Monetary policy goals, institutions, strategies, and instruments

The debate between the Banking School and the Currency School The institutional framework of monetary policy II: the design of the central bank legislation 7.1... The monetary policy fra

GOALS, INSTITUTIONS, STRATEGIES,

AND INSTRUMENTS

Peter Bofinger

in collaboration with

Julian Reischle and Andrea Schachter

OXFORDUNIVERSITY PRESS

Monetary Policy

in order to ensure its continuing availability

OXJORD

UNIVERSITY PRESS

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford.

It furthers the University's objective of excellence in research, scholarship,

and education by publishing worldwide in

Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi

Kuala Lumpur Madrid Melbourne Mexico City Nairobi

New Delhi Shanghai Taipei Toronto

With offices in Argentina Austria Brazil Chile Czech Republic France Greece

Guatemala Hungary Italy Japan South Korea Poland Portugal

Singapore Switzerland Thailand Turkey Ukraine Vietnam

Oxford is a registered trade mark of Oxford University Press

in the UK and in certain other countries Published in the United States

by Oxford University Press Inc., New York

© Peter Bofinger 2001 The moral rights of the author have been asserted

Database right Oxford University Press (maker)

Reprinted 2006 All rights reserved No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press,

or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department,

Oxford University Press, at the address above

You must not circulate this book in any other binding or cover And you must impose this same condition on any acquirer

ISBN 0-19-924057-4

When I had to hold my first classes on monetary policy many years ago, Iwas amazed by the fact that there are almost no textbooks or comprehen-sive treatises on monetary policy As a result I wrote such a book, togetherwith Julian Reischle and Andrea Schachter, which appeared in German in

1996 The positive assessment of this book by the academic community(Fase 1998) has encouraged me to consider an English version The last fouryears having been an extremely active and exciting time for monetary the-ory and policy, the outcome now differs in many respects from the originalGerman text Nevertheless, it is still shaped by the same guiding principles.While the focus is on policy-relevant issues, the book always tries to pro-vide the necessary theoretical background In contrast to textbooks on'Money, Banking, and Financial Markets', which normally describe andanalyse only the domestic monetary policy, this book adopts a comparativeapproach which always includes the practices in the most important cur-rency areas of the world With the high degree of international financialmarket integration, a student of monetary policy has to know more thanwhat is going on in his or her own country

As this book is the outcome of a long process of work, it is difficult to thankall those who have made it possible There is no doubt that I have greatlybenefited from working as an economist at the Deutsche Bundesbank in theyears 1985-90 In the last nine years, my students at Wiirzburg Universityinspired me with their critical questions on the available paradigms of mone-tary economics and gave me important feedback on draft material It wasalways very important to be able to discuss my ideas with my colleagues andassistants I am especially grateful to Christoph Harff, Oliver Hulsewig,Kai Pfleger, Julian Reischle, Andrea Schachter, Nicolas Schlotthauer, andTimo Wollmershauser Annemarie Reussner, Andre Geis, Heiko Miiller, andChristiane Klemp provided valuable technical assistance I would also like tothank Hans-Joachim Jarchow, Dieter Nautz, Lars Svensson, and the mem-bers of the Ausschuss fur Geldtheorie and Geldpolitik of the Verein furSocialpolitik for insightful and useful comments on different parts of thisbook A very special note of thanks is due to many central banks, above all theDeutsche Bundesbank, the European Central Bank, and the Bank of Japan,which supported this book by providing data and other important information

Wiirzburg P.B.

are available on the internet at www.monetary-policy.net.

All comments on the book should be mailed to Peter Bofinger at

bofinger@t-online.de.

Part I Theoretical Fundaments of Monetary Policy

1 What is money?

1.1 Introduction

1.2 A microeconomic approach to the definition of money

1.3 Defining money by its functions

1.4 Defining money by its statistical properties

1.5 The importance of a 'correct' definition of money

Appendix 1.1 Definitions of money in the euro area,

the United States, Japan, and the UK

Appendix 1.2 The concept of (net) financial assets

(flow of funds analysis)

2 The demand for money

2.1 Introduction

2.2 The quantity theory of money and the Cambridge approach

2.3 The interest rate as a determinant of the money demand

2.4 The money demand Junction ofCagan

2.5 Wealth as a determinant of the demand for money

2.6 Empirical demand Junctions for money

Appendix 2.1 Cointegration

3 The money supply process: starting point of

the transmission process

3.1 Introduction

3.2 Creation and control of the monetary base by the central bank

3.3 Consolidated balance sheet of the banking system and

the supply of the money stock

3

3 4 11 14 15 15 17

20 20 21 23 32 32 34 38

40 40 41 46

3.4 The mechanistic multiplier process

3.5 A price-theoretic model of the money supply

Appendix 3.1 A 'Poole analysis' for shocks

Monetary policy transmission

4.1 Introduction

4.2 Limited knowledge about the transmission process

4.3 The quantity theory channel

4.4 Interest rate channels

4.5 Expectations channels (the 'Phillips curve')

4.6 Summary

Appendix 4 1 The original Poole model

Appendix 4.2 The concept of the 'price gap'

48 53 65 71 71 73 74 80 95 115 116 121

5.

6.

7.

Part II Domestic Aspects of Monetary Policy

The ultimate goal and the final targets of monetary policy

5.1 Introduction

5.2 The long-term view: only price stability matters

5.3 Inflation and output growth in the short run

5.4 Operational issues of a 'stability-oriented monetary policy'

5.5 Summary

Appendix 5.1 Different variants of the nominal GDP targeting

Appendix 5.2 Calculating the welfare losses of

sub-optimal money holdings

The institutional framework for monetary policy I:

'rules versus discretion'

6.1 Overview

6.2 Rules versus discretion in monetary policy

6.3 The main arguments for limiting the discretion of central bankers

6.4 Rules that are imposed on central banks from outside

6.5 Summary of the traditional debate

6.6 Time inconsistency': a new argument in support of rules?

Appendix 6 1 The debate between the Banking School and

the Currency School

The institutional framework of monetary policy II:

the design of the central bank legislation

7.1 Overview

7.2 Independence of monetary policy

127 127 131 148 153 160 161 162

164 164 165 165 169 174 174 202

205 205 207

9

10

11

7.3 Accountability of central banks

7.4 Liability of incompetent central bankers

Appendix 7.1 The functions of central banks and the technology

of the payments systems

Strategies ('simple rules') for a stability-oriented

monetary policy

8.1 The function of simple rules

8.2 "Simple rules', intermediate targets, and indicators of

9.1 A comparison of the performance of central banks

9.2 The monetary policy framework of the Deutsche Bundesbank

9.3 The monetary policy framework of the European Central Bank

9.4 The monetary policy framework of the Federal Reserve System

9.5 The monetary policy framework of the Bank of Japan

9.6 The monetary policy framework of the Bank of England

The instruments of monetary policy

10.1 Basic requirements

10.2 Monetary base targeting v interest rate targeting

10.3 Significance of individual monetary policy instruments

10.4 Money market management by the European Central Bank

10.5 Money market management by the Federal Reserve System

10.6 Money market management by the Bank of England

10.7 Money market management by the Bank of Japan

10.8 The New Political Economy and monetary

policy instruments

Seigniorage and inflation tax

11.1 Introduction

11.2 Monetary seigniorage and fiscal seigniorage

1 1 3 Macroeconomic stability and seigniorage financing

1 1 4 Optimal seigniorage

220 228 234

240 240 245 248 257 268 274 282

293 293 295 300 307 311 317 321 321 323 336 348 358 361 364 365 369 369 370 372 382

Part III Monetary Policy in an Open Economy

12 Important building blocks of open-economy macroeconomics

12.1 Introduction

1 2.2 The exchange rate as an operating target of monetary policy

1 2.3 The exchange rate as an intermediate target of

monetary policy

1 2.4 The control of the price level in an open economy

13 Monetary policy strategies in an open economy

429 447 450

Divisia and simple-sum Ml in the United States, 1960-2000

Divisia M3 and simple-sum M3 in the United States, 1960-2000

Velocity for Germany, 1970-1998

Velocity for the United States, 1975-1998

The speculative demand for money

The Baumol-Tobin model

Ml and M3 as percentages of total financial assets (TFA) of

private households in Germany, 1955-1998

The 'macroeconomic' money (= credit) market

The market for monetary base

The money supply process

Money market rates and short-term bank rates in Germany,

1970-1998 (monthly average)

Money demand shocks

Credit (= money) supply shocks

Multiplier shocks

Aggregate demand and supply under the gold standard

Monetary growth and inflation in industrial countries, 1960-1999

Real interest rates and real GDP growth in Germany, 1963-1999

Demand shocks in the IS/LM model

The LM-curve in the price-theoretic money supply model

Restrictive monetary policy in the 'banking view'

Money market rates and net interest payments in Germany,

1971-1998

Profits and net interest payments of German enterprises,

1971-1998

Wages and inflation in the euro area, 1983-2001

A modified Phillips curve for the United States

Long-term and short-term Phillips curves

The Phillips curve for Germany, 1965-1999

The Phillips curve for the United States, 1965-1999

The Phillips curve for the euro area, 1983-2000

IS shocks

LM shocks

Supply shocks

Price gaps

Welfare costs of inflation

Real and net real interest rates in Italy, 1982-1999

Inflation and real GDP, 1972-1982

Demand shocks

8 10 22 23 26 29 33 58 59 61 64 66 67 69 76 77 81 85 87 89 92 92 96 99 101 112 113 113 118 119 120 123 133 140 145 149

5.5 Supply shocks: price-level targeting and nominal GNP targeting

6.1 The social loss function of the Barro-Gordon model

6.2 Different solutions to the Barro-Gordon model

7.1 Inflation and central bank independence, 1955-1988

7.2 Real growth and inflation, various countries, 1955-1988

7.3 Ex post accountability

8.1 Four visions of rationality

8.2 Monetary targeting and demand shocks

8.3 Monetary targeting and supply shocks

8.4 The Bank of England's 'fan-chart' for inflation (RPIX)

8.5 The neutral real interest rate

8.6 Monetary base growth and inflation targets in the United States

A8.1 Hypothetical yield curves

A8.2 Empirical yield curves

A8.3 Real GDP growth rates and the yield structure, United States,

1958-1998

A8.4 Bond yields and inflation rates, United States, 1958-1998

A8.5 The reciprocal yield structure and real short-term rates,

United States, 1958-1998

9.1 The Taylor rule for Germany, 1970-1999

9.2 Inflation and the growth rate of M3 in the euro area:

historical data, 1982-1999

9.3 The Taylor rule for the euro area, January 1999-July 2000

9.4 The Taylor rule for the United States, 1970-1999

9.5 The Taylor rule for Japan, 1970-1999

9.6 Real effective exchange rates, USA, Japan, Germany, and

United Kingdom, 1980-1999 by month

9.7 Real exchange rates and export performance in Japan, 1980-1998

9.8 The Taylor rule for the United Kingdom, 1970-1999

10.1 Interest rate and monetary base targeting

11.1 Maximum seigniorage with linear money demand

11.2 Maximum seigniorage with semi-logarithmic money demand

11.3 Surprise inflation and seigniorage

12.1 Euro-US$ spot rates and three-month forward rates,

13.2 Yen-US$ exchange rates and monthly interventions by

the Bank of Japan, 1975-1999

13.3 Real interest rates and real exchange rates in France, 1979-1998

13.4 The Deutschmark-French franc exchange rate in the EMS and

French foreign exchange reserves, 1979-1986

13.5 Tolar-Deutschmark rate and Slovenian reserves, 1992-2000

150 178 184 217 218 222 244 254 255 261 270 281 282 283 284 287 291 299 303 306 311 312 314 315 318 325 375 378 381 400 402 412 413 424 425 426

Al.l Definitions of money

A1.2 Financial assets and liabilities in Germany, 1998

3.1 Consolidated financial statement of the European System of

Central Banks, 2 June 2000

3.2 Simplified balance sheet of the Eurosystem, 2 June 2000

3.3 Consolidated financial balance sheet of the euro area MFIs,

including the Eurosystem, end-May 2000

3.4 A very simplified central bank balance sheet

3.5 A very simplified consolidated balance sheet of

the banking system

3.6 The mechanistic multiplier model

A3 1 Results of the shocks in the money supply process

4.1 A numerical example

4.2 Autoregression of the inflation rate, 1976-1998

5.1 Numerical inflation targets of central banks

5.2 Output gaps of more than 1% in G7 countries, 1982-2000

5.3 Net real interest rates with different inflation rates

5.4 Real value of a nominal payment of €1,000 which is due in

10 years

6 1 Existing currency boards

7.1 Governing bodies of central banks

7.2 Indices of central bank independence

7.3 Inflation and central bank independence, 1996-2000

7.4 Instruments of monetary policy transparency

8.1 The 'potential formula' and the Bundesbank's monetary targets

8.2 Alternative definitions of inflation targeting

8.3 Historical short-term and long-term real interest rates

8.4 The difference between estimates of the output gap

(IMF minus OECD)

9.1 Inflation performance (average annual CPI inflation)

9.2 Annual social welfare loss (according to equation 5.1) (%)

9.3 The Bundesbank's performance with monetary targeting

9.4 A monetary policy reaction function for the United States

10.1 The market for money in the IS/LM model and the money market

10.2 Minimalist approaches to monetary base targeting and

interest rate targeting

10.3 Interest rate targeting using various instruments

10.4 Key elements of reserve requirements

10.5 Suitable instruments for monetary base targeting and

interest rate targeting

16 18 41 43 46 48 49 53 70 105 109 130 131 139 144 173 216 219 220 226 251 259 271 273 293 294 296 309 324 335 338 343 348

10.6 Variability of money market rates

10.7 Monetary policy instruments of the four main central banks

12.1 Simplified central bank balance sheet

1 2.2 Simplified central bank balance sheet after massive sterilized

intervention

12.3 OLS estimates for uncovered interest parity (equations (12.14)

and (12.15)), February 1986-December 1998

1 2.4 Uncovered interest parity under flexible and fixed exchange rates

12.5 Numerical example of the Ricardo-Balassa effect

13.1 Critical inflation differentials for a fixed nominal peg:

the ERM I experience

13.2 The dual disequilibrium ofThailand, 1990-1998

1 3.3 Deutschmark interventions in the Exchange Rate Mechanism of

the EMS, 1979-1994

349 367 389 391 396 399 408 420 421 427

The money multiplier of the euro area

Lags in monetary policy

Bank of England Bank of Japan Covered interest rate parity Commonwealth of Independent States Consumer price index

All groups consumer price index excluding credit services European Community

European Central Bank Error correction mechanism European Currency Unit European Monetary System European Monetary Union Euro overnight index average of interbank overnight rates Excess Reserves

Exchange Rate Mechanism European System of Central Banks European Union

U.S Federal Reserve System Federal Open Market Committee Federal Reserve Bank

Group of the seven leading industrial countries (United States, Japan, Germany, France, Canada, United Kingdom, Italy) Gross domestic product

Gross national product Harmonised index of consumer prices International Monetary Fund Curve describing equilibrium of investment and saving Japanese government bonds

Monetary conditions index Monetary financial institutions Monetary Policy Committee Non-inflation accelerating rate of unemployment Curve describing equilibrium on the money market Money stock aggregates

Minimum reserve Net domestic assets

Survey of Professional Forecasters Sveriges Riksbank (Swedish Central Bank) Uncovered interest rate parity

to be cultivated by the latter is trust and confidence in the initiated group's prehension of the esoteric knowledge (Brunner 1981: 5) 1

com-Two decades later, after Alan Greenspan's impressive performance aschairman of Federal Reserve System since August 1987, most observerswould totally agree with Brunner's dictum Nevertheless, this view is notuniversally accepted Alan Blinder, after serving as a vice-chairman of theFederal Reserve Board and as a member of the Council of EconomicAdvisers, came to the following assessment:

Having looked at monetary policy from both sides now, I can testify that central banking in practice is as much art as science Nonetheless, while practising this dark art, I have always found the science quite useful (Blinder 1997: 17)

I hope this book will show that there is indeed a solid core of a science ofmonetary policy which is able to provide a set of techniques that are indis-pensable for successful central banking.2 Of course, mystique plays a roletoo, but I will leave this sphere of monetary policy to other writers.The book is structured in three main parts Parts I and II discuss the the-ory and the implementation of monetary policy from the perspective of

a relatively large central bank such as the European Central Bank, theFederal Reserve, and the Bank of Japan Thus, the focus is on the domesticissues of monetary policy In Part III an open-economy perspective isadopted This procedure gives the reader a comprehensive overview of the

1 A similar statement can be found at the end of a treatise by the Swiss economist Jurg Niehans (1988: 336): 'However, economics must not fall victim to the illusion that central banking will ever become a science— Whatever progress is made in monetary theory, central banking will probably remain an art.'

2 See Alfred North Whitehead (1978: 338): "The condition for excellence is a thorough training of technique.'

aspects that are relevant from a closed-economy perspective before she isconfronted with the complexities that arise in an open economy.

Part I presents the theoretical building blocks that are needed for an

understanding of monetary policy, but it does not serve as a completesurvey on all issues of present-day monetary theory Chapter 1 starts withalternative definitions of money and the microeconomic functions ofmoney Theoretical approaches to the demand for money and their empir-ical evidence are discussed in Chapter 2 Chapter 3 presents traditionaltheories to the money supply process and a price-theoretic model that isrequired for an analysis of the operating procedures of all major centralbanks From this starting point, Chapter 4 deals with the most importantchannels for the transmission of monetary policy impulses: the quantitytheory, the interest rate channel, and the expectations channel

Part //discusses monetary policy from a domestic perspective Chapter 5

starts with the ultimate goals of central banking It shows above all that anobligation to pursue price stability as the main target of monetary policy issufficiently flexible to cope with demand and supply shocks in the short andmedium term Chapter 6 deals with the debate on 'rules versus discretion' inmonetary policy It comes to the conclusion that the arguments of the tradi-tional debate are not strong enough to justify the imposition of rigid rules.The same applies to the more recent debate which is based on the timeinconsistency of a central banks' plans Chapter 7 is focused on the triad

of independence, accountability, and sanctions While most central banksnow enjoy a high degree of independence, a systematic process of account-ability is still lacking and direct sanctions are almost always absent

In Chapter 8 and the subsequent chapters I switch from the institutionalframework (the hardware of central banking) to strategic aspects (the soft-ware of central banking) First, I discuss some of the 'simple rules' that havebeen developed in the last decades and ask in what way they could be used

as a 'heuristic', facilitating the difficult decision processes of central bankersand their dialogue with the public Chapter 9 assesses the performance ofthe Bundesbank, the European Central Bank, the Federal Reserve System, theBank of England, and the Bank of Japan, above all the aspect of whether thepolicy decisions have been guided by some specific rule Chapter 10 dealswith the instruments of monetary policy After a general discussion whichleads to a minimalist approach, Chapter 11 provides a short discussion ofdifferent concepts of 'seigniorage' and the role of seigniorage financing inprocesses of hyperinflation

Part HI is devoted to the international context in which all central banks

are operating In Chapter 12 the most important building blocks of economy macroeconomics are presented: the mechanics of foreign exchange

open-market intervention, the covered and uncovered interest rate parity theory,and the purchasing power parity theory Chapter 13 discusses the policyoptions for central banks in both large and small currency areas For centralbanks in smaller countries, it is important to observe that the combination

of interest rate and exchange rate target has to be compatible with

uncov-ered interest parity and with domestic macroeconomic conditions I show

under which conditions a fixed nominal exchange rate target can serve as arule for monetary policy In addition, more flexible solutions (crawlingpegs) are discussed

For a very busy reader, the main lessons of this book can be summarized

in ten 'dos and don'ts':

1 Create the central bank constitution so that decision-makers have along-term time horizon This requires a high degree of monetary pol-icy independence from the political sphere

2 Define a price stability target over the medium term which is flexibleenough to accommodate supply shocks and leaves it up to the discre-tion of the central bank whether it is suitable to react to short-termdemand shocks

3 Use a short-term interest rate as the operating target of monetary policy

4 In order to keep the management of short-term rates simple, use ing facilities for the provision of liquidity

stand-5 Use all information on private inflation expectations as an importantindicator of future price developments

6 Rely on the self-stabilizing mechanism of private inflation tations if the inflation rate is low Follow a policy of 'interest ratesmoothing'; or, in a free variation of Milton Friedman's (1968: 12) les-son no.l, 'Prevent interest rate changes from being a major source ofeconomic disturbance'.3

expec-7 Avoid negative real short-term interest rates under all circumstances

8 If the knowledge about the outgap is good, follow an interest rate icy that is determined by the Taylor rule; otherwise, a Taylor rule on thebasis of the inflation rate provides good guidance only in situationsinvolving demand shocks

pol-9 Avoid exchange rate targets that are incompatible with uncoveredinterest parity

10 Avoid a major real appreciation by a policy of sterilized interventionsthat is compatible with the tenet no 9

3 Friedman (1968: 12) said: "The first and most important lesson that history teaches about what monetary policy can do-and it is a lesson of the most profound importance-is that monetary policy can prevent money itself from being a major source of economic disturbance.'

We will see that in the last decades a negative performance of monetarypolicy was almost always related to a violation of one these basic princi-ples Thus, what matters in the first instance is a good technology of cen-tral banking The art of monetary policy comes into play if a very good or

an excellent performance is to be achieved

I think are useful for practical monetary policy Of course, any attempt to define what is 'relevant' is highly arbitrary, and therefore I do not pretend that this part

of the book gives a comprehensive presentation of the state of the art of monetary theory; for this purpose the reader is referred to the book of Walsh (1998).

What is Money?

What this chapter is about

• The definition of money raises a difficult classification and aggregation problem.

• A simple microeconomic solution to this problem can be found if currency in tion is used as a 'crystallization point' for the definition Adding perfect substitutes to this item leads to the concepts of the monetary base and the money stock M1.

circula-• The standard approach in the literature uses the functions of money as a point of departure However, this approach is prone to a circularity In addition, only the means

of payment function provides a useful criterion for the definition of money It leads again to the monetary base and M1.

• The broader concepts of money (M2 and M3), which are simply derived by ric tests, lack a sound theoretical basis.

economet-1.1 INTRODUCTIONThe definition of 'money' is a natural starting point for any comprehen-sive monograph or textbook on monetary theory and monetary policy.Unfortunately, although a myriad of books and articles have been written

on these issues, the definition of money can still be regarded as an almostunresolved issue:

If a principal purpose of definition is to bring 'peace at mind' monetary nomics has failed us The definition of its central concept has been unsettled for many decades and is more controversial than ever (Osborne 1992: 602)

eco-In fact, the variety of instruments used on today's financial marketsmakes it extremely difficult to resolve the two main problems in establish-ing a satisfactory definition of money The first concerns how a borderlinecan be drawn between those financial assets that are regarded as moneyand those that are simply 'other financial assets' This problem is referred to

as the 'classification problem' of the definition of money Second, if such aborderline has been drawn, what are the weights that are used for aggre-gating the different 'monetary assets' into a single asset called 'money'?This problem is referred to as the 'aggregation problem' of the definition ofmoney The most common approach to the solution of these problems is to

define 'money' by the functions of'money': this goes back to John Hicks,

who wrote:

Money is what money does Money is defined by its functions (Hicks 1967: 1)

It is obvious that this approach is prone to a circularity that is one of themain reasons for the unsatisfactory state of the 'definition of money' If it

is not clear what 'money' is, it is also not possible to describe the functionsof'money' Therefore, in addition to this approach, I present two other ways

of defining money:

1 a microeconomic approach, which uses currency in circulation as a

'crystallization point' for the definition of more comprehensive cepts of money;

con-2 an econometric approach, which defines 'money' according to the

stat-istical properties of alternative monetary aggregates

1.2 A MICROECONOMIC APPROACH TO

THE DEFINITION OF MONEY

A relatively simple approach to the definition of money starts with theobvious It assumes that currency in circulation must be a component ofany definition of money In other words, no one would accept a definition

of money that excludes this asset With this crystallization point, simplemicroeconomic logic can be used to derive two important concepts ofmoney If currency in circulation is regarded as money, then all assets thatare perfect substitutes can also be regarded as money

1.2.1 Defining the monetary base and Ml

From the perspective of commercial banks, a perfect substitute for currency

in circulation are reserves that these banks hold with the central bank If abank is in need of currency, it can always convert its reserves with the cen-

tral bank (R) into currency (C) at negligible transaction costs And if a bank's

cash balances are too high, it can always exchange them for reserves.Deposits with other banks are another perfect substitute but in aggregationover all banks they add up to zero This perfect substitutability betweenreserves and currency leads to a first important concept of 'money': the

monetary base (B) This is defined as follows:

There are many prominent economists who consider the monetary basethe most relevant concept of money.1 As a more detailed analysis of theprocess of money supply will show (Chapter 3), the monetary base differsfrom all other concepts of money in that it is a completely exogenous vari-

able Many economists regard vcogeneity as a decisive quality of'money'.2

From the perspective of private non-banks, an almost perfect substitute

to currency in circulation are sight deposits Today, especially with dispensers, it is almost always possible to transfer an overnight deposit (D)into currency at very low transaction costs This leads to a second import-

cash-ant concept of money, the money stock ML This is defined as:

Thus, for the money stock Ml and for the monetary base, the tion problem can be solved rather easily by the criterion of perfect sub-stitutability This also gives an answer to the aggregation problem If theassets included in each of two definitions of money can be regarded asperfect substitutes, their weights should be a factor of one

classifica-1.2.2 Broader monetary aggregates (simple-sum aggregates)

In monetary policy the money stock M1 is regarded as an important cator, but most central banks look also at more broadly defined monetaryaggregates, which are labelled M2, M3, or even M4 The European CentralBank (ECB) uses a 'reference value' for the money stock M3 as a 'pillar' ofits stability-oriented monetary policy strategy (Chapter 9) The ECB definesthese broader aggregates as follows:

indi-M2 = M1 + Deposits with agreed maturity up to 2 years3

+ Deposits redeemable at notice up to three months;4

M3 = M2 + Repurchase agreements + Money market

fund shares/units and money market paper

+ Debt securities issued with maturity up to 2 years

1 See e.g Tobin (1980: 320): 'If there is a quantity that is the monetary anchor of the economy, this

is it.' Or Osborne (1992: 606): The money of the monetary theory is the monetary base.'

2 See James Tobin (1980:319): '[T]he nominal supply of money is something to which the economy must adapt, not a variable that adapts itself to the economy-unless the policy authorities want it to.' And Milton Friedman (1971: 2): 'Suppose that the quantity [of money] that people hold at a par- ticular moment of time happens to exceed the quantity they wish to hold Individuals will then seek to dispose of their excess money balances — However, they cannot as a group succeed— One man can reduce his money balances only by persuading someone else to increase his.'

3 Other central banks call such (or similar) assets 'time deposits'.

4 Other central banks call such (or similar) assets 'savings deposits'.

All these assets have been held with monetary financial institutions

(MFIs) in the euro area This definition includes the Eurosystem (i.e the

European Central Bank and the national central banks that participate inthe European Monetary Union), credit institutions, and money marketfunds that are located in the euro area

From a microeconomic perspective, the usage of these broader monetaryaggregates is not so easy to justify It is obvious that not all assets that areincluded in M3 and M2 can be regarded as perfect or close substitutes tocurrency in circulation This becomes evident also from the fact that cur-rency and overnight deposits are normally non-interest-bearing, while allthe other assets included in M2 and M3 bear positive interest rates whichcan sometimes deviate considerably from zero For such broader monetaryaggregates, therefore, it becomes difficult to solve the classification andaggregation problem

• If some of the components of a specific monetary aggregate are nolonger perfect substitutes for currency, it is difficult to explain why theseassets are regarded as 'money' while other assets are excluded The bor-derline between 'money' and 'other financial assets' becomes arbitrary

• With an imperfect substitutability of assets, it is also not clear whichweights should be used in order to calculate a monetary aggregate Thewidespread practice of most central banks simply to sum up components

that are not perfect substitutes ('simple-sum aggregates 1 like M2 or M3)

is difficult to reconcile with elementary microeconomic principles.Thus, all broader simple-sum monetary aggregates raise a serious classi-fication and aggregation problem

1.2.3 TheDivisia index

An interesting solution for the aggregation problem of more broadlydefined aggregates are the so-called Divisia monetary aggregates based onthe Divisia index.5 The main idea behind this index is a weighting schemefor the components of the broader monetary aggregate that reflects the dif-ferences in their moneyness It is assumed that these differences can bemeasured by the 'user costs' of financial assets.6 User costs are defined as

the costs of using a more liquid asset with a low yield instead of a less

5 Francois Divisia (1889-1964) was a French statistician The Divisia approach was proposed mainly

by William Barnett (1980, 1982) A survey on the use of Divisia indices in monetary theory and policy

is provided by Barnett et al (1992) For a comprehensive theoretical analysis, see Reischle (2000).

6 A model-based reasoning for the following concept of the so-called 'user costs' is provided by Barnett (1978).

It is important to note that it is not the user costs that serve as the weightsfor the different assets, but their shares at the sum of the 'expenditures forholding monetary assets' Because statistical indices have to be calculatedfor two different points in time, the weight in the 'time discretionary'Divisia index formula (1.6) consists of the (simple) average of the weights

for the two moments, t - 1 and t.

While the Divisia index offers a convincing answer to the aggregationproblem, it can contribute little to the solution of the classification prob-lem This is due to the fact that the concrete reference asset (long-termbonds with 10 or 30 years' maturity?; government bonds or corporatebonds?) has to be chosen in an arbitrary way Therefore the dividing linebetween 'money' and 'other financial assets' still remains somewhatblurred In practice, this problem becomes even more obvious since Divisiaindices are normally calculated for the components of the traditionalsimple-sum aggregates, i.e Ml and M3

with the weighting scheme s* t defined as

liquid asset with a higher yield For this purpose, a reference asset-which

should not be liquid, at least in principle-has to be defined Its yield (R)

serves as a benchmark In practice, the benchmark is normally the yield oflong-term bonds, which are regarded as the most 'illiquid' asset The usercosts (TT,) of all other assets can be calculated as the discounted differencebetween the yield of an asset (if) and the reference yield:

Taking the user costs as the 'price' of holding a monetary asset, it isstraightforward to apply index theory to solve the aggregation problem.The big advantage of the Divisia index over other statistical index for-mulas-for example Laspeyres', Paasche's, or Irving Fisher's 'Ideal Index'-

is the easy interpretation of its general principle: the growth rate of the(Divisia) monetary aggregate (D) corresponds to the weighted sum of thegrowth rates of the different assets Using logs, for f different monetary

assets nij, one obtains

and

Source: Federal Reserve Bank of St Louis.

A main problem of this definition of money is that it defines the ness of different assets by interest differentials As these differentials varyconsiderably over time, the index would imply that the moneyness of, say,

money-a non-interest-bemoney-aring sight deposit is higher in times of high long-terminterest rates than in times of low interest rates Whereas one cannot fullyexclude the fact that the liquidity of a monetary asset varies over time, onegets inconclusive results in times of an inverse yield curve According tothe user-cost concept, this would mean that the liquidity (or moneyness)

of long-term bonds is higher than that of short-term time deposits Inpractice, to solve the difficult conceptual problem of the definition of the'illiquid' reference asset, one uses a long-term bond interest rate and takes

an additional charge, high enough to avoid negative user costs In addition

to the problem of the correct measure for liquidity, it is unsatisfactory tohave no information about the value of the money stock but only about itschange Although the change in monetary aggregates (or, better, theirgrowth rates) is of special importance in the discussion of monetary policy,knowledge about the money stock is helpful in assessing the liquidity situ-ation of an economy Cumulating the growth rates starting at a fixed point

in time can help to overcome this problem, but the obtained stock valuedepends crucially on the (arbitrarily) chosen starting value and the calcu-lation procedure of the growth rates because the Divisia index normally ispath-dependent

In practical monetary policy, Divisia indices have played no major role.Neither the ECB nor the Bank of Japan publishes regular data for Divisiamonetary aggregates The lack of interest in this concept is due partly to thefact that monetary targeting in general is not a very attractive solution formost central banks In addition, in its communications with the generalpublic, a central bank has to take into account that the rationale of suchaggregates is very difficult to understand Nevertheless, the Bank of Englandcalculates and publishes data for Divisia monetary aggregates as one of many

indicators in its Inflation Report Figures 1.1 and 1.2 show that in the United

States the difference between the simple-sum Ml aggregate and a Divisia Mlaggregate is very limited However, for M3 in periods with high interest rates(e.g 1979-82) the Divisia index behaves differently from the simple-sumindex Thus, for a consequent monetary targeting of a broader monetaryaggregate, the differences between the two aggregates are not trivial

1.2.4 Conclusion

In sum, only the monetary base and the money stock M1 have a soundmicroeconomic foundation For the broader monetary aggregates M2 and

Source: Federal Reserve Bank of St Louis.

M3, no convincing solution to the classification problem has yet beenfound The Divisia index at least provides an answer to the aggregationproblem, but it has not been very attractive for policy-makers Thus, thecurrent practice of the ECB, which attaches a 'prominent role' to the (sim-ple-sum) money stock M3, relies on a rather arbitrary solution to the clas-sification and aggregation problem of the definition of money.

1.3 DEFINING MONEY BY ITS FUNCTIONS

Without a definition of'money', it does not seem very promising to speak

of the functions of 'money' Nevertheless, the definition of 'money' by itsfunctions is very popular in the literature Most authors avoid the problem

of circularity by the implicit assumption that the functions of 'money' are

so obvious that they need no theoretical derivation Traditionally, the

func-tions of money are presented in the form of the so-called triad of the money

functions, which comprises:

1 the function of the means of payment,

2 the function of the store of value, and

3 the function of the unit of account

1.3.1 Money as a means of payment

The function of money as a means of payment is certainly its most ant function An economy in which an asset is generally accepted as ameans of payment is much more efficient than an economy where this isnot the case Without 'money', all economic transactions have to be carried

import-out as an exchange of goods against goods This would require a double

coincidence of wants: The shivering baker needs to find a hungry taylor.'

Whenever a bilateral exchange is impossible, economic agents have toincur search and information costs to build up indirect chains of exchange.7

A theoretical alternative to the instrument of money is an economy with

a perfect capital market In this case, each agent can always bridge the gapbetween the goods he can sell today (loaves of bread) to one agent and the

7 See Jevons (1882: 1): 'Some years since, Mademoiselle Zelie, a singer of the Theatre Lyrique at Paris, made a professional tour round the world, and gave a concert at the Society Islands In exchange

for an air from Norma and few other songs, she was to receive a third part of the receipts When

counted, her share was found to consist of three pigs, twenty-three turkeys, forty-four chickens, five thousand cocoa-nuts, besides considerable quantities of bananas, lemons, and oranges At the Halle in Paris, as the prima donna remarks in her lively letter, printed by Mr Wolowski, this amount of live stock and vegetables might have brought four thousand francs which would have been good remuneration for five songs In the Society Islands, however, pieces of money were very scarce; and as Mademoiselle could not consume any considerable portion of the receipts herself, it became necessary in the mean time to feed the pigs and the poultry with the fruit.'

goods he needs today (a coat) from another agent, by borrowing on thecapital market Thus, the use of money can be explained only for situationswith imperfect information and with transaction costs Of course, this is thereal world.

Defining money by the function of the means of payment leads to a lar result as that derived by the microeconomic approach in Section 1.2

simi-For transactions among banks, the means of payment is the monetary base,

i.e deposits with the central bank Again deposits with other banks canalso be used but they cancel in aggregation For transactions among

non-banks (and among banks and non-banks), only the components of the

money stock Ml can be used as a means of payment: with all other ponents of the money stock, M2 and M3, direct payments either amongbanks or among non-banks are not possible Thus, this function wouldclearly exclude broader monetary aggregates from a definition of money

com-1.3.2 Money as a store of value

If money is used as a means of payment, it necessarily also serves to someextent as a store of value Friedman and Schwartz (1970) describe money

as a 'temporary abode of purchasing power' As a person who receives anasset for payment cannot spend it simultaneously, he or she always has tohold it for a certain period of time Thus, all assets that are used as a means

of payment should be qualified at least to some extent as a store of value

In periods of hyperinflation, all agents will try to keep the time period ing which they hold money as short as possible, which aggravates theinflationary process But this also reduces the willingness to accept a cer-tain asset as a means of payment Thus, in many high-inflation countries,instead of the national currency, the US dollar has been used as a means

dur-of payment and as a store dur-of value ('dollarization') In other words, an assethas to provide some quality as a store of value in order to function as awidely accepted means of payment

The function of store of value is of little help if it is used as the sole terion for the moneyness of a financial asset, since almost any financialasset (and even many non-financial assets) can be used as a store of value.The result would be an extremely broad definition of 'money'.8 The stat-

cri-istical concept for such a broad definition is the concept of monetary wealth, i.e the sum of all financial assets that are held in an economy (see

Appendix 1.2) While such a solution would overcome the classification

8 See Brunner and Meltzer (1971:803): 'Defining money as a temporary abode of purchasing power does not distinguish between properties of assets or between a monetary and a barter economy in a manner independent of the exchange function.'

problem, it would still be confronted with a difficult aggregation problem.However, in the academic literature, and also in practical monetary policy,such a broad aggregate plays no decisive role.

1.3.3 Money as a unit of account

The third function of money is somewhat different from the first and thesecond function The functions of the means of payment and the store of

value refer to concrete assets (currency, time deposits, savings deposits) The function of the unit of account is provided by an abstract currency unit

that is the legal currency for a certain area For the member countries of theEuropean Monetary Union (EMU) the euro became the legal currency on

1 January 1999 The former national currency units are still in use until theend of the year 2001, but legally they are non-decimal subunits of the euro

It is obvious that this function of 'money' cannot be used to solve theclassification problem of money, since all financial assets have to bedenominated in some currency unit Nevertheless, it is useful to presentthis function of money in some more detail

As a numeraire for the exchange of goods and services, money (or, more

precisely, a common monetary unit) helps to reduce transaction and mation costs Without money, the prices of all goods would have to be

infor-expressed in terms of the prices of other goods Thus, with n goods, an

economy would have (n/2)(n - 1) relative prices Such a situation would

be associated with high transaction costs, because it would be very difficult

to make price comparisons: is a shirt that costs 36 apples in one shopcheaper than the same shirt that costs 12 bottles of beer in another? If aneconomy has a common unit of account or numeraire, it is possible toexpress the prices of all goods in the terms of the nth good which serves as

a numeraire Thus, the (n/2)(n — 1) relative prices are reduced to n - 1

absolute prices Like the function of the means of payment, the function of

the numeraire leads to an enormous reduction of transaction costs

As a standard of deferred payments, money helps to organize the

intertemporal exchange of goods If a worker wants to consume less thanshe earns in the present and consume more than she earns in the future, shecan organize such a transfer by using financial markets or banks In a debtcontract which is used for that purpose, a standard of deferred paymentshas to be specified It defines the value of the debt at the time of repaymentand can be different from the means of payment that has to be used for thedischarge of that debt For instance, if A today lends 100 euro to B for oneyear, A and B could agree to use the US dollar as the standard of deferredpayments, which means that the debt (including the interest payment of,

say, 5%) would be defined in the actual euro exchange rate of the dollar.

At a rate of 1.00 dollar per euro, the debt would be 105 dollars For the charge of the debt, they could agree to use the euro Thus, in one year Bwould have to pay the amount of 105 dollars times the actual euroexchange rate of the dollar If it is 1.10 dollars, the repayment would be95.45 euros

dis-As a standard for measuring increases in wealth, money helps to

aggre-gate the different assets in a balance sheet (which can be denominated indifferent currency units) This function is of special importance in the context

of taxation and enterprise accounting An interesting aspect of that function

is that the measured increase in wealth can be different if the balance sheetsfor two different years are made up, for instance, in dollars or in euros.1.4 DEFINING MONEY BY ITS STATISTICAL PROPERTIES

A 'pragmatic' approach to definition of'money', and thus to the solution ofthe classification and aggregation problem, is adopted by many academicsand central banks Instead of trying to define the 'relevant' money stockfrom theoretical considerations and testing hypotheses based on it empir-ically, the opposite procedure is applied: the relevant money stock isdefined by its statistical properties For instance, the European CentralBank prefers the money stock M3 to the money stock Ml mainly for thefollowing reason:

The available empirical evidence suggests that broad monetary aggregates (i.e measures of money that include a wide spectrum of deposits, embracing time and savings deposits, as well as close substitutes for them, such as marketable short- term bank liabilities) exhibit the properties required for the announcement of a ref- erence value In the past, the demand for euro area broad money has been stable over the long run Broad aggregates have been leading indicators of developments

in the price level This contrasts with the empirical properties of euro area narrow money, which, although controllable using short-term nominal interest rates, exhibited neither stability with nor significant indicator properties for the price level 9 (ECB 1999a:48)

The problems of this approach for the definition of money are veryclearly described by D K Osborne:

Belief in this proposition [the stability of the demand for moneyl has led some

monetarists to define money as that (set of liquid assets) which has a stable demand

function This definition is rarely stated in so many words, but it is implicit in some

of the best works on money demand — Whether implicit or explicit, the definition

9 With the same arguments, the Deutsche Bundesbank (1995) also favoured M3 over Ml.

puts the cart before the horse We have to define and identify money before we can

test the stability of its demand This stability (if it exists) is to be demonstrated empirically, not deemed true axiomatically The theory says that money does

a,b,c, — If we then define money as that which comes closest to doing a,b,c, ,

we have accepted the theory as true (or true enough for our purpose) and forgone all opportunities to test it (Osborne 1992: 603)

Thus, instead of deriving hypotheses from a theory and confrontingthem with the challenge of falsification, which according to the researchprinciples laid down by Karl Popper (1959) is a fundamental requirement forany scientific progress, this approach fosters an immunization of theoriesagainst reality Nevertheless, it explains why broad monetary aggregatesare more popular than the money stock M1 in monetary policy, although thenarrow aggregate relies on a much more solid theoretical basis

1.5 THE IMPORTANCE OF A 'CORRECT'

DEFINITION OF MONEY

In spite of the considerable theoretical and statistical difficulties that areinvolved in arriving at a 'correct' definition of money, many central bankshave been able to pursue efficient and successful monetary policies Thisimplies that a 'correct' definition of money is not always necessary for theconduct of an efficient monetary policy As we shall see in the discussion

of different transmission processes of monetary policy (Chapter 4), thisprocess can be analysed either in terms of quantities of money or in terms

of relative prices, above all the structure of interest rates The first approachrests on the Quantity Theory of Money and the monetarist paradigm, whilethe second is based mainly on Keynesian approaches

It is clear that the correct definition of money is of crucial importance to

a monetary policy strategy that strictly follows a monetarist philosophy.However, we shall see that no central bank has ever followed a monetaristapproach for a prolonged period of time (Chapter 9) Thus, the profoundproblems of a theoretically correct definition of money that have plagued,and will continue to plague, academics at universities can be taken some-what more lightly by policy-makers at the central banks

Appendix 1.1 Definitions of money in the euro area,

the United States, Japan, and the UK

A comparison of the definitions of money in the euro area, Japan, theUnited Kingdom, and the United States is provided by Table A 1.1 This

M1 M2 M3 M4

Euroarea(ECB)

Japan

Currency in circulation + Overnight deposits

M1 + Deposits with agreed maturity up to 2 years + Deposits redeemable at notice up to 3 months Currency in circulation + M1 + Quasi-money Deposit money

United Kingdom According to ECB According to ECB

United States Currency + checkable

deposits

M1 + Household holdings

of savings deposits, time deposits, and retail money market funds

M2 + Repurchase agreements + Money market fund shares/units and Money market paper + Debt securities up to 2 years Not reported, but: M2 + Certificates of deposit (CDs) According to ECB

M2 + Institutional money funds -i- Managed liabilities

of depositories, namely large time deposits, repurchase agreements, and Eurodollars

Notes + Coin + Bank deposits + Building society deposits (including CDs, Sterling, commercial paper, and other short-term paper)

Sources: European Central Bank fwww.ecb.intl: Bank of Japan (www.boj.co.jp): Bank of England (www.bankofengland.co.uk): Federal Reserve Board (www.bog.frb.fed.us).

shows that Ml is defined in rather similar ways by all four central banks;for the broader aggregates, however, the definitions are more differentreflecting the varying degrees of importance of alternative financial assets

in these economies

Appendix 1.2 The concept of (net) financial assets

(flow of funds analysis)

The broadest 'monetary aggregate' is an aggregate that includes all financialassets in an economy While this approach would solve the classificationproblem of defining money, it would be rather difficult to find a convincingsolution to the aggregation problem, since such an aggregate would includecash as well as government bonds with a maturity of thirty years Here Ipresent the concept of (net) financial assets, since it provides importantinsights into fundamental macroeconomic accounting identities

As a simple-sum measure, many countries publish data on financialassets and liabilities of their economy (flow of funds accounts) For a bet-ter understanding of these relationships, we start with a very simplifiedbalance sheet of a private agent:

Assets

Real assets (RA) Financial assets (FA)

Liabilities Liabilities (/.)

Equity (E)

Assets include only two categories: real assets and financial assets On theliabilities side, the balance sheet includes liabilities and the equity (or totalwealth) of the private agent An additional concept can be derived by

calculating net financial assets (NFA) as the difference between financial

assets and liabilities:

Equity (or total wealth) is then the sum of net financial assets and realassets:

For a closed economy, aggregate total wealth can be calculated as the sum

of the total wealth of its inhabitants: