenged theoretically and empirically (Rabin and Thaler, 2001),
expected utility theory remains the economist’s workhorse for dealing with

Economics of gambling

371

diﬃcult problems involving uncertainty. Its main advantages are simplicity
and ability to generate hypotheses for empirical testing. Expected utility
theory runs into some problems when trying to explain gambling behaviour. First, there is the puzzle that seemingly risk-averse individuals simultaneously gamble and buy insurance (Friedman and Savage, 1948). Only a
risk-averse person would be prepared to pay a premium to reduce a risk by
purchasing insurance. Someone who buys a lottery ticket eﬀectively pays to
take on risk, since the expected returns are known to be negative. This is
risk-loving behaviour. So, how can a person be both risk-averse and riskloving at the same time? One answer, which economists quickly reject, is
that many people are ignorant of the terms of the bet. Irrationality is not
an appealing assumption and evidence can be marshalled to demonstrate
that people who gamble make the best use of available information when
gambling, in accordance with the rational expectations hypothesis. In the
case of lottery tickets, Forrest et al. (2000a) could not reject rational expectations on the part of lottery players in the UK.
Friedman and Savage proposed that people may be risk-averse when
faced with the prospect of a substantial loss out of a modest asset portfolio
(for example, the risk of fire or theft to one’s property). Faced with a small
chance of a life-changing gain, such as from winning the jackpot prize in a
lottery draw, people may be risk-loving.
This resolution of the gambling–insurance puzzle leads to a further
problem. If people are risk-loving with respect to lotteries, then a profitmaximizing lottery operator would only oﬀer one prize, the jackpot. A riskloving person oﬀered two gambles with the same average value and
diﬀerent spreads would prefer the one with greater spread. He or she would
be prepared to pay more for a larger spread gamble than a smaller one.
Hence, lottery operators would make larger profits by oﬀering a single large
prize than by oﬀering several large ones. Yet, lottery draws usually oﬀer a
range of prizes, not just the top jackpot prize. Horse race betting pays out
winnings for second and third placed horses.
The primary diﬃculty with applying expected utility theory to gambling
behaviour is that it views gambling as a speculative investment. But it is
plausible to argue that gambling confers direct consumption benefits to the
player. These benefits would comprise the fun involved in talking to friends
and work colleagues about how to spend a lottery jackpot prize, and associated anticipated pleasure of giving notice to an employer, and the thrill
involved with betting on a horse or sports team. Such dreams are used in
lottery advertising, hence the slogan ‘your ticket outta here’ found in
New Jersey. This thrill of gambling gives added pleasure to watching a
horse race or sports fixture on television or at the venue. The ‘fun’ associated with gambling, even where players lose on average, has been used to

372

Economics uncut

motivate consumer behaviour in lotteries in a number of studies such as
Forrest et al. (2002), Kearney (2002) and Farrell and Walker (1999). The
emphasis in such studies is on consumption benefits from gambling rather
than returns to gambling as an investment.
Many people engage in a lot of ‘gambling-like’ activities that do not
involve real money, such as playing bridge. This is additional evidence on
the associated consumption value of gambling.
Given that individuals decide to gamble, their consumer demand for
particular gambling products can be couched in terms of the traditional economic variables of prices, income, price of substitute goods and market size.
The notion of a price of a gamble is subtle since it is linked to expected
value of winnings. Generally, the price of a gamble can be expressed as
the price paid to make a $1 wager. That is, how much should be spent on
gambling, on average, to make a $1 return. This price is termed the takeout (one minus expected value of a gamble). In casino gaming the take-out
is also the house advantage. Within casino gaming, Eadington (1999)
shows that the lowest house advantage occurs in blackjack, baccarat and
craps, while the highest house advantage accrues to keno, American
roulette and slots. Within casino gaming, though, competition influences
the size of house advantage and it is notable that Indian casinos placed on
Indian reservations tend to oﬀer larger house advantage than Las Vegas
casinos, and are highly profitable, partly by virtue of their remote locations
(Anders, 2003).
The demand for casino gambling has recently been investigated in a traditional economic formulation by Thalheimer and Ali (2003). They
obtained data on slot machine wagering at 27 venues present at various times
over the period 1991–98, giving 153 observations. They measure demand for
slot machine gambling by amounts wagered annually on slot machines at
US casinos located either on riverboats or at racetracks (‘racinos’) as a proportion of market area population. This measure is expressed at constant
prices and in logarithms and then used as a dependent variable in a least
squares regression. From their estimates, Thalheimer and Ali find that
higher take-outs lead to lower amounts of slot machine wagering with
elastic demand at relatively high take-out rates. Beyond US $16 500 annual
per capita income, higher income is associated with greater slot machine
wagering so that casino gambling is a normal good at higher income levels.
Access is an important variable determining slot machine gambling.
Greater access from a market area, in terms of shorter distances travelled
to casinos, raises slot machine wagering. The presence of competing riverboats or of Indian casinos, as substitutes, lowers slot machine wagering.
These results are very much in accord with prior economic reasoning. The
eﬀects of regulation are pronounced as casinos with stricter loss limits and

Economics of gambling

373

those with greater restrictions on boarding times have lower amounts of
slot machine wagering.
Demand for lottery tickets has been investigated by a number of
researchers (Forrest, 2003). The empirical procedure established by Gulley
and Scott (1993) takes the nominal price of a lottery ticket as fixed but estimates the eﬀective price of a unit stake. This is computed as one minus the
expected value associated with winning the jackpot prize. In general,
expected value (EV) is given by:
EVp*JACKPOT *SHARE expected value of smaller prizes
where p is the probability of winning the jackpot (1/14 000 000 in the
6/49 game operated in several jurisdictions, including the UK), JACKPOT
is the value of the jackpot and SHARE is the expected share of the jackpot
of a winning ticket, dependent on the distribution of number selections of
other bettors. More specifically:
EV [1/N][R(1t)N][1epN]expected value of smaller prizes
where N is level of sales for current drawing, R is the rollover from previous drawing, when the jackpot is not won and is carried forward, t is percentage of wagers not returned to bettors in prizes.
The empirical procedure is to estimate EV, compute 1 EV as the
eﬀective price and then estimate lottery sales. The rollover value plays a critical role in the determination of expected value and hence eﬀective price.
Rollovers lower the eﬀective price of play. The rollover is like ‘free’ money
in the prize fund but it must be noted that the benefit of rollover is dispersed
more thinly the higher are sales so, in the limit, eﬀective price with rollover
converges to (and does not fall below) take-out rate as sales increase
(Forrest, 2003). When rollovers occur, demand for lottery tickets rises in
response to the eﬀective price reduction. Generally, the take-out is rather
high in lottery games: a typical game will only return around 50 cents out
of a $1 stake to the bettor in the form of prizes.
Various studies, summarized in Forrest (2003), have estimated the price
elasticity of demand for lottery tickets, both for US states and for the UK.
Results suggest that this elasticity is moderately above one in absolute value
for US states and around one in the UK (Forrest et al., 2000b). A key
diﬀerence between these jurisdictions is that the UK National Lottery is
operated by a private franchise, whereas US state lotteries are typically run
by the state authorities. In the UK case, unitary elasticity is consistent with
revenue-maximizing behaviour by the operator and is therefore in accord
with UK government objectives. In US state lotteries, the absolute price

374

Economics uncut

elasticity that maximizes revenues for the state is 1.2 and evidence of
whether this is actually achieved is mixed (Gulley and Scott, 1993).
However, referring back to the consumption motive for gamble, the size
of the predicted jackpot prize, rather than eﬀective price, may be the
primary economic variable aﬀecting sales. Evidence by Forrest et al. (2002)
was unable to distinguish between these two eﬀects as the expected jackpot
prize and eﬀective price were highly correlated in UK lottery sales data.
A feature of consumer demand that has been largely ignored in the economic literature on gambling is the increased demand for variety that goes
with rising incomes and rising demand for leisure. Racing, casino gaming
and sports betting can be viewed as entertainment products that compete
with other leisure pursuits. A night out at a casino in the USA or Britain is
an alternative to a visit to the cinema or a show or a sports event. This raises
the possibility of an ‘option value’ of gambling; I may not actually visit a
nearby casino but if I tire of my normal leisure pursuits the option of
casino gambling is available.
Therefore, it is reasonable to suggest that gambling products, on the
whole, are normal goods. For example, Paton et al. (2004) estimate empirically that income (average earnings) has a significantly positive impact on
betting demand in the UK. As consumer incomes rise, the demand for
leisure will grow and the demand for gambling will continue to rise. Of
course, this still leaves scope for demand variation within the sector, some
of which is traceable to demographic influence. In the UK, for example, it
is clear that demand for horse and greyhound betting is falling, while
demand for soccer and sports betting is on a rising trend as younger gamblers prefer betting on individual and team sports. Paton et al. (2004) argue
that the introduction of the UK National Lottery did not reduce demand
for bookmaker betting but there was evidence of price-driven substitution
between purchases of lottery tickets and bookmaker betting. However, the
latter conclusion is based on use of monthly betting data, which are excessively aggregated for the purpose of the authors’ study. Within the UK
National Lottery, Forrest et al. (2002) find some evidence that the main
lottery draws and instant scratchcard games are partial substitutes but
own-game characteristics are the largest influence on sales. Hence, sales of
instant scratchcards do not seriously cannibalize lottery draw sales.
A further type of substitution involving lotteries that needs to be considered is the impact on charitable giving. When the UK National Lottery
was launched in 1994, various charities complained that this would
result in lower charitable donations. Tanner (2002) finds that, when asked,
‘Which reason most explains why you play the UK National Lottery?’,
only 3 per cent of respondents replied ‘yes’ to the option, ‘to contribute to
the good causes that the Lottery supports’, notwithstanding the tendency

Economics of gambling

375

of the National Lottery operator to use examples of funding for ‘good
causes’ in television advertising. Tanner also used survey data to determine
the participation and extent of charitable giving using information on
peoples’ age, income, education and wider macroeconomic conditions. She
finds that, comparing survey responses before and after the 1994 UK
National Lottery launch, neither the number of people giving to charity
nor the level of donations changed significantly following the Lottery’s
introduction.
In the USA, Levitsky et al. (2000) oﬀer evidence on monthly gaming revenues by source from 1988 through to 1995 from Clark County, Nevada,
which contains Las Vegas. They show that although slot machines contributed to county gaming revenue at an increasing rate over time, the
impact of table games on gaming revenues was declining. Siegel and
Anders (2001) find evidence of revenue displacement from lotteries to
casinos in Arizona, especially for games oﬀering big prizes. Thalheimer and
Ali (1995) find that the introduction of US state lotteries reduced horse
betting demand.
What these empirical studies show is that when gambling is deregulated,
consumers switch demand for gambling between diﬀerent products.
Demand for gambling as a whole tends to increase, but there are also cannibalization eﬀects as some older gambling products tend to be displaced
by new products. Hence, intra-sectoral variations in demand for gambling
do occur and over time some sectors decline while others grow.
In addition to cannibalization eﬀects, there are also conventional substitution eﬀects as the impact of a price change in one sector spills over into
demand in another sector. Such impacts are diﬃcult to discern in gambling
since the eﬀective price of a bet for, say, fixed-odds horse racing, may not
vary much over time. In the lottery sector, though, variations in eﬀective
price occur due to rollovers. Forrest et al. (2004) model daily bookmaker
betting turnovers by sector and find that a lottery rollover induces substitution away from horse race, soccer and numbers betting. Paton et al.
(2003) model monthly bookmaker betting and also conclude that there is
strong evidence of positive cross-price elasticities across gambling products; when take-outs fall in one sector, sales will increase in that sector but
demand falls in other sectors. Hence, intra-sectoral variations in demand
for gambling do occur and over time some sectors decline while others
grow. Against the rise in gambling revenues due to market expansion, must
be oﬀset some negative eﬀects as take-outs fall in competing sectors. The
size of these negative eﬀects is still open for debate, particularly as some
household-based evidence concludes that lottery spending does not substitute for other forms of gambling (Kearney, 2002). Kearney finds instead
that household lottery gambling crowds out other household consumption

376

Economics uncut

by around 2 per cent. Moreover, larger proportional reductions were found
among low-income households.
Overall, though, the demand for gambling is rising, as shown in Table 13.1.
Given the general rise in demand for gambling, how should authorities
regulate gambling? Some commentators argue that increased demand
would generate social costs in excess of social benefits and further liberalization of gambling should not be encouraged. Legalization of gambling is
considered next.
3 Why tolerate gambling?
Legalization of gambling typically meets with moral, religious and political opposition. In Tennessee, a referendum was carried out in November
2002 on a decision over whether to enable federal government legislature to
install a state lottery. This referendum was conducted after a 17-year-long
campaign by Senator Steve Cohen and resulted in a 58:42 majority approving progress towards legislation. There were several important ingredients
to the successful Tennessee pro-lottery campaign. First, Tennessee is surrounded by states that already oﬀer lotteries. It was recognized that potential state tax revenues were being shifted across state borders as lottery
players travelled to buy lottery tickets and the pro-lottery campaigners
made this a powerful pragmatic argument for lottery adoption. It was
alleged that an annual loss of US $76 million of potential Tennessee revenues was incurred to the Kentucky lottery alone in the form of crossborder sales. Second, Tennessee had a fiscal deficit and its citizens appeared
reluctant to vote for increases in sales or income tax. Hence, proceeds from
lottery sales could in principle raise state revenues and alleviate any
future increases in state taxes. Third, the pro-lottery campaigners drew
upon the examples of Florida and Georgia and pledged that net revenues
from a Tennessee state lottery would be devoted entirely to support specific educational programmes. These included university scholarships
along the lines of the Georgia lottery-financed, merit-based HOPE scheme
(HOPE is the acronym for Helping Outstanding Pupils Educationally).
Pre-kindergarten nursery education would also benefit from lottery finance.
Hence, the pro-lottery campaigners tried successfully to tap into the altruism of the Tennessee electorate, even though such altruism did not seem to
extend to voting for tax raises.
This example shows some of the key economic and political influences
upon lottery adoption. As shown by Kearney (2002), lottery adoption in
the USA can be traced geographically from the north-east, with the first
state lottery in New Hampshire in 1964, through to the west, then to the
midwest and eventually to the south. A bandwagon eﬀect on US lottery
adoption seems clearly pronounced.

Economics of gambling

377

Sauer (2001) oﬀers an interest group model of gambling regulation,
which seems to be well-suited to the US experience of lottery and casino
legalization. He notes that gambling markets are, ‘the most restricted and
politicised markets in the American economy’. For simplicity, Sauer’s
model has two groups. One has a taste for gambling and the other opposes
gambling on religious and moral grounds. It is notable that vehement
opposition to legalization of gambling in the USA has come from leaders
of church groups, especially Baptists. The opposition group gains satisfaction (welfare) from restrictions on gambling. But state lotteries, and also
casinos, oﬀer a useful means of raising revenue. Note that the image of
casinos as associated with ‘hard’ gambling makes legalization of casinos
more diﬃcult than legalization of lotteries, which (allegedly) represent
‘soft’ gambling and possibly capture a broader set of customers.
According to Sauer, states with higher tax burdens are more likely to be
lottery states. Among lottery states, those with higher tax burdens get lotteries earlier than others. In equilibrium, the marginal value to the antigambling group of restricting gambling equals the marginal welfare cost on
the pro-gambling group. A rise in the consumption value of gambling,
which could come from rising incomes as argued above or a change in
tastes, reduces the equilibrium level of restrictions. A rise in social costs of
restrictions, for example from crime, would raise the level of restrictions.
For the anti-gambling lobby, there are some trade-oﬀs to consider.
Religious groups often solicit contributions via church bingo and lotteries.
Church people like education and transport facilities just as pro-gambling
citizens do and will be prepared to reduce opposition to lotteries if the
opportunity cost of lottery finance rises.
The advantage to state governments of legalized gambling is that, in the
context of growing demands on government revenues, removal of prohibition permits high marginal returns as new tax revenue flows in. Extra tax
revenue is easier to capture from a new source than by raising existing taxes.
It is notable that some states (Arizona, Colorado, California), which had
large-scale property tax revolts in the 1980s subsequently passed bills to
install state lotteries.
In summary, Sauer states that, ‘as long as the shadow price of
Government revenue remains high, political equilibrium requires that high
cost revenue sources such as gambling markets will be utilised as a means
of public finance’.
Empirical evidence oﬀered by Erekson et al. (1999) is consistent with
Sauer’s application of public choice theory to lottery adoption. Their evidence shows that the probability of lottery adoption in 50 US states in any
year over 1962 to 1990 varies positively with the size of state government
deficit in the previous year, level of per capita income and the percentage

378

Economics uncut

of contiguous states with established lotteries. The probability of lottery
adoption is lower, the greater is state education spending per pupil and the
greater is the percentage of state population that is Protestant. Intriguingly,
a variable representing incidence of Baptist or fundamentalist religious
groups, as a sub-set of Protestants, was not included in the study although
earlier studies had revealed these as being statistically negatively significant
in explaining lottery adoption (Caudill et al., 1995). Alternatively, other
evidence suggests that Catholics are more prone to support legalized gambling, including lottery adoption (Alm et al., 1993). Hence, legislation to
adopt lotteries is likely to be passed when fiscal stress is high, particularly
if neighbouring states have already adopted lotteries. These two important
economic forces were clearly paramount in the majority decision by the
Tennessee electorate to vote for lottery adoption.
The public choice model of gambling regulation has great appeal and
force in explaining US state lottery adoption. However, gambling services
are not homogeneous. In the USA, it is possible for citizens to simultaneously vote for lottery adoption yet voice fears, and reject, legalization of
Internet betting. The public choice model is incomplete in that tolerance of
one form of gambling (lotteries) may coexist with resistance to other forms
(casino gambling on the Internet).
In the UK, progress is currently (2005) being made towards further gambling deregulation as the government has produced a White Paper outlining specific proposals. Particularly notable are proposals to relax licensing
restrictions on casino openings and to permit casinos to oﬀer entertainment and to install more slot machines than are permitted at present.
Customers would have immediate access to casinos rather than facing a
24-hour wait for membership as at present. The eventual consequence is
likely to be a growth in the number of casinos operating in the UK but this
will take time. Unlike the USA, religious opposition has not been a strong
factor in the regulation of gambling in the UK. Nevertheless, the UK government softened its initial plans to introduce a substantial number of
‘resort casinos’ modelled along the lines of the large Las Vegas hotel-casino
leisure complexes. The number of resort casinos is initially to be limited to
just one compared to 24 in earlier proposals. This is likely to confer monopoly rents to the fortunate owner of this facility. Also, the capacity for large
jackpot slot machines has been substantially reduced compared to initial
plans. In line with the public choice theory of Sauer (2001), noted earlier,
this climbdown followed increased lobbying by groups concerned with
problem gambling. Such lobbying was supported by elements of the industry likely to be adversely aﬀected by casino deregulation, in particular,
bingo operators and amusement arcade owners. UK governments have typically taken a pragmatic approach to regulation of gambling with an under-

Economics of gambling

379

lying motive to satisfy, but not induce, latent demand and, further, to
extract tax revenues from gambling sources in the form of specific levies on
gambling products. Until October 2001, gamblers placing bets in British
retail outlets were faced with a tax of 9 per cent on winnings, part of which
constituted a levy to help finance the horse race industry.
4 Issues
The growth of the gambling sector generates a number of policy issues. In
this section we oﬀer a quick review of some interesting problems of regulation facing policy-makers around the world.
4.1 Gambling as economic regenerator?
For most economists, the fact that gambling is provided to meet consumer
demand is in itself welfare-enhancing in an economy, provided that consumers are not systematically misinformed. The evidence on lottery sales
from Forrest et al. (2000a) and Kearney (2002) points to the likelihood that
consumers of lottery products are making well-informed purchases. If purchases of gambling products add to consumer utility, for those who wish to
participate, then this could be taken as suﬃcient reason to permit the availability of gambling.
Some would go further than the consumer welfare argument and
propose that gambling facilities confer wider social benefits in the form of
higher economic growth, higher employment rates and economic regeneration in previously rundown urban areas. Such claims are often wildly exaggerated. The cited expenditure and income eﬀects consist of direct and
indirect employment eﬀects, such that total employment eﬀects are multiplied up from the first round impact, and gains in value added. Anders
(2003) points to a study paid for by the Arizona Indian Gaming
Association on the economic impacts of Indian casinos in Arizona. It was
proposed that these casinos generated 9300 jobs and had a total impact of
US $468 million.
These numbers should be treated with caution. The sizes of these multipliers are most likely biased upwards for a number of reasons. First, consultancy reports often deliver numbers that validate the purpose for which
the consultants were employed and which might facilitate repeat business
for the firms involved. Beneficial estimates are thereby inflated. Second,
impacts are based on predicted rather than actual gaming revenues. Third,
price eﬀects are typically ignored. Prices of local services and of property
may increase. In the case of Las Vegas, wider economic benefits were
obtained by synergy between gambling, tourism, hotel accommodation
and the entertainment industry as large corporations developed the concept
of a destination gambling resort with wider facilities than just casinos

380

Economics uncut

(Eadington, 1999). The rapid growth and huge popularity of Las Vegas is
testimony to the skill applied to this project, particularly as the earlier
image of Las Vegas was of a crime-ridden city ruled by mob gangsters.
But Las Vegas is a hard act to follow. In Britain, local authorities and
leisure companies have worked on a scheme to take advantage of forthcoming casino deregulation to set up a Las Vegas-style resort casino in
Blackpool, a down-at-heel working class seaside resort on the north-west
coast of England. Blackpool lacks the hot sunshine of Las Vegas but, more
importantly, it also lacks the exclusivity applied to gambling in Nevada.
For US citizens, Las Vegas and Atlantic City are the only resort-based gambling centres available. In England, any attempt to render Blackpool an
exclusive destination for casino gambling would be met with a legitimate
outcry from competitors. As noted above, the UK government has set out
proposals to restrict the number of resort casinos to eight. This would most
likely lead to economic rents for large casino operators with consumer
welfare reduced as take-out rates are raised above levels that would be associated with greater competition.
4.2 Social costs
Social costs of gambling such as crime, bankruptcy, divorce and suicide are
much cited side-eﬀects of gambling. Grinols and Mustard (2001) claim that
a county with a casino had about 8 per cent higher crime rates than a
county without a casino four years after the opening of a casino. However,
we should be wary of inferring too much from this claim. Additional crime
is stimulated by any new leisure facility, not just a casino. This is because
new visitors arrive and represent new targets for criminals. Also, new criminals will move into an area where new spending occurs. This will apply for
a new sports stadium as much as a casino. To then impute all social costs
of crime to gambling seems unreasonable, unless the investigator controls
for all possible sources of crime. It is apparent that the Grinols and
Mustard study is lacking in a full set of control variables. Other researchers
oﬀer more modest estimates on the relationship between gambling and
crime (see Stitt et al., 2003).
The most serious social costs considered in the gambling literature relate
to pathological or problem gambling. According to Grinols (2001), problem
gamblers (defined as a broader category than pathological gamblers) represent 2.8 per cent of the US population and generate social costs in the form
of bankruptcy, suicide, depression and support costs from counselling and
medical services of the order of US $18 000 per problem gambler per annum.
The existence of these costs must be acknowledged. For instance, Stitt
et al. (2003) compared the percentage increase of personal bankruptcies per
capita for casino jurisdictions with the percentage increase in bankruptcies

Economics of gambling

381

for paired non-casino jurisdictions over 1989 to 1998. They found that the
introduction of casino gaming was associated with a significant increase in
bankruptcy rates in five out of eight counties with casino openings.
But are these kinds of costs really ‘social’? It is hard to see how gambling
can confer social costs of the same order of magnitude as narcotics,
smoking or alcoholism. The external eﬀects such as death by secondary
smoking or drunk driving are not so clear in the case of gambling. Much
of the social cost element involves distress within households as problem
gamblers run down household assets to support their habit and incur
depression and other illnesses that impair family life. Walker and Barnett
(2000) make the extra leap to argue that problem gambling does not really
impose social costs. The costs of personal bankruptcy, depression and
divorce amount to intra-household transfers rather than external, social
costs. Of course, the costs of dealing with medical and other disorders associated with problem gambling are external to the extent that counselling
and other services are provided to help problem gamblers. But most of the
costs identified by Grinols as ‘social’ are actually transfers within households according to Walker and Barnett. On this view, the social costs of
gambling are not severe enough to warrant continued opposition to new
forms of gambling.
Although Walker and Barnett have a point, problem gambling is recognized by governments and industry sources as coming from a physiological
addiction and various governments have set up funding to research into
problem gambling. Support for problem gambling in the forms of additional research and medical and counselling facilities to help problem gamblers were identified as an essential feature of the UK government’s
proposals to deregulate gambling.
A strand of economics research relevant to concerns with ‘problem gambling’ emanates from the Becker–Murphy rational addiction theory that we
discussed in Chapter 1. Recall that in this approach, consumer demand for
addictive goods depends on both past consumption and expected future
consumption. Consumption is aﬀected by exogenous shocks such as
changes in taxation or product market regulation. Following these shocks,
two equilibria may be derived: a high use equilibrium (‘addiction’) and a
low or zero use equilibrium (‘abstention’).
The rational addiction theory seems applicable to some, but not all, gambling products. The wide incidence of participation in lotto games suggests
that lotto demand is not consistent with rational addiction (Forrest, 2003).
Informal analysis of scratchcard purchases, however, does reveal some consumers with high rates of purchase.
Mobilia (1993) tested the rational addiction hypothesis using a large data
set of racetracks in the USA over the 1950–86 period. Using take-out as the

382

Economics uncut

price variable and real amounts wagered by attendees she found, in line
with studies of rational addiction for cigarettes, that the long-run price
elasticity was –0.7, more than double the short-run elasticity. An increase
in current take-out rate was observed to reduce the real amount bet per
attendee in past and future years. Similarly, Nichols (1998) modelled
monthly wagering in the Iowa riverboat casino industry and found that
long-run price elasticity (again, using take-out as the price variable) was
almost double the short-run measure.
The findings of Mobilia and Nichols may suggest that, if certain gambling products are deregulated or if take-out is reduced, that the incidence
of addiction, and hence social costs, would increase. There are three
important caveats, however. First, the long-run price elasticities reported in
these studies are not particularly large; the largest estimate is –1.2 in the
Nichols study. Second, the Becker–Murphy model posits utility-maximizing behaviour on the part of addicted consumers, regardless of any social
disapproval, and if these are in some way constrained by government policy
to have their pleasure denied, then consumer surplus is lost. Third, the relationship between increased incidence of addiction and social costs needs to
be carefully quantified, rather than assumed, and careful empirical research
is needed before welfare and policy judgements can be made.
4.3 Illegal betting
If prohibition of some forms of gambling is sustained, it does not follow
that problem gambling is suppressed. The activity simply goes underground into the illegal sector where it is less easily monitored and problems
such as uncontrolled debt and risk of physical violence become magnified.
The scale of illegal activity in the USA is vast. With an estimated volume
of wagering at US $100 billion and gross revenues of US $10 billion, the
illegal sports betting market dominates the legal market. This is not surprising. The federal US government declared sports betting illegal in 1992,
although Delaware, Nevada and Oregon were exempted, and now the legal
sports betting market is concentrated in Las Vegas.
Strumpf (2003) provides a fascinating account of illegal sports betting in
New York. His study considers betting records of six convicted, illegal bookmakers. It appears that these bookmakers operated like orthodox firms.
They operated incentive contracts to regulate employee behaviour and
developed relationships of trust and repeat business with their customers.
These in turn gambled large sums of the order of between US $200 000 and
US $400 000 per bettor in the larger operations. The bookmakers were essentially price-takers, accepting spreads and odds from Las Vegas. The bookmakers incurred substantial financial risk and could be characterized as
risk-neutral or risk-averse. Bettors were subject to, and were prepared to

Economics of gambling

383

accept, some price discrimination in that less fair prices were oﬀered to bets
on particular local teams such as the New York Yankees. Hence, sentiment
was a feature of betting in this part of the illegal sector.
If this sector were legalized, nationwide operations would probably
absorb small, illegal bookmakers but would, initially at least, wish to tap
into the customer relationships established with the bettors involved in the
illegal sector. Tax revenue would undoubtedly increase, but an illegal fringe,
with tax avoidance, could persist as occurred in the alcoholic beverages
sector in the USA following the end of Prohibition. The personal relationships between illegal bookmakers and their customers have recently been
used by Internet bookmakers to develop information on creditworthiness
and supervision of credit lines so as to oﬀset (legal) bettor default.
The sports betting sector is surely ripe for legalization, yet there is little
pressure within the USA for this to happen. Under a special legal provision,
Atlantic City was oﬀered a two-year period when casinos were opened
there, so as to open sports books, but supply was not forthcoming. The
sports betting market in Nevada is highly competitive and margins are very
low. Hence, the pressure for legal entry into this market is weak.
4.4 The threat of the Internet?
Not all betting is about to be conducted on the Internet. According to Strumpf
(2003), ‘many (illegal) bettors would be unable to place an internet wager
simply because they can’t operate a computer’. Similarly, in the UK, traditional betting in the form of visits to a retail outlet (betting shop) will continue
to be the mainstay of the bookmaking business for some time to come.
Nevertheless, the role of the Internet in shaping the gambling industry
has already been substantial. The Informa Media Group predicts that
Internet betting in the UK is forecast to grow to £1.4 billion in 2006 from
£250 million in 2002. Gambling will be given a large boost from technological change as betting opportunities become available from computers,
3G cell phones and interactive TV. According to Datamonitor, interactive
TV gambling in Europe will generate US $11 billion in revenue by 2006.
In the USA, there is no legal recourse of Internet gambling operations
to bettor default. Interestingly, predictions suggest that Europe, and not
North America, will take the dominant share of e-gambling business over
the next few years. A key reason for this is that, in the USA, Internet gambling is essentially illegal. It is unclear whether Internet gambling is technically illegal in the USA. The Federal Wire Act, originating in 1962,
prohibits the transmission of wagering information on sports across state
lines. Confusion exists over the scope of this Act and whether it extends to
Internet betting, rather than wireless technology. US Congress has made
attempts to pass Internet prohibition acts without success. But most US

384

Economics uncut

credit card companies will not honour Internet gambling and this overrides
any law. Internet casinos are absent in the USA and legal Internet casinos,
such as those located in the Isle of Man or Australia, will refuse to accept
wagers from the USA due to its confused legal status. Some locations, such
as Belize, do have Internet sites that take wagers from the USA, particularly
in the form of sports betting, and these will proliferate and generate
increasing volumes of Internet business in the immediate future.
US Internet gambling is unlikely to be legalized in the near future.
Standard concerns about problem gambling and underage gambling have
been raised. Although it is clear that it would be welfare-improving to legalize and regulate Internet gambling, rather than have money flow abroad or
into the illegal sector, any campaign to legalize Internet gambling in
the USA would not carry the same level of support as found for lottery
adoption. In part, this is due to a distaste for Internet gambling by nonparticipants (puritanism) and also a perception shared by the public and
policy-makers that Internet gambling intrudes upon private space by permitting ‘living room casinos’.
In the UK and Australia, more eﬀort has been made to legalize and regulate Internet gambling. Indeed, the UK government explicitly recognized
the legality of Internet gambling in its recent White Paper. The problem is
how to harness Internet gambling to raise tax revenues. Paton et al. (2002)
show how one policy response to the growth of Internet gambling may
benefit consumers, producers and government in a welfare-improving
manner. Prior to October 2001, the UK government levied a tax on betting
as a proportion of winnings. In October 2001, this was replaced by a
15 per cent tax on bookmakers’ gross revenues. The rationale for this
change was that British bookmakers had shifted some parts of their operations oﬀshore, to the Channel Islands and Gibraltar, where tax regulations
were far less stringent. These bookmakers established Internet facilities for
betting following the lead of Victor Chandler’s bookmaking operation. As
a result, the UK government was faced with potentially severe reduction in
tax revenue from betting, as custom was likely to move oﬀshore. Given
Strumpf’s quote above, the leakage could well have taken a long time to
occur but the UK government held negotiations with leading bookmakers
and agreed a switch in tax policy. In return, the bookmakers agreed to move
their oﬀshore operations back onshore. This switch of tax policy towards
profits taxation rather than sales taxation could perhaps be imitated in
other jurisdictions. Thus far, bookmakers have claimed increased turnover
following the change in tax regime. It is worth noting that, through the
period in the late 1990s when oﬀshore Internet betting was developed, some
evidence suggests that the UK fixed-odds soccer betting market moved
closer to informational eﬃciency. Objective probabilities of game out-

Economics of gambling

385

comes became more closely aligned with ex ante betting odds oﬀered by
leading bookmakers (Forrest and Simmons, 2002).
This new tax policy still leaves the UK government with problems. It
plans to legalize online gaming, permitting punters to identify onshore and
oﬀshore sites. Online operators will be subject to approval and monitoring
to ensure compliance with regulations. UK applicants for legalized online
gaming must be registered as a UK company, have a UK server and a co.uk
domain address. This seems fine but the government has still not addressed
the issue of tax treatment of Internet gambling.
Furthermore, the UK government still faces the threat of migration of
gambling business, and hence potential tax revenue, oﬀshore. Looking
at casinos, the current UK position is that profits tax and gaming duty add
7.5 per cent to overall UK tax on casino turnover. Other European locations, such as the Channel Islands, oﬀer much lower tax rates. There, all that
an oﬀshore gambling operation has to pay to establish itself is a small licence
fee. Further, Alderney in the Channel Islands has uncapped the number of
gambling licences for sale (Wilding, 2002). Hence, the UK may have forestalled the movement of its established bookmaking operations oﬀshore but
there is still scope for gambling to be established in low-tax destinations.
4.5 Gambling as social engineering
Lottery revenues are essentially an implicit sales tax on consumers; players
buy tickets and, after costs, revenue is returned to the government to spend
as they consider appropriate. US states operate their own lotteries and in
these cases the revenue goes direct to government coﬀers. In the UK, a
private franchise operates the lottery but taxation is levied on the operator’s
revenue from lottery sales. Of a £1 lottery ticket, 12 pence goes directly to
general tax revenue as ‘lottery duty’ while a further 28 pence is devoted to
‘Good Causes’ including promotion of sports, cultural and arts and heritage facilities.
As noted in Section 3, some US states have earmarked lottery revenues
for specific purposes such as educational programmes. The Tennessee government aims to use its forthcoming lottery to promote college scholarships
and to finance pre-kindergarten nurseries. Such schemes have been judged
as wasteful and misdirected by several economists.
Rubenstein and Scafidi (2002) and Borg and Stranahan (2000) show that
lottery taxation is highly regressive. Spending the proceeds of lottery taxation on college scholarships, as in the Georgia HOPE programme, makes
this regressivity worse. The benefits of HOPE scholarships are shown to
have been received disproportionately by higher income and more educated
households. This is not surprising as most college students come from this
type of background. Lower income and non-white households tended to

386

Economics uncut

have higher purchases of lottery products while receiving lower benefits as
compared with higher income and white households.
Dynarski (2000) shows that Georgia’s HOPE programme widened the
gap in college attendance between blacks and whites and between those
from low- and high-income families. Given lower marginal tax rates, both
sales and income taxes would be much more eﬃcient revenue sources
than the implicit lottery tax. State income tax could be raised from 3 to
3.39 per cent or sales tax could be increased from 4 to 4.69 per cent. Either
option would be less regressive than the use of lottery revenues.
The use of lottery revenues to support particular purposes violates the
horizontal equity principle of public finance. It is also wasteful. Farrell and
Walker (1999) estimate that the deadweight welfare loss from UK lottery
taxation devoted to ‘Good Causes’ is of the order of 35 per cent of the
revenue raised. This suggests that this means of raising money for ‘Good
Causes’ is particularly ineﬃcient. Since lottery products display income
inelasticity this form of taxation is also regressive.
Where lottery revenues are devoted to particular projects, there is evidence from the UK National Lottery that this simply crowds out existing
expenditure (Bailey and Connolly, 1997). In Florida, money formerly dedicated to education was used elsewhere and replaced by lottery revenues with
no net increase in education spending. If lottery-financed public spending
is indeed additional, this is itself cause for concern. Forrest and Simmons
(2003) observe that expenditure on sports projects and other ‘Good Causes’
channelled via lottery funding escapes formal evaluation of the costs and
benefits of projects. Funds may be directed towards grandiose projects
(such as the Millennium Dome in London), which would not be funded if
in competition with other sectors for the use of general taxation.
5 Conclusion
These are interesting times for the gambling industry. Rising incomes in
advanced economies and associated rising demand for leisure has generated increased demand for gambling products. Consumers demand variety
in leisure and gambling opportunities have proliferated in the last decade
across the USA and Europe. Within the gambling sector, some parts have
declined, such as horse racing in the USA and greyhound racing in the UK
whereas others, such as legal sports betting in the UK, have grown. On the
supply-side, technological change is a major factor in shaping new gambling opportunities. It appears that the USA is setting its face against legalization of Internet betting whereas other jurisdictions are permitting
Internet betting opportunities to benefit consumers of gambling and, of
course, to claim some part of the revenues for taxation. Gambling does
confer some externalities, such as increased crime and problem gambling,

Economics of gambling

387

and policy-makers need to identify the sources and scope of such external
eﬀects and design appropriate programmes to support individuals and
communities that are adversely aﬀected.
From a welfare economics standpoint it is surely better to legalize and regulate gambling rather than have revenues flow overseas or to illegal operators,
who currently thrive in the USA. Sauer’s public choice model of gambling
helps us understand how and when US states adopt lotteries but does not
explain why US citizens, and their politicians, can simultaneously vote for
lottery adoption yet show resistance to legalization of Internet gambling.
The continued growth of demand for gambling is a sure bet. Policymakers will need to reconsider how to manage this growth in demand for
gambling as an acceptable leisure pursuit. The opportunities for
researchers to investigate problems and issues in this field will grow alongside the latent demand for its various products.
References
Alm, J., M. McKee and M. Skidmore (1993), ‘Fiscal pressure, tax competition and the introduction of state lotteries’, National Tax Journal, 46 (4), 463–76.
Anders, G. (2003), ‘Reconsidering the economic impact of Indian casino gambling’, in
Leighton Vaughan Williams (ed.), The Economics of Gambling, London: Routledge.
Bailey, S. and S. Connolly (1997), ‘The National Lottery: a preliminary assessment of net
additionality’, Scottish Journal of Political Economy, 44 (1), 100–112.
Borg, M. and H. Stranahan (2000), ‘Lottery funded merit scholarships: some lessons from
the Florida Bright Futures scholarship program’, University of North Florida, working
paper.
Caudill, S., J. Ford, F. Mixon and T. Peng (1995), ‘A discrete-time hazard model of lottery
adoption’, Applied Economics, 27 (6), 555–61.
Christiansen Capital Advisors (2003), Gross Annual Wager Data, retrieved from
www.ccai.com.
Department for Culture, Media and Sport (2001), The Gambling Review Report, London.
Dynarski, S. (2000), ‘Hope for whom? Financial aid for the middle class and its impact on
college attendance’, National Tax Journal, 53 (3), 629–61.
Eadington, W. (1999), ‘The economics of casino gambling’, Journal of Economic Perspectives,
13 (3), 173–92.
Erekson, O., G. Platt, C. Whistler and A. Ziegert (1999), ‘Factors influencing the adoption of
state lotteries’, Applied Economics, 31 (7), 875–84.
Farrell, L. and I. Walker (1999), ‘The welfare eﬀects of Lotto: evidence from the UK’, Journal
of Public Economics, 72 (1), 92–120.
Forrest, D. (2003), ‘Time-series modelling of Lotto demand’, in Leighton Vaughan Williams
(ed.), The Economics of Gambling, London: Routledge.
Forrest, D. and R. Simmons (2002), ‘Globalisation and eﬃciency in the fixed-odds soccer
betting market’, University of Salford, working paper.
Forrest, D. and R. Simmons (2003), ‘Sport and gambling’, Oxford Review of Economic Policy,
19 (4), 598–611.
Forrest, D., O.D. Gulley and R. Simmons (2000a), ‘Testing for rational expectations in the
UK National Lottery’, Applied Economics, 32 (2), 315–26.
Forrest, D., O.D. Gulley and R. Simmons (2000b), ‘Elasticity of demand for UK National
Lottery tickets’, National Tax Journal, 53 (4), 853–63.
Forrest, D., O.D. Gulley and R. Simmons (2004), ‘Substitution between games in the UK
National Lottery’, Applied Economics, 36 (7), 645–51.

388

Economics uncut

Forrest, D., O.D. Gulley and R. Simmons (2005), ‘The relationship between betting and
lottery play: a high frequency time-series analysis’, Lancaster University Discussion Paper.
Forrest, D., R. Simmons and N. Chesters (2002), ‘Buying a dream: alternative models of
demand for lotto’, Economic Inquiry, 40 (3), 485–96.
Friedman, M. and L. Savage (1948), ‘The utility analysis of choices involving risk’, Journal of
Political Economy, 56 (4), 279–304.
Grinols, E. (2001), ‘Cutting the craps and cards: right thinking about gambling economics’,
University of Illinois, working paper.
Grinols, E. and D. Mustard (2001), ‘Business profitability versus social profitability: evaluating industries with externalities – the case of casinos’, Managerial and Decision Economics,
22 (1), 143–62.
Gtech Corporation (2000), The Vital Signs of Legalized Gambling in America: Gtech’s 8th
Annual National Gaming Survey, West Greenwich, RI.
Gulley, O.D. and F. Scott (1993), ‘The demand for wagering on state-operated lottery games’,
National Tax Journal, 45 (1), 13–22.
Kearney, M.S. (2002), ‘State lotteries and consumer behavior’, National Bureau of Economic
Research Working Paper No. 9330.
Levitsky, I., D. Assane and W. Robinson (2000), ‘Determinants of gaming revenue: the extent
of changing attitudes in the gaming industry’, Applied Economics Letters, 7 (4), 155–8.
Mobilia, P. (1993), ‘Gambling as a rational addiction’, Journal of Gambling Studies, 9 (3),
121–52.
Nichols, M. (1998), ‘Empirical analysis of casino gambling addiction’, University of Nevada,
Reno, working paper.
Paton, D., D. Siegel and L. Vaughan Williams (2002), ‘A policy response to the e-commerce
revolution: the case of betting taxation in the UK’, Economic Journal, 112 (480), 296–314.
Paton, D., D. Siegel and L. Vaughan Williams (2003), ‘The demand for gambling: a review’,
in Leighton Vaughan Williams (ed.), The Economics of Gambling, London: Routledge,
pp. 247–63.
Paton, D., D. Siegel and L. Vaughan Williams (2004), ‘A time series analysis of the demand
for gambling in the United Kingdom’, National Tax Journal, 57 (4), 847–61.
Rabin, M. and R. Thaler (2001), ‘Anomalies: risk aversion’, Journal of Economic Perspectives,
15 (1), 219–32.
Rubenstein, R. and B. Scafidi (2002), ‘Who pays and who benefits? Examining the distributional
consequences of the Georgia Lottery for education’, National Tax Journal, 55 (2), 223–38.
Sauer, R. (2001), ‘The political economy of gambling regulation’, Managerial and Decision
Economics, 22 (1), 5–15.
Siegel, D. and G. Anders (2001), ‘The impact of Indian casinos on state lotteries: A case study
of Arizona’, Public Finance Review, 29 (2), 139–47.
Stitt, B.G., M.W. Nichols and D. Giacopassi (2003), ‘Does the presence of casinos increase
crime? An examination of casino and control communities’, Crime and Delinquency, 49 (2),
253–84.
Strumpf, K. (2003), ‘Illegal sports bookmakers’, University of North Carolina at Chapel Hill,
working paper.
Tanner, S. (2002), ‘Charitable giving and the National Lottery’, (www.ifs.org.uk/
charities/lottery.shtml).
Thalheimer, R. and M. Ali (1995), ‘The demand for pari-mutuel horserace wagering and
attendance with special reference to racing quality, and competition from state lottery and
professional sports’, Management Science, 45 (1), 129–43.
Thalheimer, R. and M. Ali (2003), ‘The demand for casino gaming’, Applied Economics, 35
(8), 907–18.
Walker, D. and A. Barnett (2000), ‘The social costs of gambling: an economic perspective’,
Journal of Gambling Studies, 15 (3), 213–21.
Wilding, P. (2002), ‘Prospect – Cashing in on Internet gaming: El Dorado for the Channel
Islands’, (www.gamblingcontrol.org).

14 Economics of rock ’n’ roll
Simon W. Bowmaker, Ronnie J. Phillips
and Richard D. Johnson

The unique features of the music industry oﬀer an opportunity for the
economist to answer the basic questions of who gets what and why. At the
level of the ﬁrm, the industry is characterized by the entrepreneurial spirit
of the creative musician who seeks to produce music eﬃciently and of high
quality in order to satisfy consumer demand. At this level, the production
of music has much in common with a highly competitive market. However,
the perfectly competitive model found in textbooks ultimately fails to
explain the allocation of resources in the music industry because a distribution system that contains elements of monopoly stands between the
producers and consumers of music. Added to this mixture is rapid technological change that promises to bring new proﬁt opportunities for those
willing to take risks and be entrepreneurial, whilst at the same time undermine the traditional way of doing business.
This ﬁnal chapter proceeds as follows. Section 1 provides a brief overview
of the history of the music industry, revealing how technology has always
been both its friend and foe. Next, Section 2 illustrates the size and scope
of today’s global music market. Section 3 examines, ﬁrst, the complex relationship between the major record companies (the ‘Majors’) and the independents (the ‘Indies’), second, the economics of recording contracts, and
third, how technological change is aﬀecting the record label/artist relationship. Section 4 analyses the ‘superstar phenomenon’ in the music industry.
Which characteristics determine the success of an artist? To what extent
can superstar eﬀects explain the recent acceleration in concert ticket prices
in the USA? Section 5 investigates the economics of legal and illegal
copying of music. We draw upon economic theory to show how copying
does not necessarily have to be harmful to the copyright owner, before
examining the empirical evidence relating to the impact of piracy on legitimate music sales. Section 6 concludes the chapter.
1 A brief history of the music industry
The rise of popular music on a global basis is essentially a post-World War
II phenomenon. Despite the inventions of the radio and the phonograph,
record sales remained relatively low in the ﬁrst half of the twentieth
century. The ﬁrst signiﬁcant stimulus to growth occurred in 1948 when
389

390

Economics uncut

Columbia Records introduced the 12-inch vinylite long-playing (LP)
record onto the market in the USA. LPs played at 33 revolutions per
minute (rpm) and were much superior to the old 45 rpm and 78 rpm discs
in terms of sound quality and durability, as well as being cheaper to
produce (Strobl and Tucker, 2000). The emergence of new low-cost recording equipment in the 1950s allowed many Indies to compete with the
Majors of the era (Columbia, RCA and Decca). This provided an opportunity for a greater variety of artists to release music and increased demand
as consumers exploited the wider choice of music available (Strobl and
Tucker, 2000).
The 1960s was a creative decade in which many of the old formulas and
structures changed. Hi-ﬁ stereo recordings were ﬁrst introduced and the
development of tape formats that would eventually transform the industry
picked up pace in the mid-1960s with the appearance of the ‘4-track’ and
‘8-track’. The 8-track achieved a dominant position over the 4-track until
the cassette tape, introduced by the Philips company in 1963, eventually
took over in the 1970s. Cassette tapes were smaller, more convenient, relatively cheap to produce and, in combination with the development of
accompanying compatible software, transformed the production and consumption of music. The ‘album’ began to take precedence over the ‘single’
for artists and also for consumers as the growing teenage population spent
much of its historically high levels of disposable income on music (Strobl
and Tucker, 2000).
Growth of the industry continued until the recession of the early 1980s
when the demand for music fell. Many in the industry at this time believed
that the availability of home cassette recorders was responsible for much of
the loss in sales and there were attempts to tax blank recording cassette
tapes. Yet, some analysts suggest that the decline was predominantly related
to poor quality of the recordings, inferior software, the global recession
and a shortage of creative work. In addition, the diﬃculties may have been
exacerbated by internal or X-ineﬃciencies and a lack of proper cost controls throughout the industry (Strobl and Tucker, 2000).
The downturn in growth only proved temporary, however, and a period
of high growth followed, largely due to recording improvements and the
expanded use of digital formats such as the compact disc (CD). Because of
the low cost of production and strong market demand for high quality
recordings, record companies focused upon low-risk opportunities for
proﬁt. This took the form of issuing CDs of back catalogues and compilations of works by established artists. At the same time, the development
and growth of global satellite television channels catering for music lovers
only, such as MTV, also contributed to increased demand for music (Strobl
and Tucker, 2000).

Economics of rock ’n’ roll

391

By the late 1990s, a culture of ‘free’ music was emerging as the use of
the Internet expanded. In 1999, Shawn Fanning, then an 18-year-old
student at Northeastern University in Boston, USA, created a computer
programme called Napster, which allowed Internet users to search for
songs in MP3 format (‘near’ CD-quality sound) on the hard drives of
other users, and then download this music to their own hard drives at no
cost. Napster was an example of a peer-to-peer (P2P) network, although
it did operate with a central server, listing addresses so that users could ﬁnd
one another.
However, within only a few months of its existence, Napster was sued by
the Recording Industry Association of America (RIAA) on behalf of the
major record labels for copyright infringement. A ‘guilty’ verdict was
returned and in July 2001 it was forced to shut down, but not before it had
had a signiﬁcant impact on the music industry as this chapter later illustrates.
2 Size and scope of global music market
By 2004, global sales of physical recorded music (audio and video) were
estimated to be worth US $33.6 billion (International Federation of the
Phonographic Industry [IFPI], 2005a). This represents a fall of almost 10
per cent in constant dollar terms since 1999 (see Figure 14.1). The IFPI
suggests three reasons for the recent decline in global audio sales; ﬁrst, continued sales substitutions from downloading and CD ‘burning’ from unauthorized sources of music on the Internet; second, increased competition
from other entertainment sectors, such as DVD videos and video games;
third, the presence of adverse economic conditions.
Total global unit1 sales in 2004 were 2.8 billion, 76 per cent of which were
accounted for by sales of CDs. Figure 14.2 illustrates the spectacular
growth in the popularity of the CD format in recent years and the corresponding decline in other formats, particularly vinyl and cassette.
Table 14.1 shows that in 2004, the UK had the highest per capita sales
of recorded music in both volume and value terms, at 3.2 units and
US $58.2 respectively, followed by the USA at 2.8 units and US $41.5
respectively.
While there has been a recent decline in sales of physical musical products, legitimate online music services (where consumers pay to download
from the Internet) are experiencing signiﬁcant growth. The spur to this
was the success of Apple’s iTunes Music Store, which was launched
to Macintosh users in April 2003 and sold an average of 500 000 downloads per week over the ﬁrst six months, reaching a total of 13 million
by the mid-October when the service was expanded to include PC users
(IFPI, 2004).

392

1973

1974

IFPI.

Figure 14.1

Source:

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1992

Real value in US$ of global recorded music sales (1973–2004)

US$ price values are calculated based on 2004 average realized price by format.

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Note:

US$ (billions)

40.0

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

393

1974

1973

Figure 14.2

IFPI.

500

1000

1500

2000

2500

3000

3500

Source:

millions

4000

Single

Vinyl LP

Cassette

Music Video

1994

1993

1992

1991

1990

1989

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

Global recorded music sales by format (units, millions, 1973–2004)

1988

4500

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

394

Economics uncut

Table 14.1

USA
Japan
UK
France
Germany
Canada
Australia
Italy
Spain
Netherlands

Top ten music markets (2004)
Per capita sales
(units)

Per capita sales
(US$)

2.8
2.0
3.2
2.1
2.2
1.9
2.4
0.7
1.0
1.8

41.5
40.6
58.2
32.8
26.1
21.3
36.0
11.2
14.2
31.1

Source: IFPI.

3 Economics of record contracts
On the supply-side, the music industry has a highly concentrated, oligopolistic structure currently dominated by four Majors: Sony BMG, EMI,
Universal and Warner. In 2004, they had a combined global market share
of 76.2 per cent (IFPI, 2005a), the remainder being accounted for by
numerous Indies whose continued presence reﬂects two factors. First,
they have an ability to expand markets and specialize in market niches (for
instance, Jamaican music and West Coast rap). Second, many Indies have
relied upon the Majors to bring their music to market because of the
latter’s economies of scale in the distribution of physical music products.
Indeed, one consequence of this symbiotic relationship between the
Majors and Indies has been the reduction of potential competition (Frith,
1987). Digital technologies and the Internet are, however, reducing costs
and barriers to entry and represent an opportunity for the Indies to
loosen the Majors’ stranglehold on the distribution of music (see Fox,
2002).
Aside from their ‘market niche’ role, the Indies also act as
‘Schumpeterian innovators’ (Silva and Ramello, 2000), meaning that they
create musical innovation, discover new artists and take on the related risks
of failure, thereby beneﬁting the Majors. Manuel (1993) comments that:
The Majors can aﬀord to take the risks, but generally avoid doing so; rather they
prefer to wait until a group or artist has made a name on an Indie label, and then
they acquire that act from the Indie, thus letting the Indies bear the cost of
research and development.

Economics of rock ’n’ roll

395

The story of Sun Records, an Indie that ﬁrst recorded Elvis Presley, is a
classic example. Sam Phillips, the owner of Sun, sold the rights to Elvis to
RCA, a Major. What RCA was able to do with Elvis was utilize its ﬁnancial
resources and distribution network to make Elvis a bonaﬁde megastar. As
Phillips recognized, he was unable to provide as much for Elvis as RCA
could. Though some later questioned Phillips’s decision, in the 1950s’ environment he made the business decision that he could use the money to ﬁnd
another star of Elvis’s stature and although he was not responsible for
developing Elvis into a superstar, Phillips did play an important role in the
development of the ‘Memphis Sound’ (Gillett, 1996).
The issue of a record company taking on risk leads us to the economics
of a recording contract. What exactly is the nature of risk in signing a new
artist? How does a record company (Major or Indie) spread risk? What
does the incentive structure look like in a recording contract? How are contracts enforced? In answering these questions, we deal with the implications
of prevailing contract practices, rather than the diﬀerence between actual
and ﬁrst-best contract terms. Indeed, we show that contracts in creative
industries are structurally bounded far away from ﬁrst-best.
3.1 Division of risk
For a record company, the decision on whether to sign a musician is complicated by the ‘nobody knows’ character of new musical output (Caves,
2000). Music is an example of an experience good and an individual’s
utility from consuming a particular piece of music is highly subjective.
Record company executives and producers may have deep knowledge
about what has commercially succeeded in the past but their ability to forecast the potential sales of a new artist is severely limited. The expected
payoﬀ for all parties is highly uncertain, leading Caves (2000) to suggest
that the organizational problem encountered in the music industry relates
to ‘symmetrical ignorance’, not asymmetrical information.
There is a diﬀerent form of asymmetry that is important to note,
however. The production of an album proceeds through a number of
stages, with costs completely sunk at each stage. For example, when Artist
X delivers the unﬁnished album to Label Y, Artist X’s input to the process
is sunk, but Label Y can still adjust or even withhold its input. If both
parties receive new information about the album’s market potential, then it
is eﬃcient for Label Y to hold decision rights about whether and how to
proceed (Artist X has no more choices open). It is this asymmetry that
accounts for the common use in music of the option contract, by which
Label Y buys from Artist X the option to proceed with the album’s
production once it has digested any new information about its potential
(Caves, 2000).

396

Economics uncut

In practice, the artist will agree to deliver master tapes to the record
company for a series of albums. The record company holds an option to
distribute the albums for a speciﬁed period of time and commits to pay royalties to the artist based upon revenues generated by each recording. For
each album, the record company may pay an advance to the artist to cover
the cost of recording the album and a proportion of expected earnings
from royalties. This advance against royalties can be viewed as absorption
of risk by the record company and a guarantee of at least a minimum
royalty income for the artist (Caves, 2000). However, should an album’s royalties fail to cover the advance paid to the artist, this shortfall is charged
against any albums that are subsequently released. This is known as ‘crosscollaterization’.
Having provided an advance to the artist, the record company holds
an option not to release an album should the master tape be deemed of
unacceptable quality. If the record company does release the master tape as
a proper album, then a contract will typically stipulate an increased future
advance and royalty rate, but also a commitment for the artist to deliver
another album (Caves, 2000).
Risk-averse artists are sometimes able to secure a stronger commitment
from the record company to release albums even if new information about
market potential is unfavourable. Highly successful, established artists
often bargain for even ﬁrmer obligations for the company to release
their albums; for instance, they may ask for the master tape to be only ‘technically satisfactory’ or ‘in the artist’s previous style’ rather than ‘commercially satisfactory’ (Caves, 2000).
Overall, the options within a record contract are distinctly one-sided
because the artist is bound to the label and is not permitted to quit and
record for another label until the contract expires. This arrangement was
recently challenged by successful musicians in California who would like to
see a limit placed on the duration of record contracts (Ordonez, 2002). Due
to a 1987 amendment to California labour law, musicians in this state are
speciﬁcally exempt from a general seven-year limit on labour contracts.
The label seeks a long-term contract with the artist because of the high
‘stiﬀ ratio’, that is, the proportion of recordings that are loss-makers. Vogel
(2001) notes that only one in ten albums generate a proﬁt, which is worse
than the ﬁlm industry where on average three in ten ﬁlms ‘make money’.
Ordonez (2002) quotes SoundScan ﬁgures, which reveal that of the 6455
new albums released by major labels in the USA in 2001, only 113 were
proﬁtable.2 So, a long-term contract is preferable for the label so that it can
generate suﬃcient proﬁts from the successes to cover the stiﬀs’ losses.
In a study funded by the RIAA, Wildman (2002) examined over 6000
contracts signed over the 1994 to 2000 period between artists and the ﬁve

Economics of rock ’n’ roll

397

Majors at the time. His principal ﬁndings were that: 1) the odds of success
for a new artist were very low; 2) record labels invest considerable sums of
money to record and market albums by unproven artists and 3) artists who
have a hit album renegotiate their contracts and require signiﬁcant upfront
commitments by the record company.
In particular, the Wildman study provides information on the complex
and multifaceted aspects of recording contracts. There is no such thing as
a ‘standard’ contract and there exists substantial variation in contract
terms among newly signed artists. The number of albums potentially
required to be delivered over the contract varied between one and ten,
about three-quarters of artists signed to contracts will release albums ﬁve
to seven years later, but only about 10 per cent will still be under contract
after this time. The average upfront commitment (execution advance plus
recording funds) for a ﬁrst album in a contract in the period 1994 to 1996
was nearly $300 000 in year 2000 US dollars. For similar contracts signed
in 2000, this ﬁgure had risen to over $450 000. Royalties ranged from 4 per
cent to 20 per cent with a mean of 15.2 per cent. The study also found that
only one contract of 244 signed from 1994 to 1996 was negotiated without
the artist’s legal counsel and that virtually all contracts renegotiated after a
hit album added terms favouring the artist.
3.2 Incentives and enforcement
The options contract provides a strong incentive for the label to promote a
new artist’s ﬁrst album because it secures terms of access to the artist’s
future albums. Typically, a label can contract a successful artist for long
enough to recover any losses generated on earlier albums and promotion of
one album may have the positive spillover eﬀect of stimulating demand
for the artist’s previous works (Caves, 2000). Moreover, should the artist
prove a big success, the built-in schedule of royalties speciﬁed in the
contract may not rise suﬃciently quickly for the artist to reap the rewards
and are instead collected by the label (Caves, 2000). On the other hand, as
a contract’s expiration date approaches, the label’s incentive to promote a
successful artist fades because of the fall in expected rents picked up by the
label from future albums.
The royalty contract based on the label’s sales revenue is somewhat
diﬀerent from the terms under which authors and visual artists receive
payment. Although one of the principal roles of a label is to ﬁnance physical production and distribution of an artist’s album, substantial components of promotional costs (in addition to all recording costs) are recouped
from the artist’s royalties. If artist and label have agreed on the royalty rate
schedule but not on the label’s promotional eﬀorts, then a moral hazard
problem arises: the label can shift costs to the artist without being forced to

398

Economics uncut

renegotiate the royalty rate and this may result in the label’s over-promotion
of an album relative to the expenditures that would maximize both parties’
combined proﬁts (Caves, 2000). This incentive to over-promote will
increase in line with the label’s conﬁdence that it can recover all costs it has
advanced to the artist. It will decrease as the artist’s eﬀective royalty rate
increases because the, ‘higher the rate, the less of a promotion-induced
extra dollar of revenue lands in the company’s pocket’ (Caves, 2000).
The royalty contract would appear to provide the artist with an incentive
to under-invest in the amount of time spent in the studio recording an
album. This is a reasonable assumption given that it is the label that collects
the lion’s share of additional revenue once the advanced recording costs are
recouped. However, in practice, artists may have a tendency to over-invest
in studio time because of a taste for perfectionism (‘art for art’s sake’
[Caves, 2000]), while the album’s uncertainty of success provides an economic motivation. Should the album be a ‘miss’, and its costs cannot be
recovered, the artist does not collect any royalties but the burden of the loss
falls predominantly on the label. On the other hand, should the album be
a ‘hit’, the artist earns royalties, his or her royalty rate may well increase,
leading to a debt–equity moral hazard problem: ‘If inﬂating the studio
costs raises the chances of a big win, while the downside loss falls entirely
or mostly on the label, the artist could select a heavier investment in album
production costs than the label would prefer’ (Caves, 2000).
Another form of moral hazard is encountered in relation to contract
enforcement. Speciﬁcally, once the contract has been agreed, it is very
diﬃcult for the artist to monitor the professional conduct of the label, particularly in terms of royalty payments. The label holds the actual sales data
that determine the royalties and while most contracts allow for auditing of
the label’s accounting, it is a costly exercise in terms of both time and
money. In response to this lack of transparency, labels have recently been
reducing the scope for ‘opportunistic transactions’. For instance, in 2002,
BMG announced plans to adopt a ‘fairer, more transparent’ royalty system
that would cut the length of a standard contract from 100 pages to 12. In
the same year, Universal revealed that it would be adding provisions to contracts that permitted auditors previously denied access to manufacturers’
records. In 2003, Warner promised revisions to simplify royalty calculations, to attach interest to royalties unearthed in audits and to reimburse
audit costs when unpaid royalties exceeded 10 per cent.
Overall, Caves (2000) notes that, ‘the recording contract seems peerless
in the scope for governance disputes’, and concludes that, ‘enforcement of
contracts in the creative industries depends heavily on the power of
repeated interactions among parties who value their reputations for cooperative behaviour’.

Economics of rock ’n’ roll

399

3.3 Technological change and copyright ownership
How is the introduction of new information and communication technologies such as Napster aﬀecting the relationship between artist and record
label? Regner (2003) examines the relative marginal eﬃciency and relative
indispensability of labels and artists in a pre- and post-Napster world to
determine whether label or artist ownership of copyright is more eﬃcient.
He argues that in the pre-Napster world, the relative marginal eﬃciency
and indispensability are in favour of the labels because of their more
eﬃcient promotional skills and their indispensable command over the retail
distribution network. Whilst established artists might have suﬃcient power
in the production process to demand that copyrights revert back to them,
for aspiring musicians without any reputation, there is no option but to rely
upon the label to promote and distribute their music. Taking the market as
a whole, label ownership of copyright is therefore more eﬃcient in the
pre-Napster world.
In the post-Napster world, however, information technology advances
further and alternative ways to promote and distribute music emerge.
Artists can promote themselves via the Internet (for example, with digital
updates of their existing fan base or through the information externalities
of ﬁle-sharing networks) and cheaper electronic distribution becomes
possible (direct selling of music by artists in online shops or the direct
distribution of MP3 ﬁles). So, marginal eﬃciency turns in favour of the
artists in the post-Napster scenario.
Figure 14.3 depicts this technology-induced relationship change between
artist (A) and record label (L). In the pre-Napster scenario, artists would
generally be positioned in the lower right section of the diagram because of
the high indispensability and marginal eﬃciency of the labels in terms of
promotion and distribution. However, technological innovation and
advance mean that in the post-Napster scenario, the position of artists
changes and there is a movement across the ownership threshold into
eﬃcient artist ownership.
The extent of this movement depends upon the exact impact of technology and on artist type. The ﬁgure shows that established artists will ﬁnd it
easier to cross the threshold than new artists, typically because they are
cash-endowed. Indeed, recently there have been a number of established
artists releasing their music on self-owned labels, including Prince and
The Eagles.
Building on Regner’s model, Halonen-Akatwijuka and Regner (2004)
introduce the concept of a mentor, which is an established musician who
can provide new artists with an alternative to label promotion in the postNapster scenario. A mentor ‘adopts’ a new artist in his or her own ﬁeld who
can be credibly recommended and promoted. For example, the mentor

400

Economics uncut

Relative
indispensability
A/L

Ownership
threshold

A
L

Established
artists
New
artists

Relative marginal efficiency
L/A
Source: Regner (2003).

Figure 14.3

Technology’s impact on the record label–artist relationship

would support the aspiring artist by linking to the new artist website from
his or her own frequently visited websites and endorsing him or her there.
In addition, the new artist could be invited to tour with the established artist
as a support act. Essentially, the mentor is acting as a venture capitalist by
having faith in, promoting and ﬁnancing the project of the new artist.
Again, in this scenario label ownership of copyright appears less likely.
Halonen-Akatwijuka and Regner suggest that this arrangement has
implications for the quality of music in the future. With a movement toward
artist ownership, the relative contribution of the artist in the total value of
the music increases. Therefore, the authors suggest that this will lead to an
improvement in the quality of music compared with the era of label ownership, which is characterized by relatively low quality music that is well
packaged and promoted.
However, they emphasize that label ownership remains feasible provided
that the label retains an element of indispensability. Their model only considers the ‘singer-songwriter’ type who provides all the artistic input, but if
this assumption is relaxed, then label ownership is reasonable. This is particularly relevant in the case of ‘boygroups’ where the artists only sing, with
the remainder of the artistic inputs (such as songwriting and choreography)
and promotion being provided by the label. Therefore, when examining the

Economics of rock ’n’ roll

401

impact of technological change on copyright ownership, it is important to
distinguish between music production under label ownership where label
inputs are crucial and music production under artist ownership where artistic inputs are vital.
4

The winner takes it all: economics of rock superstars

4.1 Ability and record sales
In Chapter 12, with reference to Rosen’s 1981 seminal contribution, we
examined the ‘superstar phenomenon’ in sport, where small diﬀerences in
ability translate into large diﬀerences in earnings.3 Rosen concluded his
paper by remarking on Alfred Marshall’s reason for why a talented early
nineteenth century opera singer named Elizabeth Billington earned less
than a ‘superstar’ salary. Marshall had suggested that:
. . . as long as the number of persons who can be reached by a human voice is
strictly limited, it is not very likely that any singer will make an advance on the
£10 000 said to have been earned in a season by Mrs Billington at the beginning
of the last century.

In response, Rosen commented that: ‘Even adjusted for 1981 prices,
Mrs Billington must be a pale shadow beside Pavarotti. Imagine her income
had radio and phonograph records existed in 1801! What changes in the
future will be wrought by cable, video cassettes, and home computers?’
Ten years after Rosen’s contribution, Hamlen (1991) tested for the presence of superstar eﬀects in popular music. He collected data from Billboard
ratings on 107 singers over the 1955–87 period and estimated the following
log-linear equation:
lnRSB1 lnHARB2 lnDURB3 lnF

(14.1)

where ln denotes the natural log; RS is the value of total record sales;
ln HAR is a quantitative measure of a singer’s voice quality, represented by
the high frequency harmonic content that vocalists use when they sing the
word ‘love’ in one of their songs; ln DUR is the number of years taken for
the singer to build up record sales and is a composite of the eﬀects of rising
prices and incomes during the period; ln F is a vector of dummy variables
that represent the singer’s qualitative attributes: whether the singer was
female (SX), black (RC), mostly wrote his/her own songs (SG), was a crossover country singer rather than mainstream (COUNTRY), had a wide voice
range (RG), starred in at least one movie (MV), was recognized for instrumental backing or a band (BD), and whether the singer or group had its

402

Economics uncut

career curtailed (CS) due to premature death or a group break-up at the
height of its fame. The estimated demand function and associated t-ratios
in parentheses are reported in Table 14.2.
As the table shows, Hamlen found that career longevity (ln DUR) is the
strongest predictor of total record sales, followed by being female (SX).
Being black (RC) was found to have a negative eﬀect upon record sales,
while the main measure of a singer’s ability in this study, ln HAR, was the
only other statistically signiﬁcant explanatory variable. Hamlen notes that
this does indicate that consumers do recognize quality in their pop singers,
but because the coeﬃcient is signiﬁcantly less than one (0.14) he rejects the
argument that this is the Marshall–Rosen ‘superstar phenomenon’ at work.
Of course, it is debatable whether the harmonic content of a singer’s voice
is the appropriate measure of artistic quality for singers of non-classical
music (Schulze, 2003).
Hamlen (1994) extended his earlier approach by incorporating the work
of MacDonald (1988) and Adler (1985). MacDonald proposed a multimarket superstar model based on a dynamic ‘information accumulation
process’, which he believed was appropriate for singers and musicians.
Singers start their careers by competing in an entry-level market, where
they share the revenues generated by lower-income consumers who have
lower opportunity costs of time and can aﬀord to listen to relatively poor
quality singers. However, there are a small number of relatively high quality
Table 14.2 OLS regression to identify superstar eﬀects in popular music
Dependent variable ln RS (record sales) (t-statistics in parentheses)
Constant
ln HAR (voice quality)
ln DUR (experience)
SG (songwriter)
BD (band)
RG (voice range)
RC (black)
SX (female)
MV (movie)
CS (career curtailed)
COUNTRY
Notes:
n107; F 35.1; R2 0.79.
*Signiﬁcant at 1% level.
Source: Hamlen (1991).

0.63871 (2.137)
0.14246 (2.826)*
1.22282 (13.546)*
0.03183 (0.126)
0.33482 (1.205)
0.14114 (0.525)
0.74115 (2.295)*
1.00579 (3.797)*
0.20500 (0.794)
0.04762 (0.203)
0.49654 (1.622)

Economics of rock ’n’ roll

403

singers in the entry-level market who achieve above-average success and
have the chance to compete successfully in more select, higher-level
markets where consumers have higher incomes and higher opportunity
costs of time. Each higher market level is characterized by consumers being
willing to pay a premium to listen to only the higher quality singers who
have been ‘ﬁltered out’ from the lower-market levels. This multi-market
process proposed by MacDonald can therefore produce the superstar phenomenon.
Hamlen notes that the work of Adler (1985) can undermine MacDonald’s
model. First, one assumption behind the model is that the quality ﬁlter
works perfectly. Yet, in practice, we may observe some high quality singers
doing poorly and some poor quality singers doing well. Second, Adler suggests that the success and failure of singers depends upon ‘factors other than
talent’. These other factors fall into measurable and non-measurable categories: for instance, sex or race in the former case, charisma, good looks, or
luck in the latter case. It is the existence and value of these factors, according to Adler, that determine the consumer’s demand for ‘variety’ and complicate the relationship between quality (ability) and success.
Hamlen examined the 1955–87 recorded music market within
MacDonald’s multi-level market superstar process and allowed for Adler’s
‘other factors’, which lead to an imperfect ﬁltering mechanism for quality.
He divided the market into the lower-end market for singles and the higherend market for albums, with the singles market being viewed as an entrylevel quality ﬁlter for the album market. The same explanatory variables
were used as in his 1991 study but with the following additions: eight
diﬀerent musical styles were selected to replace the COUNTRY variable,
and variables were also included to take account of the price of singles and
albums, the consumer budget on all albums, and the year in which the
singer(s) released their ﬁrst single or album (to allow for the fact that in the
1955–87 period some singers had more years in which to secure hit records).
Hamlen employed Tobit regressions since the number of hit singles or
albums is non-negative and because it allows estimation of the interaction
between singles and album markets. His single-stage reduced regression
produced coeﬃcients for quality of voice in singles and albums of 0.123
and 0.105 respectively. He suggests that because these coeﬃcients are less
than one, ‘there is no evidence of the superstar phenomenon, at least as
deﬁned with respect to proportionalities’.
His two-stage and multiple-stage simultaneous Tobit regressions helped
to separate the interactive inﬂuence of the singles and album markets. In the
albums market, he found that voice quality is not signiﬁcant but the level of
endogenous hit singles is signiﬁcant, providing evidence that the singles
market acted as a ‘quality’ ﬁlter for the albums market. However, he also

404

Economics uncut

found support for Adler’s view that this ﬁlter will be an imperfect one. For
instance, some black singers who were successful in the singles market
achieved, on average, one less hit record (0.857) in the albums market. While
this could be oﬀset by an additional hit single (0.758), Hamlen notes that,
‘it still represents an obstacle to success and a distortion in the relationship
between success and ability’. (See Chung and Cox, 1994 for another test of
the theories of Rosen and Adler in relation to recorded music sales.)
4.2 Superstar concerts
Krueger (2004) examines the relevance of the superstar phenomenon to
explain why concert ticket prices in the USA escalated between 1997 and
2003, and why ticket sales and the number of shows performed by top
artists fell over the same period.
Figure 14.4 illustrates the average price of a concert ticket (total revenue
divided by total tickets sold each year in the USA) for all concerts performed between 1981 and 2003, as well as the (ticket-weighted) average
high and low price of a ticket. The ﬁgure also reveals, according to
Krueger’s calculations, what the average price would have been had it
increased in line with the US CPI-U.
Concert ticket prices in the USA rose slightly faster than inﬂation from
1981 to 1986 (a compound 4.6 per cent a year compared with 3.7 per cent
a year), but grew much faster than inﬂation from 1996 to 2003 (8.9 per cent
a year compared with 2.3 per cent a year). Further, Figure 14.4 also shows
that the cost of the highest priced ticket at a concert has risen even faster
than the average priced ticket. Krueger estimates that weighted by total
ticket sales, the average high price ticket increased by 10.7 per cent a year
from 1996 to 2003, while the average price of the lowest price ticket rose by
6.7 per cent a year.
The number of concerts performed in the USA by artists featured in the
Rolling Stone Encyclopedia increased in the 1980s, reached a peak in the
ﬁrst half of the 1990s, and then fell by 16 per cent from 1996 to 2003.
Krueger also notes that from the late 1980s until 2000, the number of
concert tickets sold by these artists ﬂuctuated around 30 million a year, but
declined to 22 million tickets by 2003. Total ticket revenue to these artists
(in 2003 US dollars) increased until 2000, since which there has been a
10 per cent fall in revenue.
Krueger computes the total dollar value of ticket sales each year relative
to the USA total to obtain the share of revenue captured by the top 1 per
cent and top 5 per cent of all performers (ranked by their total concert
revenue). Figure 14.5 reveals the increasing skewness of concert revenues
in the 1980s and 1990s. In 1982, the top 1 per cent of artists received
26 per cent of concert revenue but by 2003, this ﬁgure had more than

405

Average

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

Average price per concert ticket in USA and inﬂation rate (1981–2003)

Krueger (2004).

Figure 14.4

Source:

Price (US$)

CPI

High

Low

2002

2000

1999

2003

2001

406

1984

1983

Top 2–5%

Top 1%

1995

1994

1993

1992

1991

1988

1987

1986

1985

Distribution of concert ticket revenue to artists in USA (1982–2003)

Krueger (2004).

1982

Figure 14.5

Source:

Everyone else

1989

1990

100

2003

2002

2001

2000

1999

1998

1997

1996

Economics of rock ’n’ roll

407

doubled to 56 per cent. Similarly, the top 5 per cent generated 62 per cent
of concert revenue in 1982 and 84 per cent in 2003.
Krueger also selects artists who were the highest revenue generators in
the 1996 to 1999 period (and also revenues in surrounding periods) to
report the revenue, number of shows performed, and average price for these
top artists just prior to the price take-oﬀ (1994–95) and just after (2000–01).
These ﬁgures are shown in Table 14.3. Krueger notes that the number of
concerts performed by these stars dropped by 18 per cent while revenue per
show rocketed by 60 per cent due to an increase in price and increase in
tickets sold per show.
He suggests that for the superstar phenomenon to account for these
quantity and price trends, two observations must hold: ﬁrst, superstar
artists have a backward-bending labour supply curve that led to a decrease
in concerts despite the increase in revenues per show; second, superstar
eﬀects gained strong momentum in the post-1996 period. Krueger tests for
the presence of superstar eﬀects using two samples of data: all artists featured in a Pollstar database (containing 232 911 reports on concerts) and a
subset of artists listed in the Rolling Stone Encyclopedia of Rock & Roll
(containing information on 1786 artists, 1275 of whom having performed
at least one concert represented in the Pollstar database).
For each artist (i) in each year (t), Krueger computes the following
dependent variables in logarithms: the (ticket-weighted) average price, total
revenue and revenue per show. The full regression model is as follows:
lnYit pSi Xit t it

(14.2)

where ln Yit is the log average price (or log total revenue or log revenue per
show), Si is the measure of star quality, which Krueger captures by the
number of millimetres of print columns (including photographs) devoted to
each artist in the Encyclopedia, Xit is a vector of covariates (number of supporting acts, years of experience of the band, and dummies for genre and
foreign status), t is a set of 22 unrestricted year ﬁxed eﬀects, and it is an
error term.
The coeﬃcient on star quality, p, has a p subscript, which speciﬁes the
time period (1981–96, 1987–91, 1992–96 or 1997–2003). This allows the eﬀect
of star quality to vary across time periods, which is captured in Krueger’s
regressions by interacting the amount of print with dummies indicating the
four periods. Therefore, his ‘rising-return-to-superstardom’ hypothesis is
essentially one of whether we observe a discrete jump in p after 1996.
The results for the full sample on the three models (average price, total
revenue and revenue per show) reveal large and increasing superstar eﬀects.
He reports that a 200 millimetre increase in print in the Encyclopedia is

408

102
20

147
64
76
54
37
141
40
40
35
33
131

33
28
4
16
12
1
1
1
48.4

$151 000
54 200

50 200
35 700
28 600
24 200
24 200
23 100
21 500
20 800
18 500
14 200
10 700

9 264
6 196
5 410
1 652
1 516
368
325
141
23 894

281
221
1 352
103
126
368
325
141
494

341
558
376
448
654
164
538
520
529
430
82

$1 480
2 710

23.32
22.99
88.19
35.45
28.86
82.14
27.51
11.94
40.41

29.49
31.39
23.73
29.64
40.66
23.27
42.65
27.74
39.46
36.54
21.55

$67.50
201.65

8 454
338
10 300
47 000
49 100
5 989
6 687
60 100
31 196

11 800
49 600
22 500
45 900
21 800
21 300
32 900
37 500
23 900
38 400
129 000

$4 837
27 700

18
2
9
48
67
25
8
118
39.5

43
62
11
59
38
40
40
18
58
51
110

1
4

470
169
1 144
979
733
240
836
509
789

274
800
2 045
778
574
533
823
2 083
412
753
1 173

$4 837
6 925
70.7
3.1
85.5
9.3
2.7
71.6
0.0
55.0
65.7
54.5
16.0
45.5
92.9
125.0
200.0
458.3
2 400.0
700.0
11 700.0
18.3

76.5
39.9
21.3
89.7
9.9
7.8
53.0
80.3
29.2
170.4
1 105.6
28.91
8.7
50.00
94.5
105.78
90.4
65.20
2 745.3
43.37
3 139.3
34.35
1 529.4
59.70
1 960.5
50.070 42 436.8
50.020
30.6

42.76
39.84
48.60
47.34
56.70
30.50
62.46
60.45
46.12
64.37
43.72

$89.22
483.61

99.0
80.0

67.3
23.6
15.4
848.4
480.2
34.8
157.6
260.5
59.8

19.6
43.4
443.5
73.6
12.3
225.0
53.0
300.6
22.0
75.0
1 335.8

226.8
155.5

24.0
117.5
19.9
83.9
50.3
58.2
117.0
319.3
23.8

45.0
26.9
104.8
59.7
39.4
31.0
46.4
117.9
16.9
76.1
102.8

32.2
139.8

Total
Number Rev. per Average
revenue of shows show
price

Percentage change

96.8
48.9

Total revenue Number Rev. per show Average
(in thousands) of shows (in thousands) price

2000–01

Note: Computations based on Pollstar database. All dollar ﬁgures converted to 2001 US dollars based on CPI-U.
Source: Krueger (2004).

The Eagles
Barbara
Streisand
Reba McEntire
Jimmy Buﬀett
George Strait
Aerosmith
Elton John
Phish
Eric Clapton
Metallica
Rod Stewart
Janet Jackson
Dave Matthews
Band
Pearl Jam
Beastie Boys
Luciano Pavarotti
Bruce Springsteen
Ozzy Osbourne
Paul Simon
Mariah Carey
Kiss
Average

Total revenue Number Rev. per show Average
(in thousands) of shows (in thousands) price

1994–95

Table 14.3 Concert revenue and prices in USA in 1994–95 and 2000–01 for artists with highest revenue per show in 1996–99

Economics of rock ’n’ roll

409

associated with 5 per cent higher prices in the 1980s, 11 per cent higher
prices in the early 1990s and 15 per cent higher prices over the 1997 to 2003
period. Krueger notes that while the impact of star quality upon annual
revenue and revenue performance is much greater, the increase in return to
star quality was much slower after 1996 than before. For instance, 200 millimetre increases in print in the 1981–86, 1987–91, 1992–96 and 1997–2003
periods were associated with increases in total revenue of 70 per cent, 90 per
cent, 104 per cent and 110 per cent respectively. For the superstar phenomenon to account for the escalation in concert ticket prices after 1996, the
eﬀects need to be growing at an increasing rate.
Similar ﬁndings from Krueger’s analysis of the sub-sample are reported.
Star quality was associated with a rise in prices throughout the 1981 to 2003
periods, but the greatest increase took place in 1992–96, a period before
prices sky-rocketed. Moreover, his results for total revenue and revenue per
show lend even less support for the superstar explanation: the impact of
print in the Encyclopedia on these outcomes fell in the 1997–2003 period
compared with the 1992–96 period. Krueger concludes that, ‘because
revenue per performance should be the driving force in the superstar model,
these results strongly suggest that accelerating returns to superstardom are
not the explanation for the rapid price growth after 1996’.
4.3 David Bowie economics
One potential reason for the acceleration in the price of concert tickets in
the USA after 1996 is Baumol and Bowen’s Cost Disease (1966). This
theory proposes that prices ought to rise faster than overall inflation in lowproductivity growth sectors due to cost increases. Concerts may be viewed
as a slow-productivity growth sector in the sense that it takes an artist of
today as least as much time, eﬀort and labour input to perform a particular song as it took another artist to perform it 30 years ago. However,
Krueger argues that it is improbable that the concert industry underwent a
‘discrete jump’ in costs compared to other industries after 1996. His
favoured explanation for the concert ticket price phenomenon is that musicians have experienced a large fall in their income from record sales, which
are a complementary good to concerts. He terms his hypothesis Bowie
Theory, after rock singer David Bowie made the following prediction in the
9 June 2002 edition of The New York Times:
Music itself is going to become like running water or electricity. So it’s like, just
take advantage of these last few years because none of this is ever going to
happen again. You’d better be prepared for doing a lot of touring because that’s
really the only unique situation that’s going to be left. It’s terribly exciting. But
on the other hand it doesn’t matter if you think it’s exciting or not; it’s what’s
going to happen.

410

Economics uncut

While this does not necessarily provide the answer for every creative
musician, it does indicate that an appropriate and innovative model is
necessary to survive in the age of the Internet.
Krueger argues that a famous band has some monopoly power because
of its inimitable sound and style. Therefore, he suggests that in the past when
greater concert attendance converted into higher record sales, the artists
had an incentive to ensure their tickets were priced below the proﬁtmaximizing price for concerts alone. However, as already noted, new technology has emerged in recent years that allows consumers to download
music without having to buy an album. This has weakened the
concert/album relationship, with the result, according to Krueger, that concerts are now priced like single-market monopoly products.4
His hypothesis is formally explained by assuming that an artist is a ﬁrm
with two complementary outputs, concert seats and albums, denoted good
1 and good 2 respectively. The artist also exerts monopoly power in both
markets. The complementarity characteristic means that the artist’s
demand curve, D1 (p1, p2) and D2 (p1, p2), depends on both prices, while
costs are assumed to be independent of each other and are a function only
of each speciﬁc good produced, C1(D1) and C2(D2). Proﬁt maximization is
achieved by the artist choosing both prices, p1 and p2, as follows:
Maxp1D1(p1, p2) p2D2(p1, p2) C1(D1) C2(D2)

(14.3)

(p1, p2)

Meanwhile, equation (14.4) below denotes the proportionate mark-up of
concert tickets over marginal cost, where the ij’s represent the value of the
own- or cross-price elasticities of demand:
p1 C1
1 (p C2)D212
2
p1
11
p1D111

(14.4)

Krueger argues that the second term in equation (14.4) has fallen in recent
years because, beginning in the late 1980s, an increase in concert attendance
has had a much weaker eﬀect on record sales. As a result, ‘artists and their
managers do not need to feel as constrained when they set concert prices’.
Indeed, he ﬁnds some tentative empirical support for Bowie Theory, pointing to the observation of Oberholzer and Strumpf (2004) that jazz fans are
much less likely to download music from the Internet than are pop and rock
fans and from 1996 to 2003, jazz concert prices in the USA rose by only 20
per cent, while those of rock and pop concerts jumped by 99 per cent.
Krueger also suggests that Bowie Theory can help explain the concert ticket
price growth recently experienced in Canada and Europe.

Economics of rock ’n’ roll

411

5 Copying music: more economics of Napster and all that
We now turn to the economics of copying music and investigate how this
often illegal activity, spurred by technological advances, is aﬀecting music
sales. We distinguish between three types of copying:5
1.

Private copying – where an individual makes copies on a unit scale of
a CD, for instance, using home equipment such as a CD-R. In many
countries private copying is not illegal provided that the copy is made
for personal use and not distributed to third parties.
Commercial piracy – involves issuing onto the market physical unauthorized copies of existing oﬃcially released music, typically at a
lower price. To the best of our knowledge, this is illegal in almost every
country.
Online piracy – refers to unauthorized downloading of music from the
Internet. This takes two main forms. First, the distribution of ﬁles from
unlicensed Internet sites known as web and FTP sites. Second, peerto-peer (P2P) activities on ﬁle-swapping services such as Napster,
which involve direct transfers of music ﬁles between users (‘peers’).

5.1 Size of copying market
By deﬁnition, private copying of music is intended for private or personal
use and so it seems unreasonable to speak of a ‘private copying market’.
Obtaining data on the extent of private copying is almost an impossible
task, but the IFPI does provide ﬁgures relating to commercial piracy and
online piracy.
It estimates that in 2004 the value of the global commercial piracy
market was US $4.6 billion.6 One and a half billion units of pirated music
were sold, which represents an increase of over 190 per cent compared with
1991, but a fall of 35 per cent from its 1998 peak (IFPI, 2005b). CDs comprised 74 per cent of global pirate sales in 2004 with a roughly equal split
between pressed discs and CD-R (38 per cent and 36 per cent respectively).
Pressed discs are made on factory production lines and dominate the pirate
music market in Asia and Russia, while CD-R piracy refers to the reproduction of albums using CD burning software and is most common in
Latin America and Southern Europe with rapid recent growth in Eastern
Europe (IFPI, 2003). The remaining 26 per cent of global pirate market
unit sales in 2004 was accounted for by cassette sales (25 per cent) and DVD
music video sales (1 per cent). Cassettes are the principal pirate format in
the Middle East, and they are still popular in Asia, Africa and Eastern
Europe (IFPI, 2002).
The IFPI estimates that in 2004 the value of the global pirate market is
equal to the entire legitimate music markets of the UK, Netherlands and

412

Economics uncut

Table 14.4 Domestic music piracy levels in 2004
Over 50%
North America
Europe
Bulgaria
Czech Republic
Estonia
Greece
Latvia
Lithuania
Romania
Russia
Serbia/Montenegro
Turkey
Ukraine
Asia
China
India
Indonesia
Malaysia
Pakistan
Latin America Argentina
Brazil
Central America
Chile
Colombia
Ecuador
Mexico
Paraguay
Peru
Uruguay
Venezuela
Middle East
Egypt
Kuwait
Lebanon
Australasia
Africa

Morocco

25–50%

10–24%

Less than 10%

Croatia
Cyprus
Hungary
Italy
Poland
Portugal
Slovakia

Belgium
Finland
Netherlands
Slovenia
Spain

Philippines
Taiwan

Hong Kong Japan
South Korea Singapore
Thailand

Canada, USA
Austria
Denmark
Germany
Iceland
Ireland
Norway
Sweden
Switzerland
UK

Israel
Bahrain
Oman
Qatar
Saudi Arabia UAE
Australia
New Zealand
Nigeria
South Africa
Zimbabwe

Note: Domestic music piracy levels are calculated as pirate units divided by legal units
plus pirate units.
Source: IFPI.

Economics of rock ’n’ roll

413

Spain combined. It notes that there has been a marked recent increase in
the number of countries whose illegal music sales exceed legal sales (see
Table 14.4). A total of 31 countries had piracy levels equivalent to or
greater than their legitimate music markets in 2004 compared to 23 in 2003
(IFPI, 2005b).
Silva and Ramello (2000) note that, ‘unauthorized sound reproduction,
in manufacturing and commercial terms, ﬁnds fertile soil in countries where
per capita income is low, the market is more backward and, as a result, the
legislative measures to control illegal activities are imperfect or altogether
non-existent’. This is generally true for the countries featured in Table 14.4
that have domestic piracy levels in excess of 50 per cent and supports
Burke’s (1996) observation that, ‘economic development rather than copyright legislation seems to be the most important feature distinguishing low
from high piracy nations’.
In terms of online music piracy, the IFPI estimates that in January 2005,
there were 870 million infringing music ﬁles on the Internet at any one time.
Around 90 per cent of these were found on P2P networks, with the remainder on FTP and websites. The IFPI notes that this represents a sharp
fall from the 1.1 billion unauthorized ﬁles on the Internet in April 2003
(IFPI, 2005c).
5.2 Economic theory of music copying
Unauthorized copying of music may be expected to hurt the copyright
owner such as a record company because it is unable to appropriate any of
the value created by its work. Prospective consumers have a reduced incentive to purchase the product (an album, for instance) when they have the
option of using unauthorized copies. Economic theory does, however,
provide a number of reasons why copying does not necessarily have to
harm the copyright owner.
5.2.1 Indirect appropriability The term ‘indirect appropriability’ was
ﬁrst coined by economist Stan Liebowitz in 1985 and refers to how copyright owners can collect revenue from unauthorized copiers by charging
higher prices for the originals from which the copies were made. Liebowitz
(2002) provides a simple example. Assume that every CD buyer makes a
single audiocassette copy for use in the car. Nobody makes copies from borrowed CDs and it is impossible to prevent individuals from engaging in
unauthorized copying. How would this aﬀect the CD’s copyright holder if
it intends to sell prerecorded cassettes as well as CDs? Liebowitz argues that
each consumer’s willingness to pay for the original CD will be greater than
it would otherwise be; ﬁrst, each original CD will have a copy made from
it, and second, each consumer will place some value on the ability to copy.

414

Economics uncut

Therefore, the copyright owner is able to charge a higher price for the CD
to capture some of this additional value.
Boldrin and Levine (2002) show that if the copyright holder faces an
elastic demand curve, as the price of CDs falls over time due to copying
technologies, output will increase more than proportionately and proﬁts
that can be indirectly appropriated in the original price will also increase.
However, Klein et al. (2002) point out that this result, ‘clearly conﬂicts with
actual record-company pricing . . . if Boldrin and Levine were correct, why
are record companies not pricing CDs as low as possible?’ Liebowitz (2002)
acknowledges that, ‘just because indirect appropriability might help create
proﬁts does not mean that it will succeed in any particular case’.
Indeed, the potential for indirect appropriability to work will depend
upon the variability in the number of copies made of each original. In an
environment characterized by widespread copying and variability in the
number of copies being made, it will be virtually impossible for the seller to
identify precisely which originals should have the higher price. This is why
commercial piracy and online piracy (where thousands of copies are made
from an original) are viewed as more harmful to copyright holders than
private copying (where only one or two copies are made from an original).
The above refers to the concept of implicit indirect appropriability.
However, a number of countries impose a levy on copying equipment such
as blank audio tapes, which is a form of explicit indirect appropriability.
Further, in Europe, a technology named digital rights management (DRM)
(see Liebowitz, 2002) is being used, which allows devices such as iPods
(Digital Audio Recorders [MP3 players] with non-removable memory) to
either limit copying or charge consumers a fee for each recorded work
downloaded from the Internet.
5.2.2 Exposure eﬀects, space-shifting, network externalities and price
commitments The ‘exposure eﬀect’ is another means by which the copyright owner can potentially beneﬁt from copying (or at least not be harmed
by it). It refers to the process by which a consumer’s online ‘sampling’ of
individual songs will increase demand for the full album recording. In other
words, this argument views online piracy as a complement to a CD purchase, rather than a substitute for it. Indeed, Robert Hall, an economist
from Stanford University and one of the experts employed by Napster in its
defence in court, argued: ‘The exchanges of music facilitated by Napster
stimulate the demand for the plaintiﬀ’s CDs by allowing consumers to
sample CDs and develop an interest in CDs that they subsequently purchase.’
The exact impact of sampling may depend upon whether the album is
released by a Major or Indie label. Major releases tend to receive consider-

Economics of rock ’n’ roll

415

able levels of promotion via radio and television, and so the sampling
eﬀect may be relatively small. By contrast, Indie releases are less likely to
receive such promotion and so it is possible that they can beneﬁt from the
‘exposure eﬀect’. It may even stimulate demand for live performances from
these artists.
The ‘exposure eﬀect’ argument does have certain ﬂaws. As noted by Klein
et al. (2002), a record company that sought to use Napster, for example, as
an advertising platform could simply release promotional samples of its
products to the Internet to distribute free. Moreover, Liebowitz (2002)
points out that because better sampling reduces ‘mistakes’ in CD purchasing, ‘consumers may discover that they do not need to purchase as many
CDs since their thirst for music can be quenched with fewer of them’.
Another argument used by Napster in its defence was the ‘space-shifting
eﬀect’, which refers to the scenario where an individual at an oﬃce computer is merely using Napster to play a CD he or she has already purchased.
This is a form of indirect appropriability and Napster drew comparisons
with two similar court cases: ﬁrst, the Diamond case of the late 1990s,
where the court held that copying a ﬁle already stored on a computer user’s
hard drive to a portable listening device was non-infringing personal use,
and second, the Betamax case of the early 1980s, where the activity of
‘time-shifting’ (taping a television programme for later viewing) was judged
by the US Supreme Court to be ‘fair use’. In the Napster case, however, the
court rejected the analogous ‘space-shifting argument’.
The concept of a network externality in relation to music piracy is
relevant if the number of albums, both legitimate and illegitimate, in circulation determines an artist’s fame and popularity. As with the exposure
eﬀect, information costs are reduced and new consumers are attracted to
buy legitimate albums. The underlying mechanism at work here is Adler’s
1985 ‘snowball eﬀect’, which proposes that the more people already know
a particular artist, the easier it is for other people to acquire artist-speciﬁc
knowledge.
Meanwhile, Takeyama (1997) argues that copying may stimulate demand
for legitimate albums by allowing record companies to credibly commit to
not reduce their prices in the future. Unauthorized copying at time t takes
away a small proportion of the future demand for an album, enabling the
record company to keep its price stable until t1, at which point the album
will change in terms of its perceived quality. Then, the record company can
charge a monopolistic price for its high quality albums without fostering
expectations of lower prices in the short run. Takeyama shows that this
strategy can increase a record company’s proﬁts and social welfare because
both high and low quality demand is satisﬁed at time t.
In summary, the economic theory of music copying suggests that two

416

Economics uncut

opposing eﬀects are in operation. On the one hand, the activity can reduce
legitimate demand for albums as potential consumers substitute into
pirated versions. On the other hand, copying may raise a consumer’s willingness to pay and stimulate legitimate demand via various positive
inﬂuences. It would appear, therefore, that the balance of these eﬀects is an
empirical issue.
5.3 Music copying and empirical evidence
Perhaps the ﬁrst empirical work relating to music piracy was the evidence
produced for the hearings on the preliminary injunction against Napster
in the autumn of 2000. Two reports that focused on whether or not
Napster was increasing or decreasing sales of CDs received particular
attention: the ‘Fine Report’, prepared by Michael Fine, Chief Executive of
Soundscan, and the ‘Jay Report’, prepared by Deborah Jay, a marketing
consultant.
The ‘Fine Report’ examined CD sales in the USA from the ﬁrst quarter
of 1997 to the ﬁrst quarter of 2000 at: 1) all ‘brick-and-mortar’ retailers;
2) those within one mile of any US college or university; 3) those within the
top 40 most wired US universities and 4) those near universities that prohibited use of Napster after the ﬁrst quarter of 2000. The underlying
assumption was that students at college and university campuses were
heavier users of Napster (and the Internet generally) than the average
consumer, and so any signiﬁcant diﬀerence between CD sales to the two
groups would be due to Napster’s impact.
It was found that from the ﬁrst quarter of 1999 to the ﬁrst quarter of
2000 (Napster actually emerged around halfway through this 12-month
period), national sales of CDs grew by 6.6 per cent, sales near colleges and
universities fell by 2.6 per cent, sales near the most wired colleges and universities dropped by 6.2 per cent and sales near colleges and universities
where Napster was prohibited after the ﬁrst quarter of 2000 fell by
8.1 per cent. From these ﬁgures, the ‘Fine Report’ concluded that Napster
was harming sales of legitimate CDs.
However, earlier data from the report cast doubt on the validity of this
conclusion. From 1998 to 1999 (a year prior to the birth of Napster), CD
sales near colleges and universities dropped by around 5 per cent, while
increasing elsewhere by approximately 5 per cent. These ﬁgures appeared
to lend support to Napster’s defence that online sales of CDs near college
and university campuses were replacing ‘brick-and-mortar’ sales rather
than illegal downloading of songs via Napster.
The ‘Jay Report’ presented survey evidence from college and university
students to examine their reasons for using Napster and its impact on their
music purchases. Aside from the usual caveats attached to survey evidence,

Economics of rock ’n’ roll

417

Liebowitz (2002) notes speciﬁc ﬂaws in the survey design of the ‘Jay
Report’. For instance, two of the categories of answers, ‘buy fewer CDs’
and ‘make my own CDs’, to the question of why a student uses Napster
were classiﬁed by Jay as indicating that Napster downloads are displacing
CD sales. While the ﬁrst answer obviously matches this description, the
second is ambiguous. A student may have made a CD to sample music for
later purchase, but Jay would classify that answer as an example of Napster
decreasing CD sales. Unfortunately, the ‘Jay Report’ does not provide separate ﬁgures for each of these two answers to determine the severity of the
problem, only reporting that 22 per cent of survey respondents either purchased fewer CDs or made their own CDs, while 8.4 per cent purchased
more CDs. Overall, the report concluded that 41 per cent of Napster users
utilized it in ways that displaced CD sales and, as noted earlier, the court
took the side of the recording industry.
More recent empirical attempts by economists to measure the impact of
illegal downloading on legitimate music sales include Liebowitz (2003),
Oberholzer and Strumpf (2004) and Zentner (2004). Liebowitz investigates
the eﬀect of a wide variety of factors that could explain the fall in music
sales in the USA between 1999 and 2002. These factors include changes in
demographics, in income, in prices of albums and other entertainment substitutes, in recording format and listening equipment, in the ‘quality’ of
music or in musical ‘taste’ and in music distribution. Using basic regression
analysis, he found that these factors cannot fully explain the observed
decline in sales and suggested that, ‘downloads are causing signiﬁcant harm
to the record industry’.
Zentner (2004) estimates the eﬀect of downloads on music purchases
using October 2001 individual-level cross-section data of 15 000 people
from seven European countries: France, Germany, Italy, Netherlands,
Spain, Sweden and the UK. From a simple comparison of means, he
reported a positive correlation between downloading and music purchasing,
even after controlling for many individual-level characteristics. However,
Zentner found diﬃculty in isolating the causal eﬀect of downloads on music
purchases due to the simultaneity between tastes for music and peer-to-peer
usage. To overcome this, he used measures of Internet sophistication and
connection speed as instruments and concluded from his results that downloading may explain a reduction in the probability of buying music of 30
per cent and a fall in music sales of 7.8 per cent in the countries under study.
However, Zentner’s sample omits those under 16-years-old, who are typically among the most active downloaders and heaviest buyers of music. In
addition, his sample does not contain information relating to the intensity
of downloading or music purchases, thus making it diﬃcult to determine
the precise impact of downloading on music sales.

418

Economics uncut

Instead of using survey data, Oberholzer and Strumpf (2004) use ﬁgures
taken directly from ﬁle sharing networks to estimate the impact of downloads on music sales. They analyse a dataset from the USA, which includes
0.01 per cent of the world’s music downloads from the ﬁnal 17 weeks of
2002. To estimate a relationship between downloading and sales, they
matched downloads to the album they were released on, for which they had
concurrent weekly sales ﬁgures. To establish causality, Oberholzer and
Strumpf instrumented for downloads using technical features related to ﬁle
sharing (such as network congestion or song length) and international
school holidays, both of which are assumed to be exogenous to sales. They
found that ﬁle sharing has only a limited eﬀect on music sales. Even in their
most pessimistic speciﬁcation, 5000 downloads are needed to displace a
single album sale.
To estimate the eﬀect of commercial piracy on CD sales, Hui and Png
(2003) examine panel data for CDs from 28 countries over the 1994 to 1998
period. Using a ﬁxed eﬀect speciﬁcation, per capita legitimate music
demand was regressed against average national CD price, personal disposable income, CD player ownership levels, worldwide MTV subscriptions
and piracy levels. Due to the possible endogeneity of CD prices, Hui and
Png searched for an instrument that would capture variations in CD prices
while not being correlated with quantity demanded. As data on supply-side
shifters were not available, they used the price of non-tradeable goods as
their price instrument, arguing that the impact of music demand on the
price of non-tradeable goods would be very small and their price would
reﬂect the cost of land and labour. The latter determine the cost of retailing, which in turn inﬂuences the price of CDs.
Hui and Png argued that the quantity of pirated CDs might be endogenous as it depends on both the price of and demand for the legitimate CD.
Two groups of instruments were chosen that would be correlated with the
pirated quantity but not legitimate sales. The ﬁrst group (piracy rates of
music cassettes and business computer software) were used because they
would be inﬂuenced by the same set of national characteristics such as
pirated products usage and culture. The second group of instruments
(unemployment rate and total consumer expenditure) were used because
they are exogenous variables that shift the pirate quantity, but are not
aﬀected by the legitimate demand for CDs.
Hui and Png’s two-stage least squares (2SLS) estimates indicated that, on
a per capita basis, a one-unit increase in CD piracy was associated with a
decline in legitimate CD sales of 0.42 units. However, they suggest that the
eﬀect of piracy is considerably less than estimated by the music industry.
Assuming that the industry did not adjust CD prices, they estimate that in
1998, actual losses amounted to 6.6 per cent of sales, or 42 per cent of

Economics of rock ’n’ roll

419

industry estimates. On the other hand, Hui and Png acknowledge that
in reality music publishers would have reduced prices to mitigate the
eﬀect of piracy and so the true revenue loss would have been greater than
6.6 per cent.
In summary, it would appear from the empirical evidence that the negative eﬀects of both commercial and online piracy outweigh their positive
eﬀects. In other words, piracy does lead to a fall in legitimate sales of music.
However, as we have shown, the extent of the negative impact may be
smaller than industry estimates. This seems a reasonable assumption given
that trade groups such as the IFPI have an incentive to inﬂate piracy rates
in order to support their position.7
6 Conclusion
This chapter applied economic analysis to the music industry. We showed
how the industry is undergoing great change in both how music is distributed and in the tastes of the consumer, largely due to technological advances.
From our investigation into the impact of music piracy, it is easy to understand why artists and record companies are at present concerned about
maintaining their livelihood. However, historically, technology has been
responsible for the growth of the mass consumption of music. Further, it has
allowed today’s superstars to earn huge sums of money compared with
musicians of 200 years ago who could not even make a living through creating music unless they were lucky enough to have a patron who would commission compositions or live performances. Record companies ﬁrst feared
radio until they realized it could become a spur to sales. Like the invention
of the recording tape and the compact disc, the Internet is more likely to ultimately result in an increase in demand for music once new business models
are in place. Indeed, the success of Apple’s iTunes Music Store has demonstrated that consumers are willing to pay for music downloads from the
Internet. Those industries that do not innovate in a market economy, or
ineﬀectively compete with other similar products, will decline. The same is
true of music.
Notes
1.
2.
3.
4.

The IFPI calculates units as the total album equivalent. Three singles are counted as one
album and sales of all formats are included. Multiple disc packaged products are
counted as one unit, except in some countries.
Major album releases only become profitable after selling 500 000 copies.
We should note that diﬀerences in talent are perhaps easier to measure in sport than in
music.
The same result is obtained if we assume a competitive market where records and concerts are joint products. If income is lowered from one, then the other has to pay a larger
share of the common cost.

420
5.
6.

Economics uncut
Space does not permit us to discuss bootlegging, the unauthorized recording and distribution of previously unreleased music. An excellent discussion and analysis of this activity can be found in Naghavi and Schulze (2001).
The IFPI’s value estimate is for commercial physical pirate product only. It does not
include illegal downloading via the Internet and is calculated at pirate selling prices. The
IFPI argues that commercial losses to the music industry from piracy are substantially
greater.
It should also be noted that, to the best of our knowledge, no empirical evidence exists
that suggests that piracy is having an adverse eﬀect upon artists’ incentives to create
music.

References
Adler, M. (1985), ‘Stardom and Talent’, American Economic Review, 75 (1), 208–12.
Adler, M. (2004), ‘The Economics of Superstars: A Review with Extensions’, unpublished
manuscript.
Baumol, William J. and William G. Bowen (1966), Performing Arts: The Economic Dilemma,
New York: The Twentieth Century Fund.
Boldrin, M. and D. Levine (2002), ‘The Case Against Intellectual Property’, American
Economic Review: Papers and Proceedings, 92 (2), 209–12.
Burke, A.E. (1996), ‘How Eﬀective Are International Copyright Conventions in the Music
Industry?’, Journal of Cultural Economics, 20, 51–66.
Caves, Richard E. (2000), Creative Industries: Contracts between Art and Commerce,
Cambridge, MA: Harvard University Press.
Chung, K. and R. Cox (1994), ‘A Stochastic Model of Superstardom: An Application of the
Yule Distribution’, Review of Economics and Statistics, 76, 771–5.
Fox, M. (2002), ‘Technological and Social Drivers of Change in the Online Music Industry’,
First Monday (http://www.ﬁrstmonday.dk/issues/issue7_2/fox/).
Fine, M. (2000), ‘SoundScan Study on Napster Use and Loss of Sales’ (http://www.riaa.com/
news/ﬁlings/pdf/napster/ﬁne.pdf).
Frith, S. (1987), ‘The Industrialization of Music’, in James Lull (ed.), Popular Music and
Communication, Newbury Park: Sage.
Gillett, Charlie (1996), The Sound of the City: The Rise of Rock and Roll, New York: Da Capo
Press.
Hall, R.E. (2000), Expert Report, A&M Records, Inc. v. Napster, Inc. 114F. Supp. 2d896
(http://newscorporate.ﬁndlaw.com/hdocs/docs/napster/napster/hall_report_ﬁnal.pdf).
Halonen-Akatwijuka, M. and T. Regner (2004), ‘Digital Technology and the Allocation
of Ownership in the Music Industry’, CMPO, University of Bristol, Working Paper,
04/096.
Hamlen, W.A. (1991), ‘Superstardom in Popular Music: Empirical Evidence’, Review of
Economics and Statistics, 73 (4), 729–33.
Hamlen, W.A. (1994), ‘Variety and Superstardom in Popular Music’, Economic Inquiry, 32,
395–406.
Hui, K. and I. Png (2003), ‘Piracy and the Legitimate Demand for Recorded Music’,
Contributions to Economic Analysis and Policy, 2 (1), 1–22.
International Federation of Phonographic Industry (IFPI) (2002), ‘Commercial Piracy Report
2002’, IFPI.
International Federation of Phonographic Industry (IFPI) (2003), ‘Commercial Piracy Report
2003’, IFPI.
International Federation of Phonographic Industry (IFPI) (2004), ‘Online Music Report’,
IFPI.
International Federation of Phonographic Industry (IFPI) (2005a), ‘The Recording Industry
in Numbers 2005’, IFPI.
International Federation of Phonographic Industry (IFPI) (2005b), ‘Commercial Piracy
Report 2005’, IFPI.

Economics of rock ’n’ roll

421

International Federation of Phonographic Industry (IFPI) (2005c), ‘Digital Music Report’,
IFPI.
Jay, D.M. (2000), ‘Expert Report of Deborah M. Jay’, Napster Litigation.
Klein, B., A.V. Lerner and K.M. Murphy (2002), ‘The Economics of Copyright “Fair Use” in
a Networked World’, American Economic Review: Papers and Proceedings, 92 (2), 205–8.
Krueger, A.B. (2004), ‘The Economics of Real Superstars: The Market for Rock Concerts in
the Material World’, Princeton University, Working Paper.
Liebowitz, S. (2002), ‘Policing Pirates in the Networked Age’, Policy Analysis, No. 438, 15
May, Washington, DC: The Cato Institute.
Liebowitz, S. (2003), ‘Will MP3 Downloads Annihilate the Record Industry?: The Evidence
So Far’, University of Texas, Dallas, Working Paper.
MacDonald, G. (1988), ‘The Economics of Rising Stars’, American Economic Review, 78,
155–66.
Manuel, Peter (1993), Cassette Culture, Chicago, IL: University of Chicago Press.
Marshall, Alfred (1947), Principles of Economics, 8th edition, New York: Macmillan.
Naghavi, A.J. and G.G. Schulze (2001), ‘Bootlegging in the Music Industry: A Note’,
European Journal of Law and Economics, 12 (1), 57–72.
Oberholzer, F. and K. Strumpf (2004), ‘The Eﬀect of File Sharing on Record Sales: An
Empirical Analysis’, Mimeo, University of North Carolina, Chapel Hill.
Ordonez, J. (2002), ‘Pop Singer Fails to Strike a Chord Despite the Millions Spent by MCA’,
The Wall Street Journal, 26 February.
Pareles, J. (2002), ‘David Bowie, 21st Century Entrepreneur’, The New York Times, 31 October
Issue 908, 54–6.
Regner, T. (2003), ‘Innovation in Music’, in Wendy J. Gordon and Richard Watt (eds), The
Economics of Copyright Developments and Research, Cheltenham, UK and Northampton,
MA, US: Edward Elgar, pp. 104–17.
Rosen, S. (1981), ‘The Economics of Superstars’, The American Economic Review, 71 (5),
845–58.
Schulze, G.G. (2003), ‘Superstars’, in Ruth Towse (ed.), The Handbook of Cultural Economics,
Cheltenham, UK and Northampton, MA, US: Edward Elgar, pp. 431–6.
Silva, F. and G.B. Ramello (2000), ‘Sound Recording Market: The Ambiguous Case of
Copyright and Piracy’, Industrial and Corporate Change, 9, 415–42.
Strobl, E.A. and C. Tucker (2000), ‘The Dynamics of Chart Success in the U.K. Pre-recorded
Popular Music Industry’, Journal of Cultural Economics, 24, 113–34.
Takeyama, L. (1997), ‘The Intertemporal Consequences of Unauthorized Reproduction of
Intellectual Property in the Presence of Demand Network Externalities’, Journal of Law
and Economics, 40 (2), 511–22.
Vogel, Harold L. (2001), Entertainment Industry Economics, 5th edition, Cambridge
University Press.
Wildman, S.S. (2002), ‘An Economic Analysis of Recording Contracts’, a study funded by the
Recording Industry Association of America (RIAA).
Zentner, A. (2004), ‘Measuring the Eﬀect of Online Music Piracy on Music Sales’, University
of Chicago, Working Paper.

Index
ability
record sales and, music industry
celebrity 401–4
rent of, free agency and salary caps,
sport policy issues 353
abortion
background 315
cause, as 327–8
access, fertility rate and 328–9
crime and 333–7
endogenous pregnancy 329–31
shotgun weddings and 331–3
demand
anti-abortion activities effects 326
price and income elasticities 323
joint products and 2
out-of-wedlock births increase and
7–8
outcome, as 321
access laws 325
anti-abortion activities 325–7
consent laws 325
cultural factors 327
demand and supply, basic model
321–4
public subsidies 324–5
sequential decision-making 334
services
availability 317–20
usage 316–17
terminology 315–16
US ban on 337–8
abortion rate
UK and USA 318, 319
worldwide 316
access
abortion
fertility rate and 328–9
laws, abortion as outcome and 325
see also availability
accidents
drug use and 61–2
addiction
drug use and 58–61

‘good’ versus ‘bad’ 5
marriage and 5
pornography demand analysis and
182–3
theories 16–24
see also drug addiction; rational
addiction theory
Adeokun, L. 202
Adler, F. 55
Adler, M. 402, 403, 415
adoption
market possibility for infertile couple
291
prices and 7
supply and demand and 306
Adult Video News 174
advertising
drug prohibition and 46–7
drug prohibition and crime 55
prostitution and 210
prostitution, decriminalization and
legalization 198
Africa
abortion, service use 317
see also South Africa
age
prostitution and 6, 216
suicide rate by, international 238–9
agency
free, sport policy issues 353–8
AGI, see Alan Guttmacher Institute
Ainslie, G.W. 24
Akerlof, G.A. 210, 331–3, 336
Alan Guttmacher Institute (AGI)
316–17
Albert, A. 220
alcohol, see drugs
Ali, M. 372, 375
Alm, J. 378
altruistic suicide
Durkheim approach 231–2
see also suicide
American Society for Reproductive
Medicine (ASRM) 304–5

423

424

Index

Anders, G. 372, 375, 379
Anderson, E. 310
Anderson, T.L. 82, 86
anomic suicide
Durkheim approach 230
see also suicide
antitrust, see competition
Apple iTunes Music Store 391, 419
appropriability
indirect, music industry copying and
413–14
Arizona Indian Gaming Association
379
Arlidge, John 189
arrests
drug offences, USA 49
prostitution and related offences
and, USA 200
see also police
Arrow, Kenneth 180, 191
artificial insemination, see donor
insemination
Asia
abortion service use 317
ASRM, see American Society for
Reproductive Medicine
Assane, D. 375
assisted reproduction
contract, market for
donation only, with 299
legal positive price, with 300
demand for and supply of children,
microeconomics of before
reproductive markets 293–4
economic issues 298
demand 301–3
price 305–10, 311
risk in demand and supply 303–5
supply 298–301
ethics 310
market evolution 295–7
market possibilities 291–3
market size 297–8, 311
services in Europe and USA 293
see also birth; marriage; sex
asymmetric information
effects, prostitution 210
Atkinson, S. 361
Australia
drug addiction statistics 12, 16

drug prohibition 52
gambling, gross annual wagering
368
prostitution 197, 204
suicide, celebrity 258
Austria
drug liberalization 72–3
prostitution, decriminalization and
legalization 197
see also Europe
availability
abortion services 317–20
guns, suicide and 251–3
see also access
Azzi, C. 268
Bailey, S. 386
Bainbridge, William 265
Bakalar, James B. 58
balance
competitive, sport
product nature 348
revenue-sharing and collective sale
of broadcasting rights 360–62
bankruptcy
social costs of gambling and 380
Bardsley, P. 31
bargaining
family members, among, decisionmaking in marriage and 152–3
Barnett, A. 381
Barros, C.P. 345
Barros, Pedro Pita 282, 285–6
baseball
free agency and salary caps and 355
see also sport
basketball
free agency and salary caps and 353,
355–6
see also sport
Baumol and Bowen’s Cost Disease 409
Beccaria, Cesare 101
Becker, Gary S.
abortion, basic model of demand
and supply and 321
behavioural economics xxix–xxxi
crime and
capital punishment xxiii–xxiv
economic theory of 102, 107, 110
fear economy xxii–xxiii

Index
terrorism xxiv–xxv
private protection against 130
criminology and xxv
discrimination and xvii–xviii
drugs and
addiction xxvi–xxvii, 17, 18, 19, 37
liberalization, sin taxes 83, 86
prohibition xxvii–xxviii
family and xviii–xix
divorce xix
economics of xx–xxi
inter-racial marriage xxi–xxii
marriage contract xix–xx
same-sex marriage xx
household production model,
demand for religion and 268
relationship formation model 156
suicide and xxviii–xxix
terrorism xxix
see also rational addiction theory
behaviour
addictive, characteristics of 11
cartel, sport 351–2
economic principles applied to 1–3
behavioural economics
Becker on xxix–xxxi
Belarus
suicide, statistics 237
see also Europe
Belgium
assisted reproduction, risk in supply
and demand 304
drug liberalization 69
see also Europe
benefits
consumption, gambling 371–2
marriage 145
children 146
love and sex 145–6
public goods 147–8
time 148–50
religion
demand and 268
spillover 273
social, pornography as economic
good and 188–9
Benjamin, D.K. 82, 86
Bennet, T. 56
Bennett, T.H. 59
Benson, Bruce L.

425

crime and
criminal justice system resource
allocation 124
police expenditure and rate of 119
private protection 130, 132
drug liberalization and
enforcement and asset seizures 78
free market 91
government regulation 81
sin taxes 84, 85, 87
Benson, C. 212
Bentham, Jeremy 101
Bergstrom, T. 155, 157
Berman, E. 277
betting
described 369
illegal
gambling policy and 382–3
Internet 384–5
USA 382–3
Internet
UK 383, 384–5
USA 383–4
see also gambling
The Bible 279
Billboard 255
Billington, Elizabeth 401
Birdzell, L.E. Jr. 86
birth
live, probability in any cycle 303
out-of-wedlock increase, abortion
and 7–8
sequential decision-making 334
unborn fetus, drug use and
62–3
see also assisted reproduction
Bitler, M. 335
Blair, R.D. 28
Blinder, Alan xxxi
Blume, L. 155
Bold, F. 281
Boldrin, M. 414
Boonchalaksi, W. 198
Borg, M. 385
Bowie, David 409
Bowie Theory (Krueger) 409–10
Brainerd, E. 249
Bretteville-Jensen, A. 32
Brien, M. 160
Britain, see United Kingdom

426

Index

broadcasting rights
collective sale of, sport policy 362–4
Brown, G.F. Jr. 55
Brown, M. 152
Browning, M. 152
Bruce, Steve 281
Brumm, H.J. 53
BSkyB 363–4
budget
constraints, indifference curves and
in marriage 147
Burke, A.E. 413
Bush, George W. 337
business cycle
suicide and 249–51
Byrne, J.M. 124
Cameron, Samuel
drug addiction and, price elasticity
of demand 31
prostitution and
demand for 202
economic welfare considerations
210
empirical studies 212
legalization 197
market entry 203
primary and secondary markets
distinguished 194
Canada
assisted reproduction, risk in supply
and demand 304
divorce trends 141, 143
drug liberalization, taxation and 85–6
marriage trends 139
religion, market product and
competition 283
Cann, D.E. 283
capital
drug addiction and 19, 22–3
human, stock loss, suicide and 253–5
religious, demand and 270–71
capital punishment
Becker on xxiii–xxiv
demand for crime prevention and
113–16
number of executions, USA 115
public opinion, USA 117
supporters’ reasons for favouring,
USA 118

see also punishment
Caputo, M.R. 83
Carael, M. 202
Carmelli, D. 17
cartel
behaviour, sport 351–2
see also competition
casinos
economic regeneration and
UK 380
USA 379, 380
reasons for gambling and, USA 372
taxation and, UK 385
toleration of gambling and, reasons
for 378–9
see also gambling
The Catcher in the Rye (Salinger) 211
Caudill, S. 378
Caulkins, J.P. 34, 48
cause
abortion as 327–8
access, fertility rate and 328–9
crime and 333–7
endogenous pregnancy 329–31
shotgun weddings and 331–3
Cave, M. 363
Caves, Richard E. 395, 396, 397–8
CD, see compact disc
CDC, see Centers for Disease Control
celebrity
music industry
ability and record sales 401–4
Bowie Theory (Krueger) 409–10
effects in popular music, OLS
regression identifying 402
superstar concerts 404–9
sport
football, effects on salaries in Italy
353–5
relative demand and supply,
teachers compared 355
suicide 255–8
Centers for Disease Control (CDC)
297–8
Chadwick, Edwin 102
Chaloupka, F.J. 31, 33, 34, 35, 95
Chapman, M. 196
characteristics
addictive behaviour 11
individual

Index
egg donor compensation and
selection effect of 309
sperm donor compensation and
selection effect of 308
pornography industry
production and consumption
173–5
profits 176–7
size distribution of firms 175–6
prostitution industry 193–4
drugs and crime link 195
marriage link 194
migration link 195
pornography link 194–5
provision 195–7
religious denominational, USA 276
charitable giving
gambling and 374–5
Chaves, M. 283
cheating
sport 258–9
Cheng, L.S. 17
Cherlin, A. 143
Cheung, H. 250
childbirth, see birth
children
adoption, prices and 7
birth, out-of-wedlock increase,
abortion and 7–8
demand for, microeconomics before
reproductive markets 293–4
divorce and 164
drug liberalization and free market
and 93
marriage gain, as 146
pornography and 175, 178, 184
suicide and 247–8
see also family
China
abortion, cultural factors and
demand for 327
prostitution, decriminalization and
legalization 197
suicide 2
Chiswick, C.U. 269
Chriqui, J.F. 74
Christiansen Capital Advisors 367
cigarettes, see drugs
Clark, A. 207
Cleland, J. 202

427

cloning
market possibility for infertile couple
292
Cloninger, D.O. 53
Coase, R. 158
Cobain, Kurt 255–8
Coelen, S.P. 322
cohabiting
marriage compared 144–5
Cohen, Steve 376
collective sale
broadcasting rights, sport policy and
362–4
see also sale
Collins, A. 202, 212
Colombia
drug prohibition and crime 55
Columbia Records 390
commercial piracy
defined 411
see also music industry
commitment
transitions into and out of marriage
and 161
commodity
euphoria as, relative theory of
addiction and 18–20
religion as 264, 286–8
compact disc (CD)
copying market, size of 411
copyright ownership, indirect
appropriability and 413
global music market, size of 391, 393
music industry history and 390
sales
commercial piracy effects 418–19
Fine report 416
Jay report 416–17
see also music industry
competition
drug liberalization and free market
and 91
drug prohibition and 46–7
religious market and
doctrine 285–6
product 280–83
sport, cartel behaviour 351–2
see also monopoly
competitive balance
sport

428

Index

product nature 348
revenue-sharing and collective sale
of broadcasting rights 360–62
Connolly, S. 386
consent
laws, abortion as outcome and 325
consumption
benefits, gambling 371–2
drug addiction and 20–24, 35–8
drug liberalization and free market
and 87–90
illicit drugs, statistics 12–16
pornography and 173–5
economic good, as 189–90
see also demand
contract
assisted reproduction
demand and 301–3
donation only, with 299
legal positive price, with 300
supply and 298–301
dissolution of, divorce as 144
marriage, Becker on xix–xx
marriage and 144–5
options, music industry, incentives
397
record, music industry 394–5
copyright ownership,
technological change and
399–401
enforcement 398
incentives 397–8
risk, division of 395–7
royalty, music industry, incentives
397–9
control, see regulation
conventional demand studies
price elasticity of demand for illicit
drugs and 32–4
see also demand
convictions
drug, USA 77
sentences to prison, jail and
probation 78
time served 79
Cook, P.J. 251
Cooper, A. 182
copying market
music industry, size 411–13
see also market

copyright
ownership, music industry
copying and, indirect
appropriability 413
record contract, technological
change and 399–401
Core Institute 248
costs
economic, suicide and 253–8
social
gambling policy and 380–82
illicit drug use 38
suicide, of 4
sunk, prostitution and 2
see also prices
Cragg, J.G. 203
Crancer, A. Jr. 61
cricket
cheating 359
see also sport
crime
abortion and 333–7
Becker on xxii–xxv
criminal justice resources allocation
122
probability and severity of
punishment and 126–9
rationing 122–6
drug prohibition and 52–5
drug use effect 55–7
drug-related, European approach to
reducing harm and 71
economists’ interest in 102–3, 132
gambling and, social costs 380
music piracy
CD sales, effects on 418–19
commercial piracy defined 411
economic theory 413
empirical evidence 415–18
exposure effect 414
indirect appropriability 413–14
levels of, IFPI ‘priority territories’
412
market size 411–13
network externality 415
online piracy, defined 411
price commitments 415
space-shifting effect 415
types of, defined 411
penal reform, history 101–2

Index
pornography and 173
prostitution and 195
suicide and, USA 251
supply of offences 110–13
time allocation between legal and
illegal activities 112
victimless, prostitution as 7
violence, of, drugs and 5
crime prevention
demand 113
capital punishment 113–16
standard empirical model of
crime, extensions 116–22
crime rate
murder
international 106, 108
USA 54
UK 104
USA 105
falling 129–32
world, compared 103–9
criminal
prison time for, USA 128
criminal justice system
incentives for police and prosecutors
6
resources allocation 122
probability and severity of
punishment and 126–9
rationing 122–6
criminology
Becker on xxv
Crockett, A. 283
Cronje, Hansie 359
culture
abortion and 324, 327
marriage and 144
reproductive market and 308–10
see also ethnicity; religion
Culyer, A. 39
Cunningham, C.L. 17
Cutler, D. 237, 240, 247–8
cycle
business, suicide and 249–51
reproductive, probability of live
birth in 303
Datamonitor 383
Dave, D. 34
Davidson, M. 298

429

Davis, D.H. 59
Daw Namoro, S. 27
DEA, see Drug Enforcement Agency
death penalty, see capital punishment
decision-making
marriage, in 150–51
bargaining among family
members 152–3
family decision-making 151–2
spending decisions on private
goods 153–4
spending decisions on public
goods 154–5
spending versus saving 155
relationship-making, male and
female 206
sequential
pregnancy and pregnancy
resolution 331
premarital sex, abortion, marriage
and childbirth 334
Della Giusta, M. 202
Deloitte and Touche 357
demand
abortion
anti-abortion activities effects 326
basic model 321–4
price and income elasticities 323
adoption prices and 306
assisted reproduction 301–3
risk in 303–5
children, for, microeconomics before
reproductive markets 293–4
cigarettes, price elasticity of 37
crime prevention 113
capital punishment 113–16
standard empirical model of
crime, extensions 116–22
drug liberalization policy and
Europe 69–72
USA 73–4
drug prohibition and 45
gambling 374
illicit drugs, price elasticity of
36–7
evidence 30–38
theory 27–9
pornography 181
addiction 182–3
Internet effect analysis 185–8

430

Index

price 181
variety-seeking 183
relative, teachers and sports stars 355
religion 264–5
benefits 268
determinants of 268–70
religious capital 270–71
scarcity 265
suicide and 241
see also consumption
Denmark
assisted reproduction 292, 304
drug addiction statistics 16
pornography 184, 189
see also Europe
DeSimone, J. 33
detection
pornography, Internet effect and 186
see also police
di Tomasso, M.L. 202
DiClemente, C.C. 17
DiIulio, J.J. 133
Dills, A. 51
DiNardo, J. 33, 48, 95
discrimination
Becker on xvii–xviii
Ditton, J. 16
divorce
Becker on xix
contract dissolution, as 144
factors precipitating 158
relationship changes 158–9
laws, effect on 137
policy interventions and 163–5
reasons for 137
trends in 137–8, 141–4, 165–6
see also family; marriage
divorce rate
France and Canada 141
Spain, Sweden and UK 142
USA 138
Dixit, Avinash K. 243, 244
doctrine
religious, competition and 285–6
Doeringer, P. 194
Dominican Republic
prostitution, decriminalization and
legalization 197
Donohue, J.J. III 129, 333–7
donor insemination

evolution of reproductive markets
and 295–6
market possibility for infertile couple
291
supply and 299–300
double-kinked demand curve
price elasticity of demand for illicit
drugs and 28–9
see also demand
drug addiction
addiction theories 16–18
melioration 20–21
rational 21–4
relative 18–20
addictive behaviour, characteristics
11
Becker on xxvi–xxvii
demand for illicit drugs, price
elasticity of 36–7
evidence 30–38
theory 27–9
drug prohibition, welfare economics
approach 38–40
hyperbolic discounting and other
extensions to rational addiction
24–7
illicit drug consumption statistics
12–16
‘primrose path’ to 20
see also addiction
Drug Enforcement Agency (DEA) 48,
49, 70, 73
drug liberalization
economists and 68–9
European policy 69
demand-side 69–72
supply-side 72–3
free market 87, 95
children and 93
competition 91
constraints 91–2
consumption in 87–90
problems with 93–5
risk 91
sanctions 92–3
supply 90–91
government monopoly 79–81, 95
government regulation 81–2, 95
taxes 82–7, 95
US policy 69

Index
demand-side 73–4
supply-side 74–9
drug prohibition
Becker on xxvii–xxviii
crime and 52–5
drug use effect 55–7
economic incentives of, problems of
illicit drugs and 44, 63
price and quantity of drugs under
46–53
demand 45
supply 45–7
user welfare and
drug quality 56–8
use and accidents 61–2
use and addiction 58–61
use and labour market outcomes
60–61
use and unborn fetus 62–3
welfare economics approach,
addiction and 38–40
Drug Use Forecasting (DUF) 34
drugs
abortion 315
alcohol
prohibition, USA 68
suicide and 248–9
cigarettes, price elasticity of demand
for 37
control, spending, USA 76
convictions for, USA 77
sentences to prison, jail and
probation 78
time served 79
illicit
consumption statistics 12–16
demand for, price elasticity of
27–38
global use 13
prevalence of use in USA,
Australia and Europe 14, 15
problems and economic
incentives, drug prohibition 44,
63
quantity seized 72, 75
social costs of use 38
prostitution and 195
quality, user welfare and prohibition
56–8
quantity and prices, under

431

prohibition 45–53
suicide and 248–9
value of 5
violence and 5
DUF, see Drug Use Forecasting
Duke, S.B. 55, 57
Durkheim, Emile 230–31, 232
Dynarski, S. 386
Eadington, W. 372, 380
economic costs
suicide and 253–8
economic good
pornography as
consumption, substitute or
complementary goods 189–90
social benefits 188–9
economic welfare
prostitution and 210–11
see also welfare
The Economist 193, 194–5
economists
crime and, interest in 102–3, 132
drug liberalization and 68–9
pornography and, interest in 171–2,
190–91
sport and, interest in 345–6, 364
suicide and, interest in 229, 232–3,
261
economy
fear, crime and, Becker on xxii–xxv
growing, falling US crime rate and
129–30
regeneration of, gambling and
379–80
Edlund, L. 193–4, 203, 205, 209, 218
education
abortion demand and, basic demand
and supply model 323–4
prostitution and 196
reproductive market and 308–9
eﬃciency
religion and 273
Egan, Timothy 174
egoistic suicide
Durkheim approach 231
see also suicide
Ehrenberg, R.G. 268
Ehrlich, Isaac xxiv, 110–12, 113–14,
116

432

Index

Ekelund, Robert B. 272
Elster, Jon 246
EMCDDA, see European Monitoring
Centre for Drugs and Drug
Addiction
employment
drug use and 60–61
pornography and 174
endogenous pregnancy
abortion and 329–31
see also pregnancy
enforcement
music industry record contract 398
England, see United Kingdom
Erekson, O. 377
Estonia
crime rate 107
suicide statistics 237
see also Europe
ethics
assisted reproduction 310
ethnicity
inter-racial marriage, Becker on
xxi–xxii
out-of-wedlock births and shotgun
weddings 332
prostitution and, satisfaction
determinant 220
reproductive market and 307–8
see also culture
euphoria
commodity, as, relative theory of
addiction and 18–20
Europe
abortion, service use 317
assisted reproduction services 293
drug addiction statistics 12, 16
drug liberalization policy 69
demand-side 69–72
supply-side 72–3
drug prohibition 51–2
drug-related crime, approach to
reducing harm and 71
drugs, quantity seized 75
Internet betting and taxation 385
music copying 414, 417
prostitution 195, 198, 199, 201
sport 345–6, 347, 362–5
suicide 233–8, 249–50
see also individually named

countries
European Monitoring Centre for
Drugs and Drug Addiction
(EMCDDA) 12, 14, 15, 16, 70–72
Evans, J.T. 172
expected utility theory
gambling and 369–71
see also utility
exposure eﬀect
music industry copying and 414
externalities
negative
Kurt Cobain’s death 256–8
social outcome and, pornography
172–3
network, music industry copying
and 415
positive, Kurt Cobain’s death 255–6
Fajnzylber, P. 53
family
Becker on xviii–xx, xxi–xxii
religion and 269–70
see also children; divorce; marriage
family economics
Becker on xx–xxi
Fanning, Shawn 391
Farrell, L. 372, 386
Farrelly, M.C. 33
fatalistic suicide
Durkheim approach 230–31
see also suicide
Faugier, J. 212
FBI, see Federal Bureau of
Investigation
FCC, see Federal Communication
Commission
FDLE, see Florida Department of
Law Enforcement
FDOC, see Florida Department of
Corrections
fear economy
crime and, Becker on xxii–xxv
Federal Bureau of Investigation (FBI)
73
Federal Communication Commission
(FCC) 284–5
Fertility Alternatives 303–4
fertility rate
abortion access and 328–9

Index
Fine, Michael 416
Finer, L.B. 319, 320
Finke, R. 264, 265, 275, 284, 285
Finland
assisted reproduction, risk in supply
and demand 304
drug addiction statistics 16
drug liberalization 69
suicide statistics 233
see also Europe
firearms
availability of, suicide and 251–3
firm
size
distribution of, pornography and
175–6
optimal, supply of religion and
277–8
Fisher, Irving 68
Fisher, T. 203
Florida Department of Corrections
(FDOC) 128
Florida Department of Law
Enforcement (FDLE) 128
football
dominance in sport, Europe and
USA compared 346
free agency and salary caps and
353–6, 357–8
Internet betting and, UK 384
revenue-sharing and collective sale
of broadcasting rights 359,
362–4
see also sport
Forrest, D.
gambling and
consumer behaviour 371, 372
demand 373–4
economic regenerator, as 379, 381
Internet threat 384–5
social engineering, as 386
substitution effects 375
Forrestal, James 257
Fort, Rodney 355, 356, 361
fostering, see adoption
Fountain, D. 61
France
abortion, drugs for 315
assisted reproduction, supply and
299

433

divorce trends 141, 143
drug addiction statistics 16
drug liberalization 70–72, 72–3
marriage trends 139
prostitution 202, 209
sport, revenue-sharing and collective
sale of broadcasting rights 362
see also Europe
free agency
sport policy issues 353–8
free market
drug liberalization and 87, 95
children 93
competition 91
constraints 91–2
consumption 87–90
problems 93–5
risk 91
sanctions 92–3
supply 90–91
Frejka, T. 328, 329
Freud, Sigmund 264
Friedberg, L. 164
Friedman, David 274
Friedman, Milton xvii, 53, 371
Frith, S. 394
Froese, P. 285
Fukuyama, Francis 311
Gabriel, Peter 247
gains, see benefits
Gallup International Millennium
Survey 265, 266–7
gambling
betting described 369
consumption benefits 371–2
expected utility theory and 369–71
income and demand 374
market size 375–6
policy issues 386–7
economic regenerator, as 379–80
illegal betting 382–3
Internet threat 383–5
social costs 380–82
social engineering, as 385–6
price of a gamble 372–4
scope worldwide 367–9
substitute goods price 374–5
technological change and 367, 386
toleration of, reasons for 376–9

434

Index

see also casinos; lottery
Garceau, L. 304
Garoupa, Nuno 282, 285–6
gas
suicide and 242
gender, see sex
George, David 171
Germany
drug addiction statistics 16
prostitution 194–5, 197
religion, regulated 284
sport, revenue-sharing and collective
sale of broadcasting rights 362
see also Europe
Gibb, John 196
Gibbons, J.C. 113
Gillett, Charlie 395
Ginsburg, Douglas xxviii
Gittings, K.J. 114
Glaeser, E.L. 237, 240, 247–8
Gohmann, S.F. 325
Goldstein, P.J. 53
good
economic, pornography as
consumption, substitute or
complementary goods 189–90
social benefits 188–9
inferior, prostitution as 7
goods
private, spending on, decisionmaking in marriage and 153–4
public
marriage gain, as 147–8
spending on, decision-making in
marriage and 154–5
substitute
complementary or, pornography
as economic good and 189–90
euphoria commodity, relative
theory of addiction and 19–20
gambling and, price 374–5
prostitution and 204–5
religion and 269–70
government
monopoly, drug liberalization and
79–81, 95
regulation, drug liberalization and
81–2, 95
see also policy
Greece

drug liberalization 69, 72–3
suicide statistics 233
see also Europe
Greenberg, S.W. 55
Grinols, E. 380
Grinspoon, Lester 58
Gross, A.C. 55, 57
Grossman, M.
drug addiction and
conventional demand studies 33
demand, effective prices 28
rational demand studies 31, 35, 37
theories 17, 22
drug liberalization and
sin taxes 83, 86
drug prohibition and, use and
accidents 61–2
Gruber, J. xxvi, 26, 335
Grundberg, K. 221
Gtech Corporation 368
Guerin, J.G. 299
Guest, P. 198
Gulley, O.D.
gambling and
consumer behaviour 371, 372
demand 373–4
economic regenerator, as 379
substitution effects 375
guns, see firearms
Haas-Wilson, D. 325
Haddad, L. 152
Hafner, H. 258
Halonen-Akatwijuka, M. 399–401
Hamermesh, D. 232, 243, 244, 248
Hamermesh–Soss model
microeconomics of suicide and 243–5
Hamlen, W.A. 401–4
Hammersley, R. 16
Hardaway, C.K. 283
Harrison, P.M. 107
Harwood, H. 61
Hausman, J. 353
Hebert, Robert F. 272
Heineck, G. 268
Hemingway, Ernest 255
Henshaw, S.K. 317, 319, 320
Herrnstein, R.J. 20–21
Hewitson, G. 300, 301
HFEA, see Human Fertilization and

Index
Embryology Authority
Hirschman, E.C. 310
Hoddinott, J. 152
Home Oﬃce 195, 204, 221
Honduras
prostitution, decriminalization and
legalization 197
HOPE (Helping Outstanding Pupils
Educationally) 376, 385–6
Horney, M. 152
Horton, K. 251
household production
marriage gains and 148–50
Huang, W. 250
Hughes, D.M. 221
Hui, K. 418–19
Hull, B.B. 281
human capital
stock loss, suicide and 253–5
see also capital
Human Fertilization and Embryology
Authority (HFEA) 310
Hume, David xxviii, 2, 232
Humphries, P. 223
Hungary
abortion, basic demand and supply
model 322
prostitution, decriminalization and
legalization 197
religion, regulated 284
see also Europe
hyperbolic discounting
rational addiction and, drug
addiction 24–7
Iannaccone, Larry
addiction studies xxvi
religion and
capital 271
competition and doctrine 286
demand for 264
product and market competition
280–81, 283
regulated religion 284, 285
risk 278
sacrifice and stigma 275, 277
scarcity and benefits of, 265, 269,
270
Ibrahimo, M. 345
IFPI, see International Federation of

435

the Phonographic Industry
Impossibility Theorem (Arrow)
180–81, 191
incentives
economic, problems of illicit drugs
and, drug prohibition 44, 63
music industry record contract
397–8
police and prosecutors, for 6
income
abortion demand and, basic demand
and supply model 323
drug addiction and 23
drug use and 60–61
gambling and 374
prostitution and 6–7, 202, 209
high compensating differential
205–9
religion and 269
sport policy issues, caps 353–8
suicide and 249
see also revenue
indirect appropriability
music industry, copying in 413–14
inferior good
prostitution as 7
see also good
Informa Media Group 383
information
asymmetric, effects, prostitution and
210
interdependence
mutual, outcome uncertainty and,
nature of sport product 349–51
International Federation of the
Phonographic Industry (IFPI)
391, 394, 411–13, 419
International Union of Sex Workers
(IUSW) 197, 203
Internet
gambling and
UK 383, 384–5
USA 383–4
worldwide scope 367
online piracy, defined 411
pornography and 174–5, 176
supply and demand analysis
185–8
prostitution and 210
see also technology

436

Index

Iran
prostitution, decriminalization and
legalization 198
Iraq
prostitution, decriminalization and
legalization 198
Ireland
religion, market product and
competition 283
suicide, supply 243
see also Europe
Isaacs, A. 196
Italy
sport 353–5, 362
see also Europe
IUSW, see International Union of Sex
Workers
IVF (in vitro fertilization)
evolution of reproductive markets
and 296
market possibility for infertile couple
291
risk in demand and supply and 304
Jaﬀee, J.H. 59
jail, see prison
Japan
drug prohibition 51–2
prostitution, decriminalization and
legalization 197
religion, regulated 285
Jay, Deborah 416
Jesus Christ 257
Jobes, D.A. 258
joint products
economics of sex, marriage and
abortion and 2
Judaism 101 144, 145
Kaestner, R. 61, 62, 94
Kahane, L.H. 324, 325, 326
Kane, Thomas 329–31
Karberg, J.C. 107
Katz, M.L. 331–3, 336
Katzive, L. 317
Kay, J. 284
Kearney, M.S. 372, 375–6, 376, 379
Kenkel, D. 17
Kesenne, S. 361
Khan, F. 16

King, J. 74
Kleck, Gary 251
Kleiman, Mark A.R. 28, 81
Klein, B. 414, 415
Klerman, Jacob A. 328–9
Klock, S.C. 298
Kolczykiewicz, M. 305
Koo, L. 258
Korn, E. 193–4, 203, 205, 209, 218
Köszegi, B. 26
Krueger, A.B. 404–10
Kutchinsky, B. 189
Kuziemko, I. 51
labour market
outcomes, drug use and 60–61
see also market
Lady Chatterley’s Lover (Lawrence)
178, 179
Laibson, D.I. xxvi, xxx, 26
Laixuthai, A. 95
Latvia
suicide, statistics 237
see also Europe
law, see legislation
Lawrence, D.H. 178
learning
transitions into and out of marriage
and 160–61
Lederman, D. 53
legislation
abortion, USA 337
access, abortion as outcome and 325
assisted reproduction, USA 297, 305
changes, effect on divorce 137
competition, UK 363
consent, abortion as outcome and
325
drug liberalization
Netherlands 70
UK 70, 73, 82
drug prohibition, USA 48
gambling, USA 383
gun, USA 251
pornography, USA 183
sport, USA 352
Lehrer, E.L. 269
Lemieux, T. 95
Leonard, G. 353
Lerner, A.V. 414, 415

Index
Lester, David 241, 250
Levine, D. 414
Levine, P. 335
Levitsky, I. 375
Levitt, Steven D.
crime and
abortion xxii, xxv, 129, 333–7
capital punishment 114–16
measurement error extent, model
for determining 120
prison xxv
private protection 130
crime rate and, fall in US 131, 132
drug prohibition and, price and
quantity of drugs 51
Liebowitz, Stan 413, 414, 416, 417
Lillard, L. 160
Lithuania
crime rate 107
see also Europe
Livermore, G. 61
Loazya, N. 53
Lott, John R. 131, 336–7
lottery
UK
reasons for gambling and 373–5
social engineering, as 385, 386
USA
reasons for gambling and 375
social engineering, as 385–6
toleration of gambling and,
reasons for 376–8
worldwide scope of gambling and
368
see also gambling
Lucifora, C. 353–5
Ludwig, J. 251
Luksetich, W.A. 29
Lundberg, S. 152, 155, 291
Luxembourg
drug liberalization 69
see also Europe
MacCoun, R.J.
drug liberalization and
free market 88–9
government monopoly 81
government regulation 82
drug prohibition and, price and
quantity of drugs 47

437

prostitution and
legal availability and extent of 221
supply and population density
201–2
MacDonald, G. 402, 403
MacDonald, Margaret 196–7
MacDonald, Z. 28, 39
Maclean’s 85
Malthus, Thomas 293–4
Manser, M. 152
Manuel, Peter 394
Manzoni, Alessandro 101
Marcotte, D. 249
Marizco, M. 222–3
market
assisted reproduction contract, for
donation only, with 299
legal positive price, with 300
free, drug liberalization 87, 95
children and 93
competition 91
constraints 91–2
consumption in 87–90
problems with 93–5
risk 91
sanctions 92–3
supply 90–91
gambling, size 375–6
labour, outcomes and drug use 60–61
music, size of global 391–4
music industry copying, size 411–13
prostitution, structure 203–5
religious 280
product and competition 280–83
competition and doctrine 285–6
regulated religion 283–5
reproductive, see assisted
reproduction
sex, marriage equilibrium 208
market entry
prostitution 203
Markowitz, S. 248, 249
marriage
addiction and 5
budget constraints and indifference
curves in 147
cohabiting compared 144–5
decision-making in 150–51
bargaining among family
members 152–3

438

Index

family decision-making 151–2
spending decisions on private
goods 153–4
spending decisions on public
goods 154–5
spending versus saving 155
gains 145
children 146
love and sex 145–6
public goods 147–8
time 148–50
inter-racial, Becker on xxi–xxii
joint products and 2
prostitution and 194
economic welfare considerations
211
foregone market opportunity
202–3
high compensating wage
differential 205–9
reasons for 137
sequential decision-making 334
same-sex, Becker on xx
sex market equilibrium 208
shotgun, abortion and 331–3
transitions into and out of 155
changes 158–9
commitment 161
learning 160–61
policy interventions 163–5
relationship-breaking 157–8,
162–3
relationship-making 155–7
trends in 137, 139–40, 165–6
see also assisted reproduction;
divorce; family; sex
marriage contract
Becker on xix–xx
see also contract
marriage rate
France and Canada 139
Spain, Sweden and UK 140
USA 138
Marshall, Alfred 11, 401
Martin, G. 258
Marx, Karl 264
Mast, B.D. 78, 84, 87, 132
Mathews, Jay 296
Mathios, A.D. 17
Matthews, R. 212

May, T. 195, 205
Maynard, A. 29
McAleer, Phelim 195
McElroy, M. 152
McIntosh, J. 233, 237
McIntyre, R.J. 322
McKee, M. 378
McKenzie, Richard B. 68
Medoﬀ, M.H. 259, 324
melioration addiction theory
drug addiction and 20–21
see also addiction; drug addiction
Merck & Co., Inc. 59
Merz, Charles 57
Michael, R.T. 321
migration
prostitution and 195, 209
Miles, Sarah 247
Miller, A.S. 270
Mincer, J. 321
Mirer, T. 243, 250
Miron, Jeﬀrey A.
drug liberalization and, free market
89
drug prohibition and
crime 52–3
price and quantity 48, 51
supply 47
user welfare 57
Mobilia, P. 31, 381, 382
Mocan, H.N. 114
Model, K.E. 95
Moﬀatt, P.G. 202, 211, 212–13
Monitoring the Future (MTF) 33, 35
monopoly
government, drug liberalization and
79–81, 95
see also competition
Monroe, Marilyn 257
Moore, M.H. 27
Morgan, J.P. 57
Morris, C. 255
Morrison, Norma 257
MTF, see Monitoring the Future
murder rate, see crime rate
Murphy, Kevin M.
drugs and
addiction xxvi, 17, 18, 19, 37
liberalization, sin taxes 83, 86
music industry copying and

Index
exposure effects, space shifting,
network externalities and price
commitments 415
indirect appropriability 414
see also rational addiction theory
Murray, J.E. 270
music industry
copying
economic theory 413
empirical evidence 415–18
exposure effect 414
indirect appropriability 413–14
market, size of 411–13
network externality 415
price commitments 415
space-shifting effect 415
types of, defined 411
global market, size and scope 391–4
history 389–91
record contract 394–5
copyright ownership,
technological change and
399–401
enforcement 398
incentives 397–8
risk, division of 395–7
rock star
ability and record sales 401–4
Bowie Theory (Krueger) 409–10
superstar concerts 404–9
suicide and 255–8
technological change, influence of
419
unique features 389
see also compact disc
music market
global size of 391–4
see also market
Mustard, D. 380
Musto, David F. 48
myopic demand studies
price elasticity of demand for illicit
drugs and 34–5
see also demand
Napster
copyright ownership and
technological change 399
history 391
music copying and 414–17

439

NARAL, see National Abortion
Rights Action League
Nash, John 152
National Abortion Rights Action
League (NARAL) 325
National Basketball Association
(NBA) 355–6
National Comorbidity Survey (NCS)
248, 249
National Football League (NFL) 356,
359
National Household Survey of Drug
Abuse (NHSDA) 12, 33
National Longitudinal Survey of
Youth (NLSY) 62
National Research Council 58, 62
National Survey on Drug Use and
Health (NSDUH) 12
NBA, see National Basketball
Association
NCS, see National Comorbidity
Survey
Neale, W.C. 348, 351
Neely, Richard 124
Netherlands
drug addiction 16, 34
drug liberalization 70, 88–9
prostitution 197, 223
see also Europe
network externality
music industry copying and 415
The New York Times 58, 174, 257,
409
New Zealand
prostitution, decriminalization and
legalization 198
NFL, see National Football League
NHSDA, see National Household
Survey of Drug Abuse
Nichols, M. 382
Nirvana 255–6
Nisbet, C.T. 32
NLSY, see National Longitudinal
Survey of Youth
Norberg, K.E. 237, 240, 247–8
North America, see United States of
America
Norway
assisted reproduction, risk in supply
and demand 304

440

Index

NSDUH, see National Survey on
Drug Use and Health
Nurco, D.N. 59
Oberholzer, F. 418
The Observer 189
occupation
suicide by, UK 254
O’Connell Davidson, J. 204
Oﬃce of Fair Trading (OFT) 363
Ohsfeldt, R.L. 325
Olds, K. 284
Olekalns, N. 31
Olson, D.V.A. 283
online piracy
defined 411
see also music industry
Ono, Y. 279
options contract
music industry, incentives 397
see also record contract
Ordonez, J. 396
Orphanides, A. 26–7
Osborn, A. 70
Ostrum, B.J. 83
Oswald, A. 207
O’Toole, Laurence 183
outcome
abortion as 321
access laws 325
anti-abortion activities 325–7
consent laws 325
cultural factors 327
demand and supply, basic model
321–4
public subsidies 324–5
labour market, drug use and 60–61
social, pornography and, negative
externalities 172–3
uncertainty, nature of sport product
mutual interdependence and
349–51
owner utility and 348–9
owner
utility and outcome uncertainty,
nature of sport product 348–9
ownership
copyright, music industry
indirect appropriability and 413
technological change and 399–401

Pacula, R.L. 17, 33, 74
Paley, William, 101–2
Parker, Dorothy 229
Pascal, Blaise 243, 265, 278
Paton, D. 374, 375, 384
Peru
drug prohibition and crime 55
Peters, S.A. 202, 211, 212–13
Philipson, T. 130
Phillips, D.P. 257
Phlips, Louis 11
Pindyck, Robert S. 243, 244
Piore, M.J. 194
piracy
commercial, defined 411
online, defined 411
see also music industry
Plath, Sylvia 255
Platt, S. 251
Plotz, David 295
Png, I. 418–19
Poland
crime rate 107
religion, market product and
competition 283
see also Europe
police
arrests
drug offences, USA 49
prostitution and related offences,
USA 200
detection of pornography, Internet
effect and 186
incentives for 6
policy
anti-suicide, economics of 258–61
capital punishment
public opinion, USA 117
supporters’ reasons for favouring,
USA 118
divorce and 163–5
drug addiction and 26–7, 38–40
drug liberalization
Europe 69–73
USA 69, 73–9
gambling 386–7
economic regenerator, as 379–80
illegal betting 382–3
Internet threat 383–5
social costs 380–82

Index
social engineering, as 385–6
marriage and, transitions into and
out of 163–5
pornography regulation 183–4
prostitution 221–4
sport 352–3, 364–5
broadcasting rights, collective sale
of 362–4
cheating 358–9
free agency and salary caps 353–8
revenue-sharing 359–62
see also government
Pollak, R. 152, 155
pornography
defining
economic perspective 179–81
legal and philosophical
perspective 177–8
demand analysis of 181
addiction 182–3
price 181
variety-seeking 183
economic good, as
consumption, substitute or
complementary goods 189–90
social benefits 188–9
economists’ interest in 171–2,
190–91
Internet effect, supply and demand
analysis 185–6
detection and punishment 186
marginal productivity of
preventers and consumers
187–8
prevention 186–7
production and consumption 173–5
profits 176–7
prostitution and 194–5
regulation
methods 183–4
problems 184
size distribution of firms 175–6
social outcomes, negative
externalities 172–3
Portugal
crime rate 107
drug addiction statistics 12, 16
drug liberalization 69
see also Europe
Posner, E.A. 292

441

Posner, Richard A. xxviii, 130, 292
pregnancy
endogenous, abortion as cause and
329–31
pregnancy resolution and, sequential
decision-making 331
Prelec, D. 20–21
Presley, Elvis 395
Preston, I. 359
Pretty Woman (film) 196
prevention
crime, demand for 113–22
pornography, Internet effect and
186–7
see also regulation
prices
abortion, basic demand and supply
model 322–3
adoption and 7
assisted reproduction 305–10, 311
drug addiction and 18–19, 24, 35–8
drug quantity and, under
prohibition 46–53
demand 45
supply 45–7
drugs, USA 50
effective, demand for illicit drugs
and 27–8
elasticity
abortion demand and 323
cigarette demand and 37
illicit drug demand and 27–38
gambling and 372–4
substitute goods 374–5
legal positive, market for assisted
reproduction contract with
300
music industry copying and 415
pornography demand analysis and
181
prostitution and 6–7, 210
Punternet 214–18
religion and 269
reproductive markets and, demand
in 302–3
suicide and 241–3
superstar concert, USA 404–10
see also costs
prison
population rate, international 109

442

Index

sentences to, drug convictions, USA
78
time
criminals, for, USA 128
served for drug convictions, USA
79
private goods
spending on, decision-making in
marriage and 153–4
see also goods
private protection
crime, falling US rate and 130–32
probation
sentences to, drug convictions, USA
78
Prochaska, J.O. 17
product
nature of, sport 348–51
religious market and 280–83
theological, space 282
production
euphoria commodity, relative theory
of addiction and 19
household, marriage gains and
148–50
joint products, economics of sex,
marriage and abortion and 2
pornography 173–5
demand analysis 182–3
Internet effect and 187–8
profit
pornography and 176–7
prosecutors
incentives for 6
prostitutes
numbers
Europe 198, 199
USA 201
prostitution
age and 6
arrests for related offences and, USA
200
defined 193
economic analysis
asymmetric information effects
210
economic welfare considerations
210–11
extensions 209
high compensating wage

differential 205–9
market entry 203
market structure 203–5
theories 202–3
economic principles, application of
193, 224–6
empirical studies 211
previous work 212
Punternet 212–20
industry characteristics 193–4
drugs and crime link 195
marriage link 194
migration link 195
pornography link 194–5
provision 195–7
inferior good, as 7
policy 221–4
prices and wages and 6–7
services, price of 214
age of provider and 216
determinants, OLS regression
identifying 217
duration in minutes and 215
sunk costs and 2
victimless crime, as 7
world, overview 197–202
see also sex
protection
private, against crime, falling US
crime rate and 130–32
provision, see supply
public goods
marriage gain, as 147–8
spending on, decision-making in
marriage and 154–5
see also goods
public subsidies
abortion 324–5
punishment
pornography, Internet effect and 186
probability and severity of, criminal
justice system resource
allocation and 126–9
see also capital punishment
Punternet
background 212–13
dataset, exploratory analysis 213–14
price determinants 214–18
satisfaction determinants 218–20
see also prostitution

Index
Quirk, James 356, 361
Rabin, M. 370
race, see ethnicity
Rachlin, H. 19
Radin, Margaret Jane 310
Rahman, A. 317
Ramello, G.B. 394, 413
Rao, V. 212
Rascher, D.A. 361
Rasmussen, D.W. 78, 84, 85, 87
rational addiction theory
addictive behaviour and 24
drug addiction and 21–4, 27–8
hyperbolic discounting and other
extensions 24–7
gambling and 381–2
pornography and 182
see also addiction; drug addiction
rational demand studies
price elasticity of demand for illicit
drugs and 35–8
see also demand
rationing
criminal justice system resources
122–6
Read, D. 26
record contract
music industry 394–5
copyright ownership,
technological change and
399–401
enforcement 398
incentives 397–8
risk, division of 395–7
see also contract
Recording Industry Association of
America (RIAA) 391, 396
Regner, T. 399–401
regulation
drug supply and 46
drugs, USA, spending 76
government, drug liberalization and
81–2, 95
pornography 183–4
Internet and 186–7
religion 283–5
see also prevention
relationship
breaking, transitions into and

443

out of marriage and 157–8,
162–3
change in
divorce and 158–9
transitions into and out of
marriage and 158–9
making, transitions into and out of
marriage and 155–7
see also divorce; family; marriage
relative addiction theory
drug addiction and 18–20
see also addiction; drug addiction
religion
commodity, as 264, 286–7
demand 264–5
benefits 268
determinants of 268–70
religious capital 270–71
scarcity 265
divorce and 164
gambling and 377
market 280
competition and doctrine 285–6
product and competition 280–83
regulated religion 283–5
marriage and 144, 145, 164
pornography and 182
risk 278–9
exclusivity strategies 279–80
suicide and 243
supply 271–2
external effects and free-riding
273–4
optimal firm size 277–8
sacrifice and stigma 274–7
see also culture
rent of ability
free agency and salary caps, sport
policy issues 353
reproductive market, see assisted
reproduction
resources
criminal justice system, allocation of
122
probability and severity of
punishment and 126–9
rationing 122–6
Reuter, Peter
drug liberalization and
free market 88–9

444

Index

government monopoly 81
government regulation 82
drug prohibition and, price and
quantity of drugs 47
prostitution and
legal availability and extent of 221
supply and population density
201–2
revealed preference principle
value and 1
revenue
superstar concerts, USA 408
distribution to artists 406
sharing, sport, collective sale of
broadcasting rights and 359–64
see also income
Reynolds, Helen 222
RIAA, see Recording Industry
Association of America
Richardson, J.T. 275
rights
broadcasting, collective sale of, sport
policy 362–4
risk
assisted reproduction demand and
supply 303–5
crime and 112
division of, music industry record
contract 395–7
drug liberalization and free market
and 91
religion and 278–9
exclusivity strategies 279–80
sex and 145–6
risk aversion
gambling and 370–71
Robins, L.N. 59
Robinson, W. 375
Roche, C.M. 221
rock ’n’ roll, see music industry
rock star, see celebrity
Rodriguez Andres, A. 249
Rolling Stone Encyclopedia of Rock &
Roll 404, 407, 409
Romania
abortion, fertility rate and access to
328
see also Europe
Rose-Ackerman, S. 272
Rosen, C. 255

Rosen, S. 196, 353, 357, 401
Rosenberg, N. 86
Rosenthal, R.W. 247
Rosenzweig, M.R. 294
Ross, S.F. 359
royalty contract
music industry, incentives 397–9
see also record contract
Rubenstein, R. 385
Rubin, P.H. 114
Russia
crime rate 107
suicide statistics 237
see also Europe
sacrifice
religion and, supply 274–7
Saﬀer, H. 33, 34
Saint-Paul, G. 292
salary, see income
sales
CD
commercial piracy effects 418–19
Fine report 416
Jay report 416–17
collective, of broadcasting rights,
sport policy and 362–4
global recorded music by format 393
record, ability and, music industry
celebrity 401–4
Salinger, J.D. 211
Samaritans 237–40, 246, 259
Sanderson, A. 357
Sargeant, M. 212
satisfaction
prostitution and, Punternet 218–20
Sauer, R. 377, 387
Savage, L. 371
saving
spending versus, decision-making in
marriage and 155
Sawkins, J.W. 268, 269, 270
Scafidi, B. 385
scarcity
religious demand and 265
Schelling, Thomas C. 11, 231
Schmidtke, A. 258
Schopenhauer, Arthur xxviii
Schultz, T.P. 294
Schulze, G.G. 402

Index
Scitovsky, Tibor 270
Scotland, see United Kingdom
Scott, F. 373, 374
Scully, G.W. 353
Seaman, P.T. 268, 269, 270
Sen, A.K. 179, 180, 190–91
services
abortion
availability 317–20
usage 316–17
assisted reproduction, Europe and
USA 293
prostitution
price of 214, 215, 216, 217
sex
joint products and 2
love and, as marriage gain 145–6
marriage market equilibrium 208
premarital, sequential decisionmaking 334
same-, marriage, Becker on xx
suicide rate by
Europe and USA 235–6
international 238–9
see also assisted reproduction;
marriage; prostitution
Sexton, Anne 255
Sharpe, J.R. 178
Shughart, William F. 83
Siegel, D. 374, 375, 384
Sigel, L.Z. 185
Silva, F. 394, 413
Silver, J. 177–8
Silverman, L.P. 32, 55
Simmons, R.
gambling and
consumer behaviour 371, 372, 374
economic regenerator, as 379
Internet threat 384–5
lottery 373, 375
social engineering, as 386
sport and, free agency and salary
caps 353–5
Simon, W. 202, 212
Singapore
prostitution, decriminalization and
legalization 197
Sirtalan, I. 31
size
firm

445

distribution, pornography and
175–6
optimal, supply of religion and
277–8
market
assisted reproduction 297–8, 311
copying, music industry 411–13
gambling 375–6
music, global 391–4
Skidmore, M. 378
Sloan, F. 31
slot machine gambling
reasons for gambling and 372–3
see also gambling
Smith, Adam 86, 101, 232–3, 283
The Smiths 246
Snare, Annika 189
soccer, see football
social benefits
pornography as economic good and
188–9
see also benefits
social costs
gambling policy and 380–82
illicit drug use 38
see also costs
social engineering
gambling policy as 385–6
social outcomes
pornography, negative externalities
172–3
see also outcome
Soss, N. 232, 243, 244, 248
South Africa
crime rate 107
see also Africa
space-shifting eﬀect
music industry copying and 415
Spain
assisted reproduction, surrogate
motherhood 292
divorce trends 142
drug addiction statistics 16
marriage trends 140
sport, revenue-sharing and collective
sale of broadcasting rights 362
suicide statistics 233
see also Europe
Spectator 195, 196
spending

446

Index

decision-making in marriage and
private goods, on 153–4
public goods, on 154–5
saving versus 155
sperm or egg donation, see donor
insemination
sport
cartel behaviour 351–2
economists’ interest in, reasons for
345–6, 364
illegal betting and, USA 382–3
policy issues 352–3, 364–5
broadcasting rights, collective sale
of 362–4
cheating 358–9
free agency and salary caps 353–8
revenue-sharing 359–62
product, nature of 348–51
team and league objectives 346–8
see also individually named sports
Spruill, N.L. 32, 55
Stack, S. 251, 253, 257, 258
Staiger, Douglas 329–31, 335
Stanley, L. 361
Stark, Rodney
religion and
demand for 264
product and competition 283
regulated religion 284, 285
sacrifice and stigma 275
scarcity and benefits of 265
Stern, S. 160
Stigler, George J. xxvi, 18, 19, 21, 27
stigma
prostitution and 203
religion and, supply 274–7
Stolinsky, D. 251
Stonebraker, R.J. 277
Stout, J.E. 298
Stranahan, H. 385
STRIDE, see System to Retrieve
Information on Drug Evidence
Strobl, E.A. 390
Strom, S. 202
Strumpf, K. 382, 383, 418
subsidies
public, abortion and 324–5
substitute goods
complementary or, pornography as
economic good and 189–90

euphoria commodity, relative theory
of addiction and 19–20
gambling and, price 374–5
prostitution and 204–5
religion and 269–70
see also goods
suicide
alcohol and drugs and 248–9
anti-suicide policy, economics of
258–61
Becker on xxviii–xxix
business cycle and 249–51
children and 247–8
China 2
cost of 4
Durkheim approach
altruistic suicide 231–2
anomic suicide 230
egoistic suicide 231
fatalistic suicide 230–31
economic costs for society
celebrity suicides 255–8
human capital stock loss 253–5
economic principles applied to 2
economists’ interest in 229, 232–3,
261
gas and 242
gun availability and 251–3
microeconomics 240–41
demand 241
Hamermesh–Soss model 243–5
supply 241–3
occupation, by, UK 254
optimal reduction, ‘old’ and ‘young’
suicides 260
para-suicide signalling games 247
preferences 245
higher and lower 245–6
long-run and short-run 245
temporary and permanent 246
statistics 233–7, 242, 252, 254
problems with 237–40
suicide rate
Europe and USA 234
by sex 235–6
international and firearm 252
age and gender, by 238–9
Sullivan, D.H. 268
Sullivan, E. 202, 212
Sullum, J. 58

Index
Sun Records 395
superstar, see celebrity
supply
abortion
anti-abortion activities effects 326
basic model 321–4
adoption prices and 306
assisted reproduction 298–301
risk in 303–5
children, of, microeconomics before
reproductive markets 293–4
criminal offences 110–13
donor insemination 299–300
drug addiction and, future research
40
drug liberalization and
European policy 72–3
free market and 90–91
US policy 74–9
drug prohibition and 45–7
pornography, Internet effect analysis
185–8
prostitution 195–7
relative, teachers and sports stars
355
religion and 271–2
external effects and free-riding
273–4
optimal firm size 277–8
sacrifice and stigma 274–7
suicide and 241–3
surrogate motherhood and 300–301
risk in 305
surrogate motherhood
ethics and 310
evolution of reproductive markets
and 296
market possibility for infertile couple
292
risk in supply and demand 305
supply and 300–301
see also assisted reproduction
Sutton, M. 32
Swan, G.E. 17
Sweden
assisted reproduction, adoption 291
abortion, drugs for 315
divorce trends 142
drug liberalization 70
marriage trends 140

447

prostitution policy 221–2
suicide statistics 233
see also Europe
Switzerland
drug liberalization 69, 81
prostitution, decriminalization and
legalization 197
see also Europe
System to Retrieve Information on
Drug Evidence (STRIDE) 33, 34,
35
Szymanski, S. 345, 359, 361
Taiwan
suicide, business cycle and 250
Takeyama, L. 415
Tanner, S. 374–5
Tauras, J.A. 33
taxation
drug liberalization and 80–81, 82–7,
95
drug supply and 46
gambling and 377–8, 384–5
pornography and 184
prostitution and 209, 222
technology
change in
copyright ownership and, music
industry record contract
399–401
gambling and 367, 386
influence on music industry 419
record label/artist relationship,
impact on 400
see also Internet
terrorism
fear economy and, Becker on
xxiv–xxv
suicide and, Becker on xxix
Thailand
prostitution policy 222, 223
Thaler, R.H. 24, 370
Thalheimer, R. 372, 375
Thew, H. 212
Thomas, D. 155
Thornton, Mark 68–9
time
allocation of, between legal and
illegal activities 112
marriage gains and 148–50

448

Index

prison
criminals, for, USA 128
served for drug convictions, USA
79
Titmuss, Richard M. 299
Tollison, Robert D. 272
Troilo, G. 183
Tschirhart, J. 361
Tucker, C. 390
Tullock, Gordon 68
UEFA 358
UK, see United Kingdom
Ukraine
suicide, statistics 237
see also Europe
unborn fetus
drug use and 62–3
see also birth
United Arab Emirates
prostitution, decriminalization and
legalization 198
United Kingdom (UK)
abortion
drugs for 315
service use 317, 318, 319
abortion rate 318, 319
assisted reproduction
reproductive markets, evolution of
295
risk in supply and demand 304
surrogate motherhood 292
crime rate 103, 104, 107
divorce trends 142, 143
drug addiction statistics 16
drug liberalization
demand-side policy 69–70
government regulation 82
supply-side policy 72–3
drug prohibition, use and addiction
59
gambling
changes in 386
economic regenerator, as 380
Internet 383, 384–5
National Lottery and reasons for
gambling 373–5
National Lottery as social
engineering 385, 386
social costs of 381

toleration of, casinos and reasons
for 378–9
marriage trends 140
music industry, size of global market
and 391
pornography
child 175, 178
defining 178
Internet effect 185
regulation problems 184
prostitution
drugs and crime link 195
market structure 204
policy 221, 223
provision 196–7
Punternet 212–20
statistics 202
sport
free agency and salary caps 357
revenue-sharing and collective sale
of broadcasting rights 362,
363–4
suicide
economic costs 253
occupation, by 254
preferences 246
statistical problems 237–40
supply 241–3
see also Europe
United Nations Oﬃce for Drugs and
Crime (UNODC) 12, 13
United States of America (USA)
abortion
access laws 325
access to, fertility rate and 328–9
anti-abortion activities 326
ban on 337–8
crime and 333–7
demand and supply, basic model
321–2, 323
endogenous pregnancy 330–31
methods 315, 316
public subsidies 324–5
service availability 319–20
service use 317, 318, 319
shotgun weddings and 331–3
abortion rate 318, 319
assisted reproduction
adoption 291
market size 297–8

Index
price, supply and demand and
306–7, 308–9
reproductive markets, evolution of
295, 296, 297
risk in demand and supply 303–5
services 292, 293
surrogate motherhood 292
capital punishment
number of executions 115
public opinion 117
supporters’ reasons for favouring
118
crime
capital punishment 113–16,
117–18
criminal justice resources
allocation, probability and
severity of punishment and
127–9
criminal justice resources
rationing 123, 124–5, 125–6
standard empirical model of
crime, extensions 119
crime rate
fall in, 129–32
statistics 103, 105, 107
divorce
laws, effect on rate 137, 164
policy interventions 163
trends 137–8, 143–4
drug addiction
consumption statistics 12, 16
conventional demand studies 32–4
rational demand studies 35–8
drug convictions 77, 78, 79
drug liberalization
economists and 68–9
free market 88
government monopoly 79, 80
policy 69, 73–9
taxation and 83–4, 87
drug offences, arrests 49
drug prohibition
crime and 53–5
drug quality and 56–7
price and quantity of drugs under
48, 50, 51–2
use and accidents 61–2
use and addiction 59
drugs

449

control, spending 76
prices 50
quantity seized 75
gambling
casinos, reasons for gambling and
372
changes in 386, 387
economic regenerator, as 379, 380
gross annual wagering 368
illegal 382–3
Internet 383–4
lottery and reasons for gambling
375
lottery and reasons for toleration
of 376–8
lottery as social engineering 385–6
social costs of 380–81, 381–2
murder rate 54
music industry
global market, size of 391
history 390
music copying, empirical evidence
416–17, 417–18
record contract, division of risk
and 396–7
superstar concert prices 404–10
pornography
defining 177–8
Internet effect 185
negative externalities 172–3
regulation methods 183
regulation problems 184
statistics 174
prison
sentences to, drug convictions 78
time for criminals 128
time served, drug convictions 79
prostitutes, numbers, as percentage
of population 201
prostitution
arrests for related offences and 200
decriminalization and legalization
198
policy 222–3, 224
satisfaction determinants 220
statistics 200, 201, 202
religion
demand 264
market product and competition
283

450

Index

regulated religion 283–5
risk, exclusivity strategies 279–80
supply, sacrifice and stigma 275–6
sport
economists’ interest in, Europe
compared 345–6, 364–5
free agency and salary caps 353,
355–6
policy issues generally 352–3
revenue-sharing and collective sale
of broadcasting rights 359
team and league objectives,
Europe compared 347
suicide
alcohol and drugs and 248–9
anti-suicide policy, economics of
259
celebrity 257, 258
children and 247
economic costs 253
gun availability and 251–3
statistics 233, 234–7, 240
suicide rate 234
by sex 235–6
UNODC, see United Nations Oﬃce
for Drugs and Crime
US Department of Justice 55, 125, 126,
127–8
USA, see United States of America
use
abortion services 316–17
drug
accidents and 61–2
addiction and 58–61
effect, crime and 55–7
global 13
labour market outcomes and 60–61
prevalence in USA, Australia and
Europe 14, 15
social costs of 38
unborn fetus and 62–3
user
drug, ‘primrose path’ to addiction 20
utility
drug addiction and, rational
addiction theory 22–3
owner, outcome uncertainty and,
nature of sport product 348–9
transferable, relationship-making
and 157

utility function
assisted reproduction, demand and
301–2
crime and 110
religion and 268–9
suicidal children and 247–8
utility theory
expected, gambling and 369–71
Vakil, F. 32
value
drugs, of 5
relationship-breaking and 157–8
relationship-making and 155–7
religion and 270
revealed preference principle and 1
Van Leeuwen, B. 26
Van Ours, J.C. 34
Varian, H. 155
variety-seeking
pornography demand analysis and
183
Vaughan, W. Jr. 20
Vaughan Williams, L. 374, 375, 384
victimless crime
prostitution as 7
see also crime
violence
drugs and 5
see also crime
Voas, D. 283
Vogel, Harold L. 396
Vogel, R.J. 28
Vrooman, J. 352, 356
wages, see income
Wagstaﬀ, A. 29
Wales, T. 152
Walker, D. 381
Walker, I. 372, 386
Waters, T. 31
wedding, see marriage
Weiss, Y. 155
welfare
drug user, prohibition and
drug quality 56–8
use and accidents 61–2
use and addiction 58–61
use and labour market outcomes
60–61

Index
use and unborn fetus 62–3
economic, prostitution and 210–11
welfare economics
drug prohibition and addiction and
38–40
West, R. 16
White, M.D. 29
Whitley, J.E. 336–7
Whitman, Douglas G. 80–81
WHO, see World Health Organization
Wilding, P. 385
Wildman, S.S. 396–7
Williams, H.C.S. 268, 269, 270
Willis, R. 155
Winick, C. 172
Wolfenden, J. 221
Wolfgang, M.E. 257
Woolley, F. 155
world
abortion rate 316
crime rate, compared 103–9
gambling scope 367–9
illicit drug use 13
murder rate 106, 108
music market, size of 391–4
prison population rate 109
prostitution, overview 197–202
recorded music sales

451

by format 393
real value of 392
suicide rate 252
age and gender, by 238–9
World Health Organization (WHO)
202
Wright, R. 56
Yang, Bijou 241, 250
Yellen, J.L. 331–3, 336
Youssef, H.146
Youth Risk Behavior Survey (YRBS)
248–9
YRBS, see Youth Risk Behavior
Survey
Yuan, Y. 48
Yunker, J. 114
Zaleski, P.A. 277
Zavodny, M. 335
Zech, C.E. 277
Zedlewski, E. 132
Zelizer, Viviana 291
Zentner, A. 417
Zervos, D. 26–7
Zillmann, D. 173
Zinberg, N.E. 59
Zwiebel, J. 57

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.6
Linearized                      : Yes
XMP Toolkit                     : 3.1-702
Modify Date                     : 2006:11:10 11:26:30-05:00
Create Date                     : 2006:11:03 20:50:36-05:00
Metadata Date                   : 2006:11:10 11:26:30-05:00
Format                          : application/pdf
Title                           : Economics Uncut : A Complete Guide to Life, Death and Misadventure
Creator                         : Bowmaker, Simon W.
Document ID                     : uuid:a3742d3f-dc1e-40d4-8a08-a6d6350aaf3e
Instance ID                     : uuid:c8a6af06-4118-447c-886a-6e3eb87bb23d
Producer                        : Acrobat Distiller 5.0.5 for Macintosh
Has XFA                         : No
Page Count                      : 485
Author                          : Bowmaker, Simon W.

EXIF Metadata provided by EXIF.tools

Economics Uncut : A Guide To Life, Death And Misadventure Simon W. Bowmaker Death, Edward Elgar

Navigation menu

Versions of this User Manual:

Views

Navigation