The Deutsche Bank Guide To Exchange Rate Determination FX Broker DB

User Manual:

Open the PDF directly: View PDF .
Page Count: 160 [warning: Documents this large are best viewed by clicking the View PDF Link!]

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 1

Deutsche Bank@

May 2002

FX Research

Global Markets Research

Michael R. Rosenberg

Head of Global FX Research

David Folkerts-Landau

Managing Director, Head of

Global Markets Research

Exchange-Rate Determination

in the Short, Medium, & Long Run

The Deutsche Bank

Guide to

Exchange-Rate

Determination

A Survey of

Exchange-Rate

Forecasting Models and

Strategies

Bandwagon Effect/

Trend-Following Behavior

Real

Interest-Rate

Differentials

Purchasing Power Parity

Capital

Flows

Fiscal

Policy Portfolio-Balance

Considerations

Current

Account

Trends

Relative

Economic

Growth

Monetary

Policy

Investor Positioning

FX Options Market

Positioning

Risk Appetite

Investor Sentiment

Net Foreign Assets

Productivity Trends

Savings/Investment

Balance Trends

Persistent Trend in

Terms-of-Trade

Exchange Rate

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

2Deutsche Bank Foreign Exchange Research

Introduction .................................................................................................... 4

Exchange Rate Determination in the Short Run ........................................ 7

Currency Forecasting Using Technical Analysis .........................................11

Sentiment and Positioning Indicators ....................................................... 21

Currency Options Market ......................................................................... 22

Order Flow and the Determination of Exchange Rates ........................... 24

Investor Positioning and the Trend in Exchange Rates ............................. 27

Risk Appetite Shifts and Currency Trends................................................. 28

Exchange Rate Determination in the Long Run ....................................... 31

Purchasing Power Parity ........................................................................... 33

The Macroeconomic-Balance Approach to Long-Run

Exchange-Rate Determination ............................................................ 42

The Real Long-Run Equilibrium Exchange Rate ....................................... 43

Productivity Trends and Exchange Rates .................................................. 46

Terms of Trade and Exchange Rates......................................................... 50

Net International Investment and the Equilibrium Exchange Rate........... 51

Long-Term Cycles in Exchange Rates....................................................... 53

Overshooting Exchange Rates ................................................................. 60

Exchange Rate Determination in the Medium Run.................................. 63

International Parity Conditions.................................................................. 65

Real Interest-Rate Differential Model ....................................................... 67

Forward-Rate Bias Strategy...................................................................... 72

Current-Account Imbalances and the Determination of

Exchange Rates................................................................................... 77

Capital Flows and Exchange Rates........................................................... 86

Mundell-Fleming Model ........................................................................... 96

Monetary Approach ................................................................................ 104

Portfolio-Balance Approach ..................................................................... 108

Fiscal Policy .............................................................................................115

Economic Growth....................................................................................119

Central-Bank Intervention ....................................................................... 124

Anticipating Currency Crises in Emerging Markets ............................... 129

References................................................................................................... 155

Table of Contents

Datastream International, Inc. is the source of the majority of data

used in the charts and tables in this publication. Other sources are

noted individually.

Market sentiment data is by permission of Consensus, Inc.,

Con-

sensus, National Futures and Financial Weekly,

(1) (800) 383-1441 or (1) (816) 373-3700

www.consensus-inc.com

Sources:

May 5, 2002

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 3

"If you think writing about the fortunes of the stock

market is tricky, try getting your arms around cur-

rencies."

Bill Gross

PIMCO

Investment Outlook, January 2002

"Having endeavored to forecast exchange rates

for more than half a century, I have understand-

ably developed significant humility about my abil-

ity in this area...."

Alan Greenspan

Remarks Before the Euro 50 Roundtable

Washington D.C., November 30, 2001

Deutsche Bank Guide to Exchange-Rate Determination

"Explaining the yen, dollar and euro exchange rates

is still a very difficult task, even ex-post."

Kenneth Rogoff

Economic Counselor and Director of Research,

International Monetary Fund

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

4Deutsche Bank Foreign Exchange Research

Getting the exchange rate right is a critical objective of all

international investors. Unfortunately, getting the exchange

rate right on a reasonably consistent basis is far from easy.

As anyone involved in the business of currency forecast-

ing can attest, it can be a humbling experience.

Currency forecasts can go awry for a variety of reasons.

For instance, if one's expectation of the direction in which

fundamental-based forces are heading is flawed, so will

be one's forecast of a currency's future path. Even if one's

interpretation of the underlying fundamental forces were

correct, currency forecasts might still go awry if short-term

technical forces carried exchange rates far from their fun-

damental equilibrium path.

Scores of empirical studies have found that fundamental-

based models tend to perform poorly in terms of explain-

ing exchange-rate trends, particularly over short-term peri-

ods. However, fundamental-based models tend to work

better over medium and especially longer-run horizons.

Unfortunately, most fund managers, whose performances

are evaluated over relatively short time spans these days,

are often not willing to risk significant sums of capital on

the basis of longer-term, fundamental-based projections.

That is why many market participants have recently turned

their attention away from longer-run fundamental-based

forecasting approaches in favor of shorter-term forecast-

ing tools such as technical-based trend-following trading

rules. In addition, there has recently been significant inter-

est in flow, sentiment, and positioning indicators to deter-

mine the exposure of market participants to the individual

currencies. Such indicators are often used as contrarian

indicators to determine whether a currency is significantly

overbought or oversold, and thus ripe for a correction.

Given the wide variety of forecasting approaches, we

thought it would be useful to put together a guidebook

that summarized each of those approaches in an easy-to-

read format. Our intention was to create a user-friendly

format where the written text was purposely kept to a

minimum and where the charts and tables—about 400 in

all—would tell the story.

This guidebook recognizes that the tools required for short-

term investors differ significantly from those needed for

medium and long-term currency managers. Hence, the

guidebook devotes separate chapters to the determina-

tion of exchange rates over short, medium, and long-term

horizons.

The adjacent schematic diagram provides a convenient il-

lustration of the layout of this guidebook and highlights

the myriad of channels through which fundamental and

technical forces jointly affect exchange rates. Some of those

channels will tend to exert a more profound impact on ex-

change rates in the short run, while others will tend have a

more profound impact in the medium or long run. We ex-

plore each of those channels in the chapters that follow.

In the chapter entitled "Exchange Rate Determination in

the Short Run," we investigate the potential risks and re-

wards of using a variety of short-run forecasting tools in

formulating short-term FX strategies. These include mov-

ing-average trend-following trading rules, sentiment and

positioning surveys, FX dealer customer-flow data, infor-

mation embedded in currency option prices, and risk ap-

petite indices.

We find that moving-average trading rules would have gen-

erated significant risk-adjusted excess returns over rela-

tively long periods for most major currency pairs, although

losing trades tend to occur far more frequently than win-

ning trades, in many cases by a factor of 3 to 1. The high

frequency of losing trades suggests that moving-average

trading rules can be risky over short periods and that an

investor would need considerable risk capital on hand to

absorb such losses to stay in the game until exchange rates

eventually become more highly trended.

The evidence on flow, sentiment, and positioning surveys

suggests that these indicators should be viewed more as

contemporaneous rather than as leading indicators of ex-

change-rate movements. We argue that such indicators can

be useful as confirmation indicators in conjunction with

traditional trend-following trading rules in formulating short-

term FX strategies.

In the chapter entitled "Exchange Rate Determination in

the Long Run," we explore the fundamental forces that

give rise to long-term cycles in exchange rates. The chap-

ter begins by noting that deviations from estimated PPP

values have tended to be large and persistent, suggesting

that fundamental forces other than relative national infla-

tion rates have played a key role in driving the long-term

path that exchange rates have taken. We investigate a va-

riety of fundamental variables that have had some suc-

cess in explaining the long-term path that currencies have

taken, including relative productivity growth, persistent

trends in a country's terms of trade, long-term trends in

net foreign asset and liability positions, and long-term

trends in national savings and investment.

Introduction

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 5

In the chapter entitled "Exchange Rate Determination in

the Medium Run," we investigate a wide range of cyclical

forces that have caused currencies either to rise or fall on

a medium-term basis relative to their long-run equilibrium

path. In many cases, these medium-term deviations from

the long-run equilibrium path have been quite sizeable and

persistent. We find that medium-term trends are influenced

by a variety of macroeconomic indicators such as the trend

in real interest-rate differentials, current and capital-account

balances, relative monetary and fiscal policies, relative eco-

nomic growth, and portfolio-balance considerations.

Finally, in the chapter entitled "Anticipating Currency Cri-

ses in Emerging Markets," we set out to identify those

economic and financial variables that have had success in

correctly predicting whether an emerging-market currency

might be vulnerable to a speculative attack, and whether it

is possible to construct an early-warning system that can

successfully pinpoint when a speculative attack might oc-

cur. Empirical research finds that crises-prone currencies

typically display a number of classic symptoms that warn

of an impending attack. Those symptoms include exces-

sive real appreciation of the emerging-market currency,

weak domestic economic growth, rising unemployment,

an adverse terms of trade shock, deteriorating current-ac-

count balances, excessive domestic credit expansion, bank-

ing-system difficulties, unsustainably large government

budget deficits, overly expansionary monetary policies, a

high ratio of M2 money supply to reserves, foreign-ex-

change reserve losses, falling asset prices, and/or a huge

build-up in short-term liabilities by either the private or public

sector.

The overall conclusion one draws from a variety of empiri-

cal studies is that the success of early-warning systems in

terms of generating correct out-of-sample projections is

mixed. While most early-warning models can point to a

significant number of correctly predicted crises, those

same models also have a tendency to generate a sizable

number of false alarms and missed crises.

Perhaps all that one can say after reviewing all the differ-

ent approaches to exchange-rate determination is that no

single approach has a monopoly on being right all of the

time. Some strategy systems such as moving-average trad-

ing rules and forward-rate bias strategies appear to have a

long-run track record of success, but one needs to be mind-

ful that there were a number of periods in the past when

significant losses were incurred from following these strat-

egies. Likewise, some key fundamental variables may have

closely tracked the trend in exchange rates in the past, but

there is no guarantee that they will continue to do so in

the future. If divergent trends begin to set in, fund manag-

ers must decide whether to disregard the trend in those

key fundamental variables or not.

Although many fund managers might prefer to follow a

rigid rule or trading system for formulating FX strategies,

one should not sell short a judgment-based approach to

currency investment management. In the end, successful

FX management is based as much on "art" as it is on "sci-

ence."

Bandwagon Effect/

Trend-Following Behavior

Real

Interest-Rate

Differentials

Purchasing Power Parity

Capital

Flows

Fiscal

Policy Portfolio-Balance

Considerations

Current

Account

Trends

Relative

Economic

Growth

Monetary

Policy

Investor Positioning

FX Options Market

Positioning

Risk Appetite

Investor Sentiment

Net Foreign Assets

Productivity Trends

Savings/Investment

Balance Trends

Exchange-Rate Determination

in the Short, Medium, and Long Run

Persistent Trend in

Terms-of-Trade

Exchange Rate

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

6Deutsche Bank Foreign Exchange Research

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 7

A Stylized Model of Exchange-Rate Overshooting

at the End of a Long-Term Uptrend

Currency’s

Value

Time

Economic Fundamentals

Exchange-Rate Determination

in the Short Run

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

8Deutsche Bank Foreign Exchange Research

Economists have come up with a wide range of theories

to explain how exchange rates are determined. The over-

whelming body of evidence from scores of empirical stud-

ies indicates that fundamental-based models, while use-

ful in explaining long-term trends, have not met much suc-

cess in explaining short-term exchange-rate trends. Indeed,

the evidence suggests that for short-term horizons, a ran-

dom walk characterizes exchange-rate movements better

than most conventional fundamental-based exchange rate

models.

One of the reasons why researchers have not been able

to unearth any significant relationship between changes

in macroeconomic variables and changes in exchange rates

over short periods is that exchange rates often exhibit much

greater variability than do macroeconomic time series in

the short run. The often chaotic behavior of exchange rates

is capable of generating so much noise that it may ob-

scure any discernable relationship between macroeco-

nomic time series and the short-term movement of ex-

change rates.

Bandwagon effects are also capable of causing exchange

rates to wander away from fundamental equilibrium val-

ues in the short run. Survey studies find that FX market

participants tend to have extrapolative expectations over

short-term horizons and mean-reverting or regressive ex-

pectations over longer horizons. If investors have extrapo-

lative expectations over short horizons, they may tend to

accentuate and perpetuate exchange-rate movements in

the short run far beyond the path justified by fundamen-

tals alone. Indeed, if a significant number of investors en-

gaged in extrapolative/trend-following trading strategies,

exchange rates might tend to overshoot on both the up-

side and downside, which could further obscure the rela-

tionship between macroeconomic fundamentals and the

short-term movement of exchange rates.

Because exchange rates can and often do deviate signifi-

cantly from any semblance of fair value in the short run,

investors have looked for alternative forecasting tools to

help them formulate currency investment strategies over

short-term horizons. Short-run forecasting tools that have

attracted interest in recent years include technical-based

trend-following trading rules, sentiment and positioning sur-

veys, FX dealers' customer order flow data, information

embedded in currency option prices, and risk appetite in-

dices.

Investors who concentrate their energies on such tools

presume that the market exhibits a tendency to tip its hand

ahead of time as to which direction it intends to take ex-

change rates in the future. While technical models have

been found to be profitable in the past, most of the other

short-term indicators are relatively new to the FX arena

and thus only a limited time series is available to test their

predictive power. What evidence we do have, however,

suggests that in most cases, these indicators are more

useful as contemporaneous rather than as leading indica-

tors of exchange-rate movements. Nevertheless, they may

prove useful as confirmation indicators that can be used in

conjunction with traditional technical-based trend-follow-

ing trading rules in formulating short-term investment strat-

egies.

Short-Run Forecasting Tools

Short-Run Forecasting Tools: A Checklist

Short-Term Trend

Down

Moving-Average Crossover Trading Rule

Market Sentiment

Speculative Positioning

Order Flow

Option Market Sentiment

Risk Appetite Indices

Neutral Over-

sold

Over-

bought

Momentum

(Consensus Inc. Index of Bullish Opinion)

(Net IMM Contracts)

(DB Customer Order Flow Database)

(Risk Reversals)

Forecasting Tool

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 9

FX market participants typically fall into one of two camps:

(1) Shorter-run technically oriented traders or

(2) Longer-run fundamental-based investors.

Shorter-run technically oriented traders do not base their

investment decisions on fundamental considerations.

Rather, they rely on trend-following trading rules to deter-

mine their position taking: they buy when the currency is

rising, and they sell when the currency is falling. In con-

trast, longer-run fundamental-based investors base their

investment decisions largely on valuation considerations.

If a currency is believed to be mispriced relative to its fair

value, fundamental-based investors would buy those cur-

rencies that are believed to be undervalued and would

sell those currencies that are believed to be overvalued.

Knowing precisely what exchange-rate level represents a

currency's true equilibrium or fair value is not an easy task.

Different exchange-rate models can and often do yield quite

different estimates of a currency's long-run equilibrium

value. In most cases, the marketplace will have only a rough

idea of where a currency's long-run equilibrium value lies.

Because of this, fundamental-based investors will not set

their sights on an imprecise point estimate of fair value,

but rather on an equilibrium range or band. Within this

equilibrium range or band, fundamental-based investors

will presume that the true but unknown equilibrium ex-

change rate, q, lies somewhere between an upper bound,

qU, and a lower bound, qL.

The qU-qL band has been referred to as the "band of agnos-

ticism" in academic writings (see DeGrauwe, 1996). When

exchange rates trade inside the qU-qL band, fundamental-

based investors tend to be agnostic in terms of their cur-

rency positioning, accepting the fact that the actual ex-

change rate is probably trading close to its fair value. Ex-

change-rate movements within the qU-qL band are viewed

as noise and therefore not worthy of serious investment

consideration. Fundamental-based investors would thus

not feel compelled to take on either aggressive long or

short currency positions when exchange rates are trading

inside the band. Instead, they would more likely adopt a

neutral stance toward currency positioning.

When exchange rates fluctuate inside the band of agnosti-

cism, trading tends to be dominated by short-term techni-

cally oriented traders since fundamental-based investors

will refrain from joining the fray until the actual exchange

rate moves outside of the band. When the actual exchange

rate moves outside the qU-qL equilibrium range, fundamen-

tal-based investors will tend to shed their agnosticism and

become more willing to take on aggressive long positions

if the actual exchange rate falls below qL and aggressive

short positions if the actual exchange rate rises above qU.

In times of greater market uncertainty, however, the band

of agnosticism is likely to widen since investor confidence

regarding estimates of fair value is likely to be less strongly

held than in tranquil environments. In such cases, techni-

cally oriented traders would tend to dominate trading ac-

tivity over even wider ranges (q'U-qL'>qU- qL). That, in turn,

would likely lead to even greater FX volatility in the short

run.

One of the key problems for fundamental-based investors

is that even if the exchange rate moved outside of the

band of agnosticism, there is no guarantee that it would

move back inside the band anytime soon. Indeed, funda-

mental-based investors run the risk that an overvalued

currency might get even more overvalued if the exchange

rate moved deeper into overvalued territory, and vice versa.

Since large financial resources are likely to be needed for

investors to position themselves against an overshooting

exchange rate, one has to wonder how many fund manag-

ers would be willing to risk their clients' capital on the

basis of long-run valuation considerations, particularly if

clients evaluate their fund manager’s investment perfor-

mance over a relatively short time span. If fund managers

view it as simply too risky to take on long or short cur-

rency positions on the basis of long-run valuation consid-

erations, then there might be several occasions when ex-

change rates could wander far from any semblance of fair

value, and yet very few investors would be willing to risk

their clients' capital to bring the exchange rate back into

line with fair value.

Why FX Market Participants Focus Their Energies on Short-Run Rather

Than Long-Run Strategies

The "Band of Agnosticism"

Index of Investor Willingness to

Make Currency Bets Based on

"Fundamentals"

Real Exchange Rate

q’

Analyzing the Behavior of Fundamental-Based Investors When

Expectations of Real Long-Run Equilibrium Exchange Rates Are Loosely Held

q’

Source: Adapted from DeGrauwe (1996)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

10 Deutsche Bank Foreign Exchange Research

When FX dealers were asked recently what role funda-

mental factors played in the determination of exchange

rates, 97% of the respondents felt that fundamentals

played no role on an intra-day basis. However, over me-

dium and longer-term horizons, FX dealers felt that funda-

mental forces did play an important role, with 57.4% of

the respondents believing that exchange rates reflect fun-

damental value on a medium-term (within six months)

basis, and 87% believing exchange rates reflect fundamen-

tal value on a long-term (over six months) basis.

When asked to rank the most important determinants of

exchange rates on an intra-day basis, FX dealers indicated

that bandwagon effects, speculative forces, and over-re-

action to news were the principal driving forces in the very

short run. On a medium-term basis, economic fundamen-

tals and technical trading increase in importance from the

dealer community's perspective, but FX dealers also con-

tinued to assign importance to speculative forces as a key

determinant of medium-term trends in exchange rates. For

longer horizons, FX dealers believed that economic funda-

mentals were the dominant factor driving exchange rates.

However, a not insignificant number (11.3%) believed that

technical trading was important even in the long run.

The FX dealer survey also asked dealers whether they

thought that exchange rates were more predictable on an

intra-day basis or on a medium-term (up to six months) or

long-term (beyond six months) basis. FX dealers were

asked to assign a rating of 1 if there was no predictability,

a rating of 5 if there was a high predictability and a rating

of 2, 3, or 4 if there was low, medium, or better than aver-

age predictability, respectively. Since FX dealers tend to

trade on an intra-day basis, one might have thought that

they would assign a high rating to exchange-rate predict-

ability on an intra-day basis. This was not the case. Indeed,

62% of the dealer respondents gave ratings of 1 or 2, to

the predictability of exchange rates in the short run, while

only 11% gave ratings of 4 or 5.

For medium and longer time horizons, the confidence in

exchange-rate predictability increases, with 30.4% of deal-

ers assigning a ranking of 4 or 5 on a medium-term basis

and 35.1% assigning a ranking of 4 or 5 on a longer-term

basis. The question that we need to ask ourselves is: if FX

dealers are more confident in predicting exchange rates

on a medium/long-term basis rather than on an intra-day

basis, why then do traders concentrate their energies on

very short-run trading? The answer might be that traders

are in a better position to evaluate and manage FX risk on

a short-term basis, which overrides their greater confidence

in medium/long-term exchange-rate predictability.

FX Dealers’ Perception of the Role of Fundamentals

in Explaining Exchange-Rate Movements

FX Dealer Survey Question—Do You Believe Exchange-Rate

Movements Reflect Changes in Fundamental Value on an:

Intraday Medium-Run Long-Run

Basis Basis Basis

(up to 6 months) (beyond 6 months)

Yes 3% 57.8% 87%

No 97% 42.2% 12%

No Opinion 0% 0.0% 1%

Source: Yin-Wong Cheung, Menzie D. Chinn, and Ian W. Marsh,

“How Do UK-Based Foreign Exchange Dealers Think Their Market Operates?”,

NBER Working Paper 7524, February 2000.

FX Dealers’ Perception of the Most Important Factor

That Explains Exchange-Rate Movements

FX Dealer Survey Question—Select the Single Most Important

Factor that Determines Exchange Rate Movements on an:

Intraday Medium-Run Long-Run

Basis Basis Basis

(up to 6 months) (beyond 6 months)

Bandwagon Effects 29.3% 9.5% 1.0%

Over-reaction to News 32.8% 0.7% 0.0%

Speculative Forces 25.3% 30.7% 3.1%

Economic Fundamentals 0.6% 31.4% 82.5%

Technical Trading 10.3% 26.3% 11.3%

Other 1.7% 1.5% 2.1%

Source: Yin-Wong Cheung, Menzie D. Chinn, and Ian W. Marsh,

“How Do UK-Based Foreign Exchange Dealers Think Their Market Operates?”,

NBER Working Paper 7524, February 2000.

FX Dealers’ Perception of the Predictability of

Exchange-Rate Movements

FX Dealer Survey Question—On a Scale of 1 to 5, Indicate If You

Believe the Market Trend Is Predictable on an:

Intraday Medium-Run Long-Run

Basis Basis Basis

(up to 6 months) (beyond 6 months)

1 (Least Predictable) 21.6% 5.9% 17.2%

2 40.3% 20.7% 16.4%

3 26.9% 43.0% 30.6%

4 9.0% 18.5% 20.9%

5 (Most Predictable) 2.2% 11.9% 14.2%

Source: Yin-Wong Cheung, Menzie D. Chinn, and Ian W. Marsh,

“How Do UK-Based Foreign Exchange Dealers Think Their Market Operates?”,

NBER Working Paper 7524, February 2000.

How FX Dealers View the Determination of Exchange Rates in the Short Run

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 11

Technical models generate exchange-rate forecasts by

extrapolating past sequences of exchange-rate movements

into the future. For example, if a currency begins to edge

higher and rises above some critical value, a technical

model will typically issue a recommendation to go long

that currency, the presumption being that the newly formed

trend will continue to carry the currency higher in the fu-

ture. Similarly, a sell signal would be issued if the currency

began to edge lower and fell below some critical value.

Trend-following trading rules will be profitable as long as

the ensuing trend does indeed move in the same direc-

tion as the preceding trend, which would be the case if

exchange rates moved in broad, well-defined cycles. But

that is not to say that exchange rates must always move in

large swings for trend-following trading rules to be profit-

Technical Analysis:

The Advantage of Trading

With and Not Against the Trend

Exchange Rate

Time

Long-Run Equilibrium

Exchange Rate

Technical Analysis:

Even though a currency overshoots its

long-run equilibrium level, a technician

will recommend maintaining long positions

as long as the trend in the exchange

rate is up.

Fundamental Analysis:

May prematurely recommend

selling a currency that has

overshot its long-run

equilibrium level.

Trendline

Market Price

Today’s

Spot

Rate

able. What matters for long-term profitability is that large

exchange-rate upswings and downswings must occur on

a frequent enough basis to overcome those occasions

when currency movements are not highly trended.

Fundamental-based models tend to focus on whether a

currency lies above or below its long-run equilibrium or

"fair" value. Technical models, on the other hand, are not

interested in whether a currency lies above or below its

fair value. Rather, technical models are only interested in

whether the trend in the exchange rate is upward or down-

ward. As long as a confirmed uptrend (or downtrend) is in

place, a recommended long (short) position will be main-

tained even if the prevailing trend carries the exchange

rate well above (below) its long-run fair value.

Currency Forecasting Using Technical Analysis—

The Advantage of Trading with and Not Against the Trend

Source: Rosenberg (1996)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

12 Deutsche Bank Foreign Exchange Research

Identification and Confirmation of a Valid Uptrend/Downtrend

Exchange Rate

Time

Trendline

Market Price

Peak

Exchange Rate

Time

Peak

Trough

Trendline

Peak

Trough

Peak

Market Price

Identification and Confirmation

of a Market Reversal

Exchange Rate

Time

Trendline

Market Price

Head-and-Shoulders Reversal Pattern

Exchange Rate

Time

Neckline

Market Price

Head

Left Shoulder

Right Shoulder

Trend-following trading rules come in many forms. But what

all trend-following trading rules have in common is that

they seek to identify the direction in which the

broad

trend

in exchange rates is heading. One can visualize the trend

in exchange rates as a series of primary and secondary

waves. The primary wave refers to the large, broad moves

in exchange rates that carry the underlying trend in cur-

rency values either upward or downward. The secondary

waves refer to the temporary corrections or partial

retracements of the primary trend that typically take place

over the course of longer-run exchange-rate cycles. What

trend-following trading rules attempt to do is to identify

the direction in which the primary waves are heading.

In a rising market, each rally and partial retracement will

be higher than its predecessor. As long as advancing prices

achieve successively higher peaks and troughs, they indi-

cate that buying pressure is overcoming selling pressure.

In such a case, the uptrend will be presumed to be intact,

until a reversal is signaled. The opposite would be the case

in a declining market.

Markets reverse in many different ways, but all reversals

have one thing in common—all reversals of valid uptrends

(downtrends) must be preceded by the failure of market

prices to achieve successively higher peaks and troughs

(lower troughs and peaks). In the case of a reversal of a

previous valid uptrend, when the wavelike series of rising

peaks and troughs is broken, it indicates that selling pres-

sure is finally beginning to overcome buying pressure.

One of the most recognizable reversal patterns is the head-

and-shoulders pattern. A head-and-shoulders pattern is

essentially a series of three successive rallies with the

second stronger than the first, and the third weaker than

the second. Because the third rally fails to carry as far as

the second, the string of successively higher peaks is bro-

ken. This is an initial indication of weak demand, and the

eventual drop in market prices to levels below the preced-

ing trough (or neckline) is confirmation that a reversal is

taking place. A recent Federal Reserve Bank of New York

study (Osler and Chang, 1995) found that head-and-shoul-

ders patterns have been successful in anticipating rever-

sals in trend for the yen and Deutschemark.

The Identification of FX Market Trends and Reversals

Source: Rosenberg (1996)

Source: Rosenberg (1996) Source: Rosenberg (1996)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 13

A wide variety of trend-following trading rules abounds,

but by and large they all share the basic property of ex-

trapolation. A chartist who monitors the behavior of mar-

ket prices on a bar chart with a ruler and pencil will as-

sume that an uptrend is intact as long as advancing prices

achieve successively higher peaks and intervening declines

fail to fall below preceding troughs.

More sophisticated extrapolation techniques can be de-

signed with the assistance of a computer. One popular

computer-based technical model is the filter rule, which

issues a buy recommendation if an exchange rate rises by

above its most recent trough, and issues a sell recom-

mendation if the exchange rate falls by

below its most

recent peak. Another popular computer-based technical

model is the moving-average crossover model. By con-

structing longer-run moving averages of daily exchange-

rate movements, one can more easily isolate the primary

trend from short-run noise. If the short-run moving aver-

age of the exchange rate rises above its long-run moving

average, it is an indication that buying pressure is over-

coming selling pressure, and vice versa.

In each case, predictions of the likely future path that ex-

change rates might take are being generated by the ex-

trapolation of the prevailing trend in exchange rates into

the future. Essentially, it really doesn't matter which trend-

following rule you use. Since a trend is a trend, all trend-

following trading rules should roughly generate the same

directional forecast.

Because all of these models generate forecasts by extrapo-

lating the recent past trend of exchange rates into the fu-

ture, buy or sell signals will be issued only after a currency

has already started rising or falling. Although this means

that trend-following trading rules will not capture the very

top and bottom of market moves, they may nevertheless

be profitable if the ensuing exchange-rate movements per-

sist long enough and carry far enough to generate signifi-

cant excess returns.

Filter-Rule Strategy

DM/US$

Time

Moving-Average Crossover Strategy

DM/US$

Time

SRMA

LRMA

All Trend-Following Trading Rules Share

the Same Property: Extrapolation

DM/US$

Time

Down Trendline

Up Trendline

Trendline Penetration

Trend-Following Trading Rules and the Principle of Extrapolation

Source: Rosenberg (1996)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

14 Deutsche Bank Foreign Exchange Research

Investors Who Follow a Trend-Following Trading Rule

May Be Vulnerable to Whipsaws Caused by False Signals

DM/US$

Time

DM/US$

An Example of a False Signal

and a Whipsaw Loss

Time

Market-Swing

Requirement

Market-Swing Requirement

of Profitable Trend-Following Trading Rules

When exchange-rate swings are sustained and pro-

nounced, the profitability of most technical trading rules

will not be seriously undermined if the trading rule tends

to be a bit late in drawing attention to a shift in market

direction. Unfortunately, this market-swing requirement

could prove in many cases to be a tall order. In markets

that exhibit little overall price variation or are not highly

trended, an investor who adheres to a trend-following trad-

ing rule will find his portfolio vulnerable to potential whip-

saw losses caused by frequent false signals.

Consider the two exchange rate series depicted below.

The series on the left illustrates the working of a profitable

filter rule trading strategy. A recommendation to buy

dollars is issued at point

when the DM/US$ exchange

rate rises

from its recent trough at point

. The ensuing

market action then carries the DM/US$ exchange rate to a

peak at point

, but a recommendation to sell dollars is

issued only at point

, when the exchange rate has fallen

from its peak level. An investor who rigidly adhered to

this

filter rule trading strategy would have bought dol-

lars at

and sold dollars at

, with the spread between

and

representing the profit margin per dollar invested.

The series on the right illustrates how losses can be in-

curred using a filter rule trading strategy. The series on the

right behaves in a similar fashion as the series on the left,

except for the fact that the dollar's upswing is less pro-

nounced following the recommendation to buy dollars at

point

1. That is, the swing in the DM/US$ exchange rate

from

1 to

1 falls short of the swing from

in the

chart on the left. At

1, the dollar's upward momentum is

shown to lose steam far earlier than was the case in the

series on the left. Indeed, the DM/US$ actually begins to

lose ground soon after its initial rise and a sell signal is

eventually issued at point

1. Since the selling price (

lies below the purchase price (

1), a loss is incurred with

the spread between

1 and

1 representing the margin of

loss per dollar invested. Hence, the buy recommendation

1 proved to be a false buy signal, with the investor

whipsawed in the process.

Potential Pitfalls from Following Technical Trading Rules

Source: Rosenberg (1996)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 15

Studies on the long-run profitability of technical-based trad-

ing rules often find that simple moving-average rules would

have generated significant excess returns over relatively

long periods for most major currency pairs. We simulated

more than 2500 sets of moving-average crossover trading

rules for seven currencies versus the U.S. dollar—the Deut-

schemark, Japanese yen, Swiss franc, British pound, and

the dollar-bloc currencies—over the 1986-2002 period, and

unearthed the optimal moving-average trading rules listed

in the table below. The criterion used to choose the opti-

mal trading rule in each case was to identify the rule that

yielded the highest Sharpe ratio (after adjusting for trans-

action costs) for the entire 16-year period.

For each exchange rate, there were a number of moving-

average crossover trading rules that yielded similar attrac-

tive Sharpe ratios, and we list the top 10 trading rules for

each exchange rate. In each case, there was always one

rule that eked out the highest Sharpe ratio and those are

shown in the column on the left. (For example, the best

performing trading rule for the Deutschemark/U.S. dollar

exchange rate was a 1-day/32-day moving-average cross-

over trading rule, which yielded an average annual total

return of 5.01% with a Sharpe ratio of 0.45.) The moving-

average crossover trading rules are ranked for each ex-

change rate and are listed from left to right in the table

below.

The 10 Best Moving-Average Crossover Trading Rules for 1986-2002

(Average Annual Total Returns and Sharpe Ratios for January 1986-April 2002 )

US$ Exchange Rate #1 #2 #3 #4 #5 #6 #7 #8 #9 #10

DEM Mov. Avg. Days 1/32 1/33 1/31 1/28 1/30 1/29 1/34 1/20 1/35 1/19

Total Return (%) 5.01 4.89 4.38 4.20 3.96 3.88 3.98 3.58 3.43 3.13

Std. Dev. (%) 11.06 11.06 11.06 11.06 11.06 11.06 11.06 11.06 11.06 11.06

Sharpe Ratio 0.45 0.44 0.40 0.38 0.36 0.35 0.36 0.32 0.31 0.28

JPY Mov. Avg. Days 8/59 8/60 7/61 8/61 7/60 9/60 9/62 9/61 8/62 8/63

Total Return (%) 9.19 9.16 8.83 8.86 8.76 8.79 8.63 8.63 8.53 8.49

Std. Dev. (%) 11.75 11.75 11.75 11.75 11.75 11.75 11.75 11.75 11.75 11.75

Sharpe Ratio 0.78 0.78 0.75 0.75 0.75 0.75 0.73 0.73 0.73 0.72

GBP Mov. Avg. Days 1/19 1/18 1/20 1/21 1/22 1/16 1/24 1/15 1/17 1/23

Total Return (%) 5.80 5.45 5.36 5.32 5.28 5.20 4.91 4.82 4.80 4.49

Std. Dev. (%) 9.46 9.47 9.46 9.46 9.46 9.47 9.45 9.47 9.47 9.45

Sharpe Ratio 0.61 0.58 0.57 0.56 0.56 0.55 0.52 0.51 0.51 0.48

CHF Mov. Avg. Days 1/57 1/59 1/58 1/56 1/60 1/55 1/69 1/54 1/61 1/65

Total Return (%) 8.63 8.44 8.40 8.40 7.87 7.79 7.50 7.24 7.21 6.84

Std. Dev. (%) 11.65 11.64 11.66 11.65 11.63 11.66 11.63 11.66 11.63 11.64

Sharpe Ratio 0.74 0.72 0.72 0.72 0.68 0.67 0.65 0.62 0.62 0.59

CAD Mov. Avg. Days 14/199 14/200 14/197 13/197 14/198 13/198 15/195 15/193 15/194 14/196

Total Return (%) 1.88 1.88 1.86 1.83 1.82 1.81 1.75 1.66 1.64 1.59

Std. Dev. (%) 4.81 4.81 4.81 4.81 4.81 4.81 4.81 4.81 4.81 4.81

Sharpe Ratio 0.39 0.39 0.39 0.38 0.38 0.38 0.36 0.34 0.34 0.33

AUD Mov. Avg. Days 13/39 13/38 13/40 1/16 13/42 13/41 14/43 13/37 13/43 13/45

Total Return (%) 3.53 3.44 3.41 2.99 2.73 2.54 2.45 2.44 2.37 2.34

Std. Dev. (%) 10.30 10.31 10.28 10.29 10.28 10.28 10.28 10.31 10.28 10.28

Sharpe Ratio 0.34 0.33 0.33 0.29 0.27 0.25 0.24 0.24 0.23 0.23

NZD Mov. Avg. Days 10/17 11/15 10/18 10/15 11/17 14/196 10/16 11/20 11/16 11/14

Total Return (%) 5.20 4.99 5.10 4.76 4.67 4.28 4.28 4.41 3.85 3.38

Std. Dev. (%) 10.79 10.80 10.78 10.80 10.79 10.14 10.80 10.78 10.80 10.81

Sharpe Ratio 0.48 0.46 0.47 0.44 0.43 0.42 0.40 0.41 0.36 0.31

Note: A Sharpe Ratio measures the amount of return on an investment (less the return of a risk-free asset) per unit of risk, which is proxied by its standard deviation.

Datastream is the source of the underlying exchange-rate data.

Empirical Evidence on the Profitability of Moving-Average Trading Rules

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

16 Deutsche Bank Foreign Exchange Research

Given the Sharpe ratios shown in the table on the preced-

ing page, one might wonder why more investors do not

rely on trend-following trading rules more regularly. One

reason is that investors might have a fairly high Sharpe

ratio threshold for committing capital to any particular trad-

ing strategy. For instance, many investors might not em-

brace a trading strategy rule unless it generated a Sharpe

ratio of at least 0.50. And in some cases, an even higher

Sharpe ratio might be required for an investor to add sig-

nificant risk capital to a particular trading rule. If that were

the case, only the yen and Swiss franc would qualify as

currencies worthy of trading from a technical perspective.

While the estimated Sharpe ratios may have been high on

average for the 1986-2002 period, one should note that

the Sharpe ratios were highly volatile when viewed on a

one-year rolling basis. As the charts below show, there

were periods when risk-adjusted return performances were

high and other periods when risk-adjusted return perfor-

mances were downright poor. This was clearly the case

for several key exchange rates in 2001, with one-year roll-

ing Sharpe ratios falling into negative territory for the Deut-

schemark, British pound, and Swiss franc.

Except for the yen and the dollar-bloc currencies, total re-

turn performances were not that impressive in 2001. In

most cases, currencies were largely range-bound versus

the U.S. dollar in 2001 and therefore generated a consider-

able number of false signals that yielded small but fre-

quent losses. On the other hand, the yen’s downtrend—

particularly in late 2001—was clearly evident and exploit-

able by following a moving-average crossover trading rule.

Deutschemark Moving-Average Crossover

Trading Rule Risk-Adjusted Performance

(1993-2002)

-3

-2

-1

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

(One-Year Rolling Sharpe Ratio)

Japanese Yen Moving-Average Crossover

Trading Rule Risk-Adjusted Performance

(1993-2002)

-2

-1

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

(One-Year Rolling Sharpe Ratio)

British Pound Moving-Average Crossover

Trading Rule Risk-Adjusted Performance

(1993-2002)

-3

-2

-1

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

(One-Year Rolling Sharpe Ratio)

Swiss Franc Moving-Average Crossover

Trading Rule Risk-Adjusted Performance

(1993-2002)

-2

-1

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

(One-Year Rolling Sharpe Ratio)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 17

Most academic studies as well as our own work docu-

ment the fact that moving-average trading rules have

tended to generate attractive risk-adjusted returns over the

long run. What is not often pointed out in these studies is

the frequency with which winning and losing positions tend

to occur. As the table below illustrates, moving-average

trading rules have tended to generate far more losing than

winning trades on average. For example, in the case of

the Deutschemark, Swiss franc, cable, and the A$, roughly

75% of the recommended trades generated by optimized

moving-average trading rules end up as losing trades. Only

25% of the trades proved to be winning ones. In the case

of the C$ and NZ$, roughly two-thirds of the recommended

trades resulted in losses. Only in the case of the yen was

the win/loss ratio close to 50%.

From a longer-run perspective, having a trading rule that

generates such frequent losing trades does not pose a

serious problem if the losing positions are cut quickly—so

only a small loss is taken—and profits are allowed to ride

on the less frequent correctly predicted winning trades.

Indeed, this is often what is found to be the case. As shown

in the table below, the average profit on the less frequent

winning trades has tended to exceed the average loss on

the more frequent losing trades by a fairly hefty margin.

This margin of difference has been sufficient to allow mov-

ing-average trading rules to be profitable in the long run.

Frequency of Winning and Losing Trades Generated by Moving-Average

Trading Rules

Profit and Losses from Optimal Moving-Average Trading Rules

(Risk-Return Analysis & Winning versus Losing Trades of Selected Currencies versus the U.S. Dollar)

(January 1986-April 2002)

DEM JPY GBP CAD AUD NZD CHF

Optimal Moving-Average Trading Rule

Number of Days (SRMA/LRMA) 1/32 8/59 1/19 14/199 1/16 10/17 1/57

Average Annual Return 5.0% 9.2% 5.8% 1.9% 3.5% 5.2% 8.6%

Standard Deviation of Returns 11.1% 11.8% 9.5% 4.8% 10.3% 10.8% 11.7%

Sharpe Ratio 0.45 0.78 0.61 0.40 0.34 0.48 0.74

Total Recommended Trades 342 81 452 35 218 165 217

Winning Trades 91 37 126 13 61 58 57

Winning Trade Percentage 27% 46% 28% 37% 28% 35% 26%

Losing Trades 251 44 326 22 157 107 160

Losing Trade Percentage 73% 54% 72% 63% 72% 65% 74%

Average Profit on Winning Trades 2.97% 5.92% 2.27% 3.69% 2.03% 2.11% 4.63%

Average Profit on Losing Trades -0.77% -1.72% -0.62% -1.03% -0.63% -0.84% -0.84%

Ratio of Profits/Losses 3.86 3.44 3.66 3.58 3.22 2.51 5.51

Note: A Sharpe Ratio measures the amount of return on an investment (less the return of a risk-free asset) per unit of risk,

which is proxied by its standard deviation.

Datastream is the source of the underlying exchange-rate data.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

18 Deutsche Bank Foreign Exchange Research

The charts below reveal the distribution of returns that each

of our optimal moving-average crossover trading rules has

generated over the past 16 years. As shown, the over-

whelming majority of recommended trades resulted in ei-

ther small gains or small losses. However, note that the

distribution of returns is skewed heavily to the right in most

cases. This indicates that large positive returns do occur

from time to time; it is just that the frequency of big win-

ning trades tends to be quite low. These highly skewed

return distributions raise the question whether conventional

measures of risk such as standard deviation, Sharpe ra-

tios, and information ratios accurately convey the asym-

metric risks facing technically oriented investors, particu-

larly over short horizons.

Distribution of Returns from Moving-Average Trading Rules

Distribution of Returns of DEM Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

100

150

200

-2.9

-2.0

-1.1

-0.20.71.6 2.5 3.4 4.35.26.1 7.0 7.9 8.89.7

10.6

11.5

12.4

13.1

Frequency

Return(%)

Distribution of Returns of JPY Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

-5.4 -2.1 1.3 4.6 8.0 11.4 14.7 18.1 21.4 24.5

Frequency

Return(%)

Distribution of Returns of GBP Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

100

150

200

250

300

-4.1

-2.8

-1.4

-0.1

1.2

2.5

3.8

5.2

6.5

7.8

9.1

10.4

11.8

13.1

14.4

15.7

17.0

18.3

19.7

21.0

22.3

23.6

Frequency

Return(%)

Distribution of Returns of CAD Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

-2.46 0.81 4.07 7.34 10.6 13.6

Frequency

Return(%)

Distribution of Returns of AUD Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

-2.4 -1.7 -1.0 -0.4 0.3 1.0 1.7 2.3 3.0 3.7 4.4 5.1 5.7 6.4 6.8

Frequency

Return(%)

Distribution of Returns of CHF Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

100

120

140

-5.73

-4.13

-2.52

-0.91 0.69 2.3 3.9 5.51 7.12 8.72

10.33

11.93

13.54

15.15 16.2

Frequency

Return(%)

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 19

The charts below provide a different way of looking at the

short-term risks facing technically oriented investors in the

FX arena. These charts show the actual gains and losses

on the recommended trades that our optimal moving-av-

erage trading rules have generated in recent years. What

these charts reveal is that there were often frequent runs

of successive losing trades that had to be absorbed be-

fore a sizeable winning trade was generated. While posi-

tive returns were generated in most cases for the entire

sample period, clearly there were uncomfortably lengthy

intervals when frequent losses were incurred.

Our analysis raises several interesting issues. First, should

technical-based moving-average trading rules be viewed

as short-term forecasting tools or instead would it be bet-

ter to view such trading rules as long-term forecasting

tools? The results here suggest that technical-based mov-

ing-average trading rules work best in the long run, not in

the short run. Indeed, the odds that a moving-average trad-

ing rule recommendation will generate a profit appear to

be no better than 25%-35% in most cases in the short

run. Second, it is possible that since the frequency of los-

ing trades is so large, this might actually dissuade inves-

tors from using technical models in formulating currency

investment strategies. If investors shy away from using

technical models because of the high frequency of short-

term losses, this might explain why excess returns in the

long run from technical-based models are not completely

arbitraged away.

In order to successfully trade currencies using a technical-

based moving-average trading rule, an investor needs to

have staying power—considerable risk capital on hand to

absorb possible frequent short-term trading losses—and

patience. There might be other technical tools that could

be used to cut down on the number of frequent losses on

short-term trading positions, but one needs to be mindful

of the fact that attempts to add "filters" or other types of

technical bells and whistles to limit the frequency of false

signals might hinder the upside potential of correctly pre-

dicted winning trades.

Should Moving-Average Trading Rules Be Viewed as Short or Long-Run

Forecasting Tools?

Profit and Losses on Individual DEM Trades

(Generated by Moving-Average Crossover Strategy)

(Most Recent 140 Trades)

-4

-2

12-27-95 04-21-97 03-27-98 01-14-99 11-08-2000 10-02-2001

Returns(%)

Trade Dates

Profit and Losses on Individual JPY Trades

(Generated by Moving-Average Crossover Strategy)

(1986-2002)

-10

-5

01-01-87 12-27-89 02-02-94 01-23-98 03-07-2002

Returns(%)

Trade Dates

Profit and Losses on Individual GBP Trades

(Generated by Moving-Average Crossover Strategy)

(Most Recent 140 Trades)

-2

-1

06-12-97 06-01-98 01-27-99 01-03-2000 01-12-2001 09-24-2001

Returns(%)

Trade Dates

Profit and Losses on Individual CAD Trades

(Generaged by Moving-Average Crossover Strategy)

(1986-2002)

-5

11-03-86 06-23-89 11-02-90 09-16-94 04-25-96 01-14-97 08-06-99

Returns(%)

Trade Dates

Datastream is the source of the

underlying exchange-rate data.

Datastream is the source of the

underlying exchange-rate data.

Datastream is the source of the

underlying exchange-rate data.

Datastream is the source of the

underlying exchange-rate data.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

20 Deutsche Bank Foreign Exchange Research

Most academic studies have concluded that exchange-rate

movements closely approximate a random walk process.

A random walk (serial correlation) test seeks to investi-

gate whether there exists a positive, negative, or zero lin-

ear relationship between today’s change in the exchange

rate and yesterday’s change in the exchange rate. Evidence

of a positive linear relationship would indicate the exist-

ence of trend persistence since a positive change in a

currency's value yesterday would tend to be followed by a

positive change today.

Estimated serial correlation coefficients could vary between

+1, 0, and -1 depending on whether there exists a strong

positive relationship, no relationship, or a strong negative

relationship between successive exchange-rate changes.

The weight of evidence generally supports the view that

for most currencies, the estimated serial correlation coef-

ficients are often quite small and in many cases statisti-

cally insignificant from zero.

Although researchers often find support for the view that

exchange-rate movements tend to fluctuate randomly on

a daily basis, they also find evidence that exchange-rate

changes are positively serially correlated when viewed on

a monthly basis. Thus, although exchange rates may fluc-

tuate randomly over very short time spans (i.e., daily), they

tend to rise and fall on a trend basis on a medium/long-

term (i.e., monthly) basis. If so, this would imply that trend-

following trading rules could be devised to profit from these

medium/long-term trends that exchange rates follow.

One of the problems with serial correlation tests is that

they seek only to determine whether a stable "linear" rela-

tionship exists between successive exchange-rate move-

ments. Although it might be the case that successive ex-

change-rate changes are linearly independent, they might

nevertheless exhibit significant positive nonlinear depen-

dence, which traditional serial correlation tests would not

detect. This might explain why researchers have found that

exchange rates follow a random walk, yet at the same time

have found that trend-following technical trading rules have

been profitable. It may be the case that there does exist

some form of serial dependency in successive exchange-

rate movements; it's just that this serial dependence is

not linear. If there exists some form of nonlinear depen-

dence in successive exchange-rate movements, it would

appear that trend-following trading rules are capturing this

form of serial dependence.

Random Walk Tests and the Profitability of Technical Trading Rules

(a)

Positive Serial Correlation

Serial Correlation Tests:

Examining Whether a Stable Linear Relationship Exists between

Current and Previous Changes in an Exchange Rate

(b)

Zero Serial Correlation (c)

Negative Serial Correlation

∆

t-1

∆

t-1

∆

t-1

Source: Adapted from Fogler (1978)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 21

Fund managers have turned to sentiment and positioning

surveys in recent years as forecasting tools. In most cases,

exchange rates move in the same direction that sentiment

and positioning surveys are moving in the short run. For

example, when sentiment toward the dollar becomes in-

creasingly bullish—as reported in the weekly Consensus

Inc. Survey of Bullish Market Opinion—the dollar tends to

rise. Similarly, when long U.S. dollar speculative positions

are undertaken on the IMM (money-market futures ex-

change), the dollar tends to rise as well.

What does the empirical evidence say about the success

of sentiment/positioning surveys in predicting exchange-

rate movements? Our analysis, listed in the table below,

finds that changes in sentiment and positioning surveys

are statistically significant, for the most part, in explaining

contemporaneous changes in exchange rates, but that

lagged values of sentiment and positioning survey data

are statistically insignificant in predicting current changes

in exchange rates. This suggests that sentiment/position-

ing survey data cannot be relied upon as forecasting tools.

What does the empirical evidence say about the useful-

ness of sentiment/positioning survey data as contrarian

indicators? The trends in market sentiment and specula-

tive positioning could be viewed as potential overbought

or oversold indicators if the rise or fall in each of those

series appeared stretched relative to historical norms.

Unfortunately, the empirical evidence suggests that there

is no statistically significant relationship between over-

stretched sentiment and positioning surveys and subse-

quent changes in exchange rates. Based on these find-

ings, it could prove risky to adopt a contrarian position sim-

ply because investors' FX exposures appeared to be at

extreme levels. History is replete with examples where

over-stretched markets have stayed over-stretched for a

considerable time. Perhaps all that one can say is that if

one or more of those indicators moved into significantly

overbought or oversold territory, then the

odds

of a rever-

sal would have increased, but that it would not be pos-

sible to issue a definitive signal that an imminent reversal

was at hand.

Although trends in sentiment and positioning surveys may

not help us in forecasting future movements in exchange

rates, they may nevertheless be useful as trend confirma-

tion indicators. Because of the strong positive contempo-

raneous correlation between the sentiment and position-

ing data, on the one hand, and the trend in exchange rates

on the other, sentiment and positioning data could be used

in conjunction with a moving-average model to confirm

whether an exchange-rate uptrend or downtrend is intact.

Market Sentiment and the U.S. Dollar

(Consensus Inc. Index of Bullish Opinion)

(1998-1999)

100

102

Jan-98 Apr-98 Jul-98 Oct-98 Jan-99 Apr-99 Jul-99 Oct-99

US$ Index----

nalysts Bullish on the US$(4-Week Avg.)(%)____

Sources: Datastream;

Based on Consenus Inc. Index of Bullish Market Opinion

Market Sentiment and the U.S. Dollar

(Consensus Inc. Index of Bullish Opinion)

(April 2000-November 2001)

100

102

104

106

108

Apr-00 Jul-00 Oct-00 Jan-01 Apr-01 Jul-01 Oct-01

US$ Index----

nalysts Bullish on the US$(4-Week Avg.)(%)____

Based on Consenus Inc. Index of Bullish Market Opinion

By permission of Consensus, Inc., (1) (816) 373-3700,

Consensus, National Futures and Financial Weekly,

www.consensus-inc.com

Speculative Positioning and the U.S. Dollar

(IMM Derived Net Non-Commercial FX Positions)

(April 2000-November 2001)

100

102

104

106

108

-150

-100

-50

100

Apr-00 Jul-00 Oct-00 Jan-01 Apr-01 Jul-01 Oct-01

US$ Index----

um of Net Contracts vs. USD(000s)____

Assessing the Statistical Significance of

Sentiment/Positioning/Flow Variables as

Explanatory Variables of Exchange-Rate Movements

(January 15, 1999-November 2, 2001 Weekly Data)

Lagged

Contemporaneous One Period

Consensus Inc. Bullish Market Opinion (US$)

Regression Coefficient 0.12 0.09

(t-Statistic) (3.90) * (1.15)

IMM Net Non-Commercial FX Positions (US$)

Regression Coefficient 0.23 0.01

(t-Statistic) (5.65) * (0.24)

*Indicates a statistically significant value at the 95% confidence level.

By permission of Consensus Inc.

Are Sentiment and Positioning Indicators Useful in Predicting FX Movements?

Source: IMM, Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

22 Deutsche Bank Foreign Exchange Research

Investors often try to infer the market's expectation of fu-

ture interest-rate movements by examining interest-rate

futures contracts for various maturities or by examining

the implied future path in interest rates embedded in yield-

curve slopes. Similarly, investors often try to infer the

market's expectation of future exchange-rate movements

by examining the level of domestic/foreign interest-rate

spreads, or the implied path of long-dated forward ex-

change rates.

Information on future interest-rate and exchange-rate move-

ments can also be gleaned from the options market.

Whereas interest-rate futures, yield-curve slopes, and yield

spreads provide point estimates regarding expected fu-

ture values, the options market provides a different piece

of information—the market's expectation of the probabil-

ity distribution of future interest rates and exchange rates.

For example, in the FX market, data on implied volatility

for a range of maturities and strike prices are readily avail-

able. Implied volatility provides us with a measure of the

marketplace's uncertainty regarding future exchange-rate

movements. If implied volatility is rising, it "implies" greater

uncertainty about future exchange-rate movements, and

vice versa.

One could construct a forward implied volatility curve to

glean expected future volatility movements by stringing

together implied volatilities for a range of maturities with

the same strike price. If the forward implied volatility curve

were steeply upward sloping, the market would be pricing

in the expectation of a future jump in currency volatility. If

the forward implied volatility curve were flat, the market-

place would be pricing in no change in the level of cur-

rency volatility in the future.

Does the forward implied volatility curve accurately antici-

pate future jumps in exchange-rate volatility? Because vola-

tility jumps often occur suddenly and without warning, it

is highly unlikely that the forward implied volatility curve

would be able to consistently forecast future jumps in the

exchange rate. Indeed, the Bank of England examined

whether implied forward volatility curves anticipated the

unwinding of the infamous yen carry (long dollar/short yen)

trade in October 1998, and concluded that "Forward vola-

tility curves…. did not expect the increase in volatilities

(that occurred) in October 1998. Although (historical) vola-

tility had increased throughout the summer, the forward

volatility curve suggested that it would fall back towards

previous levels." Indeed, in mid-September 1998, histori-

cal volatility was running around 22%-26% and forward

volatility curves were expecting implied volatilities to ease

toward 14% in the October-November 1998 period. Instead,

historical volatility soared to over 40% in early October

when the yen carry trade was suddenly unwound.

Can One Extract Information from the Currency Options Market

to Predict Future Exchange-Rate Movements?

Japanese Yen/U.S. Dollar Exchange Rate

(1997-1998)

110

120

130

140

150

Jan-97 Apr-97 Jul-97 Oct-97 Jan-98 Apr-98 Jul-98 Oct-98

Yen/US$

Japanese Yen/U.S. Dollar Implied Volatility

(1997-1998)

Jan-97 Apr-97 Jul-97 Oct-97 Jan-98 Apr-98 Jul-98 Oct-98

1-Month Volatility----(%) 12-Month Volatility____

Source: Datastream

Source: DRI

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 23

Assessing the Statistical Significance of

Sentiment/Positioning/Flow Variables as

Explanatory Variables of Exchange-Rate Movements

(January 15, 1999-November 2, 2001 Weekly Data)

Lagged

Contemporaneous One Period

Option Market Sentiment (Euro Risk Reversals)

Regression Coefficient -0.02 0.03

(t-Statistic) (-0.90) (1.23)

Option Market Sentiment (Yen Risk Reversals)

Regression Coefficient 0.11 0.04

(t-Statistic) (3.06) * (1.19)

*Indicates a statistically significant value at the 95% confidence level.

Source: Datastream; DB

Euro Risk Reversals & US$/Euro Exchange Rate

Euro Calls(+)/Puts(-) Trading at Premium vs. USD

0.82

0.84

0.86

0.88

0.90

0.92

0.94

0.96

0.98

-1.0

-0.5

0.0

0.5

1.0

1.5

Jul-00 Oct-00 Jan-01 Apr-01 Jul-01 Oct-01 Jan-02 Apr-02

US$/Euro----

alls(+)/Puts(-) Trading at Premium____

Yen Risk Reversals & Yen/US$ Exchange Rate

Yen Calls(+)/Puts(-) Trading at Premium vs. USD

(140)

(135)

(130)

(125)

(120)

(115)

(110)

(105)

(100)

-2.0

-1.0

0.0

1.0

2.0

3.0

Jul-00 Oct-00 Jan-01 Apr-01 Jul-01 Oct-01 Jan-02 Apr-02

Yen/US$(reverse scale)----

alls(+)/Puts(-) Trading at Premium____

FX traders often use risk reversals to glean information on

whether the market might be attaching a higher probabil-

ity to a large currency appreciation than to a large currency

depreciation, or vice versa. A risk reversal is a currency

option position that consists of the purchase of an out-of-

the-money (25

delta

) call and the simultaneous sale of an

out-of-the-money (25

delta

) put, both in equal amounts and

both with the same expiration date. Risk reversals are

quoted in terms of the implied volatility spread between

the 25

delta

call and 25

delta

put. For example, if the im-

plied volatility on the call were 2% larger than the implied

volatility on the put, the risk reversal would be quoted at

+2.0%. If the implied volatility on the put were 2% greater

than the implied volatility on the call, the risk reversal would

be quoted at -2%.

A risk reversal quoted at +2% would indicate that the

market was attaching a higher probability to a large cur-

rency appreciation than to a large currency depreciation.

This would indicate that the market was willing to pay more

to insure against the risk that the currency will rise sharply

than it was willing to pay to insure against the risk that the

currency will fall sharply.

The key issue for traders and investors is whether the level

or trend in currency risk reversals can be used to correctly

anticipate future exchange-rate movements. The evidence

indicates a high contemporaneous correlation between the

trend in risk reversals and the trend in exchange rates, but

no statistically significant relationship exists between

lagged risk reversal data and future exchange-rate move-

ments. Therefore, risk reversals are capable of confirming

an exchange rate’s trend, but not predicting it.

Nor is there evidence that overly stretched risk reversal

measures can function as a consistently reliable contrary

indicator. Indeed, the Bank of England's study on the un-

winding of the yen carry trade in the fall of 1998 found that

dollar/yen risk reversals failed to provide an early warning

of the dramatic unwinding of long dollar/short yen posi-

tions that was about to occur.

Inferring Information about Future Exchange-Rate Movements

from Currency Risk Reversals

Yen/US$

Profit

Payoff Schedule of 25 Delta

Risk Reversal at Maturity

(-)

(+)

Source: Datastream; DB Source: Datastream; DB

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

24 Deutsche Bank Foreign Exchange Research

In recent years, there has been increased interest on the

part of investors and academicians in FX-dealer customer-

flow data. One of the characteristics that distinguishes the

FX market from the world equity market is that the FX

market has considerably less transparency. Equity-market

disclosure requirements mandate that all trades are in-

stantly posted. Thus, volume and price data are instantly

available to all parties. Not so in the FX market. No such

disclosure requirements exist, which means that order-flow

information is not immediately available to all parties.

Indeed, large FX dealers are in a unique position in that

they are able to observe large trades that could move the

market before other parties are aware of such trades. A

recent survey of FX dealers indicated that most traders

believe that large FX dealers have a comparative advan-

tage over smaller dealers because they have a large cus-

tomer base that gives them privileged information about

customer orders.

One could envisage a marketplace where there exists both

private and publicly available information. Publicly available

information consists largely of macroeconomic-related

data, which is available to all market participants at the

same time. However, there may be microeconomic-related

information that is important for exchange rates, but is not

publicly available. Examples of microeconomic-related in-

formation are shifts in risk appetite, liquidity needs, hedg-

ing demands, and institutional portfolio rebalancings. Such

microeconomic-related FX flows give rise to buy or sell

orders that influence the short-term trend in exchange

rates, just as investor responses to macroeconomic-related

information do. Large FX dealers can use the information

gleaned from the change in their customer order flow to

drive their own short-run strategies, which can then add to

the upward or downward pressure on the exchange rate's

value.

Viewing the impact of order flow in this manner, it is evi-

dent that order flow serves as a proximate determinant

and not as the ultimate determinant of short-term ex-

change-rate movements. The ultimate determinant of ex-

change rates is the joint interplay of macroeconomic and

microeconomic information underlying the change in or-

der flow.

Given the potential impact that order flow can have on

exchange rates in the short run, a large FX dealer's order

book can provide valuable information to both traders and

investors who wish to keep abreast of underlying private

investor shifts in portfolio behavior, whether those shifts

are of a macro or micro-economic nature.

Information

Not Publicly

Available

Portfolio

Shifts Order

Flow

Change

Exchange Rate

The Effect of Information on Order Flow

and Exchange Rates

Publicly

Available

Information

Reasons for Competitive Advantage

of Large FX Players

(Based on an FX Dealer Survey)

Survey Response (%)

Large Customer Base 33.3

Better Information 22.9

Deal in Large Volumes 14.8

Ability to Affect Exchange Rate 9.4

Offer New FX Products 6.2

Access to Global Trading Network 4.7

Experienced Trades 4.2

Lower Costs 2.9

Smaller Counterparty Risk 0.5

Other 1.0

Source: Cheung and Chinn, “Traders, Market Microstructures and

Exchange Rate Dynamics”, unpublished paper.

FX Dealer Order Flow and the Determination of Exchange Rates

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 25

How FX Dealers and Fund Managers View the Importance of

Flow Analysis Relative to Fundamental and Technical Analysis

According to a recent survey of professional FX dealers

and fund managers, both FX dealers and fund managers

treat the analysis of foreign-exchange flows as a distinct

form of analysis, independent from more traditional fun-

damental and technical approaches. A recent survey by

Gehrig and Menkhoff (2002) indicates that FX market par-

ticipants in general attach the greatest weight to technical

analysis (40.2%), followed by fundamental analysis

(36.3%), and finally by flow analysis (23.5%). Among the

participants surveyed, FX dealers were the ones who as-

signed the most weight to flow analysis, with 26.2% of

them viewing it as their most important source of informa-

tion. By comparison, only 16.8% of fund managers re-

garded flow analysis as their most important source of

information, with 46.2% considering fundamental analy-

sis as their most important source.

FX market participants generally regard the analysis of FX

flows as more useful for short-term rather than long-term

forecasting. As shown in the pie chart on the bottom right,

25.4% of all respondents believe that flow analysis pro-

vides valuable information only on an intra-day basis. An-

other 37.3% believe that the information provided by flow

analysis can be useful up to a few days. Combined, 62.7%

of all respondents believe that flow analysis should be lim-

ited to forecasting horizons up to only a few days. In terms

of the longer-run usefulness of flow analysis for currency

forecasting, 21.9% of FX respondents believe that flow

analysis can be useful for several weeks, while only a

modest 15.4% believe that flow analysis can be useful for

up to several months into the future.

Survey of All FX Market Participants

The Importance of Fundamental, Technical & Flow

Analysis in Currency Investment Decision

Fundamental Analysis--36.3%

Technical Analysis--40.2%

Flow Analysis--23.5%

Survey of Fund Managers

The Importance of Fundamental, Technical and

Flow Analysis in Currency Investment Decision

Fundamental Analysis--46.2%

Technical Analysis--37.0%

Flow Analysis--16.8%

Survey of FX Dealers

The Importance of Fundamental, Technical & Flow

Analysis in Currency Investment Decision

Fundamental Analysis--32.4%

Technical Analysis--41.4%

Flow Analysis--26.2%

The Usefulness of Flow Analysis

at Different Forecasting Horizons

(Survey of All FX Participants)

Intraday--25.4%

A Few Days--37.3%

A Few Weeks--21.9%

More than 2 Months--15.4%

Source: Gehrig and Menkhoff (2002) Source: Gehrig and Menkhoff (2002)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

26 Deutsche Bank Foreign Exchange Research

Like positioning and sentiment data, the empirical evidence

suggests that FX customer order flow has a contempora-

neous—rather than a predictive—relationship with ex-

change-rate movements. The chart below indicates that

cumulative trends in order flow and the trend in the dollar

closely parallel one another on a contemporaneous basis.

(Note that since both exchange rates and customer flow

data can fluctuate erratically on a daily basis, we have

smoothed out the erratic daily fluctuations in these time

series to discern whether a strong positive correlation ex-

ists between order flow and exchange rates.)

There is no evidence, however, of a statistically significant

relationship between lagged order flow and future ex-

change-rate movements. This finding is supported by vari-

ous academic studies that have investigated the impor-

tance of order flow on exchange rates.

Such findings should not come as a surprise. After all, if

there are more dollar buyers than dollar sellers on a sus-

tained basis in the FX market, both cumulative net dollar

purchases and the dollar's value should rise in a parallel

fashion. One should expect dollar purchases in the cur-

rent period to drive the dollar higher now, not at some

point in the future.

Empirical work on customer order flow's impact on ex-

change rates indicates that certain customers' FX orders

may exert a greater impact on exchange rates than other

customers' FX orders. Research by Fan and Lyons (1999)

indicates that real-money accounts (i.e., unleveraged fund

managers) played a more important role in driving the euro

than did corporate or hedge fund players. In the case of

the yen, hedge funds and corporates played a relatively

more important role.

Assessing the Statistical Significance of

Sentiment/Positioning/Flow Variables as

Explanatory Variables of Exchange-Rate Movements

(January 15, 1999-November 2, 2001 Weekly Data)

Lagged

Contemporaneous One Period

Market Positioning in the US$ (DB Flow Data)

Regression Coefficient 0.69 -0.40

(t-Statistic) (1.97) * (-1.13)

Market Positioning in the Euro (DB Flow Data)

Regression Coefficient 0.066 0.057

(t-Statistic) (2.88) * (1.82)

Market Positioning in the Yen (DB Flow Data)

Regression Coefficient 0.187 -0.024

(t-Statistic) (1.75) (-0.22)

*Indicates a statistically significant value at the 95% confidence level.

The Impact of Real Money, Hedge Fund, and Corpo-

rate Customer Order Flows on Exchange Rates

(January 1993-June 1999, Monthly Data)

Real Money Leveraged Non-Financial

(Unhedged) (Hedged) Corporate

Flow Flow Flow

Euro (R2 = 0.27)

Regression Coefficient 1.5 0.6 -0.2

(t-Statistic) (4.6) (1.6) (-0.5)

Yen (R2 = 0.34)

Regression Coefficient 1.1 1.8 -2.3

(t-Statistic) (1.9) (4.9) (-3.5)

*Source: Richard K. Lyons,

The Microstructure Approach to Exchange Rates

, 2001.

Market Positioning and the U.S. Dollar

(DB Flow Momentum Indicator, Moving Average)

(April-November 2001)

102

103

104

105

106

107

108

-10

-5

Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01

US$ Index----Net USD Buying/Selling(index)___

Source: DB, Datastream

Customer Order Flow and Exchange Rates—Empirical Evidence

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 27

Surveys of global fund-manager positioning are another

tool that FX market participants have turned to in recent

years for short-term exchange-rate forecasting. The monthly

Russell-Mellon survey summarizes the portfolio positions

of 28 global fixed-income managers in selected curren-

cies and bond markets, with attention normally focused

on the 25th, median, and 75th percentile exposures. The

trends in investor exposure to the dollar, euro, and yen for

the 1996-2001 period are shown in the charts on the left.

In an attempt to better detect underlying trends in currency

positioning, we have calculated a net weighted-average

exposure index for investor allocations to the dollar, euro,

and yen in the Russell-Mellon survey. We make the heroic

assumption that the range of all reported currency expo-

sures can be defined by a normal distribution, and that the

25th, median, and 75th percentile exposure fit comfortably

inside this normal distribution. We then calculate a net

weighted-average exposure for each currency, which is

plotted against each of the respective currencies in the

diagrams on the right. These charts suggest that the use-

fulness of fund-manager positioning in predicting exchange-

rate movements may be tenuous at best.

Investor Positioning and the Trend in Exchange Rates

Source: Russell Mellon, Datastream

Investor Exposure to the U.S. Dollar

(DB Net Weighted-Average Exposure to the US$)

(Russell-Mellon Survey)

-20

-10

1996 1997 1998 1999 2000 2001 2002

25 and 75 Percentiles----

Net Weighted Exposure Index____

Source: Russell Mellon

Investor Exposure to the Euro

(DB Net Weighted-Average Exposure to the Euro)

(Russell-Mellon Survey)

-30

-20

-10

1996 1997 1998 1999 2000 2001 2002

25 and 75 Percentiles----

Net Weighted Exposure Index____

Source: Russell Mellon

Investor Exposure to the Japanese Yen

(DB Net Weighted-Average Exposure to the Yen)

(Russell-Mellon Survey)

-20

-15

-10

-5

1996 1997 1998 1999 2000 2001 2002

25 and 75 Percentiles----

Net Weighted Exposure Index____

Source: Russell Mellon

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

28 Deutsche Bank Foreign Exchange Research

Surveys of investor positioning can also be used to gauge

investors' appetite to take on currency risk. We have con-

structed a composite currency exposure index using the

net weighted-average exposures of the Russell Mellon sur-

vey. The Deutsche Bank Currency Exposure Index is a

simple average of the absolute value of the net weighted-

average overweight/underweight exposures that global

fund managers maintain toward the dollar, euro, and yen.

A high index value indicates that overall investor position-

ing is aggressive, while a low index value indicates that

overall investor positioning is not aggressive.

The Deutsche Bank Currency Exposure Index reveals that

investors willingly took on aggressive currency exposures

in 1996-97, with overweight/underweight positions aver-

aging 6-10 percentage points above or below their bench-

mark weights. Over the 1997-2002 period, however, the

appetite to take on greater currency exposures appears to

have fallen sharply. At present, the average overweight/

underweight currency exposure that investors are main-

taining relative to their benchmark index amounts to just

two or three percentage points relative to market norms

as investors have chosen to position their portfolio closer

to their benchmarks.

What accounts for this drop-off in investors' currency-risk

appetite? We believe that investors have been caught in a

bind in terms of weighting the conflicting forces of valua-

tion and trend. For instance, investors may have felt com-

pelled to underweight the dollar and overweight the euro,

given the dollar's huge overvaluation and the euro’s huge

undervaluation. But trend-following models have been ar-

guing that investors should take the opposite position.

Given this conflict between valuation and trend, investors

have opted to take on neutral currency exposures and have

been unwilling to shift from this stance for the past few

years. Perhaps investors are waiting for a more definitive

sign that the dollar's long-term appreciating trend has con-

vincingly reversed.

Deutsche Bank Currency-Exposure Index

1996 1997 1998 1999 2000 2001 2002

(Composite Exposure Index)

Source: Based on the Russell Mellon Survey

Currency Exposure and the Decline in Investor Appetite for FX Risk

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 29

Currency values can be affected by shifts in investors' over-

all appetite to take on risky positions. At any point in time,

investors' overall appetite for risk could lean in favor of

either risk taking or risk aversion. During periods of high

investor confidence, investors tend to increase their expo-

sure to risky assets. During periods of heightened risk aver-

sion, however, investors tend to reduce their exposure to

risky assets. They may close out positions in foreign in-

vestments and return to home base, or they may cut sig-

nificant overweight or underweight exposures relative to

their benchmarks, particularly if such exposures are lever-

aged.

It has been suggested that certain currencies may be more

vulnerable than others during risk-averse periods. For ex-

ample, countries that run large current-account deficits

might see their currencies weaken if foreign capital were

to exit during a period of heightened risk aversion. Like-

wise, high-yield currencies could be vulnerable if inves-

tors sought to unwind risky long high-yield/short low-yield

carry trades in a risk-averse environment.

How does one measure investors' overall appetite for risk?

A number of approaches exist, but by and large they all

seek to capture increases in volatility, wider credit spreads,

trends in commodity prices, and/or shifts in yield curves.

We have constructed the Deutsche Bank Risk Appetite

Index using the components shown in the table below.

There appears to be a strong contemporaneous relation-

ship between the change in the dollar's value and the

change in Deutsche Bank's Risk Appetite Index. As with

the other sentiment/positioning/flow measures, no lead-

ing or predictive relationship is evident, which is to be ex-

pected. After all, when investors’ appetite for risk falls sud-

denly and sharply, it should affect all markets at the same

time. Implied volatility should jump upward, credit spreads

should spike higher, the yield curve should immediately

and perhaps dramatically alter its slope, and currency val-

ues should change at roughly the same time.

This contemporaneous relationship between exchange

rates and risk aversion is evident in the way the Japanese

yen/U.S. dollar exchange rate and the EMBI/U.S. Treasury

yield spread (which is considered to be a measure of

emerging-market risk) have behaved during recent peri-

ods of heightened risk aversion. In the past seven years,

there have been two occasions when the dollar has

tumbled by 20% or more over a very short time span—

January-April 1995 and August-October 1998. In both in-

stances, the EMBI spread exploded to the upside. During

both periods, the EMBI spread and the yen were respond-

ing to heightened risk aversion on a global basis as inves-

tors sought to withdraw from leveraged carry trades that

had suddenly become highly vulnerable.

DB Risk Appetite Index and the U.S. Dollar

-8

-6

-4

-2

-4

-2

1995 1996 1997 1998 1999 2000 2001 2002

US$ Index(Deviation from Trend)----DB Risk Appetite Index____

Components of the DB Risk Appetite Index

Component Reason for Inclusion

G3 Implied Three-Month FX Volatility Captures FX market uncertainty and risk.

(average of USD/JPY, EUR/JPY

and EUR/USD volatility)

VIX Index Captures uncertainty and risk in the U.S.

(CBOE OEX Volatility Index, and, therefore, global equity markets.

expected volatility

in the S&P100 index)

U.S. High Yield Bond Spread An indicator of risk in U.S. credit markets.

JP EMBI+ Composite Index An indicator of emerging-market volatility,

uncertainty, and risk.

Journal of Commerce Metals Index Proxy for global-growth risks on the

assumption that risk appetites will be

higher during periods of stronger growth.

G3 Yield Curve Proxy for global growth risks, liquidity, and

(10-year bonds less cash rates) safe-haven flows into government bonds.

Construction of the RAI: The DB RAI index is a simple average of the deviation in

each of the above indicators from their trend (standardised). In this way, the index

overcomes problems associated with, for example, the structural widening in credit

spreads in recent years. Periods spent in Risk Taking or Risk Aversion territory have

been determined from the Risk Appetite Index using a standard MACD technique,

designed to pick up shifts in the trend in the RAI.

Source: DB, Datastream

Risk Appetite Shifts and Currency Trends

Emerging-Market Credit Risk & the U.S. Dollar

(1994-1998)

-20

-10

200

400

600

800

1000

1200

1400

1600

1800

2000

1994 1995 1996 1997 1998 1999 2000 2001 2002

Yen/US$(3-mo. % chg.)----EMBI Yield Spread(bps.)____

Source: Bloomberg, Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

30 Deutsche Bank Foreign Exchange Research

Researchers have found that carry trades (long positions

in high-yield currencies funded by short positions in low-

yield currencies) work best in benign risk-appetite envi-

ronments. That is, when investors have a strong appetite

to take on risk, they are more apt to engage in riskier strat-

egies such as carry trades, and if their level of confidence

is high, they may engage in such trades in a leveraged

manner. In contrast, during periods of high risk aversion,

investors are more likely to cut risky positions, particularly

positions that might be highly leveraged. Thus, if the incli-

nation to take on carry trades tends to move positively

with investors’ appetite to take on risk, we would expect

to find a close relationship between the profitability of carry

trades and the trend in risk appetite indicators.

Indeed, as the chart below reveals, the rolling returns from

being long the three highest yielding currencies in the in-

dustrial world and being short the three lowest yielding

currencies in the industrial world closely tracks the trend

in the Deutsche Bank Risk Appetite index. Given this tight

fit, fund managers who undertake carry trades on a fre-

quent basis might find it useful to initiate carry trades when

the risk appetite index is trading upward, and to cut such

positions when the risk appetite index is trending down-

ward.

Risk Appetite and the Profitability of Carry Trades

Deutsche Bank Risk Appetite Index

and the Profitability of Carry Trades

-10

-5

-15

-10

-5

1998 1999 2000 2001 2002

Fwd.-Bias Strategy Return(%)----DB RAI(index)____

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 31

Exchange-Rate Determination in the

Long Run

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

32 Deutsche Bank Foreign Exchange Research

In the long run, a currency’s value should gravitate toward

its real long-run equilibrium value. If we were able to esti-

mate which exchange-rate level represented long-run fair

value, we would be able to identify the likely path that an

exchange rate should take on a long-term basis. Unfortu-

nately, there is no uniform agreement among economists

on what exchange-rate level represents a currency's true

long-run equilibrium level. The purchasing power parity

(PPP) approach to assessing long-run fair value probably

has the widest following among economists and strate-

gists, but it is also widely recognized that the PPP approach

has serious limitations. That is because historically, the

deviations from estimated PPP values have been both large

and persistent, which suggests that fundamental forces

other than relative national inflation rates have been play-

ing an important role in driving the long-term path of ex-

change rates.

The failure of PPP to hold over medium-term and, in some

cases, long-term horizons has led economists to consider

alternative approaches to assess long-term value in the

FX markets. For instance, the IMF favors the macroeco-

nomic-balance approach to long-term exchange-rate deter-

mination, in which the long-run equilibrium exchange rate

is defined as the rate that would equalize a country's sav-

ings-investment balance with its underlying current-account

balance. According to this approach, if there is a shift in a

country's national savings, investment, or underlying cur-

rent account, then the real long-run equilibrium exchange

rate should adjust accordingly. This approach is quite simi-

lar to the Fundamental Equilibrium Exchange Rate (FEER)

approach pioneered by John Williamson.

More recently, economists have sought to identify the long-

run equilibrium paths of currencies with econometric mod-

els, relating key economic variables to long-term trends in

exchange rates. These studies have found that long-term

trends in relative productivity growth, sectoral-productiv-

ity differentials, persistent trends in a country’s terms of

trade, the net external investment position of a country as

a percentage of its GDP, and trends in national savings and

investment have had some success in explaining the long-

term path of exchange rates.

A Stylized Model of the

Long-Term Trend in a Currency’s Value

Real Exchange Rate

Time

Long-Run

Equilibrium

Path

Exchange-Rate Determination in the Long Run

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 33

The purchasing power parity (PPP) approach to exchange-

rate determination contends that the long-run equilibrium

value of a currency is completely determined by the ratio

of domestic prices relative to foreign prices. PPP is built

on the notion of arbitrage across all tradable goods and

services. Arbitrage ensures that the prices of similar goods

in two countries will be equalized, when expressed in ei-

ther local or foreign currency terms. For example, if the

U.S. and Euroland produced a similar basket of goods, ar-

bitrage should ensure that the prices of those goods would

be the same in both regions. If the price of the U.S. basket

suddenly rose relative to the price of the Euroland basket,

the dollar would need to fall to offset the relative price

change in order to keep the prices of the two baskets equal-

ized.

The evidence on PPP suggests that there are often signifi-

cant and persistent departures from PPP in the short and

medium run. One key reason is that arbitrage in tradable

goods might be limited by transaction costs, tariffs, and

entry and exit barriers. If such barriers to arbitrage were

significant, a large "zone of inaction" might exist whereby

firms would be unwilling to arbitrage traded-goods prices,

and thus significant departures from PPP would exist over

a certain range of exchange-rate movements. Real ex-

change rates would fluctuate randomly inside this zone of

inaction, but if the real exchange rate moved outside of

the transaction bands, arbitrage would finally become prof-

itable enough to bring the real exchange rate back inside

the zone of inaction.

According to a recent survey, U.K.-based foreign-exchange

dealers believe that PPP can be a useful tool for long-run

currency forecasting, but view it as a poor forecasting tool

for the short and medium term. But even for long-run fore-

casting, there does not appear to be overwhelming sup-

port for PPP. According to the survey, only 44.3% believed

that PPP could be useful in predicting exchange-rate move-

ments beyond six months, while 34.9% believed it was

not useful, and 20.8% had no opinion. Survey of U.K.-Based FX Dealers

on the Validity of PPP as a Forecasting Tool

Question: Do you think PPP can be used to predict

exchange-rate movements?

Intraday Medium Run Long Run

(within 6 Months) (over 6 months)

Yes 4.8 16.4 44.3

No 87.4 67.3 34.9

No Opinion 7.8 16.3 20.8

Source: Cheung, Chinn and Marsh, NBER Working Paper 7524, Feb. 2000.

Purchasing Power Parity

Exchange Rate Changes Are Posited to Reflect

Changes in Relative Prices

Relative Price Level

(P /P )

GUS

1.0

DM2000/

US2000

Exchange Rate

(DM/US$)

2.0

DM2000/

US1000

PPP

45 degrees

Purchasing Power Parity

Source: Rosenberg (1996)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

34 Deutsche Bank Foreign Exchange Research

According to the relative version of PPP, the exchange rate

over time will move to offset differences in national infla-

tion rates. If, for example, the U.S. inflation rate averaged

5% higher than the Euroland inflation rate, the dollar would

need to fall by 5% per annum versus the euro to offset the

gap in relative inflation rates. Thus, if PPP held at all times,

we should expect the nominal exchange-rate path to par-

allel the inflation-differential path.

A country's real exchange rate can be defined as the nomi-

nal exchange rate adjusted for changes in relative inflation

rates. The dollar's real value would be unchanged if the

dollar's nominal value moved exactly in line with changes

in U.S./foreign inflation differentials. Thus, if the U.S. in-

flated at a rate of 5% per annum above the Euroland infla-

tion rate, and the dollar fell by 5% per annum versus the

euro at the same time, the dollar's real value versus the

euro would be unchanged over time. Indeed, if PPP holds

over time, then the real exchange rate would be constant

over time as well.

Many analysts use the trend in real exchange rates as a

gauge to assess a country’s long-run competitiveness. If a

country inflated at a rapid rate relative to its trading part-

ners, and its currency failed to decline in tandem, the

currency's real value would, as a consequence, rise. The

rise in the real exchange rate could be viewed as a mea-

sure of the loss in that country's competitiveness versus

its trading partners. If, instead, the nominal exchange rate

declined to exactly offset the difference in relative infla-

tion rates, then the real exchange rate would have been

constant and there would have been no change in trade

competitiveness.

Real Exchange Rate Determination

q = e - (p* - p)

where:

q = change in the real exchange rate

e = change in the nominal exchange rate

(p* - p) = change in relative price levels

Purchasing Power Parity and the Real Exchange Rate

Relative Price Levels and

Nominal and Real Exchange Rates

Relative Price

Level

(P /P )

EUS

(P /P )

EUS

Time

U.S. Dollar’s

Value

(e )

Time

Real

Exchange Rate

(q)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 35

PPP misalignments occur when nominal exchange rates

fail to move in line with relative inflation rates. Consider

what would happen if the U.S. inflation rate averaged 5%

per annum faster than Euroland’s inflation rate but the U.S.

dollar/euro exchange rate was fixed. By preventing a de-

cline in the dollar's nominal value that would offset the

relatively higher U.S. inflation rate, the dollar's real value

q$ = e$ - (pE - pUS)

would be pushed steadily higher as the gap between the

fixed nominal value for the dollar, e$, and a falling inflation

differential, (pE - pUS), steadily widened. The more the U.S.

inflates relative to Euroland, with no offsetting dollar de-

preciation, the less competitive the U.S. would be.

PPP misalignments can also occur if the nominal exchange

rate rises or falls by large amounts relative to modestly

changing or unchanging inflation differentials. For example,

if the U.S. and Euroland inflated at similar rates, but the

dollar's nominal value nevertheless soared versus the euro,

then the dollar's real value

q$ = e$ - (pE - pUS)

would rise as the gap between a rising dollar, e$, and an

unchanged inflation differential, (pE - pUS), steadily widened.

The consequences of a steadily rising real value for the

dollar would be a long-run decline in U.S. competitiveness.

Real Appreciations and Exchange-Rate Misalignments

The Response of Real Exchange Rates

to Changes in Relative Prices

When Nominal Exchange Rates are Fixed

Relative Price

Level

(P - /P )

E- US

(p /p )

EUS

Time

U.S Dollar’s

Nominal Value

(e )

Time

Real

Exchange Rate

(q )

The Response of Real Exchange Rates

to Changes in Nominal Exchange Rates

When Relative Prices Levels are Stable

Relative Price

Level

(P -/P )

EUS

(p /p )

EUS

Time

U.S. Dollar’s

Nominal Value

(e )

Time

Real

Exchange Rate

(q )

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

36 Deutsche Bank Foreign Exchange Research

Evidence for both developed and developing countries

suggests that PPP misalignments tend to build up gradu-

ally over time. A key reason why this pattern tends to re-

cur so often is that in the early years of an overly appreci-

ated currency, firms and households find a way to adjust,

and this, at least temporarily, works to temper the harsh

effects of real currency appreciation on the domestic

economy. For example, firms may try to maintain market

share by cutting their profit margins, while households may

attempt to maintain their spending by running down their

savings. Because of these and other adjustments, real

growth might not show any visible signs of weakness and

the current-account balance might not deteriorate signifi-

cantly in the early years of a misaligned exchange rate.

Since an overly appreciated currency is not likely to give

rise initially to a deteriorating trend in domestic economic

activity or in the current-account balance, it is highly un-

likely that the market will feel compelled to wage an attack

on the overvalued currency in the early years of real ap-

preciation. The larger a PPP misalignment grows and the

longer it lasts, however, the greater the likelihood that it

will eventually lead to a visible deterioration in economic

fundamentals. Only when economic fundamentals dete-

riorate beyond a certain threshold level (point

in the dia-

gram below) will market participants become emboldened

to wage an attack on the misaligned currency and push it

back toward its real long-run equilibrium level.

PPP Misalignments and Their Eventual Corrections

PPP

Time

% Overvalued

(0)

% Undervalued

(+)

(-)

Weighted-Average Trend

in Fundamentals

Time

Real Exchange Rate

Economic Fundamentals

PPP Cycle of Exchange-Rate Misalignment and Correction

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 37

Historically, PPP misalignments exceeding +/- 20% have

been rare and short-lived in the industrial world. When-

ever the DM/US$ exchange rate has risen or fallen by more

than 20% relative to its PPP equilibrium level, the result-

ing misalignment has proven to be transitory.

For the past two years, the dollar has been overvalued by

25%-30%, which is the largest PPP misalignment that the

dollar has experienced since the 1984-85 dollar bubble.

This suggests that the dollar has been in an upside over-

shoot, and since exchange rate misalignments of this size

tend not to be sustainable, one should expect that the

DM/US$ exchange rate will eventually move back inside

its +/- 20% PPP band.

There are signs that the overvalued dollar has been taking

its toll on U.S. competitiveness since 2000. A recent sur-

vey conducted by the Federal Reserve Bank of Philadel-

phia suggests that over 45% of U.S. firms have been nega-

tively affected in some manner by the dollar's overvalua-

tion in the past two years. Further evidence can be found

in the rising chorus of complaints by U.S. industry, as re-

flected in the letter, reprinted below, from leading manu-

facturing associations to the U.S. Treasury Secretary.

The Honorable Paul O’Neill

Secretary of the Treasury

Washington DC 20220

June 4, 2001

Dear Mr. Secretary:

We are writing to tell you that at current levels the exchange value of the dollar is having a strong

negative impact on manufacturing exports, production, and employment. A growing number of American

factory workers are now being laid off principally because the dollar is pricing our products out of

markets – both at home and abroad. Small firms are being affected as well as large ones. As you

balance your responsibilities for international monetary stability and domestic economic growth, we

ask that you take into account the growing burden an overvalued dollar is imposing on U.S. manufacturing.

Since early 1997 the dollar has appreciated by 27 percent. Industries such as aircraft, automobiles and

parts, paper and forest products, machine tools, medical equipment, steel, and other capital goods –

as well as consumer goods producers — are being affected very significantly. No amount of cost

cutting can offset a nearly 30 percent dollar markup.

The total effect on the U.S. economy is staggering. These output losses are particularly serious at this

time, as they coincide with a general economic slowdown. The economic fundamentals have changed

dramatically in the last six months. Production and profitability are down, and manufacturing employment

has fallen by more than a half million jobs since mid-2000. Yet, in the face of slowing economic growth,

declining interest rates, and rising manufacturing unemployment, the dollar has remained high.

In our view, a clarification of Treasury policy is in order, to be certain that it is not seen as endorsing an

ever stronger dollar irrespective of the economic fundamentals. We urge the Treasury to make it clear

that the value of the dollar should be consistent with economic reality and market conditions. This

policy should be buttressed by a commitment to further reductions in interest rates and to cooperating

in exchange markets as appropriate. Moreover, it is vital that the Treasury not condone currency

manipulation by trading partners seeking to make their exports more competitive.

Mr. Secretary, 18 million workers and their families depend directly on the continued strength and

competitiveness of American manufacturing. Many more Americans rely on stockholdings in our

companies for their retirement income. We would like to meet with you to describe more fully the

effects the value of the dollar is having on us. We hope you will be able to accommodate our request.

Respectfully,

Jerry J. Jasinowski, President John W. Douglass, President and CEO

Aerospace Industries Association National Association of Manufacturers

W. Henson Moore, President and CEO Don Carlson, President

American Forest and Paper Association AMT – The Association for Manufacturing Technology

Stephen Collins, President Christopher M. Bates, President & CEO

Automotive Trade Policy Council Motor Equipment Manufacturers Association

The Effect of the Dollar’s Rise on U.S. Business

(Philadelphia Fed Survey)

Question: "Which of the following best

characterizes the effect of the rise in the

value of the dollar on your business?"

Substantial Negative--18.1%

Some Negative Effect--27.7%

No Effect--43.6%

Some Positive Effect--10.6%

Substanial Positive--0.0%

PPP Case Study: Dollar Overvaluation in 2001-02

Source: Datastream

Source: Federal Reserve

Bank of Philadelphia

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

38 Deutsche Bank Foreign Exchange Research

Different approaches to calculating PPP are likely to yield

different estimates of the dollar's PPP fair value. We favor

a long-run averaging approach that allows us to circum-

vent the "base period" problem, in which PPP estimates

can be distorted upward or downward if a particular year is

chosen as the base period. We currently find the dollar to

be significantly overvalued—by over 20% versus the

euro—using either relative consumer or producer price

inflation rates to generate long-run average PPP estimates.

The Economist Magazine's Big Mac Index currently esti-

mates that the dollar is only 10% overvalued versus the

euro. The price of a McDonald's Big Mac hamburger in

Germany and France is roughly equal to the price of the

Big Mac in the U.S., but in Italy and Spain, Big Macs are

20% cheaper than in the U.S.

March 15, 2002 PPP % Over/

Currency (vs. US$) Spot Rate Estimate Under-Valued

Swedish Krona 10.32 7.02 -47.0

Canadian Dollar 1.59 1.24 -27.7

Australian Dollar 0.53 0.72 -27.5

N.Z. Dollar 0.44 0.60 -27.2

Euro 0.88 1.18 -25.4

Norwegian Krone 8.78 7.11 -23.5

Deutsche Mark 2.21 1.80 -22.9

Danish Krone 8.41 7.07 -18.9

Swiss Franc 1.66 1.45 -14.5

Japanese Yen 129.06 117.32 -10.0

British Pound 1.43 1.55 -8.0

US$ Index 117.18 96.32 21.7

CPI-Based Purchasing Power Parity (PPP) Estimates March 15, 2002 PPP % Over/

Currency (vs. US$) Spot Rate Estimate Under-Valued

Swedish Krona 10.32 7.73 -33.4

Euro 0.88 1.14 -22.7

Australian Dollar 0.53 0.67 -22.1

Swiss Franc 1.66 1.38 -20.0

Deutsche Mark 2.21 1.85 -19.8

N.Z. Dollar 0.44 0.52 -15.9

Norwegian Krone 8.78 7.75 -13.3

Danish Krone 8.41 7.46 -12.8

Canadian Dollar 1.59 1.43 -11.2

Japanese Yen 129.06 117.36 -10.0

British Pound 1.43 1.46 -2.6

US$ Index 117.18 104.22 12.4

PPI-Based Purchasing Power Parity (PPP) Estimates

Big Mac PPP Estimates

(Economist Magazine’s Big Mac Index)

-60

-40

-20

Switz.

Denmark

Britain

U.S.

Argentina

France

Japan

Mexico

Sweden

Germany

Euro area

S. Korea

Canada

Taiwan

Chile

Spain

Italy

Singapore

Brazil

Australia

Poland

N.Z.

Czech Rep.

Hong Kong

Indonesia

Hungary

Russia

Thailand

China

S. Africa

Malaysia

Philippines

Over(+) Under(-) Valued

U.S. Dollar PPP Estimates in 2002

Source: www.economist.com

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 39

PPP estimates are often better guides to medium/long-

run exchange-rate movements when traded-goods price

indices are used to derive PPP estimates rather than when

non-tradable goods price indices are used. For example,

the yen tends to move much more in line with the trend in

Japanese/U.S. relative export prices than it does with the

trend in relative consumer prices.

Since PPP operates on the principle of arbitrage, interna-

tional trade should ensure that tradable goods prices in

different countries are broadly similar when measured in a

single national currency.

The Yen and Relative Unit Labor Costs

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

100

150

200

250

300

350

76 78 80 82 84 86 88 90 92 94 96 98 00 02

Japan/U.S. ULCs----

en/US$____

The Yen and Relative Producer Prices

0.60

0.80

1.00

1.20

1.40

1.60

1.80

100

150

200

250

300

350

76 78 80 82 84 86 88 90 92 94 96 98 00 02

Japan/U.S. Wholesale Prices----

en/US$____

The Yen and Relative Consumer Prices

0.50

0.60

0.70

0.80

0.90

1.00

1.10

1.20

100

150

200

250

300

350

76 78 80 82 84 86 88 90 92 94 96 98 00 02

Japan/U.S. Consumer Prices----

en/US$____

The Yen and Relative Export Prices

0.20

0.40

0.60

0.80

1.00

1.20

100

150

200

250

300

350

76 78 80 82 84 86 88 90 92 94 96 98 00 02

Japan/U.S. Export Prices----

en/US$____

Purchasing Power Parity and the Yen

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

40 Deutsche Bank Foreign Exchange Research

There exists a huge volume of empirical work on whether

or not PPP represents a valid tool for forecasting purposes.

The weight of evidence suggests that although there are

often significant and persistent departures from PPP in both

the short and medium term, exchange rates do exhibit a

tendency to gravitate toward their PPP values in the long

run.

The consensus among academic researchers is that the

speed of convergence to PPP is quite slow—PPP devia-

tions appear to dampen at a rate of roughly 15% per year.

Empirical Evidence on PPP

This places the half-life of PPP deviations at around 4½

years for exchange rates in industrial nations. In other

words, for any given deviation of an exchange rate from its

estimated PPP value, roughly half of that deviation should

be removed in 4½ years' time. As the charts below show,

there is no positive relationship between changes in ex-

change rates and changes in relative inflation rates on a

one-year basis. However, as the time horizon is length-

ened, to six years and beyond, a strong positive relation-

ship becomes apparent.

The Impact of Relative Inflation Rates on Exchange Rates

Over Different Time Horizons

Adapted from Peter Isard, Hamid Faruqee, G. Russell Kincaid, and Marin Fetherston,

“Methodology for Current Account and Exchange Rate Assessments”,

IMF Occasional Paper 209, 2001.

-10

-20

-30

-40

-50 -50 -40 -30 -20 -10 0 10 20 30 40 50

-10

-20

-30

-40

-50 -50 -40 -30 -20 -10 0 10 20 30 40 50

-10

-20

-30

-40

-50 -50 -40 -30 -20 -10 0 10 20 30 40 50

-10

-20

-30

-40

-50 -50 -40 -30 -20 -10 0 10 20 30 40 50

% Change in Exchange Rate % Change in Exchange Rate

% Change in

Relative Inflation

% Change in

Relative Inflation % Change in

Relative Inflation

% Change in

Relative Inflation

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 41

PPP and Investment Strategy

Source: Datastream, Bloomberg

Since many empirical studies find that exchange rates ex-

hibit a tendency to gravitate toward their PPP levels in the

long run, long-term investors and issuers would stand to

benefit by positioning themselves for an eventual correc-

tion of a large PPP misalignment. If investors could count

on a rapid return to PPP, they could purchase undervalued

currencies in the hope that the undervaluation would soon

be corrected via an appreciation of the currency. Similarly,

they would look to sell overvalued currencies.

For instance, the euro's gross PPP undervaluations ver-

sus both the dollar and the yen do not appear to be sus-

tainable, given the long-run track record of these two

crossrates. Yet long-dated euro/dollar and euro/yen forward

exchange rates are not anticipating a significant correction

in the euro’s PPP misalignment over the next five years.

With conventional PPP analysis arguing for a long-term ap-

preciation of the euro versus both the dollar and the yen,

this would seem to present an opportunity for investors

and issuers to exploit the divergent trends in PPP and long-

dated forward rates.

Because the deviations from PPP can be large and persis-

tent over short and medium-term horizons, it can be dan-

gerous to pursue strategies that are based solely on an

expected rapid return of exchange rates to their estimated

PPP values. The major risk to such strategies is that the

deviations from PPP may widen further instead of narrow-

ing. Investors who embrace a PPP approach to investment

strategy would have to have substantial risk capital on hand

to absorb losses that could be incurred if the departures

from PPP proved to be sizable and persistent.

One way to manage FX risk using a PPP approach to in-

vestment strategy would be to gradually take on long po-

sitions in undervalued currencies (or short positions in over-

valued currencies) as misalignments build. Modest posi-

tions would be undertaken when the misalignments were

small, and if the deviations from PPP grow and persist,

more aggressive positions could then be undertaken.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

42 Deutsche Bank Foreign Exchange Research

According to the macroeconomic-balance approach to long-

run exchange-rate determination, a currency's real long-

run equilibrium value is defined as that rate that equalizes

a country's savings-investment balance with its underly-

ing current-account balance. The IMF uses the macroeco-

nomic-balance approach to derive quantitative estimates

of long-run fair value for all industrial-country currencies.

The IMF's approach is a three-step procedure.

1. The IMF estimates a trade-equation model to calcu-

late a country's underlying current-account position.

2. The IMF then estimates a national income model to

determine the normal or sustainable savings-invest-

ment imbalance that would prevail over the medium

run if that country were operating at its full-employ-

ment potential.

3. Finally, the IMF estimates the equilibrium exchange

rate that would equate the underlying current-account

imbalance with the sustainable savings-investment

gap.

A country's underlying current-account balance tends to

move inversely with the real exchange rate. That is, a

country's current account would tend to improve as the

currency's real value moved lower and its tradable goods

became more competitive. Hence, the underlying current-

account balance, CAB, is shown to be negatively related

to the real exchange rate in the top diagram.

A country's savings-investment gap typically moves posi-

tively with the real exchange rate. Although the overall level

of national savings does not typically respond to changes

in the exchange rate, investment spending tends to move

inversely with changes in the real exchange rate. This is

because a rising real exchange rate lowers profitability,

which in turn discourages investment spending. Thus, as

the real exchange rate rises, investment spending should

fall, and the gap between savings and investment should

therefore rise. Hence, the savings-investment, S-I, gap is

shown in the middle diagram to move positively with the

real exchange rate.

A currency's real long-run equilibrium value is determined

at the point where a country's savings-investment imbal-

ance just matches its underlying current-account imbal-

ance. The equilibrium exchange rate is shown to be q0 in

the bottom diagram where the S-I and CAB schedules in-

tersect at point

The Macroeconomic-Balance Approach

to Long-Run Exchange-Rate Determination

The Underlying Current-Account Balance

and the Real Exchange Rate

Current Account Balance

Real Exchange Rate

(-) (0) (+)

CAB

The Savings-Investment Schedule

and the Real Exchange Rate

Savings-Investment

Real Exchange Rate

(-) (0) (+)

S-I

The Equilibrium Real Exchange Rate

Is Determined When the Current-Account Gap

Equals the Savings-Investment Gap

Current-Account Balance,

Savings-Investment

Real Exchange Rate

(-) (0) (+)

S-I

CAB

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 43

If a currency's real long-run equilibrium value is determined

by the interaction of savings, investment, and external-

balance considerations, its value would be constant over

time only if there were no structural changes on the sav-

ings, investment, or external-balance fronts over the rel-

evant time period. (This is illustrated in the top diagram.)

When structural changes on those three fronts occur, how-

ever, the real long-run equilibrium exchange rate will ei-

ther rise or fall in response to those changes. A structural

rise in investment, a sustained move toward fiscal stimu-

lus, or a persistent decline in private savings should give

rise to a trend increase in the real long-run equilibrium ex-

change rate over time, as shown in the bottom diagram.

The opposite would give rise to a trend decline in the real

long-run equilibrium exchange rate.

Equilibrium Exchange Rate—Constant or Variable?

Investment-Spending

Schedule Stable Over Time

Savings minus Investment,

Current-Account Surplus

Equilibrium Real

Exchange Rate (q)

CAB

(0)

S-I

Current-Account Deficit

Structural

Shifts in

Investment

Spending

Savings minus Investment,

Current-Account Surplus

Equilibrium Real

Exchange Rate (q)

S-I

CAB

(0)

S-I

Current-Account Deficit

S-I

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

44 Deutsche Bank Foreign Exchange Research

Structural changes on a country's

internal-balance

front can

exert a powerful influence on the trend in exchange rates

over time. When a country enjoys a major investment

boom, such as the U.S. experienced in the second half of

the 1990s, it will give rise to an upward shift in the S-I

schedule and to a rise in the equilibrium exchange rate

over time. Similarly, a major move toward fiscal stimulus,

as the U.S. experienced in the first half of the 1980s, would

shift the S-I curve upward and push the currency's real

long-run equilibrium value higher, as shown in the top dia-

gram.

Structural changes on a country's

external-balance

front

can also exert a powerful influence on the trend in exchange

rates over time. Productivity changes and technological

innovations in a country's traded-goods sector can give

rise to a sustained improvement in a country's current-

account balance and thereby contribute to a trend rise in

the real exchange rate over time, as shown in the middle

diagram. Japan's large and persistent current-account sur-

pluses played a key role in driving the yen's real value higher

in the 1980s and early 1990s.

During certain periods, structural changes might be occur-

ring on both the internal and external-balance fronts at the

same time. The net impact of those changes will deter-

mine whether the exchange rate will rise or fall. For ex-

ample, in the case of Japan in 2000-01, both internal and

external-balance forces were working jointly to lower the

yen's real long-run equilibrium value. Japan's fiscal stance

was tightening, investment spending was declining, and,

at the same time, Japan's current-account surplus was

undergoing a structural deterioration as competition from

Asia intensified. As illustrated in the bottom diagram, those

forces combined to cause the yen’s equilibrium value to

decline from q1 to q2.

A Shift in Investment Behavior

Can Give Rise to an Upward Revision

in a Currency’s Long-Run Equilibrium Value

Savings minus Investment,

Current-Account Balance

Real Exchange Rate

Structural

Increase in

Investment

Spending

S-I

CAB

Upward Revision in the

Currency’s Real Long-Run

Equilibrium Value

(0)(-) (+)

S-I

A Structural Improvement in Competitiveness Can Give

Rise to an Improvement in the Current-Account Balance

and in a Currency’s Real Long-Run Equilibrium Value

Savings minus Investment,

Current-Account Balance

Real Exchange Rate

Structural

Improvement

in Current

Account

CAB

Upward Revision in the

Currency’s Real Long-Run

Equilibrium Value

(0)(-) (+)

S-I

CAB

The Combination of a Structural Decrease in Investment

and a Deterioration in the Current-Account Balance

Will Give Rise to a Downward Revision in

a Currency’s Real Long-Run Equilibrium Value

Savings minus Investment,

Current-Account Balance

Real Exchange Rate

Structural

Deterioration

in Current

Account

CAB

Downward Revision in the

Currency’s Real Long-Run

Equilibrium Value

(0)(-) (+)

S-I

CAB

Structural

Decrease in

Investment or

Gov’t Spending

S-I

Changes in Real Long-Run Equilibrium Exchange Rates

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 45

The problem for currency forecasters and investment strat-

egists is how to precisely quantify what exchange rate level

truly represents a currency's real long-run equilibrium level.

Estimates of the equilibrium exchange rate are often sen-

sitive to the assumptions that a macro-econometric mod-

eler might make. Indeed, it is quite possible that a wide

range of equilibrium exchange-rate estimates are plausible,

with each estimate dependent on the underlying assump-

tions made.

Estimates of the dollar's equilibrium value by the Institute

for International Economics show the dollar to be signifi-

cantly overvalued versus the yen and euro by 25%-30%.

A recent report by the OECD surveyed a broad range of

econometric studies that constructed estimates for the

euro's long-run equilibrium value. Although the estimates

vary widely, they tend to center around the US$/€ 1.10-

1.20 range, suggesting that the euro is significantly under-

valued versus the U.S. dollar.

Selected Estimates of the Euro's

Medium/Long-Run "Equilibrium" Value Equilibrium

Key Explanatory Exchange Rate

Study Variables/Model Estimate

(US$/€)

Wren-Lewis & Driver (1998) FEER Model 1.19-1.45

Borowski and Couharde (2000) FEER Model 1.23-1.31

Alberola et al. (1999) Ratio of Nontraded/Traded Goods Prices, 1.26

Net Foreign Assets

Chinn and Alquist (2000) M1, GDP, Short-Term Interest Rates, CPI, 1.19-1.28

Ratio of Nontraded/Traded Goods Prices

Lorenzen and Thygessen (2000) Net Foreign Assets, R&D Spending, 1.17-1.24

Demographics,

Ratio of Nontraded/Traded Goods Prices

Duval (2001) Consumption, Multi-Factor Productivity, 1.15

Real Long-Term Yield Spread,

Ratio of Nontraded/Traded Goods Prices

Clostermann and Schnatz (2000) Real Long-Term Yield Spread, Oil Price, 1.13

Government Spending,

Ratio of Nontraded/Traded Goods Prices

Teiletche (2000) Productivity, Government Spending, 1.09

Real Long-Term Yield Spread, M1,

Industrial Production

OECD GDP PPP 1.09

Gern et al. Short-Term Real Interest-Rate Differential 1.03

Schulmeister PPP for Tradeables 0.87

Deutsche Bank PPP (Long-Run Average) 1.20

Source: OECD Working Paper Number 298, June 2001 (for other than Deutsche Bank).

Equilibrium Exchange Rate Estimates

Fundamental Equilibrium Exchange Rate Estimates

(Estimates of U.S. Dollar FEERs for 2000)

Euro Japanese Yen

Range 1.135-1.387 77.3-94.5

Midpoint 1.26 86.0

Source: Simon Wren-Lewis and Rebecca Driver,

Exchange Rates for the Year 2000

, Institute for International Economics, 1998.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

46 Deutsche Bank Foreign Exchange Research

Many economists cite the rise in U.S. productivity growth

relative to productivity growth overseas beginning in the

second half of the 1990s as one of the key variables that

drove the dollar sharply higher versus most major curren-

cies. Clearly, the gains in productivity were a uniquely U.S.

phenomenon. A recent Federal Reserve Board Study found

that multifactor productivity growth in the U.S. surged in

the second half of the 1990s, while it decelerated in Japan

and was broadly unchanged in Euroland.

Fed Chairman Alan Greenspan has been a strong advo-

cate of the thesis that the dollar’s seven-year uptrend was

productivity-related. According to Mr. Greenspan, "The evi-

dent strengthened demand for the dollar, relative to the

euro, has reflected a market expectation that productivity

growth in the United States is likely to be greater than in

continental Europe in the years ahead."

The importance of relative productivity growth as a key

determinant of the dollar's value versus the euro can be

gleaned from the accompanying charts. U.S. and Euroland

labor productivity growth did not differ by much in the 1980s

and early 1990s. But beginning in 1995, U.S. productivity

growth surged relative to the trend in Euroland. Indeed,

the trend in long-term Euroland/U.S. relative productivity

growth appears to explain a large portion of the 1995-2000

slide in the euro's value versus the U.S. dollar.

U.S. and Euroland Productivity

(Derived from OECD Calculations of

Unit Labor Costs and Compensation Growth Rates)

500

1000

1500

2000

2500

3000

1983 1985 1987 1989 1991 1993 1995 1997 1999 2001

Euroland----

(Index) U.S.____

Source: OECD Economic Outlook

Productivity Growth Rates

(Percentage Changes in Multifactor Productivity)

1990-95 1996-99

(%) (%)

U.S. 0.79 1.47

Canada 0.26 0.27

France 0.89 1.12

Germany 1.02 1.07

Italy 1.32 -0.14

Japan 1.31 0.85

U.K. 1.21 0.95

Australia 1.15 2.11

Denmark 2.37 0.31

Norway 2.48 1.13

Switzerland -0.57 0.84

Source: Gust, C. and Marquez, J., “Productivity Developments Abroad”,

Federal Reserve Bulletin

, October 2000.

U.S./Euroland Productivity Ratio

and the U.S. Dollar/Euro Exchange Rate

0.90

1.00

1.10

1.20

1.30

1.40

0.6

0.7

0.8

0.9

1.0

1.1

1.2

87 88 89 90 91 92 93 94 95 96 97 98 99 00

US$/Euro----Euroland/US Productivity(ratio)____

Source: OECD Economic Outlook

Productivity Trends and Exchange Rates

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 47

There has been a great deal of debate among economists

over what contributed to the recent stellar gains in U.S.

productivity and whether those gains will prove sustain-

able in the long term. A number of researchers believe

that the absolute gains in U.S. productivity were largely an

information technology (IT) phenomenon, and since the IT

boom was largely a U.S. story, the relative gains in U.S.

productivity could be explained by IT as well. Many of those

same researchers believe that the recent gains in U.S. la-

bor productivity will prove sustainable in the long run, which

would be dollar positive. On the other side of the argu-

ment, some researchers are not convinced that the U.S.

enjoyed a productivity miracle, contending instead that

favorable cyclical factors might have exaggerated the rise

in reported productivity. If that were the case, then U.S.

productivity growth would be expected to fall back to its

long-term trend.

A recent Conference Board report demonstrates that the

relatively robust U.S. productivity gains were due not only

to the comparative advantage of IT firms in the U.S., but

also to the ability of U.S. non-IT firms to use the new infor-

mation technologies developed by the IT sector to boost

their overall productivity. This was made possible by U.S.

firms' willingness to invest in these new technologies. In

contrast, non-IT firms in Euroland and Japan were slow to

invest in new technologies, even though such technolo-

gies were as available to them as they were to non-IT firms

in the U.S. The net result is that productivity gains in the

non-IT sector of the U.S. were several times greater than

the productivity gains by comparable firms in Euroland and

Japan.

According to the Conference Board report, the failure of

Euroland firms to embrace the new technologies "can be

traced to insufficient liberalization and impediments to

market evolution." The report adds that Euroland "barriers

to change…are widespread….and opportunities are still

restricted in many information and communications tech-

nology-using industries, such as transportation, communi-

cation and banking." The report emphasized that Euroland

"regulatory rigidities inhibit reallocations of labor and capi-

tal to their most productive uses, reducing the benefits to

be obtained from investment in new technologies."

The Conference Board holds out the hope that "new com-

petitive rules in communications, banking reforms, and

even the adoption of the single currency" may make it pos-

sible for Euroland to experience stronger productivity gains

in the future, which would be positive for the euro. On a

negative note, however, the report argues that "recent com-

pilations of economic and administrative regulations sug-

gest that Europe and Japan still have some way to go be-

fore they will be able to fully exploit the new computer and

information technologies. Failure to move forward on regu-

latory liberalization and the reduction of impediments to

market evolution would perpetuate the present under-in-

vestment in information and communications technology."

The report then wryly concludes that "realizing the poten-

tial growth effects from information and communications

technology is not automatic."

One implication of the Conference Board report is that U.S.

firms may enjoy stronger productivity gains than their coun-

terparts in Euroland and Japan for a while longer. This sug-

gests that until Euroland begins to close the productivity

gap with the U.S. on a sustained basis, the euro will con-

tinue to encounter difficulty in making significant headway

versus the dollar.

Contribution of IT-Producing Industries to

Labor Productivity Growth in G-7 Countries

(1995-1999)

0.30.3

0.40.40.4

0.6

0.7

0.0

0.2

0.4

0.6

0.8

U.S. U.K. Japan Germany France Italy Canada

Percentage Contribution(%)

Source: Conference Board

Contribution of IT-Using Industries to

Labor Productivity Growth in G-7 Countries

(1995-1999)

0.4

0.20.2

0.5

0.3

0.5

1.4

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

U.S. U.K. Japan Germany France Italy Canada

Percentage Contribution(%)

Source: Conference Board

U.S. Productivity Growth Surge and the Dollar—A Medium or

Long-Term Phenomenon?

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

48 Deutsche Bank Foreign Exchange Research

A new study by McKinsey & Company, Inc. ("U.S. Produc-

tivity Growth: 1995-2000") sheds additional light on the

sources of the U.S. productivity surge. The McKinsey study

contends that U.S. productivity increased in the 1995-2000

period for both cyclical and structural reasons. The study

finds that U.S. labor-productivity growth rose from an aver-

age 1.4% per annum over the 1987-95 period to an aver-

age 2.5% per annum over the 1995-2000 period, with only

about one-half of the incremental gains projected to be

long-lasting.

The McKinsey study projects that U.S. labor productivity

growth will ease a bit toward an average 2.0% per annum

pace over the next five years as the influence of shorter-

run cyclical factors washes out. While this would repre-

sent a slowdown from the 1995-2000 pace, it would nev-

ertheless exceed the pace of productivity growth in 1987-

95.

A 0.5% per annum increase in labor-productivity growth

might not seem like a lot, but if such a gain were sus-

tained over a 10-year period, it could contribute up to $1.2

trillion in additional surpluses on the U.S. Federal govern-

ment budget balance. It could also cut the cost of a 50-

year fix to Social Security in half.

The McKinsey study finds that the gain in U.S. labor pro-

ductivity over the 1995-2000 period was more than just an

IT story. The study contends that regulatory changes, strong

competition, better management, and favorable cyclical

factors all played a positive role in boosting U.S. productiv-

ity. Interestingly, the study finds that of the 59 sectors of

the U.S. economy that were investigated, only six sectors,

representing 28% of the economy, contributed to the 1995-

2000 productivity gains. The six sectors that made sub-

stantial contributions to U.S. productivity growth were: 1)

wholesale trade, 2) retail trade, 3) security and commod-

ity brokers, 4) electronic equipment/semiconductor firms,

5) industrial equipment/computer firms, and 6) telecom

services. The other 53 sectors, representing 72% of the

economy, made either small positive or negative contribu-

tions that effectively washed each other out for a net-zero

contribution to the U.S. economy's total productivity gains.

Of concern to both the U.S. economy and the dollar is the

fact that 72% of the overall U.S. economy did not partici-

pate in the productivity gains registered in the past five

years. The dollar rose strongly because the gains regis-

tered by the other 28% were so powerful that they were

able to boost the overall U.S. economy's productivity

growth far above the pace of productivity growth in

Euroland and elsewhere. For those sectors that enjoyed

strong productivity gains, the dollar's rise has not dented

their overall competitiveness, as the positive effects of a

relative rise in productivity have effectively offset the nega-

tive effects of a rising dollar. But the other 72% that did

not participate in the productivity surge have been left to

compete in world markets with an overvalued exchange

rate without an offsetting gain in productivity. That is why

a large number of U.S. firms have begun to complain that

the dollar's strength is negatively affecting their long-run

competitiveness.

1987-95

Productivity

Growth

1995-99

Productivity

Growth

(By Sector)

U.S High- & Low-Tech Industrial Production

(Year-over-Year Percentage Changes)

-20

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Total excluding High-Tech----

(%) Computers/Communications/Semiconductors____

U.S. Productivity Gains and the Dollar—Is the Dollar Now Fairly or Over-Valued?

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 49

Productivity gains will tend to exert a greater impact on

exchange rates if those productivity gains are concentrated

in the traded-goods sector of an economy. A recent Fed-

eral Reserve Bank of New York study ("To What Extent Does

Productivity Drive the Dollar?",

Current Issues in Econom-

ics and Finance

, Volume 7, Number 8, August 2001) finds

that most of the gains in U.S. productivity in the 1990s

were concentrated in the U.S. traded-goods sector.

Looking at sectoral productivity gaps (the difference be-

tween traded-goods productivity and non-traded-goods pro-

ductivity) in the U.S., Japan, and Euroland, one finds that

the U.S. sectoral productivity gap has risen in the past 30

years as the gap between U.S. traded and non-traded pro-

ductivity growth has widened. In contrast, the sectoral pro-

ductivity gap has fallen in Japan and has been steady in

Euroland. The New York Fed's findings reveal that relative

sectoral productivity gaps explain more than two-thirds of

the movement in the yen and euro in the past decade.

U.S. Productivity Growth by Sector

(1970-1999)

1.2

0.40.4

4.6

3.2

1.0

1970-80 1980-90 1990-99

Nontraded Sector(dark)

(%) Traded Sector(light)

Source: OECD; Federal Reserve Bank of New York,

August 2001

Japan Productivity Growth by Sector

(1970-1999)

0.2

2.3

2.4

1.7

4.7

5.4

1970-80 1980-90 1990-99

Nontraded Sector(dark)

(%) Traded Sector(light)

Source: OECD; Federal Reserve Bank of New York,

August 2001

Euroland Productivity Growth by Sector

(1970-1999)

0.9

1.2

1.8

2.9

3.0

4.0

1970-80 1980-90 1990-99

Nontraded Sector(dark)

(%) Traded Sector(light)

Source: OECD; Federal Reserve Bank of New York,

August 2001

U.S. Productivity and Changes in the

U.S./Euro Real Exchange Rate

(1970-1999)

1.1

0.8

-1.0

1.8

0.1

-4.9

-6

-4

-2

1970-80 1980-90 1990-99

HBS Productivity Measure(dark)

(%) Real Exchange Rate Chg.(light)

Note: Harrod, Balassa, Samuelson Productivity.

Source: OECD; Federal Reserve Bank of New York,

August 2001.

U.S. Productivity and Changes in the

U.S./Japan Real Exchange Rate

(1970-1999)

1.5

0.4

-1.6

1.9

0.6

-2.3

-3

-2

-1

1970-80 1980-90 1990-99

HBS Productivity Measure(dark)

(%) Real Exchange Rate Chg.(light)

Note: Harrod, Balassa, Samuelson Productivity.

Source: OECD; Federal Reserve Bank of New York,

August 2001.

Productivity Growth in the Traded-Goods Sector and the Determination of

Exchange Rates

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

50 Deutsche Bank Foreign Exchange Research

The Norwegian Krone and the Price of Oil

(since January 1999)

(4.80)

(4.60)

(4.40)

(4.20)

(4.00)

(3.80)

1999 2000 2001 2002

Brent Oil(US$/bbl.)----Nkr/DM(reverse scale)____

Changes in a country's terms of trade can, at times, exert

a strong influence on the direction that exchange rates

take. For instance, in commodity-oriented industrial econo-

mies such as Australia, New Zealand, and Canada, where

commodity exports make up a large part of GDP (relative

to the rather small shares in the G3 economies), one finds

a strong positive relationship between the trend in relevant

commodity-price indices and the trend in the A$, NZ$, and

C$, respectively.

In the case of the euro, the price of oil has been found to

be an important explanatory variable, with higher oil prices

contributing to a weaker euro, and vice versa. A recent

Bank of Canada study found that a similar relationship held

between energy prices and the C$, with higher energy

prices bearish for the C$ and lower energy prices bullish

for the C$.

A recent study by the central bank of Norway (Norges Bank)

finds that there is a non-linear relationship between oil

prices and the Norwegian krone. A change in oil prices

tends to have a larger impact on the krone's value when

the price of oil is below $14 a barrel or above $20 a barrel.

When oil prices fluctuate inside the $14-$20 range, the

impact of oil price changes on the krone's value is found

to be insignificant.

The Australian Dollar and Commodity Prices

Reserve Bank of Australia Commodity Price Index

100

110

120

0.40

0.50

0.60

0.70

0.80

0.90

1.00

84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

RBA Commodity Price Index(US$-Terms)----US$/A$____

The Euro and the Price of Oil

(since April 1997)

(1.30)

(1.20)

(1.10)

(1.00)

(0.90)

(0.80)

1997 1998 1999 2000 2001 2002

Brent Oil(US$/bbl.)----US$/Euro(reverse scale)____

The Canadian Dollar and Commodity Prices

Bank of Canada Commodity (excl. Energy) Index

100

110

120

130

140

(1.65)

(1.60)

(1.55)

(1.50)

(1.45)

(1.40)

(1.35)

(1.30)

1995 1996 1997 1998 1999 2000 2001 2002

BoC Non-Energy Commod. Price Index----C$/US$(reverse scale)____

Commodity Exports as a Percentage of GDP

(for Selected Industrial Countries)

0.5

1.8

3.7

5.2

12.0

12.1

19.4

N.Z. Canada Australia UK Germany U.S. Japan

(%)

Source: Datastream Source: Datastream

Source: OECD

Terms of Trade Changes and Exchange Rates

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 51

Several researchers have noted that there exists a posi-

tive long-run relationship between a country's net interna-

tional investment position as a percentage of GDP and its

real effective exchange rate. There are several reasons why

this relationship tends to hold.

First, exchange rates typically do not adjust to equilibrate

the current-account balance in each and every period. The

reason for this is that a deficit country may be able to at-

tract sufficient capital inflows to finance its deficit for a

considerable time. Indeed, if those inflows are consider-

able, the deficit country's currency might, in fact, rise even

as the current-account deficit widens—witness the U.S.

in 1981-85 and again in 1995-2002.

If a country's current-account deficit rises persistently over

time, however, there is likely to come a point when it can

no longer be easily financed. When that point is reached, a

currency's real value could then come under considerable

downward pressure. Hence, although the exchange rate

and the current-account imbalance may not appear to be

positively related on a monthly or quarterly basis, cumula-

tive trends in current-account imbalances and exchange

rates tend to be more closely aligned.

A second reason for the positive relationship between the

trend in net international investment positions and the trend

in real exchange rates is the effect of the transfer of wealth

implied by external-account imbalances. The cumulative

trend in current-account imbalances gives rise to shifts in

the residence of wealth from deficit to surplus countries,

and such wealth transfers give rise to portfolio adjustments

that put upward pressure on surplus-country currencies

and downward pressure on deficit-country currencies over

time.

Given that most investors prefer to keep the bulk of their

wealth in domestic and not foreign-currency assets—the

so-called home-country bias—investors in surplus coun-

tries will find themselves accumulating more foreign-cur-

rency assets than they desire relative to their domestic

holdings. Over time, surplus country investors will attempt

to rebalance their portfolios in favor of domestic-currency

assets. As they do, the surplus country's currency will tend

to rise relative to the deficit country's currency. The Japa-

nese yen and the Deutschemark—to a lesser degree—

underwent long-term appreciations from the 1980s until

the mid-1990s as their cumulative current-account sur-

pluses rose.

The Japanese Yen and the U.S./Japan

Cumulative Current-Account/GDP Differential

(50)

(40)

(30)

(20)

(10)

100

150

200

250

300

77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96

Cum. Curr.-Acct. as % of GDP Diff.---- (%)

/US$____

The Deutschemark and the U.S./German

Cumulative Current-Account/GDP Differential

(50)

(40)

(30)

(20)

(10)

1.00

1.50

2.00

2.50

3.00

3.50

85 86 87 88 89 90 91 92 93 94 95 96

Cum. Curr.-Acct. as % of GDP Diff.---- (%)DM/US$____

Source: Datastream Source: Datastream

Net International Investment and the Equilibrium Exchange Rate

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

52 Deutsche Bank Foreign Exchange Research

A final reason why the trends in net international invest-

ment positions and real exchange rates are positively re-

lated is that heavily indebted nations need to service their

net external debts, and that can only be done if the debtor

country runs a trade surplus to generate the income to

service those debts. In order for the debtor country to run

a trade surplus consistently over time, the real exchange

rate of a debtor country would need to decline. Similarly, a

country with a large cumulative current-account surplus

would need to see its currency's real value rise over time

to insure that the surplus and the associated net foreign

asset position did not continue rising to unsustainably high

levels.

Although the U.S. has seen its net international investment

position deteriorate sharply in the past decade, the dollar

has nevertheless remained extraordinarily resilient. One

explanation is that, although U.S. external liabilities exceed

U.S. external assets by a wide margin, the U.S. has tended

to earn a higher return on its external assets than it has

paid out on its external liabilities. The result has been that

net investment income flows, until recently, posted a mod-

est surplus. While U.S. net investment income flows have

recently swung into a modest deficit position, the deficit

does not appear to be sizeable enough at this stage to

seriously damage the dollar's long-run prospects.

U.S. Balance on Investment Income

(Four-Quarter Moving Sum)

-30

-20

-10

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

(US$ bn.)

U.S. Net Investment Position

(at Current Cost and Market Value)

-2000

-1500

-1000

-500

500

80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99

Market Value(light)

(US$ bn.) Current Cost(dark)

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 53

The dollar has exhibited a tendency to both rise and fall

over long-term cycles, with each of the dollar cycles since

the floating of exchange rates in 1973 sharing a number of

common features.

First, the magnitude of the dollar's exchange-rate move-

ments in each cycle has generally been large, often far

exceeding the forecasts of portfolio managers, market pun-

dits and forward-exchange rates.

Second, the period of each dollar cycle has tended to be

long, often surpassing by a wide margin the typical length

of U.S. and foreign business cycles. That would suggest

that cyclical factors alone cannot account for the tendency

of the dollar to move in large and protracted swings.

Third, at the end of each major cycle, there has been a

tendency for the dollar to overshoot its long-run equilib-

rium level, often in a climactic fashion.

Finally, when the dollar has overshot its fair value at the

end of a long cycle, the deviation in the dollar's value from

its estimated PPP value has often given rise to serious

imbalances on either the internal or external-balance fronts

in the U.S., Europe, or Japan. When those imbalances

reached a critical level, it often set market forces in motion

to bring the overshoot to an end, and to plant the seeds

for a 180-degree shift in the dollar's long-term trend.

Long-Term Cycles in Exchange Rates

Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

54 Deutsche Bank Foreign Exchange Research

Structural changes on a country's internal or external-bal-

ance front are largely responsible for the long-term cycles

in exchange rates that we have witnessed. For instance,

the U.S. investment boom of the second half of the 1990s

and the structural rise in U.S. productivity during that pe-

riod are largely responsible for the dollar's persistent rise

over the 1995-2002 period. As illustrated in the first dia-

gram below, the U.S. savings-investment schedule, S-I,

gradually shifted upward and to the left, giving rise to a

persistent uptrend in the dollar's real long-run equilibrium

value over time. Likewise, the persistent rise in Japan's

current-account surplus in the 1980s and early 1990s played

a key role in lifting the yen's equilibrium value over time,

as illustrated in the second diagram.

In both the dollar's and the yen's case, the upward revi-

sions in the real long-run equilibrium exchange rates ap-

pear to have taken place gradually, and not instantaneously

as several theories would have suggested. This would ex-

plain why the up and down cycles in exchange rates en-

dure for years at a time.

Factors Contributing To Long-Term Cycles

The Positive Effects over Time of Structural Increases in Investment

on a Currency’s Long-Run Equilibrium Value

Savings minus Investment,

Current-Account Balance

Real Exchange Rate

Persistent

Structural

Increases in

Investment

Spending

S-I

CAB

Upward Revision in

the Currency’s Real

Long-Run Equilibrium

Value over Time

(0)(-) (+)

S-I

Time

Real Exchange Rate

The Positive Effects over Time of Structural Increases in a Current-Account Surplus

on a Currency’s Long-Run Equilibrium Value

Savings minus Investment,

Current-Account Balance

Real Exchange Rate

Persistent

Structural

Increases in

Current-Account

Balance

CAB

Upward Revision in

the Currency’s Real

Long-Run Equilibrium

Value over Time

(0)(-) (+)

S-I

CAB

Time

Real Exchange Rate

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 55

Most long-term exchange-rate cycles are not driven by a

single shock to the internal or external-balance schedules,

but rather by a series of separate, yet related shocks that

reinforce one another in driving an exchange rate higher or

lower over time. For instance, the dollar's long-term uptrend

between April 1995 and November 2000 reflected a com-

bination of favorable factors including: (1) "New Economy"

forces that helped propel U.S. investment spending and

productivity growth higher, (2) a structural shift in capital

flows from emerging markets and Euroland to the U.S., (3)

higher oil prices, and (4) relatively tight U.S. monetary and

credit conditions up until 2001.

In the case of the euro, its long-term decline since the

early 1990s also reflected a multitude of problems includ-

ing: (1) pervasive structural rigidities in European labor and

product markets, (2) an adverse fiscal/monetary policy mix

in Euroland, (3) higher oil prices, (4) an adverse structural

shift in Euroland capital flows, and (5) excessive pessimism

regarding the potential problems associated with the launch

of the new euro notes and coins in early 2002.

U.S. Dollar Major Currency Index

(1995-2002)

100

110

120

1995 1996 1997 1998 1999 2000 2001 2002

(Index)

U.S. Dollar/Euro Exchange Rate

(and the Euro’s Downtrend since September 1992)

0.80

1.00

1.20

1.40

1.60

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

US$/Euro

Long-Term Cycles in Exchange Rates: The Dollar and Euro in the 1990s

Source: Datastream Source: Datastream

The Dollar’s Long-Term Uptrend

A Stylized Diagram of the Fundamental Forces that Have Contributed to

Increases in the Dollar’s Real Long-Run Equilibrium Value

US$ Real

Exchange Rate

Correction

of Initial

Undervaluation

U.S. Investment Boom

Time

PPP=q

Higher Oil Prices

Structural Shift in Portfolio Flows

Tighter Fed Policy

The Euro’s Secular Decline

A Stylized Diagram of the Fundamental Forces

That Have Contributed to the Euro’s Long-Term Slide

The Euro’s

Value

Purchasing Power Parity

Euroland’s Adverse Policy Mix

Time

Higher Oil Prices

Euroland’s Structural Rigidities

Adverse Structural Shift

in Capital Flows

Euro’s Confidence Crisis/

Excessive Pessimism

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

56 Deutsche Bank Foreign Exchange Research

U.S. Nonresidential Fixed Investment and

Equipment & Software Investment

(Year-over-Year Percentage Changes)

-15

-10

-5

88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

Equipment & Software----

(%) Fixed Investment____

Source: Datastream

The dollar's rising trend versus most major currencies over

the 1995-2002 period had, for the most part, strong funda-

mental underpinnings. The U.S. economy underwent a

major investment boom over this period, particularly when

compared with the weak investment performance in Eu-

rope and Japan. A large portion of the U.S. boom was con-

centrated in the information technology (IT) arena where

U.S. industry held a dominant, global position. The IT-led

investment boom, in turn, contributed to a surge in U.S.

productivity growth relative to productivity growth else-

where. Since exchange-rate trends are often strongly in-

fluenced by relative productivity trends, the relative surge

in U.S. productivity growth had the effect of raising the

dollar's equilibrium value.

In addition to the rise in productive investment that oc-

curred during the 1995-2000 period, there was probably

also a considerable amount of excess investment during

the height of the IT bubble period, particularly in the

Internet, telecom, and PC areas. This excess investment

spending, which is shown in the diagram below as left-

ward shifts in the savings-investment schedule to S-I3, is

now in the process of being unwound. Indeed, the S-I curve

now appears to be shifting back down to perhaps S-I2, sug-

gesting that the dollar's equilibrium value should be shift-

ing downward as well. At the time of this writing, how-

ever, the dollar's actual value does not appear to have fallen

nearly as much as the apparent decline in its equilibrium

value.

Real Gross Non-Residential Fixed Capital

Formation in the U.S., Euroland, and Japan

(1992-2000)

-40

-20

100

120

140

1992 1993 1994 1995 1996 1997 1998 1999 2000

Europe---- Japan....

(Cumulative % Change) U.S.____

The 1990s U.S. Investment Boom and the Dollar

Increase in

Capital Inflow

(equal to the gap

between

U.S. domestic

investment

and savings)

The Increase in Investment Spending and the Dollar

Increase in

U.S.

Investment

Spending

Real

Appreciation

of the

U.S. Dollar

Increase in

Future U.S.

Productivity

Growth

The Increase in U.S. Investment Spending

Has Given Rise to an Upward Revision

in the Dollar’s Long-Run Equilibrium Value

Savings minus Investment,

Current-Account Balance

Real Exchange Rate

Structural

Increase in

Investment

Spending

S-I

CAB

Upward Revision in the

Currency’s Real Long-Run

Equilibrium Value

(0)(-) (+)

S-I

Excess

Investment

Spending

S-I

“New Economy” Phenomenon

Excessive Investment?

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 57

The Dollar’s Dramatic Rise in 1999-2000 May Have Been Driven by a Sudden

Upward Revision in the Dollar’s Real Long-Run Equilibrium Level

Source: Datastream

The dollar’s dramatic surge between the fall of 1999 and

the fall of 2000 may have been propelled by an upward revi-

sion in the market's expectations regarding the U.S.

economy's long-run growth prospects and a subsequent

upward revision in the dollar's real long-run equilibrium value.

The U.S. economy had been growing faster than the

Euroland economy for much of the 1990s, but it wasn't

until late in the decade that observers began to take no-

tice that "something special" was taking place that was

distancing the U.S. from the rest of the world. As the ac-

companying charts show, in annual surveys from 1992-99,

professional economists had been projecting a relatively

bland long-term U.S. economic outlook, with real GDP

growth expected to average 2.5% per annum and produc-

tivity expected to rise by a modest 1.5% per annum over

each ensuing 10-year period.

Normally, one would think that shifts in expectations re-

garding the U.S. long-term growth outlook would take place

gradually over a number of years. But in 1999, expecta-

tions seemed to change overnight. In the 2000 and 2001

surveys, professional economists’ long-term projections

were pushed up sharply. Economists were now forecast-

ing long-term U.S. real GDP growth of 3.3% and U.S. pro-

ductivity growth of 2.5% per annum.

It was as if economists suddenly embraced the notion that

the U.S. was now enjoying a productivity miracle, brought

about by the "New Economy" phenomenon, and that the

gains in productivity would be permanent. Quite likely, a

literature search would show a quantum leap in the num-

ber of academic articles that appeared with the term "New

Economy" in the title beginning in 1999. (For a discussion

of the New Economy's impact on the FX market, see

Liesbeth Van de Craen and Peter Vanden Haute,

The New

Economy and the Foreign Exchange Market

Assuming the marketplace embraced the same long-term

views of the U.S. economy as those held by professional

forecasters, it is not hard to explain why the dollar rose 50

pfennigs versus the Deutschemark between autumn 1999

and autumn 2000. We believe that the suddenly improved

long-term growth outlook for the U.S. economy compelled

market participants to revise up sharply their notion of the

dollar's real long-run equilibrium value. Indeed, the dollar's

rise from the autumn of 1999 to the autumn of 2000, which

followed the upward revision in the U.S. long-term growth

outlook, roughly matched, in terms of magnitude, the rise

that occurred in the prior 4½ years.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

58 Deutsche Bank Foreign Exchange Research

The "synthetic" euro's real trade-weighted value had been

declining on a trend basis for much of the decade heading

into its 1999 launch date. Yet many pundits expected that

the new single currency would be strong, believing that

the euro would embrace all of the hard-currency attributes

of the Deutschemark and would discard all the soft-cur-

rency attributes of its depreciation-prone ERM members.

This did not happen. By 2001, the euro was trading close

to its 20-year low versus the yen and British pound, and—

except for the 1985 dollar bubble—the euro was trading

close to a record low versus the U.S. dollar.

Perhaps those pundits failed to recognize that the euro's

long-term behavior, being a weighted average of all EMU

member currencies, resembled that of a soft and not a

hard currency. In fact, its persistent decline versus a wide

range of currencies suggests that negative fundamental

developments intrinsic to Euroland probably played a large

role in the euro's long-term slide.

US$/Euro Exchange Rate

(1980-2002)

0.60

0.80

1.00

1.20

1.40

1.60

1.80

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

US$/Euro

Synthetic Euro Real Trade-Weighted Index

(1980-2002)

100

110

120

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Bank of England Index

Average

Canadian Dollar/Euro and

Australian Dollar/Euro Exchange Rates

(1980-2002)

0.80

1.00

1.20

1.40

1.60

1.80

2.00

2.20

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

A$/Euro----

$/Euro____

The Euro Relative to the Euroland Currencies

(European Currencies versus the U.S. Dollar)

100

120

140

160

180

75 77 79 81 83 85 87 89 91 93 95 97 99 01

France.... Italy/Spain__ __

(Index) Euro____ DM----

British Pound/Euro Exchange Rate

(1970-2002)

0.50

0.60

0.70

0.80

0.90

1.00

70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02

£/Euro

Japanese Yen/Euro Exchange Rate

(1970-2002)

100

200

300

400

500

600

70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02

/Euro

The Euro's Long-Term Struggle

Source: Datastream

Source: Datastream Source: Datastream

Source: Datastream

Source: DatastreamSource: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 59

Structural rigidities in Europe's labor and product markets

undoubtedly have adversely affected Europe's long-term

growth potential and thus probably contributed to a down-

ward revision in the market's assessment of the euro's

real long-run equilibrium value as well. Structural rigidities

in Europe's labor market include: job protection legisla-

tion, working time regulations, overly generous early re-

tirement schemes, unemployment insurance and welfare

benefits, and barriers to hiring skilled labor from abroad.

On the product front, structural rigidities include the slug-

gishness of firms to adopt a culture of risk taking, the high

cost of telecommunications, and the unwillingness of many

European governments to allow market forces to operate

unfettered. The charts below show that European labor

markets are far more regulated than the U.S., European

unemployment benefits are far more generous, while Eu-

ropean taxes and real labor costs are higher than they are

in the U.S. as well.

Strictness of Labor Market Regulation

(Most to Least Strict)

reece

Italy

Spain

France

Sweden

EMU-11

Germany

Ireland

Finland

Portugal

Norway

Austria

Belgium

Netherlands

Switz.

Denmark

U.K.

U.S.

ggregate Index

Source: OECD

Overall Taxes as a Percentage of Earnings

enmark

Belgium

Sweden

France

Germany

Norway Italy

Finland

Ireland

Netherlands

EMU-11

Portugal

Spain U.K. U.S.

(%)

Source: OECD

Unemployment Benefit Generosity

(Most to Least Generous)

50.1

5959

6262

100

Denmark

Netherlands

Sweden

Norway

Switz.

Belgium

Finland

France

Germany U.K.

EMU-11

Spain

Ireland Italy U.S.

ggregate Index

Source: OECD

Europe’s Real Labor Costs & Total Employment

100

120

140

160

180

1970 1975 1980 1985 1990 1995 1998

Total Employment----

(1970 = 100) Real Labor Costs____

Source: OECD

2001 Global Competitiveness

(based on World Economic Forum Rankings)

Finland

U.S.

Canada

Singapore

Australia

Norway

Taiwan

Netherlands

Sweden

N.Z.

Ireland

U.K.

Hong Kong

Denmark

Switz.

Iceland

Germany

Austria

Belgium

France

Japan

Spain

Korea

Israel

Portugal

Italy

Chile

Hungary

Estonia

Malaysia

Slovenia

Mauritius

Thailand

South Africa

Costa Rica

Greece

Rankings(Most to Least Competitive)

Source: World Economic Forum

(Top-36 Countries)

European and U.S. Total Employment

100

105

110

115

120

125

130

135

84 86 88 90 92 94 96 98 00 02e

U.S.----

(1984=100) Europe____

Source: OECD Economic Outlook

Euroland's Structural Rigidities and the Euro's Real Long-Run

Equilibrium Value

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

60 Deutsche Bank Foreign Exchange Research

Since the beginning of floating exchange rates, there have

been a number of occasions when exchange rates have

significantly overshot their long-run equilibrium path, ei-

ther on the upside or the downside. We define an exchange-

rate overshoot episode as one in which an exchange rate

deviates by more than +/- 20% from its long-run purchas-

ing power parity path and/or where there exists a signifi-

cant divergence between the actual trend in exchange rates

and some long-term, key determinant—such as the trend

in real interest-rate differentials in the case of the Deut-

schemark, or the trend in relative monetary-base growth

rates in the case of the Japanese yen.

For the most part, long-term cycles have strong fundamen-

tal underpinnings. That is, we can find a strong positive

relationship between the long-term trend in exchange rates

and the long-term trend in some key fundamental indica-

tor. What we find interesting, however, is that at the end

of a typical long-term dollar cycle there has been a ten-

dency for the dollar's value to overshoot its long-run fun-

damentally driven equilibrium path by rather large amounts.

A simple model can be designed to explain why exchange

rates tend to overshoot. The model assumes that inves-

tors' expectations of exchange rates are driven by both

fundamental and technical considerations. Mathematically,

we can say that exchange rate expectations, eet, in time

period

are driven by fundamental forces, fundt, and by

the past pattern of exchange-rate movements, et-1.

It is further assumed that the market will assign weights

to the relative importance that it attaches to the current

trend in fundamental forces,

, and to the past pattern of

exchange-rate movements, 1-

. Those weights are as-

sumed to vary over time, with the market assigning more

weight to fundamentals and less weight to technicals, and

vice versa, as fundamental and technical forces vary in

terms of influencing the path that currencies take. Ex-

change-rate expectations can therefore be expressed as:

eet =

(fundt) + (1-

)et-1

A Stylized Model of Exchange-Rate Overshooting

at the End of a Long-Term Uptrend

Currency’s

Value

Time

Economic Fundamentals

Exchange Rate Expectations can be expressed as:

et = w (fundt) + (1-w )et-1

where:

eet= exchange-rate expectations

fundt= fundamental forces

et-1 = past pattern of exchange-rate movements

w = market-assigned weight to the importance of the current

trend in fundamental forces

1-w = market-assigned weight to the importance of the past pattern

of exchange-rate movements

Overshooting Exchange Rates

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 61

Typically, in the early and middle stages of an extended

exchange-rate cycle, the market tends to attach greater

weight to the trend in economic fundamentals (i.e.,

rises). This is because the trend in exchange rates and the

trend in key fundamental variables parallel each other quite

closely. In the latter stages of an extended exchange-rate

cycle, however, there often occurs some disconnect be-

tween the trend in the exchange rate and the trend in key

fundamental variables.

What causes this disconnect to occur is the tendency of

the market at the end of a long cycle to assign less weight

to fundamental forces and more weight to the prevailing

trend in exchange rates (i.e., 1-

rises). For instance, after

a long cycle of currency appreciation, the market gets ac-

customed to being long the appreciating currency and

tends to shrug off any adverse underlying fundamental

developments as being just a temporary phenomenon.

Hence, the market will begin to attach less weight to the

deteriorating trend in fundamentals (i.e.,

will gradually

fall) and more weight to the past favorable trend in the

exchange rate (i.e., 1-

will gradually rise).

By attaching greater weight to expectations that the pre-

vailing trend will remain strong and less weight to the de-

terioration of fundamental forces, the trend in exchange

rates and the trend in fundamentals becomes discon-

nected. Eventually, a point will be reached when the mar-

ket finally recognizes that the deterioration in fundamen-

tals is a permanent, and not just a temporary phenom-

enon. At that point, market expectations will shift, result-

ing in greater weight being assigned to deteriorating fun-

damental forces and less weight to the prospect that the

exchange-rate overshoot will persist. It is at this stage that

the unwinding of long speculative positions gives rise to a

rapid, and at times violent, correction of the exchange-rate

overshoot.

There have been several extraordinary episodes of major

currency overshooting since the beginning of floating ex-

change rates—the Deutschemark in the mid-1980s, the

yen in 1995 and 1998, and the dollar in general in 2000-02.

In the case of the Deutschemark, there was a significant

disconnect between the trend in U.S./German real long-

term interest-rate differentials and the DM/US$ exchange

rate towards the end of the dollar's remarkable run up-

ward during its bubble period. There was simply no way to

explain the dollar's surge in 1984-85 on the basis of real

yield spreads, and this overshoot in the dollar's value was

eventually corrected in 1985-87.

In the case of the yen, there was a similar disconnect be-

tween the trend in U.S. and Japanese relative monetary-

base growth and the trend in the yen's value versus the

dollar in 1995 and again in 1998. The dollar overshot to the

downside in 1995 and overshot to the upside in 1998, and

in both cases the overshoots were subsequently corrected.

Source: Datastream

The Market Attaches Less Weight

to Fundamental Forces During

an Exchange-Rate Overshooting Period

Market-Assigned

Weight

(w)

Time

Market-Assigned

Weight

(1-w)

Beginning of

Exchange-Rate Cycle Overshooting

Period

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

62 Deutsche Bank Foreign Exchange Research

It is widely recognized that financial markets need stabiliz-

ing speculators to ensure that financial market prices do

not wander too far from fair value. By buying at price lev-

els that are perceived to be low and selling at price levels

that are perceived to be high, speculators will ensure that

asset prices do not become too stretched relative to fair

value. If there is a fairly elastic supply of speculative capi-

tal in the foreign-exchange market, a significant increase

in commercial demand for dollars (perhaps generated by a

foreign acquisition of a U.S. firm) need not give rise to a

dramatic rise in the dollar's value. Instead, the increased

commercial demand for dollars may give rise to only a

modest increase in the dollar's value from e0 to e1 as shown

in the diagram below.

However, consider a situation where there is an absence

or a shortage of stabilizing speculative capital. This might

occur if hedge funds, investors, and foreign-exchange trad-

ers are unwilling or unable to commit relatively large sums

of capital for speculative purposes. In such a case, the

supply of speculative capital in the foreign-exchange mar-

ket would appear to be relatively inelastic. In that environ-

ment, it may only take a modest increase in the commer-

cial demand for dollars to push the dollar's value sharply

higher from e0 to e2. In other words, when there is a short-

age of stabilizing speculative capital, exchange rates are

likely to be more volatile and prone to overshooting, even

amid modest changes in the commercial demand for dol-

lars.

Such behavior might explain why the dollar has remained

significantly overvalued against a wide range of curren-

cies in the past few years. With many investors, hedge

funds, and foreign-exchange traders unwilling or unable to

risk sufficient capital to bet against the dollar, there simply

has not been enough speculative capital in the market to

reverse the dollar's overshoot.

The tendency of speculative capital to dry up when mar-

ket prices move far out of line with fair value is not unique

to the foreign-exchange market. Indeed, most speculative

bubbles, particularly at their peaks, are not fed by exces-

sive speculation, but instead are driven by an absence of

stabilizing speculation. When an absence of stabilizing

speculative capital causes exchange rates to overshoot

their long-run equilibrium value, central banks often inter-

vene and take on the role of a stabilizing speculator.

How a Shortage of Stabilizing Speculative Capital Can Give Rise to

Exchange-Rate Overshooting

An Absence of Currency Speculation

Can Lead to Greater Exchange-Rate Volatility

and Exchange-Rate Overshooting

Quantity of Dollars

Exchange Rate

Commercial Demand for Dollars

Inelastic Speculative

Supply of Dollars

Elastic Speculative

Supply of Dollars

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 63

Exchange-Rate Determination

in the Medium Term

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

64 Deutsche Bank Foreign Exchange Research

Exchange rates are simultaneously determined by long-

term structural, medium-term cyclical, and short-term

speculative forces. Since long-term structural forces often

change at a glacial pace, they are likely to exert a greater

influence on a currency's long-run equilibrium path than

on a currency's short or medium-term path. A variety of

structural forces operate jointly to influence a currency's

long-run equilibrium path. These include the long-term trend

in national inflation rates, relative productivity growth, long-

term trends in net foreign asset and liability positions, per-

sistent trends in a country's terms of trade, and long-term

trends in national savings and investment.

Medium-term cyclical forces often cause a currency to ei-

ther rise or fall relative to its long-run equilibrium path. In

many cases the medium-term deviations from the long-

run equilibrium path can be quite sizeable, and in some

instances they can be quite persistent. In the long run,

these medium-term cyclical forces tend to wash out, and

once they do, there is a tendency for the exchange rate to

return to its long-run equilibrium path.

Nevertheless, in the interim, fund managers must find a

way to adjust their portfolios to account for these devia-

tions from long-run equilibrium levels or else risk serious

underperformance over medium-term horizons. Since in-

vestment performance is often evaluated over relatively

short time spans today, it is important for fund managers

to channel their energies towards getting the direction that

exchange rates take over short and medium-term horizons

right.

Medium-term trends in exchange rates tend to be influ-

enced by the direction of macroeconomic variables. The

key medium-term determinants of exchange rates include

trends in real interest-rate differentials, current and capital

account trends, relative monetary and fiscal policies, rela-

tive economic growth, and portfolio-balance considerations.

The success that these variables have had in explaining

the medium-term trends that exchange rates have taken

in the past is at best mixed. In many cases, one can find

evidence of a strong correlation between one or more of

these variables and the trend in exchange rates over sev-

eral cycles. But something often happens to break this

tight cyclical linkage. When it does, the subsequent use-

fulness of that variable as an explanatory variable becomes

seriously undermined.

Overall, we find that shifts in fiscal and monetary policies

play the most important role in driving a currency's value

on a medium-term basis, with many of the major exchange-

rate cycles in the 1990s attributable to shifts in fiscal/mon-

etary policy mixes of individual countries. We find that the

trend in the ratio of Japan's monetary base to the U.S.

monetary base has done the best job of explaining the

Japanese yen/U.S. dollar exchange rate's medium-term cy-

clical path over the past 12 years. In the case of the Deut-

schemark (and now the euro), the trend in the U.S./Ger-

man real long-term yield spread has done the best job of

explaining the medium-term cyclical paths that the Deut-

schemark/U.S. dollar exchange rate has taken over the past

25 years.

The Yen and Japanese/U.S. Monetary Policies

0.90

1.00

1.10

1.20

1.30

1.40

1.50

100

120

140

160

180

89 90 91 92 93 94 95 96 97 98 99 00 01 02

Japan/U.S. Monetary Base(ratio)----

en/US$____

Note: Y2K Adjusted,

Japan Reserve Require. Rate Change Adj. Series

Exchange-Rate Determination in the Medium Term

DM/US$ Exchange Rate and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI)

-1000

-800

-600

-400

-200

200

400

600

1.00

1.50

2.00

2.50

3.00

3.50

75 77 79 81 83 85 87 89 91 93 95 97 99 01

U.S. Less German Real Interest Rates(bp)----DM/US$____

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 65

International parity conditions are financial arbitrage con-

ditions that would exist in an ideal world. The key interna-

tional parity conditions are (1) purchasing power parity, (2)

covered interest-rate parity, (3) uncovered interest-rate

parity, (4) the Fisher effect, which focuses on inflation as

the key determinant of nominal interest rates, and (5) the

forward exchange rate as an unbiased predictor of the fu-

ture spot rate. Parity conditions show how inflation differ-

entials, interest-rate differentials, forward exchange rates,

and expected changes in exchange rates are linked inter-

nationally. Parity conditions tell us that high (low) inflation

countries should see their currencies depreciate (appreci-

ate) over time, and that forward exchange rates should

function as unbiased predictors of future spot rates. These

international parity conditions form the building blocks of

many of the major macro models of exchange-rate deter-

mination.

Empirical evidence indicates that most of the key interna-

tional parity conditions rarely hold in either the short or

medium term. There are often significant departures from

purchasing power parity, interest-rate differentials often

fail to explain future exchange-rate changes, and the for-

ward exchange rate has been found to be a poor predic-

tor of the future spot exchange rate. Why study interna-

tional parity conditions if they fail to hold? Actually, if inter-

national parity conditions held at all times, there would

be no profitable arbitrage opportunities available for inter-

national investors to exploit by moving capital from one

market to another. It is only when international parity con-

ditions fail to hold that profit opportunities from cross-

border investments become available.

Ex Ante

Purchasing Power Parity

According to

ex ante

purchasing power parity (PPP), the expected

change in the spot exchange rate should equal the difference in

expected national inflation rates.

Ex ante

PPP tells us that a high-

inflation country should see its currency depreciate and that a

low-inflation country should see its currency appreciate over time.

Covered Interest-Rate Parity

According to the covered interest-rate parity condition, an invest-

ment in a foreign-currency deposit completely hedged against

exchange-rate risk should yield exactly the same return as a com-

parable domestic-currency deposit. Since a hedged foreign-cur-

rency investment will have the same risk characteristics as a do-

mestic-currency investment, we should expect the yield on the

domestic investment, iD, to equal the foreign interest rate, iF, less

the forward discount, FD. Indeed, arbitrage should insure that

this would always be the case. The empirical evidence in support

of covered interest-rate parity is quite robust.

Uncovered Interest-Rate Parity

According to the uncovered interest-rate parity (UIP) condition,

the expected return on an uncovered foreign-currency investment

should equal the expected return on a comparable domestic-cur-

rency investment. The expected return on a domestic-currency

investment, iD, is known with certainty, while the expected return

on an uncovered foreign-currency investment, iF- ee, is not known

with certainty because the actual change in the exchange rate, e,

may turn out to be different from the expected change in the

exchange rate, ee. The UIP condition implies that investors do not

need to be compensated—in the form of a risk premium—for

taking on the risk that their expectations may prove to be wrong.

In the absence of a risk premium, the yield spread between for-

eign and domestic currency deposits, iF - iD, should adjust to equal

the expected change in the exchange rate, ee. The empirical evi-

dence is not highly supportive of UIP in either the short or me-

dium run.

e = p - p

eee

Expected Change

in Exchange Rate

Foreign-Domestic

Expected

Inflation

Differential

i - i = FD

Foreign-Domestic

Interest-Rate

Differential

Forward

Discount

e = i - i

Foreign-Domestic

Interest-Rate

Differential

Expected

Change in

Exchange Rate

International Parity Conditions—

The Building Blocks of Macro Exchange-Rate Modeling

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

66 Deutsche Bank Foreign Exchange Research

Fisher Effect

According to the Fisher effect, the nominal interest rate, i,

in a given country will equal the real interest rate, r, plus

the expected inflation rate, pe. If the real interest rate in

the foreign country is equal to the domestic real interest

rate, rF = rD, then the yield spread between two countries,

iF - iD, should equal the expected inflation differential be-

tween the two countries, pFe - pDe. Empirical evidence sug-

gests that real interest rates often differ among countries.

Thus, nominal yield spreads will not necessarily reflect dif-

ferences in national inflation rates.

Forward Rate as an Unbiased Predictor of the

Future Spot Rate

If covered interest-rate parity holds such that iF- iD=FD, and

uncovered interest-rate parity holds such that ee= iF- iD, then

the forward discount, FD, will equal the expected change

in the spot exchange rate, ee. Empirical evidence suggests,

however, that the forward exchange rate is a poor and bi-

ased predictor of the future spot exchange rate.

International Parity Conditions—

How Spot Exchange Rates, Forward Exchange Rates,

and Interest Rates Are Linked Internationally

If all of the key international parity conditions held at all

times, the expected change in the spot exchange rate

would equal (1) the forward discount, (2) the interest-rate

differential, and (3) the expected inflation differential. These

conditions will hold only in an ideal world. Empirical stud-

ies actually reject most of the parity conditions (covered

interest-rate parity being the exception). There are often

large and persistent deviations from PPP and UIP, and the

forward exchange rate has been found to be both a poor

and a biased predictor of the future spot rate. When these

parity conditions fail to hold, profitable investment oppor-

tunities become available.

e = FD

Forward

Discount

Expected

Change in

Exchange Rate

i - i = p - p

FD F D

Foreign-Domestic

Interest-Rate

Differential

Foreign-Domestic

Expected

Inflation

Differential

Expected

Change in

Exchange Rate

Foreign-Domestic

Expected

Inflation

Differential

p - p

Foreign-Domestic

Interest-Rate

Differential

i - i

Forward

Discount

e = p - p

eee

i - i = FD

e = i - i

i - i = p - p

FD F D

e = FD

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 67

Derivation of the Real Interest-Rate Differential Model

of Exchange-Rate Determination

(1) ee$ = iF - iUS Uncovered Interest-Rate Parity

(2) ee$ = qe$ - (peF - peUS)Ex Ante Purchasing Power Parity

(3) qe$ = (iF - peF ) - (iUS - peUS)(2) minus (1)

(4) qe$ = rF - rUS Real Interest-Rate Parity

(5) qe$ = q$ - q$Long-Run Exchange-Rate Adjustment

(6) q$ - q$ = rF - rUS Combine (4) and (5)

(7) q$ = q$ + (rUS - rF)RID Model

(8) q$ = q$ + (rUS - rF) - (φUS - φF)Extended RID Model

International capital flows have become increasingly mo-

bile in the past 20 years as barriers to cross-border move-

ments of capital have been gradually lifted in most mar-

kets. As cross-border capital flows have risen in importance,

interest-rate differentials have come to play a more impor-

tant role in the determination of exchange rates. Indeed,

empirical evidence has shown that relative current-account

trends have become less important in explaining exchange-

rate movements in the past two decades, while interest-

rate differentials have become more important.

To analyze the role that interest-rate differentials play in

the determination of exchange rates, many international

economists use the real interest-rate differential (RID)

model as their principal analytical tool. The RID model is a

formal model that links the trend in real exchange rates to

the trend in real interest-rate differentials.

The RID model is fairly simple to derive. If it is assumed

that both uncovered interest-rate parity (equation 1) and

ex ante purchasing power parity (equation 2) hold in the

long run, then the expected long-run change in the dollar's

real value will equal the real long-term yield spread (equa-

tions 3 and 4). If it is assumed that the expected change in

the dollar’s real value in the long run equals the difference

between the dollar’s actual real value and its long-run equi-

librium level (equation 5), it can then be shown in equation

7 that the dollar's actual real value, q$, can be expressed

as a function of two key variables—the dollar's real long-

run equilibrium value, q$, and the real long-term interest-

rate differential, rUS-rF. According to this model, the dollar's

real value would rise if there were an upward revision in

the market's assessment of the dollar's real long-run equi-

librium value and/or if there were a rise in the U.S./foreign

real long-term interest-rate differential.

The RID model can be extended to incorporate shifts in

risk premiums on U.S., φUS, and foreign assets, φF. The ex-

tended model is shown in equation 8, where a rise in the

risk premium on U.S. assets tends to exert downward pres-

sure on the dollar’s value, while a rise in the risk premium

on foreign assets tends to exert upward pressure on the

dollar’s value.

The Real Interest-Rate Differential Model

of Exchange-Rate Determination

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

68 Deutsche Bank Foreign Exchange Research

Empirical studies have found that the trend in real long-

term yield spreads does a better job of explaining exchange-

rate movements than the trend in nominal short-term yield

spreads. There are two reasons for this. First, real yields—

rather than nominal yields—better reflect structural shifts

in savings and investment and the easing/tightening moves

by central banks. Second, changes in long-term interest

rates are likely to have a more profound impact on exchange

rates than changes in short-term rates because changes

in short-term rates can be and often are transitory, while

changes in long-term rates tend to be sustained. Taking

both of these factors together, the historical evidence sup-

ports a stronger linkage between exchange rates and real

long-term yield spreads than between exchange rates and

nominal short-term yield spreads.

The accompanying diagram shows that the magnitude of

an exchange-rate response to a given change in real inter-

est-rate spreads is far greater when real long-term yield

spreads change (from q0 to q5) than when real short-term

yield spreads change (from q0 to q1). That is because ex-

change rates tend to respond more vigorously to changes

in real interest-rate spreads that are expected to persist.

The diagram also suggests that large changes in exchange

rates may result from fairly modest changes in real long-

term interest-rate differentials. The diagram illustrates as

well that in the long run, the response of an exchange rate

to a change in real yield spreads will only be transitory.

In theory, the rise in the real exchange rate in response to

the change in real yield spreads should be instantaneous,

followed by a gradual decline in the real exchange rate back

to its long-run equilibrium value as real yield spreads re-

turn to their normal levels. In practice, both the change in

real yield spreads and the change in exchange rates tend

to be much more gradual than the diagram depicts. It is

this more gradual response that gives rise to the tight

graphical relationship that one finds when comparing the

trend in the dollar's value with the trend in U.S./foreign

real long-term yield spreads.

Real Exchange-Rate Response to a Rise in Real

Short- and Long-Term Interest-Rate Differentials

U.S. Dollar’s Real Value

Time

Source: Adapted from Clark . (1994).

et al

q= q

An increase in real

interest-rate differentials

expected to last for

five years.

The Real Interest-Rate Differential Model—Theory

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 69

The trend in real long-term interest-rate differentials has

played a pivotal role in driving the dollar's value versus the

Deutschemark in the past few decades (as shown below

in the charts of the various periods).

The decline of the dollar in the late 1970s, the rise of the

dollar in the first half of the 1980s, and the dollar's decline

after 1985 can be attributed to changes in U.S./German

real long-term yield spreads.

Over the 1993-98 period, an adverse shift in real yield

spreads led to the dollar's decline in 1994-95, while the

dollar's rise in 1995-98 can be attributed to a shift in the

U.S./German real yield spread that was favorable to the

dollar.

In the 1999-2000 period, the relationship broke down when

the trend in Deutschemark (now the euro) completely di-

verged from the trend in the U.S./German (now the U.S./

European) real long-term yield spread. Although real yield

spreads moved against the U.S., the dollar nevertheless

surged against the euro, with the dollar’s gains likely at-

tributable to a rise in the dollar’s real long-run equilibrium

level.

We believe that the breakdown in the real yield spread/

exchange rate linkage in 1999-2000 will prove to be only a

temporary phenomenon. The indications are now that the

real-interest-rate-spread/exchange-rate linkage has been re-

established in 2001-02.

The Real Interest-Rate Differential Model—Evidence

DM/US$ Exchange Rate and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, 1977-1990)

-800

-600

-400

-200

200

400

600

1.00

1.50

2.00

2.50

3.00

3.50

77 78 79 80 81 82 83 84 85 86 87 88 89 90

U.S. Less German Real Interest Rates(bp)----DM/US$____

DM/US$ Exchange Rate and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, 1993-June 1999)

-300

-200

-100

100

200

1.30

1.40

1.50

1.60

1.70

1.80

1.90

2.00

1993 1994 1995 1996 1997 1998 1999

U.S. Less German Real Interest Rates(bp)----DM/US$____

DM/US$ Exchange Rate and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, July 1999-2000)

-120

-100

-80

-60

-40

-20

1.80

1.90

2.00

2.10

2.20

Jul-99 Oct-99 Jan-00 Apr-00 Jul-00

U.S. Less German Real Interest Rates(bp)----DM/US$____

DM/US$ Exchange Rate and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, Aug. 2000-2002)

-150

-100

-50

100

150

2.05

2.10

2.15

2.20

2.25

2.30

2.35

Aug-00 Nov-00 Feb-01 May-01 Aug-01 Nov-01 Feb-02

U.S. Less German Real Interest Rates(bp)----DM/US$____

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

70 Deutsche Bank Foreign Exchange Research

To many FX market participants, 2001 was a highly un-

usual year. The Federal Reserve cut short-term interest

rates from 6.5% to 1.75% and U.S./foreign short-term yield

spreads narrowed dramatically, yet the dollar remained

remarkably resilient. This has led some observers to ques-

tion whether the old rules for FX forecasting still apply.

Indeed, some observers have tried to redefine the linkage

between interest rates and exchange rates by arguing that

lower interest rates may, in fact, be positive for a currency

if they raise expectations of future economic growth in

both absolute and relative terms.

From an historical vantage point, 2001 does not appear to

be that unusual. Indeed, 2001 is not the first time that

nominal U.S. short-term interest rates and the dollar's value

have moved in opposite directions. Consider the Fed tight-

ening episodes of 1972-74, 1977-80, 1986-88, and 1994-

95. In all of those cases, the dollar actually fell, and in most

cases, quite sharply.

Nominal Short-Term Yields and the Dollar’s Value

The Dollar & U.S. Short-Term Interest Rates

(January 1972-August 1974)

2.20

2.40

2.60

2.80

3.00

3.20

3.40

1972 1973 1974

3-Month Euro-$ Rate(%)----DM/US$____

The Dollar & U.S. Short-Term Interest Rates

(January 1977-March 1980)

1.60

1.80

2.00

2.20

2.40

2.60

1977 1978 1979 1980

3-Month Euro-$ Rate(%)----DM/US$____

The Dollar & U.S. Short-Term Interest Rates

(August 1986-December 1988)

1.60

1.70

1.80

1.90

2.00

2.10

1986 1987 1988

3-Month Euro-$ Rate(%)----DM/US$____

The Dollar & U.S. Short-Term Interest Rates

(1994-1995)

1.30

1.40

1.50

1.60

1.70

1.80

1994 1995

3-Month Euro-$ Rate(%)----DM/US$____

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 71

The seemingly anomalous relationship between nominal

short-term interest rates and the dollar’s value that occurred

in 1972-74, 1977-80, 1986-88, and 1994-95 can be explained

by the trend that real long-term interest-rate differentials

took in all four episodes. In all of those cycles, real long-

term yield spreads moved in favor of Germany, which sup-

ported the case for a weaker, not a stronger, dollar.

In 2001, the same thing happened, except in reverse. Even

though nominal short-term yield spreads moved in Europe's

favor, real long-term yield spreads actually moved mod-

estly in favor of the dollar.

The conclusion one should draw from this historical analy-

sis is that the level and trend in nominal short-term yield

spreads may not be the best guide for explaining medium-

term exchange-rate movements. Rather, it is the level and

trend in real long-term yield spreads that do the best job in

terms of explaining medium-term exchange-rate move-

ments.

Real Long-Term Yield Spreads and the Dollar’s Value

The Dollar and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, 1972-1974)

-1000

-800

-600

-400

-200

200

2.20

2.40

2.60

2.80

3.00

3.20

3.40

1972 1973 1974

U.S. Less German Real Interest Rates(bp)----DM/US$____

The Dollar and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, 1977-1980)

-800

-600

-400

-200

1.60

1.80

2.00

2.20

2.40

2.60

1977 1978 1979 1980

U.S. Less German Real Interest Rates(bp)----DM/US$____

The Dollar and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, 1986-1988)

-250

-200

-150

-100

-50

1.40

1.60

1.80

2.00

2.20

2.40

2.60

1986 1987 1988

U.S. Less German Real Interest Rates(bp)----DM/US$____

The Dollar and U.S./German

Real Interest-Rate Differentials

(10-Year Bond Yields Less CPI, 1994-1995)

-250

-200

-150

-100

-50

100

150

1.30

1.40

1.50

1.60

1.70

1.80

1994 1995

U.S. Less German Real Interest Rates(bp)----DM/US$____

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

72 Deutsche Bank Foreign Exchange Research

Numerous academic papers have investigated empirically

whether the forward exchange rate has served as a reli-

able unbiased predictor of the future spot exchange rate.

The forward-rate predictor hypothesis is usually tested by

regressing the change in the exchange rate, et+k, on the

forward discount, FDt,k:

et+k = a + ß (FDt,k)

If the forward exchange rate were an unbiased predictor

of the future spot rate, estimates of the regression coeffi-

cient, beta (ß), would be equal to 1.0. If ß = 1.0, then et+k

would be expected to equal FDt,k, on average, over time.

Interestingly, the academic evidence strongly rejects the

hypothesis that the forward exchange rate is an unbiased

predictor of the future spot exchange rate, with estimates

of ß that are often significantly less than one. Indeed, ß

has often been found to be less than zero. In a survey of

75 published papers on the forward rate predictor hypoth-

esis, the average estimate of ß was found to be - 0.88.

This startling finding has the following implication for cur-

rency investment strategies. Currencies that trade at a for-

ward discount, on average, weaken less than the amount

implied by the forward discount. If anything, there has been

a tendency for such currencies to appreciate, not depreci-

ate, over time. This means that the forward discount has

been a biased predictor of future changes in the spot ex-

change rate. Hence, the term "forward discount bias." The

opposite applies for currencies that trade at a forward pre-

mium. From a strategy standpoint, if these findings on past

performance were to remain valid in the future, investors

would stand to benefit by over-weighting currencies that

trade at a forward discount and under-weighting curren-

cies that trade at a forward premium.

Various interpretations have been offered to explain these

anomalous findings. One view holds that currencies that

trade at a forward discount are more risky than those that

trade at a forward premium. Hence, investors demand a

higher risk premium to buy and hold such currencies, and

the excess return earned by investing in such currencies

merely represents the reward for accepting higher risk.

Another view holds that the market simply makes repeated

expectational errors, and that investors could exploit this

by trading against the implicit views expressed in the for-

ward exchange market.

Although excess returns could have been earned in the

past by investors who sought to exploit the "bias" in the

forward discount, the risk/return tradeoff is not always at-

tractive for such strategies. Monthly returns can be and

often are quite volatile for single-currency trades. In fact,

single-currency trading strategies have not been attractive

enough to justify risking large sums to trade the bias. (A

single-currency strategy consists of buying, or going long,

only the highest yielding currency, i.e., the currency with

the largest forward discount, and funding it by borrowing

(or going short) only the lowest yielding currency, i.e., the

currency with the largest forward premium.)

Risk-adjusted excess returns from following diversified

multi-currency approaches to exploit the bias, however,

have been more impressive. Adopting a diversified strat-

egy of going long the three highest-yielding currencies in

the industrial world and funding that position by going short

the three lowest-yielding currencies has yielded estimated

Sharpe ratios averaging 0.7-0.8 for dollar, euro, and yen-

based investors. That is nearly double the Sharpe ratios

generated by buying and holding the shares of the S&P

500 index. This basket approach to trading the forward dis-

count bias is the bedrock of Deutsche Bank's Forward-

Rate Bias trading system.

Forward-Rate Bias

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 73

Our favorite approach to trading the forward-rate bias is to

adopt a diversified strategy to exploit the fact that curren-

cies trading at a forward discount tend to outperform those

currencies trading at a forward premium. The 11 curren-

cies that we include in our trading system are the U.S.

dollar, euro, Japanese yen, Swiss franc, British pound,

Canadian dollar, Australian dollar, New Zealand dollar, Dan-

ish krone, Norwegian krone, and Swedish krona. Our in-

vestment approach is straightforward: we recommend

going long the three highest-yielding currencies in the in-

dustrial world and going short the three lowest-yielding

currencies in the industrial world. Net long/short positions

are put on at the beginning of each month and then closed

at the end of each month. This process is repeated each

month over time.

The table below shows how our monthly recommenda-

tions for long and short currency positions for the 11 in-

dustrial country currencies can vary over time. At the be-

ginning of 2001, a high-yield/low-yield currency investment

strategy would have favored being long the Norwegian

krone, New Zealand dollar, and U.S. dollar, and being short

the Swedish krona, Swiss franc, and Japanese yen. By

the end of 2001, with U.S. yields having fallen below those

in most other markets, this strategy shifted in favor of be-

ing long the Norwegian krone, New Zealand dollar, and

Australian dollar, and short the Swiss franc, U.S. dollar, and

Japanese yen.

Recommended Long and Short Positions versus the U.S. Dollar

from Adhering to Long High-Yielders/Short Low-Yielders Currency-Investment Strategy

(January 2001-April 2002)

Month AUD CAD DKK EUR JPY NZD NOK SEK CHF GBP USD

Jan-01 ----short long long short short - long

Feb-01 ----short long long short short long -

Mar-01----short long long short short long -

Apr-01----short long long short short long -

May-01----short long long short short long -

Jun-01 ----short long long - short long short

Jul-01 ----short long long - short long short

Aug-01 ----short long long - short long short

Sep-01 ----short long long - short long short

Oct-01 ----short long long - short long short

Nov-01 long ---short long long - short - short

Dec-01 long ---short long long - short - short

Jan-02 long ---short long long - short - short

Feb-02 long ---short long long - short - short

Mar-02 long ---short long long - short - short

Apr-02 long ---short long long - short - short

May-02----short long long long short - short

Note: Datastream is the source of the underlying data.

Trading the Bias

Three-Month Nominal Interest Rates

(as of December 31, 2000)

0.6

3.4

4.1

4.9

5.3

5.7

5.8

6.1

6.3

6.5

7.4

Norway N.Z. U.S. Aust. U.K.Canada

Denmark Euro

Sweden

Switzerland Japan

(%) 3-Month Euro-Deposit Rates

Three-Month Nominal Interest Rates

(as of December 31, 2001)

0.1

1.81.8

2.1

3.3

3.6

3.8

4.0

4.2

4.8

6.5

Norway N.Z. Aust. U.K.

Sweden

Denmark EuroCanada

Switzerland U.S. Japan

(%) 3-Month Euro-Deposit Rates

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

74 Deutsche Bank Foreign Exchange Research

Our analysis of the risks and returns that would have been

generated for a dollar-based investor following our diversi-

fied forward-rate bias strategy indicates that relatively at-

tractive risk-adjusted returns could have been earned in

the past 16 years. Over the 1986-2001 period, the average

annual excess return from following the forward-rate bias

strategy was around 5.6%, with an annualized standard

deviation of 7.8% and a Sharpe ratio of 0.73. This Sharpe

ratio compares favorably with the annualized Sharpe ratio

on a buy-and-hold S&P 500 equity strategy of 0.3-0.4.

At the end of January 2001, the cumulative excess return

since 1986 on the forward-rate bias strategy stood at 233%.

The distribution of monthly returns indicates that the re-

turns are skewed toward positive levels, suggesting posi-

tive investment returns were available from following this

strategy over the 1986-2001 period. High values for rolling

Sharpe ratios lend additional support to the long-run at-

tractiveness of this strategy.

Risk/Reward Structure

of Forward Discount Bias Strategy

(Going Long All High-Yield Currencies and Short All Low-Yield

Currencies Relative to the U.S. Dollar)

(January 1986-April 2002)

Annual Excess Return 5.6%

Standard Deviation 7.8%

Sharpe Ratio 0.73

Note: Datastream is the source of the underlying data.

Cumulative Excess Returns of US$-Based

Forward-Rate Bias Strategy

(1986-2002)

100

150

200

250

1986 1988 1990 1992 1994 1996 1998 2000 2002

(Index, 1986=100)

Distribution of Monthly Excess Returns

of US$-Based Forward Rate Bias Strategy

(1986-2002)

-8.2 -7.1 -6.0 -5.0 -3.9 -2.8 -1.8 -0.7 0.3 1.4 2.5 3.5 4.6 5.0

Frequency

Monthly Excess Return(%)

Sharpe Ratios of US$-Based

Forward Rate Bias Strategy

(1986-2002)

-4

-2

1986 1988 1990 1992 1994 1996 1998 2000 2002

Five-Year Rolling Sharpe Ratio----

One-Year Rolling Sharpe Ratio____

Forward-Rate Bias Strategy Long-Run Track Record for US$-Based Investors

Annual Excess Percentage Returns

Generated by the DB Forward-Rate Bias Strategy

From the Perspective of U.S. Dollar, Euro, and Yen-Based Investors

(1986-2001)

------ Investors’ Base Currency ------

Year US$ Euro Yen

(%) (%) (%)

1986 -8.4 -7.2 -7.3

1987 -0.6 -0.1 0.1

1988 8.4 8.9 8.9

1989 6.6 7.1 7.2

1990 9.0 8.9 9.3

1991 8.5 7.6 8.2

1992 5.4 4.1 5.1

1993 -4.7 -5.4 -4.5

1994 5.4 5.3 5.5

1995 8.5 9.5 10.4

1996 24.2 24.6 24.7

1997 5.2 5.7 5.8

1998 -10.7 -10.2 -9.2

1999 4.4 4.8 4.7

2000 3.2 3.3 3.5

2001 9.5 9.6 9.8

Note: Datastream is the source of the underlying data.

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 75

Empirical evidence indicates that in the very long run, the

cumulative returns on dollar-denominated and foreign-cur-

rency-denominated short-term fixed-income instruments

are broadly the same. That is, high-yield markets have seen

their currencies weaken over time, while low-yield mar-

kets have seen their currencies strengthen over time.

Indeed, consider the economic consequences if the cu-

mulative returns on a persistently high-yielding market

exceeded the cumulative returns on a persistently low-

yielding market. One market could consistently outperform

another market only if (1) its real yields were consistently

higher, which would dampen that country's long-run growth

prospects, or (2) its currency became progressively more

overvalued on a PPP basis, which would dampen the

country's long-run trade competitiveness. Eventually, a de-

terioration in growth and/or competitiveness would give

rise to corrective moves on the interest-rate and/or ex-

change-rate fronts that would erase the cumulative excess

returns. As shown below, the cumulative returns of U.S.

and foreign money-market instruments in U.S. dollar terms

were virtually the same for the 1990-1999 period.

How, then, do forward-rate bias strategies generate such

attractive risk-adjusted excess returns? The answer is that

in the industrialized world, individual markets often trade

places with other markets in terms of which market has

the highest or lowest yield. When currency

's yield rises

above currency

's yield, the widening of the

A-B

yield

spread will likely contribute to a rise in currency

's value.

When currency

's yield begins to fall relative to currency

's yield, downward pressure on currency

is likely to

materialize. The forward-rate bias strategy would gener-

ate excess returns if an investor went long currency

when

its yields rose above currency

's, and would also gener-

ate excess returns if that investor went short currency

when its yields fell below currency

's, as long as the ex-

change rate moved in sympathy with the

A-B

yield spread.

Thus, the forward-rate strategy could generate significant

excess returns, even though the cumulative returns on

currencies

and

might be the same in the long run.

World Money-Market Performance

(in US$ terms)

100

120

140

160

180

200

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000

U.S.----

(Index) Non-US____

Source: Salomon Bothers Money Market Indices

Forward-Rate Bias—The Economic Rationale

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

76 Deutsche Bank Foreign Exchange Research

The monetary, fiscal, external, and financial-sector shocks

that hit Japan in 1995-98 were all largely yen-negative. As

the yen steadily weakened, investors became increasingly

confident that the yen's path was essentially a one-way

street, which encouraged fund managers to take on ag-

gressive long-dollar/short-yen positions. As the yen weak-

ened on a trend basis for 3½ straight years, consistently

remaining above its 60-week moving average, and with

the U.S./Japan short-term yield spread averaging around

500 basis points, the "Yen Carry Trade" (borrowing in yen

to invest in dollar-denominated securities) became enor-

mously popular.

The yen carry trade generated large and persistent excess

returns over the April 1995-August 1998 period. While it is

difficult to quantify precisely how large the outstanding

positions were at their peak, BIS data on foreign-exchange

market turnover suggest that the yen carry trades were

probably quite sizable, given that the yen's share of total

foreign-exchange turnover rose substantially between 1995

and 1998. As the BIS notes, "Between 1995 and 1998, the

value of transactions involving the Japanese currency (at

constant exchange rates) had grown at twice the rate of

those involving the U.S. dollar or the Deutschemark, and

60% faster than global market turnover."

Excessive optimism or pessimism can often carry an ex-

change rate to extreme levels, and such was the case for

the yen. The popularity of the yen carry trade had pushed

the dollar into substantial overbought territory versus the

yen by the summer of 1998.

Undervalued and oversold, the yen was, in retrospect, ripe

for a corrective move upward. Although Japan's fundamen-

tal backdrop had not improved much, the existence of large

highly leveraged positions in the yen carry trade by the

speculative community raised the risk that a sudden un-

winding of those positions could drive the yen's value up

sharply. When the Russian debt crisis hit in the summer of

1998, investors around the world felt compelled to unwind

all highly leveraged fixed-income and currency positions,

including the yen carry trade.

U.S./Japanese Short-Term Interest-Rate Spread

(Three-Month Euro-Deposit Rates)

(1995-2002)

100

200

300

400

500

600

700

1995 1996 1997 1998 1999 2000 2001 2002

(Basis Points)

During the ensuing melee, the yen carry trade became a

victim of the global deleveraging process. The dollar fell by

9% versus the yen between August 31 and September 7,

1998, and then fell an additional 12% October 6-8 as highly

leveraged long-dollar/short-yen positions were aggressively

unwound. From a peak of ¥/US$ 147 on August 11, 1998,

the dollar had fallen by 34 big figures in just two months to

a low of ¥/US$ 113 on October 16, 1998.

Huge losses were incurred during the dollar sell-off of 1998,

which highlights the risks of taking on leveraged positions

in carry trades. However, from a long-run perspective, the

cumulative excess returns on the yen-carry strategy have

been fairly attractive, despite those losses. Indeed, there

have been only two periods (the shaded areas in the chart

below) when the long-dollar/short-yen carry trade has suf-

fered brief, albeit large, short-run losses since April 1995.

Note that the cumulative excess returns from following

this strategy have now surpassed those recorded just be-

fore the unwinding of the yen carry trade in 1998.

Forward-Rate Bias Strategy Opportunities and Pitfalls—

The Case of the Yen Carry Trade

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 77

Conventional wisdom holds that countries that run persis-

tent current-account surpluses will see their currencies

appreciate over time, while countries that run persistent

current-account deficits will see their currencies depreci-

ate over time.

Current-account imbalances can affect exchange rates

through a variety of channels. First, the existence of major

external imbalances can influence the flow supply of and

demand for individual currencies and thereby directly in-

fluence the paths that exchange rates take.

Second, major current-account imbalances can shift the

residence of financial wealth among deficit and surplus

nations, decreasing it in the former and raising it in the

latter. Such shifts in the residence of financial wealth could

then lead to a shift in global asset preferences, which, in

turn, could influence the paths that exchange rates take.

Third, there is a general notion that there exists some limit

on the ability of countries to run sizeable and persistent

current-account deficits, because such deficits could lead

to an unending rise in debt owed by deficit countries to

foreign investors. If foreign investors view the rise in ex-

ternal debt as unsustainable, they would reason that at

some point, a major depreciation of the deficit country's

currency might be required to ensure that the current-ac-

count deficit narrowed and that the external debt stabi-

lized at a level deemed sustainable.

In such a manner, the existence of a large current-account

imbalance will tend to alter the market's notion of what

level of the exchange rate constitutes a currency's real long-

run equilibrium value, with ever-larger deficits and external

debt levels giving rise to steady, downward revisions in

market expectations of the real long-run equilibrium ex-

change rate.

Current Account Imbalances and Exchange Rates—

The Channels of Influence

Flow

Supply and Demand

for Foreign Exchange

Exchange

Rate

Transfer of Wealth

from Deficit

to Surplus Countries

External

Debt

Sustainability

Current

Account

Imbalance

Current-Account Imbalances and the Determination of Exchange Rates

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

78 Deutsche Bank Foreign Exchange Research

Most market participants view U.S. current-account trends

as important for the dollar because such trends directly

influence the flow supply of and demand for dollars in the

foreign-exchange market. Using the traditional balance-of-

payments flow model of exchange-rate determination, we

can describe how current-account flows determine the

supply of and demand for dollars.

Let's assume that the flow demand for dollars, D$, is gen-

erated by foreign demand for U.S. goods and services,

while the flow supply of dollars, S$, is generated by U.S.

demand for foreign goods and services. If current-account

flows determine the supply of and demand for dollars, then

it should be the case that when the U.S. current-account

balance is in equilibrium, the supply of and demand for

dollars will be in equilibrium as well. In terms of the dia-

gram below, the dollar's equilibrium value, q0, would be

determined by the intersection of the supply, S$, and de-

mand, D$, curves for dollars.

Balance-of-Trade Flow Model

of Exchange-Rate Determination

Dollar’s Value

Quantity of

Dollars

If, however, the U.S. were running a current-account defi-

cit, perhaps because the value of the dollar was too high

at q1 relative to its equilibrium value, q0, then the supply of

dollars, S$, on the foreign-exchange market would exceed

the demand for dollars, D$. The excess supply of dollars,

which is the foreign-exchange market counterpart of the

U.S. current-account deficit, is represented as the gap be-

tween points

and

. The dollar's value at q1 would clearly

not be sustainable since the excess supply of dollars,

on the foreign-exchange market would exert continual

downward pressure on the dollar's value. Once the dollar's

value falls toward its long-run equilibrium value, q0, the

supply of dollars would be brought back into line with the

demand for dollars, and the U.S. current account would be

brought back into balance at point

Balance of Payments Flow Model of Exchange-Rate Determination

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 79

The traditional balance-of-payments flow model posits that

exchange rates will adjust until all current-account imbal-

ances are corrected and current-account equilibrium is re-

stored. How much exchange rates must adjust to restore

current-account equilibrium depends on the elasticities of

the supply and demand curves for foreign exchange.

In the accompanying diagram, we plot two sets of supply

and demand curves for dollars—one set that is highly elas-

tic (Se$ and De$, respectively) and another set that is highly

inelastic (Si$ and Di$, respectively). Initially, it is assumed

that the U.S. is running a current-account deficit equal to

at the prevailing exchange rate, q1. The exchange-rate

change necessary to correct the U.S. current-account im-

balance will differ depending on whether the prevailing

supply and demand curves for dollars are highly elastic or

inelastic.

Elasticities of Supply and Demand

for Dollars and

Exchange-Rate Determination

Exchange Rate

Quantity of

Dollars

C´

If the supply and demand curves for dollars were highly

elastic, only a modest decline in the dollar's value to q0

would be required to restore equilibrium to the U.S. cur-

rent account. If, instead, the supply and demand curves

for dollars were highly inelastic, a large decline in the dol-

lar to q2 would be required to restore equilibrium to the

U.S. current account.

U.S. export and import price elasticities are estimated to

be quite small, perhaps averaging only around -1.0 for ex-

ports and -0.3 for imports. Based on these rather low elas-

ticities, it would likely require a rather large depreciation of

the dollar to restore a sizeable U.S. current-account deficit

to a more sustainable level. Thus, the inelastic supply and

demand curves shown below may be a more realistic pic-

ture of the potentially large dollar adjustment that might

be required to restore equilibrium to the U.S. current-ac-

count balance.

Trade Elasticities and the Exchange-Rate/Trade-Balance

Adjustment Process

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

80 Deutsche Bank Foreign Exchange Research

The dramatic appreciation of the dollar over the 1995-2001

period despite a record deterioration of the U.S. current

account is clear evidence that a current-account deficit need

not be bearish for a currency. Indeed, the same thing hap-

pened in the first half of the 1980s when the dollar soared,

yet the current-account deficit steadily widened.

Current-account imbalances are typically driven by struc-

tural changes in international competitiveness, changes in

the terms of trade, and long-term shifts in national sav-

ings-investment (S-I) balances. Consider the impact on

exchange rates and the current account if a country under-

went a major investment boom or undertook a long-term

policy of fiscal expansion. In both cases, the S-I schedule

would shift upward to the left. The increase in investment

relative to national savings would result in a wider current-

account deficit and, at the same time, put upward pres-

sure on the real exchange rate. That is pretty much what

happened to the U.S. savings-investment balance, the cur-

rent-account balance, and the dollar in both the 1980-85

and 1995-2001 episodes. Hence, when the source of the

deterioration in the current-account balance is a rise in in-

vestment relative to national savings, there is a high likeli-

hood that the currency will strengthen over time, not

weaken.

Current-Account/Exchange-Rate Linkages

and the Impact of Economic and Financial Disturbances

Economic Current Exchange

Conditions Account Rate Linkage

Expansionary Monetary Policy Deterioration Depreciation Positive

Restrictive Monetary Policy Improvement Appreciation Positive

Expansionary Fiscal Policy Deterioration Appreciation Negative

Restrictive Fiscal Policy Improvement Depreciation Negative

Exogenous Increase Improvement Appreciation Positive

in Domestic Supply

Exogenous Increase Improvement Appreciation Positive

in Foreign Demand

U.S. Savings-Investment Balance

and the Dollar

Savings minus Investment,

Current-Account Balance

U.S. Dollar

Real Exchange Rate

“New Economy”

Increase in

Investment

Spending

S-I

CAB

“New Economy”

Upward Revision in the

Dollar’s Real Long-Run

Equilibrium Value

(0)(-) (+)

S-I

If the deterioration in a country's current-account balance

were instead due to a long-run deterioration in that

country's competitiveness, the CAB schedule would shift

downward to the left relative to an unchanged S-I sched-

ule. In such a case, the deterioration of the current ac-

count would coincide with a long-term slide in the domes-

tic currency’s real long-run equilibrium value.

The message from all this is that certain economic distur-

bances may cause exchange rates and the current-account

balance to move in the same direction while other distur-

bances may cause exchange rates and the current account

to diverge. Consider the current-account/exchange-rate link-

age that arises in response to fiscal and monetary policy

changes. A country that pursues a highly expansionary

monetary policy will likely see its current account deterio-

rate and its currency weaken at the same time. The ques-

tion is whether the weakening of the currency is due to

the easier monetary policy or the current-account deterio-

ration or possibly both. In the case of a highly expansion-

ary fiscal policy, the trend in the current account and the

trend in the exchange rate may diverge as the current ac-

count deteriorates while the currency appreciates.

Are Current-Account Deficits Necessarily Bearish for Exchange Rates?

U.S. Current-Account Deficit and the Dollar

Savings minus Investment,

Current-Account Balance

U.S. Dollar

Real Exchange Rate

S-I

CAB

Downward Revision in the

Dollar’s Real Long-Run

Equilibrium Value

Long-Run Deterioration in

U.S. Trade Competitiveness

(0)(-) (+)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 81

Japan's tendency to run large and persistent current-ac-

count surpluses over time largely explains why the yen's

trade-weighted value has risen on a secular basis. From a

simple flow supply-and-demand perspective, one would

expect larger and larger surpluses to give rise to succes-

sive increases in the demand for yen over time. Indeed,

the rising trend in Japan's current-account surplus has

moved in sympathy with the rising trend in the yen's value

in the past 30 years.

Related to this is the fact that Japan's cumulative current-

account surpluses have allowed Japanese investors to

acquire a rising claim on U.S. assets. To induce Japanese

investors to acquire and hold onto those rising dollar claims,

either U.S. interest rates needed to rise relative to those in

Japan, or the dollar needed to fall versus the yen to make

dollar assets appear cheaper to Japanese investors, or

some combination of the two was required. The evidence

indicates the dollar bore the brunt of this adjustment. The

accompanying chart shows that the yen’s long-term trend

moved sympathetically with the ratio of Japan's interna-

tional investment position relative to the U.S. international

investment position, as a percentage of their respective

GDPs.

In addition to the long-term financial pressure on yen ema-

nating from Japan's persistent current-account surpluses,

Professor Ronald I. McKinnon of Stanford University makes

the case that commercial tensions between the U.S. and

Japan, including intermittent threats of an outright trade

war, may have played an important role in driving the yen's

real value higher over time. According to McKinnon, the

U.S.—concerned about Japan's large and rising bilateral

trade surplus and frustrated by the inability of U.S. firms to

penetrate Japan's relatively closed markets—pursued a

policy of coupling protectionist threats with demands, im-

plicit or explicit, for yen appreciation. With the yen the fo-

cus of both U.S. trade and exchange-rate policies, the Japa-

nese authorities had to either tolerate an ever-rising yen or

accept the wrath of American protectionists.

Repeated attempts to talk the dollar down versus the yen

created, according to McKinnon, a syndrome of an ever-

rising yen that had a pervasive influence on investor ex-

pectations in Japan. Concerned about possible large losses

on their dollar-based investments, brought about by re-

peated attempts by U.S. officials to talk the dollar down,

Japanese investors became increasingly reluctant to hold

too large a proportion of their total portfolio in foreign as-

sets. This meant that Japanese funds would effectively be

bottled up inside Japan, with no safe-passageway for

Japan's current-account surpluses to exit through. This

contributed to an ever-rising basic balance of payments

surplus in Japan (the current account plus the long-term

capital account). As shown in the accompanying chart, the

tendency of Japan's basic balance of payments to regis-

ter a persistent surplus over time played a key role in driv-

ing the yen higher over time.

The Japanese Yen and the U.S./Japan

Cumulative Current-Account/GDP Differential

(50)

(40)

(30)

(20)

(10)

100

150

200

250

300

77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96

Cum. Curr.-Acct. as % of GDP Diff.---- (%)

/US$____

The Yen and Japan’s Current-Account Balance

-10

-5

(350)

(300)

(250)

(200)

(150)

(100)

(50)

76 78 80 82 84 86 88 90 92 94 96 98 00 02

Current Acc’t(¥ tr.)----

en/US$(reverse scale)____

The Yen and Japan’s Basic Balance

-10

-5

100

120

140

160

180

86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

Basic Balance(12-mo. moving sum)(¥ tr.)----

en Index____

Japan’s Current-Account Balance and the Yen

Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

82 Deutsche Bank Foreign Exchange Research

There are essentially three ways that the U.S. can narrow

its large current-account deficit. First, the U.S. could pur-

sue policies that encourage domestic savings and/or dis-

courage domestic investment, perhaps via a tighter fiscal

and/or tighter monetary policy. Such policies would con-

tribute to a leftward shift of the savings-investment (S-I)

schedule in the diagram on the right. As illustrated, U.S.

domestic demand would need to be pared back, perhaps

significantly, to restore sustainable balance to the U.S. ex-

ternal account.

A second way to eliminate the unsustainable portion of

the U.S. current-account deficit would be to engineer a

major depreciation of the dollar to improve U.S. competi-

tiveness, and thereby boost net exports. A major dollar

depreciation would shift the export-import (Ex-Im) sched-

ule to the right to Ex-Im1 and would to narrow the deficit

from point

A’

to a more sustainable level such as point

B’

An even larger dollar depreciation would be needed to shift

the export-import curve to Ex-Im2 to narrow the deficit to

point

C’

The problem with this strategy is that it might create other

problems for U.S. policymakers, and in all likelihood, might

fail completely. That is because if the U.S. economy were

already operating at its full employment level of output Y'F,

then a dollar depreciation that shifts the export-import curve

to the right would drive the level of U.S. output past its full

employment potential. Since this would be highly inflation-

ary, the Federal Reserve would need to tighten monetary

policy, perhaps significantly. But if the Federal Reserve

were forced to push U.S. interest rates up sharply, that

would attract capital inflows to the U.S., which would then

strengthen the dollar, thereby offsetting, perhaps com-

pletely, the initial depreciation of the dollar.

A third way to eliminate the unsustainable portion of the

U.S. current-account deficit would be to combine belt-tight-

ening and dollar-depreciation initiatives. A modest belt-tight-

ening move on the fiscal and monetary policy front could

shift the S-I curve to the left to S-I3, while a modest dollar

depreciation could work to shift the export-import curve to

the right to Ex-Im3. The combination of these two policy

actions would leave the equilibrium level of output un-

changed while narrowing the U.S. current-account deficit

to a sustainable level.

Correcting the U.S. Current-Account Deficit—

Belt Tightening vs. Dollar Depreciation

Correcting the U.S. Current-Account Imbalance

via a Weakening of the Dollar

Output

Saving-Investment,

Exports-Imports

B’

C’

S-I

Ex-Im

A’

Y’

Y’’ Y’’’

Ex-Im

Correcting the U.S. Current-Account Imbalance

via Tighter Fiscal/Monetary Policies

plus a Weakening of the Dollar

Output

Saving-Investment,

Exports-Imports

C’

S-I

Ex-Im

A’

Y’

Ex-Im

Correcting the U.S. Current-Account Imbalance

via Tighter Fiscal and Monetary Policies

Output

Saving-Investment,

Exports-Imports

B’

C’

S-I

Ex-Im

A’

Y’

Y’’Y’’’

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 83

Many pundits have argued that the record deterioration of

the U.S. current-account deficit will inevitably lead to a hard

landing for the dollar. So far it has not.

One possible explanation for the dollar’s resiliency is that

the current-account deterioration might be an

equilibrium

and not a

disequilibrium

phenomenon. If one views U.S.

economic developments from a "New Economy" perspec-

tive, the revolutionary changes in information technology

may have raised both the speed limit on sustainable U.S.

GDP growth and the sustainable U.S. current-account defi-

cit.

With the return on capital on new technology investments

highly attractive, foreign investors have been more than

willing to finance a larger-than-normal gap between U.S.

savings and investment. As a result, the inflow of foreign

capital into the U.S. has exceeded the typical flow of capi-

tal that found its way to the U.S. in the past. Since, by

definition, a country's current-account deficit equals its

capital-account surplus, which also equals its savings-in-

vestment gap, it would appear that the information tech-

nology revolution in the U.S. simultaneously raised both

the speed limit on sustainable U.S. growth and the size of

the sustainable current-account deficit that the U.S. can

safely run without triggering a dollar crisis. In fact, the in-

formation technology revolution, which is presently domi-

nated by U.S. firms, may have lifted the dollar's real long-

run equilibrium level as well.

While it might be true that the size of the sustainable cur-

rent-account deficit has risen, there nevertheless must exist

some threshold level for the U.S. current-account deficit

that would trigger a decline in the dollar’s value. The key

questions for investors then are: (1) what might the new

higher threshold level be, and (2) how soon will it be be-

fore the U.S. current-account deficit crosses that thresh-

old and begins to take its toll on the dollar?

U.S. Dollar Index

100

105

110

115

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

US$ Major Currency Index

U.S. Current-Account Balance

-500

-400

-300

-200

-100

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

(US$ bn.) Annualized

U.S. Current-Account Deficit and the Dollar

Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

84 Deutsche Bank Foreign Exchange Research

Based on an analysis of 17 episodes in industrial countries

over the 1980-95 period, Catherine L. Mann (1999) found

that current-account deficits reaching 4.2% of GDP tended

to set forces in motion—including corrective currency ad-

justments—for a reversal of the current-account shortfall.

The U.S. current-account deficit as a percentage of GDP is

presently at 4.1%. If the U.S. economy recovers strongly

over the course of 2002, the U.S. current-account deficit

could widen to as much as 5% of GDP, well above the

4.2% average threshold.

Another study (Caroline L. Freund, 2000) examined a larger

sample of industrial country current-account adjustment

episodes and found that the median current-account defi-

cit/GDP ratio that triggered corrective adjustments in the

current account was around 4.9%. Those results are re-

ported in the table below.

As shown by the median current-account adjustment pro-

cess, an industrial-country deficit of 4.9% of GDP has, in

the past, led to a 19% real depreciation of the deficit

country's currency over a three-year period, with the de-

cline beginning one year prior to the peak in the current-

account deficit and ending about three years after the peak.

In conjunction with a slowdown in domestic demand, the

decline of the currency has typically helped bring about an

83% decline in the current-account imbalance. If the U.S.

conforms to the median episode, and the U.S. current-ac-

count deficit/GDP ratio rises to 5% by year-end 2002, then

the corrective downward adjustment in the dollar's value

(which typically occurs one year prior to the peak in the

deficit) should have started in early 2002.

Estimating the Threshold Level for the U.S. Current-Account Deficit

That Would Trigger a Decline in the Dollar’s Value

Current-Account Adjustment and Real Exchange-Rate Depreciation

(in 25 Episodes of Large Current-Account Adjustments, 1980-1995)

Relative to Peak Deficit Year (t0),

Current Account/GDP Ratio Three-Year Real the Year in which

at Peak 3 Years after Peak Percentage Exchange-Rate Exchange-Rate Depreciation

Date (t0) (t

3) Change Depreciation Began Ended Duration

Australia (1989) -6.2% -3.7% 40% 21% 1 5 4

Austria (1980) -4.9% 0.4% 108% 19% -1 2 3

Belgium (1981) -4.2% -0.1% 98% 51% -3 4 7

Canada (1981) -4.2% -0.4% 91% ----

Canada (1993) -3.9% 0.6% 114% 19% -1 3 4

Denmark (1986) -5.3% -1.0% 81% ----

Finland (1991) -5.5% 1.3% 124% 30% -1 2 3

France (1982) -2.2% 0.0% 100% 15% -1 3 4

Greece (1985) -8.1% -1.5% 82% 21% -2 2 4

Hong Kong (1980) -5.0% 0.6% 111% ----

Ireland (1981) -13.6% -5.6% 59% 41% -2 1 3

Israel (1982) -8.7% 4.8% 155% 14% 2 6 4

Italy (1981) -2.6% -0.8% 70% 12% -1 1 2

Italy (1992) -2.5% 2.3% 194% 28% 0 4 4

Korea (1980) -8.5% -1.8% 78% ----

New Zealand (1984) -13.6% -8.0% 41% 11% -1 1 2

Norway (1986) -6.0% 0.2% 104% 5% -3 1 4

Portugal (1981) -16.8% -2.8% 83% 8% 1 3 2

Singapore (1980) -13.3% -3.5% 74% ----

Spain (1981) -2.9% 1.1% 139% 36% -1 3 4

Spain (1981) -3.8% -1.4% 62% 17% 1 4 3

Sweden (1980) -3.5% -0.8% 77% 20% 1 4 3

Sweden (1992) -3.6% 2.1% 160% 21% 1 4 3

United Kingdom (1989) -4.4% -1.8% 60% 15% 3 5 2

United States (1987) -3.7% -1.7% 56% 34% -1 2 3

Median -4.9% -0.8% 83% 19% -1 3 3

Source: Caroline L. Freund, “Current Account Adjustment in Industrial Countries", Federal Reserve Board, International Finance Discussion Paper, No. 692, December 2000.

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 85

After the experiences of the early 1980s and the late 1990s,

when the dollar soared despite the significant widening of

the U.S. current-account deficit, one could make the case

that the U.S. current-account deficit might be overrated as

a key determinant of the dollar. In fact, one could question

whether the measurement problems associated with cur-

rent-account data makes them a reliable measure of U.S.

trade competitiveness.

The reported current-account data omit the contribution

that U.S. foreign affiliates make to the aggregate sales by

U.S. firms in international markets. When allowance is

made for the role played by U.S. foreign affiliates, one finds

that U.S. firms' penetration of foreign markets actually ex-

ceeds the penetration of foreign firms in the U.S. market.

The U.S. Commerce Department reports that worldwide

sales by U.S. companies to foreign residents (which in-

clude U.S. exports plus total sales by U.S. companies' for-

eign affiliates to non-residents) amounted to $3.17 trillion

in 1998 (latest data available), while total sales by foreign

companies to U.S. residents (which include U.S. imports

plus total sales by foreign firms' U.S. affiliates to U.S. resi-

dents) amounted to $2.81 trillion in the same period. That

means that worldwide sales by U.S. companies to foreign

residents exceeded the sales of foreign firms to U.S. resi-

dents by $363 billion. Breaking the data down, the Com-

merce Department notes that of the $3.17 trillion in U.S.

firms' worldwide sales, $930 billion represented U.S. ex-

ports of goods and services, while $2.24 trillion repre-

sented the sales by U.S. firms' foreign affiliates. Of the

$2.81 trillion in foreign companies' sales to U.S. residents,

$1.1 trillion represented U.S. imports of foreign goods and

services, while $1.7 trillion represented the sales by for-

eign firms' U.S. affiliates.

What is clear from the data is that although U.S. imports

($1.1 trillion) exceeded U.S. exports ($930 billion), leaving

a wide U.S. trade deficit of $167 billion, the sales of U.S.

firms' foreign affiliates ($2.24 trillion) far exceeded the sales

of foreign firms' U.S. affiliates ($1.71 trillion), leaving a U.S.

surplus on foreign affiliate operations of $530 billion.

Overall, despite the fact that the U.S. runs a large trade

deficit, U.S. firms have had greater success in terms of

penetrating foreign markets than foreign firms have had in

penetrating the U.S. market. The data show that, at the

margin, U.S. firms tend to accommodate foreign demand

through their overseas affiliates (71% of worldwide sales)

rather than through outright exports from the U.S. (only

29% of worldwide sales). In contrast, foreign firms, at the

margin, tend to rely more on exports to the U.S. (39% of

their U.S. sales) and a bit less on their U.S.-based affiliates

(61% of their U.S. sales) to accommodate demand in the

U.S. market.

What does this all imply for the U.S. current-account

deficit's impact on the dollar? Does the dollar need to de-

cline in order to balance U.S. exports and imports? Does

the dollar have to be penalized because U.S. firms tend to

accommodate foreign demand more through their over-

seas affiliates than via direct exports, particularly given the

fact that U.S. firms do a better job in the aggregate than

their foreign counterparts in terms of penetrating the

other's market? Or should the dollar be rewarded to re-

flect the global reach and strength that U.S. firms exhibit

in the market?

Worldwide Sales by U.S. Companies versus

Foreign Companies’ Sales to the U.S.

(Latest Data, 1998)

(US$ Billions)

Worldwide Sales Sales by Foreign

by U.S. Firms Firms to the U.S. Difference

Total $3,173 $2,810 +$363

Sales by

Foreign Affiliates $2,240 (71%) $1,710 (61%) +$530

Exports $933 (29%) $1,100 (39%) -$167

Source: U.S. Commerce Department

Is the U.S. Current-Account Balance a Reliable Measure of

U.S. Trade Competitiveness?

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

86 Deutsche Bank Foreign Exchange Research

Greater financial integration of the world's capital markets

and the increased freedom of capital to flow across na-

tional borders have increased the importance of financial

flows in the determination of exchange rates. The rise in

the dollar's value in recent years was largely driven by

strong increases in the demand for U.S. assets by over-

seas investors. For example, Euroland investors had been

structurally underweight equities for a long time, but be-

ginning in the late 1990s they started to shift their portfo-

lio allocations away from bonds and toward equities. Within

their rising equity holdings, European investors also made

the decision to increase the percentage allocated to U.S.

equities and reduce the percentage allocated to Euroland

equities. Overall, there was a marked rise in European in-

vestors' demand for U.S. equities in 1999-2001 that helped

lift the dollar versus the euro.

Foreign demand for Japanese equities has also played a

role in influencing the medium-term direction of the yen.

The yen's weakness over the 1995-98 period coincided

with declining foreign demand for Japanese equities, and

the yen's subsequent rebound in late 1998 and in 1999

coincided with a rise in foreign demand for Japanese equi-

ties. The drop-off in foreign demand for Japanese equities

in 2000-01 coincided with a sharp decline in the yen's value.

Foreign direct investment (FDI) and portfolio capital flows

were highly positive for the dollar and negative for the euro

in the past few years. As shown in the table below,

Euroland's overall balance of payments was in huge defi-

cit—because of sizable FDI and portfolio outflows from

Euroland—which is in stark contrast to the U.S. and Japa-

nese overall balance of payments surpluses.

Although the U.S. has seen its current-account deficit

double between 1998 ($217 billion) and 2000 ($435 bil-

lion), the U.S. over the same period was the recipient of

massive FDI and portfolio inflows, which exceeded the size

of the U.S. current-account deficit. As a result, the U.S.

overall balance of payments swung from a deficit of $43

billion in 1998 to a surplus of $52 billion in 2000.

In contrast, Euroland saw its current account swing from

a surplus to a deficit position, while FDI and portfolio flows

recorded large outflows as well. As a result, Euroland ran

overall balance of payments deficits averaging $173-$183

billion per annum in 1998-2000, which weighed heavily on

the euro during this period.

European Investor Demand for U.S. Equities

and the U.S. Dollar/Euro Exchange Rate

(12-Month Moving Sum of Net Equity Purchases)

(1.50)

(1.40)

(1.30)

(1.20)

(1.10)

(1.00)

(0.90)

(0.80)

-50

100

150

200

1995 1996 1997 1998 1999 2000 2001 2002

US$/Euro(reverse scale)----European Net Purchases(US$ bn.)____

The Japanese Yen and

Foreign Investor Demand for Japanese Equities

(12-Month Moving Sum of Net Equity Purchases)

-20

100

120

140

100

110

120

130

140

150

160

170

180

1995 1996 1997 1998 1999 2000 2001 2002

Foreign Net Purchases(US$ bn.)----Yen Effective Rate____

U.S., Euroland, and Japanese Balance of Payments

(1998-2000)

(US$ Billions)

Current Account FDI and Portfolio Flows Overall Balance of Payments

1998 1999 2000 1998 1999 2000 1998 1999 2000

U.S. -217 -331 -435 174 338 487 -43 7 52

Euroland 35 -7 -32 -218 -166 -144 -183 -173 -176

Japan 120 109 118 -63 -36 -60 57 73 58

Source: DatastreamSource: Datastream

Capital Flows and Exchange Rates

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 87

Despite all of the attention that capital flows receive in the

FX market, there has not been much rigorous empirical

testing to determine whether these flows have a statisti-

cally significant and quantitatively important impact on

exchange rates, at least until lately. The International Mon-

etary Fund recently published a report examining the im-

pact that capital flows have had on the euro and the yen in

the past decade. The IMF examined the trend in capital

flows and other relevant indicators to determine whether

they had statistically significant and quantitatively impor-

tant effects on exchange rates. The IMF's findings (based

on an analysis of exchange rates over the 1988-2000 pe-

riod) can be summarized as follows:

(1) "There is evidence that equity flows matter for the euro-

dollar rate, but not for the yen-dollar rate."

(2)Relative expected current and future growth rates de-

rived from Consensus Economic Forecasts are cor-

rectly signed and statistically significant in explaining

the euro, but not the yen.

(3)Relative equity returns are not statistically significant

in explaining either the euro or the yen.

(4)Bilateral current-account trends vis-à-vis the U.S. are

not statistically significant in explaining either the

euro's or the yen's trend. However, the IMF notes that

when multilateral, and not bilateral, current-account

positions were investigated, there was some evidence

of a positive effect of current-account trends on ex-

change rates.

(5)"There is little evidence that merger and acquisition

flows are important for exchange-rate determination.

The coefficients on net foreign direct-investment (FDI)

flows are correctly signed, but statistically insignifi-

cant for both the euro and yen." These results are con-

sistent with the view that "the majority of cross-bor-

der merger and acquisition flows are financed through

share-swaps that have no immediate impact on the

demand for currencies."

(6)"Net bond flows appear to have no significant effect

on the euro- or yen-dollar rate."

(7) "The movements of the euro-dollar and yen-dollar ex-

change rates are significantly correlated with the long-

term interest-rate differential, but not their short-term

equivalent." The IMF did note, however, that in the case

of the euro, the explanatory power of the long-term

spread is weaker than that for equity flows.

IMF Analysis of Factors That Explain Movements

in the Euro and the Yen versus the U.S. Dollar

(1988-2000)

Euro Yen

Coefficient Coefficient

Current Account & Capital Flows

Current Account Insignificant Insignificant

Capital Account

Net Bond Flows Insignificant Incorrect Sign

Net Equities Flows Significant Insignificant

Foreign Direct Investment Insignificant Insignificant

Traditional Underlying Factors

Long-Term Interest Differential Significant Significant

(Marginally)

Short-Term Interest Differential Insignificant Insignificant

Relative Current Growth Insignificant Insignificant

Alternative Underlying Factors

Relative Stock Returns Insignificant Insignificant

Relative Expected Growth Significant Insignificant

Note: Coefficients of regression of change in the logarithm of the U.S. dollar exchange

rate on a constant and the contemporaneous value of the explanatory variable using

quarterly data since 1988. Significant denotes that the variable was both statistically

significant and had the correct sign. All other variables were not statistically significant.

Source: Robin J. Brooks, Hali Edison, Manmohan S. Kumar, Torsten M. Slok,

"Exchange Rates and Capital Flows", International Monetary Fund,

Research

Department Series, Working Paper,

No. 01/190, May 2000, available on the IMF's

Web site (www.imf.org).

The IMF concluded that its findings were "supportive of

the new conventional wisdom that net equity flows are

important for the euro-dollar exchange rate, although long-

term interest-rate differentials also appear to matter. For

the yen, however, there is little evidence that net equity

flows have been an important determinant of the bilateral

exchange rate with the U.S. dollar." The IMF noted that the

importance of equity flows for the euro might stem from

the fact that European investors might have been diversi-

fying their portfolios away from intra-European investments

to U.S. equity investments. The IMF warned that if such

portfolio adjustments are not yet complete, then the euro

could remain weak for some time until European portfo-

lios are completely rebalanced.

Capital Flows and Exchange Rates: Empirical Evidence

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

88 Deutsche Bank Foreign Exchange Research

Correlation Between Equity Market Trends

and Exchange Rates

(Daily, Weekly, Monthly, and Quarterly Data)

(January 1983-May 2000)

Equity Market/ Correlation Coefficient

Exchange Rate Daily Weekly Monthly Quarterly

Dow Jones/

US$ Index 0.03 0.04 0.04 0.11

Dow Jones-Nikkei/

Yen/US$ 0.04* 0.07* 0.08 0.11

Dow Jones-DAX/

DM/US$ 0.17** -0.15** -0.25** -0.22**

* Statistically significant at 5% level

** Statistically significant at 1% level

Source: BIS Quarterly Review, August 2000

Correlation Between International Portfolio

Equity Flows and Exchange Rates

(Monthly Data)

(January 1983-May 2000)

Equity Market Flows/

Exchange Rate Correlation Coefficient

U.S.-Japan Equity Flows/

Yen/US$ 0.04

U.S.-European Flows/

DM/US$ 0.05

Japan-European Flows/

Yen/DM -0.26

Source: BIS Quarterly Review, August 2000

Given the surge in the U.S. equity market in the second

half of the 1990s, many observers have naturally associ-

ated the rise in the dollar's value over this period with the

increase in U.S. equity values. Appearances, however, can

be deceiving. As the Bank of International Settlements

reports in its August 2000

BIS Quarterly Review

, "There is

little evidence of a robust significant correlation between

stock market indices and the major exchange rates." The

BIS finds that the "correlation coefficient for the Dow Jones

index and the nominal effective dollar rate is positive but

very small and not statistically significant."

The BIS also reports that "a similar result holds for the co-

movement of the return on the Dow Jones relative to the

Nikkei and the percentage change of the bilateral yen/dol-

lar rate." What was particularly surprising regarding the BIS

findings was that "the correlation between the return on

the Dow Jones relative to Germany's DAX equity index

and movements in the Deutschemark/U.S. dollar exchange

rate was negative and statistically significant." In other

words, when the U.S. equity market has risen relative to

the German equity market, the dollar has tended to fall,

not rise in value versus the Deutschemark, as conventional

reasoning would have suggested.

Not only did the BIS find a weak relationship between eq-

uity indices and exchange-rate movements, it also found a

weak relationship between international portfolio equity

flows and the trend in exchange rates. The correlation be-

tween U.S./Japan equity flows and the Japanese yen/U.S.

dollar exchange rate was close to zero, as was the correla-

tion between U.S./European equity flows and the Deut-

schemark/U.S. dollar exchange rate. Contrary to conven-

tional wisdom, the BIS found that the yen tended to

strengthen versus the Deutschemark when equity flows

moved in favor of Europe, although the reported correla-

tion coefficient was not statistically significant.

Given the historically weak relationship between equity-

market trends and exchange rates, why has the foreign-

exchange market become so fixated on equity-market de-

velopments? One reason may be that in the late 1990s

the correlation between the Dow Jones Industrial Aver-

age and the dollar was in fact highly positively correlated

(even though the long-run correlation between the Dow

Jones index and the dollar has averaged close to zero).

The high correlation in the late 1990s may have led market

observers to believe that this trend would continue into

the future. This proved not to be the case over the 2000-01

period, however, as the correlation between the Dow Jones

index and the dollar plummeted.

Correlation of U.S. Equity Prices & the Dollar

(Rolling Six-Mo. Correlation of Log Differences

of Daily Dow Jones Ind. & US$ Trade Wgt Indices)

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

Correlation Coefficient (Month-end Values)

Equity-Market Trends and the Dollar

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 89

Although exchange rates and equity markets will move in

sync from time to time, the relationship between equity-

market trends and exchange rates is not stable. For in-

stance, between 1990 and 1995, the U.S. dollar fell de-

spite a strongly performing U.S. equity market, while the

Japanese yen soared despite a weakly performing equity

market. In both cases the correlation between local eq-

uity-market performances and the local currency's value

was negative. In contrast, between 1995 and 2001, the

correlation between local equity-market performances and

the local currency's value turned highly positive. The U.S.

dollar soared in tandem with a rising U.S. equity market,

at least up until late 1999, while the yen weakened in tan-

dem with a trend decline in the Japanese equity market.

Such instability makes it difficult to form judgments on pos-

sible future currency moves based solely on expected eq-

uity-market performances.

The Dollar and U.S. Equity Prices

(1990-1995)

250

300

350

400

450

500

550

600

650

1990 1991 1992 1993 1994 1995

S&P 500 Index----US$ Index____

The Yen and Japanese Equity Prices

(1990-1995)

1000

1500

2000

2500

3000

(180)

(160)

(140)

(120)

(100)

(80)

(60)

1990 1991 1992 1993 1994 1995

TOPIX Index----

en/US$(reverse scale)____

The Yen and Japanese Equity Prices

(1996-2002)

800

1000

1200

1400

1600

1800

(150)

(140)

(130)

(120)

(110)

(100)

1996 1997 1998 1999 2000 2001 2002

TOPIX Index----

en/US$(reverse scale)____

The Dollar and U.S. Equity Prices

(1995-2002)

400

600

800

1000

1200

1400

1600

100

110

120

1995 1996 1997 1998 1999 2000 2001 2002

S&P 500 Index----US$ Index____

Source: Datastream Source: Datastream

Equity-Market Trends and Exchange Rates—An Unstable Relationship

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

90 Deutsche Bank Foreign Exchange Research

The market's recent fixation with foreign direct investment

flows as an explanatory variable in projecting future ex-

change-rate movements is a phenomenon that derives

largely from recent large acquisitions of U.S. firms by Eu-

ropean corporates and the surge in the dollar that appeared

to accompany those acquisitions. But the relationship be-

tween FDI flows and the dollar is not robust over time.

If one plots the trend in the U.S. dollar trade-weighted in-

dex against the smoothed trend (four-quarter moving sum)

in U.S. net FDI over the 1988-2000 period, a careful read-

ing of that plot would reveal that U.S. net FDI flows were

actually negative from 1991 to mid-1998, yet the dollar was

strong during a considerable portion of that time. Mean-

while, quarterly data reveal that net FDI flows were quite

large in the fourth quarter of 1998, yet the dollar fell during

that period.

Since then, FDI flows have been quite volatile—large in

one quarter and then modest in the following quarter. The

trend in the dollar has clearly not moved in sympathy with

this quarterly pattern in net FDI flows during the past two

years. Indeed, the correlation between quarterly net FDI

flows and the dollar's value since 1981 is a low 0.05.

U.S. Annual Net FDI and the Dollar

(Net Foreign Direct Investment, 4-Qtr. Mov. Sum)

(1988-2002)

100

105

110

-100

-50

100

150

200

250

88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

US$ Index----

nnual Net FDI(US$ bn.)____

U.S. Quarterly Net FDI and the Dollar

(Net Foreign Direct Investment)

(1995-2002)

100

110

-40

-20

100

120

140

1995 1996 1997 1998 1999 2000 2001 2002

US$ Index(line)Quarterly Net FDI(US$ bn.)(bars)

Foreign Direct Investment Flows and Exchange Rates

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 91

A recent BIS study entitled "The Impact of Transatlantic

M&A Activity on the Dollar/Euro Exchange Rate" finds that

over the 1999-2000 period, M&A announcements of Eu-

ropean firms' acquisitions of U.S. firms had a statistically

significant impact on the dollar's value versus the euro,

and that the size of this impact was independent of how

the deal was financed. Interestingly, M&A deals involving

U.S. firms' acquisitions of Euroland firms were found not

to have a statistically significant impact on the dollar's value

versus the euro.

This asymmetric response suggests that the influence of

M&A deals on the dollar's value over the 1999-2001 pe-

riod may have less to do with the FX flows, if any, gener-

ated by such deals, and more to do with the way the mar-

ket perceived the underlying rationale for those deals. That

is, the market may have perceived the steady acquisition

of U.S. firms by Euroland firms as a sign that Euroland

firms had greater confidence in the long-run growth pros-

pects of the U.S. relative to Euroland. The fact that the

pace of transatlantic activity slowed sharply in 2001, yet

the dollar remained remarkably resilient, suggests that the

linkage between M&A activity and the dollar may be less

robust when viewed over longer time horizons.

Estimates of the Effect of M&A Flows

on the U.S. Dollar/Euro Exchange Rate

(January 1999-September 2001)

Coefficient

(t-Statistic)

Current and Lagged M&A Flows into Euroland +0.346

(1.55)

Current and Lagged M&A Flows into U.S. -0.481*

(-2.52)

Note: * denotes that the variable was statistically significant and had the correct sign.

Source: Ingo Fender and Gabriele Galati, “The Impact of Transatlantic M&A Activity

on the Dollar/Euro Exchange Rate”,

BIS Quarterly Review

, December 2001.

Merger & Acquisition Flows and the Dollar

(Cumulative Net Flow from Euroland to the U.S.)

(1998-2002)

(1.30)

(1.20)

(1.10)

(1.00)

(0.90)

(0.80)

100

150

200

250

300

1998 1999 2000 2001 2002

US$/Euro(reverse scale)(line)M&A Flows(US$ bn.)(bars)

Merger and Acquisition Flows and the Dollar—1999-2001

Source: Bloomberg, Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

92 Deutsche Bank Foreign Exchange Research

It is widely recognized that part of the dollar's strengthen-

ing trend in recent years can be explained by the dramatic

surge in foreign direct investment and portfolio flows into

the U.S. from overseas markets. Many assume that this

surge in capital inflows reflected a newfound optimism in

the health and vigor of the U.S. economy and financial

markets. While this is undoubtedly true, it may also be the

case that a significant portion of the capital flow that has

found its way to the U.S. simply reflected capital flight from

other parts of the world, notably emerging markets.

The evidence suggests that a series of crises in emerging

markets—from Asia to Russia to Brazil to Turkey to Argen-

tina—persuaded investors to exit from emerging-market

investments and look for more stable returns in the indus-

trialized markets. Given the U.S. dollar's dominant posi-

tion as a global medium of exchange and store of value,

the U.S. naturally benefited from that capital flight. Net

private capital flows into emerging markets peaked at $233

billion in 1996, just prior to the Asian financial crisis of 1997.

Since then, private capital flows to emerging markets have

plummeted, with the 2000 inflow amounting to only $9.0

billion.

As the charts below indicate, net FDI flows into emerging

markets did not change much over the 1997-2001 period.

Net FDI inflows remained stable, averaging around $170

billion per annum. Where there was a striking change in

the path of capital flows was in the bond financing and

bank lending areas. Bank lending plummeted from an in-

flow of $51.9 billion in 1998 to an outflow of $32.3 billion in

2001. Bond financing flows fell from a high of $62.3 billion

in 1996 to a mere $9.5 billion in 2001.

The Impact of Emerging-Market Capital Flight on the Dollar—1997-2001

Capital Flows into Emerging Markets

(Net Private Capital Flows)

20.1

8.9

59.6

69.6

116.8

233.2

205.7

151.3

145.6

100

150

200

250

1993 1994 1995 1996 1997 1998 1999 2000 2001e

(US$ bn.)

Source: IMF World Economic Outlook,

December 2001

Bond Flows into Emerging Markets

9.5

16.9

29.5

40.9

49.6

62.3

30.7

38.1

36.7

11.1

11.0

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001e

(US$ bn.)

Source: World Bank

Equity Flows into Emerging Markets

18.5

50.9

34.5

15.6

30.1

48.9

36.1

35.2

51.0

14.1

7.6

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001e

(US$ bn.)

Source: World Bank

FDI Flows into Emerging Markets

168.2

166.7

184.4

178.3

172.5

130.8

106.8

90.0

66.6

47.1

35.7

100

150

200

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001e

(US$ bn.)

Source: World Bank

Banking Flows into Emerging Markets

-32.3

-6.1

-23.3

51.9

45.6

32.2

30.9

9.3

4.1

16.3

5.0

-40

-20

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001e

(US$ bn.)

Source: World Bank

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 93

The 1997-2001 episode did not mark the first time that

capital flight from emerging markets has played an impor-

tant role in contributing to a major rise in the dollar's value.

A similar development occurred in the first half of the 1980s

when the dollar underwent a dramatic upside overshoot

that could not be adequately explained by U.S., European,

or Japanese economic fundamentals.

One reason the dollar might have overshot its equilibrium

level in 1982-85 was that the U.S. had become a major

recipient of massive capital flight from Latin America fol-

lowing the onset of the Latin American debt crisis in 1982.

Evidence on Latin American capital flight and dollarization

ratios indicates that a large and sustained increase in the

demand for dollars by Latin American residents occurred

at that time.

Argentina’s Dollarization Ratio

(1979-1985)

1979 1980 1981 1982 1983 1984 1985

(%)

Source: IMF

Mexico’s Dollarization Ratio

(1970-1985)

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

(%)

Peru’s Dollarization Ratio

(1970-1985)

100

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

(%)

Source: IMF

Uruguay’s Dollarization Ratio

(1970-1985)

100

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

(%)

Source: IMF

The Latin American Debt Crisis and the Surge of the Dollar—1982-85

Capital Flight from Developing Countries

(1980-1985)

147.5

136.4

123.8

99.9

85.2

75.4

100

120

140

160

1980 1981 1982 1983 1984 1985

(US$ bn.)

Source: Liliana Rojas Suarez,

"Risk & Capital Flight in Developing Countries",

IMF, 1991.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

94 Deutsche Bank Foreign Exchange Research

The U.S. was the beneficiary of a structural shift in capital

inflows in the 1995-2001 period, both from the emerging

markets and from Europe. The rise in capital flows from

emerging markets to the U.S. began with the flight of capi-

tal from Asia following the Asian financial crisis in 1997.

Capital flows to emerging markets dried up as well in 1998

following the Russian/LTCM crisis and have yet to recover.

It is assumed that most of those flows found their way to

the U.S.

The inflows from Europe came in the form of surges in

foreign direct investment into the U.S., stepped-up foreign

purchases of U.S. equities by Euroland investors, and huge

increases in European purchases of U.S. corporate bonds.

Foreign direct investment and equity flows slowed dra-

matically in 2001, but this was more than offset by a major

surge in foreign purchases of U.S. corporate bonds. These

inflows helped drive the dollar steadily higher over the 1995-

2001 period despite a record deterioration of the U.S. cur-

rent-account deficit.

Net Foreign Direct Investment in the U.S.

-100

100

200

300

400

-50

100

150

200

1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Annual(line)(US$ bn.) Quarterly(bars)

Net Foreign Purchases of U.S. Corporate Bonds

100

200

300

400

100

1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Annual(line)(US$ bn.) Quarterly(bars)

Net Foreign Purchases of U.S. Stocks

-50

100

150

200

250

-20

1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Annual(line)(US$ bn.) Quarterly(bars)

The U.S. Dollar and

Net Portfolio Investment into the U.S.

(12-Month Moving Sum of Net Portfolio Invesment)

-100

-50

100

150

200

250

300

100

105

87 88 89 90 91 92 93 94 95 96 97 98 99 00 01

Net Portfolio Investment(US$ bn.)----US$ Major Currency Index____

Financing the U.S. Current-Account Deficit

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 95

Under normal conditions, U.S. interest rates would have

been required to rise relative to foreign interest rates in

order to attract greater capital inflows to finance the steady

widening of the U.S. current-account deficit. As shown in

Figure (a) in the diagram below, the equilibrium value for

the dollar is shown to be determined at the point where

the U.S. savings-investment schedule (S-I) and the U.S.

current-account balance schedule (CAB) intersect. As

shown in Figure (b), the U.S./foreign interest-rate spread

required to finance the current-account deficit and attract

the necessary capital flows is (rUS -rF)1.

An increase in U.S. investment, such as the one experi-

enced by the U.S. during the 1990s, would have shifted

the S-I schedule to the left, which under normal circum-

stances would have required a relative rise in U.S. interest

rates in order to induce foreign investors to finance the

wider U.S. current-account deficit. Higher U.S. interest rates

would, in turn, have led to wider U.S./foreign yield spreads,

from (rUS -rF)1 to (rUS -rF)2.

It was truly remarkable during the 1995-2001 period of dollar

appreciation that the U.S. was able to finance a record de-

terioration in its current-account deficit without having to

push U.S. interest rates higher relative to interest rates

overseas.

How did the U.S. accomplish this? The U.S. was the ben-

eficiary of a major structural shift in capital flows to the

U.S. (shown in Figure (b) as a shift of the capital flow sched-

ule from

U.S. Capital Flows

1 to

U.S. Capital Flows

2) that

was independent of any change in U.S./foreign yield

spreads. Indeed, the capital flight from the emerging mar-

kets and the inflow of capital from Euroland actually al-

lowed nominal and real yield spreads to narrow during the

dollar's run-up in 1999-2001. Hence, the normally reliable,

positive correlation between trends in U.S./foreign yield

spreads and trends in the dollar’s value was broken in 1999-

2000. At the same time, the correlation between the trend

in U.S./foreign yield spreads and the U.S. current-account

deficit was broken as well.

U.S. Current-Account Deficit and the

U.S./German Short-Term Interest-Rate Spread

-800

-600

-400

-200

200

400

600

-1

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

U.S.-German 3-Mo. Rates(bp)----

urr’t Acc’t Deficit/GDP(%)____

How a Surge in U.S. Investment Spending

Typically Affects the Dollar

and U.S./Foreign Real Yield Spreads

Savings minus Investment,

Current-Account Surplus

U.S. Dollar

Real Exchange Rate

Surge in

Investment

Spending

S-I

CAB

Upward Revision in the

Dollar’s Real Long-Run

Equilibrium Value

(0)

S-I

U.S. Capital Outflows

U.S./Foreign

Real Yield Spreads

U.S. Capital Flows

Typical Rise in

U.S. Real Interest Rates

Necessary to Attract

Additional Capital Inflows

(from point to ).

(r - r )

US F 2

(0)

U.S. Capital Inflows

(r - r )

US F 1

U.S. Capital Flows

(b)

(a)

Current-Account Deficit

But the 1995-2000 Structural

Shift in Capital Inflows (from

US Capital Flow to US Capital Flow )

Allowed Real Interest-Rate Spreads

to Remain at (r -r ) .

US F 1

The U.S. Current-Account Deficit and U.S./Foreign Yield Spreads

U.S. Current-Account Deficit and the

U.S./German Real Long-Term Yield Spread

-600

-400

-200

200

400

-1

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

U.S.-German Real 10-Year Yield(bp)----

urr’t Acc’t Def./GDP(%)____

Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

96 Deutsche Bank Foreign Exchange Research

Our understanding of the impact that monetary and fiscal

policy changes have on exchange rates owes much to the

pioneering work of Robert A. Mundell and J. Marcus

Fleming. The Mundell-Fleming (MF) model has become

the textbook standard that most students today use to

study the role of monetary and fiscal policy in an open

economy.

In the MF model, the degree of capital mobility plays a

pivotal role in determining the exchange rate's response

to a change in fiscal or monetary policy. For example, an

expansionary monetary policy will lead to a depreciation

of the domestic currency in large part because the induced

decline in domestic interest rates will generate an outflow

of capital that puts downward pressure on the domestic

currency. The more sensitive capital flows are to changes

in interest rates, the greater will be the outflow of capital

in response to the interest-rate decline. Hence, the more

mobile capital is, the more the currency will depreciate in

response to a monetary expansion.

Fiscal policy's impact on the exchange rate is a bit am-

biguous. If capital is highly mobile, an expansionary fiscal

policy will likely give rise to an appreciation of the domes-

tic currency. If capital is relatively immobile, then an ex-

pansionary fiscal policy could lead to a decline in the do-

mestic currency's value. This is because an expansionary

fiscal policy typically results in a rise in domestic interest

rates and an increase in economic activity. The rise in do-

mestic interest rates should attract an inflow of capital from

abroad, which should be positive for the domestic currency,

but at the same time, the consequent rise in domestic

economic activity will contribute to the deterioration of the

trade account, which should be negative for the domestic

currency.

Whether the exchange rate rises or falls in response to a

fiscal expansion will depend on how sensitive capital flows

are to changes in interest rates. If capital flows are highly

mobile, the induced inflow of capital from abroad should

dominate the deterioration in trade, and thus the domes-

tic currency will tend to appreciate in value. If capital flows

are relatively immobile, the opposite will be the case.

The Mundell-Fleming Model of Exchange-Rate Determination

Source: Rosenberg (1996)

The Mundell-Fleming Model Transmission Mechanism

Overall

Balance of

Payments

Deficit

Domestic

Expansionary

Monetary

Policy

Decrease in

Domestic

Interest Rates

Increase in

Domestic

Economic

Activity

Deterioration

Trade

Balance

Capital

Outflow

Depreciation

Domestic

Currency

How Changes in Monetary Policy Affect Exchange Rates

The Mundell-Fleming Model Transmission Mechanism

Overall

Balance of

Payments

(?)

Domestic

Expansionary

Fiscal

Policy

Increase in

Domestic

Interest Rates

Increase in

Domestic

Economic

Activity

Deterioration

Trade

Balance

Capital

Inflow

Depreciation

Domestic

Currency

How Changes in Fiscal Policy Affect Exchange Rates

Appreciation

Domestic

Currency

Trade-Balance

Deterioration

Dominates

Capital Inflow

Dominates

Deterioration of

Trade Balance

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 97

In the MF model, an expansionary domestic monetary

policy will exert downward pressure on a domestic

currency's value while a restrictive domestic monetary

policy will exert upward pressure on a domestic currency's

value. The more mobile is capital, the greater will be the

exchange-rate response.

In an IS-LM context, an expansionary domestic monetary

policy will shift the LM (liquidity) curve downward and to

the right to LM1, resulting in a decline in domestic interest

rates and a rise in domestic demand. The decline in do-

mestic interest rates will cause capital to flow abroad, while

the rise in domestic demand will cause the trade balance

to deteriorate. The combined deterioration of the trade and

capital accounts will lead to a deterioration of the overall

balance of payments and, in the process, exert downward

pressure on the domestic currency.

Under conditions of low capital mobility, the interest-rate-

induced outflow of capital should be modest (from K0 to

K1). A modest capital outflow in turn should result in a

modest decline in the domestic currency's value (from E0

to E1). Under conditions of high capital mobility, the inter-

est-rate-induced outflow of capital should be large (from

K0 to K2), and the resultant decline in the domestic

currency's value (from E0 to E2) should be large as well.

Monetary Policy's Impact on Exchange Rates in the MF Model

The Effect of an Easing in Monetary Policy on an Exchange Rate

under Conditions of Low and High Capital Mobility

Output

Interest Rate

Capital Inflow

Interest Rate

B’

Low Capital

Mobility

High Capital

Mobility

Trade Deficit

Exchange Rate

Im -Ex

Capital Inflow

Exchange Rate

B’ A

Im -Ex

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

98 Deutsche Bank Foreign Exchange Research

The impact of fiscal policy on exchange rates is somewhat

ambiguous in the MF model. In an IS-LM context, an ex-

pansionary fiscal policy will give rise to a rightward shift of

the IS (investment-savings) curve to IS1, resulting in a rise

in domestic interest rates from i0 to i1 and a rise in domes-

tic demand from Y0 to Y1. The rise in domestic interest rates

will attract an inflow of capital from abroad, which should

drive the domestic currency’s value higher (from E0 to E2)

while the rise in domestic demand should cause the trade

balance to deteriorate, which should cause the domestic

currency to weaken (from E0 to E1). The equilibrium ex-

change rate could thus be either E1 or E2 following a fiscal

expansion, i.e., the resulting change is ambiguous.

Under conditions of high capital mobility, the induced in-

flow of capital should dominate the deterioration in trade,

and thus the currency should strengthen significantly. If

capital mobility is not that high, then the exchange-rate

response will be more muted. Under certain conditions

when capital mobility is extremely low, the deterioration

in trade may dominate and the domestic currency could

weaken.

Fiscal Policy's Impact on Exchange Rates in the MF Model

The Effect of an Expansionary Fiscal Policy on an Exchange Rate

under Conditions of High Capital Mobility

Output

Interest Rate

Capital Inflow

Interest Rate

High Capital

Mobility

Trade Deficit

Exchange Rate

Im -Ex

Capital Inflow

Exchange Rate

Im -Ex

Ambiguous

The Effect of an Expansionary Fiscal Policy on an Exchange Rate

under Conditions of High Capital Mobility

Output

Interest Rate

Capital Inflow

Interest Rate

High Capital

Mobility

Trade Deficit

Im -Ex

Capital Inflow

Exchange Rate

Im -Ex

Ambiguous

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 99

The specific mix of monetary and fiscal policies that a coun-

try pursues can have a profound impact on exchange rates.

In a world of high capital mobility, an expansionary do-

mestic fiscal policy will give rise to an appreciation of the

domestic currency's value. Similarly, a tight domestic mon-

etary policy will give rise to an appreciation of the domes-

tic currency's value when capital flows are highly mobile.

Thus, the combination of an expansionary fiscal and a re-

strictive monetary policy should be extremely bullish for a

currency when capital mobility is high. Conversely, the

combination of a restrictive fiscal policy and an expansion-

ary monetary policy should be extremely bearish for a cur-

rency. When monetary and fiscal policies are both expan-

sive or restrictive at the same time, the impact on exchange

rates is ambiguous.

The classic example of how an expansive fiscal/tight mon-

etary policy mix contributed to a sharp rise in a currency's

value is the case of the U.S. in the first half of the 1980s,

when President Reagan's expansive fiscal policy collided

with Fed Chairman Paul Volcker's tight monetary policy.

That policy mix helped drive U.S. real interest rates and

the dollar’s value to extraordinarily lofty levels.

Another classic episode in which a dramatic shift in the

policy mix caused equally dramatic changes in exchange

rates was the case of Germany in 1990-92. During that

period, the German government pursued a highly expan-

sionary fiscal policy to help facilitate German unification.

At the same time, the Bundesbank pursued an extraordi-

narily tight monetary policy to combat the inflationary pres-

The Monetary/Fiscal Policy Mix

and the Determination of Exchange Rates

Expansionary

Monetary

Policy

Restrictive

Monetary

Policy

Expansionary

Fiscal

Policy

Restrictive

Fiscal

Policy

Ambiguous

Domestic

Currency

Appreciates

Domestic

Currency

Depreciates

The 1992 ERM Crisis—

The Effect of Germany’s Fiscal/Monetary Policy Mix

on European Interest Rates

Output

Interest Rate

Y= Y

Output

ROE

Interest Rate

B’

A’

B’’

C’

D’

sures associated with unification. The combined expan-

sive fiscal/tight monetary policy mix drove interest rates

sharply higher in Germany, and those higher interest rates

were then transmitted to the rest of Europe via the ERM

pegged exchange-rate regime. Those higher interest rates

led to a marked slowdown in European growth and to a

marked rise in European unemployment. Recognizing that

this deterioration in the European economic climate was

unsustainable, currency speculators waged an attack on

the ERM regime and eventually toppled it. European cen-

tral banks could no longer keep interest rates high enough

to defend the ERM pegs and at the same time encourage

a rebound in economic activity.

Fiscal/Monetary Policy Mix and the Determination of Exchange Rates

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

100 Deutsche Bank Foreign Exchange Research

The (synthetic) euro's decline in the past nine years pro-

vides an example of how an adverse fiscal/monetary policy

mix can have a negative impact on a currency’s value. In

the run-up to European Monetary Union (EMU), Europe's

overall fiscal stance leaned toward significant restraint as

most governments in Europe steadily trimmed their struc-

tural budget deficits to satisfy the EMU admission criteria

on budget deficits and debt levels set out in the Maastricht

Treaty. All else being equal, fiscal restraint would have been

negative for the euro.

To counteract the negative economic consequences of

Europe's relatively restrictive fiscal policy stances, most

European central banks felt compelled to pursue relatively

easy monetary-policy stances. Real short-term interest

rates in Europe were pushed steadily lower in the 1990s.

All else being equal, the pursuit of a relatively easy mon-

etary policy should have exerted downward pressure on

the euro's value.

A tight fiscal or an easy monetary policy alone would have

been enough to depress the euro's value in the 1990s.

But the combination of the two proved to be deadly. As

illustrated in the IS-LM diagram below, the pursuit of an

easy monetary policy in Europe resulted in a rightward

shift in Europe's LM schedule from LM0 to LM1, while the

pursuit of a tight fiscal policy in Europe resulted in a left-

ward shift in Europe's IS schedule from IS0 to IS1. The inter-

section of the newly shifted IS1 and LM1 curves at point

indicates that Europe's policy mix drove European real in-

terest rates lower, and those lower real rates then contrib-

uted to a decline in the euro's value from point

A’

to point

C’

Euroland Real Short-Term Rates & the Euro

(Three-Month Euro Deposit Rate less CPI)

0.80

1.00

1.20

1.40

1.60

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

US$/Euro----Euroland Real 3-Mo. Rate(%)____

Euroland’s Budget Deficit and the Euro

0.80

0.90

1.00

1.10

1.20

1.30

1.40

-1

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

US$/Euro(line)Budget Deficit/GDP(%)(bars)

Source: OECD Economic Outlook

The Effect of Europe’s Fiscal/Monetary Policy Mix on the Euro

Output

Real Interest Rate

Euro’s Value

Real Interest Rate

C’

A’

(a) (b)

Euroland's Fiscal/Monetary Policy Mix in the 1990s and the Long-Term

Decline in the Euro's Value

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 101

The Japanese yen’s response to changing monetary/fiscal

policy conditions during 1999-2001 provides another ex-

ample of the influence of policy mixes on exchange rates.

In 1998-99, a number of prominent economists offered

policy prescriptions to help Japan out of its decade-long

economic slump. Adam Posen, of the Institute for Interna-

tional Economics in Washington, D.C. wrote a book en-

titled

Restoring Japan's Economic Growth

(1998), in which

he recommended that the Japanese government should

undertake a major fiscal stimulus initiative to reverse the

decline in Japanese economic activity. Posen argued that

Japan's fiscal stance had been too austere in the 1990s,

particularly after the 1997 hike in Japan's consumption tax.

Therefore, massive fiscal stimulus was needed to offset

the contractionary forces underway in Japan. In an IS-LM

context, Posen's solution amounted to shifting the IS curve

as far to the right as possible.

Paul Krugman of Princeton University offered a different

policy prescription. Krugman (1998) argued that Japan had

fallen into a liquidity trap and stressed that the Japanese

economy needed a negative real interest rate to restore

economic growth. Since nominal interest rates were al-

ready near zero and could not move into negative territory,

Krugman asserted that Japan needed to generate expec-

tations of higher inflation to drive real interest rates into

negative territory. That could be accomplished only if the

Bank of Japan made a credible commitment to expand

the rate of growth of the money supply now and in the

future. In an IS-LM context, Krugman's solution amounted

to shifting the LM curve downward and to the right, so

that it intersected the IS curve in negative real interest-

rate territory.

A combination of fiscal and monetary stimulus probably

would have been the most desirable policy, allowing

Japan's real growth to expand by more than would have

been possible with either fiscal or monetary stimulus alone.

In addition, the impact of such a policy mix would have

been broadly neutral for the yen. An expansive fiscal policy

stance would have been yen supportive, while an expan-

sive monetary policy stance would have been yen nega-

tive, leaving the yen's value largely unchanged.

Posen’s Solution for Resolving

Japan’s Economic Problems

A Massive Fiscal Stimulus

Output

Real Interest Rate

Easy Monetary Policy and Negative Real Interest Rates

Output

Real Interest Rate

Krugman’s Solution for Resolving

Japan’s Economic Problems

Japan’s Monetary/Fiscal Policy Mix Debate and the Yen’s Value in 1999-2000

Output

Real Interest Rate

Preferable 1999 Japanese

Fiscal/Monetary Policy Mix

A Combination of Fiscal and Monetary Stimulus

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

102 Deutsche Bank Foreign Exchange Research

In the end, the Japanese authorities chose to follow the

policy recommendations suggested by Posen and engi-

neered a powerful fiscal stimulus over the 1998-99 period.

The Bank of Japan did not adopt an aggressive policy of

monetary easing, as recommended by Krugman, but rather

leaned toward restraint over much of the 1999-2000 pe-

riod. So instead of Japan's LM curve shifting to the right, it

appears that the LM curve shifted to the left in 1999 and

remained there in 2000.

The result of this expansive fiscal/tight monetary policy

mix was that Japan's LM curve shifted leftward while its

IS curve shifted rightward, which had the effect of driving

real interest rates higher instead of lower. As shown in the

IS-LM diagram, the increase in output associated with this

policy mix is normally quite modest (shown as a rise from

Y1 to Y4), while the rise in real interest rates tends to be

large. There was indeed a very sharp increase in real inter-

est rates in Japan relative to U.S. real rates in 1999-2000,

which contributed to a rise in the yen's value over that

period.

The Effect of Japan’s 1999

Expansive Fiscal/Restrictive Monetary Policy Mix

on the Yen

Output

Real Interest Rate

Yen’s Value

Real Interest Rate

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 103

The 1999-2000 period of yen strength ended when Japan's

fiscal/monetary policy mix swung back in the direction of

significant fiscal restraint and aggressive monetary ease

in 2001-2002. Such a policy mix is normally bearish for a

currency, but this policy mix was especially bearish for the

yen, given the magnitude of the policy changes that were

implemented.

On the fiscal-policy front, Japan's structural budget deficit

as a percentage of GDP (the actual budget deficit as a

percentage of GDP adjusted for cyclical developments) is

expected to contract by a combined 1.7 percentage points

in 2002-03, according to IMF estimates. This would repre-

sent a greater tightening in Japan's fiscal stance than the

tightening that occurred in 1997, which was widely cred-

ited with pushing the Japanese economy into a recession

in 1997-98.

Given that Japan is presently saddled with huge budget

deficits, a large outstanding stock of government debt, and

adverse demographics that are expected to wipe out the

social security system's surpluses, the Japanese govern-

ment will need to pursue relatively restrictive policies, not

just in 2002, but for several years to come while it puts its

fiscal house in order. Such a policy stance should exert

downward pressure on the yen, not just in 2002, but for

many years to come.

Japan's monetary-policy stance is adding to this down-

ward pressure on the yen's value. Japan's monetary policy

became highly reflationary in late 2001 as economic con-

cerns began to mount. The Bank of Japan started adding

liquidity at an extremely rapid clip, with the level of BoJ

current-account deposits being targeted at the upper end

of a ¥10-¥15 trillion target range. This infusion of liquidity

boosted monetary-base growth to more than a 30% year-

over-year rate by early 2002, its highest pace in the past

three decades. This surge in Japanese monetary-base

growth relative to U.S. monetary-base growth contributed

to the yen's weakening trend.

Monetary/Fiscal Policy Mix and the

Determination of Exchange Rates

Expansionary

Monetary

Policy

Restrictive

Monetary

Policy

Expansionary

Fiscal

Policy

Restrictive

Fiscal

Policy

Ambiguous

Domestic

Currency

Appreciates

Domestic

Currency

Depreciates

Japan’s Fiscal Policy Trends

(Changes in Structural Budget Balance as % GDP)

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

1995 1996 1997 1998 1999 2000 2001 2002 2003

(Percentage Point Change)

Contractionary Fiscal Policy

Expansionary Fiscal Policy

Bank of Japan Monetary-Base Growth

-10

89 90 91 92 93 94 95 96 97 98 99 00 01 02

(Year-over-Year % Change)

Japan’s Monetary/Fiscal Policy Mix in 2001-02 and the Value of the Yen

Source: Datastream

Source: OECD, IMF Estimates

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

104 Deutsche Bank Foreign Exchange Research

Historically, changes in monetary policy have had a pro-

found impact on exchange rates. In the case of the dollar,

relatively easy monetary policies pursued by the U.S. dur-

ing the Carter Administration contributed to a sharp de-

cline of the dollar in 1977-78. The dramatic tightening in

U.S. monetary policy by the Volcker-led Federal Reserve in

the early 1980s helped drive the dollar dramatically higher

between 1981 and 1985. In the case of the yen, a rela-

tively tight monetary policy by the Bank of Japan in the

early 1990s contributed to a major strengthening of the

yen during that period. Then beginning in 1995 and con-

tinuing until 1998, the Bank of Japan pursued an extraordi-

nary easy monetary policy that contributed to a weaker

yen. More recently, the Bank of Japan has been engaging

in a very dramatic easing in monetary policy and the yen

has fallen commensurately. Finally, in the case of many

emerging-market currencies, overly easy monetary poli-

cies have often been the catalyst for currency-crisis epi-

sodes.

The monetary approach asserts that changes in the sup-

ply of and demand for money are the primary determi-

nants of exchange-rate movements. Although the mon-

etary approach is widely regarded as an incomplete theory

of exchange-rate determination because it ignores other

important explanatory variables, it nevertheless correctly

warns that the pursuit of overly expansionary monetary

policies will exert downward pressure on a currency's

value, and vice versa.

The monetary model of exchange-rate determination can

be derived from a basic model of the demand for money.

If it is assumed that purchasing power parity (PPP) holds

at all times, the equilibrium exchange rate can be shown

to be completely determined by trends in relative money-

supply growth, relative GDP growth, and relative interest-

rate differentials. According to the monetary model, a rela-

tive increase in domestic monetary growth, a relative de-

crease in domestic GDP growth, and a relative rise in do-

mestic interest rates will exert downward pressure on a

domestic currency's value.

Derivation of the Flexible-Price Monetary Model

of Exchange-Rate Determination

(1) mS = mD = m = p + b1y - b2i Money supply and demand

as a function of prices,

output, and interest rates

(2) p = m - b1y + b2i Price-level determination

(3) e$ = (pF - pUS) Purchasing power parity

assumption

(4) e$ = (mF-mUS) - b1(yF-yUS) + b2(iF-iUS)

Flexible-price monetary model

of the exchange rate

Source: Rosenberg (1996) Source: Rosenberg (1996)

Source: Rosenberg (1996)

How Real Economic Forces Influence Exchange Rates

in the Flexible-Price Monetary Model

of Exchange-Rate Determination

Change in

Real

Economic

Activity

Change in

Price

Level

Change in

Exchange

Rate

Change in

Real

Money

Demand

The Monetary Approach to Exchange-Rate Determination

The Direct Effect of Money-Supply Changes

on Exchange Rates

in the Flexible-Price Monetary Model

of Exchange-Rate Determination

Change in

Money

Supply

Change in

Price

Level

Change in

Exchange

Rate

The Impact of Money-Supply Changes

on Price Levels and Exchange Rates

Domestic Price Level

Foreign Currency’s

Value

Domestic

Money Supply

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 105

While the monetary model has not been highly success-

ful, broadly speaking, in explaining exchange-rate trends,

it nevertheless has done a fairly good job of explaining the

yen's medium-term cycles over the past 12 years. Since

1989, the yen has tended to strengthen on a trend basis

when Japanese monetary-base growth has decreased

relative to the U.S., and vice versa.

For example, monetary conditions in Japan were quite lax

in 1988-89 and this helped foster a weakening of the yen

during that period. Over the 1990-95 period, Bank of Ja-

pan policy was especially tight compared to a rather loose

Federal Reserve. That relatively tight BoJ stance helped

contribute to a stronger yen in 1990-95. The BoJ then turned

easy again over the 1995-98 period and the yen weak-

ened in tandem. BoJ policy then turned restrictive in 1998-

2000 and the yen strengthened in response. Beginning in

2001 and carrying over into 2002, BoJ policy has turned

easy again, and the yen has weakened accordingly. All in

all, the yen’s medium-term cycles of the past dozen years

were largely determined by the path that the Japan/U.S.

monetary-base ratio took.

Some observers might find it difficult to accept that BoJ

policy was too tight in 1998-2000 when Japanese short-

The Yen and Japanese/U.S. Monetary Policies

0.90

1.00

1.10

1.20

1.30

1.40

1.50

100

120

140

160

180

89 90 91 92 93 94 95 96 97 98 99 00 01 02

Japan/U.S. Monetary Base(ratio)----

en/US$____

Note: Y2K Adjusted,

Japan Reserve Require. Rate Change Adj. Series

Japan’s Monetary Conditions Index

-6

-4

-2

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

(%)

Restrictive

Accommodative

Japan’s Monetary Policy Function

(Taylor Rule)

-4

-2

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

Model Estimates(+/-1 Std. Error Band)----

(%) Discount Rate____

Source: Datastream Source: Datastream

Source: Datastream

The Monetary Model’s Track Record—

Japanese Monetary Policy and the Yen

term interest rates were hovering around zero. We would

argue, however, that nominal interest rates were a mis-

leading guide as to how tight the Bank of Japan truly was.

In fact, Deutsche Bank's monetary conditions index for

Japan shows a shift from an accommodative policy stance

to a restrictive policy in 1998-2000, and our Taylor Rule

model suggests that negative nominal rates would have

been more appropriate than the zero interest-rate policy

that the BoJ has been maintaining up until now.

Japan’s GDP Implicit Price Deflator

-3

-2

-1

89 90 91 92 93 94 95 96 97 98 99 00 01 02

ear-over-Year % Change

Japan’s Real Short-Term Interest Rate

(Three-Month Euro-¥ Rate less Core Inflation)

-3

-2

-1

1995 1996 1997 1998 1999 2000 2001 2002

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

106 Deutsche Bank Foreign Exchange Research

A variety of other indicators were also suggesting that BoJ

policy was actually quite tight over the 1998-2000 period.

For example, if one accepts the monetarist proposition that

"inflation is always and everywhere a monetary phenom-

enon", then it must also be true that deflation is always

and everywhere a monetary phenomenon. If so, the exist-

ence of outright deflation in Japan would be evidence that

BoJ policy had been too tight. Indeed, Japan's GDP defla-

tor had been declining for nearly four years and had fallen

to -2%, a near record pace of outright deflation.

Another indicator suggesting that BoJ policy was too tight

in 1998-2000 was the rising trend in real short-term inter-

est rates. With Japan's GDP deflator falling and nominal

short-term interest rates stable at around zero percent, this

had the effect of driving Japan's real short-term interest

rates higher on a trend basis. A declining money multiplier

and weak broad money-supply growth were also indica-

tive of an overly tight monetary-policy stance. The tighten-

ing of credit conditions in Japan in 1998-2000, including

the decline in Japanese bank lending, the decline in Japa-

nese banks' willingness to lend, and the steady erosion in

equity prices and land values, were still more signs of fi-

nancial restraint.

Japan’s Money Multiplier

(Broad Liquidity/Monetary Base)

89 90 91 92 93 94 95 96 97 98 99 00 01 02

Money Multiplier(6-mo. mov. avg.)

Japan’s Broad Liquidity Money-Supply Growth

(1995-2002)

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

1995 1996 1997 1998 1999 2000 2001 2002

(Year-over-Year % Change)

Japanese Bank Lending

(1982-2002)

-10

-5

1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

(Year-over-Year % Change)

Japan’s Equity Market

(1986-2002)

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

Nikkei Price Index

Japanese Land Prices

(Nationwide Index)

100

110

120

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

(Index)

Japanese Banks’ Willingness to Lend

(Tankan Survey)

-40

-30

-20

-10

85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

(Diffusion Index)

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 107

Monetary policy typically affects economic activity through

a variety of channels: (1) an interest-rate channel, (2) a bank-

lending or credit channel, (3) an exchange-rate channel, (4)

an inflation-expectations channel, and (5) a wealth-effect

channel. Under normal circumstances, an easier monetary

policy will lead to lower short-term interest rates, and as

lower short rates are transmitted to lower rates across the

entire yield curve, investment spending will tend to rise.

If, however, nominal short-term interest rates are already

at zero, as has been the case in Japan, an easier policy

stance will not be able to drive short-term interest rates

any lower, thereby rendering the interest-rate channel in-

effective.

The bank-lending channel in Japan has also been rendered

ineffective as the Bank of Japan's attempt to ease policy

has been blunted by the financial problems besetting the

banking industry. Indeed, since Japanese bank lending has

outright contracted in the past four years, this channel has

been effectively closed. And if bank share prices are any

indication of what is in store, the banking system's prob-

lems are not likely to go away anytime soon.

The exchange-rate channel has also failed to operate prop-

erly in Japan. Normally an easy monetary policy gives rise

to a depreciation of the exchange rate and that, in turn,

contributes to a rise in net exports and therefore economic

activity. Up until recently, the yen was fairly strong, sug-

gesting that monetary policy in Japan had been tight in

the first place. Given the yen’s firmness in 1999-2000, it is

obvious that this channel did not function properly during

that period. However, the yen’s weakness in 2001-02 may

render this channel more effective in the future.

When nominal short-term interest rates are zero, monetary

policy can still be effective if the money supply is expanded

at a sufficiently rapid rate so as to raise the expected infla-

tion rate. A higher expected inflation rate will then lower

the

real

short-term interest rate, even if

nominal

rates are

zero. In theory, those lower real interest rates should stimu-

late investment spending as outlined in the conventional

interest-rate channel. Many outside observers have rec-

ommended that the Bank of Japan engineer a quantitative

easing in policy and adopt a positive inflation target. But

Japan's deflationary pressures remain intact, thereby ren-

dering the inflation-expectations channel inoperative over

the last decade.

Finally, the wealth-effect channel in Japan has not been

operative for some time now. Normally, an easier mon-

etary policy can be transmitted to the domestic economy

through rising equity prices and real estate values, which

then increases net wealth and, in turn, helps boost con-

sumption. But with the Tokyo stock market and real estate

values down significantly on a trend basis, there has been

a trend decline in net financial wealth over time. Indeed,

the inability of equity and real estate prices in Japan to

rally raises questions whether monetary policy has been

truly easy even when interest rates are near zero.

With the traditional channels of monetary policy not func-

tioning properly, the Bank of Japan's policy actions have

not been as accommodative as would have been the case

under normal conditions. The end result of all this is that

the Bank of Japan has been too tight, perhaps inadvert-

ently, and this gave rise to a stronger-than-desired Japa-

nese yen.

Why Monetary and Credit Conditions in Japan Became Too Tight in the

Past Decade

The Channels of Monetary Policy

How Monetary Policy Changes Are Transmitted to Economic Activity

Expansionary

Monetary

Policy

Stimulate

Investment

Spending

Lower

Nominal

Interest Rates

Stimulate

Investment

Spending

Increase in

Economic

Activity

Increase in

Bank

Loans

Exchange

Rate

Depreciation

Stimulate

Net

Exports

Expansionary

Monetary

Policy

Expansionary

Monetary

Policy

Expansionary

Monetary

Policy

Expansionary

Monetary

Policy

Increase in

Economic

Activity

Increase in

Economic

Activity

Increase in

Economic

Activity

Increase in

Economic

Activity

Rise in

Inflation

Expectations

Rise in

Equity Prices

Rise in Land &

Housing Prices

Lower

Real

Interest Rates

Rise in

Value of

Financial Wealth

Interest Rate

Channel

Bank Lending

Channel

Exchange Rate

Channel

Inflation

Expectations

Channel

Wealth Effect

Channel

Adapted from Frederick Mishkin, “The Channels of Monetary Transmission: Lessons for Monetary Policy”,

NBER Working Paper, No. 5464, February 1996.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

108 Deutsche Bank Foreign Exchange Research

The portfolio-balance model posits that exchange rates are

determined by the supply of and demand for all financial

assets. The monetary model is actually a subset of the

broader portfolio-balance model in that the monetary model

narrowly focuses on the supply of and demand for money

as the key determinant of exchange-rate movements. The

portfolio-balance model broadens the menu of assets that

can determine the path that exchange rates take. In addi-

tion to relative changes in money supply and demand, the

portfolio-balance model focuses on relative changes in bond

supply and demand (domestic as well as foreign) as key

determinants of exchange-rate movements.

In the portfolio-balance model, global investors are as-

sumed to hold a diversified portfolio of domestic and for-

eign bonds. Their desired allocation to domestic and for-

eign bonds is assumed to vary in response to changes in

expected return and risk considerations. In the portfolio-

balance framework, a steady increase in the supply of do-

mestic bonds outstanding, generated by a continued wid-

ening of the government budget deficit, will be willingly

held only if asset holders are compensated in the form of

a higher expected return or risk premium. The higher risk

premium could come in the form of higher domestic inter-

est rates or an immediate decline in the domestic

currency's value, or some combination of the two. The

major insight one draws from this is that in the long run,

governments that run large budget deficits on a sustained

basis will eventually see their currencies decline in value.

Portfolio-Balance Model

Transmission Mechanism

Increase in

Supply of

Government

Debt

Outstanding

Increase

Risk

Premium

Increase in

Domestic

Interest

Rates

Decline in

Domestic

Currency’s

Value

The Supply and Demand for Domestic Bonds and the Risk Premium

Risk Premium

Supply of

Domestic Bonds e

Domestic

Currency’s Value

Risk Premium

Portfolio-Balance Model of Exchange-Rate Determination

Source: Rosenberg (1996)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 109

According to the portfolio-balance model, a rising net ex-

ternal debt burden will, in the long run, have negative con-

sequences for domestic interest rates and/or the domes-

tic currency's value. That is because, in the long run, for-

eign investors will demand a higher risk premium to buy

and hold increasing claims on the debtor country's assets.

That higher risk premium could come in the form of wider

domestic/foreign interest-rate spreads and/or a weaker do-

mestic currency. If the external debt becomes particularly

burdensome, the domestic currency may have to weaken

sharply, thereby allowing foreign investors to buy the debtor

country's assets at fire-sale prices.

The diagram below illustrates what normally happens when

a country undergoes a rapid deterioration in its net exter-

nal debt position. The risk premium on a debtor country's

assets, Ø1, is determined at point

, where the outstand-

ing stock of external debt, DS1, is willingly held by foreign

investors. If the outstanding stock of debt rises to DS2, for-

eign investors will demand a higher risk premium, Ø2, to

buy and hold that debt (at point

). The result would be a

widening in domestic/foreign yield spreads and/or a weak-

ening of the domestic currency.

Rising Net External Debt and the

Risk Premium on U.S. Dollar-Denominated Assets

(A Stylized Diagram of the 1997-2002 Period)

Risk Premium

on U.S. Assets

U.S. Net External Debt

Foreign Demand

for

US$-Denominated

Assets

External Debt, Portfolio Balance & Exchange-Rate Determination—Theory

Cumulative

Current-Account

Deficits

Rising

Net External

Debt Ratio

Increase in

Domestic

Interest

Rates

Decline in

Domestic

Currency’s

Value

Current-Account Imbalances and Exchange Rates—

The Channels of Influence

Increase

Risk Premium

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

110 Deutsche Bank Foreign Exchange Research

Just how long can a country run large and persistent cur-

rent-account deficits (or surpluses) before there is a re-

quired adjustment in the exchange rate to correct the grow-

ing external imbalance? There must be some limit on the

ability of a country to run sizable and persistent current-

account deficits, because the financing of such deficits

would lead to an unending rise in debt owed by deficit

countries to foreign investors. Since that debt must be

serviced, there is a risk that a deficit country could fall into

a "debt trap" with rising debts adding to debt-service obli-

gations, which then worsen the current-account deficit.

Once entrenched in a debt trap, the deteriorating current-

account trend adds to the rising external debt, which mag-

nifies the debt-service burden further, and so on.

The trend in the outstanding external debt as a percent-

age of GDP would rise without limit, unless this was off-

set by a rising trade surplus as a percentage of GDP. In the

absence of an improving trade position, the market would

be faced with the task of adding more and more risky debt

of the deficit country to its portfolios. At some point, the

market would probably demand a higher risk premium to

induce it to buy and hold onto that rising debt. Eventually,

there may come a time when the market perceives the

debt of the deficit country as so risky that it refuses to add

any more of that debt to its portfolios.

Debt-Trap Dynamics

Trade

Deficit Current-Account

Deficit External Debt

Outstanding

Debt

Service

Real Currency Appreciation vs.

the Change in Net Foreign Assets/GDP Ratio

(in Selected OECD Countries, 1986-90 vs. 1981-85)

Rate of Real

Currency Appreciation

Change in Net Foreign Assets/GDP Ratio

40%

30%

20%

10%

-10%

-20%

-0.15 -0.05-0.10 0.0 0.05 0.10 0.15

Australia

U.S.

U.K. Finland

Sweden

Netherlands

Canada

Denmark

Italy France Spain

Belgium

Germany

Japan

Norway

Source: Obstfeld and Rogoff (1994).

Thus, although it may be difficult to quantify precisely, there

will clearly be a limit on just how long a country can run

large and persistent current-account deficits (or surpluses).

It is unlikely that an efficient forward-looking market would

be willing to finance an unending string of current-account

deficits until the deficit country became hopelessly en-

gulfed in a debt trap. Rather, an efficient forward-looking

market would look far into the future, and if it judged the

future path of the net external debt/GDP ratio to be unsus-

tainable, then the exchange rate would adjust sooner to

ensure that the trade balance would improve by a suffi-

cient amount to offset the rising debt-service burden and

thus stabilize the net foreign debt/GDP ratio at a level that

was deemed sustainable.

Using the sustainability criterion to determine a currency's

real long-run equilibrium level, the equilibrium real exchange

rate could be defined as that rate that will ensure a stable

net external debt/GDP ratio in the long run. One would

then expect to find a positive (negative) relationship be-

tween a currency's real value and the trend in the net ex-

ternal asset (liability) position of that country, and long-run

equilibrium would be attained only when the net invest-

ment position stabilized. Obtsfeld and Rogoff (1994) pro-

vide evidence suggesting that there has been a significant

positive relationship between an increase in a country's

net foreign asset position as a percentage of GDP and the

trend in its real effective exchange rate for 15 OECD coun-

tries.

Debt Trap Dynamics and the Equilibrium Exchange Rate

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 111

The U.S. suffered a mammoth deterioration in its net ex-

ternal debt position in 2000, according to the latest annual

estimates by the U.S. Commerce Department. The U.S.

net external debt, which is the difference between the level

of foreign-owned assets in the U.S. and the level of U.S.-

owned assets abroad, soared in 2000 to $1.84 trillion, from

$1.1 trillion the previous year on a current cost basis. On a

market value basis, U.S. net external debt soared from

$1.52 trillion in 1999 to $2.19 trillion in 2000.

Note that the U.S. net external debt did not change much

over the 1997-99 period despite a record deterioration in

the U.S. current-account deficit. The reason for this was

that favorable valuation factors on U.S.-owned assets

abroad offset increases in foreign holdings of U.S. assets.

Unfortunately, negative valuation adjustments in 2000,

coupled with record increases in foreign holdings of U.S.

assets, contributed to a near-50% rise in the U.S. debt

burden, the worst annual deterioration on record.

The portfolio-balance model would have predicted that the

dollar should have fallen and U.S./foreign yield spreads

should have widened in response to the deterioration in

the U.S. net external debt position. This did not happen.

Instead, U.S./foreign yield spreads actually narrowed, not

widened, while the dollar's value soared, not weakened

as the portfolio-balance model would have predicted. This

would suggest, if anything, that the risk premium on dol-

lar-denominated investments must have fallen, not risen

in recent years.

U.S. Net Investment Position

(at Current Cost)

-2000

-1500

-1000

-500

500

82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00

(US$ bn.)

U.S. Net Investment Position

(at Market Value)

-2500

-2000

-1500

-1000

-500

500

82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00

(US$ bn.)

U.S. External Debt Problem & the Dollar—Theory vs. Practice

Source: Datastream Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

112 Deutsche Bank Foreign Exchange Research

Perhaps the diagram below better characterizes the recent

trend in the U.S. risk premium. In this diagram, let's as-

sume that the U.S. net external debt position during 1997-

99 stood at

U.S. Net External Debt

1997-99. Since the U.S. net

external debt position did not change much in 1997-99,

we assume that the outstanding stock of foreign-held U.S.

debt remained fixed at that level. Given the prevailing

For-

eign Demand for U.S. Assets

1 that existed in 1997, the U.S.

risk premium is assumed to have initially stood at Ø1 in

1997.

Between 1997 and 1999, foreign demand for U.S. assets

began to rise. This rise was attributable to increased Euro-

pean investor demand for U.S. equities and bonds, record

FDI flows in the U.S., and capital flight from emerging mar-

kets to the U.S. as investors looked for a refuge from the

Asian and Russian financial crises. This increased demand

for U.S. assets is shown as a rightward shift in the Foreign

Demand for U.S. Assets schedule to

Foreign Demand for

U.S. Assets

2. That shift in overseas demand helped drive

the risk premium on dollar-denominated assets down from

Ø1 to Ø3 (or from point

to point

C).

As the U.S. net external debt soared from

U.S. Net Exter-

nal Debt

1997-99 to

U.S. Net External Debt

2000 in 2000, one

might have expected the risk premium on dollar denomi-

nated assets to have risen to perhaps Ø4 (or point

), but

that was not the case. With U.S./foreign yield spreads con-

tinuing to narrow and the dollar remaining firm, the risk

premium on dollar-denominated assets must have fallen

further in 2000. This implies that although the U.S. net ex-

ternal debt position rose in 2000, the demand for U.S. as-

sets by foreign investors must have risen commensurately,

say to

Foreign Demand for U.S. Assets

Recent Trends in the Risk Premium on US$-Denominated Assets

Rising Net External Debt and the

Risk Premium on U.S. Dollar-Denominated Assets

(A Stylized Diagram of the 1997-2001 Period)

Risk Premium

U.S. Net External Debt

U.S.

Net External

Debt

1997-99

U.S.

Net External

Debt

2000

Foreign Demand for

US$-Denominated Assets

Foreign Demand for

US$-Denominated Assets

Foreign Demand for

US$-Denominated Assets

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 113

Relative

Increase in

U.S. Equity

Market

Capitalization

Ratio

Relative

Increase

U.S. Wealth

Relative

Rise in

U.S.

Aggregate

Demand

Relative

Rise in

U.S. Real

Interest Rate

How a Relative Rise in the U.S. Equity Market’s Capitalization Ratio

Would Help Raise the Dollar’s Value

Relative

Rise in U.S.

Household

Propensity

to Consume

Capital

Inflow

into U.S.

Rise in

U.S. Dollar’s

Value

Analysts attribute the rise in the dollar's value versus the

euro in the second half of the 1990s to a wide range of

factors including strong productivity gains in the U.S., struc-

tural shifts in portfolio flows from Euroland to the U.S., in-

creased M&A flows from Euroland to the U.S., structural

rigidities in Euroland, "New Economy" forces in the U.S.,

relative economic growth rates, policy mixes, and euro-

changeover risk. The portfolio-balance model offers an ad-

ditional explanation—the increase in wealth generated by

the U.S. equity-market rally in the second half of the 1990s

and its impact on U.S. domestic demand.

While it is true that both the U.S. and Euroland equity mar-

kets rallied strongly in the second half of the 1990s, U.S.

equity-market capitalization as a percentage of GDP rose

by a larger amount than the comparable capitalization ra-

tio in Euroland. According to the IMF, U.S. market capitali-

zation rose from about 80% of GDP in 1994 to 160% in

1999, while the comparable capitalization ratio in Euroland

rose from 30% of GDP to 90% of GDP. Thus, the capitali-

zation ratio in the U.S. rose a full 100%, while Euroland's

capitalization rose just 60%. Since U.S. households have

a higher propensity to consume out of equity wealth than

do their Euroland counterparts, the greater increase in U.S.

wealth relative to Euroland had a larger impact on the rise

in U.S. demand relative to Euroland demand in the second

half of the 1990s.

Shifts in Equity Market Capitalization Ratios & the Trend in Exchange Rates

Euroland & U.S. Equity-Market Capitalization

(as a Percentage of GDP)

100

120

140

160

180

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

Euroland----

(%) U.S.____

Adapted from Guy Meredith,

"Why Has the Euro Been So Weak?", IMF

Source: FIBV

Euroland-U.S. Equity-Market Capitalization

Differential and the Euro

0.80

0.90

1.00

1.10

1.20

1.30

1.40

-100

-90

-80

-70

-60

-50

-40

-30

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001

US$/Euro----(%) Euroland-U.S. Capitalization Ratios___

Adapted from Guy Meredith,

"Why Has the Euro Been So Weak?", IMF

Source: FIBV

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

114 Deutsche Bank Foreign Exchange Research

At the outset of the euro's launch, EMU optimists pre-

dicted that global investors and central banks would at-

tempt to build their holdings of euro assets at the expense

of their dollar holdings. It was widely felt that such portfo-

lio shifts would be extremely bullish for the euro.

What was not considered by EMU optimists was the pos-

sibility that issuers might also find the euro as an attrac-

tive vehicle in which to fund their operations. If the issu-

ance of euro-denominated debt exceeded the demand for

that debt, then the euro would weaken, not strengthen.

That is pretty much what happened to the euro over the

1999-2001 period. Issuance of euro-denominated debt by

non-residents soared between 1999 and 2001, both in

absolute terms and as a share of all foreign-currency de-

nominated debt. As shown in the accompanying chart, the

share of international debt denominated in euros rose from

17% prior to the euro's launch to around 27% in 2000.

Most of that increase in debt was purchased by Euroland

residents, as little was purchased by international inves-

tors.

Overall, the increased issuance of euro-denominated debt,

the increased demand for foreign equities by euro-area resi-

dents, and the stepped-up acquisition of U.S. firms by Eu-

ropean firms resulted in a large flow of capital out of

Euroland, which acted to depress the euro's value over

the 1999-01 period.

Euro-Denominated Debt Issuance and the Value of the Euro

Increased

Investment in

Non-Euro Assets

by Euroland

Residents

How Euro-Area Portfolio Shifts Might Have

Contributed to a Trend Decline in the Euro’s Value

Capital

Outflow

from

Euroland

Decline in

Euro’s

Value

Increase in

Euro-Denominated

Borrowing

by Non-Residents

Euro-Denominated Share of

International Debt Issues

(Percentage of Total Issues in All Currencies)

1994 1995 1996 1997 1998 1999 2000

Current Currency Values----

(%) Constant Currency Values____

Adapted from Guy Meredith,

"Why Has the Euro Been So Weak?", IMF

Source: Detkon and Hartmann (2000)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 115

There is no universal agreement among economists on

the role that fiscal policy plays in the determination of ex-

change rates. The reason for disagreement is that fiscal

impulses are transmitted to exchange rates through a vari-

ety of channels, some of which generate positive influ-

ences on a currency's value, while others generate nega-

tive influences.

For example, in the Mundell-Fleming model, an expansion-

ary fiscal policy typically gives rise to higher domestic in-

terest rates and an increase in domestic economic activ-

ity. The rise in domestic interest rates will induce a capital

inflow that should contribute to a rise in the domestic

currency's value, but the consequent rise in domestic eco-

nomic activity will also contribute to a deterioration of the

trade account, which should put downward pressure on

the domestic currency's value. How sensitive capital move-

ments are to the rise in domestic interest rates will deter-

mine whether the induced capital inflow will dominate the

deterioration of the trade balance or vice versa.

In the portfolio-balance model, a steady string of rising bud-

get deficits will result in a rising stock of government debt

outstanding. If risk-averse investors balk at buying and

holding an ever-rising stock of government debt, then in-

vestors would need to be induced to buy and hold that

debt through higher domestic interest rates, an immedi-

ate decline in the domestic currency's value, or some com-

bination of the two.

Perhaps one could combine the Mundell-Fleming and port-

folio-balance models into a single integrated framework

by asserting that a stimulative fiscal policy is likely to be

positive for a currency in the short run, but negative in the

long run. After the currency’s initial appreciation during the

period when the stimulative fiscal policy is put into place,

budget deficits will rise and total debt outstanding will grow.

As the government's debt obligation rises, the market will

begin to wonder how that debt will be financed, and if the

mountain of debt is sizable, the market may believe that

pressure will eventually be brought to bear on the central

bank to monetize the debt. That clearly would lead to a

rapid reversal of the initial currency appreciation.

Alternatively, the market may believe that the government's

fiscal stance will eventually have to shift toward signifi-

cant restraint to restore longer-run balance to its fiscal po-

sition. A reversal of the fiscal stance that initially drove the

currency higher would, by definition, set forces in motion

for a reversal of the initial currency appreciation.

The Short- and Long-Run Response of Exchange Rates

to Changes in Fiscal Policy

Expansive

Fiscal

Policy

Increase in Real

Interest-Rate

Differential

Currency

Appreciates

Government

Debt

Buildup

Currency

Depreciates

Short-Run Response

Long-Run Response

Central Bank

Monetizes

Debt

Fiscal Stance

Turns

Restrictive

Fiscal Policy and Exchange Rates

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

116 Deutsche Bank Foreign Exchange Research

U.S./German Real Interest-Rate Differentials

and the U.S. Federal Budget Deficit

(1977-1987)

-800

-600

-400

-200

200

400

600

76 77 78 79 80 81 82 83 84 85 86 87

Budget Deficit as a % of GDP----U.S.-German Real 10-Yr.(bp)____

The U.S. Dollar and U.S. Defense Spending

100

110

120

130

140

150

74 76 78 80 82 84 86 88 90 92 94 96 98 00 02

Defense Spending(% of GDP)----US$ Index____

One of the best examples of how a stimulative fiscal policy

can have a favorable impact on a currency is U.S. Presi-

dent Reagan's Economic Recovery and Tax Act of 1981.

The dollar rose sharply following the passage of the act as

it contributed to both a rise in the dollar's real long-run

equilibrium level and a significant rise in U.S./foreign real

interest-rate differentials.

The U.S. 1981 tax bill included cuts in marginal tax rates

for individuals and investment tax credits and accelerated

depreciation allowances for corporations to help spur busi-

ness investment. From an exchange-rate determination

standpoint, the business-oriented tax cuts helped boost

the dollar's real long-run equilibrium value because they

made U.S. industry more competitive. But the dollar got

its real boost when U.S. government spending soared at

the same time that taxes were cut. That pushed the U.S.

budget deficit as a percentage of GDP from 1.6% up to a

high of 6.1% in just four years in the early 1980s. Along

with an extremely tight monetary policy during the early

1980s, this helped drive U.S./European real yield spreads

up sharply, thereby sending the dollar up sharply as well.

Until the last decade, trends in U.S. military spending as a

percentage of GDP were highly positively correlated with

trends in the dollar's value. Higher military spending was

generally dollar supportive, while lower military spending

was generally dollar bearish. Academic research posited

that when U.S. military spending rose, it was likely due to

increased uncertainty in the global political landscape, with

the dollar benefiting from its role as a safe-haven currency

during troubled times. However, the close relationship

between military spending and the dollar appears to have

broken down in recent years.

The U.S. Dollar and the

U.S. Federal Budget Deficit

(1976-1990)

100

110

120

130

140

150

76 77 78 79 80 81 82 83 84 85 86 87 88 89 90

Deficit as a % of GDP----US$ Index____

U.S. Fiscal Policy and the Dollar

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 117

Japan’s Budget Deficit

(Excluding Social Security)

7.1

6.7

7.0

7.8

6.7

5.3

6.5

6.0

4.7

4.5

1.6

0.7

1.5

1.4

2.0

2.5

3.5

3.1

4.0

6.2

83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02

(% of GDP)

Source: OECD Economic Outlook,

2001-02 OECD Projections

Japan’s Mounting Debt Problem

(Assuming Unchanged Fiscal Policies)

100

150

200

250

300

1990 1995 2000 2005 2010 2015 2020 2025

Assuming 3.5% Real Interest Rate____ 1.5%----

Net Debt(% of GDP)

Source: OECD Economic Outlook,

June 1999

Japan’s Fiscal Stance and GDP Growth

(Annual Percentage Change in Japan’s Structural

Balance as a Percent of GDP)

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

91 92 93 94 95 96 97 98 99 00 01e 02e 03e 04e

GDP Growth(%)----

Fiscal Stance(Led One Year)(%)____

Expansionary

Restictive

Japan's fiscal deficit—at both the central and local author-

ity level—has been swelling in recent years, partly because

of past stimulus efforts and partly because government

revenue growth has been weak during Japan's decade-

long slump. Japan already has the industrial world's larg-

est government debt outstanding as a percentage of GDP,

and that debt burden seems certain to mount in future

years, according to OECD projections. Even those debt

projections probably understate Japan's true debt burden

by a sizeable amount, given Japan's seriously under-funded

social security and public pension systems, as well as other

off-balance-sheet liabilities that the government will be

obliged to underwrite.

The path for Japanese debt is clearly not sustainable. In

the long run, the Japanese government will have to do

what every other government in the industrial world has

had to do when faced with a similar set of circumstances.

That is, the Japanese government will have to engineer a

multi-year plan of fiscal consolidation—either by raising

taxes and/or by curtailing government spending—to bring

the budget deficit/GDP ratio down to sustainable levels.

Changes in Japan's fiscal stance have historically had a

marked impact on Japan's GDP growth rate, with stimula-

tive policies boosting GDP growth higher and restrictive

policies dragging GDP growth lower. Therefore, as Japan's

fiscal stance is tightened, pressure will come to bear on

the BoJ to turn more accommodative to sustain economic

growth.

As a result of these anticipated policy changes, Japan's

fiscal/monetary policy mix appears destined to shift toward

a stance of significant fiscal restraint and considerable

monetary ease. Hence, Japan's policy mix should there-

fore be highly yen negative. Indeed, since such a policy

mix will need to be put in place for a number of years, the

yen appears destined to weaken on a long-term basis.

Japan’s Fiscal Policy Mix and the Yen

Source: OECD Economic Outlook

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

118 Deutsche Bank Foreign Exchange Research

Euroland’s Budget Deficit and the Euro

0.80

0.90

1.00

1.10

1.20

1.30

1.40

-1

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

US$/Euro(line)Budget Deficit/GDP(%)(bars)

Source: OECD Economic Outlook

Canada’s Budget Deficit and the C$

(1.60)

(1.50)

(1.40)

(1.30)

(1.20)

-4

-2

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

C$/US$(reverse scale)(line)Budget Deficit/GDP(%)(bars)

Source: OECD Economic Outlook

Australia’s Budget Deficit and the A$

0.50

0.55

0.60

0.65

0.70

0.75

0.80

-2

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

US$/A$(line)Budget Deficit/GDP(%)(bars)

Source: OECD Economic Outlook

Sweden’s Budget Deficit and the Krona

(11)

(10)

(9)

(8)

(7)

(6)

(5)

-5

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Skr/US$(reverse scale)(line)Budget Deficit/GDP(%)(bars)

Source: OECD Economic Outlook

A number of currencies—notably the Canadian dollar, euro,

Australian dollar, and Swedish krona—weakened on a trend

basis over the past 10 years as their respective govern-

ments engineered major fiscal consolidations to rein in

excessively wide budget deficits.

In all cases, pressure was brought to bear on the respec-

tive central banks to pursue highly accommodative mon-

etary policies during those extended periods of fiscal con-

solidations. Therefore, the policy mix in each of those cases

exerted significant downward pressure on the respective

currencies.

The Experience of Other Countries with Restrictive Fiscal Policies

in the Past 10 Years

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 119

Is a strongly growing domestic economy bullish or bear-

ish for a currency? According to the monetary model of

exchange-rate determination, an increase in domestic eco-

nomic activity will give rise to an increase in the demand

for domestic money, and the increased demand for money

(everything else being equal) should then lead to an ap-

preciation of the domestic currency. In contrast, in the

Mundell-Fleming and Balance of Payments Flow models,

an increase in domestic economic activity will give rise to

an increase in import demand, which should cause the

trade balance to deteriorate. The deterioration of the trade

balance should then exert downward pressure on the do-

mestic currency.

One could combine the conclusions drawn from these two

models by arguing that in the short/medium run, a strongly

growing domestic economy should be bullish for a domes-

tic currency because increases in economic activity are

often associated with attractive investment opportunities

that should attract significant inflows of capital from abroad.

But in the long run, the balance of payments may place a

constraint on how rapidly a domestic economy can grow

relative to growth overseas. If strong domestic demand

gives rise to a significant deterioration of the trade bal-

ance that is deemed unsustainable, then either domestic

demand will have to be restrained or the domestic

currency's real value will have to fall to bring the trade/

current-account imbalance back to sustainable levels.

How Changes in Economic Activity Influence Exchange Rates

in the Balance of Payments Flow (Mundell-Fleming) Model

of Exchange-Rate Determination

Increase in

Domestic

Economic

Activity

Deterioration

Trade

Balance

Decrease in

Domestic

Currency’s

Value

Deterioration

in Balance

of Payments

How Changes in Economic Activity Influence Exchange Rates

in the Monetary Model

of Exchange-Rate Determination

Increase in

Domestic

Economic

Activity

Increase in

Demand

for Money

Increase in

Domestic

Currency’s

Value

Economic Growth and Exchange Rates

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

120 Deutsche Bank Foreign Exchange Research

Trends in economic activity have been a key determinant

of exchange-rate trends in recent years. In the case of the

dollar, the trends in the U.S. unemployment rate and U.S.

consumer confidence were, until recently, highly correlated

with the trend in the dollar. However, in both cases, there

was a disconnect between those economic indicators and

the dollar in 2001.

The trend in the euro has also been found to be highly

correlated with the difference in U.S./Euroland

expected

GDP growth rates since the euro's inception in 1999.

Expectations of U.S./European GDP Growth

and the Euro

0.80

0.90

1.00

1.10

1.20

-1.2

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

1999 2000 2001 2002

US$/Euro----German-U.S. GDP Expectations(%)____

Source: Consensus Economics Inc.;

Forecasts of U.S. and German GDP growth

for the following year (three-month average)

Euroland/U.S. Real GDP Growth Differentials

and the Euro

(Year-over-Year Percent Changes in Real GDP)

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

-6

-4

-2

91 92 93 94 95 96 97 98 99 00 01 02

US$/Euro----(%) EMU-US GDP Diff.(led 2 qtr.)____

U.S. GDP Growth and the Dollar

(1993-2002)

100

105

110

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

US$ Major Currency Index----Real GDP(y-o-y % chg.)____

U.S. Consumer Sentiment and the Dollar

(Conference Board Consumer Confidence Index)

(1.50)

(1.40)

(1.30)

(1.20)

(1.10)

(1.00)

(0.90)

(0.80)

100

120

140

160

1994 1995 1996 1997 1998 1999 2000 2001 2002

US$/Euro(reverse scale)----Consumer Confidence Index____

U.S. Unemployment and the Dollar

(1995-2002)

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

3.5

4.0

4.5

5.0

5.5

6.0

1995 1996 1997 1998 1999 2000 2001 2002

US$/Euro----U.S. Unemployment Rate(%)____

How Changes in Economic Activity Have Influenced Recent Trends in

Exchange Rates

Source: Datastream

Source: DatastreamSource: Datastream

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 121

The Yen & Japan’s Leading Economic Indicator

(OECD Composite Leading Indicator for Japan)

(6)

(4)

(2)

(150)

(140)

(130)

(120)

(110)

(100)

(90)

(80)

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

Leading Indicator(%)----

en/US$(reverse scale)____

U.K./German Unemployment and the Pound

2.00

2.20

2.40

2.60

2.80

3.00

3.20

3.40

1.0

1.5

2.0

2.5

3.0

3.5

1996 1997 1998 1999 2000 2001 2002

DM/£----

erman/U.K. Unemployment Rate(ratio)____

Trends in Global Economic Growth

and the Australian Dollar

-30

-20

-10

-6

-4

-2

89 90 91 92 93 94 95 96 97 98 99 00 01 02

US$/A$(y-o-y %)----

ECD Leading Economic Indicator(y-o-y %)____

Other currencies have also exhibited a tendency to move

in tandem with relative economic performances. For in-

stance, there has been a close relationship in recent years

between the yen and the OECD's leading indicator of Japa-

nese economic activity, with weaker Japanese growth

prospects negative for the yen and vice versa. In the case

of the British pound, the ratio of German/U.K. unemploy-

ment rates has served as a good guide to sterling's pros-

pects. All things being equal, a decline (rise) in the U.K.

unemployment rate relative to the German unemployment

rate has been bullish (bearish) for sterling.

Finally, in the case of Australia, the A$'s fortunes have of-

ten been tied to global economic prospects, with strong

global growth positive for the A$ and vice versa. Evidently,

strong global growth boosts the world's demand for com-

modities, which is positive for Australia's trade balance,

and thus the A$.

Source: Datastream Source: Datastream

Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

122 Deutsche Bank Foreign Exchange Research

It is not too difficult to unearth the reasons for the dollar's

rising trend over the past six years. The U.S. grew two to

three times faster than either the German or Japanese

economies, with 'New Economy' forces in the U.S. widely

credited for the relative strength of the U.S. economy, and

thus the dollar.

According to the ‘New Economy’ view, the surge in U.S.

investment spending and the consequent rise in U.S. pro-

ductivity helped push U.S. growth significantly above the

pace of overseas growth, which, in turn, raised the dollar's

real long-run equilibrium value versus the euro. This view

appears to be shared by many economists. However, what

is open to dispute is: (1) by how much did the dollar's real

long-run equilibrium value actually rise, and (2) did the

dollar's rise in 2000-01 push it above its new long-run equi-

librium level?

(Note that the German economy did not fare much better

than Japan over the past 10 years. Given that the 1990s

have often been characterized as Japan's lost decade, they

could just as easily have been characterized as Germany's

lost decade as well.)

Structural Rigidities, Long-Term Growth, and Long-Term Trends

in Exchange Rates

U.S., German, and Japanese Real GDP Growth

Annual Percentage Growth Rates

(1993-2002) 1993-2002

1993 1994 1995 1996 1997 1998 1999 2000 2001p2002pAverage

U.S. 2.7 4.0 2.7 3.6 4.4 4.3 4.1 4.1 1.2 2.3 3.3

Germany -1.1 2.3 1.7 0.8 1.4 2.0 1.8 3.0 0.6 0.9 1.2

Japan 0.5 1.1 1.5 3.6 1.8 -1.0 0.7 2.2 -0.4 -1.0 0.9

Source: IMF

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 123

How Germany and Japan Rank in Independent Surveys of Global Competitiveness

Babson College A. T. Kearney

World Institute Japan Center Global Foreign PriceWaterhouse

Economic Management for Economic Entrepreneurship Direct Investment Coopers

Rank Forum Development Research Monitor Confidence Index Opacity Index

1U.S. U.S. U.S. Brazil U.S. Singapore

2 Singapore Singapore Singapore Korea China U.S.

3 Luxembourg Finland Netherlands U.S. Brazil Chile

4 Netherlands Netherlands Finland Australia U.K. U.K.

5 Ireland Switzerland Hong Kong Norway Mexico Hong Kong

6 Finland Luxembourg Norway Canada Germany Mexico

7 Canada Ireland Sweden Argentina India Italy

8 Hong Kong Germany Australia India Italy Hungary

9 U.K. Sweden U.K. Italy Spain Israel

10 Switzerland Iceland Switzerland U.K. France Uruguay

11 Taiwan Canada Canada Germany Poland Greece

12 Australia Denmark Germany Denmark Canada Peru

13 Sweden Australia N.Z. Spain Singapore Egypt

14 Denmark Hong Kong Belgium Israel Thailand Lithuania

15 Germany U.K. Denmark Finland Australia South Africa

16 Norway Norway Japan Sweden Czech Rep. Japan

17 Belgium Japan Iceland Belgium S.Korea Colombia

18 Austria Austria Ireland France Netherlands Argentina

19 Israel France Taiwan Singapore Taiwan Taiwan

20 N.Z. Belgium Austria Japan Japan Brazil

21 Japan N.Z. France Ireland Hungary Pakistan

Euroland and Japan’s Competitiveness Rankings

World Economic Forum Rankings

(1996-2001)

Country 2001 2000 1999 1998 1997 1996

Finland 1 6 11 15 19 16

U.S. 212334

Canada 375548

Netherlands 84971217

Ireland 11 5 10 11 16 26

U.K. 12984715

Germany171525242522

Austria 18 18 20 20 27 19

Belgium 19 17 24 27 31 25

France 20 22 23 22 23 23

Japan 21 21 14 12 14 13

Spain 22 27 26 25 26 32

Portugal 25 23 27 26 30 34

Italy 26 30 35 41 39 41

Luxembourg NA 3 7 10 11 5

Source: World Economic Forum, www.weforum.org

The persistent underperformance of the German and

Euroland economies has, according to the BIS, given rise

to an asymmetry in the response of global investors to

favorable and unfavorable data releases out of Euroland.

When Euroland data releases are disappointing, the mar-

ket tends to react negatively and the euro declines in value.

However, when Euroland data releases are favorable, the

investment community has been reluctant to reward the

euro by driving it higher. Evidently, investors have become

accustomed to longer-term disappointing data out of

Euroland, and therefore treat favorable data releases as

transitory events and unfavorable data releases as more

permanent events. Given the market's asymmetric behav-

ior, the euro has tended to trade with a sustained down-

ward bias.

Market participants often place the blame for Euroland’s

and Japan's sluggish growth on structural rigidities that

inhibit their long-term economic potential. As shown in the

table below, the major European economies and particu-

larly Japan fare poorly in various independent surveys of

global competitiveness, entrepreneurship, attracting for-

eign direct investment, and opacity/transparency. The fact

that Europe's and Japan's poor standing is consistent

across a wide range of independent surveys suggests that

structural rigidities could act as a major drag not only on

Europe's and Japan's long-term growth prospects, but on

the euro's and yen's value as well.

Indicators of Global Competitiveness & Long-Term Value in FX Markets

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

124 Deutsche Bank Foreign Exchange Research

The argument most often made to justify official interven-

tion in the FX markets is that the exchange rate is simply

too important a price to be left to the market. Intervention

advocates contend that financial markets cannot be trusted

to get the exchange rate "right". They argue that interven-

tion is necessary to keep exchange rates broadly in line

with their long-run fundamental equilibrium value. Thus,

the case for intervention relies largely on the belief that

the authorities can do a better job than the market in terms

of driving exchange rates towards their long-run equilib-

rium value. This assumes that government officials have

better information and greater knowledge than the mar-

ket, which, of course, is highly debatable.

No precise rules govern the conduct of FX intervention,

but the IMF has established guidelines to prevent one

country from unfairly manipulating its exchange rate to gain

an unfair advantage over others.

Numerous empirical studies have been conducted to as-

sess whether intervention operations have a statistically

significant and quantitatively important impact on exchange

rates. In most studies, the results have been negative. In

those cases where a statistically significant relationship

between intervention and exchange rates has been found,

it has generally been the case that the impact was not

quantitatively important, i.e., the effect of the intervention

effort was neither sizable nor lasting. The major reason for

such findings is that intervention operations are typically

small relative to the volume of foreign exchange that

changes hands on a daily basis in the FX markets. Daily

intervention efforts typically average in the hundreds of

millions or perhaps only several billion. This contrasts with

average daily FX turnover of $1.2 trillion. While episodes of

successful intervention can be found, no systematic rela-

tionship exists between official intervention efforts and the

trend in exchange rates.

Why FX Intervention Advocates Believe

That the Market Cannot Be Trusted

to Get the Exchange Rate "Right"

• FX market may fail to use "all" existing information.

• FX market may be dominated by noise (trend-fol-

lowing) traders.

• Excessive speculation.

• Excessive risk aversion (shortage of speculation).

• FX market may be using a defective model of ex-

change-rate determination.

• Market perceptions of the future may be flawed or

inappropriately skewed.

• FX market may be pre-occupied with extraneous

information.

• FX market may be subject to persistent mood

swings, shifting from excessive optimism to exces-

sive pessimism, and then back.

Acceptable Forms of FX Intervention

That Help Foster Orderly Market Conditions

• Simple smoothing operations to limit erratic short-

run fluctuation in exchange rates.

• Counter excessive speculation or market overreac-

tion to news.

• Trend-breaking operation to put an end to an unde-

sirable uptrend or downtrend in a currency's value.

• Counter excessive risk aversion.

• Intervention to target a currency's value to some

specific level or range.

• Intervention to prevent large and persistent

misalignments.

Central Bank Intervention

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 125

Dollar Over/Undervaluations & U.S. Intervention

(DM/US$ PPP Estimates, 1998-2002)

-1000

-500

500

1000

1500

2000

-20

-10

1998 1999 2000 2001 2002

Intervention(US$ mn.)(bars)

Over/Under PPP Level(DM/US$)(line)

Dollar Overvalued -- Sell Dollars

Dollar Undervalued -- Buy Dollars

U.S. intervention policy has been highly changeable over

the past 30 years. Official attitude toward the dollar has

swung from benign neglect, to active encouragement of

a weaker dollar, to active intervention to stop the dollar

from falling, and most recently to acknowledge that a

strong dollar is in the U.S. national interest.

Intervention was most pronounced in the late 1970s, the

late 1980s, and in 1994-95. Since 1995, the Fed has inter-

vened on only two occasions—in mid-1998 (to guide the

dollar lower versus the yen) and in the fall of 2000 (to guide

the dollar lower versus the euro).

Dollar Over/Undervaluations & U.S. Intervention

(1976-1980)

-4000

-2000

2000

4000

6000

-40

-20

1976 1977 1978 1979 1980

Intervention(US$ mn.)(bars)

Over/Under PPP Level(DM/US$)(line)

Dollar Overvalued--Sell Dollars

Dollar Undervalued--Buy Dollars

Dollar Over/Undervaluations & U.S. Intervention

(1981-1986)

-4000

-2000

2000

4000

6000

-40

-20

1981 1982 1983 1984 1985 1986

Intervention(US$ mn.)(bars)

Over/Under PPP Level(DM/US$)(line)

Dollar Overvalued--Sell Dollars

Dollar Undervalued--Buy Dollars

Dollar Over/Undervaluations & U.S. Intervention

(1987-1998)

-4000

-2000

2000

4000

6000

-40

-20

87 88 89 90 91 92 93 94 95 96 97 98

Intervention(US$ mn.)(bars)

Over/Under PPP Level(DM/US$)(line)

Dollar Overvalued--Sell Dollars

Dollar Undervalued--Buy Dollars

Changeability of U.S. Exchange-Rate Policy

Administration/Period Policy

Nixon 1969-71 Benign neglect.

1971-73 Cut dollar loose from gold.

Devaluated dollar.

Carter 1977-78 Encouraged decline of dollar.

1978-80 Major intervention to arrest dollar’s decline.

Reagan/ 1981-84 Benign neglect/laissez-faire attitude toward

Bush dollar’s rise.

1985-86 Encouraged decline of dollar (Plaza Accord).

1987-92 Promoted dollar stability (Louvre Accord).

Clinton 1993-96 Encouraged decline of dollar.

1997-00 Strong-dollar policy.

Bush 2001- Strong-dollar policy.

U.S. Intervention Policy

Source: Federal Reserve Bank, Datastream Source: Federal Reserve Bank, Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

126 Deutsche Bank Foreign Exchange Research

In Japan, the Ministry of Finance has the responsibility for

determining the country's intervention policy, while the

Bank of Japan carries out that policy in the FX market.

Japanese authorities have consistently pursued a policy

of "leaning against the wind" in the conduct of intervention

operations. Whenever the yen has risen sharply against

the dollar, the BoJ has bought dollars and sold yen to mod-

erate the yen's rise. When the yen has fallen sharply in

value versus the dollar, the BoJ has sold dollars and pur-

chased yen to moderate the yen's decline. As a result,

there has been a close parallel movement between the

yen's value and the BoJ's holdings of foreign-exchange

reserve, with reserves rising whenever the yen has risen,

and vice versa.

Some analysts have found that the BoJ has tended to in-

tervene more aggressively when the yen has risen sharply

than when the yen has fallen sharply. That is, BoJ inter-

vention policy tends to be asymmetric, with greater em-

phasis placed on moderating yen strength than on tem-

pering yen weakness.

BoJ intervention was particularly aggressive in 1995 when

the yen soared to an all-time high of ¥/US$ 79.75. The BoJ's

FX reserves soared by nearly $75 billion between mid-1995

and mid-1996 to encourage an outright decline in the yen's

Japan’s Foreign Exchange Reserves & the Yen

100

120

140

160

100

200

300

400

500

1994 1995 1996 1997 1998 1999 2000 2001 2002

Yen/US$(line)Reserves(US$(bn.)(bars)

BoJ/MoF Intervention Policy

The Yen & Japanese Short-Term Interest Rates

(Three-Month Euro-Deposit Rates)

(1980-1998)

100

150

200

250

300

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998

3-mo. Rate(%)----Yen/US$____

Source: Datastream

value. At the same time, short-term rates were guided

lower from 1.75% in April 1995 to 0.50% in September

1995 to help encourage a decline in the yen's value.

The BoJ was also aggressive in 2001 as it attempted to

guide the yen lower. Indeed, in September 2001 alone,

the BoJ intervened to the tune of around $25 billion, and

the entire intervention effort was nonsterilized, i.e., its dollar

purchases were directly added to Japan's monetary base.

The intervention effort succeeded in driving the yen lower,

in part, because it forced investors with long yen/short

dollar exposures to unwind their positions and, in part,

because the monetary effects were quite substantial, as

Japan's monetary-base growth soared on a year-over-year

basis from 9.0% to 14.2% during that month.

The BoJ has frequently geared the overall thrust of its do-

mestic monetary policy actions with an eye toward mod-

erating movements in the yen's value. The BoJ has often

cut short-term interest rates to dissuade investors from

purchasing yen assets when the yen was rising sharply in

value, and vice versa. This would explain why Japanese

interest rates and the yen often move inversely, with the

yen rising at the same time that interest rates are falling,

and vice versa.

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 127

Economists have identified three

direct

channels through

which central bank intervention might have an immediate

impact on exchange rates.

First, intervention that alters the flow supply of foreign

exchange relative to the demand for foreign exchange can

directly affect the short-term trend in exchange rates. In-

tervention operating through that channel can be effective

only if the volume of intervention is sizable relative to the

daily turnover in the foreign-exchange market.

Second, nonsterilized intervention that alters the supply

of money relative to the private sector's demand for money

can directly affect the medium-term trend in exchange

rates. Intervention operating through this monetary chan-

nel can be effective only if the volume of intervention is

sizable relative to the outstanding stock of domestic money

holdings.

Third, sterilized intervention that alters the supply of do-

mestic bonds relative to the supply of foreign bonds in

private portfolios can also have a direct impact on the

medium-term trend in exchange rates. Intervention oper-

ating through that channel can be effective only if the vol-

ume of intervention is sizable relative to the stock of pub-

licly traded domestic and foreign bonds held in private

portfolios.

Unfortunately, the weight of evidence suggests that the

volume of intervention is often quite small relative to the

daily turnover of foreign-exchange activity, the stock of

money held by the private sector, and the stock of publicly

traded domestic and foreign bonds in private portfolios.

Thus, most studies conclude that the direct effect of inter-

vention on exchange rates is either statistically insignifi-

cant or quantitatively unimportant.

Central-Bank Intervention’s Role

in Regulating the Flow

Supply of and Demand for Foreign Exchange

Exchange Rate

Quantity of Foreign Exchange

D’

S’

The Portfolio-Balance Channel Transmission Mechanism

(The Impact of Sterilized Intervention on Exchange Rates

Change in

Exchange

Rate

Expectations

of Future

Monetary

Growth

Sterilized

Intervention

Change in

Risk

Premium

Change in

Expected

Future

Exchange

Rate

Change in

Domestic

Interest Rate

Change in

Foreign

Interest Rate

Direct Central Bank Intervention Channels—

The Balance of Payments Flow, Monetary, & Portfolio-Balance Channels

The Monetary Channel Transmission Mechanism

(The Impact of Nonsterilized Intervention on Exchange Rates

Change in

Monetary

Base

Change in

Exchange

Rate

Expectations

of Future

Monetary

Growth

Nonsterilized

Intervention

Multiple

Expansion of

Domestic

Money Supply

Change in

Expected

Future

Exchange Rate

via

Monetary

Channel

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

128 Deutsche Bank Foreign Exchange Research

Economists have also identified a number of

indirect

chan-

nels through which intervention can alter investors' expec-

tations and positioning and, in so doing, help push cur-

rency values in the direction desired by monetary officials.

First, a central bank may use intervention as a signal of its

monetary-policy intentions. To the extent that expectations

of monetary growth are altered, intervention could have

an immediate impact on exchange rates.

Second, a central bank may use intervention as a signal to

the market that an exchange rate is deviating too far from

its long-run equilibrium value. To the extent that interven-

tion helps anchor market expectations to the true long-run

equilibrium exchange rate, it can have a stabilizing influ-

ence.

Central Bank Intervention—

Signaling Channel Transmission Mechanism

Change in

Exchange

Rate

Convey Official

Intention of

Future

Monetary-Policy

Actions

to the Market

Central Bank

Intervention

Revision

of Market

Expectations

of Future

Exchange

Rate

Market

Efficiency

Channel

Superior

Information

Channel

Third, a central bank may want to take advantage of the

element of surprise and intervene when exchange rates

have overshot their equilibrium level and the market is

heavily overbought or oversold. If trend-followers dominate

foreign exchange activity and drive exchange rates far from

their true long-run equilibrium values, a surprise round of

intervention, appropriately timed, could force traders with

vulnerable long or short positions to unwind their positions.

Such actions could serve to break the prevailing trend and

possibly lead to a trend reversal if trend-followers could be

persuaded to jump in and help reinforce the new trend in

exchange rates.

As was the case above, the evidence in support of inter-

vention operations working through one of these indirect

channels is mixed. There have clearly been selected epi-

sodes of successful intervention operations, but most stud-

ies conclude that central banks cannot unduly influence

exchange rates on a sustained basis.

Central Bank Intervention—

Noise-Trading Channel Transmission Mechanism

Change in

Exchange

Rate

Central Bank

Intervention

Traders

Adjust Their

Short-Run

Currency

Positions

Indirect Central Bank Intervention Channels—

Signaling and Noise-Trading Channels

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 129

Anticipating Currency Crises in

Emerging Markets

Note: This section draws, in part, from Michael Rosenberg’s “Currency Crises in Emerging Markets”, Merrill Lynch, 1998.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

130 Deutsche Bank Foreign Exchange Research

The first half of the 1990s witnessed a strong and grow-

ing interest on the part of global fund managers in emerg-

ing-market investments. But a series of financial crises

during the second half of the decade significantly reduced

global fund managers' appetite for such investments. In-

deed, net private capital flows into developing countries

swung from an inflow of $233.2 billion in 1996 to a mere

$8.9 billion in 2000, according to IMF estimates, as over-

seas investors shied away from emerging-market invest-

ments following the Mexican peso crisis of 1994-95, the

Asian currency crisis of 1997, and the Russian crisis of

1998.

While the potential rewards from investing in emerging

markets can be quite attractive, these crisis episodes high-

light that the risks can be quite great as well. From a stra-

tegic standpoint, the issue is whether these currency cri-

ses could have been adequately anticipated ahead of time

to allow investors to hedge or unwind their emerging-mar-

ket currency exposures before the collapse in currency

values actually took place.

Views on the underlying causes of currency crises differ

greatly. One school of thought contends that currency cri-

ses tend to be precipitated by deteriorating economic fun-

damentals. If so, and the trend in those economic funda-

mentals weakens steadily and predictably, then it should

be possible to construct an early-warning system to an-

ticipate when a currency might be vulnerable to a specu-

lative attack.

An alternative school of thought argues that while evidence

of deteriorating economic fundamentals might explain a

relatively large number of currency collapses, there might

also be cases where economies with relatively sound fun-

damentals could see their currencies come under attack

because of a sudden adverse shift in market sentiment or

mood swing that is totally unrelated to economic funda-

mentals. If attacks can occur out-of-the-blue when eco-

nomic fundamentals are not offering any warning of an

impending attack, then there may simply be no way for

fundamental-based models to predict the onset of currency

crises.

Currency Crises in Emerging Markets—Introduction

Capital Flows into Emerging Markets

(Net Private Capital Flows)

20.1

8.9

59.6

69.6

116.8

233.2

205.7

151.3

145.6

100

150

200

250

1993 1994 1995 1996 1997 1998 1999 2000 2001e

(US$ bn.)

Source: IMF World Economic Outlook,

December 2001

Mexican Peso/U.S. Dollar Exchange Rate

(1994-1995)

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

Jan-94 Apr-94 Jul-94 Oct-94 Jan-95 Apr-95 Jul-95 Oct-95

Peso/US$

Asian Currencies & the 1997-98 Currency Crisis

(Currency Indices, January 1997=100)

100

120

140

160

180

200

220

100

200

300

400

500

600

700

Jan-97 Apr-97 Jul-97 Oct-97 Jan-98 Apr-98 Jul-98 Oct-98

Korean Won---- Thai Baht....Indonesian Rupiah____

Source: Datastream

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 131

Crisis-prone currencies typically display a number of clas-

sic symptoms that warn of an impending speculative at-

tack. These symptoms include:

• Excessive real appreciation of the emerging-market

currency.

• Weak domestic economic growth.

• Rising unemployment.

• A deteriorating current-account balance.

• Excessive domestic credit expansion.

• Banking-system difficulties.

• Unsustainably large government budget deficits.

• Overly expansionary monetary policies.

• A high ratio of M2 money supply to reserves.

• Foreign-exchange reserve losses.

• Falling asset prices.

• A huge buildup in short-term liabilities by either the

private or public sector.

While not all of these symptoms are present in each and

every currency crisis, most crises display a number of these

symptoms, enough to at least warn of the possibility of an

impending speculative attack.

Typically, analysts will look for evidence of deteriorating

fundamentals to determine whether an emerging-market

currency might be vulnerable to a speculative attack. One

can visualize that there exist critical threshold levels for

economic vulnerability indicators where, if economic fun-

damentals deteriorate beyond those threshold levels, an

attack on the currency will inevitably be waged.

The seeds of an inevitable currency collapse are planted

when a policy shift or an economic or financial shock gives

rise to a serious deterioration in one or more key economic

variables. Once economic fundamentals deteriorate be-

yond critical threshold levels (point

in the middle dia-

gram), a currency will naturally collapse under its own

weight.

Deteriorating Economic Fundamentals and the Onset of Currency Crises

Deteriorating Economic Fundamentals

and the Onset of Currency Crises

Economic

Fundamentals

Time

Policy

Variable

Financial

Shock

Time

Nominal

Exchange

Rate

A’’

(+)

(-)

(b)

(a)

(c)

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

132 Deutsche Bank Foreign Exchange Research

Since speculators, investors, and borrowers tend to be

forward looking, they will recognize before the fact that a

steady deterioration in economic fundamentals will lead

to an inevitable currency collapse. By looking ahead and

reasoning backward, speculators, investors, and borrow-

ers will not wait for point

to be reached to capitalize on

or to hedge against a possible currency collapse. Instead,

forward-looking market participants will begin to sell the

vulnerable currency well ahead of the inevitable collapse

at point

. By doing so, heavy selling pressure will occur

well before point

is reached. Such selling might trigger a

currency collapse in the vicinity of point

Forward-Looking Speculators Bring Forward

the Day of Reckoning

Economic

Fundamentals

B’

Time

Nominal

Exchan

Rate

Time

Forward-Looking Investors May Wage an Attack Before Fundamentals

Deteriorate Beyond Threshold Levels

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 133

In a world where there is complete certainty regarding the

future path that economic fundamentals will take, it should

be possible to pinpoint exactly when a speculative attack

might occur. However, in a world where there is less than

complete certainty, it might not be possible for a leading-

indicator model to correctly anticipate the onset of all cur-

rency crises.

Consider the case where economic fundamentals do not

deteriorate in a steady and predictable manner. Assume

that the trend in economic fundamentals is gradually weak-

ening, but there comes a point (such as point

in the

diagram below) where a sudden, unexpected improvement

in one or more key fundamental indicators permanently

alters the long-term trend in economic fundamentals in a

favorable direction. The danger here is that prior to point

being reached, an early detection system might have pre-

maturely warned of an inevitable currency collapse at point

when no such warning was truly called for.

In contrast to situations where leading-indicator models

might falsely warn of an attack that does not in fact occur,

there might be situations where leading-indicator models

fail to correctly warn market participants of an actual at-

tack. Consider a situation where the trend in economic

fundamentals appears to be sound and displays no sign of

serious deterioration right up until the time of a specula-

tive attack. Just prior to the point of attack, the trend in

economic fundamentals is shown to suddenly deteriorate

in the diagram below, perhaps because of an adverse

terms-of-trade shock, a sudden collapse of a large corpo-

ration or bank that undermines business confidence, or

because of contagion effects stemming from financial cri-

ses in neighboring markets. The deterioration in the funda-

mental backdrop and the attack on the now vulnerable,

but previously sound, currency occur virtually simulta-

neously with little or no warning.

One additional problem that model builders face is that

not all leading indicators of currency crises point in the

same direction at the same time. In many instances, we

may find that certain economic indicators deteriorate

steadily and predictably and therefore warn of an impend-

ing crisis, while other indicators may exhibit no sign of

deterioration and therefore suggest no need for a correc-

tive exchange-rate adjustment.

The simple fact is that when a country's economic funda-

mentals point in different directions, there is likely to be

considerable uncertainty whether the general trend in a

broad grouping of fundamental indicators argues for an

inevitable currency collapse or not. And when there is a

high level of uncertainty, it is difficult to generate signifi-

cant confidence in a mechanical, fundamental-based model

that purportedly predicts the onset of currency crises.

Uncertainty and the Timing of Speculative Attacks

A Sudden Deterioration in Economic Fundamentals

Could Trigger a Speculative Attack with Little Advance Warning

Economic

Fundamentals

A’

Time

Source: Rosenberg (1998) Source: Rosenberg (1998)

Uncertainty about the Future Course of Economic Fundamentals

Makes It Difficult to Pinpoint When an Attack Might Occur

Economic

Fundamentals

Time

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

134 Deutsche Bank Foreign Exchange Research

A number of economists have tried to make the case that

an exchange-rate regime might not have a unique equilib-

rium, but instead might face multiple equilibria. The dia-

gram below illustrates the possible range of equilibrium

settings that might exist for a particular exchange rate. At

the right end of the fundamental economic spectrum, a

unique equilibrium exists since countries with strong fun-

damentals are unlikely to see their currencies come under

speculative attack. At the left end of the fundamental eco-

nomic spectrum, a unique equilibrium also exists since

countries with weak fundamentals are highly likely to see

their currencies come under speculative attack. In the

middle, a so-called "gray area" exists where exchange-rate

crises may or may not occur.

It is in this gray area where multiple equilibria may exist—

economic fundamentals are either not strong enough that

a speculative attack can be completely avoided or not weak

enough that a crisis is an inevitable outcome. Operating in

the gray area, speculative attacks are possible but not in-

evitable. Exchange rates may be stable in the gray area for

a while, but they can easily be pushed over the edge by a

sudden adverse shift in market sentiment or mood swing.

Should a currency operating in the gray area come under

speculative attack, domestic economic fundamentals could

deteriorate unexpectedly if business confidence suddenly

sours or if foreign-currency debt burdens suddenly rise. If

the fundamental economic deterioration is truly serious,

the currency could be pushed out of the gray area toward

the "weak fundamentals" area where attacks are inevitable.

By so doing, the initial attack, which occurred when the

currency was still in the gray area, would prove to have

been a self-fulfilling prophecy.

Multiple Equilibria and Self-Fulfilling Expectations

Multiple Equilibria and the

Probability of Exchange-Rate Crises

Weak Fundamentals

No AttackAttack Inevitable

"Gray Area"

(Vulnerable to Attack but Not Inevitable)

Strong Fundamentals

Probability of Attack on Currency

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 135

If market participants were truly forward looking, interest-

rate differentials would tend to widen as an impending

currency crisis drew nearer and investors demanded a

higher return or risk premium to cover the expected cur-

rency loss on their investment. However, a careful reading

of the history of speculative attacks, particularly the crises

of the 1990s, suggests that this has not been the case.

For example, interest-rate differentials failed to widen

measurably in anticipation of the ERM currency crisis of

1992-93, the Mexican peso crisis of 1994-95 or the Asian

currency crisis of 1997-98. Not only did interest-rate differ-

entials fail to widen, but the consensus forecasts of pri-

vate economists, the risk assessments by credit rating

agencies, and the economic outlook reports of the IMF

and the World Bank also failed to provide adequate warn-

ing of the impending currency crises. Even in those cases

where doomsayers might have correctly anticipated the

onset of an impending currency crisis, the eventual ex-

change-rate collapse often exceeded the direst projections

of doom and gloom analysts.

If anything, the evidence seems to be more consistent

with the view that market participants are more often than

not taken by surprise when an attack occurs. Indeed, in-

vestors and borrowers often tend to be leaning the wrong

way in terms of their portfolio positioning at the time of an

attack. Once an attack is waged, investors and borrowers

are immediately forced to reposition their portfolios to avoid

excessive currency losses, and such portfolio reposition-

ing works to intensify selling pressure on the currency

under attack. It is this massive liquidation of vulnerable

positions that is largely responsible for the seemingly ex-

cessive exchange-rate moves that typically occur during

speculative attacks.

Can the Marketplace Adequately Anticipate the Onset of Currency Crises?

Mexican/U.S. Short-Term Yield Spread

Prior to the 1994 Mexican Peso Crisis

(91-Day Cetes less U.S. T-Bill Rate)

400

600

800

1000

1200

1400

1600

03-Jan-94 04-Mar-94 05-May-94 06-Jul-94 06-Sep-94 07-Nov-94

(Basis Points)

Asian/U.S. Short-Term Yield Spreads

Prior to the 1997 Asian Currency Crisis

(Local 3-Month Rates less U.S. T-Bill Rate)

200

400

600

800

1000

1200

01-Jul-96 30-Aug-96 31-Oct-96 01-Jan-97 04-Mar-97 05-May-97 04-Jul-97

Thailand---- Malaysia....

(Basis Points) Korea____

European/German Short-Term Yield Spreads

Prior to the 1992 ERM Crisis

(3-Month Euro-Rates less Euro-DM)

100

200

300

400

01-Jan-92 03-Mar-92 04-May-92 03-Jul-92 03-Sep-92

Spain----

(Basis Points) U.K.____

Source: Datastream

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

136 Deutsche Bank Foreign Exchange Research

Why does the market fail to adequately anticipate the on-

set of currency crises? To answer this question one needs

to develop a model of investor and borrower behavior that

explains why market participants might assign decreasing

weight to the probability of a currency collapse at a time

when the actual probability of a currency collapse is rising.

A variant of the model developed by Frankel and Froot

(1990) may help explain this phenomenon.

The Frankel-Froot model can be applied to the case of an

emerging-market economy that is maintaining a fixed ex-

change rate and where there may be some question

whether the fixed exchange-rate regime will prove to be

sustainable. Let's assume that investors’ expectations re-

garding the outlook for an emerging-market currency (eeEM)

that is pegged to the U.S. dollar are determined as a

weighted-average of two possible outcomes: (1) the fixed

exchange-rate regime continues to survive (eeEM=0), and

(2) the fixed exchange-rate regime comes under attack,

where the currency is expected to collapse at a rate equal

to eeEM= k (

being the expected percentage rate collapse

of the emerging-market currency). Investors are assumed

to assign a certain probability, w, to the existing exchange-

rate regime staying intact and another probability, 1-w, to

the regime collapsing. Let's assume that the interest-rate

differential between the emerging-market country and the

U.S., iEM - iUS, equals the market's "average" expected change

in the emerging-market exchange rate. Mathematically, the

average change in the expected exchange rate can be ex-

pressed as the weighted-average probability of the two

possible exchange-rate outcomes:

eeEM= iEM - iUS = w(0) + (1 - w)k

Investors are assumed to vary the weights,

and 1-

that they attach to the two possible exchange-rate out-

comes depending on how successful the authorities are

in preserving a stable, fixed exchange-rate regime. The

longer the fixed exchange-rate regime survives, the more

confident investors will be that the regime will continue to

survive in the future. In other words, investors are prone

to extrapolate favorable past behavior into the future much

like chartists do.

Hence, the longer a fixed exchange-rate regime survives,

the more weight investors will assign to the likelihood of

no exchange-rate change in the future (i.e.,

will rise) and

less weight to the likelihood of a currency collapse (i.e., 1-

will decline). As

rises and 1-

declines, the interest-

rate differential, iEM - iUS, will tend to decline over time. This

implies that the market will assign less weight to the pos-

sibility of a future currency collapse over time.

Thus, the long-term success of an existing exchange-rate

regime might in and of itself encourage investors to as-

sign less weight to the likelihood of a crisis even if the

existing regime is enjoying unwarranted excess credibil-

ity. As the exchange-rate misalignment builds, the market

"expectation" of a possible currency collapse could diverge

sharply from the actual risk of such a collapse.

The Longer a Fixed Exchange-Rate Regime Stays Intact,

the Less Weight Investors Assign to the Likelihood of a Crash

Market-Assigned

Weight to the

Likelihood of a

Currency Crash

Time

The Market’s Assessment of the Risk of an Imminent

Currency Crisis Might Differ Greatly from the True Risk

Probability of a

Currency Crisis

Exchange-Rate

Misalignment

True Probability

of a Crisis

Market’s Assessment of the

Risk of a Currency Crisis

Widening Gap

Between

Perception & Reality

Exchange-Rate Expectations and the Anticipation of Currency Crises

Source: Rosenberg (1998)Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 137

Currencies that undergo large and persistent real appre-

ciations are often vulnerable to a speculative attack. In-

deed, empirical studies find that overly appreciated real

exchange rates are the best single leading indicator of cur-

rency crises in emerging markets. How do real apprecia-

tions get started? This question can be answered with the

assistance of a simple stylized diagram of the equilibrium

and disequilibrium forces that contribute to large and per-

sistent real exchange-rate appreciations over time.

The Typical Path to a Currency Crisis—Persistent Real Appreciation,

Overvaluation, and Collapse

We break the typical causes of a sustained real exchange-

rate appreciation episode into four phases:

1. The correction of a currency's initial undervaluation

(from point

to point

2. A rise in the real long-run equilibrium exchange rate

(from point

to point

with the combined rise in the

real exchange rate from point

to point

considered

an equilibrium phenomenon),

3. Inflation inertia that contributes to a cumulative rise in

inflation that is not offset by a change in the nominal

exchange rate, and

4. A surge in capital inflows that drives the nominal ex-

change rate, and thus the real exchange rate, higher

(from point

to point

, with the rise in the real ex-

change rate from point

to point

considered to be

a disequilibrium phenomenon).

Typical Causes of Real Exchange-Rate Appreciation

Real

Exchange Rate

Time

Rise in Real Long-Run

Equilibrium Exchange Rate

Inflation Inertia and

Excessive Surge

in Capital Inflows

Correction of

Initial Undervaluation

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

138 Deutsche Bank Foreign Exchange Research

Exchange-Rate Targeting, Inflation Inertia, and Currency Overvaluation

One of the primary factors responsible for disequilibrium

increases in the real exchange rate has been inappropriate

exchange-rate policies. For instance, not allowing the nomi-

nal exchange rate to depreciate to offset a persistently high

underlying inflation rate can lead to a disequilibrium rise in

the real exchange rate. An overvalued exchange rate will

threaten the competitiveness of an emerging-market coun-

try, and make its currency vulnerable to a speculative at-

tack.

Many emerging-market countries use the nominal ex-

change rate as an anchor to help bring their inflation rates

down to those prevailing in the rest of the world. The in-

tent is to buy credibility with a hard-currency posture. By

pegging the nominal exchange rate or limiting its flexibil-

ity, the authorities are compelled to adjust their monetary

policies to satisfy their nominal exchange-rate objective.

As a result, they are restricted from pursuing overly ex-

pansionary policies for purely domestic reasons.

In theory, a rigid nominal exchange-rate peg should allow

an emerging-market nation to bring its inflation rate down

rapidly to match the inflation performance of hard-currency

industrial countries. In practice, however, this seldom hap-

pens. Instead, backward-looking expectations, particularly

if the emerging-market country has a long history of failed

stabilization efforts, and structural impediments—such as

wage-indexation programs—work to prevent prices and

wages from falling rapidly into line with those in the indus-

trial world.

Combining an inertia-laden decline in emerging-market/in-

dustrial-country inflation differentials with a pegged nomi-

nal exchange rate, will give rise to a disequilibrium increase

in the real exchange rate. Policy decisions that permit a

buildup in the cumulative inflation gap without an offset-

ting depreciation of the nominal exchange rate are a recipe

for a currency crisis.

Inflation Inertia and Real Exchange-Rate Appreciation

Nominal

Exchange

Rate

Time

Inflation

Differential

Time

Real

Exchange

Rate

p-p

EM US

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 139

Surges in Capital Inflows and Real Appreciation

Surges in capital inflows have also been an important

source of real appreciation for many emerging-market cur-

rencies. Surges in capital inflows affect the real exchange

rate via two channels. First, where there is scope for nomi-

nal exchange-rate flexibility, a surge in private capital in-

flows will directly drive the value of the nominal exchange

rate higher and, in the process, drive the real exchange

rate higher in tandem. Second, a surge in capital inflows

can indirectly drive the real exchange rate higher if the

authorities intervene to stabilize the nominal exchange rate

and the resulting increase in the central bank's foreign-

exchange reserves fuels a major increase in the domestic

money supply (assuming the intervention move is not com-

pletely sterilized). An increase in domestic money-supply

growth will then tend to fuel a rise in domestic demand

and in the process drive domestic inflation higher. A rise in

domestic inflation without an offsetting depreciation of the

domestic currency will then drive the real exchange rate

higher. Hence, both directly and indirectly, a sustained

surge in capital inflows is likely to drive the real exchange

rate higher in emerging markets.

In many episodes of significant real appreciation in emerg-

ing markets, excessive optimism on the part of interna-

tional investors and domestic borrowers has led to exces-

sive surges in capital inflows that have caused real ex-

change rates to overshoot their long-run equilibrium levels

by a wide margin. At times, a speculative bubble in the

real exchange rate might form, making the emerging-mar-

ket currency vulnerable to a possible collapse in the event

of a speculative attack. When a currency moves along a

bubble path, its rise into deeply overvalued territory is

clearly not sustainable in the long run, but positive feed-

back trading and bandwagon effects are capable of sus-

taining an exchange-rate overshoot for short and medium-

term periods.

Capital Inflows and the Real Exchange Rate

Surge in

Capital

Inflow

Increase in

Foreign-Exchange

Reserves

Increase

Money Supply

Increase

Inflation

Nominal

Exchange-Rate

Appreciation

Real

Exchange-Rate

Appreciation

Positive Feedback Trading and

Excessive Surges in Capital Inflows

Surge in

Capital

Inflow

Real

Appreciation Excess

Returns

Increase

Foreign-Exchange

Reserves

Expectation

Additional

Above-Average

Returns

Increase

Money Supply

Increase

Economic

Activity

Lower

Budget

Deficit

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

140 Deutsche Bank Foreign Exchange Research

Capital Flows, Excessive Leverage, and Crisis Vulnerability

An excessive surge in capital inflows can plant the seeds

of a currency crisis by contributing to: (1) an unwarranted

real appreciation of the emerging-market currency, (2) a

huge build-up in external indebtedness, (3) a financial as-

set or property-market bubble, (4) a consumption binge

that contributes to an explosive growth in domestic credit

and/or the current-account deficit, or (5) over-investment

in risky projects and questionable activities.

It is no accident that the ERM, Mexican, and Asian cur-

rency crises were all preceded by a surge in capital in-

flows. In each case, the surge in capital inflows led to a

buildup of huge, highly leveraged, long speculative posi-

tions by local as well as international investors in the cur-

rencies that eventually came under heavy attack. For ex-

ample, in the run-up to the ERM crisis, investors, believing

that European yield convergence would occur as European

Monetary Union (EMU) approached, took on highly lever-

aged long positions in the high-yielding European curren-

cies and short positions in the lower-yielding Deutsche-

mark. Likewise, in the run-up to the Mexican peso crisis,

investors and banks were highly leveraged and made ex-

tensive use of derivative products in taking on speculative

long peso/short dollar positions. And in the run-up to the

Asian currency crisis, it was Asian firms and banks that

were highly leveraged as they took on a huge volume of

short-term dollar-denominated debt to fund local activities.

The seemingly excessive currency collapses (at least rela-

tive to underlying economic fundamentals) that followed

in each of those episodes simply reflected the sudden

unwinding of those leveraged positions at the time of each

attack.

When investors assign a smaller and smaller probability to

an imminent currency collapse, they often tend to become

overconfident that no attack will occur, and thus appear

willing to take on highly leveraged, long speculative posi-

tions in currencies that are in reality vulnerable to a sud-

den collapse. What actually triggers a collapse is less im-

portant than the fact that at the time of an attack investors

tend to be leaning, perhaps heavily, the wrong way in terms

of their portfolio positioning. When the attack comes, fear

takes over, as highly leveraged investors need to unwind

their vulnerable long positions. It is the unwinding of those

long positions that causes exchange rates to depreciate

violently at the time of an attack.

Excessive Surges in Capital Inflows

Often Plant the Seeds of Future Currency Crises

Excessive Surge

Capital Inflow

Real

Appreciation

Increase in

External Indebtedness

Current-Account

Deficit

Excessive Growth

in Domestic Credit

Equity and/or

Property Bubble Currency

Crisis

Over-investment

in Risky Projects

Source: Rosenberg (1998)

Investors and Borrowers Are Often Caught Leaning the Wrong Way

at the Time of an Exchange-Rate Collapse

Net Long Positions

in Emerging-Market

Currency

A’

Time

Nominal

Exchange

Rate

Time

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 141

Real Exchange-Rate Misalignment and Collapse

A currency that undergoes significant real appreciation will

be vulnerable to a speculative attack only if the actual real

exchange rate (q) rises relative to its real long-run equilib-

rium value (q). In such cases, the resulting overvaluation

of the real exchange rate relative to its long-run equilib-

rium level should be viewed as a "disequilibrium" rise in

the real exchange rate that would call for an eventual cor-

rection of the misaligned exchange rate.

Since the real long-run equilibrium exchange rate, q, is not

directly observable, it will never be known for sure whether

a rise in the actual real exchange rate, q, is an "equilibrium"

or a "disequilibrium" phenomenon. This will be especially

true for small and medium-sized real appreciations. How-

ever, for large and sustained real appreciations, the odds

of a persistent rise in q being a purely equilibrium phe-

nomenon are likely to be small. Instead, large and persis-

tent real appreciations are more likely to be disequilibrium

phenomena in need of correction. Indeed, in a recent ex-

haustive study of real appreciation episodes in 93 coun-

tries over the 1960-94 period, Goldfajn and Valdes (1996)

found that virtually all currencies that underwent real ap-

preciations exceeding 35% from trough to peak eventu-

ally collapsed either through a speculative attack or a sur-

prise devaluation.

Views on what real exchange-rate level constitutes a

currency's real long-run equilibrium value often differ among

market participants. Hence, views on the magnitude of

the prevailing real exchange-rate misalignment will differ

as well. Given such high levels of uncertainty, the market

will generally presume that a sustained increase in the

real exchange rate is an equilibrium phenomenon unless

convinced otherwise. Hence real exchange-rate

misalignments are unlikely to be corrected until there are

visible signs of fundamental economic deterioration.

In a typical real appreciation/misalignment episode, real

appreciation tends to proceed gradually. As the exchange-

rate misalignment builds, there will probably be no correc-

tive adjustment early on, since economic fundamentals

likely will display little sign of significant deterioration. But

as the real exchange-rate misalignment grows and persists,

the trend in economic fundamentals will gradually begin

to deteriorate. When the deterioration in fundamentals

reaches a point that convinces speculators that remedial

action is called for, then and only then will they attack the

overly appreciated currency.

Evidence presented by Goldfajn and Valdes (1996) suggests

that real appreciations tend to build up gradually and then

end abruptly. They find that there is a large asymmetry

between the average duration of real appreciation episodes

and the time that it takes to unwind the appreciation of

the real exchange rate. Goldfajn and Valdes estimate that

the average duration of real appreciation episodes is about

twice as long as the time that it takes to unwind the real

appreciation.

Real Exchange-Rate Misalignments Tend to

Build Up Gradually and End Abruptly

Economic

Fundamentals

A’

Time

Real

Exchange

Rate

Time

Source: Rosenberg (1998)

Probability That a Real Appreciation Episode

Will Come to an End without a

Speculative Attack or Surprise Devaluation

0.00

0.06

0.11

0.23

0.32

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

15 20 25 30 35

Probability

Source: Goldfajn and Valdes (1996).

Real Currency Appreciation(%)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

142 Deutsche Bank Foreign Exchange Research

Current-Account Imbalances and Currency Crises

Countries that run large and persistent current-account

deficits are likely to see their currencies become vulner-

able to a speculative attack. That appeared to be the case

for European countries prior to the ERM crises of the early

1990s.

When a country is running a large and persistent current-

account deficit, it must attract continual capital inflows to

finance the deficit. Over time, this will give rise to a large

and growing external debt. If this external debt burden

becomes inordinately large, it might raise questions about

the solvency of the indebted nation. If the market sus-

pects that the indebted nation does not have the ability to

generate sufficient trade surpluses in the future to repay

its debt, they might reason that exchange rates will need

to be adjusted downward to boost the competitiveness of

the indebted nation.

Italy’s Current-Account Balance

(1987-1992)

-2.5

-2.1

-1.6

-1.5

-0.8

-0.3

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

1987 1988 1989 1990 1991 1992

(% of GDP)

Source: OECD Economic Outlook, June 1998.

U.K. Current-Account Balance

(1987-1992)

-1.7

-1.4

-3.5

-4.3

-3.5

-1.1

-5

-4

-3

-2

-1

1987 1988 1989 1990 1991 1992

(% of GDP)

Source: OECD Economic Outlook, June 1998

Spain’s Current-Account Balance

(1987-1992)

-3.7

-3.8

-3.7

-2.9

-1.1

0.0

-4.0

-3.0

-2.0

-1.0

0.0

1987 1988 1989 1990 1991 1992

(% of GDP)

Source: OECD Economic Outlook,

June 1998

Sweden’s Current-Account Balance

(1987-1992)

-3.5

-2.0

-2.7

-1.6

-0.3

0.0

-4.0

-3.0

-2.0

-1.0

0.0

1987 1988 1989 1990 1991 1992

(% of GDP)

Source: OECD Economic Outlook, June 1998.

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 143

While current-account imbalances may have played an

important role in triggering the Mexican peso and Asian

currency crises in the 1990s, most international econo-

mists would argue that the violent character of those cur-

rency-crisis episodes could be explained only by financial-

sector weaknesses, and not by external-balance consider-

ations.

What does the empirical evidence suggest about the role

of current-account imbalances in triggering currency cri-

ses in emerging markets? Frankel and Rose (1996) exam-

ined the fundamental forces that triggered currency

crashes in over 100 emerging-market countries over the

1971-92 period, and found that current-account imbalances

were not statistically significant in predicting currency

crashes. Sachs, Tornell, and Velasco (1996) found that cur-

rent-account imbalances were not helpful in determining

which currencies were vulnerable to contagion effects fol-

lowing the Mexican peso collapse in 1994-95. However,

Goldstein et al. (2000) and the IMF (1998) found that cur-

rent-account deficit/GDP ratios have, on average, been sig-

nificantly larger in periods leading up to a crisis than dur-

ing tranquil periods. The latter results suggest that current-

account imbalances may have some predictive power in

anticipating currency crises.

Mexico’s Current-Account Balance

(1990-1994)

-7.1

-5.8

-6.7

-4.6

-2.8

-8

-6

-4

-2

1990 1991 1992 1993 1994

(% of GDP)

Source: OECD Economic Outlook, June 1998.

Malaysia’s Current-Account Balance

(1992-1997)

-5.1

-4.9

-9.9

-7.8

-4.8

-3.8

-12

-10

-8

-6

-4

-2

1992 1993 1994 1995 1996 1997

(% of GDP)

Source: National government data.

Thailand’s Current-Account Balance

(1992-1997)

-0.8

-8.1

-8.0

-5.6

-5.1

-5.7

-10

-8

-6

-4

-2

1992 1993 1994 1995 1996 1997

(% of GDP)

Source: National government data.

Indonesia’s Current-Account Balance

(1992-1997)

-2.4

-3.4-3.4

-1.7

-1.5

-2.3

-4

-3

-2

-1

1992 1993 1994 1995 1996 1997

(% of GDP)

Source: National government data.

Korea’s Current-Account Balance

(1992-1997)

-1.8

-4.8

-1.8

-1.0

0.3

-1.3

-6

-5

-4

-3

-2

-1

1992 1993 1994 1995 1996 1997

(% of GDP)

Source: OECD Economic Outlook, June 1998.

Philippine’s Current-Account Balance

(1992-1997)

-4.5

-4.7

-4.4

-4.6

-5.6

-1.6

-6

-5

-4

-3

-2

-1

1992 1993 1994 1995 1996 1997

(% of GDP)

Source: National government data.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

144 Deutsche Bank Foreign Exchange Research

Overly Expansionary Monetary Policy and Currency Crises

Countries that pursue overly expansionary monetary poli-

cies are likely to see their currencies become vulnerable

to a speculative attack. History is replete with examples in

which domestic credit creation by a central bank in excess

of money demand has planted the seeds of a speculative

attack.

If a central bank wished to peg its currency to a hard cur-

rency, it would need to pursue monetary policies consis-

tent with that exchange-rate objective. An overly expan-

sionary monetary policy would be inconsistent with the

maintenance of stable nominal exchange rates in the long

run. Examples of inconsistent policies would be situations

in which a central bank felt compelled to monetize a large

and persistent budget deficit or had to inject massive sums

of liquidity to prop up an ailing banking system. In both

cases, such policy stances would inevitably lead to a col-

lapse of a pegged exchange-rate regime.

We can illustrate the inherent inconsistency between the

pursuit of a stable nominal exchange rate, on the one hand,

and an excessively expansionary monetary policy, on the

other, with a simple model. We start by assuming that

changes in a currency's value, e, are inversely related to

changes in that country's monetary base, M:

e = - M (1)

That is, an expansion of the monetary base will lead to an

equiproportionate decline in a currency's value, while a

contraction of the monetary base will lead to an

equiproportionate rise in a currency's value. In this simpli-

fied model, a fixed exchange rate can be maintained over

time only if the monetary base is kept constant over time.

That is, if:

M = 0 (2)

then, from (1) above

e = 0 (3)

A country's monetary base consists of domestic credit

issued by the central bank, D, and the central bank's hold-

ings of foreign-exchange reserves, R:

M = D + R (4)

If a central bank attempted to monetize a large budget

deficit by expanding domestic credit, D, that would lead to

a corresponding rise in the monetary base, M, unless the

central bank’s holdings of reserves, R, were run down by

an amount equal to the rise in D. If the central bank's hold-

ings of foreign-exchange reserves, R, were run down by

an amount equal to the increase in domestic credit, D, the

monetary base would be held constant and the nominal

exchange rate would remain fixed.

Since a central bank's holdings of foreign-exchange re-

serves are finite, there is a limit to how much a central

bank can continuously run down its reserves to offset a

continuous expansion of domestic credit. At the point

where the central bank's holdings of foreign-exchange re-

serves approach zero (shown as t1 in the diagram above),

an exchange-rate peg would collapse on its own accord

because at that point the monetary base would begin to

expand at the same rate as domestic credit was expand-

ing. That is, if

R = 0 (5)

then changes in the monetary base, M, would be entirely

determined by changes in domestic credit, D:

M = D (6)

With both D and M rising at the same rate, the rise in M

would no longer be consistent with the maintenance of a

stable exchange rate. Once M starts to expand, the ex-

change rate would begin to decline at the same rate that

M is rising. In practice, if forward-looking investors saw

that the central bank’s FX reserves were falling to danger-

ously low levels, they would, in all likelihood, wage an at-

tack on the vulnerable currency prior to point t1 being

reached.

Overly Expansionary Monetary Policies, Reserve Loss,

and the Inevitability of an Exchange-Rate Collapse

Time

Monetary

Base

Time

Foreign-

Exchange

Reserves

Domestic

Credit

Nominal

Exchange

Rate

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 145

M2/FX Reserve Ratio as a Leading Indicator of Currency Crises

The ratio of M2 money supply to the central bank's hold-

ings of foreign-exchange reserves has been found to be a

reliable leading indicator of currency crises in emerging

markets in a number of recent empirical studies. Kaminsky,

Lizondo, and Reinhart (1998) found that large increases in

a country's M2/reserve ratio tended to precede currency

crises in a majority of cases studied. In addition, Sachs,

Tornell, and Velasco (1996) found that countries with rela-

tively high M2/reserve ratios tended to be more vulner-

able to the contagion effects associated with the Mexican

peso crisis (the so-called Tequila effect) than those coun-

tries with relatively low M2/reserve ratios. Furthermore,

Goldstein and Hawkins (1998) found that Asian countries

had considerably higher M2/reserve ratios than their Latin

American counterparts in mid-1997, just prior to the Asian

financial crisis. This made the Asian currencies more vul-

nerable to a speculative attack. Finally, the International

Monetary Fund (IMF), in its May 1998 World Economic

Outlook report, noted that the M2/reserve ratio displayed

a "remarkable pattern around the time of a crisis". Accord-

ing to the IMF, the M2/reserve ratio tended to rise sharply

in the 24 months leading up to a crisis and then to plum-

met sharply once a crisis got underway.

What makes the M2/reserve ratio attractive as a leading

indicator is that it captures the extent to which a central

bank has sufficient foreign-exchange reserves on hand to

absorb a run by deposit holders out of local-currency de-

posits into foreign currency. If the M2/reserve ratio is rela-

tively high, it would suggest that a central bank may not

have sufficient resources to directly satisfy all of the de-

mands of deposit holders if everyone attempted to move

out of domestic-currency deposits at the same time. With-

out sufficient reserve backing, the emerging-market cur-

rency would be vulnerable to an attack. In contrast, if an

emerging-market country had a relatively low M2/reserve

ratio, that would imply that the central bank had sufficient

reserves to back the domestic-currency liabilities of the

banking system. As the IMF notes, the M2/reserve ratio

"can be viewed as an indicator of investor confidence in

the domestic financial system."

Latin American vs. Asian M2/Reserve Ratios

(Just Prior to the 1997 Asian Currency Crisis)

6.26.2

4.94.9

4.2

3.6

3.5

1.8

Chile BrazilArgentina Mexico

Philippines ThailandIndonesia Korea

Mid-1997 M2 Money Supply/FX Reserves(ratio)

Source: IMF

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

146 Deutsche Bank Foreign Exchange Research

High Unemployment as a Leading Indicator of Currency Crises

A deterioration in the domestic economic situation can also

act as an early-warning sign of a currency crisis. For in-

stance, substantial increases in unemployment rates to

unacceptably high levels appeared to be a key factor in

precipitating the collapse of two major fixed exchange-rate

regimes in the 20th century—the collapse of the Gold Stan-

dard in the 1930s and the collapse of the ERM in 1992-93.

If a country's unemployment rate rises to levels that are

deemed to be politically unacceptable, the monetary au-

thorities may face a major policy dilemma: interest rates

may need to be lowered to stimulate growth and alleviate

the unemployment problem but, at the same time, inter-

est rates may need to be raised to defend a potentially

vulnerable exchange rate.

Recognizing that policymakers may face a growing di-

lemma in an environment of weak domestic demand,

speculators will reason that the authorities' commitment

to defend the exchange rate may gradually wane as the

unemployment rate marches steadily higher. As the cred-

ibility of an exchange-rate commitment wanes, higher in-

terest rates may be needed to persuade skeptical inves-

tors that the existing exchange-rate regime is still viable.

Unfortunately, since higher interest rates will act to slow

domestic demand and aggravate the unemployment prob-

lem, attempts to boost credibility through higher interest

rates may actually backfire and undermine the credibility

of the exchange-rate peg even further.

Since forward-looking market participants will recognize

that a high and rising unemployment rate will erode the

credibility of an existing exchange-rate regime, they are

likely to test policymakers' resolve and wage an attack

before the authorities voluntarily abandon their exchange-

rate commitment. In fact, if policymakers are put to the

test by a speculative attack, they may abandon their de-

fense of the existing exchange-rate regime because an

interest-rate defense likely will prove to be too costly.

Source: Rosenberg (1998)

Unsustainably High Real Interest Rates Can Give Rise to

Weaker Economic Activity and Rising Unemployment,

Thereby Making a Currency Vulnerable to a Speculative Attack

Economic

Activity

A’

Time

Real

Interest

Rate

Time

Nominal

Exchan

Rate

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 147

Debt Crises and Currency Crises

When an emerging-market country's government or pri-

vate sector amasses a large volume of short-term debt, it

places its currency in a vulnerable position where a self-

fulfilling panic can lead to a sudden and massive specula-

tive attack. For example, a country that issues a lot of short-

term local-currency debt could run into problems if, for

whatever reason, domestic and foreign creditors refuse to

roll over the existing stock of debt.

In such a case, the local monetary authorities might feel

compelled to push domestic short-term interest rates up

sharply to induce the investment community to buy and

hold the large stock of short-term debt. Unfortunately, high

and rising short-term interest rates can place the highly

indebted public or private sectors of an emerging market

in a vulnerable position directly—by increasing the cost of

servicing the large stock of local-currency debt—or indi-

rectly—by leading to a slowdown in domestic economic

activity, which acts to reduce public and private sectors'

revenue intake and thereby makes it even more difficult to

service the debt.

A rising local-currency debt-service burden poses three

major risks for an emerging-market currency. First, it might

raise questions about long-term solvency, which could

spark a sudden panic withdrawal of funds by local and for-

eign creditors. Second, it might raise fears that the local

authorities will attempt to reduce the real value of the debt

burden by inflating. Third, a large local-currency debt-ser-

vice burden could constrain the local authorities from pur-

suing a tight monetary policy that might be necessary to

defend their currency from speculative attack. All of these

risk factors highlight the fact that currency vulnerability is

likely to rise when local-currency debt-service burdens are

large and growing.

Currency vulnerability is also likely to be high when an

emerging-market economy amasses a large volume of

short-term foreign-currency debt. Indeed, short-term for-

eign-currency debt has the potential to impose a greater

burden on an emerging-market country than local-currency

debt because the local authorities cannot reduce the real

burden of a foreign-currency debt obligation by inflating.

In addition, given that the local-currency value of a foreign-

currency debt obligation is sensitive to changes in the ex-

change rate, currency depreciation could be extremely

costly, since that would directly raise the local-currency

value of the foreign-currency debt. Crisis-prone currencies

are therefore likely to be in an especially vulnerable state

when short-term foreign-currency debt obligations are high

and rising.

A large buildup in short-term foreign-currency debt, by it-

self, will not precipitate a currency crisis if foreign credi-

tors are willing to extend new loans and roll over any exist-

ing loans. However, if for whatever reason fear strikes in

The Local-Currency Value of Short-Term Foreign-Currency Debt

Tends to Soar Following a Major Currency Collapse

Time

US$ Value

of Dollar-

Denominated

Debt

Time

Nominal

Exchange

Rate

Local-Currency

Value

of Dollar-

Denominated

Debt

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

148 Deutsche Bank Foreign Exchange Research

the hearts of foreign creditors, and foreign funds are with-

held, the emerging-market currency could come under

heavy downward pressure. If so, the local-currency value

of the short-term foreign-currency debt could soar, possi-

bly throwing otherwise viable local entities into insolvency.

The importance of short-term foreign-currency debt as a

measure of financial vulnerability is highlighted by the fact

that debt crises played an important role in the currency

crises in both Mexico and Asia. Indeed, the buildup of a

large volume of short-term foreign-currency debt by the

public sector in Mexico and the private sector in Asia is

widely credited with planting the seeds of the Mexican

peso and Asian currency crises. In both situations, the lo-

cal-currency value of the foreign-currency debt soared as

the peso and Asian currencies came under heavy down-

ward pressure.

The local-currency value of foreign-currency debt is highly

sensitive to changes in the exchange rate. A sudden sharp

depreciation of an emerging-market currency will instantly

raise the local-currency value of all short-term foreign-cur-

rency-denominated debt, and by so doing, will dramati-

cally increase the burden of servicing that debt. As indebt-

edness soared in Mexico and Asia, default risk rose sharply

and, as a result, overseas private capital was no longer

forthcoming. With no funds forthcoming, the trend in eco-

nomic fundamentals deteriorated suddenly and markedly.

The stylized diagram below demonstrates that a sharp rise

in the local-currency value of short-term foreign-currency

debt places a nation's economy in a vulnerable position if

soaring debt burdens make it difficult for local firms to ser-

vice their debts. Note that right up until the time of a specu-

lative currency attack, a highly indebted nation may not

show any visible signs of economic weakness. Once an

attack takes place and the local-currency value of short-

term foreign-currency debt soars, the trend in economic

fundamentals is shown to suddenly deteriorate as firms

go bankrupt, banks fail, and the government's fiscal bal-

ance deteriorates.

A country's economic vulnerability might not reveal itself

prior to a debt-induced attack, and there may be few signs

suggesting that an attack is imminent. An emerging-mar-

ket economy might successfully leverage off of relatively

cheap short-term foreign-currency debt and post an im-

pressive growth performance for a long while without trig-

gering a crisis, much as Asia did in the decade leading up

to the crisis in 1997. In fact, Asia's solid growth performance

might have deluded the majority of market participants into

believing that Asia's underlying fundamentals were sound

when, in fact, they were all the time vulnerable if the value

of the Asian currencies, for whatever reason, suddenly

declined.

This has important implications for modeling currency cri-

ses in emerging markets. Signs of fundamental economic

deterioration may not be visible going into a crisis. While

evidence of deteriorating fundamentals may be useful in

predicting certain currency crises, there may be few reli-

able guideposts available to predict when a debt-driven

currency crisis might occur. A large foreign-currency debt

burden may be a sign of currency vulnerability in the long

run, but there may simply be no way of knowing before-

hand when an attack might actually be waged.

A Sharp Increase in the Local-Currency Value of

Short-Term Foreign-Currency Debt Could Lead

to a Sudden Deline in Economic Activity

Economic

Fundamentals

A’

A’’

Time

Local-Currency

Value of Dollar-

Denominated

Debt

Time

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 149

Banking Crises and Currency Crises

Many countries that have faced serious currency crises in

recent years have also experienced severe banking crises

at the same time. This is no mere coincidence. Recent

empirical work conducted by Kaminsky and Reinhart (1997)

finds that banking crises are a good leading indicator of

currency crises. According to Kaminsky and Reinhart, more

than 50% of all banking crises in the past 25 years have

been followed by a currency crisis within three years, while

about 25% of all banking crises were followed by a cur-

rency crisis within one year. The fact that banking crises

tend to lead currency crises does not imply that banking

crises are the primary cause of currency crises. What the

evidence does suggest is that a common set of funda-

mental economic factors determines whether a banking

system and a currency might both be vulnerable to a cri-

sis. When the trend in economic fundamentals is inexora-

bly deteriorating, banking systems have tended to experi-

ence greater stress early on, followed by the currency

market.

Not all currency crises are preceded by a banking crisis.

However, when banking crises are present, they do tend

to aggravate and accentuate the crisis on the currency front.

This is true of both developed and developing countries.

Banking system problems in Scandinavia in the early 1990s

and in Japan in the second half of the 1990s accentuated

their exchange-rate depreciations during those periods.

Banking crises in Mexico and Asia are widely credited with

accentuating the slide of the Mexican peso and Asian cur-

rencies during their crisis periods in the 1990s.

In a typical banking/currency crisis, the seeds of a banking

crisis are often planted when an emerging-market country

takes steps to liberalize its financial markets. When finan-

cial liberalization is accompanied by other economic re-

forms, capital inflows from abroad often increase substan-

tially. This surge in capital inflow is then intermediated by

the banking sector to the rest of the emerging-market

economy via a lending boom. The lending boom typically

gives rise to an asset-price bubble in the equity and/or prop-

erty market and often encourages over-investment or over-

consumption. Given that emerging-market banks on aver-

age are less regulated and not as adequately supervised

as their counterparts in the industrial world, they are more

likely to engage in questionable and risky lending activi-

ties.

The excesses in a typical banking crisis are ultimately un-

wound when the asset-price bubble bursts and bankrupt-

cies climb. As bankruptcies climb, this gives rise to an in-

crease in nonperforming loans on emerging-market banks'

books. If nonperforming loans are large relative to the ag-

gregate loan portfolio, banks may become technically in-

solvent, which in turn may prompt a run on the banking

system. In a panic phase, depositors are prone to flee both

weak and healthy banks, thereby placing a strain on the

entire banking system.

A banking crisis is likely to lead to a significant downturn

in economic activity. As economic activity weakens, there

will be great pressure on the monetary authorities to take

corrective action to stimulate monetary policy. If the reso-

lution of a banking crisis requires a costly fiscal bailout,

the monetary authorities may be called upon to monetize

part or all of any increase in deficit financing. Hence, a

banking crisis may indirectly set off a currency crisis by

requiring the monetary authorities to pursue an excessively

expansionary monetary policy.

Excessive Surges in Capital Inflows

Often Plant the Seeds of a Future Banking Crisis

Surge in

Capital

Inflow

Asset

Price

Bubble

Increase in

Questionable

and Risky

Loans

Bank

Lending

Boom

Bubble

Bursts

Increase in

Non-

performing

Loans

Over-

consumption

and

Over-

investment

Banking

Crisis

Banking Crises Often Lead Currency Crises

Banking

Crisis

Decline in

Economic

Activity

Dramatically

Easier

Monetary

Policy

Costly

Fiscal

Bailout

Decline in

Central

Bank’s

Holdings

Foreign-

Exchange

Reserves

Decline in

Credibility

of Existing

Exchange-

Rate

Regime

Capital

Flight

Currency

Crisis

Source: Rosenberg (1998) Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

150 Deutsche Bank Foreign Exchange Research

Contagion—How Currency Crises Spread from One Market to Another

Each of the major currency crises of the 1990s has had an

element of contagion that saw speculative pressures

spread rapidly from one currency to another. For example,

in the ERM crisis of 1992-93, the collapse of the Italian lira

spread rapidly to the U.K., Spain, and Scandinavia, and then,

after a one-year respite, spread to France and Belgium.

During the Mexican peso crisis of 1994-95, speculative

pressures spread rapidly to a number of markets, with Ar-

gentina the most adversely affected. In the case of the

Asian currency crisis of 1997-98, the collapse of the Thai

baht spread both north and east to affect most countries

in the region.

While economists agree that currency crises often con-

tain a powerful contagion element, there is disagreement

as to the underlying causes of contagion. There are actu-

ally several channels through which contagion effects can

cause speculative attacks to spread from one currency to

another. First, strong bilateral or third-party trade links can

be an important source of contagion. For example, if the

currency of country A declines and this adversely affects

the competitiveness of country B, this might raise the vul-

nerability of country B's currency to speculative attack.

Second, similar macroeconomic performances might lie

behind contagious speculative attacks. If country A's cur-

rency came under attack because of poor current-account

balance or economic growth prospects, this might lead

investors to believe that country B, which might have similar

weak macroeconomic prospects, could also see its cur-

rency come under attack.

Third, financial linkages in the form of common lenders/

creditors, market liquidity, and investor risk appetite can

be important sources of contagion. For example, asset-

allocation shifts by global investors might play a role in the

spillover of currency crises from one country to another.

Global fund managers with exposure to the emerging

markets often diversify their investments among several

emerging markets. If one of the currencies in a fund

manager's emerging-market portfolio were to come un-

der heavy speculative attack, the fund manager—con-

cerned about the possibility of fund withdrawals by indi-

vidual investors—might decide to build up cash positions

and to reduce the fund’s overall exposure to the entire

basket of emerging-market currencies in his portfolio. If

other fund manager portfolios are rebalanced in a similar

fashion, this may place downward pressure on a large num-

ber of emerging-market currencies.

The Channels of Contagion

Crisis

Country “A”

Bilateral

Trade Linkages

Level of

Market Liquidity

Common

Lender

Trade with a

Common Third Party

Crisis

Country “B”

Global Reduction in

Appetite for Risk

Source: Rosenberg (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 151

Fourth, a general decline in global investors' appetite for

risk might generate contagion among emerging-market

currencies if investors reduce their exposure to all risky

assets. This typically occurs during times of global finan-

cial or economic stress.

Finally, if policymakers in an individual emerging market

abandon the defense of a pegged exchange rate because

the cost of defending the exchange rate against attack

proves to be too great, then this action might contain infor-

mation on the ability and willingness of policymakers in

other countries to withstand the pressures of speculative

attacks.

According to a recent IMF analysis of emerging-market

contagion effects, trade linkages probably played a small

role in the spread of currency crises from one country to

another in the 1990s. In contrast, financial linkages appear

to have played a larger role, particularly through changes

in investor risk appetite and the existence of common lend-

ers.

Contagion and Its Causes

(Contagion Effects after the Mexican, Thailand, and Russian Financial Crises)

Bilateral Trade Trade with Level of

with Initially a Common Market Global

Affected Third Common Reduction in

Country 1Party 2Lender 3Liquidity 4Appetite for Risk 5

Mexico (December 1994)

Argentina — Low Yes, little exposure Low Moderate decline in risk appetite in January 1995.

Brazil — Low Yes, little exposure High

Thailand (July 1997)

Hong Kong SAR — Low No High Modest decline in risk appetite in May 1997, but not sustained.

Indonesia — Low Yes, moderate exposure Low

Malaysia Low High Yes, moderate exposure Moderate

Philippines Low Moderate No Low

South Korea — Moderate Yes, high exposure Moderate

Russia (August 1998)

Brazil — — No High Marked decline in risk appetite in August and September.

Hong Kong SAR — — No High

Mexico — — No Moderate/High

Sources: IMF World Economic Outlook (December 2001); Kaminsky and Reinhart (2000); Kumar and Persaud (2001); and Glick, and Rose, 1999.

1 Exposure through bilateral trade is measured by the share of the country’s total exports destined to the initial crisis country.

2 Trade with a common third party in the same commodities is measured as the percentage of total exports competing with the top exports of initial crisis country.

3 For a discussion of how Bank of International Settlements data can be used to identify common bank lender clusters, see Kaminsky and Reinhart (2000).

For bonds, see J.P. Morgan’s EMBI+ weights.

4 Market liquidity is roughly proxied by the country’s representation (its share) in the global mutual funds’ emerging market portfolio.

High, moderate, and low classifications are comparisons with respect to other emerging markets.

5 For a description of the methodology used to estimate risk appetite, see Kumar and Persaud (2001).

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

152 Deutsche Bank Foreign Exchange Research

Devising an Early-Warning System to Anticipate Currency Crises

Since currency crises tend to be triggered in countries with

a number of economic problems, not just one problem, an

ideal early-warning system should be broad-based, incor-

porating a wide range of symptoms that crisis-prone cur-

rencies might exhibit prior to a speculative attack. For in-

stance, a typical crisis-prone currency might exhibit a num-

ber of symptoms such as an overly appreciated real ex-

change rate, a major terms-of-trade shock, a deteriorating

current account, weak economic activity and rising unem-

ployment, an overly expansionary fiscal policy, an overly

expansionary monetary policy, an excessive increase in

bank lending, declining foreign-exchange reserves at the

central bank, a high ratio of M2 money supply to central

bank reserves, falling equity and property prices, a sharp

increase in nonperforming loans on bank books, and/or a

sharp increase in short-term debt incurred by the domes-

tic private or public sector.

Kaminsky and Reinhart (1998) argue that an ideal early-

warning system should focus on the behavior of a group

of indicators, and not on the behavior of a single indicator

to predict the onset of currency crises. They found that

currency crises in Latin America and Asia over the 1970-

95 period were preceded by a multitude of economic and

financial problems, not just one. Examining the behavior

of 15 macroeconomic indicators prior to the onset of the

currency crises in Latin America, Kaminsky and Reinhart

found that in 44% of the crisis episodes, 80%-100% of

the indicators were flashing a warning signal prior to each

crisis episode. In another 39% of the crisis episodes, 60%-

80% of the indicators were flashing prior to each attack.

Similar results were found for currency crisis episodes in

Asia and elsewhere.

The table below provides an extensive list of indicators

that can help detect whether potentially vulnerable cur-

rencies have become "symptomatic" in terms of display-

ing signs that a currency crisis might be imminent. Con-

sidering the broad range of symptoms that a potentially

vulnerable currency might exhibit, a rule of thumb might

be that the probability of an imminent currency crisis would

be considered to be high if a relatively large number of

indicators started flashing. An absence of flashing indica-

tors would suggest that the probability of an imminent cri-

sis was low.

Indicators of Currency Crisis Vulnerability

Symptoms Indicators

Real Exchange Rate Overvaluation Behavior of real exchange rate relative to trend,

exports, imports, trade balance.

Terms of Trade Shock Price of exports, price of imports, commodity prices.

Current Account Imbalance Real exchange rate, savings, investment.

Weak Economic Activity/ Real GDP, output gap, unemployment rate,

Rising Unemployment real interest rate.

Overly Expansionary Fiscal Policy Government spending, budget deficit,

volume of credit extended to public sector.

Overly Expansionary Monetary Policy Domestic credit expansion, reserves,

money-supply growth, money multiplier.

High Ratio of M2 to Reserves M2, international reserves.

Banking Crises Equity and property prices, non-performing loans,

lending/deposit rate spread, bank share prices.

Debt Crises Total debt, domestic debt, foreign-currency debt,

short-term/total debt.

Contagion Foreign growth, foreign interest rates, crises elsewhere.

Performance of Leading Indicators

Prior to Currency Crises

Percentage of Indicators Showing

Anomalous Behavior Prior to Crisis

Quintiles Latin America Asia Others

80%-100% 44% 28% 32%

60%-79% 39% 42% 39%

40%-59% 8% 28% 18%

20%-39% 8% 0% 11%

less than 20% 0% 0% 0%

100% 100% 100%

Source: Kaminsky and Reinhart (1998)

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 153

Kaminsky, Lizondo, and Reinhart (1998) examined the be-

havior of 16 macroeconomic and financial indicators around

the time of a currency crisis for 76 currency-crisis episodes

in 15 developing and five industrial countries over the 1970-

95 period. They established a ranking for each indicator in

terms of the largest number of good signals issued, the

fewest number of false signals issued, and the average

lead time between the point in time when a signal was

first issued and the onset of a crisis.

How Well Do Leading Indicators Perform in Anticipating the Onset of

Currency Crises?

The real exchange rate was rated the best indicator, both

in terms of issuing the highest percentage of good signals

and the fewest false signals. Other indicators that per-

formed well in terms of issuing a large number of good

signals with few false signals included banking crises, the

ratio of M2 to reserves, the trend in equity prices, exports

and output. All of the indicators studied provided a warn-

ing window of 1- 1½ years, with banking crises providing

the longest lead time.

Leading Indicators of Currency Crises

Ranked by Highest Percentage of Good Signals Issued

Actual Good Signals/

Indicator Possible Good Signals

Real Exchange Rate 25

International Reserves 22

M2/Reserves 21

M2 Multiplier 20

Banking Crisis 19

Terms of Trade 19

Exports 1

Stock Prices 1

Output 16

Excess M1 Balances 16

Bank Deposits 16

Real Interest Rate 15

Domestic Credit/GDP 14

Lending Rate/Deposit Rate 13

Real Interest-Rate Differential 11

Imports 9

Source: Kaminsky, Lizondo, and Reinhart (1998).

Leading Indicators of Currency Crises

Ranked by Lowest Percentage of Bad Signals Issued

Actual Bad Signals/

Indicator Possible Bad Signals

Real Exchange Rate 5

Banking Crisis 6

Exports 7

Stock Prices 8

Output 8

Excess M1 Balances 8

Domestic Credit/GDP 9

M2/Reserves 10

Real Interest Rate 11

Real Interest-Rate Differential 11

Imports 11

International Reserves 12

M2 Multiplier 12

Terms of Trade 15

Bank Deposits 19

Lending Rate/Deposit Rate 22

Source: Kaminsky, Lizondo, and Reinhart (1998).

Leading Indicators of Currency Crises

Ranked by Longest Lead Time between

Date of Signal and the Onset of a Crisis

Indicator Months of Lead Time

Banking Crisis 19

Real Exchange Rate 17

Real Interest Rate 17

Imports 16

M2 Multiplier 16

Output 16

Bank Deposits 15

Excess M1 Balances 15

Exports 15

Terms of Trade 15

International Reserves 15

Stock Prices 14

Real Interest Rate 14

M2/Reserves 13

Lending Rate/Deposit Rate 13

Domestic Credit/GDP 12

Source: Kaminsky, Lizondo, and Reinhart (1998).

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

154 Deutsche Bank Foreign Exchange Research

The Deutsche Bank Alarm Clock (DBAC) is an economet-

ric model that estimates the probabilities of exchange-rate

and local interest-rate events. The model's recommenda-

tions are published every month in the

Deutsche Bank

Alarm Clock Monthly

DBAC simultaneously estimates exchange-rate and inter-

est-rate events over the course of the next 30 days, auto-

matically measuring the extent that interest-rate pressures

affect the exchange rate and vice versa. The model covers

the period starting in 1985, and includes macroeconomic

fundamentals and financial vulnerability indicators for 19

emerging markets in a pooled sample to produce the esti-

mates.

On the currency side, DBAC reports estimated probabili-

ties of an exchange-rate devaluation of 5% or more, 10%

or more, 15% or more, 20% or more, and 25% or more.

On the interest-rate side, DBAC reports the probability of

a 500 basis point increase in interest rates. These prob-

abilities are then compared to what is termed the “action

trigger”, which indicates when a given probability estimate

is signaling a potential crisis.

The action triggers were derived from the maximization of

in-sample return over cost of funds for a specific invest-

ment strategy. The choice of the action trigger is depen-

dent on one's risk preference or investment strategy, and

therefore different preferences or investment rules may

imply different action triggers. However, for two typical

strategies—long/short and long/out positions in all of the

countries included in the sample—the optimal action trig-

gers turn out to be about the same.

In the following charts, whenever the probability line

crosses above the trigger level, the model recommends

taking defensive action; when the probability line is below

the trigger level, the exchange rate is not expected to move

sharply enough to warrant defensive action.

Deutsche Bank Alarm Clock Model for Emerging-Market Currency and

Interest-Rate Risk

Brazilian Real Devaluation Probability

(Probability of 5% Exchange-Rate Devaluation)

Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02

Probability(%)

DBAC Trigger Level

Indonesian Rupiah Devaluation Probability

(Probability of 5% Exchange-Rate Devaluation)

Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02

Probability(%)

DBAC Trigger Level

Venezuelan Bolivar Devaluation Probability

(Probability of 15% Exchange-Rate Devaluation)

Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02

Probability(%)

DBAC Trigger Level

South African Rand Devaluation Probability

(Probability of 5% Exchange-Rate Devaluation)

Sep-00 Dec-00 Mar-01 Jun-01 Sep-01 Dec-01 Mar-02

Probability(%)

DBAC Trigger Level

Source: Deutsche Bank Alarm Clock

Source: Deutsche Bank Alarm ClockSource: Deutsche Bank Alarm Clock

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 155

References

Aziz, Jahangir, Francesco Caramazza and Ranil Salgado

(2000), "Currency Crises: In Search of Common Elements",

IMF World Economic Outlook Supporting Studies

, 2000,

pp. 86-128.

Bank For International Settlements (1999), "The Yen Carry

Trade and Recent Foreign Exchange Market Volatility",

BIS

Quarterly Review

, March 1999, pp. 33-37.

Berg, Andrew, Eduardo Borensztein, Gian Maria Milesi-

Ferretti and Catherine Pattilo (1999),

Anticipating Balance

of Payment Crises: The Role of Early Warning Systems

IMF Occasional Paper No. 186, Washington, D.C.

Bernard, Henri J. and Gabriele E.B. Galati (2000), "The Co-

Movement of US Stock Markets and the Dollar",

BIS Quar-

terly Review

, August 2000, pp. 31-34.

Bernhardsen, Tom and Oistein Roisland (2000), "Factors

That Influence the Krone Exchange Rate",

Norges Bank

Economic Bulletin

, Fourth Quarter 2000.

Bilson, John F.O. (1984), "Purchasing Power Parity as a Trad-

ing Strategy",

Journal of Finance

, Vol. 39, No.3, pp. 715-

25.

Brooks, Robin, Hali Edison, Manmohan S. Kumar and

Torsten Slok (2001), "Exchange Rates and Capital Flows",

IMF Working Paper

, WP/01/190.

Cheung, Yin-Wong, Menzie D. Chinn and Ian W. Marsh

(2000), "How Do UK-Based Foreign Exchange Dealers Think

Their Market Operates",

NBER Working Paper

, No.7524,

February 2000.

Chinn, Menzie D. and Ron Alquist (2000), "Tracking the

Euro's Progress",

International Finance

, Vol. 3, No. 3, No-

vember 2000, pp. 357-373.

Clostermann, Jorg and Bernd Schnatz (2000), "The Deter-

minants of the Euro-Dollar Exchange Rate: Synthetic Fun-

damentals and a Non-Existing Currency",

Deutsche

Bundesbank Discussion Paper

, 2/00.

Cooper, Neil and James Talbot (1999), "The Yen/Dollar Ex-

change Rate in 1998: Views From Options Markets",

Bank

of England Quarterly Bulletin

, February 1999, pp. 68-77.

Copeland, Laurence S. (2000),

Exchange Rates and Inter-

national Finance

, Third Edition, Pearson Education Limited,

Essex.

DeGrauwe, Paul (1996),

International Money: Postwar

Trends and Theories

, 2nd Edition, Oxford University Press,

Oxford.

Dominguez, Kathryn M. and Jeffrey Frankel (1993),

Does

Foreign Exchange Intervention Work?

, Institute For Inter-

national Economics, Washington, D.C.

Dornbusch, Rudiger (1976), "Expectations and Exchange

Rates Dynamics",

Journal of Political Economy

, Vol. 84, pp.

1161-76.

Dunis, Christian and Pierre Lequeux (2001), "The Informa-

tion Content of Risk Reversals",

Derivatives Use, Trading

and Regulation

, Vol. 7, Number 2, pp.98-117.

European Central Bank (2002), "Economic Fundamentals

and the Exchange Rate of the Euro",

ECB Monthly Bulle-

tin

, January 2002.

Fender, Ingo and Gabriele Galati (2001), "The Impact of

Transatlantic M&A Activity on the Dollar/Euro Exchange

Rate",

BIS Quarterly Review

, December 2001, pp. 58-68.

Fleming, J. Marcus (1962), "Domestic Financial Policies

Under Fixed and Floating Exchange Rates",

IMF Staff Pa-

pers

, Vol.9, November 1962, pp. 369-79.

Frankel, Jeffrey (1979), "On the Mark: A Theory of Floating

Exchange Rates Based on Real Interest Rate Differentials",

American Economic Review

, Vol. 69, September 1979, pp.

610-22.

Frankel, Jeffrey and Kenneth Froot (1990), "Chartists, Fun-

damentalists and the Demand For Dollars", in

Policy Is-

sues in Interdependent Economics

, edited by T. Courakis

and M. Taylor, Clarendon Press, Oxford.

Frankel, Jeffrey A. and Andrew K. Rose (1996), "Currency

Crashes in Emerging Markets: Empirical Indicators",

NBER

Working Paper

, No. 5437.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

156 Deutsche Bank Foreign Exchange Research

Froot, Kenneth A. and Richard H. Thaler (1990), "Anoma-

lies: Foreign Exchange",

Journal of Economic Perspectives

Vol.4, No.3, Summer 1990, pp. 179-192.

Gehrig, Thomas and Lukas Menkhoff (2002), "The Use of

Flow Analysis in Foreign Exchange,"

CEPR Discussion Pa-

per Series

, No. 3221.

Goldfajn, Ilan and Rodrigo O. Valdes (1996), "The Aftermath

of Appreciations",

NBER Working Paper

, No. 5650.

Goldstein, Morris, Graciela L. Kaminisky and Carmen M.

Reinhart (2000),

Assessing Financial Vulnerability: An Early

Warning System For Emerging Markets

, Institute For In-

ternational Economics, Washington, D.C.

Goldstein, Morris, and John Hawkins (1998), "The Origin

of the Asian Financial Turmoil",

Reserve Bank of Australia,

Research Discussion Paper

, No. 9805.

Gruen, David (2001), "Some Possible Long-Term Trends in

the Australian Dollar",

Reserve Bank of Australia Bulletin

December 2001, pp.30-41.

Gust, Christopher and Jaime Marquez (2000), "Productiv-

ity Developments Abroad",

Federal Reserve Bulletin

, Oc-

tober 2000, pp. 665-681.

Hallwood, C. Paul and Ronald MacDonald (2000),

Interna-

tional Money and Finance

, 3rd Edition, Blackwell Publish-

ers, Malden, Massachusetts.

IMF,

World Economic Outlook Reports

, May 1998, Decem-

ber 2001.

Isard, Peter and Hamid Faruqee, editors (1998), “Exchange

Rate Assessment: Extensions of the Macroeconomic Bal-

ance Approach”,

IMF Occasional Paper

, No. 167.

Isard, Peter, Hamid Faruqee, G. Russell Kincaid and Mar-

tin Fetherston (2001), “Methodology For Current Account

and Exchange Rate Assessments”, I

MF Occasional Paper

No. 209.

Kaminsky, Graciela L. and Carmen M. Reinhart (1997), "The

Twin Crises: The Causes of Banking and Balance of Pay-

ments Problems",

University of Maryland Working Paper

No. 37.

Kaminsky, Graciela L. and Carmen M. Reinhart (1998), "Fi-

nancial Crises in Asia and Latin America: Then and Now",

mimeo.

Kaminsky, Graciela L., Saul Lizondo and Carmen M.

Reinhart (1998), "Leading Indicators of Currency Crises",

IMF Staff Papers

, Vol. 45, No. 1, March 1998, pp. 1-48.

Koen, Vincent, Laurence Boone, Alain de Serres and Nicola

Fuchs (2001), "Tracking the Euro",

OECD Economic Depart-

ment Working Paper

, No. 298.

Kritzman, Mark P. (1998), "The Forward Rate Bias",

AIMR

Conference Proceedings, Currency Risk in Investment

Portfolios

, Charlottesville, VA.

Krugman, Paul (1998), "It's Baaack: Japan's Slump and the

Return of the Liquidity Trap",

Brookings Papers on Economic

Activity

, 1998, Vol. 2.

Laidler, David and Shay Aba (2002), "Productivity and the

Dollar: Commodities and the Exchange Rate Connection",

C.D.

Howe Institute Commentary

, No. 158, February 2002.

Lyons, Richard K. (2000), "Multiple Currency Investment

Strategies To Take Advantage of the Forward Discount Bias",

class project for International Finance 285 course, (http://

haas.berkeley.edu/~Lyons/285.html).

Lyons, Richard K. (2001),

The Microstructure Approach to

Exchange Rates

, MIT Press, Cambridge, Mass.

MacDonald, Ronald and J. Nagayasu (2000), "The Long-

Run Relationship Between Real Exchange Rates and Real

Interest Rate Differentials: A Panel Study",

IMF Staff Pa-

pers

, Vol. 47, No. 1, pp. 116-128.

MacDonald, Ronald and Jerome L. Stein (1999),

Equilib-

rium Exchange Rates

, Kluwer Academic Publishers,

Dordrecht, Netherlands.

MacDonald, Ronald and Philip Swagel (2000), "Business

Cycle Influences on Exchange Rates: Survey and Evidence",

IMF World Economic Outlook Supporting Studies

, 2000,

pp.129-160.

Mann, Catherine L. (1999),

Is the US Trade Deficit Sustain-

able?

, Institute For International Economics, Washington,

D.C.

DB Guide to Exchange-Rate Determination

May 2002 Deutsche Bank @

Deutsche Bank Foreign Exchange Research 157

Mark, Nelson C. (2001),

International Macroeconomics and

Finance: Theory and Econometric Methods

, Blackwell Pub-

lishers, Malden, Mass.

McGuckin, Robert H. and Bart Van Ark (2001),

Making the

Most of the Information Age: Productivity and Structural

Reform in the New Economy

, The Conference Board, Inc.,

New York.

McKinsey Global Institute (2001),

US Productivity Growth

1995-2000, McKinsey and Co., Washington D.C.

McKinnon, Ronald I. (1979),

Money in International Ex-

change: The Convertible Currency System

, Oxford Univer-

sity Press, New York.

McKinnon, Ronald I. And Kenichi Ohno (1997),

Dollar and

Yen: Resolving Economic Conflict Between The United

States and Japan

, MIT Press, Cambridge, Mass.

Meredith, Guy (2001), "Why Has the Euro Been So Weak?"

IMF Working Paper

, WP/01/155.

Meese, Richard and Kenneth Roguff (1983), "Empirical

Exchange Rate Models of the Seventies: Do They Fit Out

of Sample",

Journal of International Economics

, Vol. 14,

No. 1, pp. 3-24.

Mundell, Robert A. (1963), "Capital Mobility and Stabiliza-

tion Policy Under Fixed and Flexible Exchange Rates",

Ca-

nadian Journal of Economics and Political Science

, Vol. 29,

November 1963, pp. 475-85.

Neely, Christopher J. (1997), "Technical Analysis in the For-

eign Exchange Market: A Layman's Guide",

Federal Re-

serve Bank of St. Louis Review

, September/October 1997,

pp. 23-38.

Obstfeld, Maurice and Kenneth Rogoff (1994), "The

Intertemporal Approach to the Current Account",

NBER

Working Paper

, No. 4893.

Osler, Carol and P.N. Kevin Chang (1995), "Head and Shoul-

ders: Not Just A Flaky Pattern",

Federal Reserve Bank of

New York Staff Reports

, No. 4.

Pilbeam, Keith (1998),

International Finance

, 2nd Edition,

Macmillan Press LTD, London.

Posen, Adam S. (1998),

Restoring Japan's Economic

Growth

, Institute For International Economics, Washing-

ton, D.C.

Rogoff, Kenneth (1996), "The Purchasing Power Parity

Puzzle", J

ournal of Economic Literature

, Vol. XXXIV, June

1996, pp. 647-668.

Rosenberg, Michael R. (1998),

Currency Crises in Emerg-

ing Markets: A Guide To Speculative Attack Models and

Early Warning Systems

, Merrill Lynch Research Report.

Rosenberg, Michael R. (1996),

Currency Forecasting: A

Guide To Fundamental and Technical Models of Exchange

Rate Determination

, Richard D. Irwin Co., Chicago, IL.

Sachs, Jeffrey, Aaron Tornell and Andres Velasco (1996),

"Financial Crises in Emerging-markets: The Lessons From

1995",

Brookings Papers on Economic Activity

, 1996, No.

1, pp. 147-215.

Sarno, Lucio and Mark P. Taylor (2001), "Official Interven-

tion in the Foreign Exchange Market: Is It Effective and, If

So, How Does It Work?",

Journal of Economic Literature

Vol. XXXIV, No. 3, September 2001, pp. 839-868.

Tille, Cedric, et al. (2001), "To What Extent Does Productiv-

ity Drive the Dollar",

Federal Reserve Bank of New York

Current Issues in Economics and Finance

, Vol. 7, No. 8,

August 2001.

Williamson, John, editor (1994),

Estimating Equilibrium

Exchange Rates

, Institute For International Economics,

Washington, D.C.

DB Guide to Exchange-Rate Determination May 2002

Deutsche Bank @

158 Deutsche Bank Foreign Exchange Research

FX Weekly 7 December 2001

Deutsche Bank @

14 Deutsche Bank Foreign Exchange Research

Global Markets Research

David Folkerts-Landau

Managing Director, Global Head of Research Tel: +44 20 754 55502

Stuart Parkinson, Chief Operating Officer Tel: +44 20 754 57303

Peter Garber, Global Strategist Tel: +1 212 469 5466

Credit Derivatives Research and Strategy

John F. Tierney, Global Head Tel: +1 212 469 6795

Credit Research

Simon Adamson, Co-head, European High Grade Credit Research Tel: +44 20 754 58715

David Bitterman, Co-head, US High Yield Credit Research Tel: +1 212 469 2599

Nuj Chiaranussati, Head of Asian Credit Research Tel: +65 6423 5930

Anja King, Co-head, European High Grade Credit Research Tel: +44 20 754 59260

Helen Rodriguez, Head, European High Yield Research Tel: +44 20 754 53824

Andrew W. Van Houten, Co-head, US High Yield Credit Research Tel: +1 212 469 2777

Economic Research

Ciaran Barr, Chief Economist UK Tel: +44 20 754 52088

Ulrich Beckmann, Co-head, European Economics Tel: +49 69 910 31729

Marco Annunziata, Chief Economist, Emerging Europe Tel: +44 20 754 55506

Ivan Colhoun, Chief Economist, Australia Tel: +61 2 9258 1667

Peter Hooper, Chief US Economist Tel: +1 212 469 7352

Atsushi Mizuno, Chief Economist, Japan Tel: +81 3 5156 6316

Carlo Monticelli, Co-head, European Economics Tel: +44 20 754 72884

Michael Spencer, Chief Economist, Asia Tel: +852 2203 8305

Emerging Markets Research

Jose Luis Daza, Head of Emerging Markets Strategy Tel: +1 212 469 8985

Leonardo Leiderman, Head of Emerging Markets Economics Tel: +1 212 469 5894

Fixed Income and Relative Value Research

Jamil Baz, Global Head Tel +44 20 754 54017

Foreign Exchange Research

Michael Rosenberg, Global Head Tel +1 212 469 4776

Indices

Fergus Lynch, Head of Index Development Tel: +44 20 754 58765

Quantitative Flow Research

Robin Lumsdaine, Head/Global Econometric Strategist Tel: +44 20 754 54858

Securitization Research

Karen Weaver, Global Head Tel: +1 212 469 3125

Main Offices

London

Winchester House

1 Great Winchester Street

London EC2N 2DB

United Kingdom

Tel: +44 20 7545 8000

Frankfurt

Grosse Gallusstrasse 10-14

60311 Frankfurt

Germany

Tel: +49 69 9100-0

New York

31 West 52nd Street

New York, NY 10019

United States of America

Tel: +1 212 469 5000

Hong Kong

55/F, Cheung Kong Center

2 Queen’s Road, Central

Hong Kong

Tel: +852 2203 8888

Sydney

Grosvenor Place

Level 18,225 George Street

Sydney NSW 2000

Australia

Tel: +61 2 9258 1661

Singapore

5 Temasek Boulevard

#08-01 Suntec Tower Five

Singapore 038985

Tel: +65 6224 4677

Tokyo

Sanno Park Tower

2-11-1, Nagatacho,

Chiyoda-ku, Tokyo 100-6171

Japan

Tel: +81 3 5156 6000

"210 %"

+(



Publication Address:

Deutsche Bank AG London

Winchester House

1 Great Winchester Street

London EC2N 2DB

Internet:

http://research.gm.db.com

Ask your usual contact for a

username and password.

The information and opinions in this report were prepared by Deutsche Bank AG or one of its affiliates (collectively “Deutsche

Bank”). This report is based upon information available to the public. The information herein is believed by Deutsche Bank to be

reliable and has been obtained from sources believed to be reliable, but Deutsche Bank makes no representation as to the

accuracy or completeness of such information. Deutsche Bank may be market makers or specialists in, act as advisers or lenders

to, have positions in and effect transactions in securities of companies mentioned herein and also may provide, may have

provided, or may seek to provide investment banking services for those companies. In addition, Deutsche Bank or its respective

officers, directors and employees hold or may hold long or short positions in the securities, options thereon or other related

financial products of companies discussed herein. Opinions, estimates and projections in this report constitute the current

judgement of the author as of the date of this report. They do not necessarily reflect the opinions of Deutsche Bank and are

subject to change without notice. Deutsche Bank has no obligation to update, modify or amend this report or to otherwise notify

a reader thereof in the event that any matter stated herein, or any opinion, projection, forecast or estimate set forth herein,

changes or subsequently becomes inaccurate or misleading. Prices and availability of financial instruments also are subject to

change without notice. This report is provided for informational purposes only. It is not to be construed as an offer to buy or sell

or a solicitation of an offer to buy or sell any financial instruments or to participate in any particular trading strategy in any

jurisdiction in which such an offer or solicitation would violate applicable laws or regulations. The financial instruments discussed

in this report may not be suitable for all investors and investors must make their own investment decisions using their own

independent advisors as they believe necessary and based upon their specific financial situations and investment objectives. If a

financial instrument is denominated in a currency other than an investor’s currency, a change in exchange rates may adversely

affect the price or value of, or the income derived from, the financial instrument, and such investor effectively assumes currency

risk. In addition, income from an investment may fluctuate and the price or value of financial instruments described in this report,

either directly or indirectly, may rise or fall. Furthermore, past performance is not necessarily indicative of future results, and

should not therefore form the basis of a decision whether or not to invest in any securities issued by any company mentioned

herein. Unless governing law permits otherwise, all transactions should be executed through the Deutsche Bank entity in the

investor’s home jurisdiction. In the U.S. this report is approved and/or distributed by Deutsche Bank Securities Inc., a member of

the NYSE, the NASD and SIPC. In the United Kingdom this report is approved and/or communicated by Deutsche Bank AG

London. In jurisdictions other than the U.S. and the U.K. this report is distributed by the Deutsche Bank affiliate in the investor’s

jurisdiction, and interested parties are advised to contact the Deutsche Bank office with which they currently deal. Additional

information relative to securities, other financial products or issuers discussed in this report is available upon request. No part of

this material may be copied or duplicated in any form or by any means, or redistributed, without Deutsche Bank’s prior written

Subscribers to research via

email receive their

electronic publication on

average 1-2 working days

earlier than the printed

version.

If you would like to receive

this or any other product

via email please contact

your usual Deutsche Bank

representative.

Global Markets Global Markets Global Markets

The Deutsche Bank Guide To Exchange Rate Determination FX Broker DB

Navigation menu

Versions of this User Manual:

Views

Navigation