The making of Monetary Policy in the UK 1975-2000

The Making of Monetary Policy in the UK, 1975–2000

David Cobham

JOHN WILEY & SONS, LTD

The Making of Monetary Policy in the UK, 1975–2000

SERIES IN FINANCIAL ECONOMICS AND QUANTITATIVE ANALYSIS Series Editor:

Stephen Hall, Imperial College Management School, UK

Editorial Board: Robert F. Engle, University of California, USA John Flemming, European Bank, UK Lawrence R. Klein, University of Pennsylvania, USA Helmut L¨utkepohl, Humboldt University, Germany

Developments in Forecast Combination and Portfolio Choice Christian Dunis, Allan Timmermann and John Moody (edited by) Stable Paretian Markets in Finance Stefan Mittnik and Svetlozar Rachev Financial Innovation Philip Molyneux and Nidal Shamroukh Nonlinear Modelling of High Frequency Financial Time Series Christian Dunis and Bin Zhou Macroeconomic Modelling in a Changing World: Towards a Common Approach Chris Allen and Stephen Hall (edited by) Forecasting Financial Markets: Exchange Rates, Interest Rates and Asset Management Christian Dunis (edited by) Quantitative Financial Economics: Stocks, Bonds and Foreign Exchange Keith Cuthbertson Maximum Entropy Econometrics: Robust Estimation with Limited Data Amos Golan, George Judd and Douglas Miller

The Making of Monetary Policy in the UK, 1975–2000

David Cobham

JOHN WILEY & SONS, LTD

C 2002 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, Copyright
West Sussex PO19 8SQ, England

Telephone

(+44) 1243 779777

Email (for orders and customer service enquiries): [email protected] Visit our Home Page on www.wileyeurope.com or www.wiley.com All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London, W1T 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to [email protected], or faxed to (+44) 1243 770571. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Other Wiley Editorial Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 22 Worcester Road, Etobicoke, Ontario, Canada M9W 1L1

Library of Congress Cataloging-in-Publication Data British Library Cataloguing in Publication Data Cobham, David P. The making of monetary policy in the UK, 1975-2000 / David Cobham. p. cm.—(Series in financial economics and quantitative analysis) Includes bibliographical references and index. ISBN 0-471-62387-3 (alk. paper) 1. Monetary policy—Great Britain. I. Title. II. Series. HG939.5 .C574 2001 339.5 3 094109045—dc21

2002071361

A catalogue record for this book is available from the British Library ISBN 0-471-62387-3 Typeset in 10.25/12 pt Times by TechBooks, New Delhi, India Printed and bound in Great Britain by TJ International, Padstow, Cornwall, UK This book is printed on acid-free paper responsibly manufactured from sustainable forestry in which at least two trees are planted for each one used for paper production.

Contents List of Figures

ix

List of Tables

xi

Series Preface

xiii

Preface

xv

Acknowledgements

xvii

List of Abbreviations

xix

1

Analytical Introduction: Alternative Monetary Frameworks 1.1 The Rationale for Intermediate Targets in Monetary Policy 1.2 Types of Monetary Policy Target 1.3 Monetary Policy Rules 1.4 Conclusion

1 1 3 9 10

2

Historical Introduction 2.1 A Macroeconomic Narrative, 1975–2000 2.2 Monetary Frameworks and Phases 2.3 Analysing the Determination of the Money Supply in the UK 2.4 Instruments of Monetary Control

11 11 15 20 23

3

Monetary Targeting, 1977–1986 3.1 Monetary Aggregates and the Type of Targets Used 3.2 Target Ranges and Economic Forecasts 3.3 Target Ranges and Credit Counterpart Forecasts 3.4 Financial Innovation and Monetary Targeting 3.5 Targets and Outturns for Monetary Growth 3.6 Monetary Growth, Target-Setting and Financial Innovation 3.7 Conclusion: The Abandonment of Monetary Targets Appendix 3.1 Monetary Targets and Outturns

27 27 29 32 37 42 47 49 50

4

The Interregnum and the Lawson Boom, 1985–1990 4.1 Explanations of the Lawson Boom 4.2 Macro Policy Decisions in the Second Half of the 1980s

53 53 61

vi

Contents

4.3 4.4 4.5

Financial Liberalisation versus Excessive Depreciation A Monetary Policy Reaction Function with Results for 1985–1990 Conclusions

65 67 70

5

The ERM Interlude, 1990–1992 5.1 Into the ERM 5.2 Out of the ERM 5.3 Alternative Explanations of the 1992–1993 EMS Crises 5.4 The UK Case in the Light of Other Explanations 5.5 A Counterfactual Analysis 5.6 Conclusions

6

The 1993–1997 ‘New Framework’ for Monetary Policy 6.1 The New Monetary Policy Framework 6.2 Evaluation of the New Framework 6.3 The Credibility of Policy under the New Framework 6.4 The post–ERM Reaction Function 6.5 Conclusions

93 93 95 104 105 105

7

Monetary Policy under the MPC, 1997–2000 7.1 Inflation Targeting with Instrument Independence 7.2 Central Bank Independence 7.3 The Workings of the MPC Regime 7.4 The MPC’s Reaction Function 7.5 Conclusions

107 107 108 111 117 118

8

Interest Rate Decisions and the Exchange Rate 8.1 The Official Concerns Underlying Interest Rate Changes 8.2 A Preliminary Look at Misalignments 8.3 A Preliminary Look at Large Exchange Rate Changes 8.4 Interim Conclusions Appendix to 8.1 Interest Rate Changes

119 119 124 127 129 130

9

Monetary Policy and the Exchange Rate 9.1 Monetary Policy and the Exchange Rate, by Phases 9.2 The Major Misalignments 9.3 Large Exchange Rate Changes 9.4 Sterling Crises 9.5 Conclusions

137 137 153 155 157 157

Interest Rate Smoothing 10.1 International Evidence 10.2 Large Changes and Reversals in the UK 10.3 Explanations of Smoothing 10.4 Evidence on Smoothing for the UK 10.5 Evaluating the Explanations of Interest Rate Smoothing 10.6 Conclusions

161 161 162 166 168 173 189

10

73 73 77 80 89 91 92

11

Contents

vii

Conclusion: Competence, Commitment and Monetary Policy Performance 11.1 Demand and Supply Shocks, Inflation Expectations and Control 11.2 Competence and Commitment 11.3 Interest Rate Smoothing and the Performance of Monetary Policy 11.4 The External Dimension 11.5 Conclusions

191 192 194 195 196 196

References

199

Index

207

Figures 3.1 3.2 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 6.1 6.2 6.3 7.1 7.2 8.1 8.2 8.3 9.1 10.1 10.2

Broad money velocity M0 velocity Monetary growth Income, exports and investment Income, exports and imports Income, exports and consumption The sterling oil price and the exchange rate Actual versus pre-ERM counterfactual interest rate Actual versus post-ERM counterfactual interest rate Actual versus MPC counterfactual interest rate Actual versus Taylor rule counterfactual interest rate Inflation, 1992–1997 Bond yield spreads, 1990–1997 UK policy rate, 1992–1997 Inflation, 1997–2000 Bond yield spreads, 1994–2000 Policy rate (daily data), 1975–2000 UK exchange rates, 1975–1998 UK real exchange rate, 1975–1998 Relative unit labour costs, 1991–2000 Policy rate (end quarter), 1975–2000 Policy rate (daily data), 1997–2000

40 41 54 60 60 61 63 68 69 69 70 96 97 101 112 112 122 125 126 151 171 172

Tables 1.1 2.1 2.2 2.3 3.1 3.2 3.3 3.4 3.5 3A.1 3A.2 3A.3

Alternative monetary policy targets Macroeconomic outcomes Macroeconomic policy indicators Monetary policy regimes £M3 targets and other indicators £M3 and counterparts, forecasts and outturns Main financial innovations in the UK in the 1970s and 1980s Changes in velocity £M3 targets and outturns Monetary targets, 1976–1979 Monetary targets for £M3 under the MTFS, 1980–1983 Monetary targets for M1 and PSL2 under the MTFS, 1982–1983 and 1983–1984 3A.4 Monetary targets for £M3 and M0 under the MTFS, 1984–1985 to 1986–1987 3A.5 Later targets and monitoring ranges for M0 and M4 4.1 Contributions to growth of demand 6.1 Monthly Monetary Meetings, 1994–1997 6.2 The degree of agreement between Governor and Chancellor, January 1994 to May 1997 7.1 Cukierman’s index of CBI for the UK 7.2 Grilli et al.’s index of political independence for the UK 7.3 Grilli et al.’s index of economic independence for the UK 8.1 UK interest rate changes, 1975–2000 8.2 Frequency of particular concerns 8.3 Large misalignments 8.4 Large exchange rate changes 8A.1 Explicit concerns underlying interest rate changes 9.1 Phases of monetary policy 9.2 Exchange rates, inflation and interest rates by phases 9.3 Summary of Section 9.1 9.4 Large exchange rate changes 9.5 Interest rates in the months before sterling crises 9.6 Interest and exchange rates over and after sterling crises 10.1 Very large policy rate changes 10.2 Different types of policy rate change

8 12 13 19 30 33 38 41 43 50 50 50 51 51 59 99 100 109 109 110 120 122 127 128 130 138 139 152 156 158 159 163 163

xii

10.3 10.4 10.5 10.6 10.7 10.8 10.9

List of Tables

Large policy rate changes Policy rate reversals Disaggregated data on smoothing by change Disaggregated data on smoothing by quarter Data on smoothing by quarters in percentages MPC discussions of interest rate decisions, 1997–2000 MPC members’ dissenting votes, June 1997–December 2000

164 165 168 170 171 174 183

Series Preface This series aims to publish books which give authoritative accounts of major new topics in financial economics and general quantitative analysis. The coverage of the series includes both macro and micro economics and its aim is to be of interest to practitioners and policymakers as well as the wider academic community. The development of new techniques and ideas in econometrics has been rapid in recent years and these developments are now being applied to a wide range of areas and markets. Our hope is that this series will provide a rapid and effective means of communicating these ideas to a wide international audience and that in turn this will contribute to the growth of knowledge, the exchange of scientific information and techniques and the development of cooperation in the field of economics. Stephen Hall Imperial College Management School, UK

Preface Over the quarter of a century with which this book is concerned, the UK has had an extraordinarily diverse experience of monetary policy and monetary regimes. Monetary policy has been transformed, from attempts to control broad money from the supply side with the use of indirect controls on bank lending, to an almost exclusive focus on interest rates in a context of inflation targeting. The exchange rate has at times been fixed, at other times almost perfectly flexible, and at other times again more or less managed. Meanwhile the real economy has experienced large variations in growth, together with what most observers have seen as a sharp rise and then a gradual decline in the NAIRU; inflation has varied between 25% and 2%. This is a book about the making of monetary policy in the UK, about how and why the monetary regimes changed over the period, and how and why the monetary authorities took the decisions they did about monetary growth, interest rates and the exchange rate. The book makes intensive use of the “official sources”, particularly the publications of the Bank of England; it is assumed that these sources have to be read critically, but nevertheless that for most of the period they convey a large part of the truth about the way economic developments were viewed and decisions were made. The first two chapters contain analytical and historical background. Chapter 1 examines the main monetary frameworks that have been implemented or proposed, that is the different types of monetary policy target, with particular reference to the issues of stabilisation properties, influence on inflation expectations and ease of control which are emphasised in the standard literature; it also introduces the Taylor monetary rule. Chapter 2 provides an outline historical narrative of macroeconomic and monetary policy, together with a classification of phases according to monetary regimes; it discusses how the determination of the money supply was regarded in the UK, especially in the first decade of the period, and reviews the evolution of the instruments of monetary control over the period. Chapters 3 to 7 investigate the different monetary frameworks implemented in the UK: Chapter 3 discusses the operation of monetary targets, Chapter 4 the second half of the 1980s when there was little or no explicit monetary framework, Chapter 5 the ERM membership phase of 1990–92, Chapter 6 the initial inflation targeting phase of 1993–97, and Chapter 7 the phase of inflation targeting with Bank of England control of interest rates, 1997–2000. While the open economy issues are an important element in all of the preceding analysis, they are examined more directly in Chapters 8 and 9. Chapter 8 presents an analysis of the monetary authorities’ concerns with respect to each interest rate change over the period, on the basis of explicit statements in the official sources; this shows the importance of exchange rate concerns, particularly in the years up to 1989. It also takes a preliminary look at “large” misalignments and “large” exchange rate changes. Chapter 9 looks in more detail at the

xvi

Preface

response of monetary policy to the movement of the exchange rate in each phase of policy, and then at “large” misalignments, “large” exchange rate changes and sterling crises. Chapter 10 considers the extent of the smoothing of interest rates by the UK monetary authorities, and investigates the reasons for smoothing with particular reference to the post1997 period, for which the minutes of the Monetary Policy Committee provide evidence that is not available for the other phases. Chapter 11 concludes: it returns to the issues highlighted in the analytical introduction of Chapter 1, and argues that a full assessment of the performance of monetary policy under different frameworks needs also to take account of the competence and the commitment of the monetary authorities. Much of the work on this book was done while I was a Houblon-Norman research fellow at the Bank of England during 2001, and I am grateful to the Houblon-Norman Trust and the Bank for that opportunity. I am also grateful for their help and assistance to the secretaries at the MPC unit: Janette Parry, Elaine Kalli, Bernie Camp and Stephanie Smith. While I was at the Bank I benefited from discussions with and/or comments by Bill Allen, Chris Allsopp, Peter Andrews, Kosuke Aoki, Ravi Balakrishnan, Charlie Bean, Alec Chrystal, Roger Clews, Rebecca Driver, Andy Haldane, Ben Martin, Gordon Midgley, Ed Nelson, Kalin Nikolov, Sushil Wadhwani, Peter Westaway and Geoff Wood inside the Bank; and from Alex Cobham, Gus O’Donnell and Simon Wren-Lewis outside. I am grateful to Chris Jensen-Butler at St Andrews for his encouragement and support, and to my colleagues there for comments and discussions on a range of issues. Finally, I should also like to record my gratitude for the insights I have gained over the years from discussions with, from comments by and/or from the writings of Mike Artis, Charles Goodhart, Antonio Gramsci, David Laidler, Mike Sumner and George Zis. None of these people (nor the Houblon-Norman Trust nor the Bank of England) bear any responsibility for the errors, omissions, methods or judgements in what follows.

Acknowledgements Various chapters in this book draw in part on previously published papers by the present author, revised versions of sections of which are reproduced by permission of the copyright holders as follows: Chapter 2: “The money supply process”, in Surveys in Monetary Economics, Volume II, edited by C. Green and D. Llewellyn, Blackwell, 1991, pp. 44–78, copyright Blackwell Publishing; and “Monetary Policy”, in M. Artis and D. Cobham (eds), Labour’s Economic Policies 1974–79, Manchester University Press, 1991, pp. 38–55, copyright Manchester University Press. Chapter 3: “UK monetary targets 1977–86: picking the numbers”, in D. Cobham, R. Harrington and G. Zis (eds), Money, Trade and Payments: Essays in honour of D.J. Coppock, Manchester University Press, 1989, pp. 38–56, copyright Manchester University Press; and “Financial innovation and the abandonment of monetary targets: the UK case”, in R. O’Brien and T. Datta (eds), International Economics and Financial Markets, Oxford University Press, 1989, pp. 245–66, copyright The AMEX Bank Review. Chapter 4: “The Lawson boom: excessive depreciation versus financial liberalisation”, Financial History Review, April 1997, vol. 4, pp. 69–90, Cambridge University Press. Chapter 5: “Inevitable disappointment? The ERM as the framework for UK monetary policy 1990–92”, International Review of Applied Economics, May 1997, vol. 11, pp. 209–224, copyright Taylor and Francis (http://www.tandf.co.uk); “Causes and effects of the EMS crises of 1992–93”, Journal of Common Market Studies, 1996, vol. 34, pp. 583–602, copyright Blackwell Publishing; and “The German currency union and the crises in the European Monetary System”, in S. Frowen and J. H¨olscher (eds), The German Currency Union of 1990, Macmillan, 1997, pp. 36–58, copyright Palgrave. Chapter 6: “The post-ERM framework for monetary policy in the UK”, Economic Journal, July 1997, vol. 107, pp. 1128–41, copyright Blackwell Publishing. Permission from the Bank of England to reproduce Figure 4.5 (from the Bank of England Quarterly Bulletin, September 1986, p. 331) is also acknowledged.

Abbreviations BEIR:

(Bank of England) Inflation Report; references to, for example, 1994a/b/c/d refer to numbers 1, 2, 3 and 4 for 1994. BEQB: Bank of England Quarterly Bulletin; references to, for example, 1986a/b/c/d refer to numbers 1, 2, 3 and 4 for 1986. EMS: European Monetary System ERM: Exchange Rate Mechanism (of the EMS) FSBR: Financial Statement and Budget Report GEMSU: German Economic, Monetary and Social Union (1990) HC: House of Commons report (Hansard) IFS: International Financial Statistics (IMF) M0, M1, £M3, M3, M4: measures of the money supply—see Chapter 3, Section 3.1 MMM: Monthly Monetary Meetings; references to the Minutes of the MMMs are given in the form MMM, June 1996, §36 (i.e. paragraph 36) (the day of the month is given only where there were two meetings in the same month) MPC: Monetary Policy Committee; references to the Minutes of MPC meetings are given in the form MPC, August 1999, §20 NAIRU: non-accelerating inflation rate of unemployment PSL2: measure of the money supply—see Chapter 3, Section 3.1

1 Analytical Introduction: Alternative Monetary Frameworks The purpose of this chapter is to provide an analytical review of the main monetary frameworks or policy regimes that have been implemented or proposed, as a background to the discussion of the operation and performance of those frameworks in the rest of the book. The chapter starts with a brief survey of the rationale for intermediate targets for monetary policy, drawing on the targets and instruments literature and the rules versus discretion literature. That discussion is then used to compare alternative types of monetary policy target, from monetary targets through nominal income and inflation targets to exchange rate targets. Finally, monetary policy rules of the kind popularised by Taylor (1993) are examined.

1.1 THE RATIONALE FOR INTERMEDIATE TARGETS IN MONETARY POLICY There are two strands of literature which might offer a justification for intermediate targets in monetary policy: the targets and instruments literature initiated by Poole (1970), and the rules versus discretion literature which goes back to Simons (1936) but was transformed by Kydland and Prescott (1977). This section briefly reviews them both, in order to provide a background for the evaluations of alternative targets in Section 1.2.1 The background to Poole’s research was the debate over Milton Friedman’s (1960) monetary rule, under which the money supply should grow each year at a constant rate equal to the underlying rate of productivity growth. Provided the demand for money was stable, this rule would preclude the discretionary manipulation of monetary policy for inappropriate purposes, ensure price stability, and at the same time incorporate a limited counter-cyclical effect, since in booms (recessions) money would grow more slowly (rapidly) than nominal income. Poole contrasted this ‘monetarist’ policy with the ‘Keynesian’ policy of a fixed interest rate within a static closed economy IS-LM model,2 where the monetary rule can be understood as a constant money supply while the ‘Keynesian’ policy is represented as a horizontal rather than upwardsloping LM curve. He then analysed the outcome for income in the context of stochastic shocks to expenditure and money demand under these contrasting policies. The underlying assumption was that policy cannot respond quickly enough to shocks, but different types of policy rule involve different automatic responses which can be ranked against each other. 1 For a more detailed discussion see Fischer (1990), Friedman (1990) and Fischer (1994) or, for a simpler exposition, Cobham (1998: chapter 16). 2 Moreover, given the current macroeconomic context, Poole’s main concern was the stabilisation of income rather than the control of inflation. However, the main results carry over to an aggregate demand/aggregate supply (AD/AS) model with short run non-neutrality.

2 The Making of Monetary Policy in the UK, 1975–2000 In the presence of expenditure shocks that shift the IS curve, the monetary rule with its upward-sloping LM curve leads to a smaller variance of income than the Keynesian alternative. On the other hand, in the presence of money demand shocks that shift the LM curve, income varies under the monetary rule but remains constant under the fixed interest rate policy as variations in money demand are absorbed by variations in supply. In the full stochastic analysis where both shocks occur, which type of rule is preferable depends on the relative variance of the two types of shocks and also on their covariance. But in general the optimal policy will be a combination of the two rules in which money supply and interest rate both vary with each other and with income in a specific way, rather than a fixed level of the money supply. Poole’s analysis was developed in a different direction by Benjamin Friedman (1975). His basic argument was that the monetary authorities do not in fact control the money supply as such. What they do in the US institutional set-up (and comparable arguments apply elsewhere3 ) is to manipulate the interest rate and the level of bank reserves in an effort to set the money supply at some desired level M∗ . Policy should be thought of as a two-stage process. First the authorities choose M∗ as the level needed to produce some desired level of the ultimate target of policy (e.g. nominal income), on the basis of some ex ante assumptions about the likely shocks to the economy and the financial system. In the second stage they set their policy instruments—the interest rate and bank reserves—so as to generate money supply M∗ , but the generation of M occurs in the context of the actual (ex post) shocks. Friedman is able to show that the optimal policy is to use the money supply as an information variable, which provides information about the size and the nature of the shocks affecting the economy; that information can then be used to set the interest rate and bank reserves in such a way as to attain the target level of income. The conclusion of this strand of the literature is that, in terms of the short-run stabilisation properties of alternative policies, there is no good justification for an intermediate target such as that implied by M. Friedman’s monetary rule. However, Poole’s seminal analysis raised the issue of the response of the economy under different types of policy rule to different kinds of shocks, and the general proposition that the monetary authorities should ‘lean into’ portfolio or asset price shocks, but not aggregate demand shocks, can be found in recent work such as Cecchetti et al. (2000). The second strand of literature is the debate over rules versus discretion. This debate started with Simons’ (1936: 5) argument for rules to prevent ‘discretionary (dictatorial, arbitrary) actions’ by the monetary authorities, in order to preclude inappropriate political pressures on and/or actions by the authorities. M. Friedman, who had argued in his (1948) and (1953) papers that the lags in the implementation and operation of policy were such that discretionary policy was likely to be destabilising, proposed his monetary rule in 1960. The development of the natural rate hypothesis by Friedman (1968) and Phelps (1967) implied that there was scope for policy to affect real income only in the short run, and the rational expectations hypothesis (Lucas, 1972) implied that there was no systematic scope for policy even in the short run. But it was the concept of time-inconsistency developed by Kydland and Prescott (1977) which finally provided a convincing justification for formal rules or targets. If the authorities prefer low inflation to high inflation but would also like to set output (unemployment) above (below) the natural rate, and the private sector understands this, private

3 See Chapter 2, Sections 2.3 and 2.4, on the instruments of monetary control in the UK. The similarity between the US and UK contexts is that in both cases the authorities act on one or more policy instruments whose relationship to the (growth or level of the) money supply is not purely mechanical but subject to other influences.

Analytical Introduction: Alternative Monetary Frameworks 3 expectations of inflation will be higher than they would have been otherwise.4 The only possible long-run equilibrium, where the authorities have no incentive to stimulate output and the private sector knows that, will be at a ‘high’ level of inflation but at the natural rate of output.5 Both the authorities and the private sector would prefer to be at the natural rate with a lower level of inflation, but this is not attainable unless some ‘precommitment device’ is introduced. One form of the latter is the use of an intermediate target for monetary policy as a way of raising the cost to the authorities of reneging on the aim of low inflation: if the authorities announce they are going to keep monetary growth, say, within some particular target range and then fail to do so, they will ‘lose face’ and suffer damage to their political as well as economic credibility; their willingness to accept such damage convinces the private sector that they are indeed serious about the aim of low inflation. The second kind of solution to the time-inconsistency problem is the introduction of a monetary ‘constitution’ which gives the power to decide on monetary policy to a body which has no incentive to increase output, such as an independent central bank with a mandate to pursue price stability. This strand of literature does therefore produce a justification for intermediate targets, a justification originally applied to monetary targets but applicable also (mutatis mutandis) to the other main types of target. It should be noted, however, that this justification was constructed for the case where the monetary authorities are not independent and have specific preferences over the level of output. In the case of an independent central bank, a target may still be useful but its role will be more that of providing accurate information about the intentions of the central bank rather than that of persuading the private sector of the bank’s bona fides.6 The more accurate and precise the private sector’s expectations about inflation, the lower the unemployment cost of keeping inflation low.

1.2 TYPES OF MONETARY POLICY TARGET There are four different types of target to be considered here:7 a monetary target, in which the monetary authorities announce in advance a target for the growth of some particular monetary aggregate over some particular period; a nominal income target, where they announce a target for the growth of nominal income; an inflation target, where they announce a target for some measure of the rate of inflation; and an exchange rate target, where they announce a target for the exchange rate, usually defined as a central parity with a margin of permissible fluctuation either side of it. In analysing the purely domestic targets (monetary, nominal income and inflation) it will be assumed that they are being pursued in a regime of exchange rate flexibility.8 Monetary targets were introduced in most of the main developed countries (including the UK) in the mid-1970s; in some they continued to be used well into the 1990s, but in many 4 The basic point is also valid in Bean’s (1998) analysis. Here the monetary authorities are not aiming to set output above the natural rate in response to the existence of distortions that make the natural rate lower than the (optimal) Walrasian equilibrium rate. However, for reasons that relate in part to the political (electoral) process, they have a general preference for higher output. 5 It is assumed that the authorities’ loss function is quadratic in inflation, so that at a high level of inflation the authorities’ dislike of inflation is so great that they cannot gain by increasing output at the expense of higher inflation. 6 See Bean (1998) for the argument that delegating monetary policy to an independent central bank in itself solves the time-inconsistency problem. 7 Monetary, nominal income and exchange rate targets are ‘intermediate’ targets, whereas an inflation target is a target for the (presumed) ‘final’ objective of policy. 8 The alternative possible assumption, that of a closed economy, may be acceptable for the US economy but is not appropriate for the medium-sized open economy of the UK with which this book is concerned.

4 The Making of Monetary Policy in the UK, 1975–2000 (including the UK) they were abandoned in the early or mid-1980s. Nominal income targets have been discussed, but not implemented.9 Exchange rate targets were widely used in Europe in the 1980s and 1990s, in the form of fixed parities within the Exchange Rate Mechanism (ERM) of the European Monetary System. Inflation targets were introduced in a number of countries (including the UK), mainly small and relatively open economies, in the early 1990s. For each kind of target there are some important operational choices that have to be made. The authorities have to decide the precise definition of the variable to be targeted—which monetary aggregate? which price index? which anchor currency? They have to decide what period an announced target should cover—one year? several? They have to decide whether the target should be expressed as a range or as a single number. And they have to decide if there should be any exemptions, that is special circumstances under which they would not feel obliged to try to hit the target (and which they would announce in advance). The two strands of literature discussed in the previous section can be used to provide a framework for evaluating the different possible kinds of monetary policy target. With reference to the targets and instruments literature, an efficient target will act automatically to limit the variability of income in the face of shocks to expenditure and money demand: it will have good short-run stabilisation properties. On a broader view than Poole’s, and in an open economy AD/AS model, the analysis should cover supply as well as demand shocks. It will also be useful to distinguish foreign from domestic demand shocks, and to consider portfolio shocks that affect the exchange rate.10 One important qualification to Poole’s method is that, as argued by B. Friedman (1975), the monetary authorities cannot just ‘set’ the money supply at the desired level—any more than they can ‘set’ any of the other possible target variables at the desired level. In analysing the performance of alternative targets under shocks it is therefore necessary to think in terms of how the authorities must respond in the short run if they are determined to keep to an unchanged target, as opposed to the medium term where they may respond by changing their target. With reference to the time-inconsistency literature, an efficient target should convey clear information to the private sector about the authorities’ intentions, and it should do so in such a way that the private sector can easily check whether the authorities are abiding by their stated intentions; this will depend in part on the availability of good quality statistics. Where the central bank is not independent, that process of verification is essential to the resolution of the time-inconsistency problem and hence to the attainment of low inflation; but where the bank is independent, the emphasis will be on the provision of clear and accurate information to feed into the private sector’s expectations of inflation. While these two strands of analysis reveal some interesting characteristics of the different targets, it is also important not to lose sight of the issue of control: how precisely and over what time horizon can the targeted aggregate be controlled by the authorities? Clearly the best hypothetical stabilisation properties and the most powerful potential effects on expectations are ultimately worthless if the target is not practically feasible. The stabilisation properties of alternative targets can be considered informally11 in terms of an AD/AS model in which the AD curve is derived from the interaction of an IS and 9 See Meade (1978) and Brittan (1982) for early advocacy of nominal income targets, and Bean (1983) for an evaluation which predates the abandonment of monetary targets. 10 Shocks to equity prices will not be treated here, given the low importance of such shocks in UK monetary policymaking over the period. 11 See Artis and Currie (1981) for an example of a more formal comparison between monetary and exchange rate targets.

Analytical Introduction: Alternative Monetary Frameworks 5 an LM curve, the AS curve is vertical in the rational expectations long run but upwardsloping for unanticipated short-run shocks, and the foreign exchange market involves interest parity with a foreign exchange risk premium.12 The pursuit of a monetary target involves keeping the money supply, with its contribution to the LM and AD curves, unchanged in the face of shocks. A nominal income target means that the authorities attempt to keep nominal income unchanged, i.e. they shift the AD curve so that it intersects the short-run AS curve (whether the latter has shifted or not) at points which describe a rectangular hyperbola. An inflation target has to be considered within this framework as a price level target: the authorities shift AD to make it intersect the short-run AS at the original price level. Finally, under an exchange rate target the authorities have to use the interest rate and their foreign currency reserves so as to keep the exchange rate unchanged; in the analysis that follows it is assumed that, under conditions of high capital mobility, the effectiveness of sterilised intervention is limited, so that the emphasis is on interest rate changes rather than foreign currency sales or purchases. For expenditure shocks Poole had already shown (in a closed-economy model) that a monetary target is stabilising in the face of a domestic expenditure shock, that is, keeping the money supply at a constant level reduces the fluctuation in income compared to what would occur if the interest rate were kept constant; in AD/AS terms the AD curve moves less in response to an expenditure shock under a monetary target than if the interest rate is held constant. This stabilisation effect will be even stronger for the other two types of domestic target: if the authorities are targeting nominal income growth they will react to an unexpected rise in, say, investment expenditure by raising the interest rate by more than it would otherwise have risen, in order to dampen the effect on income; and if they are targeting inflation they will take similar action in order to prevent a rise in income which would lead to higher inflation at a later date. In an open economy with a flexible exchange rate there may be a smaller rise in the interest rate accompanied by an appreciation of the exchange rate;13 in this case the sectoral distribution of the impact will be different—with the traded goods sector more, and the non-traded goods sector less, severely affected—but not its overall result. Under an exchange rate target, on the other hand, if the authorities respond to the upward pressure on the exchange rate by preventing the rise in the interest rate, then domestic income will not be damped as under the other targets and will vary more widely than it would have done otherwise. In the case of a money demand shock Poole’s analysis showed that income will vary more under a monetary target than with a constant interest rate, and this will also be valid for an open economy (though the change in the interest rate may be partly replaced by a change in the exchange rate). However, if the authorities are targeting nominal income or inflation they will respond to an increase in the demand for money (which would otherwise raise the interest rate and discourage expenditure) by loosening policy: these targets are therefore stabilising in the face of money demand shocks. An exchange rate target will also be stabilising here, ceteris paribus, provided that the authorities relax policy to prevent appreciation. 12 The determination of the foreign exchange risk premium is not well understood, and the term often serves as an all-encompassing residual. The reason for including a premium here is to make it possible to consider exogenous shocks to the exchange rate; it can be thought of as determined by the country’s net external debt position and by an exogenous shock term. 13 Ceteris paribus, i.e. in particular if the expected long-run equilibrium of the exchange rate and the risk premium are not affected, the incipient rise in the differential between domestic and world interest rates must be covered (for continued interest parity) by a rise in the expectation of depreciation, which in turn requires that the exchange rate should immediately appreciate relative to the long-run equilibrium.

6 The Making of Monetary Policy in the UK, 1975–2000 A foreign expenditure shock, i.e. a change in the demand for exports, will be treated in a similar way to a domestic expenditure shock under monetary, nominal income and inflation targets (it has a similar effect on the IS curve in terms of the Poole analysis). Under an exchange rate target an increase in export demand may appreciate the exchange rate directly14 as well as via a rise in the interest rate; if the authorities find themselves actually lowering the interest rate to maintain the parity income will rise by more than it would have done otherwise. A foreign asset price shock affecting the exchange rate, in the form of an increase in foreign interest rates causing the demand for assets to switch away from the domestic towards the foreign country’s assets, will tend to raise the domestic interest rate and depreciate the exchange rate. The same effects will occur for a portfolio shock affecting the exchange rate, in the form of a rise in the foreign exchange risk premium (as the result, for example, of a loss of confidence in the currency). Under a monetary target there is no reason for policy to respond to either of these shocks, and the new equilibrium will have a higher interest rate and higher income. Under a nominal income target or inflation target policy will tighten to prevent the depreciation and the consequent rise in aggregate demand leading to a rise in prices and income, that is the interest rate will rise by more and the exchange rate depreciate by less than under the monetary target: policy will ‘lean into’ the shock. Under an exchange rate target policy will tend to move further in the same direction: the authorities will have to raise the domestic interest rate to prevent any depreciation against the foreign currency, so that income will fall, and if the shock leads to a depreciation of the expected long-run equilibrium exchange rate this effect may be even larger. A negative supply shock (e.g. short-run upward pressure on the level of wages) shifts the short-run AS curve to the left, with less output being produced at a given price. All the kinds of target considered here are demand-side policies, so none of them can handle supply shocks easily. However, there are some important differences. Under a monetary target the authorities have no reason to respond to a supply shock, so that output will fall and prices rise. Under an inflation target, however, the authorities are committed to maintaining the price level; they must therefore act to reduce aggregate demand, in which case the fall in output will be larger. Under a nominal income target the authorities may have to relax or tighten policy, depending on the slope of the AD curve, but since prices are allowed to rise the fall in output must be less than under an inflation target. Finally, under an exchange rate target there appear to be two effects with opposite signs. On the one hand, a negative supply shock which raises prices will push up the interest rate; if the authorities relax policy to prevent this leading to an appreciation then the fall in output will be partially offset. On the other hand, the rise in prices will adversely affect the current account balance, and this may tend to depreciate the (expected) long-run equilibrium, and hence the current value, of the exchange rate; if the authorities tighten to prevent this, the fall in output will be larger. It is not obvious a priori which of these effects will dominate. For the effects of the different targets on inflation expectations what is important is that the targets should be visible and verifiable: agents should be able to apprehend easily the information the targets contain, and to check easily whether the authorities are adhering to their targets. Monetary targets are easily enough understood by well-informed agents, but it may be doubted whether other economic agents can so easily grasp the implications for inflation of a particular announced target range or number. Targets for the monetary base and 14 If it causes an appreciation of the expected long-run equilibrium and/or a reduction in the risk premium (as the result of an improvement in the external debt position of the domestic economy).

Analytical Introduction: Alternative Monetary Frameworks 7 other narrower measures are relatively verifiable, in the sense that the monetary statistics are published at high (monthly or even weekly) frequency, without much delay and without much revision; but the statistics for the broader aggregates are not available on a weekly basis and are subject to more delay and more revision.15 Nominal income targets are more easily interpretable (agents need to understand something of the split between income and prices, but do not need to think about velocity). However, they are much less verifiable, because the data are published only with substantial delays and at low frequency (i.e. quarterly rather than monthly), and are liable to be revised by substantial amounts at later dates. Inflation targets obviously convey the clearest information about the government’s inflation objectives, and in that sense should affect expectations most strongly. Inflation statistics are published with moderate (monthly) frequency but are not susceptible to significant later revision, so that inflation targets have good verifiability as well as visibility. Exchange rate targets are probably more easily interpreted than monetary but less than inflation targets: agents need to know something about the inflation rate in the anchor country to interpret the meaning of an exchange rate target for domestic inflation. On the other hand they are highly verifiable, with good, non-revised data being published on a daily, or higher, frequency basis. The various possible targets also differ in the authorities’ ability to control them precisely and rapidly. For monetary targets it is again necessary to distinguish between different measures. While for some economists the monetary base is, or could be, under appropriate conditions, susceptible to precise and short-term control, most central bankers have taken the opposite view that the base (which for the UK consists largely of notes and coin in circulation with the public) could be directly controlled only at the risk of creating unacceptable tensions in the financial system (Goodhart, 1994). But in any case it is clear that over the period with which this book is concerned the conditions required for monetary base control to work were not in existence. Other narrow measures of money such as M1 (basically current account deposits) are open to control only indirectly, via effects on the arguments in the demand functions (nominal income and interest rates). Measures of broad money such as £M3 or M4 (basically current and time deposits in banks or in banks and building societies), on the other hand, are in principle open to control from the supply side if the credit counterparts framework is accepted (see Sections 2.3 and 2.4 of Chapter 2). But even so the control is far from precise and there are significant lags, of the order of 12 months or more, between policy changes and changes in the main private sector counterpart, bank lending to the private sector. Neither nominal income nor inflation are susceptible to precise or immediate control by the authorities, but the lags and the imprecision are greater for inflation; in broad terms the lag between policy changes and nominal income is probably around 12 months, and that between policy changes and inflation around 18–24 months. In all these cases the lags imply that targeting is not going to be successful unless the authorities can forecast accurately both the way the targeted variable is likely to evolve on unchanged policies and the way it will respond to policy changes, so that they can implement required changes sufficiently far in advance. The exchange rate, on the other hand, poses a different set of problems. In the very short run the exchange rate may be controllable through foreign exchange market intervention, but over a longer period that is more difficult. Policy needs to be directed towards ‘steering’ the foreign exchange market towards the desired exchange rate (partly by confidence-promoting measures designed to influence the expected long-run equilibrium) and setting the interest rate 15 References

to publication lags in this and the next two paragraphs concern the statistics for the UK.

8 The Making of Monetary Policy in the UK, 1975–2000 Table 1.1 Alternative monetary policy targets Targets

Short-run stabilisation properties for shocks to: domestic expenditure money demand foreign expenditure foreign asset prices or foreign exchange risk premium supply Influence on inflation expectations: interpretability verifiability Control: precision lags

Monetary

Nominal income

Inflation

Exchange rate

good poor good poor

good good good good

good good good good

poor good poor very poor

poor

poor

very poor

??

poor good to less good*

less poor poor

very good good

good very good

poor

poorer

very poor

poor

poorer

very poor

good to very poor** good

Notes: *good for narrower, less good for broader aggregates; **good in tranquil periods, very poor in crisis periods.

at the level required to produce a balance in the foreign exchange market at the target parity, and such actions may not be straightforward. Moreover, the exchange rate may be stable for a considerable period but then move sharply and unpredictably in a direction not intended or desired by the authorities. As the literature on balance of payments crises has made clear, crises can affect currencies whose fundamentals are ‘sound’ as well as those whose fundamentals are unsound.16 The above discussion of the various possible monetary policy targets is summarised in Table 1.1, which gives rough rankings under the various headings considered. As can be seen there, monetary targets perform well for domestic and foreign expenditure shocks, but poorly for domestic and foreign asset market shocks and supply shocks; they score poorly on interpretability but better on verifiability; and the monetary aggregates are difficult to control. Nominal income targets perform well vis-`a-vis all of the demand shocks, but not supply shocks; they score better than monetary targets on interpretability but much worse on verifiability; and nominal income is even more difficult than money to control. Inflation targets do well on demand shocks but very badly on supply shocks; they score well on interpretability and verifiability; but inflation is the most difficult of the purely domestic targets to control. Finally, exchange rate targets cushion the effect of money demand shocks and possibly supply shocks, but they perform badly for domestic and foreign expenditure and foreign exchange risk premium shocks; they are easily verifiable but less interpretable; and control of the exchange rate varies widely as between tranquil and crisis periods. 16 See

the review of this literature in Chapter 5 below.

Analytical Introduction: Alternative Monetary Frameworks 9

1.3 MONETARY POLICY RULES Monetary policy rules are not intermediate targeting frameworks but lower-level rules or ‘instrument rules’ (Svensson, 1999), which specify how the instrument(s) of monetary policy should be adjusted in response to developments in the economy. One such rule, which will be covered here only briefly, is that put forward by McCallum (1988), under which the growth of the monetary base (M0) is adjusted so as to deliver some target level of nominal income: m˙ = k * − v˙ + λ(x * − x) where m˙ is the quarterly growth of MO, k ∗ is a constant, v˙ is the trend change in velocity, x and x ∗ are the actual and target levels of (log) nominal GDP, and λ is the weight on the deviation of actual nominal GDP from target. This is comparable to Friedman’s monetary rule defined for the monetary base, except that it includes an adjustment for velocity changes and a larger counter-cyclical effect in the final term. The basic problem with this rule, and the reason it is not discussed further here, is that it requires control of the monetary base. This, as already stated above, has never been undertaken in the UK (and rarely elsewhere), and arguably would be a risky approach to adopt (Goodhart, 1994). A more important monetary policy rule is that put forward by Taylor (1993), under which the policy interest rate is adjusted in response to deviations of inflation from target and of output from trend: i = π + w1 (y − y * ) + w2 (π − π * ) + r * where i is the nominal interest rate, π and π ∗ are actual and desired inflation, y and y ∗ are (logs of) actual and trend output so that y − y ∗ is the output gap, r ∗ is the equilibrium real interest rate, and w1 and w2 are weights on the deviations of inflation and output from target/trend. In fact Taylor’s original rule for the US was π −2 y − y* + +2 2 2 where π ∗ and r ∗ were each assumed to be 2, and the weights on inflation and output were each 0.5. The Taylor rule has had a marked influence on the monetary policy literature of the last decade. Taylor originally presented it both as a stylisation of the behaviour which the Federal Reserve had apparently been following for several years and as a proposal for how monetary policy should be operated. It is not intended to be applied in a mechanical way, instead it should be seen as a useful benchmark against which alternative policy actions can be assessed. At the same time it has some of the same advantages as intermediate targets, as a form of pre-commitment and as a protection against political pressures (Taylor, 1998). Its properties have been explored in a variety of analytical and empirical settings, and rules of this form, though not necessarily with Taylor’s original weights, have been found to be optimal under certain conditions.17 Taylor rules have also been used widely in empirical work: economists such as Clarida, Gal´ı and Gertler (1998) have estimated reaction functions in the form of Taylor-type rules, as a way of characterising the behaviour of the monetary authorities in different countries over different periods. Particular attention has been attached in this work to the response of the interest rate i =π+

17 See

for example Ball (1999a), Clarida et al. (1999), Svensson (1997) and papers in Taylor (1999a).

10 The Making of Monetary Policy in the UK, 1975–2000 to inflation: it has been shown analytically that for inflation to be stable in the long run the coefficient w2 needs to be positive, so that policy reacts to a rise in inflation by raising the real rate of interest.18 Some attention has also been paid to the possible role of the exchange rate within a Taylorrule framework in an open economy context. Ball (1999b) extended his backward-looking (1999a) model to the open economy, and presented an optimal rule in the form of an equation for a ‘monetary conditions index’—a weighted average of the interest rate and the exchange rate—as a function of inflation, the output gap and the lagged exchange rate. Svensson (2000) discussed open economy issues in a forward-looking model, and his optimal rules are for the interest rate as a function of the real exchange rate, the foreign interest rate and the foreign exchange risk premium, as well as inflation and the output gap. Empirical estimates of reaction functions for the UK can be found in Clarida, Gal´ı and Gertler (1998) and Nelson (2000), both of which have an open economy dimension in at least some periods, while a more thorough open economy treatment can be found in Adam, Cobham and Girardin (2001).

1.4 CONCLUSION Three of the four kinds of monetary policy targets discussed in Section 1.2 have been implemented in the UK, the exception being nominal income targets, while the inflation targeting regimes of the 1990s can also be presented as the application of a form of Taylor rule. The analysis of Sections 1.1 and 1.2, summarised in Table 1.1, shows that no single monetary policy target dominates on all the relevant criteria. But the analysis usefully draws attention to the kinds of shocks, the problems in communication and the difficulties of control which are most important for each kind of target. The next chapter gives a historical outline of the macroeconomic and monetary policies pursued between 1975 and 2000, and distinguishes between different monetary regimes. Later chapters then provide assessments of the operation and performance of the different frameworks in different subperiods (drawing inter alia on the Taylor rule estimations of Adam, Cobham and Girardin, 2001). Many of the factors emphasised in this analytical introduction turn out to have been important at different times, though they cannot fully account for the successes and failures of the various regimes.

18 For

example Clarida et al. (1999), Taylor (1999b).

2 Historical Introduction This chapter starts with a basic historical narrative for macroeconomic and monetary policy and the macroeconomic development of the economy over the period, paying more attention to episodes not covered in detail elsewhere and less to those that are. Section 2.2 examines the evolution of monetary policy in terms of the different frameworks or regimes in operation at different times, thereby identifying the phases and subperiods which will be used elsewhere in the book. Section 2.3 gives an account of the way in which the determination of the money supply was presented and analysed during the first decade or so of the period, both by the monetary authorities1 and by most other commentators, as essential background to the discussion of broad money targeting in the next chapter. Section 2.4 discusses the instruments used by the monetary authorities over the period.

2.1 A MACROECONOMIC NARRATIVE, 1975–2000 The immediate background to the state of the economy in 1975 was the heady combination of the Heath–Barber boom of 1972–73 and the first oil price shock of late 1973. Monetary and fiscal policy had been highly expansionary in 1972–73 as the government—partly influenced by the rise of unemployment, whose headline rate went above one million for the first time in 1971—embarked on a ‘dash-for-growth’ strategy according to which a sustained rise in aggregate demand was expected to produce a marked rise in investment that would shift the economy on to a higher growth path. GDP did indeed grow rapidly, at rates of 3.6% in 1972 and 7.3% in 1973, but there was little sign of a major change in investment behaviour. The growth of broad money (which owed its momentum to the 1971 liberalisation measures of Competition and Credit Control as well as to the macro stimulus) was around 25% in each of 1972 and 1973. But at the same time the current account deteriorated from the surpluses of 1970 and 1971 and there was a balance of payments crisis in June 1972, to which the government responded by allowing sterling to float. By 1973 GDP was slowing, the economy appeared to have hit a capacity constraint, inflation was rising despite the use of incomes policy, and sterling was continuing to depreciate. The government finally changed course late in 1973, cutting back on public expenditure, raising interest rates and introducing the ‘corset’ (Supplementary Special Deposit Scheme), an indirect form of control on banks’ lending. But by that time the first oil price shock was beginning to take effect, and the government was involved in a major industrial dispute with the coal miners. The Labour government that came to power in February 1974 without an overall majority in Parliament responded by trying to resolve the most immediate difficulties (such as the miners’ 1 For most of the period covered by this book the term ‘the monetary authorities’ refers to the Bank of England and the Treasury together, but from June 1997, when the Bank acquired instrument independence, the term refers to the Bank’s Monetary Policy Committee.

12 The Making of Monetary Policy in the UK, 1975–2000 Table 2.1 Macroeconomic outcomes Year

GDP growth (%)

Inflation (RPI, %)

Inflation (RPIX, %)

Unemployment (%)

Current account (% GDP)

Exchange rate (1995 = 100)

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

−1.7 −0.7 2.8 2.4 3.4 2.8 −2.2 −0.1 1.8 3.7 2.4 3.8 4.2 4.4 5.2 2.1 0.6 −1.5 0.1 2.3 4.4 2.8 2.6 3.5 2.6 2.3 3.0

16.0 24.1 16.8 15.9 8.3 13.4 18.1 11.9 8.7 4.6 5.0 6.1 3.4 4.1 4.9 7.8 9.5 5.9 3.7 1.6 2.5 3.4 2.4 3.1 3.4 1.6 2.9

24.6 16.9 15.8 8.7 12.6 17.0 12.2 8.6 5.2 4.4 5.2 3.6 3.7 4.6 5.9 8.1 6.8 4.7 3.0 2.4 2.8 2.9 2.8 2.7 2.3 2.1

2.9 4.3 5.6 6.0 †5.7 4.7 6.2 9.7 †10.3 11.1 11.2 11.5 11.6 10.6 8.7 7.3 7.1 8.9 10.0 10.5 9.6 8.7 8.2 7.0 6.3 6.1 5.5

−3.8 −1.4 −0.8 −0.1 0.7 −0.2 1.2 2.5 1.7 † 0.6 −0.1 0.1 −0.6 −1.4 −3.8 −4.6 −3.5 −1.5 −1.7 −1.7 −0.2 −0.5 −0.1 0.8 0 −1.1 −1.7

167.6 153.0 132.5 125.8 126.4 133.3 146.7 150.8 145.3 136.3 131.3 131.2 119.6 117.1 124.2 120.6 117.8 118.8 114.3 104.8 105.1 100.0 101.7 118.5 122.5 122.3 126.7

Notes and sources: † indicates break in series. GDP growth: Economic Trends Annual Supplement, various issues. RPI inflation: average of 12-month growth rates, Economic Trends Annual Supplement, various issues. RPIX inflation: average of 12-month growth rates, Bank of England. Unemployment: standardised rate, % of total (1974 –81) or civilian (1982–2000) labour force, OECD Economic Outlook, various issues. Current account: OECD Economic Outlook, various issues. Exchange rate: nominal effective, IFS.

dispute) but adopting relatively neutral macro policies in the first instance.2 It continued the incomes policies of its predecessor but tried, with little immediate success, to contain the fiscal deficit and kept monetary growth relatively low in 1974 and 1975 after the excesses of the previous two years.3 Meanwhile inflation continued to rise, reaching a peak in August 1975 (of 26.9% for the 12-month growth of the RPI), while GDP fell sharply in 1974 and again in 1975 (see Table 2.1 for data on macro outcomes, and Table 2.2 for data on macro policy indicators). 2 On the Labour government’s macro policy see Artis and Cobham (1991: chapters 1 and 16) and Allsopp (1991). 3 It should be noted, however, that if Artis and Lewis’s (1976) conjecture that the 1972–73 monetary expansion had

pushed the private sector off its money demand curve is correct, monetary growth of 10% per year, even at the current inflation rates, was not tight.

Historical Introduction 13 Table 2.2 Macroeconomic policy indicators Year

M0 growth (%)

£M3 growth (%)

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

11.8 14.2 11.5 11.0 15.0 13.1 8.5 5.6 2.5 5.9 5.5 4.7 4.0 4.7 6.8 5.9 5.3 2.4 2.4 4.9 6.4 5.9 6.7 6.2 6.1 7.2 8.1

16.3 8.3 6.6 7.6 15.8 12.8 15.6 16.7 13.0 12.7 9.3 12.5 19.1 20.6 21.1

M4 growth (%)

Short-term interest rates

Long-term interest rates

Fiscal deficit

13.2 13.9 16.9 15.4 17.1 17.5 15.9 7.9 5.1 3.5 5.1 7.3 9.9 11.2 9.8 5.4 6.6

11.4 10.2 11.1 7.7 8.5 13.0 15.1 13.0 11.4 9.6 9.3 11.6 10.3 9.3 9.9 13.3 14.1 10.9 8.9 5.3 5.2 6.3 5.8 6.5 6.8 5.0 5.8

14.8 14.4 14.4 12.7 12.5 13.0 13.8 14.7 12.9 10.8 10.4 10.5 9.9 9.5 9.4 9.6 11.1 9.9 9.1 7.9 8.1 8.3 8.1 7.1 5.5 4.7 4.7

−3.8 −4.5 −4.9 −3.2 †−4.4 −3.3 −3.4 −2.6 −2.5 −3.3 −4.0 −2.9 −2.6 −1.9 0.6 0.9 −1.5 −2.8 −6.5 −8.0 −6.8 −5.8 −4.4 −2.0 0.4 1.3 1.9

Notes and sources: †indicates break in series. MO: average of 12-month growth rates, Bank of England website, plus Financial Statistics for 2000. £M3: average of 12-month growth rates, Economic Trends Annual Supplement, 1989, plus Financial Statistics, June 1989. M4: average of 12-month growth rates, Bank of England website, plus Financial Statistics for 2000. Short-term interest rate: Treasury bill rate, IFS. Long-term interest rate: long-term government bond yield, IFS. Fiscal deficit: general government financial balance as % of GDP, OECD Economic Outlook, various issues.

Government policies began to tighten in 1975, with a more restrictive incomes policy agreed with the trade union movement and the move towards ‘cash limits’ on public expenditure (introduced in the 1976 budget). However, sterling, which had depreciated at intervals since June 1972, fell sharply in March 1976 in the beginning of a long-drawn-out crisis which was ended only in December 1976 after some sharp rises in interest rates and the government’s agreement with the IMF on a Letter of Intent which included limits on domestic credit expansion (DCE)4 and on public expenditure.5 4 DCE can be thought of as the sum of the domestic sources of monetary growth. 5 In fact the spending cuts agreed with the IMF were dwarfed by the cuts that occurred

to the introduction of cash limits. See Jackson (1991: 83–5).

during 1976/77 in response

14 The Making of Monetary Policy in the UK, 1975–2000 In 1977 the pressures that had pushed sterling down in 1976 operated in reverse, and the authorities rebuilt the foreign exchange reserves and lowered interest rates in an effort to preserve the competitive advantage gained from the depreciations of the earlier years. Income was growing at around the trend rate, and inflation (which had been pushed up again by the depreciation of 1976) was at least well below the 1975 level. The DCE target agreed with the IMF rapidly became redundant in the sense that it was obviously going to be easily fulfilled, and attention switched to the more formal target for broad money6 promulgated in the 1977 Budget. The increasing capital inflows put the attainment of this target at risk, and in October the authorities finally chose the monetary target over the implicit exchange rate target; however, the appreciation was only temporary. In 1978 macro policy was relaxed a little, but the government was unable to get trade union agreement to the 5% limit on wage rises which it favoured. The tensions involved led to the ‘Winter of Discontent’ with a series of labour disputes, particularly in the public sector. In May 1979 a new government under Margaret Thatcher came to power, committed to radical changes in many areas of policy. On the macro front the new government was more concerned to control inflation and less concerned about unemployment; it put more emphasis on monetary targets and less on fiscal policy; and one of its first actions was to change the balance between direct and indirect taxation, with cuts in income taxes and a large rise in VAT. Monetary policy was tightened, at a time when inflation was already rising and bound to rise further in the short run under the impact of the VAT rise, and shortly afterwards the second oil price shock began to affect the economy (the UK being by now a significant oil producer and roughly self-sufficient in oil). The result of these two forces was a marked appreciation of sterling, which (together with high interest rates) created a severe financial squeeze on UK industry.7 GDP fell in 1980 and again in 1981, but inflation came down sharply from 1980. Unemployment rose strongly from mid-1980. Policy responded to the evidence of severe pressure with cuts in interest rates partly facilitated by a tightening of fiscal policy in the 1981 Budget. Subsequently important modifications were introduced to the government’s Medium Term Financial Strategy, and broad money targeting was formally suspended in 1985 and abandoned in the March 1987 Budget. From 1981 sterling gradually ceased to be so obviously overvalued, but there were recurring sharp declines in the exchange rate which led the authorities to raise interest rates sharply (but temporarily); these included the crisis of January 1985 when the pound nearly went to parity with the (strong) dollar. Growth revived, rising sharply in 1983 but then falling back in 1984 to around the trend rate; there was a temporary boost to output from the ending of the miners’ strike in 1985 Q1, but growth was not fast enough to stop unemployment continuing to rise until 1986. Late 1985 and 1986 saw the ‘reverse oil price shock’ which was used by other countries largely to reduce their inflation rates further. However, the UK authorities chose instead to use the shock to allow a substantial stimulus to the (still rather sluggish) domestic economy via the associated fall in sterling in its capacity as a ‘petro-currency’, but without immediate inflationary implications. The depreciation was almost certainly permitted to go too far towards the end of 1986, and most of it was then locked in by the policy of shadowing the Deutschemark (DM) on which Lawson embarked in March 1987 in a quest for greater exchange rate stability and perhaps eventual entry into the ERM. Together with the wide range of measures of financial liberalisation that had been implemented since 1979, this depreciation provided the context for 6 This was now defined as sterling M3, the sterling content of the previously 7 See Chapter 8 for a measure of the overvaluation of sterling at this time.

targeted aggregate, M3.

Historical Introduction 15 the boom of 1986–88 (see Chapter 4). The DM-shadowing policy was abandoned a year later amidst economic and political tensions, but in the meantime it had played a role in initiating the ‘Lawson boom’ of 1987–89 which saw UK inflation returning to double figures in 1990. The response to the overheating that developed was delayed, partly because of the commitment to shadowing the DM and partly because policy-makers underestimated the growth of demand and overestimated the responsiveness of supply. They then found themselves having to rely almost entirely on rises in interest rates; fiscal policy was in effect unavailable since at that stage there was already a large budget surplus. Interest rates were increased by 5.5% between May and November 1988, and by another 2% by October 1989, in order to bring the boom and the associated rise in inflation under control. By mid-1990 the rise in economic activity had come to a halt, inflation was rising more slowly (it peaked at 10.9% in September) and unemployment had ceased to fall (it reached a trough of 5.5% in the first half of 19908 ). Meanwhile, the UK government finally agreed on membership of the ERM. The UK entered on 8 October 1990 after a tortuous domestic debate, at an exchange rate which some thought overvalued, and at a time when the ERM was coming under increasing strain from the effects of German reunification in 1990 (the German business cycle had been disconnected from the US cycle, and the loosening of fiscal policy by the German government had led the Bundesbank to raise interest rates to bring inflation under control). The UK government was unwilling to devalue within the ERM unless other countries did so too, but may itself have contributed to the growing speculative pressures on the pound which led to the UK’s suspension from the ERM in September 1992 (see Chapter 5). In the aftermath of the exit from the ERM the UK authorities allowed a ‘rebalancing’ of the economy, with sharp reductions in interest rates (as well as a significant depreciation). They also introduced a new framework for policy consisting of inflation targeting and a revised relationship between Treasury and central bank. GDP growth revived, without a large passthrough leading to a high rate of inflation, but the new framework was an imperfect constraint on political control of interest rates in the run-up to the next election (see Chapter 6). In May 1997 one of the new Labour government’s first measures was to make the Bank of England responsible for setting interest rates so as to attain an inflation target set by the government. Inflation was brought under control, and growth continued around the trend rate with a dip in 1999 (related to the deterioration of the world environment) but a resurgence in 2000 (see Chapter 7).

2.2 MONETARY FRAMEWORKS AND PHASES In the UK, as in many other countries, monetary policy became more important during the course of the 1970s, after the Bretton Woods system had broken down and inflation had become a more pressing issue for economic policymakers. The UK authorities apparently had internal targets for broad money growth from 1974: according to the Governor’s 1978 Mais Lecture, ‘Before [1976] [controlling the growth of the monetary aggregates] constituted an internal aim: I think it is not therefore entirely accidental that during each of the three years 1974–76 the growth of sterling M3 was about 10%, well below the rate of expansion of national income in current prices’ (BEQB, 1978a: 33). The first formal announcement of a monetary target 8 On the claimant count rate, as used in Economic Trends Annual Supplement, 2000 edition; the previous peak had been 11.0 in 1986 Q3, and the next peak was 10.4 in 1993 Q1.

16 The Making of Monetary Policy in the UK, 1975–2000 (for the broad aggregate M3) came from the Chancellor in July 1976 in an attempt to calm the foreign exchange markets during the prolonged sterling crisis of that year.9 More formal target ranges were set for 1977/8 and 1978/9; they coexisted with limits on DCE agreed with the IMF in December 1976, but those limits rapidly became non-binding and attention was focused on the targets for broad money.10 Monetary targets were given more importance by the government of Margaret Thatcher which came to power in May 1979, and the Medium Term Financial Strategy (MTFS) whose first version was announced in the March 1980 Budget set out targets for £M3 for four consecutive financial years (three in later versions), accompanied by projections for the Public Sector Borrowing Requirement (PSBR) as a percentage of GDP. This greater emphasis on monetary policy was accompanied by a shift of the balance of power over monetary policy from the Bank of England to the Treasury (Kynaston, 1995: 31–2, 37). While much academic argument has seen the introduction of monetary targets as a move towards monetary rules of the sort advocated by Friedman (1960) or Kydland and Prescott (1977), as discussed in Section 1.1 of Chapter 1 above, Lane (1985) has pointed out that few central bankers (or ministers of finance) thought in these terms. They did not see monetary targets as rules, or even as incipient rules, and probably did not see the need for constitutional constraints on their actions; their acceptance of the concept and implications of rational expectations was far from wholehearted; and in adopting monetary targets they did not ‘consider themselves bound to a course of action which they would not otherwise choose to follow’.11 Instead, there seem to have been three main arguments for monetary targets. First, announcing monetary targets was a way of placing a stronger emphasis on price stability, as opposed to output and employment, as the primary objective of monetary policy. Second, announced targets should give clearer information to the private sector on the monetary authorities’ intentions, and therefore help the private sector to make more accurate and less uncertain economic forecasts, which in turn would make the economy more stable. And third, announced targets might have an independent influence on expectations of inflation, causing them to fall more quickly than they would otherwise have done and thereby reducing the unemployment cost of lowering inflation. The UK monetary authorities appear to have shared these ideas, within an overall emphasis on the role of targets in providing a guarantee of stability comparable to that offered in earlier times by ‘fixed exchange rates or Gladstonian budgetary principles’.12 As the Bank itself put it in 1984, The Bank has consistently felt the need for a financial anchor, whether it be a fixed exchange rate or a monetary target, which needs to be both publicly announced and vigorously defended, in order to give a confident basis for private sector decision-makers to plan forward. But it has never, on the other hand, felt that economic relationships were sufficiently predictable, or the financial system so static, that the conduct of policy could be safely placed on a quasi-automatic basis with the adoption of constant rules. (Bank of England, 1984: 4)

On this basis three separate monetary policy regimes can be identified for the first part of the period. Between January 1975 and March 1977 there were monetary targets of some 9 The Chancellor said that the growth of M3 during 1976/7 ‘should amount to about 12%’ (BEQB, 1976c: 296). In October the Governor referred to this figure as a ‘target’ (1976d: 454). 10 See Table 3A.1 in the appendix to Chapter 3 for details of DCE targets and outturns. 11 Lane (1985: 200). See also Sumner (1980), Foot (1981) and Davis and Meek (1983) on the adoption of monetary targets in developed countries in the 1970s. 12 BEQB, 1978a: 34. See also Fforde (1983), and BEQB, 1977a: 48–9; 1977d: 461.

Historical Introduction 17 kind (for M3), but they were initially internal to the Bank rather than published and then when they were published their status was unclear; ‘informal monetary targets’ seems to be the best characterisation. Between April 1977 and May 1979 the monetary targets (for £M3) were more formal, and announced, but there is no doubt that the authorities felt they retained substantial discretion in their implementation. From June 1979 under the Thatcher government the monetary targets were meant to be a more serious commitment, and were seen as central to the whole thrust of policy; the role of discretion was intended to be smaller. Moreover, in the form of the Medium Term Financial Strategy (MTFS) (the first version of which was promulgated in the 1980 Budget), the monetary targets were specified for four (later three) years rather than just the current year, and they were accompanied by ‘projections’ for the Public Sector Borrowing Requirement (PSBR). However, as already indicated, the first two years of the MTFS saw large overshoots of the £M3 targets, which were permitted by the authorities largely on the grounds that the deflationary pressure on the economy was already very severe.13 The obvious explanation of the combination of deflation and monetary overshoot was that velocity was falling, and a downward trend in (broad money) velocity was to become firmly established and recognised as the decade went on. In response to these early overshoots the target ranges were adjusted and targets for other (narrower and broader) aggregates were added, but overshoots recurred and monetary targeting continued to appear unsatisfactory. The turning point can be located in the Budget of March 1981: efforts were made to reduce the fiscal deficit (despite the severity of the recession) with the intention that this would allow interest rates to be reduced more quickly (and that might in turn help to reduce the overvaluation of sterling); moreover the government allowed a major element of ‘base drift’ in the sense that the 1981/2 monetary target first announced in the 1980 MTFS was now to apply from the level of £M3 actually reached in February 1981 (rather than the level that would have been reached if there had been no overshoot during 1980/81), and the FSBR 1981–82 referred to other indicators of monetary tightness as well as £M3. More base drift was allowed in the 1982 Budget, following a further overshoot, but in addition the target ranges were raised by 3%14 and applied to M1 and PSL2,15 and the FSBR declared that the interpretation of monetary conditions would ‘continue to take account of all the available evidence, including the behaviour of the exchange rate’.16 The 1983 Budget did not include any major innovations, but the FSBR 1984– 85 set out target ranges for £M3 and M0 rather than M1 or PSL2. From April 1981, therefore, the monetary regime involved a significantly higher element of discretion. There is little doubt that monetary targeting was made more difficult by the movement in velocity, itself due mainly to financial innovation and liberalisation which blurred the distinction between bank and building society deposits, raised doubts about the use of the £M3 aggregate (which included only bank deposits), and resulted in a strong growth of credit. Indeed, this was the explanation stressed by the monetary authorities themselves. However, it can also be argued (see Chapter 3) that the repeated overshoots of the targets were due partly to the way 13 Part of the 1980/81 overshoot was due to the unwinding of the disintermediation which had occurred under the corset before it was abolished in June 1980. The PSBR projection was also overshot by a large margin in 1980/81 (as was the forecast for 1979/80), but less so in subsequent years. 14 From the 5–9% set out in the 1980 and 1981 MTFSs for 1982/3 to 8–12%, and similarly from 4–8% to 7–11% for 1983/4. 15 ‘Private sector liquidity 2’ was a measure of money which included building society as well as bank deposits and some other liquid assets, and was later renamed M5. See Chapter 3, Section 3.1. 16 FSBR 1982–83: 15.

18 The Making of Monetary Policy in the UK, 1975–2000 in which the authorities chose the target ranges, which emphasised short-term announcement effects and produced ranges which were not always consistent either with the authorities’ expectations of nominal income growth and velocity or with their expectations of the growth of the credit counterparts which were central to official thinking about monetary policy in this period. Whatever their cause, the repeated overshoots meant that the monetary targets were unable to fulfil much of their intended role: their ability to provide information to the private sector and to affect inflation expectations could not be sustained if the outturn for monetary growth bore no consistent relation to the announced intentions of the authorities. In addition, the authorities’ ability to control broad money from the supply side, particularly via the technique of ‘overfunding’, had run into difficulties (discussed in Section 2.4 below). There is evidence (see Chapter 3) to suggest that, even at the time of the 1985 Budget, the authorities may not have believed that the £M3 target which they announced was feasible. In any case the target was soon being overshot, and in October 1985 the Chancellor suspended the target and declared that overfunding would no longer be undertaken; these decisions meant that monetary policy had lost both its overall framework and one of its key instruments. A new target for £M3 was announced in the 1986 Budget but it too was overshot, and broad money targets were formally abandoned in the 1987 Budget. Targets for M0, a definition of the monetary base, remained but it was never clear how much the authorities actually treated M0 as a target (rather than an information variable). Thus from April 1985 the monetary regime can be characterised as one in which there were no binding targets. The monetary authorities were no doubt well aware of the lack of a framework for monetary policy at this time. Their main response was to press for UK entry into the ERM: it was argued that ERM membership would offer a more credible framework for policy than that provided by monetary targets which were frequently missed, together with an alternative nominal discipline to keep inflation under control. However, the attempt to move in that direction made by Chancellor Lawson, with the support of the Governor of the Bank of England and a range of other ministers and officials, was vetoed by Prime Minister Thatcher in November 1985 (shortly after the £M3 target had been formally suspended).17 However, the development of international policy coordination through the September 1985 Plaza meeting and the February 1987 Louvre accord seems to have encouraged Lawson to join the ERM informally, as it were, with a view to trying to demonstrate to Thatcher the merits of a peg to the Deutschemark, and he began to shadow the DM in March 1987. That policy lasted, through frequent interest rate changes and much foreign exchange market intervention, only until March 1988, at which time the recurring upward pressure on the exchange rate, together with growing tensions between Prime Minister and Chancellor, led to the decision to allow sterling to float upwards. For a month or two the authorities may have hoped to restabilise sterling at a higher level, but this too was overtaken around May by the pressing need to address the rise in inflation. Thus the DM-shadowing of March 1987 to March 1988 was succeeded by another period in which there were no explicit targets but policy was concentrated on fighting inflation without, preferably, destabilising the exchange rate. The UK finally entered the ERM under another Chancellor, John Major, in October 1990. In principle it now had an exchange rate target, but there is some doubt as to how far this was really understood by all the relevant policymakers, and it is possible to argue that the 17 On the politics of entry at this time and later see Keegan (1989: 181–2), Smith (1992: 56–62, 164–73), Lawson (1992: chapters 33, 39 and 40), Thatcher (1993: 693–8, 719–24) and Stephens (1996: 39–52).

Historical Introduction 19 Table 2.3 Monetary policy regimes Phase in months

Monetary policy framework

January 1975 to March 1977 April 1977 to May 1979 June 1979 to March 1981 April 1981 to March 1985 April 1985 to February 1987 March 1987 to March 1988 April 1988 to September 1990 October 1990 to September 1992 October 1992 to April 1997 May 1997–(December 2000)

informal monetary targets monetary targets (with discretion) monetary targets (MTFS) with less discretion monetary targets (MTFS) with more discretion no binding targets DM-shadowing no explicit target fixed exchange rate (ERM) inflation target, Bank advising Chancellor inflation target, decisions by MPC

authorities were themselves partly responsible for the financial markets’ loss of confidence in the sterling parity (see Chapter 5). Sterling was ‘temporarily suspended’ from the ERM in September 1992, in the context of a major speculative crisis. The failure of the ERM as a framework for monetary policy, for whatever reason, and the loss of confidence associated with its demise, made it imperative for the authorities to construct some alternative framework. Within a few weeks of Black Wednesday the authorities began this process, announcing the introduction of an inflation target and a new relationship between the Bank of England and the Treasury, under which the former would publish a regular report on and forecast of inflation and would make formal recommendations on interest rates which the latter (that is, the Chancellor, then Norman Lamont) could accept or reject. In April 1994 it was decided (by the new Chancellor, Kenneth Clarke) that the minutes of the monthly monetary meetings between himself and the Governor of the Bank would be published (with a delay of around six weeks). There seems little doubt that this new regime offered much improved accountability and transparency in monetary policy, but it did not prevent political factors entering into decision-making (see Chapter 6). The widespread view that Chancellor Clarke had declined to raise interest rates when it would have been appropriate in the period before the election of May 1997 can be assumed to have contributed to the decision by the incoming Labour government to hand decision-making on interest rates to a new Monetary Policy Committee at the Bank of England. The Bank, or rather the MPC, continued to publish an inflation report and inflation forecast, and the minutes of the MPC meetings were published;18 these minutes are considerably more informative than those of the Monthly Monetary Meetings between 1994 and May 1997. This, the final period covered in this book, can be characterised as inflation targeting under instrument independence. Table 2.3 gives the dates and characteristics of the various monetary regimes which have operated in the UK over the period. The ten phases include four in which there were some sort of monetary targets, though in three, if not all, of these there was much more discretion in the way they were implemented than is envisaged in the academic literature on targets discussed in Chapter 1. There were two phases in which there was some sort of exchange rate target; in the first of these policy was heavily geared to maintaining the target but in the second 18 Initially the minutes were published two weeks after the following meeting, as in the previous phase; but from October 1998 they have been published with a delay of only two weeks.

20 The Making of Monetary Policy in the UK, 1975–2000 phase, although there was a public commitment to a specific parity, policy was probably more concerned with domestic variables (see Chapters 5, 8 and 9). There have also been two phases of inflation targeting, one without and one with instrument independence. In no case was a satisfactorily functioning monetary regime cast aside, though it is possible to argue that if the authorities had operated one or other regime differently it would have been more successful. The fact that the authorities finally granted instrument independence to the Bank of England may suggest that the intermediate target solution to the problem of time-inconsistency is inadequate, but again that may have been due to the way the intermediate targets were pursued. In any case, it should not be assumed that this last regime is entirely satisfactory, for the exchange rate has been a continuing and serious problem (see Chapter 9).

2.3 ANALYSING THE DETERMINATION OF THE MONEY SUPPLY IN THE UK19 In macroeconomic and monetary economics texts used in the UK in the 1960s, the money multiplier analysis, in which the money supply is determined by the ratios of cash to deposits and bank reserves to deposits, was the normal way of analysing the determination of the money supply.20 That approach had also been applied with considerable success to US monetary history by Friedman and Schwartz (1963) and Cagan (1965). Tobin’s (1963) pioneering criticism of the multiplier analysis in the US argued that banks are much more like non-bank financial intermediaries than commonly thought, insofar as they can expand their assets only if the public is willing to hold an increased amount of bank deposits. Thus ‘there is at any moment a natural economic limit to the scale of the commercial banking industry’ (Tobin, 1963: 6) at the point where the marginal return to the banks on their assets is equal to the marginal cost to them of deposits. Tobin went on to argue that reserve requirements (typically applied to banks but not to other financial intermediaries) ensure that the banking industry operates at a lower level of output than this competitive equilibrium. This gave some scope for bank reserves to determine deposits, but even here Tobin stressed the role of considerations of profitability (depending on the preferences of depositors and the lending opportunities available to the banks) on the one hand and the observed variations in actual reserve ratios on the other. The general implication of this ‘new view’ was that the growth of monetary and non-monetary financial claims needed to be analysed in terms of the portfolio adjustments of a range of bank and non-bank financial intermediaries as well as households and firms. More specifically, ‘the quantity of money as conventionally defined is not an autonomous variable controlled by governmental authority but an endogenous or “inside” quantity reflecting the economic behaviour of banks and other private economic units’ (Hester and Tobin, 1967: viii). The immediate impact of Tobin’s work in the UK was limited. Indeed, a survey by Johnson concluded that ‘the “new view” is long on elegant analysis of theoretical possibilities, but remarkably short on testable or tested theoretical propositions . . . ’ (Johnson, 1970: 105). More generally, it may be suggested that the econometric complexity and the detailed data required to model the ‘new view’ must have deterred attempts at this time to apply it, while its concept of money as endogenous (which would have been widely understood to mean that it was passive 19 See Cobham (1991a) for a more detailed discussion. 20 See, for example, Rowan (1968) and Newlyn (1962).

Historical Introduction 21 with respect to the growth of nominal income) may also have discouraged interest from some monetary economists. Whatever the reasons, however, there are few references to Tobin (1963) in the UK literature during the decade or so after it was published.21 However, the conventional multiplier analysis was not universally accepted in the UK. It found little place in the writings of Sayers (e.g. 1964), and was largely rejected by monetary economists such as Cramp (1966, 1971), while Sheppard’s (1971) empirical work on the UK found that, although the standard cash ratio model worked reasonably well for the period 1921–39, it performed poorly for the period 1947–62. Discussion tended to emphasise specific aspects of the UK financial system, such as the use of a liquid assets ratio, and issues of official behaviour such as the interest rate which the authorities did or should charge on loans to the discount market.22 The debate was transformed by the introduction into the official publication Financial Statistics in June 1966 of a table entitled ‘Factors determining changes in money supply’, following a paper in Economica by Bell and Berman (1966), who were respectively a Treasury economist and a government statistician. Bell and Berman set out to provide an ex post analysis of changes in the money supply in terms of government borrowing operations, the balance of payments and bank lending to the non-bank private sector. Most of the paper was concerned with the accounting relationships between the budget deficit and the various sources of finance to the government, between bank lending in various forms and bank deposits, and so on. However, at the same time Bell and Berman also tried to make statements about causation, for which they were sharply rebuked by Sheppard (1968). In particular, they provided no proper basis for their conclusion that ‘a large number of variables must be taken into account in any theory of the determination of the money supply, and that concentration on any one variable (e.g. the supply of Treasury bills) can be misleading’ (Bell and Berman, 1966: 159). Moreover, their final table gave a summary of key totals for the 11 years 1954–64, both in pounds and in percentage contributions to the increase in money supply, yet the totals concerned—which included measures of the government’s deficit and borrowing requirement, the sources of its borrowing, the increases in bank deposits and in the money supply, and the balance of payments, but not the increase in bank lending to the private sector—did not constitute an accounting identity and no possible combination of the percentages given added up to 100 per cent. In fact the Financial Statistics table, which excluded the balance of payments as such but included bank lending to the private sector, was rather less vulnerable to this criticism. The format of this table remained the same until June 1969, following an article in Economic Trends on ‘Money supply and domestic credit’. The new format then introduced, entitled ‘Domestic credit expansion and changes in the money supply’, proceeded from the public sector borrowing requirement (PSBR), via the net acquisition of public sector debt by the private sector and bank lending to the private and overseas sectors, to domestic credit expansion (DCE) and thence, via overseas lending to the public sector (which subsumed the balance of payments deficit) and some minor adjustments, to the increase in money supply. This format (with minor permutations involving, in particular, a downgrading of the DCE concept) served as the basis for the main official statistical presentations on the growth of broad money (M3 or sterling M3) for most of the next two decades, although there were some important changes 21 In fact even Goodhart (1973b) did not mention Tobin (1963). However, in his introduction to the paper in Goodhart (1984), he specifically linked his views on the money supply process with Tobin’s ‘new view’. Chick (1973: chapter 5.7) also clearly associates the two papers. 22 In particular, there had been some discussion in the 1960s of an alternative liquid assets theory in which it was the supply of liquid assets rather than cash that determined aggregate bank deposits; see Coppock and Gibson (1963).

22 The Making of Monetary Policy in the UK, 1975–2000 in detail. Sheppard had considered the assertion by Bell and Berman that the government’s borrowing operations, the balance of payments and the banks’ lending operations all influenced the money supply as ‘well worth examining’ (Sheppard, 1968: 301); there can be little doubt that the continued publication of statistics in this form strongly encouraged monetary thinking in the UK in this direction. All three strands of multiplier critique—the ‘new view’, the more institutional UK debate and the new statistical presentation—can be found in the seminal paper by Goodhart (1973b).23 Goodhart set out to show that the standard multiplier analysis was misleading and to propose in its place that the determination of the money supply should be analysed in terms of a general equilibrium stock adjustment model of the financial system. The heart of his critique of multiplier analysis was not its mechanical nature,24 but its assumption that the reserve base or stock of high-powered money H was exogenous. Goodhart argued that if the monetary authorities were trying to control H at all, this was only in order to control M, in which case they would presumably vary H in attempts to offset fluctuations in the multiplier itself; that official policy targets for H would not be independent of the state of the economy; and that the authorities were typically concerned about other variables such as interest rates and the exchange rate as well as about H. For all these reasons H could not be taken as exogenous; to do so would involve incorrect assumptions about the behaviour of the authorities. Goodhart then shifted the emphasis to the factors that determine H by working through the accounting identity that relates changes in H to the budget deficit, the government’s operations in marketable debt, the redemptions of maturing marketable debt, sales of non-marketable debt and external currency flows, stressing that at least some of these items were outside the immediate control of the monetary authorities while operations in marketable debt had obvious implications for interest rates. Next he sketched a simplified model of portfolio adjustment for the private sector and the banks, in which required reserve ratios appeared only among the arguments in the banks’ demand function for reserves. From these two analyses he identified ‘the four main critical issues influencing the determination of the stock of money [as:] (i) the size of the public-sector deficit, (ii) market reactions to the authorities’ open-market operations, (iii) the elasticity of substitution between foreign and domestic assets, and (iv) the interestelasticity of demand for advances’ (Goodhart, 1973b: 199), all of which were ‘simply pushed out of sight’ by the standard treatment of H as exogenous. Finally, this list was used to provide a rationale for the presentation in the official monetary statistics of changes in the money supply in terms of the public sector deficit, sales of public sector debt to the private sector, external financing of the public sector and bank lending to the private sector, on the grounds that this format encouraged users of the data to ask the relevant questions. Many readers of Goodhart (1973b) may feel that the paper moves rather too quickly from the four critical factors to the four credit counterparts of the official statistical presentation. It seems clear that the UK monetary authorities liked the counterparts approach partly on ‘political economy’ grounds, because each of the main domestic counterparts could be related directly to a particular area of policy (fiscal policy, interest rates/debt management, credit 23 Goodhart had previously (in a paper drafted before he went to work at the Bank of England) described the UK monetary authorities as having ‘next to no control over the reserve base of the banks’ (1973a: 495), partly because of their unwillingness to allow large swings in interest rates and partly because of the scope for the banks to ‘engineer’ increases in their own liquidity, e.g. by promoting the use of commercial bills (which were counted as liquid assets for the banks) as an alternative to advances. 24 That is, the use of fixed cash-to-deposits and reserves-to-deposits ratios. This can be remedied by modelling the ratios as depending on interest rates (e.g. Brunner and Meltzer, 1964).

Historical Introduction 23 controls and other policies towards bank lending), while external monetary flows were also included in the analysis (Fforde, 1983: 201). But there can be little doubt that a simplistic appreciation of the approach unaccompanied by awareness of the case for the determination of the money supply to be treated as a ‘branch of the more general theory of portfolio adjustment in response to price changes’ (Goodhart, 1973b: 199) contributed, for example, to the popularity of the view that monetary growth is determined by the size of the budget deficit. Goodhart himself had expected a full general equilibrium stock adjustment model of the financial system to be successfully constructed and estimated at the Bank of England within a few years. In the event, however, the combination of data and econometric problems mentioned above, on the one hand, and structural change in the financial system,25 on the other, prevented the successful estimation of such a model. Instead, empirical work in the 1970s and early 1980s tried to model the determinants of the individual credit counterparts, in particular bank lending to the private sector, but without much success. In the 1980s the emphasis shifted back towards larger models in which each counterpart was represented, but these models were far from the general equilibrium portfolio adjustment models that theory indicated were appropriate.26 Finally, from the late 1980s interest in the determination of the supply of broad money largely evaporated. Economists at the Bank of England and elsewhere continued to model the demand for broad money,27 but it was treated as endogenous in Tobin’s sense and a useful information variable, but little more than that.

2.4 INSTRUMENTS OF MONETARY CONTROL Over the period with which this book is concerned there has been considerable evolution in the instruments used in monetary policy. At the start of the period the authorities were targeting broad money, which they approached in terms of the credit counterparts. It was therefore natural that they thought of control of the individual counterparts (taking full account of any ‘offsets’ between them) as a means of controlling the aggregate. Moreover, their belief that they could control broad money from the supply side in this way was one of the reasons for choosing broad money rather than the monetary base or some other narrow aggregate as the target. In the 1960s the monetary authorities had relied on direct controls over bank lending as the main method of monetary control, but this was replaced in 1971 by the ‘new approach’ of Competition and Credit Control, according to which bank lending to the private sector should be rationed by price, rather than quantity. In the event this method of monetary control was unsuccessful, mainly because the demand for bank credit turned out to be much less interestelastic, particularly in the short run, than expected,28 and in 1972 and 1973 the UK experienced record rates of growth of broad money. One response to this was the introduction in December 1973 of the ‘corset’ (Supplementary Special Deposits Scheme), a new method of indirect control over bank lending. The corset required banks to place with the Bank of England special non-interest-bearing deposits at specified and steeply rising rates, if the growth of their interest-bearing eligible 25 As highlighted in Goodhart’s note to the later reprinting 26 See Cobham (1991a: section 2.3). 27 See, for example, Fisher and Vega (1993). 28 See, for example, Goodhart (1984: chapters III and IV).

of the paper (1984: 197).

24 The Making of Monetary Policy in the UK, 1975–2000 liabilities (IBELs)—essentially the variable element of the sterling resources available to the banks for on-lending—exceeded a norm predetermined by the authorities. Thus the corset worked by raising the marginal cost of deposits to the banks; it was intended to discourage the practice of liability management that had developed rapidly (along with the parallel money markets) since 1971, in which the banks bid for deposits to fund the lending they wish to make, rather than making loans in line with the deposits they find at their disposal, as in the traditional practice of asset management.29 The scheme was in action from December 1973 to February 1975, from November 1976 to August 1977, and from June 1978 to June 1980, but was then abolished. How effective the corset was is not entirely clear.30 It seems likely that in the first two episodes the corset did not make a major contribution to restraining monetary growth, and did not provoke much disintermediation either. During the third episode, however, there was significant and growing disintermediation via the bill leak (together with some effect, perhaps, on bank lending); after the abolition of exchange controls in the summer and autumn of 1979 disintermediation increased and the (non-cosmetic) effectiveness of the corset probably diminished. Its passing has not generally been regretted. However, it has been argued that its existence, even though it was not always in operation, served to restrain the expansionary ambitions of the banks, which continued to play only a small role in the mortgage market and did little to respond to the competitive threat posed by the building societies.31 During the monetary targeting period, apart from the corset, the authorities sought to act on the other two main credit counterparts: the PSBR and net sales of public sector debt to the private sector. The PSBR had previously been considered to be in the domain of fiscal policy, but in the mid-1970s it became accepted that it should be controlled for its contribution to monetary growth; this view lay behind some of the policy measures of the 1974–79 Labour governments and, even more, those of the Thatcher government from 1979. In fact, control of the PSBR was less easy than expected, largely because of difficulties in controlling public expenditure (particularly in the 1970s under conditions of high inflation) and because of its susceptibility to unforecast cyclical variation (notably in 1979/80 and 1980/81). Borrowing from the private sector, mainly through the issue of gilt-edged securities and the sale of National Savings products, had been relatively poorly controlled in the 1960s and early 1970s, mainly because the authorities were unwilling to bring about large swings in yields.32 However, from the mid-1970s through to the early 1980s there was a series of changes in the way gilts were issued, in the types of gilts issued, and in the marketing of National Savings products. On the one hand, the authorities became more aggressive in bringing about changes in yields in order to sell gilts.33 On the other hand, they introduced new types of gilts, particularly partly-paid issues (from 1977) and index-linked gilts (in stages from 1981). After the 29 See BEQB, 1982a: 74–85, for a clear account of the scheme and its operations. 30 See BEQB, 1982a: 74–85, Spencer (1986: chapters 5 and 6) and Cobham (1991b) for the more detailed discussions

on which this paragraph is based. 31 See Dow and Saville (1988: 156–7). The banks began to compete much more vigorously with the building societies after the abolition of the corset in 1980. 32 The principal reason for this was that they believed that the demand for gilts depended heavily on expected price changes and expectations were largely extrapolative. 33 Their tactics in this respect came to be referred to as the ‘Duke of York’ strategy, on the notion that they forced interest rates up (sharply) and then let them decline (more gradually); during the decline yields would be relatively high and gilts prices would be rising, so that investors who cared about the level of yields and/or short-term capital gains would be willing to buy. See Gowland (1982: 148–52, 189–90). However, it is difficult to find examples of swings in interest rates which fit neatly into this pattern, and most interest rate movements can be explained without reference to official concern with debt sales (see Chapters 8 and 9).

Historical Introduction 25 mid- or late 1980s there was less evolution in types of gilts, but the authorities moved gradually (between the mid-1980s and the mid-1990s) from predominance of the longstanding ‘tap’ system for issuing gilts to almost complete reliance on auctions. Moreover, from 1998 gilts sales have been handled by the Debt Management Office, an Executive Agency under the Treasury, while monetary policy as such is the domain of the Bank of England and the MPC. Once the corset had been abolished the authorities could influence bank lending to the private sector only via interest rates, and more indirectly via income and business confidence, but the response of the demand for credit to interest rates is weak and lagged, and the effect of income and confidence probably even more lagged and variable. One of the results of this was that in the first half of the 1980s bank lending repeatedly overshot the authorities’ expectations (see Chapter 3). They reacted by increasing the sales of public sector debt to try to offset the impact of bank lending, and this rapidly turned into ‘overfunding’, in which the public sector borrows from the private sector more than its deficit. Other things being equal, overfunding restrains the growth of bank deposits, but it creates shortages of bank liquidity which need to be relieved if the monetary and payments system is to continue to function. The authorities acted to restore banks’ liquidity by purchasing short-term securities from the banks, initially Treasury bills but later mainly commercial bills. However, this led to the accumulation of a ‘bill mountain’ in the hands of the authorities and a distortion of the pattern of financial flows and relative interest rates, while it became increasingly unclear whether the overall effect on monetary growth was real or cosmetic.34 Overfunding was abandoned in 1985, with the change being formally announced by the Chancellor in October along with the suspension of the £M3 target, but there had been little overfunding since early that year. Once it had been abandoned, the authorities could no longer effectively target broad money, although a further target was set for 1986/87. By the mid- to late 1980s, then, monetary policy had a single instrument, the interest rate, which was used, not to try to generate a particular level or rate of growth of any monetary aggregate, but to influence aggregate demand (and to varying extents the foreign exchange and other financial markets) directly.

34 See Goodhart (1991: 297–9) and, for an official view before the policy was abandoned, the article on funding the PSBR in BEQB, 1984d: 482–92.

3 Monetary Targeting, 1977–1986 This chapter explores the formal monetary targeting which was undertaken in the UK from 1977 until it was abandoned in 1986. In particular, it examines the relative contributions of the authorities’ own targeting procedures and exogenous financial innovation to the perceived failure of monetary targeting. Section 3.1 reviews the monetary aggregates used in the UK and the type of targets set over the period. Section 3.2 compares the target ranges for £M3 with the official forecasts (so far as they can be obtained) for the growth of money GDP over comparable periods, and with the previous period growth (as known when the targets were set) of money GDP and £M3. Section 3.3 attempts to reconstruct, on the basis of the information available, the forecasts for the various counterparts of £M3 which the authorities presumably made alongside the (published) monetary targets. Section 3.4 reviews the meaning and nature of financial innovation, and discusses how it is thought to have impinged on monetary targeting in the UK. Section 3.5 examines the target and outturn for £M3 in each year of the period in turn, and discusses how both the setting of the targets and financial innovation contributed to the overshoots of the broad money targets in various years. Section 3.6 draws out the conclusions on the monetary authorities’ target-setting procedures and on the causes of the overshoots. Section 3.7 discusses the reasons for the abandonment of monetary targets in the light of the preceding analysis.

3.1 MONETARY AGGREGATES AND THE TYPE OF TARGETS USED In the mid-1970s the two main definitions of money used in the UK were narrow money M1 (notes and coin in circulation plus the UK private sector’s sight deposits at banks) and broad money M3 (which also included private sector time deposits at banks). During the years of informal monetary targeting from 1974 to 1976 the emphasis was on M3, but attention was shifted to £M3 (which excludes private sector foreign currency bank deposits) in late 1976 at the time of the agreement with the IMF on limits to a corresponding definition of domestic credit expansion (DCE). There were three main factors underlying the UK monetary authorities’ attachment to broad money.1 First, broad money had been a better predictor of the rise in inflation in the early to mid1970s than M1, though the demand for broad money was not clearly more stable than that for narrow. Second, the credit counterparts framework developed in the late 1960s and early 1970s, partly as the result of IMF pressure for limits on DCE, enabled monetary policy to be seen as part of a wider macroeconomic framework including fiscal policy in a way that emphasised the need to reduce budget deficits and finance them through non-monetary means.2 Third, that 1 See 2 See

in particular the Governor’s 1978 Mais lecture, BEQB, 1978a: 31–7, particularly pp. 36–7. Fforde (1983: 201–2) for the ‘political economy’ of M3.

28 The Making of Monetary Policy in the UK, 1975–2000 framework also allowed the authorities to believe that they had a technique for controlling the money supply as defined in this way (by acting on the various credit counterparts), whereas the alternative of M1 could have been controlled only by acting on the demand side or, possibly, by making the radical shift to monetary base control to which they were deeply opposed. In addition, the authorities may have felt that the border between M1 and non-M1 elements in M3 was not sufficiently firm to make M1 targets viable, and that it would be better to target the whole of the relevant deposits in the banking system rather than just a subset of them.3 Throughout the years of formal monetary targeting from 1977/8 to 1986/7 the main targets were for £M3, but these were supplemented for 1982/3 and 1983/4 by targets for M1 and PSL2 (a wider aggregate which, in addition to the components of £M3, included private sector holdings of shares and deposits in building societies, sterling certificates of deposit and various other liquid assets);4 and from 1984/5 by targets for M0 (a measure of the monetary base consisting of private and banking sector holdings of notes and coin plus banks’ operational deposits at the Bank of England, first introduced in 1981). The targets for M1 and PSL2 were added in the light of the difficulties experienced in hitting the £M3 targets and doubts about the stability of its velocity, while by 1984 the behaviour of M1 was ‘becoming increasingly difficult to interpret’ and it was felt that other measures of narrow money such as M0 were ‘likely to be more satisfactory indicators of financial conditions’ (FSBR 1984–85: 6).5 In 1987 £M3 was renamed M3 (while the old M3 became M3c), PSL2 was renamed M5 and attention was shifted to a new wider definition of money named M4 (originally PSL1) which is constructed on a similar basis to M3 but covers building society as well as bank deposits.6 In July 1989, when the Abbey National building society was converted into a public limited company and a bank, this would have produced a substantial (but artificial) increase in M3: instead the authorities ceased to report M3, and in effect replaced it by M4 (BEQB, 1989c: 352–3). M1 was also discontinued at that time though NIB (non-interest-bearing) M1 was still published for a few years. M4c and M5 were replaced in 1991 by the publication of a range of data for ‘liquid assets outside M4’ (BEQB, 1991b: 263–6). From that time the two main definitions in use were M0 and M4, referred to as narrow and broad money respectively.7 Target ranges continued to be set for M0 after the £M3 targets were abandoned in 1987. The target range became a ‘monitoring range’ in 1993/4 when a formal inflation target was adopted, and a similar monitoring range for M4 was also introduced at that time. Those ranges lasted (unchanged) until the Monetary Policy Committee was set up in May 1997. However, the target ranges set after 19878 were not major determinants of interest rate decisions (see Chapter 8 below) or important elements of monetary policy. The informal target for 1976/7 had been for the financial year, and the formal targets for 1977/8 and 1978/9 were set for the 12 months from April to April. However, the late overshoot of the 1977/8 target and the recognition of the pressures that could develop towards the end of a target period led the authorities to roll the April 1977 – April 1978 target forward (at the 3 However, even by 1978 the authorities were aware that the boundary between banks and building societies was becoming blurred (BEQB, 1978a: 37). 4 PSL stands for private sector liquidity. 5 See the Appendix to this chapter for comprehensive data on monetary targets and outturns, including those for M0, M1 and PSL2. 6 For further details see the article on ‘Measures of broad money’, BEQB 1987b: 212–19. 7 M2, an alternative narrow aggregate to M1, was introduced in 1982; it was revised in 1992 to become a subset of M4 and continued to be published first as M2 and then as ‘retail deposits and cash in M4’. See BEQB, 1990c: 336–7; 1992c: 317. 8 See the Appendix to this chapter for targets and outturns.

Monetary Targeting, 1977–1986 29 same rate) for the 12 months October 1978 – October 1979. The following target, announced in the June 1979 Budget of the new government, was for the 10 months from June 1979 to April 1980 (but was later extended to the period up to October 1980). For the first four years of the MTFS, the targets were for the 14 months from, for example, February 1980 to April 1981 (with an overlap that served the same function as a rolling target). In May 1985 it was announced that in future the target ranges would apply to the 12-month growth rates throughout the financial year, so that for 1985/6 the relevant period was from May 1985 to April 1986, and for 1986/7 it was from March 1986 to April 1987.9 The continuing M0 targets and later monitoring ranges (and the later M4 monitoring ranges) were also set on the 12-month growth rate basis.

3.2 TARGET RANGES AND ECONOMIC FORECASTS A remarkable feature of UK monetary targeting is the almost complete lack of official explanation for (and non-official comment on) the numbers selected for the target ranges. The monetary target ranges were officially announced in the annual Budget. The Financial Statement and Budget Report (FSBR) typically refers to the Chancellor’s Budget Statement in the House of Commons as the source of the decision, but the Budget Statements contain little explanation of the numbers chosen and no analysis of alternative target numbers. The Treasury and Civil Service Committee (1979–80) tried in July 1980 to get the Chancellor to justify his choice of 7–11% for 1980/81 but without significant result. Only towards the end of the targeting period was any explanation given: the FSBR 1985–86 contains a table with a forecast for money GDP alongside the monetary target ranges, with the statement that the table ‘shows the assumptions about the growth of money GDP on which the present ranges are based’ (p. 7); the FSBR 1986–87 has a comparable table, but no such statement. However, in both cases the forecasts for money GDP were for the growth between one financial year and the next, which was not the same period as that for the monetary targets, and by this time the authorities’ commitment to their targets was questionable (and widely questioned). On a priori grounds, or on the basis of practice in other countries,10 it might be expected that the numbers chosen for the target ranges for £M3 would be systematically related to the official forecasts for the growth of money GDP, and/or to the recent growth of money GDP and of £M3 itself. Table 3.1 provides a selection of relevant data. Column (3) gives the implicit forecast for the growth of money GDP over the target period derived from the published official forecasts for real GDP and retail price inflation in columns (1) and (2). Column (4) gives the explicit forecast for the growth of money GDP between financial years given in the FSBR for the second half of the period. Columns (5), (6) and (7) give the latest figures for the growth of money GDP, retail prices and £M3 as known at the time of the various budgets. Column (8) gives the £M3 target range, and columns (9) to (12) compare the midpoint of the target range with, respectively, the implicit and explicit forecasts for money GDP, the previous period growth of money GDP and the previous period growth of £M3. It is convenient to take first the comparison of the target with the previous outturns. Column (11) shows that there was considerable variation in the relationship between the target and previous money GDP growth, while column (12) shows some large fluctuations in the gap 9 Economic Progress Report (HM Treasury), no. 177, May 1985. 10 See, for example, Courakis (1981) and the papers in Meek (1983).

111/4 71/2 143/4 15 9 81/4 6 11/2 43/4 31/2

11/2 3 −1 −11/2 1 2 21/2 21/2 21/2 21/2

Year

1977/8 1978/9 1979/80 1980/81 1981/2 1982/3 1983/4 1984/5 1985/6 1986/7

123/4 101/2 133/4 131/2 10 101/4 81/2 63/4 71/4 6

Money GDP (implicit) (3) n.a. n.a. n.a. n.a. n.a. 9.8 7.6 7.9 8.3 6.7

Money GDP (explicit) (4) 14.6 12.9 10.3 16.6 13.4 10.2 9.7 8.3 6.6 11.0

Money GDP (5) 16.2 9.5 10.3 19.1 12.5 11.0 5.3 5.1 5.4 5.1

Retail prices (6)

Previous period outturns

9 12 111/2 12 20 141/2 10 93/4 91/2 143/4

(7)

£M3 9–13 8–12 7–11 7–11 6–10 8–12 7–11 6–10 5–9 11–15

target (8)

£M3

n.a. n.a. n.a. n.a. n.a. 0.2 1.4 0.1 −1.3 6.3

−13/4 −1/2 −43/4 −41/2 −2 −1/4 11/4 −1/4 7

1/2

explicit (10)

implicit (9)

Money GDP forecasts

−3.6 −2.9 −1.3 −7.6 −5.4 −0.2 −0.7 −0.3 0.4 2.0

Money GDP previous (11)

£M3 target minus

2 −2 −21/2 −3 −12 −41/2 −1 −13/4 21/2 −13/4

previous (12)

£M3

Notes: (1) Growth of GDP at factor cost between first halves of calendar year of budget and following calendar year, as given in FSBR. (2) Average of two estimates which span financial year, as given in FSBR. (3) = (1) + (2) (4) Growth in GDP at current market prices between forthcoming and previous financial years, calculated from financial year totals in FSBR. (5) Growth in GDP at current market prices, expenditure based, s.a., for latest four quarters available, as given in Economic Trends, July 1979 and April of all other years. (6) Growth of RPI over latest 12 months available, as given in Economic Trends, June 1979 and March of all other years. (7) For 1977/8 growth over latest 12 months, January 1976 to January 1977, as given in BEQB 1977a; for other years growth at annual rate over previous target period as given in FSBR: April 1977 to February 1978; April 1978 to April 1979; June 1979 to February 1980; February 1980 to February 1981 and similarly for remaining years. (8) Target (at annual rates) for April 1977 to April 1978; April 1978 to April 1979; June 1979 to April 1980; February 1980 to April 1981 and similarly for remaining years. (9) to (12) = midpoint of (8) minus, respectively, (3), (4), (5) and (7).

Retail prices (2)

Real GDP (1)

Forecasts

Table 3.1 £M3 targets and other indicators (%)

Monetary Targeting, 1977–1986 31 between the target and previous £M3 growth. Some of these variations can be attributed to oneoff effects such as the June 1979 rise in VAT and the post-corset re-intermediation distortion to £M3: insofar as the former was thought to have given a once-for-all boost to prices and hence money GDP in 1979/80 an unchanged stance of monetary policy would have involved a larger gap between the target for 1980/81 and previous money GDP growth; and insofar as the postcorset distortion to £M3 was not expected to be repeated an unchanged monetary stance would have involved a larger gap between the target for 1981/2 and previous £M3 growth. A further part of the variation in column (11) may be due to a change in the authorities’ expectations regarding the velocity of £M3. There is no precise information available on these expectations on a year-by-year basis, but it seems that in the late 1970s and very early 1980s the authorities reckoned on a trend rise in velocity of 1–2% a year, and then from 1982 or so they began to assume a trend fall in velocity of around 2% a year. This change would account for part of the reduction in the gap between the target and previous money GDP growth as between the first four or five years and the last five years of the period. A similar effect may have occurred with the upward revision of the target ranges in 1982 which affects the figures in column (12). However, it is highly unlikely that effects of these kinds could account for all the variations in columns (11) and (12)—what remains is presumably a variation in the intended severity of policy. Columns (9) and (10) compare the target with the money GDP forecasts. Again the figures suggest considerable variation from year to year. In this case one-off effects of the kind discussed in the previous paragraph are not relevant since the monetary authorities should have taken account of them in their money GDP forecasts. Changes in the intended severity of policy are not relevant either, because changes in policy should also have been taken into account in the forecasts. Thus the figures in these columns amount (with the signs reversed) to the change in velocity which would have occurred if the money GDP forecasts had been correct and if £M3 had grown in the middle of the target range. They can therefore be compared with the authorities’ explicit beliefs about velocity. As indicated above there is no detailed information on these beliefs but, while they probably moved from an expected trend rise to an expected trend fall as between the first and second halves of the period in a way that is consistent with part of the variation in column (9), it seems highly unlikely that they could have varied on a year-by-year basis by anywhere near as much as is implied by the figures in the table. What this analysis shows is that the numbers picked for the monetary targets do not appear to have been related at all closely either to the official forecasts for money GDP or to the previous period growth of money GDP or £M3.11 Instead, the closest relationship in the table is that between the target range in each year and that for the previous year: with the exception of the ‘standstill’ in 1979/80 and the increases of 1982/3 and 1986/7 (both of the latter also representing upward revisions to the monetary targets for those years presented in earlier versions of the MTFS), there is a consistent ‘n − 1’ relationship between the monetary targets in succeeding years.

11 It is worth noting that this is not a universal feature of monetary targeting. Cobham and Serre (1986, 1987) showed that the mean and standard deviation of the divergences between the monetary target and (a) lagged monetary growth and (b) lagged nominal income growth were considerably lower in France than in the UK. They also showed that the mean and standard deviation of the divergences between the monetary target and the monetary outturn were considerably lower in France.

32 The Making of Monetary Policy in the UK, 1975–2000

3.3 TARGET RANGES AND CREDIT COUNTERPART FORECASTS As discussed in Chapter 2, the UK monetary authorities approached the question of monetary control during this period in terms of the counterparts of £M3. However, complete sets of their forecasts for the counterparts consistent with the monetary targets are not available.12 Table 3.2 represents in the odd-numbered rows those forecasts that are available and a tentative reconstruction of those that are not, while the even-numbered rows give the outturn figures. The point of this reconstruction is to shed light on the internal consistency of policy: given the instruments of monetary control at the disposal of the authorities and their way of approaching issues of control, would the targets have appeared to them to be realisable? The reconstruction is necessarily speculative and broad-brush, but as will be seen it enables some interesting points to be made. It also provides a framework for an analysis of monetary growth as such, and it will be used for that purpose in Section 3.5. Row (1) gives the target growth of £M3, obtained by multiplying the stock outstanding at the end of the previous financial year by the midpoint of the target annual growth range (the monetary targets applied to slightly different time periods but the other published forecasts applied to financial years and that period is therefore used here). Row (3) gives the PSBR forecasts published in the FSBR. Row (7) gives the forecasts for sales of gilt-edged securities to the non-bank private sector from Coleby (1983), the only published source of such forecasts. Row (9) gives the National Savings (NS) debt sales targets from the BEQB and the 1981–5 Budget Statements. Row (5) gives an estimate of the authorities’ forecasts for total sales of public sector (PS) debt to the private sector, based partly on rows (7) and (9). Rows (11), (13) and (15) give estimates of the authorities’ forecasts for sterling bank lending to the private sector, the external and foreign currency counterparts, and net non-deposit liabilities respectively, based on a variety of considerations specific to the financial year concerned. Finally, row (17) shows the difference between the target growth of £M3 in row (1) and the sum of the counterpart forecasts, rows (3) + (5) + (11) + (13) + (15). The specific assumptions made in constructing each year’s estimates will now be discussed in turn. For 1977/8 the PSBR forecast was a little lower than the 1976/7 outturn and much lower than the outturn for the two previous years. The official forecast for gilt-edged sales from Coleby (1983) was, at £3.9bn, nearly £2bn below the 1976/7 outturn. No NS debt target is available but NS debt sales had been rising in recent years, from £0.1bn in 1974/5 to £1.0bn in 1976/7. Other public sector debt sales had averaged £0.9bn but varied widely over the last three years. These considerations suggest a forecast for total debt sales in row (5) of £5.5bn, roughly the average of the three previous years (when the PSBR had been slightly higher on average) and implying a public sector contribution to monetary growth (i.e. PSBR minus PS debt sales to the private sector) of £3bn, above that for 1976/7 but below the average for 1974/5 to 1976/7. Bank lending to the private sector had risen to £3.4bn in 1976/7 after the exceptionally low levels of late 1974 and 1975/6; however, this was thought to have been at least partly due to the prolonged sterling crisis of 1976 (BEQB, 1976c: 296), and the corset had been reintroduced in November 1976. The authorities are therefore likely to have expected a somewhat lower level in 1977/8, such as the £3bn figure in row (11). The estimated forecast for external and foreign currency counterparts in row (13) is less negative than the outturn for recent years, because the current account was expected finally to move into surplus in the first half of 1978 (FSBR 12 A

1987 request for this information by the present writer was turned down by the Treasury.

£ bank lending to private sector External and foreign currency counterparts Net non-deposit liabilities Residual

National savings (−)

Public sector debt sales to private sector (−) Gilt-edged securities (−)

PSBR

£M3

Item target outturn forecast outturn forecast outturn forecast outturn target outturn forecast outturn forecast outturn forecast outturn forecast (%£M3)

4.4 6.2 8.5 5.6 −5.5 −6.7 (−3.9) (−4.9) — (−1.1) 3.0 3.7 −0.5 +4.0 −0.8 −0.4 −0.3 (−0.8)

1977/8 4.6 5.3 8.5 9.2 −7.0 −8.5 (−5.8) (−6.2) — (−1.6) 4.0 6.3 0 −0.6 −0.8 −1.0 −0.1 (−0.2)

1978/9 4.6 6.4 8.3 9.9 −8.0 −9.2 (−6.4) (−8.3) — (−1.4) 6.0 9.3 −1.0 −2.4 −1.0 −1.2 +0.3 (+0.6)

1979/80 5.2 10.7 8.5 13.2 −8.0 −10.8 (−5.4) (−8.9) (−2.0) (−4.2) 8.0 9.2 −2.0 +0.6 −1.2 −1.5 −0.1 (−0.2)

1980/81 5.5 9.8 10.6 8.8 −10.5 −11.3 (−7.2) (−7.1) (−3.0) (−3.6) 8.0 14.9 −1.0 −0.8 −1.5 −1.7 −0.1 (−0.1)

1981/2 8.6 10.1 9.5 9.2 −9.5 −8.4 n.a. (−4.6) (−3.0) (−3.0) 13.0 14.4 −1.0 −2.6 −1.8 −2.4 −1.6 (−1.9)

1982/3 8.7 7.9 8.2 10.0 −8.2 −12.5 n.a. (−9.8) (−3.0) (−3.4) 14.0 15.2 −2.5 −0.2 −2.4 −4.7 −0.4 (−0.4)

1983/4 8.2 11.8 7.2 10.2 −7.2 −12.6 n.a. (−9.3) (−3.0) (−3.1) 14.0 18.6 −1.5 −1.7 −3.5 −2.7 −0.8 (−0.8)

1984/5 8.0 19.1 7.1 5.8 −7.1 −3.5 n.a. (−2.8) (−3.0) (−2.1) 18.0 21.4 −2.0 −2.5 −2.5 −2.0 −5.5 (−4.8)

1985/6

17.4 25.4 7.1 3.3 −4.6 −1.2 n.a. (−1.5) n.a. (−3.4) 21.0 30.3 −2.5 −2.4 −2.5 −4.6 −1.1 (−0.8)

1986/7

Note: (1) Target growth of £M3 obtained by applying midpoint of target range to outstanding stock at end of previous financial year. (3) PSBR forecast from FSBR. (5) Estimated forecast (see text for details). (7) Forecast for sales of gilts to non-bank private sector given in Coleby (1983). (9) Target for sales of National Savings debt given in BEQB and/or Budget Statements: BEQB, 1982c: 350; House of Commons Report—Hansard (HC), vol. 1000, col. 764; vol. 19, col. 735; vol. 39, col. 140; vol. 56, col. 290; vol. 75, col. 786. (11), (13) and (15) Estimated forecasts (see text). Relationships between rows: (1) = (3) + (5) + (11) + (13) + (15) + (17); (2) = (4) + (6) + (12) + (14) + (16). Outturn figures in even columns from BEQB, Financial Statistics, various issues; adjusted to consistency with definitions used in forecasts.

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18)

Row

Table 3.2 £M3 and counterparts, forecasts and outturns (£bn)

34 The Making of Monetary Policy in the UK, 1975–2000 1977–78: 12). The estimated forecast for net non-deposit liabilities in row (13) is in line with the recent trend. These forecasts sum to £4.7bn, £0.3bn (0.75% of £M3) above the growth of £M3 implied by the midpoint of the 9–13% target range; such a residual is within the margin of error that needs to be attached to this sort of exercise and can be disregarded. For 1978/9 the PSBR forecast was the same as the forecast for the previous year (which had been undershot) while the gilts sales target was much higher. The estimated forecast in row (5) for total PS debt sales is between the outturns for 1976/7 and 1977/8; it implies both a lower PS contribution to monetary growth than that forecast for 1977/8 and a narrowing of the gap between the gilts target and the total debt sales forecast to take account of the likely crowding out of sales of other forms of PS debt by sales of gilts. The forecast of £4bn for bank lending to the private sector is a little above the outturn for 1977/8 on the grounds that the economy was expected to expand more rapidly and the corset had been suspended in August 1977 (it was reimposed in June 1978 when bank lending turned out to be higher than expected—BEQB, 1978b: 169; 1978c: 344). The externals forecast in row (13) is put at zero in response to the unprecedented positive outturn for 1977/8 (associated with the attempt to hold the exchange rate down in the face of large capital inflows in the late summer and autumn of 1977). The net non-deposit liabilities forecast in row 15 is in line with the trend. These forecasts sum to £4.7bn, just £0.1bn above the growth of £M3 implied by the midpoint of the target range. For 1979/80 the PSBR forecast was again little changed on recent forecasts. The gilts sales forecast was higher than the 1978/9 outturn, which had exceeded the 1978/9 forecast. No NS debt sales target is available, but sales had risen further in recent years to £1.6bn in 1978/9. Other PS debt sales on the other hand had been very low. These considerations point to a forecast for total PS debt sales in row (5) of around £8bn, which implies a PS contribution to monetary growth of £0.3bn, rather less than the outturn for 1978/9. Bank lending to the private sector had been £6.3bn in 1978/9 with some upward trend through the year; it seems likely that the authorities would have expected it to continue at something like the same level since the corset remained in operation and had been neither tightened nor relaxed, but minimum lending rate (MLR) was raised by 2% in the June 1979 Budget ‘primarily to moderate bank lending’ (BEQB, 1979c: 261): hence the estimated forecast of £6bn.The estimate in row (13) for the external counterparts is in the centre of the recent range for this item, while that in row (15) for net non-deposit liabilities allows for some trend growth. These forecasts sum to £4.3bn, giving a residual of +£0.3bn, which is again within the margin of error and can be disregarded. For 1980/81 both a gilts sales forecast and an NS debt sales target are available; the total debt sales figure in row (5) allows for a small amount of other debt sales, and with the PSBR forecast little changed implies a slightly lower PS contribution to monetary growth. Bank lending to the private sector had turned out much higher in 1979/80 than forecast; the figure of £8bn in row (11) for 1980/81 moves some of the way towards the 1979/80 outturn, but not all the way, since the authorities expected the industrial demand for bank credit to ‘be reduced in the course of the recession’13 and the outturn of £9.2bn was later described as ‘much more than expected’ (BEQB, 1981b: 161). The figure for the external counterparts in row (13) is set (absolutely) higher than the 1979/80 forecast in the light of the 1979/80 outturn and the current account deficit which was forecast for 1980/81 (FSBR 1980–81: 27). The net non-deposit liabilities estimate in row (15) is in line with recent trends. These forecasts yield a negligible residual in row (17) of −£0.1bn. 13 BEQB, 1980b: 140. On the other hand the Budget Statement expected some upward effect on bank lending from re-intermediation following the abolition of the corset (HC, vol. 981, col. 1445).

Monetary Targeting, 1977–1986 35 For 1981/2 the PSBR forecast was much higher than that for previous years, following the large overshoot of 1980/81; the gilts sales forecast and the NS debt sales target were also much higher. The estimated total PS debt sales forecast in row 5 allows for a smaller amount of other debt sales than the estimated forecast in 1980/81, since other PS debt sales had been negative in 1980/81 and the authorities are unlikely to have aimed to overfund. The estimated bank lending forecast in row (11) is the same as that for 1980/81—the latter had been exceeded but the authorities thought the demand for credit was ‘moderating’ (BEQB, 1890d: 389; see also 1981a: 21) and no repeat of the reintermediation that had affected the 1980/81 outturn was expected; moreover the 1981/2 outturn was later described as ‘substantially larger than expected’ (BEQB, 1982b: 180). The externals estimate in row (13) represents a reversion to trend; the 1980/81 outturn had been positive rather than negative and a small current account surplus was forecast for 1981/2 (FSBR 1981–82: 28), but the authorities were emphasising that a current account surplus does not necessarily produce a positive external influence on £M3 (BEQB, 1981a: 5), and that capital outflows were running at a high rate (BEQB, 1981a: 17; see also 1981b: 162), presumably partly in response to the abolition of exchange controls. The figure for net non-deposit liabilities in row (15) again follows the trend, and the total of the counterpart forecasts yields a negligible residual of −£0.1bn. For 1982/3 no gilts sales forecast is available. In 1981/2 the authorities had overfunded but it is unlikely that they intended to do this again; on the other hand they now clearly recognised the strength of the demand for bank lending (see, for example, BEQB, 1982a: 6; 1982b: 179–80). The estimated forecast for total PS debt sales in row (5) therefore assumes the intention of full funding of the PSBR. The 1981/2 bank lending forecast had been greatly exceeded by the outturn of £14.9bn, with lending to companies lower in the first half and higher in the second half of the year partly as a result of the delay in tax payments caused by the civil service dispute. For 1982/3 the authorities thought that lending to persons might ease to some extent, but ‘this may be less likely for lending to business’ (BEQB, 1982a: 6). This suggests their forecast must have been at least the £13bn of row (11). The estimated externals forecast in row (13) is the same as that for the previous year, which had been relatively accurate: a current account surplus was expected (FSBR 1982–83: 24), but 1981/2 had seen a considerable surplus associated with a negative external counterpart and the capital outflow remained high.14 The net non-deposit liabilities figure in row (15) is again in line with the trend. In this case, however, the total of the counterpart forecasts exceeds the £M3 target by £1.6bn (indeed the total is only £0.1bn less than the £M3 growth implied by the upper bound of the target range). Such a divergence immediately prompts a reconsideration of the counterpart forecasts, but it is difficult to see how these could be made any less: it is not plausible that the authorities intended from the beginning of the year to overfund, for they had been under criticism for some time for the growth of the bill mountain associated with the overfunding of 1981/2; after years of under-predicting bank lending they were trying to take full account of its likely course (this was one of the main reasons why the £M3 target had been revised upwards this year) and might well have expected an even higher level; and the expected contractionary impact of the other two counterparts together was larger than the outturn in most recent years. It is therefore difficult to avoid the conclusion that the authorities’ forecast for the counterparts added up to a figure near the top rather than the middle of the target range, even though the latter was revised upwards, from the 5–9% specified in the 1980 and 1981 versions of the MTFS to 8–12%. 14 Including the balancing item. See BEQB, 1982a: 18, where it is also suggested that the current account surplus for 1982 would be considerably smaller than the estimated outturn of £8bn for 1981.

36 The Making of Monetary Policy in the UK, 1975–2000 For 1983/4 again no gilts sales forecast is available and the figure in row (5) assumes the intention of full funding (there had been slightly less than full funding in 1982/3). The authorities appear to have expected bank lending to remain high, though perhaps a little lower than in 1982/3 (BEQB, 1983a: 7), hence the figure of £14bn in row (11) as against the 1982/3 outturn of £14.4bn. The externals forecast in row (13) is set much higher in the light of the 1982/3 outturn and the smaller current account surplus expected in 1983/4 (FSBR 1983-84: 20). The net non-deposit liabilities figure in row (15) reflects the sharp rise in the 1982/3 outturn. This leaves a small residual of −£0.4bn. For 1984/5 it is again assumed that the authorities intended full (narrow) funding of the PSBR, despite the substantial overfunding which had occurred in 1983/4. The estimated forecast for bank lending is the same as the forecast but lower than the outturn for 1983/4 (and the two previous years’ outturns); the authorities expressed the hope that the Budget might encourage companies to fund their overdrafts (BEQB, 1984a: 7). Moreover they later described the acceleration of bank lending in the last two quarters of the financial year as ‘puzzling’ (BEQB, 1984d: 453; 1985a: 6). The externals and net non-deposit liabilities forecasts in rows (13) and (15) each show a reversion to trend levels after the unusual outturn of 1983/4, but their total is roughly equal to the outturn for both 1982/3 and 1983/4. This leaves a residual of −£0.8bn (equal to 0.8% of £M3). For 1985/6 the intention of full funding is again assumed. The forecast for bank lending to the private sector is much higher than recent forecasts, but slightly below the outturn for 1984/5; there is no indication that the authorities expected bank lending to fall back sharply from the previous level, but they had found the latter ‘puzzling’ and the March 1985 BEQB mentioned the possibility that the ‘approval of short-term capital market borrowing’ and the ‘fuller capital gains tax indexation’ introduced in the Budget might encourage companies to borrow more from the capital markets and less from the banks.15 The externals and net nondeposit liabilities forecasts show a further reversion to trend levels, with the total in line with the 1984/5 outturn. These forecasts exceed the midpoint of the £M3 target range by £5.5bn, much the largest residual in the table, and its upper bound by £3.2bn. It is possible that the authorities were a little more optimistic on bank lending, and that they expected a slightly larger negative impact from the externals and net non-deposit liabilities. It is conceivable that they planned on overfunding the PSBR to a limited extent. But it is not plausible that differences such as these from the estimated forecasts in the table could amount to £5bn or even £3bn—in other words the authorities must have known at the beginning of the year that the £M3 target which they were announcing was not realisable. Finally for 1986/7 the intention of full wide funding is assumed: the Chancellor’s Mansion House speech of October 1985 had announced the abandonment of overfunding in favour of a policy of full wide funding, that is, covering the PSBR out of the sum of external and foreign currency financing of the public sector and sales of debt to the domestic non-bank private sector (BEQB, 1985d: 518). The bank lending forecast is in line with the outturn for 1985/6; it is clear that the higher outturn in 1986/7 was above official expectations (BEQB, 1986b: 185, 187; 1986d: 476; 1987a: 28). The estimated forecast for net non-deposit liabilities is in line with recent levels. These figures leave a small residual of −£1.1bn (0.8% of £M3) relative to the midpoint of the new, upward-revised, target range of 11–15%. This analysis shows that although most of the target ranges were consistent with how the authorities appear to have expected the counterparts to £M3 to develop, there was one year— 15 BEQB, 1985a: 7. The June BEQB took the view that ‘some moderation in the pace of bank lending is still to be expected, but one cannot be wholly confident of this until it begins to become apparent’.

Monetary Targeting, 1977–1986 37 1985/6—when the authorities cannot have believed the target range was realisable, and another year—1982/3—when they must have expected monetary growth to be at the top rather than the middle of the target range.

3.4 FINANCIAL INNOVATION AND MONETARY TARGETING Financial innovation can be defined as change in the structure of the demands for and supplies of different financial claims. In particular it refers to changes in the types of deposit and loan offered by banks and other financial intermediaries, to the development of new financial claims and markets for them in the sphere of derivatives, and to new ways of carrying out financial transactions in a wide range of markets from domestic and international money and capital markets.16 Table 3.3 lists the main financial innovations that occurred in the UK in the 1970s and 1980s. The process of financial innovation is multifaceted and the individual innovations are heterogeneous; they are therefore classified in the table simply in terms of the area of their immediate impact and the date from which they became prominent. In the banking sector, there were a series of innovations from the early 1970s, involving more competition between banks on both deposits and advances, narrower margins for the banks, and lower transaction costs and costs of bank intermediation for the banks’ customers. Among the building societies, there was a growth of competition between societies and a move towards the provision of the kinds of financial services (personal loans and money transmission) previously offered only by banks. In addition, there came to be much more competition between the banks and the building societies, on both asset and liability sides, and the boundary between these two groups of intermediaries became increasingly blurred. The money markets saw the development of a range of new instruments and new markets, which in some cases had started in the early 1960s but burgeoned in the 1970s, again involving more competition between financial institutions and narrower margins on the business of financial intermediation, together with a process of ‘spectrum-filling’ in which a more complete range of assets and liabilities (notably in terms of maturities) was made available. In the domestic capital markets, a number of new instruments were introduced and a number of new specialised markets established, largely inspired by developments in the US or on the international capital markets, while the Big Bang of 1986 transformed the London Stock Exchange into a more open, efficient and internationally competitive market. Overall, these developments provided greatly improved facilities for the hedging of various risks, increased competition and lower transactions costs. Finally, the international capital markets in London saw the growth of a range of financial instruments in a development characterised by the processes of securitisation, spectrum-filling, ‘unbundling’ and ‘repackaging’ of financial instruments, and globalisation. The causes of financial innovation can be classified under four headings. First, the relevant environmental factors include the increased volatility of asset and commodity prices in the 1970s and 1980s; the larger sectoral deficits/surpluses experienced in that period, particularly those of the public sector; the greater emphasis by monetary authorities on the non-monetary

16 This section does not purport to give a comprehensive account of financial innovation. For further discussion see, in particular, the Cross Report (Bank for International Settlements, 1986), Mayer (1986) and the other papers in the Winter 1986 edition of the Oxford Review of Economic Policy, and Llewellyn (1991).

38 The Making of Monetary Policy in the UK, 1975–2000 Table 3.3 Main financial innovations in the UK in the 1970s and 1980s (classified by main area of direct impact, with date of major expansion/change) A. Banks (i) end of interest rate cartel—1971 (ii) adoption of liability management—early 1970s (iii) company sector: medium-term variable rate loans, improved terms and interest rates on large deposits, more competitive charges—1970s (iv) personal sector: cheque and credit cards, personal fixed-term loans, more competitive charges—1970s (v) ATMs—late 1970s, early 1980s (vi) banks’ entry to mortgage market—1981–82 (vii) interest-bearing retail sight deposits—1983–84 (viii) debit cards—1987 B. Building societies (i) improved terms and rates of deposits—late 1970s (ii) ATMs—early 1980s (iii) end of interest rate cartel—1983 (iv) use of wholesale money markets as source of funds—domestic 1983, eurocurrency 1986 (v) unsecured personal loans—1986 (vi) transfer to bank/plc status—1988–89 C. Money market (i) parallel markets: local authority, CDs, finance houses, interbank—from 1960s (ii) eurocurrency deposits—from 1960s (iii) reduction of privileges of discount houses—1980s (iv) sterling commercial paper market—1986 D. Domestic capital markets (i) London Traded Options Market—1978 (ii) gilt-edged: new varieties and issue techniques—from late 1970s (iii) London International Financial Futures Exchange—1982 (iv) Stock exchange: end of dual capacity and minimum commissions, opening up of membership, move to screen market—1986 E. International capital markets in London (i) eurocurrency markets—from 1960s (ii) syndicated credits—mid-1970s (iii) foreign exchange markets (major expansion)—mid-1970s (iv) fixed rate eurobonds (origin 1960s)—early 1980s (v) international equity trading—early 1980s (vi) swaps market—1982 (vii) floating rate notes (origin early 1970s)—1982–83 (viii) euronotes—1984 (ix) eurocommercial paper—1986

financing of budget deficits; and the fall in the credit ratings of major banks in the early 1980s as a result of the developing country debt crisis. Second, under the heading of deregulation there were a series of domestic policy measures from Competition and Credit Control in 1971 to the Building Societies Act of 1986, whose effect was to promote competition within and between different elements of the financial system, to desegment the system and to eliminate all direct quantity controls. On the external front deregulation included the ending of exchange controls in 1979, which was crucial to the transformation of the City of London and the reunification of what Plender and Wallace (1985) called the ‘two Cities’—domestic and international—which had existed before.

Monetary Targeting, 1977–1986 39 Third, the category of technological developments includes the introduction of electronic screens to create larger, unified and more fast-moving financial markets where transactions costs were lower, in foreign exchange, eurobonds and equities; the introduction of automated teller machines (ATMs) which lowered transactions costs (especially non-pecuniary ones) for the retail customers of banks and building societies; and the introduction of computers to provide the essential calculational back-up for the new markets in financial futures, options and swaps. Fourth, increased competition was in nearly every sphere both a cause and an effect of financial innovation. The effects of financial innovation were obviously wider, but the present concern is with the consequences for monetary targeting. Here the most important effects were the reduction in the costs of financial intermediation (roughly, the difference between the interest charged to a customer on a loan and the interest received by a customer on a deposit) and the reduction in credit rationing (particularly for households). As the Loughborough Lecture (BEQB 1986d: 499–507) argued, these could be expected to lead to an increased demand by companies and households for both debt and liquidity; the latter would include a rise in the demand for money. A second important effect was that the boundary between banks and building societies became increasingly blurred, with the banks entering the mortgage market and building societies offering deposit terms and other services comparable to those provided by banks; this in turn raised questions about the use in broad money targeting of an aggregate which included bank but not building society deposits. As noted in Section 3.1, the UK authorities abandoned £M3 in favour of M4 (which includes the latter) in 1989 when the conversion of the Abbey National building society (one of the two largest) from mutual to limited liability bank would have involved a substantial upwards shift in £M3, but the transition could have been justified earlier. A third effect was that monetary control became more difficult, since the rise in the proportion of financial intermediaries’ deposits that were interest-bearing would have reduced the interest-elasticity of the demand for money (broad money in particular).17 In addition, deregulation and liberalisation made it practically impossible to go back to the kinds of direct controls used in the 1960s. However, it should be noted that the financial innovations in the bank and building society sectors (sections A and B of Table 3.3) were not the only ones that were relevant to monetary targeting. In particular, the growth of the parallel money markets (section C) was an essential prerequisite for the switch by banks (and later building societies) from asset management to liability management. Developments in the domestic and international capital markets (sections D and E) both increased the competitive pressures on banks (through the disintermediation involved in securitisation) and made it easier for the banks to cover themselves against risk, so that they were both encouraged and enabled to offer financial services at lower costs to their customers. And many of the new financial instruments—futures, options and swaps— should be seen partly as ‘intermediate’ products that contributed to the ‘final’ products such as eurobonds which were important elements of the changed financial environment faced by companies, households and financial intermediaries. A lecture by the Bank of England Governor in 1984 had discussed structural change and its implications for monetary targeting, claiming that ‘[t]he relationships between the various [monetary] aggregates, and between them and nominal income, have been subject to considerable variation and uncertainty from year to year’ (BEQB 1984d: 476), and had raised the 17 Insofar as potential borrowers considered the cost of borrowing relative to the gain on deposits, it would also have reduced the interest-elasticity of the demand for advances.

40 The Making of Monetary Policy in the UK, 1975–2000 4 £M3 1963-81 £M3 1981-88 M4 1963-82 M4 1982-2000 3

2

1

0 63Q1 65Q1 67Q1 69Q1 71Q1 73Q1 75Q1 77Q1 79Q1 81Q1 83Q1 85Q1 87Q1 89Q1 91Q1 93Q1 95Q1 97Q1 99Q1

Figure 3.1 Broad money velocity (Sources: see notes to Table 3.4)

possibility that at some stage these problems could become so large as to imply an abandonment of monetary targeting such as had already occurred in Canada (1984d: 477). In the 1986 Loughborough Lecture, the Governor discussed the impact of structural change in more detail and announced that the Chancellor and he would be considering whether the unpredictability of the relationship between money and nominal incomes had increased to the point at which it would be better to abandon monetary targets (no broad money target was set in the subsequent Budget). However, in neither of these cases was any serious evidence on the unpredictability of the relationships presented, and the emphasis of the Loughborough Lecture was largely on the fact that velocity was now trending down instead of up (with the argument that insofar as this reflected pressures from the demand rather than the supply side it did not have the same implications for inflation).18 For monetary targeting, what is important is not so much the direction of the trend in velocity as its predictability, and its predictability over the relatively short period of a year for which the target was set. Figures 3.1 and 3.2 show the velocity of £M3, M4 and M0 on quarterly data, while Table 3.4 gives the average percentage changes over one and over four quarters in the velocity of these monetary aggregates for various periods, together with the standard deviation of the changes. The velocity of broad money (£M3) was rising in the 1960s and 1970s, with the exception of the 1971–73 episode which many observers have identified as a supply-driven reduction,19 but it began to fall in 1980 and it fell consistently through the rest of the 1980s. M4 followed broadly similar trends in the 1970s and 1980s, and was roughly constant after 1990. The velocity of M0, on the other hand, rose throughout the 1970s and 1980s to a peak in 1993, and then declined gently for the rest of the 1990s. 18 Burns (1988) also provides data on the average change in velocity over various periods, but gives no data on the variance of velocity changes. 19 See in particular Artis and Lewis (1976, 1991: chapter 4). Table 3.4 excludes the mid-1970s for the two broad aggregates for this reason.

Monetary Targeting, 1977–1986 41 35

25

15

5 69Q3 71Q3 73Q3 75Q3 77Q3 79Q3 81Q3 83Q3 85Q3 87Q3 89Q3 91Q3 93Q3 95Q3 97Q3 99Q3

Figure 3.2 M0 velocity (Sources: see notes to Table 3.4)

Table 3.4 shows that the standard deviation of one quarter changes in the velocity of £M3 was lower for the years 1977–79 than in 1963–71, and lower again in 1980–85, when monetary targeting was supposedly becoming more difficult in the presence of structural change; for the four quarter changes the standard deviation rose between 1963–71 and 1977–79, but fell back

Table 3.4 Changes in velocity

Monetary aggregate

£M3 M4

M0

Period

Average change over 1 quarter (%)

Standard deviation of change over 1 quarter

Average change over 4 quarters (%)

Standard deviation of change over 4 quarters

1963 Q1 to 1971 Q4 1977 Q1 to 1979 Q4 1980 Q1 to 1985 Q4 1963 Q1 to 1971 Q4 1977 Q1 to 1979 Q4 1980 Q1 to 1985 Q4 1986 Q1 to 1990 Q4 1991 Q1 to 2000 Q4 1969 Q3 to 1993 Q4 1994 Q1 to 2000 Q4

0.36 1.03 −0.84 −0.19 0.51 −1.22 −1.62 −0.21 1.00 −0.29

1.71 1.57 1.13 1.40 1.40 1.18 0.81 1.23 1.51 0.80

1.22 4.10 −2.51 −0.95 2.08 −4.11 −6.11 −0.99 4.09 −0.98

2.65 3.99 3.06 2.42 3.08 3.39 1.21 2.52 2.98 1.23

Sources: The velocity figures for £M3 are calculated from data on £M3 and GDP (average estimate), both seasonally adjusted, in Economic Trends Annual Supplement, 1989 edition; there are breaks in the series in 1981 Q4 and 1983 Q1. The velocity figures for M4 for 1963 Q1 to 1988 Q2 are calculated from the same source, while those from 1982 Q3 to 2000 Q4 are those given in Economic Trends Annual Supplement, 2000 edition, updated for the last two years from Financial Statistics, October 2001.

42 The Making of Monetary Policy in the UK, 1975–2000 in 1980–85.20 The variability of the velocity of M4 was typically lower than that for £M3; for the one quarter changes it was the same in 1977–79 as in 1963–71 and lower in 1980–85, while for the four quarter changes it rose in successive periods. For both one and four quarter changes the standard deviation then fell sharply in the second half of the 1980s but returned in the 1990s to around the level of the 1960s. The velocity of M0 was much less variable after 1993. Even if the likely impact on velocity of variations in the interest rate is ignored,21 these data are not obviously consistent with the claim that the relationship between the targeted monetary aggregate and nominal income had become more variable and more uncertain in the heyday of monetary targeting in the first half of the 1980s. On the other hand, the standard deviations of the change in velocity are generally high relative to the (4%) width of the target ranges.

3.5 TARGETS AND OUTTURNS FOR MONETARY GROWTH It will now be useful to go through the ten years of monetary targets in turn noting, first, whether the particular target chosen was reasonable in the light of the analysis of Section 3.2, that is, consistent with the authorities’ economic and velocity forecasts; second, whether it was feasible in the light of the analysis of Section 3.3, that is, consistent with their (estimated) counterpart forecasts; and third, what, if any, other factors may have influenced the authorities’ choice of target. The discussion then identifies as far as is possible the extent to which (i) the way in which the targets were set and (ii) unanticipated financial innovation underlie any deviation of the outturn of monetary growth from the target. To analyse the causes of deviations the credit counterparts framework will be used. For present purposes it is assumed that, before the abandonment of overfunding in 1985, the authorities were able to exercise a rough and medium term control of monetary growth from the supply side. Moreover, with the exception of the first six months of 1977/78 the exchange rate was floating, so that the demand for money had to adjust to the supply via changes in nominal income and interest rates. This means that the problem posed for the authorities’ ability to hit their monetary targets by structural change and financial innovation can be thought of in terms of the direct impact of the latter on the credit counterparts, and the impact on the authorities’ decisions on supply-side monetary control. The discussion of particular years will therefore consider, for example, whether a particular overshoot of the monetary target was due to an excessive growth of bank lending to the private sector (not offset by increased sales of public sector debt) which can be related to deregulation and innovation; or whether an overshoot was the result of a deliberate choice by the authorities who were responding to evidence that the monetary stance was tighter than the broad money figures suggested (e.g. because of an upwards shift in the demand for broad money). For 1977/8 the target was clearly reasonable in relation to the money GDP forecast and the authorities’ likely expectation about velocity; and it was feasible in terms of their expectations 20 Contemporary data are only slightly more sympathetic to the official case: for the one quarter change in the velocity of £M3 the standard deviation rises from 1.32 for 1977 Q1 to 1979 Q4 (data from Financial Statistics, January 1983) to 1.57 for 1980 Q1 to 1985 Q4 (data from Financial Statistics, July 1986); but for four quarter changes the standard deviation falls from 3.22 to 2.52 (the latter for 1981 Q1 to 1985 Q4, the nearest period available in the later issue). 21 Artis and Lewis (1984) in a long run study of the demand for money which used an unusual monetary aggregate (for which data could be obtained only up to 1983) found that the velocity of this aggregate for the years 1980–83 was very close to the levels predicted by a regression of the data on the gilt yield for 1920–57 (while the observations for 1973–76 were well away from the curve).

Monetary Targeting, 1977–1986 43 Table 3.5 £M3 targets and outturns Target period April 1977 to April 1978 April 1978 to April 1979 June 1979 to April 1980 February 1980 to April 1981 February 1981 to April 1982 February 1982 to April 1983 February 1983 to April 1984 February 1984 to April 1985 1985/6 (average of 12-month growth rates)a 1986/7 (average of 12-month growth rates)b

Target range

Outturn over target period

Outturn over financial year

9–13% 8–12 7–11 7–11 6–10 8–12 7–11 6–10 5–9

16.0% 10.5 9.6 19.1 13.7 11.1 9.5 11.9 13.8

15.6% 11.4 12.5 18.5 14.3 11.7 8.2 11.5 16.7

18.3

19.0

11–15

Source: outturn figures from BEQB, Financial Statistics, various issues. Notes: a Target suspended in October 1985; outturn for target period is average of 12-month growth rates from May 1985 (when this method was introduced) to April 1986; b Outturn for target period is average of 12-month growth rates from March 1986 to April 1987.

for the counterparts. The 9–13% range was above the previous period outturn of 9.7%, but that had been depressed by the extraordinary gilt-edged sales of 1977 Q1. Strictly the 9–13% range was not originally a target, but what the authorities expected to be consistent with the formal target for DCE agreed with the IMF in December 1976; the same range had also been given then as consistent with the (higher) DCE target for 1976/7 (BEQB, 1977a: 16; 1977b: 148–9). However, the estimates in Table 3.2 imply that the authorities’ own forecasts for DCE in both 1977/8 and 1978/9 were well below the DCE limits agreed with the IMF.22 And in any case, the 9–13% range rapidly assumed the status of a target as the balance of payments improved and attention was shifted away from DCE on to the money supply. As it turned out (see Table 3.5) the 1977/8 target was substantially exceeded, with a large undershoot on the PSBR, but an even larger overshoot (relative to the estimated forecasts in Table 3.2) on external and foreign currency counterparts (an overshoot only partly offset by unplanned overfunding). However, this was due essentially to the capital inflow associated with the conflict between monetary and exchange rate targets that developed in the autumn of 1977; the authorities chose the monetary target over the exchange rate target but they did not do enough to undo the monetary expansion that had occurred in the summer and autumn of 1977. Thus the overshoot of the target should be attributed to a weakness in control rather than to any deficiency in target-setting or to unanticipated financial innovation. For 1978/9 the target range of 8–12% was rather high in relation to the money GDP forecast, which was more than 2% lower than that for the previous year; it was clearly feasible in terms of the authorities’ counterpart forecasts. The 1978 Budget was a moderately expansionary one; it included the introduction of rolling monetary targets with the suggestion that a lower target might be set in the autumn (in the event it was not), while the 8–12% range was said to give ‘ample room’ for the expected increase in bank lending to the private sector.23 In fact, £M3 22 Those limits were £7.7bn and £6bn. In terms of Table 3.2 the authorities’ forecasts for DCE are given by the sum of rows (3) + (5) + (11) + (15). 23 Budget Statement, HC, vol. 947, cols 1191–2.

44 The Making of Monetary Policy in the UK, 1975–2000 grew within and just above the mid-point of the target range. The largest divergence (from the estimated forecasts) was that in bank lending, which occurred despite the reimposition of the corset in June 1978 in response to an unexpected spurt early in the year and seemed to have been demand-driven (BEQB, 1978b: 169; 1979b: 123). This divergence was only partly compensated by increased debt sales and an unexpected negative flow on externals. For 1979/80 the target range set by the new government in its first Budget in June 1979 was again 1% lower than that for the previous year, but money GDP was expected to grow some 3% faster (largely because of the VAT increase) and the implied velocity change is not compatible with what the authorities probably believed at the time: thus the authorities, or at least the ultimate decision-makers, must either have not believed the official inflation and GDP forecasts, or have chosen to disregard the inconsistency.24 On the other hand, the target looked feasible in terms of the counterpart forecasts. The outturn was within the range for the target period, partly thanks to the disintermediation associated with the corset,25 but above it for the financial year which included an acceleration of monetary growth in April and May 1979. In terms of the estimated forecasts (for the financial year) there were large overshoots on the PSBR and bank lending to the private sector (both related to the deepening recession), which were only partly compensated by higher than planned debt sales and a larger than expected outflow on externals. The 1980 Budget saw the introduction of the MTFS. The target range for 1980/81 was kept at 7–11%, the same as that for 1979/80, while the money GDP forecast was also little changed from the previous year, so that the target remained incompatible with the authorities’ likely expectations about velocity.26 On the other hand, the target again appeared feasible in terms of the counterpart forecasts in Table 3.2. A curious feature of this year is that the Chancellor envisaged in his Budget Statement some upward pressure on £M3 in the summer of 1980 from re-intermediation following the abolition of the corset in June 1980 which ‘cannot be precisely measured or predicted’, but which he hoped could be accommodated within the target range;27 however, the BEQB made no mention of this phenomenon until after it had occurred. In the event, £M3 grew well above the target range, at a time of deepening recession associated, in particular, with an increasingly overvalued exchange rate. There was a very large overshoot on the PSBR, a much smaller overshoot on bank lending (partly due to post-corset re-intermediation but partly related to the financial squeeze on the corporate sector), and a small external inflow instead of an outflow; these were only partly offset by higher than planned debt sales. However, it is clear that the authorities decided to allow a large overshoot, on the grounds that monetary conditions were already tight enough, if not too tight.28 The overshoot should therefore be attributed primarily to the setting of a non-reasonable target, which the authorities recognised implicitly in the course of the year. For 1981/2 the inflation and money GDP forecasts were much lower so that the lower target range of 6–10%, as set out in the 1980 version of the MTFS, was a more reasonable one,

24 There is some evidence in favour of the former possibility in the 1979 Budget Statement, HC, vol. 981, cols 243, 263, quoted in Atkinson and Hall (1983: 162–3). 25 By 1980 Q1 £M3 velocity was more than 8% higher than in 1979 Q1. 26 Burns (1988) contains an internal Treasury forecast for money GDP growth in 1980/81 of 17.1%. On that basis, as opposed to the 131/2% in Table 3.1, column (3), the £M3 target was even more incompatible with the authorities’ likely expectations about velocity. 27 HC, vol. 981, col. 1444. 28 Interest rates were reduced in July and November 1980. See BEQB, 1980d: 406–7, 1983b: 205; also FSBR 1981–82, Part II.

Monetary Targeting, 1977–1986 45 at least if the authorities were still expecting velocity to rise.29 In terms of their view of the likely development of the counterparts it was also a feasible target. However the £M3 target was overshot by a large margin: the PSBR was now below but bank lending well above the estimated forecasts, again partly due to the financial squeeze. Debt sales were above target despite the PSBR undershoot, and there was large overfunding on the year. However, as in 1980/81 the overshoot was essentially allowed by the authorities, who were by now becoming convinced of the importance of structural changes in the financial system tending to raise the demand for money and reduce the velocity of circulation.30 In addition, the high level of bank lending was related to the banks’ entry into the mortgage market, following the abolition of the corset. Thus the £M3 overshoot in this year can legitimately be classified as due to unanticipated financial innovation. This view of financial innovation (together, perhaps, with the recorded overshoots of 1980/81 and 1981/2) led the authorities to revise upwards the 1982/3 target range from the 5–9% set out in the 1980 and 1981 versions of the MTFS to 8–12%. They were probably no longer expecting the velocity of circulation to rise and may well have been expecting it to decline. But given the explicit money GDP forecast (published for the first time) of 9.8%, the new target range was consistent only with a constant or very slowly declining velocity. In addition, the authorities’ likely forecasts of the counterparts suggested a growth of £M3 towards the top of the target range. However, the 8–12% could at least be presented as substantially lower than the 1981/2 outturn;31 and it was no higher than the last target range of the previous government. In the event £M3 grew within but towards the top of the target range, by an amount close to the total of the estimated forecasts. There was an overshoot on bank lending and a very small undershoot on the PSBR, but debt sales were below forecast (giving net underfunding) and external flows were more negative than the estimated forecast. The banks were now consolidating rather than further expanding their position in the mortgage market, and on the money demand side there was a movement out of bank time deposits into term shares with building societies (BEQB, 1987b: 216). For 1983/4 the further fall in the inflation and money GDP forecasts (particularly the explicit money GDP forecast) made the 7–11% range envisaged in the 1982 version of the MTFS look more reasonable, even if the authorities were now clearly expecting some decline in velocity. That range also looked feasible in terms of the likely growth of the counterparts. In fact, £M3 grew within and near the centre of the range (above the mid-point over the target period, but below it on the financial year). Bank lending and the PSBR were both above forecast, but this was offset by higher debt sales, and there were offsetting divergences in externals and non-deposit liabilities. The banks were still consolidating their position in the 29 There is nothing in FSBR 1981–82 that indicates any other expectation, and FSBR 1982–83 refers to ‘the past upward trend in velocity’. Burns (1988) includes an internal Treasury forecast for money GDP growth in 1981/2 at 10.5%, on which basis the target was less clearly reasonable. 30 BEQB, 1982a: 21; see also FSBR 1982–83: 14: ‘The demand for liquid balances as a medium for saving, rather than spending, seems to have increased significantly in the last three years, implying a shift in velocity,’ and BEQB, 1982a: 6. 31 The following passage from Bruce-Gardyne (1986: 186–7), suggests that this kind of consideration was indeed important: ‘The last update of the MTFS, at the time of the previous Budget, might have pencilled in, for the year confronting us, a range of 6–10 per cent growth in our £M3 lodestar . . . And that might, come February of the following year, look too ambitious for our comfort. But if we could by then assert—and we usually could—that in the financial year drawing to its close £M3 looked like growing by 13 per cent, instead of the 7–11 range set for it . . . then we could make a virtue of our iniquities. Brushing aside our past intentions, we would boldly repeat the previous year’s 7–11 per cent range for £M3 . . . and emphasise the progress this would constitute not against the previous year’s intentions (which would be assigned to the memory hole) but against our latest expectations of the previous year’s performance.’

46 The Making of Monetary Policy in the UK, 1975–2000 mortgage market, and there was a further shift into building society term shares (BEQB, 1987b: 216). For the following year, 1984/5, the implicit and explicit money GDP forecasts were rather different; if the authorities expected velocity to fall by around 2% the 6–10% range set out in the 1982 and 1983 versions of the MTFS was barely reasonable on the former and not on the latter. On the basis of the estimated forecasts the target range was feasible (though they imply monetary growth clearly in the upper half of the target range). £M3 was within the target range as late as February 1985 (BEQB, 1985a: 25), but ended up (for both the financial year and the target period) above the target range. The divergence was mainly the result of large overshoots on bank lending to the private sector (particularly to firms) and the PSBR which were not fully offset by increased debt sales (although there was considerable net overfunding). The BEQB described the high level of bank lending as ‘puzzling in view of the apparent strength of corporate finances and in particular the recent very high rate of new equity issues’, but did not attribute it to any obvious form of financial innovation.32 This suggests that the monetary overshoot should be attributed primarily to short-term weaknesses in monetary control. The 1985/6 Budget was delivered under the shadow of the most serious sterling crisis since 1976. With money GDP expected to accelerate slightly, the 5–9% target range previously envisaged would have looked non-reasonable even if the authorities had expected velocity to remain constant, but by now they clearly expected it to fall. They also expected a continuing high level of bank lending to the private sector, and that was largely responsible for the major inconsistency between the probable development of the counterparts and the 5–9% target range for £M3 set in the previous year’s MTFS. In a lecture in October 1984 the Governor of the Bank of England had emphasised the need for pragmatism and discretion rather than mechanical rules in operating monetary policy (BEQB, 1984d: 474–81). Nevertheless, the 5–9% target was confirmed in the Budget. The obvious explanation is that the authorities were worried that, in the wake of the January 1985 sterling crisis, a failure to confirm the target range specified in the previous version of the MTFS would upset the financial markets. In the event, the PSBR was below forecast, but debt sales were even further below. However, bank lending was well above, so that £M3 growth was much higher even than the estimated forecasts in Table 3.2. The monetary overshoot was clear from early in the year, and was attributed by the authorities to structural changes within the monetary system (BEQB, 1984d: 474–81). The £M3 target was formally suspended in October 1985. It was announced at the same time that the technique of overfunding would no longer be used, though it seemed to have already been abandoned; for the financial year as a whole there was net underfunding (even in terms of the ‘wide’ funding notion adopted in October 1985). For this year, then, the target range was a priori non-reasonable and the estimated forecasts implied monetary growth well above the target range; in the event monetary growth turned out to be even higher than those forecasts. The overshoot can be attributed to a combination of an inappropriate target and unanticipated financial innovation. The 1986 Budget reintroduced a £M3 target at the much higher range of 11–15%. This made a generous allowance for a large fall in velocity,33 and looked feasible in terms of the 32 The authorities noted that the increase in bank lending ‘included a substantial amount of lending to borrowers wishing to take advantage of the capital investment allowances for 1984/85 before the lower rate for 1985/86 became effective. This will have exaggerated the short-term rate of growth of bank lending but the underlying pace remains uncomfortably high . . . ’ (BEQB, 1985b: 186; see also 1985a: 26). 33 The FSBR 1986–87 described the raised target range as ‘reflecting the rapid fall in velocity in recent years’, and added, ‘Illustrative ranges for future years are not given for £M3 because the uncertainties surrounding its velocity trend are at present too great.’ (12)

Monetary Targeting, 1977–1986 47 counterparts. But in the event it was greatly exceeded: there was a further undershoot on the PSBR as economic growth began to accelerate, but bank lending expanded even more rapidly than expected and the authorities could no longer offset it by overfunding. The £M3 and bank lending overshoots were seen by the authorities as resulting largely from structural change including increased competition between banks and building societies, though the FSBR for 1987–88 also mentioned high real interest rates and the end of overfunding.34 Given that there was marginal overfunding on the new wide definition, it can be argued that the overshoot was partly due to the absence of effective instruments of monetary control once overfunding had been abandoned, as well as to unanticipated financial innovation. The broader macroeconomic background, as discussed in Chapter 2, was that monetary targeting had been becoming more discretionary and less rule-based from 1981, with more and more attention to interest rates rather than monetary aggregates (interest rate decisions are discussed in more detail in Chapter 8). By 1985 the authorities’ focus was clearly elsewhere: in late 1985 (see Chapter 4) they embarked on a strategy of using the oil price fall to stimulate aggregate demand without risk of inflation, and by late 1986 they were moving towards an even greater emphasis on exchange rate stability, so that their attachment to the monetary targets in the last two years was far from wholehearted.

3.6 MONETARY GROWTH, TARGET-SETTING AND FINANCIAL INNOVATION It is now possible to draw out a number of conclusions from the preceding sections. First, it is clear from the analysis of Section 3.2 (Table 3.1) that the severity of the targets, which can be taken as a measure of the restrictiveness of monetary policy, varied considerably from year to year: they were not systematically related to previous money GDP growth or to previous £M3 growth, even if probable changes in the authorities’ expectations about the trend of velocity are taken into account. The targets were particularly severe in 1979/80, 1980/81 and 1985/6, and relatively lax in 1978/9. Second, it is also clear from Table 3.1 that some of the targets promulgated were inconsistent with the authorities’ inflation and money GDP forecasts, given their beliefs about velocity, in particular those of 1979/80, 1980/81 and 1985/6. Third, the analysis of Section 3.3 (Table 3.2) indicated that some of the targets promulgated cannot have been regarded as attainable by the authorities themselves: although most of the targets were roughly consistent with what the authorities expected in terms of the counterparts, their expectations for 1982/3 were consistent only with monetary growth at the top of the target range and those for 1985/6 only with monetary growth well above the target range. Why then did the monetary authorities choose the particular ranges set? The discussion in Chapter 2 (Section 2.2) emphasised three objectives for the introduction of monetary targets: (a) announcing monetary targets should assist the authorities to put increased emphasis on price stability rather than other goals; (b) monetary targets should give clearer information to the private sector on the authorities’ intentions, and thereby help the private sector to make more accurate and less uncertain economic forecasts; and (c) announced targets might exert an independent influence on inflation expectations. While the first objective does not constrain the target ranges, it is difficult to argue that in choosing them the authorities were 34 See

also the article on ‘Measures of broad money’, BEQB, 1987b: 212–19.

48 The Making of Monetary Policy in the UK, 1975–2000 trying to provide clear information or to influence inflation expectations: there is no systematic relationship between the money GDP forecasts and the target ranges or between the official inflation forecasts and the target ranges in Table 3.1. Instead, the most important regularity in the choice of target ranges seems to have been the ‘n − 1’ principle, under which successive target ranges were set 1% lower. The ‘n − 1’ idea was first introduced with the 1978/9 target but it was followed in 1979/80 and became firmly embodied in each version of the MTFS: the authorities were concerned to show a continuous tightening of policy, primarily with respect to the previous year’s target but also on occasions with respect to the previous year’s outturn.35 The choice of 7–11% in 1979/80—a range which was thoroughly non-reasonable in the light of official expectations at the time—can be understood in this light; similarly the even more inappropriate choice of 5-9% for 1985/6, and the failure to set the range for 1980/81 above that for 1979/80. The ‘n − 1’ principle is, of course, essentially arbitrary and does not provide a sound discipline for monetary policy over the medium term: it did not preclude the expansion of 1978/9, for example, and in the absence of the fall in velocity would not have precluded periods of greater expansion in the 1980s, for example in 1983. For the ten years of broad money targets in the UK there were substantial overshoots in six and no undershoots (Table 3.5). In principle there are two kinds of reasons why the monetary authorities might not attain a monetary target. First, they may not be fully committed to it: they may change their mind during the course of the year as to what monetary growth is appropriate, or they may be unwilling to take the steps needed to attain the target. Second, they may experience technical difficulties in making monetary growth take place at the desired rate: their instruments may be inadequate, or the process may be affected by unforeseen events or forces. Of the six overshoots, those of 1977/8 and 1984/5 (the two smallest overshoots) were the result of technical difficulties in the form of weaknesses in short-term monetary control; that of 1980/81 (the largest overshoot) was primarily due to the non-reasonableness of the target set (which the authorities recognised in the course of the year and decided not to try to attain); and that of 1981/82 (the third largest overshoot) was due to technical difficulty in the form of unanticipated financial innovation. For 1985/6 (the second largest overshoot) unanticipated innovation was important, but the target set had been both non-reasonable and non-feasible and the authorities seem to have made very little attempt to hit it, even before the decision to suspend the target in October.36 The target for 1986/7, on the other hand, was not obviously non-reasonable or non-feasible, and the overshoot was due to technical difficulties in the form of unanticipated financial innovation and inadequate instruments of control. Of the four years where the outturn for £M3 was within the target range, the 1979/80 range was clearly non-reasonable, but the resulting build-up of pressure led to a change in the official commitment only in 1980/81. The 1982/3 target was nearly non-reasonable and nearly nonfeasible; monetary growth was kept within the target range but it was near the top of the range and at a level consistent with what would have been more clearly reasonable and feasible. And the outturn for 1983/4 was the closest for any of the ten years of targeting to the midpoint of the target range, which had been both reasonable and feasible.

35 See the quotation from Bruce-Gardyne (1986) in footnote 31. 36 £M3 growth was ‘well above its target range’ early in the year,

and the authorities remarked that ‘[T]he further decline in the velocity of broad money has exceeded earlier expectations, and may be associated with continuing changes in the structure of the financial system. In these circumstances the behaviour of the exchange rate . . . necessarily assumed a somewhat greater importance as an indicator of monetary conditions’ (BEQB, 1985c: 344).

Monetary Targeting, 1977–1986 49

3.7 CONCLUSION: THE ABANDONMENT OF MONETARY TARGETS The official explanation for the abandonment of monetary targets was that structural change and financial innovation had made the relationship between money and nominal income too unpredictable. There is no doubt that there was considerable structural change and financial innovation in the first half of the 1980s. However, there is no good evidence of an increase in the unpredictability of velocity over the short term that is relevant for monetary targeting, and if the unpredictability of velocity was high enough for monetary targeting to be inappropriate in the 1980s, it must also have been high enough in earlier times as well. Moreover, although it is clear that unanticipated financial innovation played an important role in some of the overshoots of the monetary targets, it was not responsible for all or even most of them. The obvious alternative explanation for the demise of monetary targets is that the authorities had operated them in ways that were inefficient, that resulted in repeated overshoots and were therefore responsible for their becoming discredited. It is clear that the authorities’ rather arbitrary procedures for setting the target ranges contributed to many of the overshoots, while weaknesses in monetary control were an additional factor. Moreover, the authorities could have operated the targets differently: they could have set the targets on a more systematic basis, ensuring that they were both reasonable and feasible, but avoided adverse short-term effects on confidence due to the targets being more variable and not ‘n − 1’ by providing much more information in advance on the reasons for their choices, including their expectations about the likely development of velocity as well as of money GDP;37 the non-fulfilment of these expectations could then have been used to explain any deviations from the targets. And by switching earlier from £M3 to M4 they would have been able to eliminate the problems posed for policy by trying to predict the swings in competition between banks and building societies.38 However, there are several other factors that contributed to the abandonment of targets. First, there had always been some doubts in monetary policymaking circles (perhaps particularly within the Bank of England) about the rationale for broad money targets, doubts focused on the arbitrary nature of any definition of ‘money’. Such doubts were encouraged and legitimated by the genuine structural changes that were taking place, even if velocity had not become less predictable over the short term. Second, the rationale for targeting broad money had been put into question, initially by the distortions induced by repeated reliance on overfunding as the main technique for offsetting the uncontrollable element of bank lending (and the growing conviction that such control was purely cosmetic), and then by the abandonment of that technique in 1985. Third, opinion in monetary policymaking circles had shifted, through the sterling overvaluation of the early 1980s and the attempts to unwind it (see Chapters 8 and 9), towards a much greater interest in exchange rate stability (which was generally interpreted as ruling out monetary targets). The connections between different financial markets and asset prices then encouraged policymakers to view the interest rate as the focus of monetary policy, in a way which was consistent with the downgrading of monetary aggregates and which emphasised the one definite monetary instrument that the authorities now possessed. 37 This would also have allowed them to set narrower target ranges. 38 Moreover, under this sort of procedure the monetary targets would have provided more information to the private

sector and, because the authorities’ credibility would have been higher and more sustainable over the long term, they might have had more effect on inflation expectations.

50 The Making of Monetary Policy in the UK, 1975–2000

APPENDIX 3.1 MONETARY TARGETS AND OUTTURNS Table 3A.1 Monetary targets, 1976–1979 Targeted aggregate M3 (%) DCE (£bn) DCE (£bn) DCE (£bn) £M3 (%) £M3 (%) £M3 (%) £M3 (%) £M3 (%) £M3 (%)

Target period

Target range

Outturn

1976/7 1976/7 April 1977 to April 1978 April 1978 to April 1979 April 1976 to April 1977 April 1977 to April 1978 April 1978 to April 1979 October 1978 to October 1979 June 1979 to April 1980 June 1979 to October 1980

12 9 7.7 6 9–13 9–13 8–12 8–12 7–11 7–11

9.7 3.8 4.1 6.8 7.5 16.0 10.5 13.6 9.6 17.7

Sources: BEQB, Economic Trends, various issues.

Table 3A.2 Monetary targets for £M3 under the MTFS, 1980–1983

£M3 targets (%) set out in: FSBR 1980–81 FSBR 1981–82 FSBR 1982–83 FSBR 1983–84 Outturn

1980/81

1981/2

1982/3

1983/4

1984/5

1985/6

7–11

6 –10 6 –10

5–9 5–9 8–12

4 –8 4 –8 7–11 7–11 9.5

6–10 6–10 11.9

5–9 13.8

19.1

13.7

11.1

Note: target periods are February 1980 to April 1981 (at annual rate) et cetera. Sources: FSBR, BEQB, Economic Trends, various issues.

Table 3A.3 Monetary targets for M1 and PSL2 under the MTFS, 1982–1983 to 1983–1984

M1 targets (%) set out in: FSBR 1982–83 FSBR 1983–84 Outturn for M1 PSL2 targets (%) set out in: FSBR 1982–83 FSBR 1983–84 Outturn for PSL2

1982/3

1983/4

1984/5

1985/6

8–12

7–11 7–11 14.1

6–10 6–10 (17.2)*

5–9 (21.8)*

7–11 7–11 12.9

6–10 6–10 (15.8)*

5–9 (14.2)*

12.5 8–12 10.8

Note: * targets for M1 and PSL2 were abandoned in 1984; target periods are February 1982 to April 1983 (at annual rate) et cetera. Sources: FSBR, BEQB, Economic Trends, various issues.

Monetary Targeting, 1977–1986 51 Table 3A.4 Monetary targets for £M3 and M0 under the MTFS, 1984–1985 to 1986–1987

£M3 targets (%) set out in: FSBR 1984 –85 FSBR 1985–86 FSBR 1986 –87 Outturn for £M3 M0 targets (%) set out in: FSBR 1984 –85 FSBR 1985–86 FSBR 1986 –87 Outturn for M0

1984/5

1985/6

1986/7

1987/8

1988/9

6–10

5–9 5–9

3–7 3–7

2–6 2–6

11.9

13.8*

4–8 4–8 11–15 18.3

4–8

3–7 3–7

5.6

3.9

2–6 2–6 2–6 4.2

1–5 1–5 2–6 5.0

0–4 0–4 1–5 6.9

1989/90

1–5 5.9

Notes: * target for £M3 suspended in October 1985; target periods are February 1984 to April 1985 (at annual rate), then May 1985 to April 1986 (average of 12-month growth rates), then March 1986 to April 1987 (average of 12-month growth rates) et cetera.

Table 3A.5 Later targets and monitoring ranges for M0 and M4 Year M0 ranges (%) for: 1987/8 1988/9 1989/90 1990/91 1991/2 1992/3 1993/4 1994/5 1995/6 1996/7 M4 ranges (%) for: 1992/93* 1993/4 1994/5 1995/6 1996/7

Target or monitoring range

Outturn

2–6 1–5 1–5 1–5 0–4 0–4 0–4 0–4 0–4 0–4

5.0 6.9 5.9 4.4 2.2 3.0 5.2 6.6 5.8 6.7

4–8 3–9 3–9 3–9 3–9

3.6 4.1 5.1 8.4 10.1

Notes: target periods are generally March 1987 to April 1988 (average of 12-month growth rates) et cetera; * target applies only to second half of 1992/93 (November to April).

4 The Interregnum and the Lawson Boom, 1985–1990 From Chapter 3 it is clear that monetary targeting effectively ended in 1985, when the authorities set a target range in the Budget which was neither reasonable nor feasible, then suspended the target and announced the end of overfunding in October. A new target range was set for 1986/7 but the authorities no longer had adequate instruments with which to enforce it, and the target was substantially overshot. During the five and a half years from April 1985 to October 1990, when the UK entered the ERM, there was no clear and effective framework for monetary policy, although between March 1987 and March 1988 a serious attempt was made to peg the sterling–DM exchange rate. Before that there were two years of unattained (and unattainable) monetary targets, and after the DM-shadowing episode there were two and a half years of no formal or informal framework, while the authorities struggled to overcome the inflation unleashed by the boom of 1986–88. This chapter examines the conduct of policy during the subperiod in detail. Section 4.1 discusses six different explanations of the causes of the boom, rejecting four outright. Section 4.2 examines in greater depth the way in which macroeconomic policy decisions were taken and implemented over the second half of the 1980s. Section 4.3 assesses the two main rival explanations of the causes of the boom, and argues that they should be seen as partly complementary. Section 4.4 notes the results of an analysis of the monetary policy reaction function for this subperiod, and discusses how policy would have been different if it had been set in line with the reaction functions for later subperiods. Section 4.5 draws some conclusions.

4.1 EXPLANATIONS OF THE LAWSON BOOM Headline inflation as measured by the RPI rose temporarily to 7% in 1985 Q2, after fluctuating around 4.5 to 5% over 1983–84, but, by mid-1986, it had declined to below 3%. After hovering around 4% during 1987 and the first half of 1988, inflation rose once more and steeply to reach a new peak of 10.4% in 1990 Q3.1 This increase clearly lagged movements in output and capacity. GDP grew by between 3.8 and 5.2% from 1985 to 1988, but its growth was slower in 1989 and negative during the second half of 1990. Unemployment declined from a peak of 11.0% in mid-1986 (as against the previous low of 3.9% in late 1979) to a new low of 5.5% in early 1990.2 It seems clear, therefore, that the resurgence of inflation from 1988 was associated with an over-expansion of aggregate demand in the economy. A number of possible causes of 1 An initial peak of 8.2% in 1989 Q2 was followed by a small dip to 7.6% in 1989 Q4. The RPIX measure of inflation (which excludes mortgage interest payments and therefore the effect of rises in interest rates) peaked at 9.2% in 1990 Q4. 2 Data for UK claimant count, from Economic Trends Annual Supplement, 2000 edition.

54 The Making of Monetary Policy in the UK, 1975–2000 25.0

20.0

15.0

10.0

5.0

0.0 1983M7

1984M7

1985M7

1986M7

1987M7 M0

M4

1988M7

1989M7

1990M7

1991M7

£M3

Figure 4.1 Monetary growth (Source: Bank of England. Note: growth rates are percentage growth since 12 months before)

that over-expansion of demand—from the acceleration of monetary growth in the mid-1980s to the oil-price-related depreciation of 1986—will be discussed in turn.

4.1.1 Prior Monetary Expansion Some commentators have attributed the boom’s initiation to an acceleration of monetary growth from late 1985 (or early 1986), detected either at the time or somewhat later. For example, Congdon expressed concern about the apparent monetary and credit expansion in The Times both in October 1985 and January 1986. By 1988 he was arguing that the acceleration of credit and broad money displayed during early 1986 was ‘the inevitable consequence of the change in policy approach’ which had occurred during 1985 when the £M3 target for 1985/86 was suspended, the technique of overfunding was abandoned and the Chancellor ‘stopped paying much attention to credit and broad money’.3 However, this identification of the cause of the boom is not convincing. First, it is not clear that there was a significant acceleration of monetary growth during 1985–86. Table 2.2 shows that on annual data the growth of £M3 rose, reaching 21.1% in 1988 as against 9.3% in 1984. But the broader measure, M4, shows no increase in growth in 1985, an increase during 1986 which was partly reversed in 1987, and a further small increase during 1988–89; these increases took M4 growth only from 13.2% in 1984 to 17.5% in 1989. The growth of the narrow measure M0 fell in 1984, 1985 and 1986, and there was an appreciable acceleration only in 1988. Figure 4.1, 3 The 1985 and 1986 articles are reprinted in section 5 of Congdon (1992); the quotes are from a July 1988 article in The Times reprinted in section 7 (p. 156).

The Interregnum and the Lawson Boom, 1985–1990 55 which shows the 12-month growth rates of each series, reinforces the impression that, except for £M3, the acceleration of monetary growth was relatively small during 1985–86 and only became substantial, relative to recent previous years, during 1988–89.4 Moreover, the slower rise in £M3 during 1984–85, and its faster increase from late 1985, can be readily explained from the relevant credit counterparts in terms of the competition between banks and building societies. The banks had entered the mortgage market with enthusiasm in 1981 but their share of new mortgage lending fell back sharply in 1984. Their share recovered somewhat in 1985–86, and then increased strongly in 1987 and 1988 at the same time as that of other specialist mortgage lenders, so that the building societies’ share was squeezed.5 M4 includes deposits with both building societies and banks, therefore being unaffected by movements in the changing shares of these two groups of institutions. However, such movements make £M3 an unreliable guide to the underlying trend of monetary growth, as was recognised by the authorities in the Loughborough Lecture (BEQB, 1986d: 499–507) and the abandonment of the £M3 measure in 1989. In addition, it is worth emphasising that the movements in velocity associated with the 1985–88 period were not comparable with those of 1972–73. During the 1972–73 ‘Barber boom’ there was a sharp fall in the (then upward-trending) velocity of £M3 that was subsequently reversed. This suggested to many6 that a supply-side monetary expansion had pushed the private sector off its demand for money function in the short term. By contrast, the mid1980s saw broad money velocity falling only slightly more sharply than in the previous few years and it was not until 1991, by when the boom had long turned into recession and the monetary expansion had slowed dramatically, that the M4 velocity began to bottom out. Second, the problem of causality here is even more difficult than usual. When monetary targets were being pursued more seriously, say from 1977 until 1984, it was possible to argue that the authorities were aiming to control the growth of broad money over the medium term and were in possession of appropriate instruments. When overfunding was abandoned in 1985, the immediate effect was a minor apparent acceleration of monetary growth in the form of £M3 as some of the distortions associated with overfunding were unwound.7 But without overfunding the authorities had very little direct or immediate control over monetary growth, and the effect of interest rates on the demand for bank credit was known to be weak and heavily lagged.8 This means that monetary growth was much more obviously endogenous from 1985, that is responsive to, and determined by, changes in economic activity and other factors. It is therefore difficult to talk of an (exogenous) increase in monetary growth initiating a rise in aggregate demand.

4 Data for the growth of the real money supply (emphasised, for example, by Congdon, 1992: 156–7) show a sharper acceleration from early 1986, but this reflects the short-run fluctuations in inflation (a minor upturn in 1985 followed by a minor downturn in 1986) rather than decisive changes in monetary growth. 5 See the data on shares of net new mortgage lending in Buckle and Thompson (1995: 110). See also the figures in Financial Statistics (March 1989) on the credit counterparts to changes in M4 and M5: lending to the private sector undertaken by banks and building societies respectively was comparable in the second half of 1984, but bank lending rose much more rapidly from 1985 to 1987 and, in 1987–88 bank lending was more than twice the level of building society lending. 6 For example, Artis and Lewis (1976, 1981: chapter 2). 7 The amount of bills held by the Issue Department of the Bank of England fell sharply from a peak of £11,043m in December 1985 to £6,717m in September 1986, but then rose to a new peak of £11,576m in December 1986, before embarking on a more sustained fall over 1987 and 1988. 8 See, for example, Goodhart (1984: chapter IV); Artis and Lewis (1991: 160).

56 The Making of Monetary Policy in the UK, 1975–2000 4.1.2 Financial Liberalisation A number of authors, including Sargent (1991) and Lawson (1992), have drawn attention to the variety of measures of financial liberalisation and deregulation introduced between the late 1970s and the mid-1980s to free up the market for credit.9 In turn, these gave rise to major increases in competition. The measures and their effects (see also Table 3.3) ranged from the abolition of exchange controls in autumn 1979, through the discontinuation of the corset in summer 1980 and the elimination of fixed reserve requirements for banks in 1981, to the banks’ entry into the mortgage market in 1981, the collapse of the building societies’ interest rate cartel in late 1983 and the building societies’ introduction of unsecured lending from 1986. Together, they made it much easier for both the corporate and personal sectors to borrow. Indeed, an important part of the official case for the abandonment of broad money targets in 1987 was the argument that the private sector was choosing, in the light of these improved opportunities and the associated reduction in financial intermediation costs, to hold both more assets and more liabilities (see Chapter 3, Section 3.4). Financial liberalisation in this sense was clearly at the very least an important permissive factor in the late 1980s boom. Total consumer credit from the monetary sector rose annually at rates of 17% and more between 1983 and 1988, loans for house purchase increased even more rapidly, and the growth of bank lending in sterling to industrial and commercial companies accelerated from a 1983–5 average of 10.2% to 17.3% in 1986, 21.6% in 1987 and over 32% in each of 1988 and 1989.10 The phenomenon of equity withdrawal in the housing market— householders, as a group, borrowing for house purchase more than they needed and diverting the excess into consumers’ expenditure—also became more prominent. Miles’s (1992) figures show a marked rise in 1986, subsequently sustained over the rest of the decade.11 Whether financial liberalisation alone could have caused the boom is not clear. Sargent, for example, disputed this ‘consensus view’ by arguing that such a large increase in borrowing was unlikely to have occurred unless debtors believed that their ability to repay had improved. He pointed to ‘the climate of over-optimistic expectations which developed in the 1980s about the economy’s performance’, which he attributed to the improvement in UK productivity growth after the 1980–81 recession and the government’s own talk of the ‘transformation’ of the economy (1991: 79). However, it is difficult to believe that the emphasis on changed expectations could have such a large impact unless it was supported by an appropriate macroeconomic context. Individuals must have come to believe that their permanent incomes were set to rise partly, at least, because they observed increases in their current incomes. Indeed, Lawson, while expressing some sympathy for Sargent’s argument, pointed to the unexpectedly strong and steady growth from 1982 to 1988, which ‘coupled with the exceptional duration of the upswing, led too many borrowers and lenders to believe that it would go on for ever’ (1992: 631). The Bank of England’s post-mortem on the monetary policy pursued during the second half of the 1980s emphasised the extent of subsequent revisions made to the statistics on demand and output such that the extent of their growth had been initially underestimated.12 However, the Bank attributed the boom essentially to a large rise in consumer demand due to an improvement 9 See also Muellbauer and Murphy (1989). 10 Stock data from Financial Statistics, various issues. 11 In real terms (1985 prices), net cash withdrawal rose

from £5,348m in 1982, £5,566m in 1983, £7,131m in 1984 and £7,520m in 1985, to £12,571m in 1986, £12,221m in 1987 and £17,087m in 1988 (Miles, 1992: 70). 12 See also Nelson and Nikolov (2001).

The Interregnum and the Lawson Boom, 1985–1990 57 in long-term expectations, financial liberalisation and inappropriate interest rate reductions.13 This analysis therefore combined financial liberalisation and the climate of expectations, but added a policy element in the form of interest rates—though it did not analyse the reasons for the latter. A considerable amount of empirical work has been undertaken, especially by Bayoumi (1993), Carruth and Henley (1992), and Miles (1994), on the connections between consumers’ expenditure and developments in the housing market, focusing particularly on increases in borrowing and in equity withdrawal which can be related to financial liberalisation.14 In principle, it can certainly be argued that financial liberalisation should lead to an increase in consumers’ expenditure: households become enabled to exercise the choices that they had been previously constrained from making, and this increases their borrowing and, in the short run, their expenditures (although, for some, credit restrictions may have been replaced by lenders applying credit rationing). However, the magnitude of credit liberalisation’s contribution to the marked fall in the personal sector saving ratio during the second half of the 1980s has not been clearly established.15 Further discussion of this explanation is deferred to Section 4.3. 4.1.3 Excessive Relaxation in Response to the Stock Market Crash In the UK, as in the US, the monetary authorities responded to the stock market crash of October 1987 by lowering interest rates to prevent the development of a financial and economic crisis. It can be argued that interest rates were reduced too much in the UK or, and more convincingly, that they were held down for too long. Base rates were cut in three steps, each of 0.5%, between 23 October and 3 December, and did not regain their ‘pre-crash’ levels until July 1998. In this respect Walters (1990: 108) contrasted adversely the performance of the UK monetary authorities with that of the Federal Reserve. However, the boom was already well under way by autumn 1987 so that the post-crash monetary relaxation could not have been its stimulus. Moreover, since the Bank of England argued in February 1988 that the fears generated by the crash had not materialised, that the domestic economy remained buoyant and that a ‘non-accommodating stance’ for monetary policy was required (BEQB, 1988a: 8), it is difficult to put much emphasis on the crash. Policy must have been driven largely by other factors and considerations. 4.1.4 Excessive Fiscal Expansion In principle it is possible that the boom could have been due to excessive fiscal expansion. Some commentators have pointed to the particular effect upon the climate of expectations (as identified by Sargent) of the tax cuts in the 1986, 1987 and 1988 Budgets, which reduced the basic rate of income tax from 30% to 25%. However, the budget deficit, as measured by the public sector borrowing requirement, was on a falling trend from 1984 and was negative from 1987 to 1990, while the ratios of gross and net general government debt to GDP fell each year 13 See the Governor’s Durham lecture on ‘Monetary policy in the second half of the 1980s’, BEQB, 1990c: 215–20. 14 See also Aoki, Proudman and Vlieghe (2001). 15 Miles’s (1994) results suggest that about a third of the increase in consumer spending between 1986 Q1 and 1988 Q4 was due to the rise in house prices (which can be associated at least in part with credit liberalisation), one-fifth was due to anticipated and unanticipated changes in income, and another fifth was due to real interest rate movements.

58 The Making of Monetary Policy in the UK, 1975–2000 from 1984 to 1990.16 In addition, the increases in disposable incomes due to tax cuts were much smaller than the expansion of credit. Therefore, while tax cuts may have provided some limited support for the boom, fiscal policy cannot be regarded as a major causal factor. 4.1.5 Shadowing the DM In an interview with the BBC World Service in mid-May 1989, Mrs Thatcher claimed that the UK had picked up its ‘inflationary tendency’ through the shadowing of the Deutschemark.17 Sir Alan Walters, brought back as her personal economic adviser at the beginning of that month, subsequently argued that the policy of shadowing the mark was ‘quite critical in exacerbating the inflationary pressure . . . The primary motive for the expansionary policy was to contain the exchange rate of sterling at DM3.0. This was the main explanation for the inflationary policy from the latter half of 1987 through to the summer of 1988’ (1990: 112). This stance is related to the ‘Walters critique’ of the European Monetary System (EMS).18 According to the latter, the convergence of nominal interest rates within the EMS results in real interest rates that are perversely high in low inflation countries where demand could be allowed to increase. In particular, both views give a central role to the effect of interest rates on domestic demand while largely ignoring the role of competitiveness effects on demand through the exchange rate. However, the ‘new EMS’ model of Giavazzi and Spaventa (1990), which includes such exchange rate effects, shows that the short-run expansionary effects of low interest rates in countries like Spain are outweighed in the long run by competitiveness effects which accelerate the adjustment process and reduce the output costs of disinflation. In addition, Artis (1994) has shown that the perverse array of real interest rates predicted by Walters was not characteristic of the EMS during the second half of the 1980s. In the context of the UK’s DM-shadowing, Walters seems to have been arguing that this caused interest rates to be inappropriately low for longer than they would otherwise have been, thus exacerbating the boom and the consequent inflation.19 But this ignores the restraining influence exerted on UK inflation, other things being equal, by the peg to the DM (a lowinflation currency). If there is any force in the ‘law of one price’, then this peg could not in itself cause higher inflation, while higher inflation would have been restrained at least over the long term by the loss of competitiveness. However, if the peg had been fixed at too low an initial rate, then UK prices could be expected to rise in the short term towards DM prices. Thus, the important question is whether the peg had indeed been fixed at too low a rate. It should also be noted that real interest rates were strongly positive during the DM-shadowing episode, and showed no tendency to decline.20 16 Debt figures from OECD Economic Outlook (December 1996). The same source gives calculations for the general government structural balance (percentage of trend GDP) which suggest some fiscal expansion in 1986 and 1987 (more than reversed in 1988). But such figures are highly sensitive to the assumptions used for their construction. The general government financial balance (percentage of GDP) shown in Table 2.2 improved monotonically from a deficit of 4% in 1984 to a surplus of 0.6% in 1988 and 0.9% in 1989. 17 See Keegan (1989: 227); Smith (1992: 152); and Lawson (1992: 919–21). 18 Walters (1986: chapter 7; 1990: chapter 5). 19 According to Lawson (1992: 784–5, 788–9, 794–5), Thatcher was concerned mainly about the foreign exchange market intervention which the policy involved, rather than about the low level of interest rates; she was also perfectly aware of the DM-shadowing policy at the time. Thatcher’s own account (1993: 702) suggests even sterilised intervention would increase monetary growth and put downward pressure on interest rates; she claimed not to have known about the policy until November 1987 (699). 20 The movement of nominal interest rates is discussed in more detail in Chapters 8 and 9.

The Interregnum and the Lawson Boom, 1985–1990 59 Table 4.1 Contributions to growth of demand (% of GDP)

Consumers’ expenditure General government final consumption Gross domestic fixed capital formation Exports Imports GDP at market prices

1985

1986

1987

1988

2.2 0

4.0 0.4

3.2 0

4.6 0

0.6

0.3

1.4

2.4

1.3 −0.6 3.8

1.0 −1.5 4.2

1.4 −1.8 4.4

0.1 −3.0 5.2

Source: calculated from Economic Trends Annual Supplement, 2000 edition.

4.1.6 Excessive Depreciation Comparison of real exchange rate levels is notoriously difficult, but in broad-brush terms it seems likely that, after the prolonged overvaluation associated with the appreciation of 1980–81, sterling had returned by 1985 to a level where it was no longer either substantially overvalued or undervalued. Measures of the UK’s competitiveness in trade in manufactures, for example, suggest a return to levels comparable to those of 1979.21 Similarly, Williamson’s (1985) calculation of the fundamental equilibrium exchange rate (FEER) shows the actual rate falling rapidly towards the FEER over the course of 1984.22 Finally, according to the careful, but broad-brush, analysis in Chapter 8 below, the real exchange rate had been marginally undervalued in 1985 Q1 but rebounded after the January 1985 sterling crisis to become nearly overvalued in Q3, before falling back slightly in Q4. Between November 1985 and February 1987, however, sterling fell by some 14% against the dollar. The early part of this depreciation occurred within the context of a marked fall in oil prices, but sterling continued to decline after the halting, and partial reversal, of that fall. Although there was a small appreciation in early 1987, a large part of the depreciation was locked in by the rate at which sterling began to shadow the DM in March 1987. On the analysis of Chapter 8 the real exchange rate was substantially undervalued in 1986 Q4 and 1987 Q1. Such depreciation could have been expected to exert strong positive effects on UK demand and, therefore, output in the short run. On the annual data in Table 4.1 the rise in exports contributed around one-third of the expansion of GDP between 1985 and 1987. However, on quarterly data (see Figures 4.2 to 4.4) exports grew more rapidly than GDP between 1984 and 1987, particularly in late 1986 and early 1987, and they were a more important injection to aggregate demand than investment before mid-1987. The increase in demand provided by exports was not offset by rising imports before mid-1987, and exports grew more rapidly than consumption until late 1987. Thus, there is some empirical support for the proposition that sterling’s depreciation in 1986 stimulated exports and hence income. 21 On a base of average 1985 = 100, the 1979 figures are as follows: relative export prices 98.6; relative producer prices 97.0; IMF index of relative unit labour costs—actual 98.6 and normalised 88.9; and import price competitiveness 108.9. 22 See also Wren-Lewis et al. (1991).

60 The Making of Monetary Policy in the UK, 1975–2000 140

130

120 GFCF Exports GDP

110

100

90

80 84Q1

85Q1

86Q1

87Q1

88Q1

Figure 4.2 Income, exports and investment (average 1984 = 100) (Source: Economic Trends Annual Supplement, 2000 edition)

The 1986 depreciation has the right timing to explain the boom and provides the macroeconomic context in which the growth of money and deregulated credit could rise. It also gives an appropriate role to the DM-shadowing of 1987–88. It can therefore be argued that the 1986 depreciation was at the very least an important trigger—an important proximate economic determinant—of the boom and consequent inflation. But in that case the interesting question is whether the authorities knew what they were doing and, if so, why they did it.

150

140

130

120

Exports Imports GDP

110

100

90

80 84Q1

85Q1

86Q1

87Q1

Figure 4.3 Income, exports and imports (average 1984 = 100) (Source: Economic Trends Annual Supplement, 2000 edition)

88Q1

The Interregnum and the Lawson Boom, 1985–1990 61 140

130

120 Household consumption Exports GDP

110

100

90 84Q1

85Q1

86Q1

87Q1

88Q1

Figure 4.4 Income, exports and consumption (average 1984 = 100) (Source: Economic Trends Annual Supplement, 2000 edition)

4.2 MACRO POLICY DECISIONS IN THE SECOND HALF OF THE 1980s In autumn 1985 it was clear that world economic growth had been muted over the first half of that year and was unlikely to pick up significantly in the near future (see BEQB, 1985c: 341–60). The dollar had depreciated from its February 1985 peak, but this had had little effect upon the slow growth of the US economy, while the other major economies were also sluggish. The growth of the UK economy was decelerating (if allowance was made for the temporary boost due to the ending of the coal strike) and unemployment was flat. Inflation was falling and expected to continue to decline well into 1986, partly as the result of the strengthening of sterling after the January 1985 crisis, which had required increases in interest rates of over 4% before it was brought under control. The 1985/6 target for £M3 had been set at the 5–9% first given in the 1983 Medium Term Financial Strategy. As argued in Chapter 3, this target range was not consistent either with the authorities’ expectations about money GDP and velocity or with their likely forecasts of the credit counterparts. The obvious explanation for these discrepancies is that, although they knew at the start of the year that the previously announced figure was unlikely to be attained, the authorities were unwilling to increase their monetary target because of the recent sterling crisis and the low level of confidence in the financial markets. In the event, the growth of £M3 accelerated and the target was formally suspended in October, as announced in the Chancellor’s Mansion House speech along with the abandonment of the technique of overfunding. It is clear that November 1985 was, if not the first, then the most important time when Chancellor Lawson tried to persuade Prime Minister Thatcher to allow the entry of the pound sterling into the ERM.23 The key argument, to which Lawson, the Bank of England and the 23 Keegan (1989: 181); Brittan (1989: 33); Walters (1990: 99–101); Smith (1992: 61); Lawson (1992: chapters 33, 39 and 40); and Thatcher (1993: 693–8).

62 The Making of Monetary Policy in the UK, 1975–2000 Treasury all subscribed was that membership would provide a credible framework for policy and an alternative monetary discipline to replace the broad money target suspended the previous month. Mrs Thatcher vetoed the proposal. However, early in December 1985, the context within which macroeconomic policy was operating changed sharply with the arrival of a fall in oil prices. The spot price for Brent crude had fluctuated during 1985 around $27 but dropped to $22 in January 1986 and reached $13 in March.24 It was immediately obvious, with the UK being a significant oil producer, that ‘successful adjustment to lower oil prices is likely to require some exchange rate depreciation in order to improve the competitiveness of the non-oil sector’, and the Bank of England signalled its acceptance of such depreciation in mid-January.25 Sterling experienced further falls in late July and in late September, the latter occurring after the oil price had risen from its July floor of under $10 to a range of $13–15 where it was to remain for the rest of 1986. Sterling’s weakness was brought under control only with the support of the Bundesbank through concerted intervention in the foreign exchange market.26 With regard to sterling’s depreciation over 1986 as a whole, Keegan (1989: 183) represents the Chancellor as seizing the opportunity of the oil price fall to allow a devaluation which would stimulate the economy, but does not mention the likely inflationary consequences. Brittan (1989: 33) talks of sterling being ‘allowed to depreciate further without immediate visible inflationary effect, thus giving British industry a temporary competitive advantage and stoking up the boom over the election period’. Lawson himself gives the impression that the September depreciation in particular was unwanted. On his account it meant that ‘I lost the opportunity to lock in the marked fall in inflation that the oil price collapse had temporarily secured, and allowed policy to become looser just when the credit boom was starting to take off’ (1992: 651). It is clear that the authorities were aware of what they were doing. The September 1986 BEQB published a diagram,27 reproduced here as Figure 4.5, which showed the alternative combinations of effective exchange rate and sterling oil price required to deliver unchanged output, or current balance, or prices, or PSBR, over a three-year horizon. By August 1986 the economy was at a point which would produce unchanged prices but a substantial boost to output. For most of the following year the economy was in the area to the south-east of the price line where prices were expected to rise, and from 1987 Q4 to 1988 Q3 it was well to the north-west of the constant inflation line. The figure was constructed on the assumption that neither wages nor interest rates would react. But any conventional macroeconomic model of the period would have predicted a rise in wage pressure as a result of the expansion of output and the depreciation of the exchange rate and, therefore, a significant positive effect on price inflation over a period as long as three years. Thus macroeconomic policy in this period involved a deliberate choice, not merely to forgo a further fall in inflation, but to risk a rise for the sake of an expansion of output. In September 1986 the Bank of England had regarded the sterling depreciation of July–August as ‘not . . . sufficiently large . . . to offset the beneficial effects on the near-term outlook for 24 Figures supplied by the Royal Bank of Scotland. The price fell to $12 in April, rose to $14 in May and then declined to its lowest point of $9.6 in July; over the rest of 1986 it hovered around $14. Monthly prices averaged $18 in 1987, $15 in 1988 and $18 in 1989. 25 See BEQB, 1986a: 6 for the quotation, and 27–8 for the signal. Lawson (1992: 650) notes that he had himself argued in 1984 the related case for the eventual exhaustion of North Sea oil to be countered by a real depreciation of sterling, an argument consistent with Forsyth and Kay (1980). 26 Lawson (1992: 653–6). 27 In an article entitled ‘The impact of changing oil prices’, BEQB, 1986c: 331–2.

The Interregnum and the Lawson Boom, 1985–1990 63 Effective exchange rate index, 1975 = 100 90

Output down up

July 1985

80

n dowup rent r u C nce August bala n 1986 dowup s e Pric

70

up BR wn S P do

8

12 16 Spot oil price---UK Brent Blend (£ per barrel)

20

60

Figure 4.5 The sterling oil price and the exchange rate (Source: Bank of England Quarterly Bulletin, 1986c: 331. Reproduced by permission)

inflation of the fall in oil prices that has occurred over the last year’ (BEQB, 1986c: 333). But sterling fell further during September and October, and in December the Bank identified the depreciation that had occurred over the year to mid-October as ‘a fall estimated, if anything, to exceed that required to offset, over 2–3 years, the effect on the current account of halved oil prices and threatening to exert more than offsetting upward pressure on retail prices’; interest rates were raised in response (BEQB, 1986d: 451; see also 475). The rise in the oil price around the turn of the year was not accompanied by a significant recovery of sterling, which, ‘together with the comparative buoyancy of M0 and the strength of bank lending’, suggested to the Bank the ‘need for continuing caution in the conduct of monetary policy’ (BEQB, 1987a: 8).28 The policy of shadowing the DM was conceived in late February 1987 within the context of the Louvre Accord, which involved a concerted attempt to stabilise exchange rates between the world’s three major currencies (dollar, yen and Deutschemark). According to Keegan (1989: 28 Lawson’s own account stresses that he was concerned about the weakness of sterling as early as September 1986, but says that it was ‘not until the summer of 1987 that the Bank began to worry that monetary conditions may not be tight enough’ (1992: 653–5, 659).

64 The Making of Monetary Policy in the UK, 1975–2000 193), Lawson ‘hoped to prove to Mrs Thatcher, with a dry run, that she need have no fears about sterling’s chances of stability within the EMS’. More generally, however, the policy represented simply one step further in the development of the authorities’ preference for exchange rate stability as opposed to money supply targets. The March 1987 budget formally abandoned targets for broad money, put rather more emphasis than before on the growth of money GDP, but said little about the exchange rate. According to Funabashi’s account of attempted international policy coordination during the mid-1980s, at the Louvre meeting of 21–22 February 1987 Lawson had considered the current sterling rate ‘a bit weaker than it ought to have been because of declining oil prices’. Therefore the chancellor wanted ‘sterling’s midpoint [to be] a bit above where it was’ (Funabashi (1988: 186). In fact within a few weeks (by the time of the Budget), sterling appreciated by around 4% in effective terms and from just under DM2.80 to just over DM2.90. The latter, according to Keegan (1989: 193–8), was the central rate that the authorities were now attempting to maintain, a rate decided upon partly on the basis of discreet soundings with industrialists. Lawson himself mentions the post-Budget interest rate cut ‘with the pound at DM2.95’, and refers to ‘the policy of shadowing the DM, at DM3 to the pound’.29 However, the policy dilemma that was to dominate the period of DM-shadowing (and had its origin in the preceding excessive depreciation) was already manifest. As the Bank later remarked, on ‘domestic grounds, there was a reluctance to see interest rates fall too far or too fast: at the same time a substantial rise in the exchange rate would have had damaging consequences for industrial confidence’ (BEQB, 1987b: 189). The interest rate falls during the first few months of the execution of the strategy were reversed in early August when the upward pressure on sterling weakened. Interest rates were reduced again following the October 1987 stock market crash. The authorities interpreted this as easing the policy dilemma ‘since it implies a tightening of monetary conditions, by raising the cost of equity capital to firms and, through the effect of wealth reduction, by depressing consumption demand by households’ (BEQB, 1987d: 479). Monetary authorities in the US and elsewhere also reduced interest rates. In January 1988 the Governor of the Bank of England argued that the growth of domestic demand immediately before the crash had been ‘significantly stronger than many had thought—indeed, unsustainably rapid’, and played down the effects of the crash (BEQB, 1988a: 60). A small rise in interest rates in early February was followed by intense and sustained upward pressure on sterling which, after substantial intervention, culminated in the decision to allow sterling to float above DM3.00 in early March. This decision was reached in the context of extreme and near-public tensions between Prime Minister and Chancellor.30 It appears, however, that the authorities still hankered after stability of the exchange rate because the ‘tightening of monetary conditions through exchange rate appreciation was offset in part’ by interest rate reductions in March and April. ‘The combination of a stronger currency and lower interest rates did not represent an ideal response to concerns about excessive growth of domestic demand, but the strength of external demand for sterling had created a situation in which this method of tightening monetary conditions was the only one available.’ (BEQB, 1988b: 181). It is not obvious why the appreciation needed to be offset in this way if the authorities had really concluded that domestic demand needed to be restrained. From May 1988, however, the upward pressure on sterling fell away and interest rates were moved up sharply in June and July in what turned out to be the first phase of a long period of first rising, 29 Lawson (1992: 682, 683). He also notes that ‘there was no meeting at which the DM3 ceiling was formally agreed’ (789). 30 See Keegan (1989: 22–6); Smith (1992: 132–8); and, for the views of the chief protagonists, Lawson (1992: chapters 63–4) and Thatcher (1993: chapter 24).

The Interregnum and the Lawson Boom, 1985–1990 65 and then continuing high, interest rates. These persisted until the UK’s entry into the ERM in October 1990. The above discussion suggests that the authorities knowingly took a risk in allowing an excessive depreciation of sterling in 1986. The inflationary implications of a large part of that depreciation were then locked in by the rate at which it was decided to embark upon shadowing the DM in March 1987. The risk itself can be defended, in part at least, on the grounds that the UK economy clearly required a stimulus. Growth since 1980 had been unimpressive, unemployment was still rising during the first half of 1986 and, with the temporary boost from the ending of the miners’ strike now coming to an end, the immediate prospects were poor. In addition, the authorities may well have thought that they could avoid, or alternatively recoup, any excessive depreciation. Indeed, they succeeded in recouping some of it during the period between the Louvre Accord and the beginning of DM-shadowing. But why did the authorities persist with the DM-shadowing and the relatively low interest rates it required, when evidence of overheating was accumulating? Here it is not possible to rule out entirely the personal elements in the Thatcher–Lawson relationship to which Keegan has drawn attention (1989: 222–6, 230–3), but there are wider factors which should be accorded more importance. Firstly, the authorities were slow to recognise either that monetary policy was much too lax or that the economy was beginning to overheat. The growth of the only monetary aggregate for which there was still a target, M0, did not rise significantlyuntil mid-1988, but the expansion of a variety of measures of credit had been strong since at least 1986, and there are numerous indications in the BEQB that the Bank of England, at least, was aware of that growth. It is difficult to escape the impression that, with the dropping of broad money targets, the Governor’s Loughborough Lecture injunction that the ‘detailed study of liquidity and of the development of credit are essential elements in judging financial conditions’ (BEQB, 1986d: 507) had also been abandoned. With respect to the growth of demand, the Bank of England later made much of the fact that, because of inaccuracies in the statistics then available, demand and output had not at the time appeared to be expanding too fast (BEQB, 1990b: 216). However, it had previously noted with respect to April–May 1987 that the authorities ‘did not wish to see interest rates fall too sharply, with the economy already buoyant and monetary indicators suggesting prospective levels of demand which could put upward pressure on inflation in the near term’ (BEQB, 1987c: 344). Moreover, normal concepts of the lags in the effects of policy and the evidence of the sharp deterioration of the balance of trade31 should have suggested the need for caution. Secondly, it seems likely that the authorities overestimated the supply-side improvements thought to have occurred since 1979. In fact, Lawson was later to concede that he had probably, to some extent, mistaken ‘an unusually long cyclical upswing for a permanent supply-side improvement’ (Lawson, 1992: 804–5), though he was certainly not alone in this. The Bank of England also expressed at the time the view that improved supply-side performance offered ‘the chance of continuing growth at a rate sufficient to reduce unemployment’ (BEQB, 1987c: 321).

4.3 FINANCIAL LIBERALISATION VERSUS EXCESSIVE DEPRECIATION In Section 4.1 four explanations of the boom were rejected but two were retained: financial deregulation and liberalisation, and excessive depreciation in 1986. The policy decisions 31 Lawson argued at the time that at least part of that deterioration reflected private sector borrowing to finance the boom in investment (1992: 854–7).

66 The Making of Monetary Policy in the UK, 1975–2000 involved in the latter were discussed in Section 4.2, and these two rival explanations can now be brought together. It is useful to start by asking what would have happened if only one of these two factors had been present. On the one hand, if there had been liberalisation but no excess depreciation, then it is reasonable to suppose that the expansion of demand would have led earlier to a large balance of trade deficit. At the same time, competitiveness would have been squeezed more quickly32 and inflation would not have risen to such a high peak. On the other hand, if the depreciation had occurred without the prior liberalisation, consumption and investment would presumably have expanded less rapidly due to the lower availability of credit and the boom would have persisted for longer before reaching its peak. Secondly, the excess depreciation explanation can, in a sense, subsume the financial liberalisation factor. The growth of income due to the depreciation can be seen as inducing households and firms to make more use of the credit facilities that became available, in a way that fits in with the Sargent–Lawson–Bank of England arguments discussed in Section 4.1.2 It is worth noting that the household saving ratio, which had fallen from a peak of over 13% in late 1980 to an average of 9.0% during 1983–85, embarked on a distinct further decline which took it to 7.5% on average in 1986, 5.5% in 1987 and 3.9% in 1988.33 In addition, prior financial liberalisation strengthens the empirical case that there was a depreciation-induced stimulus to aggregate demand since it suggests that part of the rise in consumption (and later investment) can be seen as a by-product of the increase in exports. The net effect of the latter is not the rise in exports less imports as conventionally measured, but the rise in exports plus the export-induced rise in consumption minus the rise in imports. On the other hand, the financial liberalisation explanation cannot subsume the excess depreciation. There is no way in which the liberalisation of earlier years, which had not induced a substantial boom by 1986, could have triggered the excessive depreciation of that year, or indeed the DM-shadowing of the following year. In addition, the continuation of the boom for so long surely has to be seen in terms of policy decisions rather than financial liberalisation. The growth of credit turned down quite sharply during 1990 and the household saving ratio rose from under 4% in the second half of 1988 to 9.4% in 1991 (and over 11% in 1992–3). These movements could be seen as reflecting the end of a long process of portfolio adjustment, but it seems more natural to regard them as the lagged effect of the change in monetary policy and the cyclical position. This took interest rates from their low of under 8% in May 1988 to 13% by the end of 1988, to 14% in mid-1989 and to 15% in late 1989. Finally, some importance should surely be attached to the emphasis placed by the policymakers of the period—Lawson and the Bank of England—on the decisions they took, even though that makes them more culpable for the outturn. Although the Bank refers to interest rate cuts as well as financial liberalisation, Lawson’s own assessment points unequivocally to the excessive depreciation of 1986 rather than later policy mistakes. He also claims that Thatcher vetoed a proposed increase in interest rates in December 1986, which would have reduced the extent of the depreciation and limited the subsequent inflation.34 However, Thatcher’s memoirs make no mention of the latter episode and no echo can be found in the BEQB so it is difficult to evaluate. 32 In fact most measures of UK competitiveness did not return to their 1985 levels until early 1988. On the criteria introduced in Chapter 8 the real exchange rate was below the ‘wide band’ in 1986 Q4 and 1987 Q1, but it did not return to the ‘central range’ until 1987 Q4. 33 Data from Economic Trends Annual Supplement, 2000 edition. 34 Lawson (1992: 645, 799–800, 991–3, and, for the veto, 665–8, 681).

The Interregnum and the Lawson Boom, 1985–1990 67 These considerations suggest that the financial liberalisation and excess depreciation explanations should be seen as complementary rather than rival explanations, but some causal primacy should be accorded to the excess depreciation.

4.4 A MONETARY POLICY REACTION FUNCTION WITH RESULTS FOR 1985–1990 With the abandonment of monetary targets, policy was unequivocally focused on the interest rate as the main proximate instrument of the authorities. As discussed in Section 1.3 of Chapter 1, it has become common to estimate Taylor rules, in which the interest rate responds to inflation, the output gap and some external variables, as a way of characterising monetary policy in different countries and periods. There are two papers in the literature with results for subperiods close to those used here.35 Nelson (2000) presents some results for 1979–87 and 1987–90: in the former period the interest rate appears to respond rather weakly to inflation and the output gap, while in the latter the inflation response is replaced by a response to the sterling–DM exchange rate (with a long-run coefficient of 1.11). However, for present purposes the most relevant results are those of Adam, Cobham and Girardin (ACG) (2001), who address separately the 1985–90 (‘pre-ERM’) subperiod, the 1992–97 (‘post-ERM’) subperiod and the 1997–2000 (‘MPC’) subperiod. Their pre-ERM regressions also include dummies for the DM-shadowing episode, March 1987 – March 1988. ACG use the generalised method of moments (GMM) technique to estimate the error-invariables approach as pioneered in this context by Clarida, Gal´ı and Gertler (1998), which gives more precise results than the instrumental variables (IV) technique used by Nelson. ACG were particularly concerned to model the external influences on UK monetary policy, US as well as German. Because they were examining relatively short sample periods where there was considerable collinearity between potential regressors, they chose to estimate and then compare ‘domestic’ and ‘international’ reaction functions separately, instead of testing down from an over-parameterised general nesting model. In their domestic function the Treasury bill rate i responds to inflation π and the output gap y, whereas in the international function it responds to German and US interest rates iger and ius (they find the £/DM and £/$ exchange rates do not add significantly to the results). For the pre-ERM period their preferred equation is the international reaction function, which (on the basis of encompassing and over-identifying restriction tests) clearly dominates the domestic function. The long run relationship is: i t = 1.8† + 1.09‡igert + 0.77†igert ∗ D + 0.51‡iust − 0.36†iust ∗ D where D is a dummy variable equal to 1 in the period March 1987 to March 1988 and 0 otherwise. The (short-run) parameter on the lagged dependent variable is 0.57† and the equation standard error is 0.50% (as against the standard deviation of the Treasury bill rate of 2.04%).36 This result means that the UK interest rate was determined by German and, to a lesser extent, US interest rates, with the German rate becoming even more important and the US rate unimportant in the DM-shadowing episode. Domestic inflation and the output gap have no additional explanatory power. As ACG argue, this needs to be understood not in terms of the 35 Other results for the UK are summarised in Adam, Cobham and Girardin (2001, 36 † denotes significance at the 10% level, and ‡ significance at the 5% level.

table 2).

68 The Making of Monetary Policy in the UK, 1975–2000 16 14 Actual

Pre-ERM reaction function

12 10 8 6 4 2 0 1985M1

1987M1

1989M1

1991M1

1993M1

1995M1

1997M1

1999M1

Figure 4.6 Actual versus pre-ERM counterfactual interest rate (Source: Adam, Cobham and Girardin (2001))

monetary authorities’ preferences (presumably they have no attachment to the foreign interest rates themselves), but as a recognition by the UK authorities of the necessity to pay attention to external variables, in the light of the repeated exchange rate slides of September/October 1981, November 1982/January 1983, July 1984 and January 1985, all of which occurred after a period of interest rate cuts. In addition, the presence in the regression of foreign interest rates rather than exchange rates can be seen as a way for the authorities to acquire credibility (by showing that their policy is ‘like’ that of reputable authorities elsewhere) and (given uncovered interest parity) of stabilising the exchange rate over the medium term, while avoiding the need to respond to a noisy variable whose motion is unpredictable and difficult to explain, even ex post. ACG also report preferred regressions for the post-ERM period and the MPC period, and these are discussed in Chapters 6 and 7 respectively. The preferred regression for the post-ERM period is a joint model which includes domestic as well as international variables, while that for the MPC period is a domestic model in which the international variables figure only as instruments for inflation and the output gap. ACG interpret the rising importance of domestic variables and the declining role of external variables as reflecting gains to credibility which allowed the monetary authorities to respond to domestic variables without having to worry unduly about the exchange rate. ACG provide figures (reproduced here as Figures 4.6 to 4.9) which show how the interest rate would have moved throughout the whole of their period if it had been determined (given the actual values of inflation, the output gap and foreign interest rates) as in each of the three preferred regressions, and as in a straightforward Taylor rule.37 Three points stand out from these comparisons. First, for 1985–86, policy set on the basis of the post-ERM rule, the MPC rule or the Taylor rule would have been looser. This suggests that in those years at least, Chancellor Lawson was not obviously taking risks with inflation. Second, for the March 1987 – March 1988 DM-shadowing episode, while the MPC rule predicts an increasingly 37 Their Taylor rule rate is calculated on the basis of coefficients of 1.5 on inflation and 0.5 on the output gap, and on the assumptions of an equilibrium real interest rate of 3% and an inflation target of 2.5%.

The Interregnum and the Lawson Boom, 1985–1990 69 20

16 Actual

Post-ERM reaction function

12

8

4

0 1985M1

1987M1

1989M1

1991M1

1993M1

1995M1

1997M1

1999M1

Figure 4.7 Actual versus post-ERM counterfactual interest rate (Source: Adam, Cobham and Girardin (2001))

higher rate than the actual as from May 1987 (with a peak divergence of over 7% in March 1988), the others imply rates which were below or relatively close to the actual up to September or October 1987, with peak divergences (in March 1988) of around 2% for the post-ERM rule and 1% for the Taylor rule. Third, for the 1987–90 boom, while the MPC rule would have entailed an earlier and much larger rise in the rate than actually occurred, the post-ERM rule predicts a rate that is higher than the actual but considerably lower than that implied by the MPC rule, and the Taylor rule predicts a rise in line with the actual up to early 1989, then a further rise in early 1990 rather than mid-1989. The peak divergence for the MPC rule is nearly 10% in March 1990, while those for the post-ERM and Taylor rules are around 2% in June and 3% in October 1990 respectively. 30

25 Actual

MPC reaction function

20

15

10

5

0 1985M1

1987M1

1989M1

1991M1

1993M1

Figure 4.8 Actual versus MPC counterfactual interest rate (Source: Adam, Cobham and Girardin (2001))

1995M1

1997M1

1999M1

70 The Making of Monetary Policy in the UK, 1975–2000 18 16 14 Actual

Taylor rule

12 10 8 6 4 2 0 1985M1

1987M1

1989M1

1991M1

1993M1

1995M1

1997M1

1999M1

Figure 4.9 Actual versus Taylor rule counterfactual interest rate (Source: Adam, Cobham and Girardin (2001))

It is clear that there is a much larger difference here between the actual rate and that implied by the MPC rule than between the actual and the rates implied by the other two rules. ACG suggest that while some of this reflects the MPC’s greater ‘activism’, much of it reflects the greater scope for autonomous policy afforded by a substantial credibility gain associated with instrument independence in 1997 and the revised inflation targeting framework. In the low credibility circumstances of the 1980s, criticism of monetary policy under Lawson should be based more on the comparison with the post-ERM and Taylor rules, which imply that policy was tightened too little and too late, but suggest that the policy error was somewhat smaller.

4.5 CONCLUSIONS It has been argued in Sections 4.1 to 4.3 that the late 1980s boom was triggered primarily by the excessive sterling depreciation of 1986. This depreciation was locked in by the rate at which the DM was shadowed for a year from March 1987. Its effects were reinforced by a largely endogenous acceleration of monetary and credit growth made possible by the prior measures of financial and credit liberalisation. Moreover, the depreciation was deliberately allowed in the clear knowledge of the risks entailed. These were then exacerbated by over-optimistic expectations about the extent of the improvements to the supply-side of the UK economy and by a cavalier disregard for the growth of credit and liquidity, both of which led the authorities to continue with these policies for too long. This emphasis on the depreciation of the exchange rate as the trigger of the boom avoids the necessity to argue the difficult case that interest rates were too low: real (ex post) interest rates were strongly positive, at around 5%, throughout the period and showed no significant decline (see Chapter 9), so that it is hard to maintain that low interest rates were the cause of the boom. In addition, while according to the analysis of Section 4 interest rates set on the basis of the policy rules of later periods would have been raised earlier and by more in order to head off the

The Interregnum and the Lawson Boom, 1985–1990 71 rise in inflation in the late 1980s, rates would also have been set lower in 1985 and 1986 when inflation was low and output below capacity. The outcomes for the exchange rate, growth and inflation are not obvious; given the low level of credibility of monetary policy at this time, it may in fact have been sensible that, as Clarida et al. put it (1998: 1054), ‘Monetary policy boiled down to keep [sic] real rates steadily high over this period, even when inflation was low during the mid-1980s’. With respect to monetary frameworks, the informal exchange rate target of March 1987 – March 1988 contributed to the inflationary boom only because the operational choice of the peg was inappropriate; and the regime was too short for any further conclusions to be drawn. The lack of binding targets in the two years before DM-shadowing cannot be blamed for the boom either, insofar as monetary expansion succeeded rather than preceded the decision to peg at an undervalued rate and was at least partly endogenous. On the other hand, the postshadowing phase of no explicit targets saw a determined and ultimately successful attempt to control the inflationary boom but without provoking the kind of appreciation associated with sharp monetary tightening in the past, e.g. in 1979–80 (see also Chapter 9).

5 The ERM Interlude, 1990–1992 Over the second half of the 1980s the UK monetary authorities finally concluded that the UK should enter the ERM. This issue had been a focus of intense debate within the government for some time. In particular it had been central to the disputes between Margaret Thatcher and both Nigel Lawson (Chancellor of the Exchequer 1983–89) and Geoffrey Howe (Foreign Secretary 1983–89, Leader of the House 1989–90) which led to the resignations of the latter two in October 1989 and November 1990 respectively. Before this, however, their pressure had led Thatcher to enunciate at the Madrid European Council in June 1989 the conditions under which sterling would enter the ERM: these included the abolition of all foreign exchange controls, further moves towards completion of the Single Market and free markets more generally, and the reduction of UK inflation. The new Chancellor (from October 1989), John Major, soon became convinced in his turn of the merits of joining, and in the spring and summer of 1990 widespread expectations that the UK was likely to join in the near future led sterling to appreciate by some 6% against the DM. By now exchange controls had been abolished in the other ERM countries. In September Major revised the key remaining ‘Madrid condition’ for sterling’s entry (the other conditions were open to considerable interpretation), so that, instead of a reduction of UK inflation into line with that in ERM member countries, only a reduction in prospective inflation into line with the others was now required. The UK entered the Exchange Rate Mechanism of the EMS on 8 October 1990, only to leave it less than two years later in the crisis of Black Wednesday, 16 September 1992. Entry had taken place after many years of discussion, and the debate continued during the UK’s membership and was not silenced by its exit. This chapter starts by examining the conditions under which and the way in which the UK entered the mechanism. Section 5.2 then investigates the internal policy decisions and external events that culminated in the UK’s exit. Section 5.3 reviews alternative explanations of the EMS crises of 1992–93 of which the sterling crisis was one element. Section 5.4 considers the UK case in relation to those explanations. Section 5.5 uses the reaction functions found by Adam, Cobham and Girardin (2001) for the pre-ERM, post-ERM and MPC periods to shed further light on the UK’s experience in the ERM. Section 5.6 concludes.

5.1 INTO THE ERM The discussion of the previous two chapters has made clear a large part of the motivation for entry into the ERM: monetary targeting had been abandoned as being no longer a useful framework for monetary policy, and policy-makers had become increasingly concerned about the level of and changes in the exchange rate. In a lecture to the Institute of Economic Affairs in July 1989, for example, the Governor of the Bank of England referred to two potential economic benefits from membership of the ERM: it could provide a reduction in intra-EC exchange rate volatility and ‘an additional anchor for prices’, but he argued that any gain in the credibility of counter-inflationary policy would have to be worked for,

74 The Making of Monetary Policy in the UK, 1975–2000 and he doubted whether there would be any automatic effect on wage settlements (BEQB, 1989c: 373).1 In the immediate aftermath of the Lawson boom the authorities’ attention was focused on the control and reduction of inflation, and they regarded entry to the ERM in those circumstances as inappropriate,2 but the episode must have given added impetus to the desire to enter, as a way of avoiding another loss of control. In a lecture on monetary policy in the second half of the 1980s which represents the Bank’s main statement on the Lawson boom, the Governor argued the need to prevent another resurgence of inflation and therefore ‘not [to] allow the lessons of the second half of the 1980s to be forgotten’, and suggested that—unlike four other measures put forward by critics of the high-interest-rate policies of that period—ERM membership ‘could play an important part’, presumably in controlling aggregate demand and hence inflation, and thereby allowing lower interest rates in some situations (BEQB, 1990b: 220, 219). Official statements at the time of entry put the emphasis more on the immediate reduction of inflation. The Chancellor of the Exchequer told the House of Commons that ‘A firm exchange rate is a vital part of our policy to maintain tight monetary conditions in order to reduce inflation . . . [ERM membership will be] an additional discipline for the United Kingdom economy’.3 In part at least this emphasis was probably a reflection of conflicts over ERM entry within the government; in particular Prime Minister Thatcher’s opposition to the ERM, which was longstanding and deeply felt, could have been swayed more easily by appealing to the counter-inflationary properties of the move rather than the implied constraint on future policy, the likely increase in exchange rate stability or, indeed, the credibility effects. Thus the authorities regarded ERM membership as conferring benefits in the form of greater exchange rate stability and a contribution to counter-inflationary policy. However, it is not clear how much they saw this as affecting their own room for manoeuvre. The Governor in his IEA lecture had stressed the need for policy to be ‘continuously directed to the counter-inflationary discipline needed to sustain’ the exchange rate commitment (BEQB, 1989c: 373). But Major, in the quote given above, talked only of a discipline ‘for the UK economy’, while Thatcher ‘was determined to demonstrate that we would be looking more to monetary conditions than to the exchange rate in setting interest rates’.4 In addition, there is no recognition in the official statements of what academics would recognise as the time-inconsistency problem: the authorities did not see membership as a way of tying their own hands in order to convince the financial markets of their rectitude, which they did not see as being open to question, and, as already noted, the Governor rejected the idea of an automatic gain in credibility affecting wage settlements. Moreover, the authorities clearly did not consider ERM entry as the prelude to the UK’s acceptance of EMU.5 1 In a lecture given a year after the UK joined the ERM the Governor talked positively about membership, emphasising the successful reduction in both inflation and interest rates over that period and saying that ‘ERM membership has of course greatly reinforced the credibility’ of the UK’s ‘continuing counterinflationary commitment’ (BEQB, 1991d: 517). 2 See, for example, the Governor’s remarks reprinted in the BEQB, 1990a: 60, 64. 3 Hansard, 15 October 1990, vol. 177, col. 928. See also the Bank of England’s announcement of entry in the BEQB, 1990d: 439. 4 Thatcher (1993: 724) (the full quote is given in the text below). For contemporary comment to the effect that Thatcher, at least, had underestimated the extent to which sovereignty had been given up on entry to the ERM, see the Economist, 13 October 1990, pp. 13–14. 5 See, for example, the Governor’s speech in July 1989, BEQB, 1989c: 372–3. See also Lawson (1992, e.g. chapter 71); Thatcher (1993: chapters 24 and 25).

The ERM Interlude, 1990–1992 75 The above account suggests that the intra-governmental disputes over the ERM had not been properly resolved. This suspicion is strengthened by a consideration of the interest rate cut which accompanied the UK’s entry into the mechanism. Thatcher was keen to see a fall in interest rates, which had been at 15% (base rate) for over a year. According to the account she gave in her memoirs, I insisted against the Treasury and the Bank on a simultaneous announcement of a 1 per cent cut in interest rates. They had not disputed that the monetary and other figures warranted this; but they had wanted to delay. But I for my part was determined to demonstrate that we would be looking more to monetary conditions than to the exchange rate in setting interest rates. So on Friday 5 October we announced that we were seeking entry into the ERM, and I placed heavy emphasis on the interest rate cut and the reasons for it in presenting that day’s decision. (Thatcher, 1993: 724)

That cut was widely criticised, at the time and since, as being premature and giving the wrong signal.6 The Governor was obliged to defend it shortly afterwards, referring to evidence that ‘the conditions necessary to reduce inflation are now coming into place’, arguing that at 14% interest rates remained high and their restraining influence had been reinforced by ERM entry, and insisting that no significant easing of the policy stance had been or was intended (BEQB, 1990d: 485). Another aspect of entry which was open to criticism was the exchange rate at which it was effected. The UK entered the ERM with a central rate defined on the basis of £1 = DM2.95, just above the rate at which the market closed on Friday 5 October, and it announced this rate before it had been discussed or agreed with the European Community monetary committee (Smith, 1992: 171–2). It seems certain that the German authorities were unhappy at the time about this rate, which was presented to them as a fait accompli rather than a basis for discussion: Major . . . telephoned Karl Otto P¨ohl, the Bundesbank President, to tell him the news shortly before the official announcement at 4 p.m. The phone call was designed to impart glad tidings, but P¨ohl’s reaction was less than enthusiastic. The exchange went along the following lines: Major: We are coming in at DM2.95. P¨ohl: That is unrealistic. That is not possible. Major: But it has been decided by the prime minister. P¨ohl: I don’t care about your prime minister. (Marsh, 1994: 158)7

In addition, studies such as Wren-Lewis et al. (1991) which used Williamson’s (1985) concept of the fundamental equilibrium exchange rate (FEER) claimed that sterling was overvalued by around 10% at the time of entry, a view put forward by the National Institute for Economic and Social Research at the time and later upheld in a retrospective on the UK’s ERM experience by Barrell et al. (1994). On the other hand, research based on purchasing power comparisons, for example by the OECD, suggested that sterling was not overvalued at the time of entry.8 At the time the authorities responded by rejecting the view that the DM2.95 parity was too high—indeed they claimed that the real exchange rate this represented was below the 6 See, for example, the Economist, 13 October 1990, pp. 13 and 29; Lawson (1992: 1009). 7 See also Stephens (1996: 174). 8 See OECD (1991, Annex 2), which concluded that sterling did not appear to be significantly

over- or undervalued in real terms compared to historical averages, and Williams (1991). See also Wright (1992) for some criticism of the FEER concept, and Allsopp (1994), one of the few who continued to argue this case after Black Wednesday.

76 The Making of Monetary Policy in the UK, 1975–2000 average of the previous decade.9 Moreover, the Chancellor (now Norman Lamont) stated in December 1991 and again in July 1992 that the UK’s move to the narrow band in the ERM, when it came, would be at an unchanged central parity of DM2.95. However, the interest rate cut and the exchange rate parity issues should be seen as connected. The longstanding critique of the EMS by Alan Walters (1986, chapter 7), which must have been well-known to policy-makers not least because Walters was Thatcher’s personal adviser on economic matters between 1981 and 1984 and again in 1989, argued that under the ERM nominal interest rates of member countries would converge, with the result that the pattern of real interest rates would be perverse: those with lower inflation rates would have higher real interest rates, and vice versa, so that inflation rates would tend to diverge rather than converge. The ‘Walters critique’ is certainly open to criticism,10 but what is important here is that it was given at least some credence by policy-makers. The implication of this argument for the UK’s putative entry into the ERM in the conditions of 1989–90 was that it might find itself having to accept a large cut in interest rates which was inappropriate for domestic reasons, a possibility made explicit by the Governor in his IEA lecture in July 1989: It would be a mistake to enter the mechanism in circumstances where our anti-inflationary policy might be compromised or undermined. This could happen if we wished to keep interest rates high for domestic reasons but, by for example committing ourselves to too low a parity, we were pushed towards lowering interest rates to keep sterling within its band. (BEQB, 1989c: 374)11

The authorities also welcomed the appreciation of the spring and summer of 1990 as a useful counter-inflationary tightening of policy (BEQB, 1990c: 326; 1990d: 442). They may therefore have calculated that an interest rate fall soon after entry was unavoidable for Walters-critique reasons, but would be more acceptable in terms of the stance of monetary policy if preceded by an appreciation;12 moreover, under those conditions the intra-governmental politics of entry would be easier since Thatcher could be allowed the interest rate cut she wanted. But if this explanation of the way in which the UK entered the ERM is correct it implies that policy-makers were focusing on the short-term impact of entry on inflation and interest rates (Allsopp, 1994: 142), without thinking, in particular, of competitiveness in the longer run. Even if sterling was not overvalued on entry, it was certainly not undervalued,13 and the revision of the Madrid conditions so that entry occurred before substantial inflation convergence meant that some competitiveness was bound to be lost in the first few years.14 In concentrating their energies on getting Thatcher’s acceptance of entry, UK policy-makers may have thought that problems associated with the manner and rate of entry could always be sorted out later. 9 BEQB, 1991a: 54; Treasury (1990–91). See also the Governor’s speech to the CBI Eastern Region on 8 October 1992, ‘Monetary policy and the ERM’, mimeo, Bank of England. 10 See, for example, Artis (1994). Walters himself presented a somewhat modified version in his (1990) book. 11 See also the Economist, 6 October 1990, p. 133, and Currie and Dicks (1990). 12 In this period the Bank of England regarded the interest rate and exchange rate as two aspects of the stance of monetary policy. For example, the Bank had commented on the interest rate cuts following the appreciation of sterling in March–April 1988 as follows: ‘. . . the combination of a stronger currency and lower interest rates does not represent an ideal response to current concerns and a different balance would be desirable if it could be achieved’ (BEQB, 1988b: 162; see also 181). Lawson (1992: 836) records that Thatcher was infuriated by this remark. 13 On the broad-brush analysis of Chapter 8, sterling was just inside the top of the ‘wide band’, i.e. it was near to but not quite overvalued. 14 In the event, over the period of UK membership (which coincided with the immediate post-unification boom in Germany) UK prices rose by 7.0%, German by 6.9% and French by 4.7%. (Consumer prices, data from OECD Main Economic Indicators.)

The ERM Interlude, 1990–1992 77 Lawson (1992: 503) argued that, had the UK entered in 1985, the pound would almost certainly have devalued with the French franc (by 6% against the DM) in April 1986. However, by 1990 the ERM was in the ‘new EMS’ phase in which realignments had become much less acceptable, and the Maastricht treaty of December 1991 made them all but impossible in non-crisis conditions. In entering the ERM, then, the UK monetary authorities were hoping to find a more credible framework for policy and an alternative nominal discipline to that of monetary targets which were seen to have failed. But the way they entered the ERM was strongly influenced by shortterm and political considerations, and it is doubtful whether all of the politicians, in particular, had really understood the constraints involved. They believed ERM membership would give a clearer indication to the markets of what government policy was intended to be, that it would tend to reduce the volatility of sterling against other ERM currencies, and that it might usefully constrain the price- and wage-setting behaviour of firms and workers. But they were not clear about the implications for their own behaviour, and they certainly did not conceive of membership as a pre-commitment on their own part designed to convince the financial markets that they would not renege on the pursuit of price stability.

5.2 OUT OF THE ERM The basic story of the UK’s membership of the ERM is easily told. Although the immediate honeymoon effect in the financial markets was short-lived (BEQB, 1991b: 29, 30), the pound remained (until the summer of 1992) reasonably stable and inflation fell sharply, from its peak of 10.4% in the quarter before entry to 3.6% in 1992 Q3.15 Although the initial interest rate cut had unsettled the financial markets, the authorities were able to make substantial cuts amounting to 3.5% between mid-February and early September 1991. By late 1991 the Treasury was hailing the unusually small differential between German and UK (short-term) interest rates as ‘a rare event . . . in part due to the perceived discipline of the ERM commitment,’ but also partly the result of ‘the substantial progress in convergence of UK inflation towards German inflation seen over the last year’ (Davies, 1991–92: 36). On the other hand GDP growth fell to 0.6% in 1990 and −1.5% in 1991, and recovered only to 0.1% in 1992; unemployment (UK claimant count) rose continuously from its trough of 5.5% in the first two quarters of 1990 to 10.2% in 1992 Q4 (and a peak of 12.9% in 1993 Q2 and Q3); house prices (especially in the south-east of England) fell sharply; and mortgage foreclosures and repossessions rose dramatically. There was official concern about the burden of outstanding debt in the housing market and in the corporate sector (e.g. BEQB, 1991d: 455, 534–7; and 1992c: 242–3), and about the possible effects on the financial system of defaults triggered by a rise in interest rates.16 Economic recovery always seemed to be around the next corner, and the authorities’ confident assertions in early 1991 that the recession would be shorter and less severe than that of 1980–81 gave way by mid-1992 to the recognition that it was the longest recession since the 1940s (BEQB, 1991a: 7; 1992b 126).17 The Danish referendum result of June 1992 jolted the financial markets’ confidence in the move to EMU at existing exchange rates, and speculative pressures on sterling, as on other 15 This is the figure for the RPI; RPIX 16 Stephens (1996: 189, 217). 17 Later figures show that GDP began

inflation fell from 9.0% to 4.2% in 1992 Q3.

to rise in 1992 Q2, but contemporary data did not indicate an unequivocal upturn until well into the following year (BEQB, 1993b: 167, referring to 1993 Q1).

78 The Making of Monetary Policy in the UK, 1975–2000 ERM currencies, began to build up, particularly after the Bundesbank raised its interest rates in mid-July. These pressures came to a head in September in advance of the French referendum on the Maastricht Treaty, and the UK left the ERM in the crisis of 16 September 1992, commonly referred to as Black Wednesday, despite a last-minute increase in interest rates and massive (and unprofitable) foreign exchange market intervention. The UK’s exit needs to be evaluated in the light of an examination of the UK authorities’ actions and statements over the months leading up to Black Wednesday. First, the UK monetary authorities resisted upwards pressure on UK interest rates, starting in the period after the Bundesbank raised its key rates by 0.5% in mid-December 1991 ( just after the Maastricht conference) (BEQB, 1992b: 36–7). During the following quarter the Bank of England did not discourage downward pressure on rates (BEQB, 1992b: 156, 159–62), although there was no actual cut—partly because of the general election campaign—until early May. Then the cut of 0.5% in official rates was not entirely reflected in interbank rates because of ‘unease over the reaction of the foreign exchange market and some expectation that German rates might be raised again, perhaps at the Bundesbank Council meeting on 21 May’; however, sterling strengthened and rates generally fell (BEQB, 1992c: 274). The rise in German interest rates in fact came on 16 July in the form of a 0.75% rise in the discount rate but no change in the Lombard rate: however, the UK ‘did not need to follow the rise in rates in some other ERM countries’, although sterling weakened (BEQB, 1992d: 388). For the next two months, but particularly from late August, the authorities continued to resist upward pressure on UK interest rates, via a range of money market and gilts market tactics,18 and cuts in interest rates on National Savings products on 20 July and 5 August in order to avoid competitive pressure leading to a rise in building society deposit and mortgage rates.19 Second, the UK authorities allowed sterling to fall towards the bottom of its wide band in the ERM after the German interest rate rise in July 1992 in order to bring about a slight easing of monetary conditions.20 The pound had reached its lower limit against the escudo by the end of July; in early August it fell below DM2.83 (as against the wide band limit of DM2.778) and had to be supported at the margin against the peseta and the escudo, and it went below DM2.795 on 22 August. Intervention in the foreign exchange market became increasingly frequent and heavy from around that time, but sterling remained below DM2.80 for most of the rest of the period up to Black Wednesday. Third, the authorities made clear their unwillingness to countenance devaluation. In early July, for example, the Chancellor repeated his earlier statement that the UK would ‘move to the narrow bands in due course at our current central rate of DM2.95’, and went on to reject devaluation out of hand on the grounds that it would raise rather than lower interest rates.21 On 3 September the government announced an ECU 10 billion borrowing programme 18 BEQB, 1992d: 389–90. These included relieving money market shortages early in the day, allotting no bills at all in the Treasury bill tender on 28 August, not cutting the tap prices of gilts for which demand had dried up, and issuing gilts designed to appeal to domestic rather than international investors. 19 On 20 July the government ‘attempted to head off a widespread rise in the cost of home loans’ by cutting the rate on a highly successful new National Savings product, the First Option Bond (Financial Times, 21 July 1990); on 5 August it cut rates on a variety of other National Savings products ‘in a move apparently aimed at avoiding mortgage rate increases by building societies’ (Financial Times, 6 August 1990). See also Stephens (1996: 218, 221). 20 BEQB, 1992d: 388. Thus the authorities did not take the kind of action (e.g. intervention, rises in interest rates, realignment) normally expected when a currency passed its divergence threshold in the ERM. 21 Treasury Bulletin, summer 1992, vol. 3, no. 3, pp. 59, 61. The latter point was in line with the article by Davies (1991–92) previously referred to. The Chancellor also ruled out interest rate cuts themselves, in or out of the ERM.

The ERM Interlude, 1990–1992 79 (which would make possible more extensive foreign exchange market intervention), and this provided a temporary lift to sterling. Official opposition to devaluation was repeated more and more strongly over the next two months, culminating in Major’s statement in Glasgow on 10 September in which he ‘insisted that the Exchange Rate Mechanism had delivered the low inflation and the relative exchange rate stability demanded by the business community. Maintaining sterling’s parity was essential to ensure that Britain preserved and improved on those gains. “As the Chancellor has made clear, there is going to be no devaluation, no realignment.”. . . “The soft option, the devaluer’s option, the inflationary option would be a betrayal of our future; and it is not the government’s policy.” ’22 Fourth, the authorities allowed it to become clear to the markets—at a time of strong domestic political pressure for lower interest rates, notably from the Eurosceptic right wing of the Conservative party—that they were desperate to see interest rates come down. The cut in the National Savings interest rate on 20 July ‘had given a clear signal to the financial markets that [the Treasury] could not accept a rise in mortgage rates. Henceforth, any threat to defend the pound with higher borrowing costs would be recognised as a bluff’ (Stephens, 1996: 218). And while the full story of the ECOFIN conference of EU ministers of finance and central bankers in Bath on 4–5 September 1992 came out only later, it was clear enough at the time that extraordinary (but unsuccessful) pressure to cut interest rates had been exerted on Helmut Schlesinger, the Bundesbank President, notably by the UK Chancellor, Norman Lamont, who was in the chair. The joint statement, which said that ministers had reaffirmed their faith in the existing exchange rate parities and that the Bundesbank had no current intention to increase its interest rates, was not enough to calm speculation in the forex markets. Matters were not helped by Lamont’s proclamation of victory over the Bundesbank, or by an interview shortly after in which Schlesinger appeared to cast doubt on the Bundesbank’s own commitment to the existing parities.23 What these actions and statements added up to was an extremely high-risk strategy. The authorities were trying to get through what they expected to be a limited period of turbulence ending with the French referendum on the Maastricht Treaty (on 20 September 1992), but to do so they were successively playing all the cards in their hand and exhausting their strategic options. There was certainly a possibility that they might have succeeded—the general turbulence might not have reached such a pitch, and/or something positive might have turned up to reduce the pressure on the pound and on UK interest rates—and in that case the brinkmanship involved in the government’s strategy would have paid off. But the failure of concerted foreign exchange market intervention in favour of the dollar and against the DM in July and August (BEQB, 1992d: 384–5, 389) should have indicated that the strength of the DM was well established and unlikely to be unwound in the near future. In addition, there was never any serious probability that the Bundesbank, which guarded its independence so jealously, could be pressurised by foreign governments into lowering its interest rates: being seen to bow to overt pressure would have affected the Bundesbank’s reputation so badly that such pressure 22 Financial Times, 11 September 1992. This speech was described in an article by Philip Stephens the next day headlined ‘Major gambles his authority on the pound’ as representing the ‘return of conviction politics’, rather than ‘the ritual pledge of a prime minister in the heat of battle with speculators on foreign exchange markets’ (Financial Times, 12/13 September 1992). See also Stephens (1996: 235). Privately, the UK government rejected devaluation of sterling except in the context of a major ERM realignment, which the Bundesbank had first suggested in late 1990 but the French consistently refused. Publicly, the UK ruled out a sterling devaluation even if the franc was devalued (Stephens, 1996: 218). 23 See Financial Times, 7 September 1992, for a contemporary report, and Norman and Barber (1992) and Stephens (1996: 228–34) for the more complete story which came out later.

80 The Making of Monetary Policy in the UK, 1975–2000 must have made a cut less, rather than more, likely. Thus the strategy followed by the UK authorities stood a good chance of failing, but they had no contingency plan for that situation, no alternative policy that could be implemented in a controlled and orderly way if what they saw as external pressures became too strong.24 Moreover, it is clear that the authorities could have followed a different strategy, which would have had different effects. In particular a rise in UK interest rates in mid-July would have indicated unequivocally to the markets that the UK’s commitment was firm and deliberate;25 and the embarrassing episode at Bath could have been avoided.

5.3 ALTERNATIVE EXPLANATIONS OF THE 1992–1993 EMS CRISES It is convenient to distinguish two types of explanation of the EMS crises. Explanations of the first type start from theories of exchange rate determination and view a crisis simply as a concentrated manifestation of factors causing a depreciation/appreciation of the currencies concerned. Such an approach requires an analysis of the fundamental forces pushing for a movement in the exchange rate together with an explanation of why these fundamental forces have a sudden rather than a more gradual impact; a roughly equivalent approach is in terms of background factors and triggers or catalysts. As regards the fundamentals this approach is able to draw on a relatively well-developed (though by no means settled) body of economic theory, but it requires some more or less ad hoc assumptions to explain the timing, and it is difficult within this framework to pose questions about the relative importance of background and trigger factors since both seem equally necessary. Explanations of the second type start from the theory of balance of payments crises, in which a given balance of payments problem with falling foreign exchange reserves leads to a speculative attack on the currency and the collapse of the fixed exchange rate (‘first-generation’ models); or in which a speculative attack is generated by expectations of a future change in the fundamentals in the form of the policy chosen by an optimising policy-maker, a change which is expected to occur anyway or to occur if a speculative attack takes place (‘second-generation’ models).26 Such an approach provides a more satisfactory analysis of the timing of a balance of payments crisis, but takes a large part of the fundamentals including initial private sector expectations as given. This section starts by considering two explanations of the first type: the German unification explanation, and the competitiveness explanation. Under the second type it considers a number of explanations, starting with two applications of the ‘second-generation’ of crises models— referred to as the ‘expected policy change’ and ‘contingent policy change’ models. It then 24 See Stephens (1996: 208–25) for a detailed account of government policy discussions over the summer and early autumn of 1992. 25 It may have been true by September that, as the Governor argued in his 8 October 1992 speech (see note 9), ‘raising UK interest rates, when the economy was so weak and inflationary pressure so subdued—especially in the context of the debt overhang—would have been regarded not just by the financial markets, but by the world as a whole, as transparently perverse. . . . far from adding to credibility, [raising rates] was always likely to bring—indeed in the event it did bring—the latent pressure to a dramatic climax.’ But that situation was one essentially created by the monetary authorities themselves by their behaviour over the preceding months. 26 These models obviously incorporate some theory of exchange rate determination, but their focus is on the nature of the crisis. See in particular Krugman (1979), Obstfeld (1986) and references in the latter. Sutherland (1995) emphasises the downgrading of foreign exchange reserve losses, the incorporation of stochastic shocks and the modelling of policy-makers’ preferences in more recent models.

The ERM Interlude, 1990–1992 81 looks at two explanations which are closer to ‘first-generation’ models: one that hinges on the monetary authorities behaving in a way that was incompatible with market equilibrium, and one that stresses the systemic and policy coordination aspects of the crises. Finally it evaluates the various approaches and concludes that a synthesis based on the incompatible policies view but subsuming some elements of other explanations offers the most convincing account of the most important aspects of the 1992–93 crises. 5.3.1 Exchange Rate Determination Explanations Among the factors cited in explanations of the EMS crises based on models of exchange rate determination the most widely discussed is German Economic, Monetary and Social Union (GEMSU). In the accelerating upheavals affecting the Soviet Union-related states in Eastern Europe at the end of the 1980s, the frontier between the Federal Republic of Germany and the German Democratic Republic was opened in November 1989. The parliamentary elections of March 1990 in the GDR were won by the ‘Alliance for Germany’, which was dominated by the (West German) Christian Democratic Union and advocated an immediate accession of the GDR to the Federal Republic. A treaty establishing economic, monetary and social union between the two countries was signed in May 1990 and came into force at the beginning of July 1990, with the introduction of the Deutschemark in East Germany. The August signing of the Unification Treaty brought the GDR formally into the FRG in October, and the first all-German elections in December 1990 were won by the CDU and its allies. This process had several major economic implications. First, the Deutschemark became the currency in the former GDR in place of the (ost-) mark, at rates of exchange which were the subject of considerable controversy.27 Second, there was a rise in private sector investment in Germany as a whole in response to the opportunities now available in the East. Third, East Germans now qualified for essentially the same social security benefits as West Germans. Fourth, there was sustained upward pressure on wages in East Germany which had initially been well below West German levels but soared towards them, apparently without regard to the much lower levels of productivity in the East. Fifth, much of East German industry was rapidly exposed as uncompetitive and was closed down, with a steep rise in unemployment. And finally there was a massive rise in the German budget deficit as a result of the flow of unemployment benefits and industrial subsidies to the East. There was much discussion at the time of the likely effects of these developments for the DM exchange rate. In particular, it was argued that unification would create a boom which would push up domestic inflation and raise interest rates (particularly if, as seemed likely, the Bundesbank reacted to prevent the incipient inflation) and therefore lead to an appreciation of the DM; at the same time the increase in demand would require a rise in the relative price of non-traded goods, which would occur more easily if the DM appreciated.28 On the other hand, Germany was likely to increase its net foreign indebtedness such that more exports would be needed in the long run to cover the increased outflow of debt servicing.29 For its impact on the EMS the best-developed model of the impact of GEMSU is that of Branson (1994), which represents a detailed exposition of his (1993) reply, drawing on earlier 27 See Hasse (1993). 28 See, for example,

Begg et al. (1990) on the pressures for DM appreciation, and Gros and Thygesen (1992: chapter 5.6) for the view that unification might not require any major exchange rate adjustment. 29 See Wyplosz (1991) and M´ elitz (1991) for a more detailed discussion of the role of foreign debt.

82 The Making of Monetary Policy in the UK, 1975–2000 work on the US dollar in the mid-1980s, to Eichengreen and Wyplosz’s (1993) dismissal of the GEMSU explanation.30 The model is a powerful but parsimonious two-country version of the portfolio balance model of exchange rate determination. As such it encompasses both the interest rate effect discussed above, which would also generate appreciation in a MundellFleming model or a Dornbusch exchange rate dynamics model, and the indebtedness effect emphasised by Wyplosz (1991) which entails a long-run depreciation. In the long run the real exchange rate is determined by equilibrium in the current account (where net exports are sufficient to offset the debt service costs of the country’s net foreign indebtedness) and equilibrium in the foreign exchange market (where the real interest differential is sufficient to offset the risk premium associated with the country’s net indebtedness). In the short run the economy moves along a saddlepath given by the arbitrage condition of equilibrium between the real interest differential on the one hand and the sum of the risk premium and the expected real exchange rate change on the other. Within this model Branson considers GEMSU as an unanticipated (but fully credible when announced) increase in Germany’s structural budget deficit. In the long run, because Germany experiences a period of current account deficits, the real exchange rate must depreciate to generate the larger net exports required to service its larger foreign debt. In the short run, however, German interest rates rise relative to US rates and there is a jump appreciation of the real exchange rate, which then depreciates along a new saddlepath towards the new long-run equilibrium. Thus the fiscal expansion involved in GEMSU (and the investment boom, which operates in the same direction) leads to a short-run real appreciation followed by a larger longrun depreciation. The assumed price inflexibility means that the real exchange rate changes have to come about through nominal exchange rate changes, hence it is predicted that the DM must appreciate in the short run and depreciate in the long run. Branson’s model is focused mainly on the DM-$ exchange rate and the German–US interest differential. The implication for the EMS is that the decision by other EMS countries to maintain their pegs to the DM meant that the DM had to appreciate in the short run even more against the dollar (and the yen) than it would otherwise have done. This meant that the other EMS countries had to accept severe deflationary pressure in the form not merely of German-related increases in their interest rates but also of the DM’s appreciation against other currencies. In Branson’s view the growing pressure for a fall in exchange rates and/or in interest rates was ‘like a balloon under increasing pressure’, bound to burst sooner or later; in the event, ‘the timing seems to have been connected’ with the Danish and French referenda (1994: 26).31 The other main exchange-rate-based explanation of the EMS crises is the competitiveness approach put forward, among others, by De Grauwe (1994, 2000), Masera (1994), Steinherr (1994), Bacchetta (1994) and Micossi and Padoan (1994). The argument is that a number of countries were suffering from declining competitiveness, because their inflation rates were higher than the EMS average and competitiveness had not been restored by devaluations, for example Italy, Spain and Portugal (though the latter had joined the ERM only in April 1992). The other country to which the competitiveness argument has been applied is the UK, which is claimed to have suffered from inadequate competitiveness mainly because it had entered the ERM at an overvalued exchange rate. For all of these countries the growing lack 30 A different and less convincing view of the causal connection between the crises and GEMSU, set out in Vaubel (1994), is discussed in Cobham (1997). 31 The Bundesbank also took the view that exchange rates should have been adjusted earlier in the 1990s, but was unwilling to attach too much importance to GEMSU on its own (Tietmeyer, 1994).

The ERM Interlude, 1990–1992 83 of competitiveness can be presented as a background factor making depreciation eventually inevitable and therefore making an exchange rate crisis more likely; as with the GEMSU explanation the timing of the crises can be explained only by appealing to external events such as the Danish and French referenda. 5.3.2 Balance-of-Payments Crises Explanations Eichengreen and Wyplosz (1993) discussed two models from this literature that might be thought to capture something of the EMS crises of 1992–93. The first of these is what might be called the ‘expected policy change model’, the type of model developed by Flood and Garber (1984) in which forex market expectations are driven not by the current fundamentals of the exchange rate but by expectations of a coming change in monetary policy. In the EMS case it can be suggested that the deepening recession in the non-German EMS economies, at a time of unusually high German interest rates, led the markets to believe that the authorities in these countries would shortly shift away from pegging to the DM towards reflation, with a monetary relaxation which would imply abandoning the peg to the DM, that is switching from a fixed exchange rate regime to a floating one. A more fully worked out application of such a model to the EMS crises is given by Ozkan and Sutherland (1994) in which the focus is on interest rates rather than unemployment. National authorities are thought of as maximising a welfare function which depends on domestic output with the latter affected by interest rates. Domestic interest rates in turn are affected by external shocks. There is some threshold level of foreign interest rates at which the government’s optimising decision is to abandon the fixed exchange rate and lower its own interest rate, and the private sector knows what this threshold is. As foreign interest rates push up domestic interest rates, the probability of that threshold being reached increases, and the consequent probability of the fixed rate regime being abandoned is incorporated in the risk premium, which makes domestic interest rates rise more rapidly and accelerates the crisis. A related argument is developed and applied to the UK by Masson (1995a, b), drawing on Drazen and Masson (1993). The basic idea is that credibility, measured by the long-term interest differential, depends partly on the government’s actions, since non-devaluation increases the probability that the government is ‘tough’ (less likely to devalue) rather than ‘weak’, but also on the level of unemployment, since higher unemployment makes it more likely that a further shock to unemployment will push the costs of non-devaluation above some critical point (which itself depends on the character of the government), and therefore makes devaluation more likely whatever the government’s preferences. The probability that the government is ‘weak’ rather than ‘tough’ is updated by a Bayesian process which implies that, so long as the government continues not to devalue, this element must decline. Masson produces empirical estimates which he interprets as indicating that the perceived probability that the UK government was ‘weak’ fell from around 0.7 in October 1990 to around 0.2 by the end of 1991 and less than 0.1 in August 1992; and that the interest differential which would exist if the government was believed to be ‘tough’ rose, due to the rise in unemployment, from around zero in October 1990 to nearly 0.8 in August 1992. The second model discussed (and preferred) by Eichengreen and Wyplosz is what might be called the ‘contingent policy change model’, the type of model developed by Obstfeld (1986) in which the markets believe that official policy will change if a speculative attack occurs since that will alter the balance of costs and benefits from maintaining the peg; thus there can be multiple equilibria and speculative attacks can be self-fulfilling. In the EMS case the crucial

84 The Making of Monetary Policy in the UK, 1975–2000 factor is that the Maastricht Treaty specified as an entry qualification for EMU the criterion that a country should have kept its exchange rate stable within the normal EMS bands for two years without severe tensions. The inclusion of this criterion means that if a country experiences a devaluation, or even just severe tensions, it may no longer qualify for entry and the benefits of maintaining the peg are sharply reduced.32 The Ozkan and Sutherland version of the expected policy change model and the contingent policy change model can encompass countries whose fundamentals are generally agreed to have been sound, such as France, as well as those with obvious difficulties such as Italy. They are also broadly consistent with many of the findings of Eichengreen and Wyplosz’s (1993) survey of European forex traders. These findings indicate the importance to such traders as reasons for the 1992 crisis of high German interest rates, of the apparent lack of public support for Maastricht, and of persistent inflation in some countries. They also indicate that doubts were first provoked by the Danish referendum result, and show the existence of contagion (that is, the observed weakness and depreciation of one currency led to expectations of difficulties for other currencies). Finally, they reveal the low emphasis given by forex traders to unemployment in assessing the prospects for a particular currency, which is taken by Eichengreen and Wyplosz as evidence in favour of the contingent rather than the expected policy change model. M´elitz (1994) provided a rather different explanation which is focused on the behaviour of the French authorities in the run-up to the July 1993 crisis. M´elitz assumes that the equilibrium exchange rate is determined by some (constant) fundamental factors, and the actual rate adjusts towards the equilibrium rate, e.g. along a saddlepath such as in the Branson model. He also assumes an arbitrage condition with the interest rate differential between French franc interest rates and those on non-EMS currencies equal to the sum of the expected exchange rate change and the risk premium (though in his short-run analysis the risk premium is in effect constant).33 If for domestic reasons the French authorities want to see the franc interest rate reduced below the sum of the foreign interest rate and the risk premium, this can occur only if either there is a fall in the foreign interest rate (which, since this is outside the EMS and outside French influence, can be excluded), or the franc is expected to appreciate. For the franc to be expected to appreciate when the equilibrium exchange rate is unchanged requires either that the German monetary authorities should lower their interest rates so that all the EMS currencies can depreciate vis-`a-vis non-EMS currencies, or that the franc should depreciate on its own against the DM and the non-EMS currencies. What seems to have happened in June and July 1993 is that the French government embarked on a policy of lowering domestic interest rates, with no intention of abandoning the DM peg but with the hope that the German authorities would accept a reduction in their rates. Initially when the French money market intervention rate went below the corresponding German rate on 18 June there was little effect, but the franc soon began to weaken.34 Growing economic and political tensions associated with the French authorities’ attempts to pressure the Germans into lowering their rates and the Bundesbank’s refusal to do so eventually precipitated the crisis of 32 Moreover, according to Eichengreen and Wyplosz, the national monetary authorities are more likely to draw the conclusion that it is not worth resisting an attack (a) if EMU is likely to come into operation sooner, so that there is less time for a country to restore its reputation after a devaluation, and (b) if countries that are not members at the start find it more difficult to join later—as argued by Alesina and Grilli (1993). 33 M´ elitz’s analysis is conducted in first difference terms so that the risk premium is not explicitly identified. 34 See M´ elitz (1994, especially figure 4.3), the statistical evidence presented in Girardin (1994) and the brief account of the crisis in Bell (1995).

The ERM Interlude, 1990–1992 85 29–30 July that led to the decision to widen the permitted margins of fluctuation in the ERM from 2.25% to 15% on 2 August. The implication of this analysis is that the French authorities made a serious error in trying to impose on the financial markets lower French interest rates without a franc depreciation—a combination that was incompatible with equilibrium—thereby creating uncertainties in a situation where exchange rates could move in one direction only and ‘coordinating’ forex market expectations in that direction. The relevance of this sort of explanation for the UK is discussed in the next section. However, it seems less applicable to Italy and to the three Nordic countries, Finland, Sweden and Norway. The latter, although not members of the EMS and therefore not entitled to draw on its credit or other facilities, had been pegging to the ECU for varying periods before September 1992 but were all floating by the end of the year. Each of these countries was suffering from severe budgetary imbalances, and Finland and Norway were undergoing banking crises. For Italy (and to a lesser extent for the others) it could be argued that the failure of the government to grasp the nettle of fiscal consolidation allowed uncertainty to grow in the financial markets. But there was no incompatibility of policy here of the kind emphasised by M´elitz for France, where the authorities were trying to impose on the markets something inconsistent with the short-run arbitrage equilibrium between the interest differential and the sum of the expected depreciation and the risk premium. For Italy and the Nordic countries it seems not that the authorities gave a confusing ‘steer’ to the forex markets, but that they were unable to give an effective steer of any kind. While all of the work surveyed so far treats the crises as a number of separate bilateral crises between the DM and other EMS currencies, Buiter, Corsetti and Pesenti (1998a) have analysed the crises as systemic phenomena. Their model (developed more formally in their 1998b) builds on the international policy coordination literature and includes structural policy spillovers both between centre and periphery and between the various periphery countries. In particular, they analyse the effects on aggregate demand in one of the periphery countries of monetary expansion and depreciation in another. They assume that the (negative) expenditureswitching effect is outweighed by the (positive) expenditure-changing effect which results from the centre country reducing its interest rate to maintain its demand and output in response to its own real appreciation. Faced with the domestic tensions created by GEMSU,35 Buiter et al. argue that the Bundesbank was unwilling to compromise on price stability, but the periphery countries were unable to agree on how to distribute between them (with or without compensation) the perceived costs to credibility of devaluation against the DM. Had they been able so to agree, a relatively small but general devaluation would have been enough to defuse the tensions, bring German interest rates down and preserve the credibility of the system. But instead the EMS countries ended up, by default, with uncoordinated and therefore larger devaluations (under crisis conditions) of a subset of the periphery countries, and eventually in July 1993 with a recognition that the ERM had ceased to function as a policy coordination device. Under this interpretation the failure to agree a general realignment in early September 1992, followed by the smaller than expected devaluation of one country (Italy) on 14 September, made clear to the markets the extent of the coordination failure and hence brought about a revision of expectations that led to the waves of speculation over succeeding months. 35 See von Hagen and Strauch (1999) for a discussion of the fiscal policy decisions that led to the adverse policy mix in Germany at this time.

86 The Making of Monetary Policy in the UK, 1975–2000 5.3.3 Evaluation and Synthesis The two explanations based on models of exchange rate determination are open to criticism at a number of points. The Branson GEMSU explanation is focused primarily on the exchange rate between the DM and the dollar (and yen) rather than that between the DM and the other EMS currencies. However, as M´elitz (1994) showed, the French authorities at least had plausible reasons for wanting to retain the peg with the DM when the German authorities first raised the possibility of a DM appreciation within the EMS in the autumn of 1990. At that point France had made substantial gains in terms of reducing inflation and improving its credibility, with the short-term interest rate premium on francs having fallen by over 2% since 1987. This improvement would have been endangered by any depreciation of the franc against the DM, with consequent effects on the course of unemployment as well as inflation. By September 1992, on the other hand, M´elitz argues that this reasoning no longer made sense. France was in the middle of a recession, its inflation rate was well below the German rate and the interest rate premium was much smaller. Thus a French devaluation might have been viewed by the markets as an appropriate policy response to an unusual external shock, and might not therefore have adversely affected the French authorities’ credibility. Indeed, Drazen and Masson (1993) have shown that in certain circumstances maintaining a tight monetary policy in the face of an adverse shock can weaken the authorities’ credibility.36 However, it should not be considered as surprising that the French authorities were apparently unwilling to gamble on the (still recent) improvement in their reputation. In any case any evaluation by non-German EMS authorities of the costs and benefits of maintaining the DM peg must depend on their perception of the length of time for which the peg would entail a serious disequilibrium; if as Branson’s model shows the short-run appreciation would be followed by a larger long-run depreciation, then any government which believed that the period of substantial overvaluation would be, say, less than two or three years could reasonably have chosen to accept the short-run costs of keeping the peg in order to obtain the longer-run benefits in terms of credibility, whereas if the period of substantial overvaluation was to be, say, ten years, abandoning the peg would be more attractive. In fact France, for example, experienced a rather small real appreciation over the period.37 In addition, while Steinherr’s (1994) argument that the shock to aggregate demand in Germany caused by unification was limited goes too far (Vines, 1994), it is clear from the changes in current accounts that in the early years of unification a large part of the increase in German demand was exported and the other EMS countries received a significant stimulus which must have offset in part the deflationary pressure they experienced through interest and exchange rate effects.38 The competitiveness explanation, on the other hand, is obviously incomplete in that it applies only to some of the countries exposed to speculative attack in September 1992. Moreover, while statistical evidence supports the claim of declining competitiveness for Italy, Spain and Portugal 36 In their two-period model a devaluation in the first period increases the expectation that the authorities are ‘weak’ rather than ‘tough’ on exchange rates (inflation) and that they are therefore more likely to devalue in the second period, but not devaluing in the first period raises unemployment in that period and, by a persistence effect, in the second period, which then lowers the minimum size of shock that would cause the authorities to devalue in the second period. Thus playing ‘tough’ in the first period may, if the signalling effect is small and the persistence effect large, actually weaken credibility in the second period. Drazen and Masson find some empirical support for this possibility in their examination of the French–German interest rate differential. 37 On IFS data the French real effective exchange rate based on relative unit labour costs fluctuated over 1988–95 between highs of 104.7 in 1988 and 104.0 in 1990 and lows of 99.5 in 1991 and 99.4 in 1992, to reach 100 in 1994 and 1995. 38 See also Tietmeyer (1994).

The ERM Interlude, 1990–1992 87 (Eichengreen, 1993; De Grauwe, 1994), it should be noted that these countries were, at least in part, using the ERM peg to assist domestic disinflation, in a way that they and other countries had done in earlier years (when exchange rates within the ERM, though more adjustable, were not typically changed by enough to compensate fully for inflation differentials). For the UK, as already discussed, the evidence of overvaluation is weaker, and in any case the choice of a high entry rate may have been a deliberate, if somewhat short-sighted, attempt to prevent domestic disinflation being offset by the Walters effect. Both GEMSU and competitiveness explanations also have difficulty in identifying both why there was no major exchange rate change before the crises, and why the crises occurred when they did. For the former point these explanations have to appeal to the idea that the forex markets’ confidence in the prospect of EMU acted as a ‘glue’ to hold the system together, and for the latter they appeal to the Danish and French referenda. However, while the ‘glue’ notion makes sense for currencies such as the French and Belgian francs it is not obvious that such perceptions would have applied to the weak-competitiveness countries (for which there had long been talk of one last realignment before EMU). And, while the referenda (including the opinion polls preceding the French result) are a convincing backdrop to the turbulence in the forex markets, it is difficult to explain all of the detailed timing by reference to them.39 Thus, it is difficult to believe that these factors are sufficient to explain why the crises occurred no sooner and no later than they did. However, the thread that links these questions is the inherent weakness of exchange rate determination models. Despite two decades of effort economists do not have good predictive models of the exchange rate, and the evidence shows that substantial misalignments can continue for several years. In that case attempts to explain the EMS crises on the basis of arguments that existing exchange rates were or had become misaligned must be unconvincing; even though it may be possible to tell an attractive story ex post it is typically not possible to reject alternative counterfactuals. Thus, while analyses such as that of Begg et al. (1990) could argue that the asymmetric shock of GEMSU would be better handled by a nominal DM appreciation than by higher German inflation with unchanged ERM exchange rates, they could not and did not predict crises in 1992 or 1993. Moreover, the fact that the widening of the bands after the 1993 crisis was not followed by major interest rate cuts and substantial depreciations by countries such as France makes clear that GEMSU did not have to lead inexorably to an appreciation of the DM vis-`a-vis the other EMS currencies. The various balance-of-payments crises explanations discussed all include an emphasis on short-run perceptions in forex markets and allow for crises to occur even if the fundamentals are sound. They are therefore not subject to the same strictures as the exchange rate determination models. One difficulty for the expected and contingent policy change models is that unemployment and interest rates rose to high levels in countries that were not subjected to speculative attack as well as in those that were. A more important problem for these models is that some of the national monetary authorities concerned were in no sense tempted to switch regimes, so that rational speculators should not have anticipated a switch. For example, the UK government in September 1992 and the French authorities in July 1993 had staked their reputations on 39 For example, the 20 September 1992 French referendum result of a small but positive majority for the Maastricht Treaty seemed to have no impact on the forex markets, and pressure on the franc eased only on 23 September when the Banque de France and the Bundesbank implemented a coordinated interest rate rise/cut. More importantly, Rose and Svensson (1994) found that realignment expectations did not rise until shortly before the September 1992 crisis, which implies that financial markets as well as policy-makers were surprised by the events.

88 The Making of Monetary Policy in the UK, 1975–2000 maintaining their ERM parities. Moreover, what is publicly known of the EU Finance Ministers’ negotiations over the weekend of 31 July – 1 August suggests intense and widespread opposition to a generalised depreciation of other currencies against the DM,40 and as it turned out the widening of the bands did not lead to a significant change in policy in countries like France. In the UK the government paid a huge political price for its commitment; since that price was predictable in advance the sincerity of the government’s desire before 16 September 1992 to maintain the peg cannot be doubted. It is also difficult to believe that a number of the other governments involved, notably the Italian and Nordic governments, were undertaking the careful cost–benefit analysis assumed in these models. On the contrary, they regarded a regime switch as a serious failure to be avoided if at all possible.41 A related problem is that in these models financial markets form expectations about government policy on the basis of their observations of a limited range of variables which excludes, in particular, both any signals coming from the low-level, day-to-day actions of the monetary authorities in the money and forex markets and the public announcements of government ministers and central bankers. These factors are difficult to quantify and to model formally, and they include a large element of ‘froth’, but there can be little doubt that they affect expectations in the very short run, that is over the period which is of crucial importance for exchange market crises.42 It is revealing that Masson’s (1995a) brief descriptive account of the UK’s membership of the ERM does not mention how the reluctance of the UK monetary authorities to raise interest rates became increasingly obvious through the summer of 1992 in a way that must have contributed to the markets’ doubts about the sterling–DM peg.43 The ‘incompatible policies’ explanation of M´elitz (1994) avoids the awkward problem of forex market expectations turning out ex post to be irrational. In addition, the fact that national monetary authorities are seen as having themselves created that uncertainty explains the coordination of the expectations of atomistic dealers and investors, and avoids the implausible attribution of the entire blame for the crisis to speculators. At the same time the M´elitz approach may be broadened to subsume the most attractive elements of the policy change models: the links between the state of the economy, the probability of a regime change, and the markets’ perception of that probability; and the way in which speculative pressure may itself affect that probability. It can also be seen as retaining a key element of the Krugman (1979) model—the idea that the initial cause of a crisis is inappropriate policy (excessive domestic credit expansion with fixed exchange rates)—but incorporating this in the context of an asset market view of exchange rate determination. The systemic explanation of Buiter et al. (1998a) is attractive precisely because it is systemic, and therefore explains phenomena largely ignored by the other explanations. It is also not open to the criticism that the forex markets’ expectations turn out to be irrational, and it identifies a trigger for the revision of expectations in September 1992. In addition, it illuminates the 40 See Marsh (1993), Goodhart (1993) and Sherman and Kaen (1994). 41 The UK government (and the Italian and Nordic governments) did

of course accept depreciation in the end, but that does not prove that the crises were initiated by speculative activity based on the understanding that the governments would choose to switch regime if attacked; it merely shows that governments can sometimes be forced to switch. 42 Avesani et al. (1995) take much more account of contemporary market perceptions (as gleaned from the Economist), and their model produces an interesting alternative measure of credibility which they show is much closer to market perceptions than Svensson’s (1991) ‘simplest measure’ of target zone credibility. 43 Insofar as rising unemployment and the apparent absence of economic recovery made it politically more and more difficult for the UK authorities to raise interest rates, it needs to be emphasised that the UK government had done a great deal to box itself into this particular corner.

The ERM Interlude, 1990–1992 89 linkages between the various member countries of the system, and in particular the relation between the centre country’s interest rate and its real exchange rate. However, it leaves no role for decisions by the periphery country national monetary authorities other than their initial reluctance to devalue. M´elitz’s explanation, on the other hand, focuses on those decisions in one particular case, and takes the high level of German interest rates in this period as given. However, as Branson and Ozkan and Sutherland as well as Buiter et al. have stressed, German interest rates were high because of GEMSU and the way in which it was implemented.44 In that case it is possible to say something more precise about the relationship between background factors and triggers. On the one hand, a crisis could have occurred along the lines of M´elitz’s analysis whatever the level of interest rates, so long as the French (and perhaps other national monetary authorities) were unwilling to accept that level. Indeed, the franc’s ERM parity survived the September 1992 crisis which coincided with the peak in German interest rates, but could not be maintained in July–August 1993, when German interest rates were both much lower and on a well-established declining trend. On the other hand, even higher German interest rates would not have produced a crisis if the French and other authorities had accepted the constraints facing them: if they had properly understood the choice between (a) high interest rates but maintaining the peg and (b) lower interest rates and abandoning the peg, they would either have accepted earlier German suggestions of exchange rate changes or have accepted the inevitable cost of keeping the peg. This suggests that while Buiter et al. provide a convincing analysis of the systemic difficulties which emerged in the wake of GEMSU, primary responsibility for the speculative crises themselves should be attributed to the trigger factor of incompatible policy behaviour by (non-German) national monetary authorities rather than to background factors such as GEMSU or the technical details of the operation and performance of monetary policy in Germany and elsewhere.

5.4 THE UK CASE IN THE LIGHT OF OTHER EXPLANATIONS A number of points can now be made about the UK’s experience in the ERM. First, the UK was not obviously suffering a lack of competitiveness that was either as large as that faced by Italy, Spain or Portugal, or sufficient to cause a forced realignment: the UK was much less uncompetitive in 1992 than in the early 1980s, for example,45 and in the past significant misalignments of currencies have persisted for considerable periods before pressures to correct them developed. Second, while German unification (given the way it was implemented) was clearly the underlying cause of the high German interest rates, the account in Section 5.2 emphasises that the UK monetary authorities made specific choices (from a range of alternatives) when faced with these high German interest rates: German unification may have been an important background factor but it was not the trigger of the sterling crisis, nor on its own a sufficient cause of that crisis. Third, with respect to the expected and contingent policy change explanations, in the UK’s case it is clear both that the authorities did not envisage any change in policy, and that they were not optimising in the sense of continuously counting the costs and benefits of alternative 44 Filc (1994) argues that the problem was exacerbated by the Bundesbank’s preoccupation with a misleading monetary aggregate which led it to deflate too rapidly. However, for reasons given later in this paragraph, factors such as this, or the Bundesbank’s delay in raising interest rates during the course of 1990, or the Bundesbank’s day-to-day actions in 1992 or 1993, should be regarded as of minor importance. 45 See Chapter 8 below.

90 The Making of Monetary Policy in the UK, 1975–2000 strategies and standing ready to switch between them when the relative net benefits changed. Indeed, they were so committed to remaining within the ERM at an unchanged parity that they failed to develop any fallback strategy at all. To convince the markets of the seriousness of their intentions they refused to consider publicly any alternative, but they were also refusing to take actions such as raising interest rates in good time, which would have been consistent with that rhetoric. And the fact that it all ended in a crisis, which was clearly perceived to be a disaster and a humiliation for the government, shows that they had not developed any alternative strategy in private either. Fourth, the systemic explanation of Buiter et al. (1998a) identifies a trigger for the general revision of expectations about the stability of the ERM only two days before Black Wednesday, but it cannot explain why sterling was a particular focus of speculative pressure, or why sterling had already weakened in the weeks preceding this. Instead, the actual sterling crisis of September 1992 seems best explained along lines similar to M´elitz’s (1994) discussion of the French franc crisis of July 1993: the UK authorities wanted lower interest rates without depreciation, and these two objectives could not be realised together with equilibrium in financial markets. The 1992 UK case differs from the 1993 French case in that German interest rates were at their peak in September 1992 but on a declining trend in July 1993. In addition, the UK authorities might just have avoided disaster if external events (the French referendum and the associated opinion polls, the turmoil hitting other ERM and non-ERM currencies, even the interviews given by the Bundesbank President46 ) had been different. On the other hand the French authorities were making a more frontal challenge to the German authorities and could have avoided disaster only in the highly unlikely eventuality that the Bundesbank gave in to pressure at its Council meeting on 29 July 1993. Nevertheless the UK case resembles the French in that the crisis can be seen as having been provoked essentially by the attempts of the national (non-German) monetary authorities to obtain lower interest rates without depreciation. The UK (and later the French) monetary authorities should have prepared an appropriate contingency plan, which might have involved a depreciation of the currency but would have made that orderly and controlled, and would have preserved at least part of the authorities’ face and credibility. Finally, it may be asked how the UK authorities could have believed that there was a possibility of pressurising the Bundesbank, whose thinking both in general and on EMS exchange rate and interest rate issues in this period was well known. There is a great deal of contact at a range of levels between the various European central banks and finance ministries. It seems highly unlikely that officials in the Treasury and the Bank of England could have been so ignorant of the attitudes and behaviour of the Bundesbank, or that they could have failed to give appropriate advice. An alternative explanation may be that the UK political authorities (notably the Chancellor, Lamont) were influenced by attitudes towards Germany and the Germans of the kind whose public expression led to Nicholas Ridley’s resignation in July 1990,47 and/or by the longstanding attachment of British politicians (particularly Conservative ones) to a ‘Westminster sovereignty’ view of the political process in which the government 46 It has sometimes been suggested that Schlesinger’s interview with Handelsblatt, an abridged version of which was distributed to other German newspapers on the evening of 15 September, was intended to force sterling out of the ERM. But well-informed journalists regarded Schlesinger as unworldly rather than Machiavellian (Norman and Barber, 1992). And in any case a sensible strategy from the UK’s point of view would have been immune to minor events of that sort. 47 Ridley was Trade and Industry Secretary at the time, but had been in the Cabinet since the early 1980s and was a close ally of Thatcher. He made a series of remarks in an interview with the Spectator magazine which were interpreted as xenophobically anti-German.

The ERM Interlude, 1990–1992 91 is supposed to have virtually absolute and unconstrained power so long as it can carry its own supporters in the House of Commons. If that explanation is correct, the UK’s exit from the ERM can be seen as more of a political, rather than an economic, failure.

5.5 A COUNTERFACTUAL ANALYSIS Section 4.4 of Chapter 4 used the reaction functions estimated by Adam, Cobham and Girardin (ACG) (2001) to provide a counterfactual analysis of policy between 1985 and 1990. A comparable analysis can now be offered for the ERM period, using the policy rules which ACG estimated for the pre-ERM, post-ERM and MPC subperiods and their simulated Taylor rule, which are graphed in Figures 4.6 to 4.9. As discussed in Chapter 4, the preferred reaction function for the pre-ERM period has the interest rate responding only to German and US interest rates, that for the post-ERM (1992–97) subperiod has the rate responding to both these international variables and the standard domestic variables (inflation and the output gap), and that for the MPC period has the rate responding to the domestic variables, with the international variables appearing only as instruments for forecasting the domestic variables. There are two main points to be made. First, while all of the rules predict a sharp fall in the interest rate between mid- or late 1990 and late 1992, the rate implied by the pre-ERM rule is consistently and increasingly above the actual, with a gap of over 4% between August 1991 and September 1992.48 This suggests that entry into the ERM may have provided a credibility gain which allowed the monetary authorities to lower the interest rate more than they would have been able to do if policy had continued to be made on the previous basis, and to take more account of domestic variables than they were able to do before. In that case the UK’s entry itself should be counted as a success. Second, the rates implied by the other three rules go below the actual, in late 1990 for the post-ERM rule, in mid-1991 for the Taylor rule, and in early 1991 for the MPC rule; and they all predict a continued fall in the rate after mid-1991, whereas the actual rate was more or less stable between then and September 1992 (after the official interest rate cut of 0.5% in September 1991 there was only one further cut before Black Wednesday, also of 0.5%, in May 1992). On the MPC rule interest rates would have fallen as much as 6.84% below the actual (in August 1992), whereas for the post-ERM and Taylor rules the divergence was rather smaller: it peaked in the former case at 4.77% in June and in the latter case at 3.45% in August.49 This suggests that membership of the ERM did indeed prevent the authorities from cutting rates in the way that later policy would have implied. However, it should be stressed that the structural constraints on monetary policy in 1990–92, notably the level of credibility, were quite different from those that existed in later subperiods, so the UK authorities might well not have been able to behave at this time as indicated by the later policy rules even if the UK had not been in the ERM.50 To put these numbers in perspective, the policy rate fell from 15% immediately before ERM entry in October 1990 to 10% in May 1992 (and remained at 10% until Black Wednesday). As regards the timing, the BEQB argued in August 1991 that ‘[w]hile the ERM represents a 48 This gap may be compared with the standard error for the pre-ERM equation (over the period for which it was estimated) of 0.50%. 49 These gaps may be compared with the standard errors for the post-ERM and MPC equations of 0.16% and 0.20% respectively. 50 Actual interest rates were well above the levels implied by the MPC and Taylor rules for most of the post-ERM period as well.

92 The Making of Monetary Policy in the UK, 1975–2000 potential constraint on policy, its requirements have not diverged from what has been seen as appropriate for domestic objectives’, and in November 1991 that the narrowing of the UK– German interest differential ‘reduced the scope for a further fall in domestic interest rates. However, the substantial reduction in interest rates which had occurred since sterling joined the ERM had already provided a significant stimulus to demand’ (BEQB, 1991c: 344–5, 1991d: 478). Thus the authorities themselves seem at the time to have taken the view that the ERM began seriously to constrain UK interest rates only towards the end of 1991, rather later than suggested by the counterfactuals. Thus, the counterfactual econometric evidence suggests that ERM entry itself improved the authorities’ ability to lower interest rates in response to the deepening recession and falling inflation—developments which, because of the lags involved, should be attributed largely to the high interest rates prevailing since late 1988, which were in turn the response to the loss of control over inflation in the Lawson boom. However, the improvement in credibility was not sufficient to allow UK interest rates to fall as much as they would have done on the basis of the post-ERM or MPC policy rules (which would have entailed UK rates going well below German rates) while the UK remained in the ERM. In that case the crucial questions which the authorities should have addressed were how large and how long a ‘sacrifice’ continued ERM membership at the existing parity would impose, and whether the benefits (in terms of exchange rate stability, further improvements in credibility and—if they wished—easier entry into EMU) would be worth the pain.

5.6 CONCLUSIONS The UK entered the ERM in 1990 in search of a better framework for monetary policy than, and an alternative nominal discipline to, the discarded framework of monetary targets. However, the details of entry—the timing, the parity and the associated interest rate policy—were dominated by short-term considerations. More importantly the politicians, in particular, do not seem to have really understood the constraints they were assuming. The period of ERM membership included a sharp fall in inflation (though probably little of this was due to entry itself ), but it ended in crisis and exit because the UK monetary authorities were ultimately not prepared to accept either of the alternatives implied by the circumstances of late 1992—that is, higher interest rates or devaluation against the DM (preferably together with other members of the ERM). Instead, they pursued an extremely high-risk strategy which had a small a priori chance of success but was always more likely to lead to a crisis in which the UK’s chosen ERM parity could not be maintained. Thus the crisis of Black Wednesday should be seen as the almost inevitable result of the failure of the UK monetary authorities (taken as a single agent) to understand the full benefits and costs involved in membership of the ERM, and their failure to understand how their own behaviour needed to be modified for membership to be sustained. Finally, as regards the monetary framework used in this subperiod, there were mistakes made in the entry into the exchange rate targeting regime, but they were not fatal. The subperiod turned out to be one of difficult external circumstances (associated with GEMSU) which could not have been entirely foreseen at the time of entry. But the key factor responsible for the failure of the framework was the monetary authorities’ inability to accept the need for some decisive action in response to the increasing tensions, either to devalue in time (preferably along with other currencies) or to raise interest rates in the summer of 1992. Insofar as that inability reflected the domestic political opposition to ERM membership, particularly in the Conservative party, which the government had not confronted with much vigour because of its own ambivalence, the failure was political as well as economic.

6 The 1993–1997 ‘New Framework’ for Monetary Policy On 16 September 1992 sterling was ‘suspended’ from the ERM, but it rapidly became clear that the authorities had no appetite for an early resumption of membership. Instead, they set about constructing an alternative framework for monetary policy to replace the monetary and then exchange rate targets which were seen as having failed. Section 6.1 of this chapter outlines the new framework, Section 6.2 provides a partial evaluation, Section 6.3 discusses some econometric research on the credibility of the new framework, and Section 6.4 discusses a monetary policy reaction function for this phase, contrasting it with policy rules for other phases. Section 6.5 concludes.

6.1 THE NEW MONETARY POLICY FRAMEWORK Black Wednesday allowed a major ‘rebalancing’ of economic policy. Within a few weeks short-term interest rates had been reduced by 3% relative to the day before Black Wednesday, and sterling had depreciated by 10–15%. At the same time the authorities were obliged to put together a new framework for monetary policy to replace that which had collapsed. Some aspects of this new framework were put in place within a few weeks, but some were developed more slowly (over the course of 1993 and early 1994). The new framework can be thought of as consisting of two parts, first a policy target and a related conception of the role of other economic variables, and second a new relationship between the political authorities (the Treasury) and the central bank, and these will be discussed in turn.1 The new policy target, to take the place of the exchange rate target under the ERM and the monetary targets of the first half of the 1980s, was an inflation target: policy was to aim for inflation to be within the range of 1–4%, and in the lower half of that range (1–2.5%) by the end of the current Parliament, that is by spring 1997 at the latest. Inflation for this purpose was measured by the 12-month increase in the RPIX, that is in the index of retail prices excluding mortgage interest payments. The target was restated by the Chancellor in June 1995 as a target of 2.5% or less for the longer term beyond the end of the current Parliament.2 The focus for policy was to be the rate of inflation expected in two years’ time, two years being thought of as the typical lag between policy measures and their effect on inflation. In pursuing the inflation target the authorities would also operate a target range (later referred to as a monitoring range) for M0, the measure of the monetary base for which formal targets had been in continuous existence since 1984, and a monitoring range for M4, the broad money aggregate including 1 See also Treasury (1992), King (1994), Bowen (1995) and the Bank of England Governor’s speech on ‘Evolution of the monetary framework’ in BEQB, 1997a: 98–103. 2 An earlier Treasury statement had referred to a longer-term target of 2% or less, ‘levels that match the best in Europe’ (Treasury, 1992: 2).

94 The Making of Monetary Policy in the UK, 1975–2000 building society deposits as well as bank deposits which had replaced £M3 in the light of the blurring of the boundary between these two sets of institutions. The authorities would also have regard to the exchange rate, but no target or monitoring range was specified. The new relationship between the Treasury and the Bank of England involved a number of elements. First, the Bank was to produce a quarterly Inflation Report which would include its forecast of inflation over the next two years. This report, which was produced from February 1993 and published alongside the Bank’s Quarterly Bulletin, was from the autumn of 1993 shown to the Treasury only in its final form (King, 1994), so that no direct censorship by the Treasury was possible. Second, the meetings which had always taken place between the Governor of the Bank and the Chancellor of the Exchequer were regularised and formalised; they became the principal forum within which interest rate decisions were discussed, although the Chancellor continued to take those decisions himself. Third, as from November 1993 the Bank was given some discretion over the timing of any interest rate change which the Chancellor decided upon; this was introduced to get rid of the perennial suspicion that interest rate changes were being made for political purposes, with reductions timed to coincide with Budgets, Conservative Party annual conferences and other such events. Fourth, from April 1994 the minutes of the monthly Monetary Meetings were published two weeks after the subsequent meeting, that is, usually about six weeks after the meeting itself. The minutes took the form of an account of monetary and economic developments which was the product of a prior meeting between Treasury and Bank officials, followed by a summary of the discussion which gave first the Governor’s views (these were provided in written form for inclusion in the minutes), then the Chancellor’s views and an account of other points raised in the discussion, and lastly the Chancellor’s summing up on the question of whether interest rates should be raised or reduced. Finally, the Treasury also published (on the day concerned) a monthly Monetary Report which gave an account of the main economic and monetary data considered at these meetings, and released (again on the day concerned) press notices to explain the reasons for any interest rate changes. Given the previous perceived failure of monetary targets, the inflation target was an obvious replacement for an exchange rate target which was no longer regarded as viable. It was also a strategy which had been adopted in a range of countries in the preceding few years, mostly small open economies.3 As discussed in Chapter 1, inflation targets, like exchange rate targets, automatically take account of money demand shocks (such as were supposed to have made monetary targets unworkable in the UK in the 1980s), and they are highly visible. They may therefore have a stronger direct influence on inflation expectations than monetary targets, and they represent in principle a stronger pre-commitment. However, inflation is more difficult to control than the monetary aggregates, so that inflation targets may be more difficult to operate effectively. They may also be inferior to nominal income targets because they lead to more variability of real output in response to adverse supply shocks, although King (1994) argued that inflation targets with a target expressed as a range enable the authorities in practice to avoid the adverse impact of supply shocks. On the other hand, the new relationship between the Bank and the Treasury should be seen as originating in the acute loss of credibility by the government in the speculative crisis of Black Wednesday. The Treasury referred to some of these changes in late 1992 as ‘steps to improve the credibility of [the government’s] anti-inflation strategy’ (Treasury, 1992: 3), while Mervyn 3 For general discussion of inflation targeting see Leiderman and Svensson (1995); Haldane (1995); Lane et al. (1995); Almeida and Goodhart (1998); and Bernanke et al. (1999).

The 1993–1997 ‘New Framework’ for Monetary Policy 95 King, Chief Economist at the Bank of England, spoke of ‘institutional changes designed to bolster the credibility of the commitment to low inflation’ (King, 1994: 123). Both these sources emphasise openness and transparency, but those could have been provided without the institutional separation of responsibilities involved in the new framework: the Treasury and the Bank could have produced a joint inflation report and forecast, and more information could have been provided on the reasoning behind interest rate decisions without the device of a Bank of England recommendation that was, or was not, accepted by the Chancellor. It therefore seems obvious that what was happening was that in order to regain credibility the Treasury felt obliged, not merely to undertake to explain its policies and actions more clearly, but also to concede some limited and informal (and therefore reversible) autonomy in monetary policy to the Bank of England, whose reputation had been less adversely affected by the events of September 1992. However, it should also be noted that the shift had the effect of strengthening the Chancellor against pressure from the Prime Minister (Stephens, 1996: 292–3), and of forcing the Bank’s views out into the open in a way that could weaken, as well as strengthen, the Bank. In fact, the authorities explicitly recognised the time-inconsistency problem for the first time in November 1992 in a lecture by the Bank of England Governor (BEQB, 1992d: 441–8) which stands in sharp contrast to the Bank’s statements on the UK’s entry into the ERM (BEQB, 1990d: 482–6) and earlier general statements on monetary policy.4 The authorities could have chosen to tackle the time-inconsistency problem by making the Bank of England independent,5 but they chose not to do so. In his June 1993 resignation speech from the post of Chancellor, Lamont revealed that he had tried for two and a half years to persuade the Prime Minister to accept central bank independence, which he argued would have enabled interest rates to be lower for a given exchange rate, made policy more credible and ensured the necessary discipline. Major’s reply took the line that it was not possible to make the Bank independent without loss of Parliamentary accountability.6 When Lawson was Chancellor, he had also prepared a plan for Bank of England independence, in 1988 (Lawson, 1992: 867–73, 1059–61), but Thatcher had shown no interest.7

6.2 EVALUATION OF THE NEW FRAMEWORK A first step in evaluating the new framework is to look at the outcome in terms of the rates of inflation experienced during the subperiod. The 12-month RPIX (the targeted measure of inflation) stood at 4.0% in September 1992 (down from the peak of 9.2% in November 1990) 4 See for example Fforde (1983); BEQB, 1986d: 499–507, 1987c: 365–70, 1990b: 215–20. A version of the timeinconsistency concept can be found in the discussion of monetary targets in Lawson’s November 1985 minute on the EMS (Lawson, 1992: 1055–8). However, it is difficult to find any echoes of it in statements from that or earlier periods by other policy-makers such as Thatcher, Howe and the Bank. Lawson’s (1988) minute to Thatcher on the case for an independent Bank of England (Lawson, 1992: 1059–61) does not contain a clear formulation of the concept either. In addition, the way the monetary target ranges were set emphasised short-term announcement effects rather than a medium-term commitment or discipline (see Chapter 3). These points suggest that Lawson’s 1985 minute should be seen as a highly personalised view and one reflecting his own hindsight, rather than as a description of the way in which policy was actually made or conceived. 5 This course was recommended at the time by the OECD (1993: 43). It is worth noting that in terms of the standard criteria used for measuring central bank independence the changes involved in the new framework made very little difference to the status of the Bank of England (see Chapter 7). 6 Hansard vol. 226, cols 283–4, 297. 7 See also Stephens (1996: 276–9). The Roll Committee’s report (Roll et al., 1993) outlined around this time ways in which the problem of accountability could be satisfactorily treated in the UK context.

96 The Making of Monetary Policy in the UK, 1975–2000 4

2.5

RPIX

1 Sept 92

Mar 93

Sept 93

Mar 94

Sept 94

Mar 95

Sept 95

Mar 96

Sept 96

RPI

Mar 97

Sept 97

Figure 6.1 Inflation, 1992–1997 (Source: Economic Trends, various issues)

and continued to fall more or less consistently to a trough of 2.0% in September 1994; for most of 1995 and the first half of 1996 it fluctuated around 2.8%, but it rose to 3.3% in October 1996 before falling back to 3.1% in December, 2.9% in February 1997, and 2.5% (the target figure) in April and May 1997 (see Figure 6.1). The headline RPI (which is affected by movements in mortgage interest rates) fell from 3.6% in September 1992 to a low of 1.2% in May 1993 but reached a high of 3.9% in September 1995; it then fell (under the impact of interest rate cuts) to 2.1% in mid-1996 but rose to 2.7% in October 1996, fell back to 2.5% in December but rose to 2.8% in January 1997 before declining to 2.4% in April, after which it rose again. Thus inflation continued to fall for a while despite the depreciation and went below 2.5% for most of 1994, but it was above 2.5% for every month from January 1995 to March 1997. In fact the target of 2.5% or less on the RPIX was just met in April and May 1997, but inflation began to rise immediately afterwards, reaching 3.0% in July and remaining above 2.5% for the rest of 1997. Since the time period here is obviously short, a second step is to look at inflation expectations. These are tracked in some detail in the Bank’s Inflation Report, which considers surveys of expectations, private sector forecasts and the expectations implied by existing bond yields. Both surveys and private sector forecasts indicate a sharp increase in expected inflation immediately after Black Wednesday, a gradual reduction over 1993, a less clear-cut downward trend in 1994, and some rise in 1995. By late 1995, the median private sector forecast was for inflation in 1996 Q4 to be 3%. Up to January 1997 the median forecast for 1997 Q4 remained stuck at 2.9%, and the median forecast at that point for 1998 Q4 was 3.3%. By April 1997 the 1997 Q4 forecast had declined to 2.7% and that for 1998 Q4 to 2.9%. The Bank of England’s calculations of the inflation expectations underlying market interest rates8 suggest that shorter-term expectations continued to fall after Black Wednesday, rose over most of 1994 (when world bond markets were disturbed) and declined over 1995 and part of 1996. Medium- and longer-term expectations 8 See

Deacon and Derry (1994), King (1995) and Breedon (1995).

The 1993–1997 ‘New Framework’ for Monetary Policy 97 3

2

1

0 1990M1

1991M1

1992M1

1993M1 UK-Germany

1994M1

1995M1

1996M1

1997M1

France-Germany

Figure 6.2 Bond yield spreads, 1990–1997 (Source: International Financial Statistics)

rose sharply in late 1992, fell in late 1993 but were higher in 1995 and much of 1996. As at the end of June 1996, the markets expected inflation to be 4.2% in June 1999, 4.6% in June 2001, and 4.9% in June 2006. By April 1997 inflation expectations had moderated a little, but the markets were still expecting inflation of 4.0% in April 2002 and 4.3% in April 2007.9 A simpler but less precise measure of the credibility of policy is provided by the interest differential on long-term government bonds. Figure 6.2 shows the UK–German and French– German differentials from 1990 to 1997.10 While the latter was lower than the former throughout, the UK–German differential had fallen sharply in the run-up to and during the ERM period, both absolutely and relative to the French–German differential. It then rose strongly following Black Wednesday, fluctuated through 1994 and 1995 around a level comparable to that experienced in the earlier part of ERM membership, and then rose from mid-1995 to a level comparable to that of the late 1980s. By contrast the French–German differential declined strongly after the Jupp´e–Chirac government finally committed itself to the Maastricht fiscal convergence criteria in late 1995. At the same time UK short-term interest rates, which had fallen below German rates when the UK left the ERM, rose above German rates in July 1994. UK rates had been below French rates throughout 1992; they went above them briefly in late 1994 and then more lastingly from July 1995.11 Thus inflation was lower than in previous phases, but the reduction seemed less secure towards the end of the phase. Inflation expectations, and credibility more widely, which had been badly affected by the exit from the ERM, improved considerably in the first year or so 9 Data supplied by the Bank of England. But see Breedon (1995) for the tendency of these estimates in the past to over-predict inflation, possibly because of an inflation risk premium. 10 Monthly data on 10-year government bond yields from International Financial Statistics. The UK–German spread had fallen rather unevenly from a peak of over 8% in late 1976 to an average of 2.8% in 1983–84, risen to a peak of 4.6% in late 1986 and fallen back to an average of 2.5% in 1989. 11 Treasury bill rates from International Financial Statistics.

98 The Making of Monetary Policy in the UK, 1975–2000 after exit; they then deteriorated for at least part of 1994 and 1995, and expectations remained stubbornly above the government’s target. At the same time economic growth was broadly satisfactory with an average of 3.0% for 1993–96 (which included the recovery from the recession of 1990–92) but a possibly worrying acceleration to 3.5% in 1997, while unemployment fell from the peak of 10.5% in 1993 to 8.2% in 1996 and 7.0% (the level of the previous trough in 1990) in 1997 (see Table 2.1). However, the post-ERM period is too short for a sound judgement of the outcome of this framework for monetary policy to be made: in particular the new arrangements were never subjected to strong pressures of the kind experienced in the Lawson boom, on the one hand, or the ensuing recession, on the other. A direct examination of the process by which these arrangements operate is therefore in order. The Inflation Report produced by the Bank of England rapidly established a solid reputation for itself, partly because of its thoroughness and partly because of its method. The latter was in line with mainstream macroeconomic thinking on inflation, which emphasised that inflation may be affected in the short term by a range of factors including costs, but is determined in the medium and long term largely by demand and monetary factors. Over this subperiod the report starts with an account of recent developments in inflation, from the various measures of retail prices through producer output prices to expenditure deflators; its second section examines developments in money and interest rates, covering narrow, broad and Divisia money, credit, interest rates and exchange rates; the third section examines the growth of domestic and external demand and of output; the fourth looks at the labour market, from earnings through unemployment to wage expectations and productivity; the fifth examines price dynamics, from the exchange rate–prices pass-through to profitability and margins; and the sixth gives the Bank’s medium-term inflation projection, alongside survey and private sector forecasts and implied financial market expectations. Thus the new arrangements clearly embodied elements of greater openness which should have made it easier for the private sector to understand and predict government behaviour. The allocation to the Bank of England of responsibility for the short-term timing of interest rate changes might also have contributed to the depoliticising of monetary policy, though in practice the Bank’s discretion here was limited (BEQB, 1997a: 102). An important change was made to the Bank’s published inflation forecasts in February 1996. Before that date the Inflation Report had presented the Bank’s ‘central projection’ for the next two years with a range either side based on the mean absolute error on inflation projections since 1985. In February 1996 the Bank replaced that presentation with a ‘fan chart’. This had a central band coloured deep red, which included the (not separately identified) central projection but was wide enough for there to be a 10% chance that inflation would turn out to be within the band. On either side of this central band there were further bands coloured successively less deeply red, showing the ranges within which inflation was expected to be with a probability of 20%, 30%, 40%, etc. (BEQB, 1996a: 46–8). The aim of the change was to ‘convey to the reader a more accurate presentation of the Bank’s subjective assessment of medium-term inflationary pressure, without suggesting a degree of precision that would be spurious’ (BEQB, 1998a: 30). While there were good reasons for reducing the attention paid to the central projection alone, however, the effect of the change was also to make it more difficult for outside observers to assess the Bank’s forecasting record, and that is one reason why no such assessment is attempted here.12 12 Other reasons concern the problem of taking account of (unpredictable and/or unpredicted) exogenous factors affecting the outturn and the short time-periods both for this framework and for the fan chart procedure. For further detail on the fan chart see the article in BEQB, 1998a: 30–7.

The 1993–1997 ‘New Framework’ for Monetary Policy 99 The core of the new framework, however, must be the way in which interest rates are set. Here the minutes of the monthly Monetary Meetings provided more information on monetary policy than had been available in the UK before. In principle, they should have shown on what grounds interest rates were being set and, particularly important for credibility, the extent to which the central bank was in agreement with the decisions taken by the Chancellor. Table 6.1 Table 6.1 Monthly Monetary Meetings, 1994–1997 Meeting 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

Date

Classification

12.1.94 2/3.2.94 2.3.94 30.3.93 4.5.94 8.6.94 6.7.94 28.7.94 7.9.94* 26.9.94 2.11.94 7.12.94 28.12.94 2.2.95 8.3.95 5.4.95 5.5.95 7.6.95 5.7.95 27.7.95 7.9.95 29.9.95 1.11.95 13.12.95 17.1.96 7.2.96 7.3.96 3.4.96 8.5.96 5.6.96 3.7.96 30.7.96 4.9.96 23.9.96 30.10.96 11.12.96 15.1.97 5.2.97 5.3.97 10.4.97 6.5.97

B2–C2 B2–C1 B2–C2 B2–C2 B3–C3 B4–C4 B4–C4 B4–C4 B5–C5 B3–C3 B4–C4 B5–C5 B3–C3 B5–C5 B4–C4 B4–C3 B5–C4 B5–C4 B5–C4 B4–C4 B4–C3 B4–C4 B4–C4 B1–C1 B2–C1 B3–C3 B1–C1 B3–C3 B3–C2 B3–C1 B3–C3 B4–C3 B4–C3 B5–C4 B5–C5 B5–C4 B5–C4 B5–C4 B5–C4 B5–C4 B5–C5

Interest rate change −0.25

+0.5 +0.5 +0.5

−0.25 −0.25 −0.25 −0.25

+0.25

+0.25

Resulting interest rate level 5.5 5.25 5.25 5.25 5.25 5.25 5.25 5.25 5.75 5.75 5.75 6.25 6.25 6.75 6.75 6.75 6.75 6.75 6.75 6.75 6.75 6.75 6.75 6.5 6.25 6.25 6.0 6.0 6.0 5.75 5.75 5.75 5.75 5.75 6.0 6.0 6.0 6.0 6.0 6.0 6.25

Source: Minutes of the Monthly Monetary Meetings. Note: *The Chancellor decided on the change only on 9 September 1994. Key to Table 6.1: C = Chancellor of the Exchequer, B = Bank of England Governor. 1 = wants cut in interest rates, 2 = bias towards cut but not justified in this case, 3 = no bias for or against cut/rise, 4 = bias towards rise but not justified in this case, 5 = wants rise in interest rate.

100 The Making of Monetary Policy in the UK, 1975–2000 Table 6.2 The degree of agreement between Governor and Chancellor, January 1994 to May 1997 B1 B2 B3 B4 B5

C1 2 2 1

C2 3 1

C3 6 4

C4

C5

8 9

5

The entry in each cell indicates the number of Monthly Monetary Meetings at which the views on interest rates expressed by the Governor and Chancellor were those defined by the rows and columns concerned. Source: Table 6.1.

provides a classification of each of the 41 meetings for which the minutes are available (January 1994 to May 1997). It identifies the views of the two protagonists on what should happen to interest rates in terms of five categories: the unequivocal recommendation of a cut in interest rates; a ‘bias’ towards a cut, that is the recognition that it is a cut rather than a rise which should be considered, but where it is argued that a cut is not currently justified; ‘no bias’, that is no preference in either direction; a ‘bias’ towards a rise, that is the recognition that it is a rise rather than a cut which should be considered, but where it is argued that a rise is not currently justified; and the unequivocal recommendation of a rise. The results are summarised in Table 6.2, which shows the number of meetings for each possible combination of the views of the Chancellor and the Governor. If they had always agreed, all meetings would be classified as lying along the diagonal, but in the table there are 17 off-diagonal meetings, in all of which the Governor took a ‘harder’ line than the Chancellor. Agreement on 24 out of 41 might be considered impressive, but the fact that all the disagreements are in the same direction suggests some systematic difference of attitude. A brief account of policy over the period is therefore useful (see Figure 6.3 for the movement of the policy rate). The post-Black Wednesday interest rate cuts had taken minimum lending rate to 8% by mid-October, 7% by mid-November, and 6% by January 1993, and a further cut of 0.5% was made in November 1993 in the light of the evidence of moderate growth and low inflation. The Chancellor made a further cut of 0.25% in February 1994, a cut opposed by the Governor who was more cautious on inflation and more confident about growth.13 By September growth was stronger and inflation still subdued, but the Chancellor and Governor agreed on a preemptive rise of 0.5%, a measure which was repeated in December and February 1995. Three months later the Governor was worried about a fall in sterling and pressed for a further rise in rates, but the Chancellor, basing his judgement on ‘a broad assessment of the economic data, rather than market expectations’, argued that growth was now at a sustainable rate and played down the inflationary implications of the recent depreciation (MMM, May 1995, §32). The Chancellor continued to reject the Governor’s advice to raise interest rates in the Monetary Meetings of June and July. The initial evidence of disagreement between the two caused a weakening of confidence in financial markets, but over the succeeding weeks new information suggested that growth was not excessive and inflation not rising, and the markets 13 At the time of the meeting it had not yet been decided to publish the minutes, and those of this meeting were published only in April. See Stephens (1996: 294–5) on this episode and the disagreements of 1995.

The 1993–1997 ‘New Framework’ for Monetary Policy 101 12 11 10 9 8 7 6 5 4 Sept 1992

Sept 1993

Sept 1994

Sept 1995

Sept 1996

Figure 6.3 UK policy rate, 1992–1997 (Source: Bank of England)

came to give the Chancellor the benefit of the doubt. The Bank also ceased to press for a rate increase. By the end of the year economic growth appeared to be slower and the prospects for inflation had improved. An agreed rate cut of 0.25% in December 1995 was followed by a second cut in January, about which the Bank was less enthusiastic. With growth apparently continuing to be below trend, there was another agreed cut in March 1996 and a further cut which the Governor opposed in June. By September, in the light of evidence of above-trend growth of demand, the Governor was calling for a 0.5% rise in interest rates but the Chancellor considered the case not yet proven. Although rates were raised by 0.25% at the end of October, the Governor continued to press in December 1996 and the first four months of 1997 for a further rise, while the Chancellor appealed to the substantial recent appreciation of sterling to justify his decision to leave interest rates unchanged (e.g. MMM, January 1997, §33; April 1997, §§38–41). A number of points emerge from this brief account and from more detailed perusal of the minutes concerned. First, the language of discussion in the monthly Monetary Meetings as revealed in the minutes was clearly economic, though this does not demonstrate that political factors were not important at an underlying level. Second, the focus of policy on inflation in two years’ time provided plenty of scope for argument between Chancellor and Governor about the likely outturn (on unchanged policies): among the range of indicators referred to there will in many situations be some pointing in each direction, and forecasts are far from perfect. Thus inflation targeting as it was operated in this subperiod was far from a mechanical or rule-based exercise and allowed wide scope for discretion. Third, there is no evidence in the minutes that monetary growth (M0 or M4) was an important influence on interest rate decisions. Fourth, the continuing importance of concerns about the exchange rate makes clear that the adoption of an inflation target had not solved the longstanding problem for the UK of how to handle its external relationships. In principle the exchange rate should have mattered only insofar as its level and/or movement affected the future course of domestic inflation, but this meant that

102 The Making of Monetary Policy in the UK, 1975–2000 it continued to be both important and a matter for debate and disagreement. Fifth, the way in which the new framework operated involved inappropriate incentives for the Chancellor, and possibly also for the Governor; such incentives were bound to compromise the ability of the decision-making process to achieve the stated inflation target. In principle, the new framework for monetary policy left the final decisions with an accountable Chancellor of the Exchequer but, for improved credibility, allowed publication of the views on policy of the Governor of the Bank, with both these individuals trying to hit the inflation target specified by the government. In practice, however, it can be argued that the framework encouraged both parties to pursue additional objectives. The Chancellor naturally retained the political objectives that have always been present in monetary policymaking in the UK, and in order to pursue them more effectively he was likely to want to recoup the power ceded to the Bank in the aftermath of Black Wednesday.14 This meant that he had an incentive not merely to establish his own reputation but also to diminish the reputation of the Bank and its influence on the markets. One way to accomplish the latter was to accustom the markets to minor disagreements between him and the Bank, in situations where his differing views could be shown to be not unreasonable ex ante; this would be particularly effective if his decisions turned out to be justified ex post. On the other hand the Bank was likely to want to maintain and perhaps expand the limited increase in autonomy which it had obtained, if only to improve the effectiveness of monetary policy in the longer term. The Bank could not compete directly with the Chancellor by following alternative policies which were then shown to be preferable, since decision-making remained firmly with the Chancellor. Its leverage depended on its power to move the markets by publicly disagreeing with the Chancellor, and in order to preserve that leverage it needed to have recourse to that power only sparingly. Thus it was more damaging for the Bank to be ‘wrong’ when the Chancellor was ‘right’ than to be wrong when the Chancellor was also wrong, and the Bank had an incentive to shade its views in the Chancellor’s direction in cases where the disagreement was small, in order to ensure that the impact of a disagreement remained a significant threat. There were a number of episodes of policymaking between 1994 and May 1997 which can be given plausible interpretations along these lines. First, the willingness of the Chancellor to raise MLR in late 1994 and early 1995—long before the likely election date and therefore well before an electorally motivated Chancellor would need to stimulate the economy—could be seen as a strategic attempt to establish a reputation as ‘tough’ on inflation in order to facilitate expansion later. Second, the Chancellor’s unwillingness to raise rates in the summer of 1995 might have started as a calculated risk. The argument was ‘finely balanced’, so that even if the markets thought the Bank was ‘right’ the impact of a different decision would not be large; and an early challenge to the Bank’s authority before it was well-established would be less risky than a later challenge. Moreover, even if the Chancellor had to raise rates the next month the experience of a public disagreement for a short period would help to reduce the impact of later disagreements (when it could be more important electorally to pursue a different course from that favoured by the Bank). In the event, the Chancellor was able to claim that he had been right and the Bank wrong: the Bank was outmanoeuvred and from that point the markets began to pay less attention to its views. Third, the four interest rate cuts between December 1995 and June 1996 might have been prompted partly by electoral considerations (with an 14 The Chancellor may also have been concerned about the creation of an alternative centre of authority to that of the Westminster government, which traditionally claims near-absolute power in the UK.

The 1993–1997 ‘New Framework’ for Monetary Policy 103 election due some 18 months after the first cut) as much as by evidence of slower growth. Fourth, the Bank’s willingness to come into line with the Chancellor after its strategic failure in the summer of 1995 and to support the December 1995 cut (though not all of the later cuts) might have reflected in part a desire by the Bank to restore the leverage of a threat of public disagreement with the Chancellor by temporarily suspending that disagreement, as well as a response to evidence of slower growth. Fifth, the Chancellor’s repeated rejection of the Bank’s advice to raise rates between December 1996 and April 1997 could be interpreted as intended to generate a substantial consumption-led expansion in time for the election, despite the probability that inflation would overshoot the government’s target, which it did shortly after the election. These interpretations cannot be proved to be correct, but—more importantly—they cannot (and could not at the time) easily be shown to be incorrect either, and the possibility that the monetary authorities were acting in such ways must therefore have fed into the expectations held by the financial markets and the private sector more generally. Under these conditions the credibility of monetary policy under these arrangements might have been higher than in the immediate aftermath of Black Wednesday, but it was bound to remain limited. At the same time there seems little reason to doubt that the Bank of England made efforts (mostly not observable) to protect and strengthen its autonomy and to further the case for central bank independence. In particular, it worked to improve its monetary expertise (notably its inflation forecasting capacity) and made use of the mechanisms of accountability such as the Inflation Report and the Governor’s contributions to the minutes of the monthly Monetary Meetings (as well as speeches by the Governor and other officials) to establish its technical reputation; and while talking about monetary policy mainly in terms of what central banks should do and how, and without referring explicitly to political decision-making, it also made small but persistent references to the concept of central bank independence. King’s (1997a) discussion of the new arrangements,15 for example, distinguishes analytically between four regimes: (a) the optimal state-contingent rule (to which no commitment is credible so that the rule is time-inconsistent); (b) a verifiable rule which is time-consistent but inevitably non-contingent; (c) pure discretion; and (d) a conservative central banker (as in Rogoff, 1985). The optimal rule (a) is not feasible (because it is time-inconsistent), and the second-best solution turns out to be the conservative central banker (d). King then proceeds to argue that the openness involved in the 1993–97 UK arrangements could improve the credibility and predictability of monetary policy, and in this environment an inflation target could be seen as giving incentives to central bankers to follow the first-best state-contingent rule. He concludes by noting that (as at the beginning of 1997) the arrangements seemed to have worked well, that all the major political parties in the UK were committed to retaining them, and that ‘[s]ome have gone further and proposed changes to give an even greater role to the Bank of England’ (1997a: 97). A lecture by the Governor in November 1996 on the evolution of the monetary framework ended by suggesting that, if the framework failed to deliver ‘a decade of growth through stability’, his audience would be able to judge whether that was ‘a result of our own technical incompetence or some failure of the political process . . . In the latter case you may just find that the Bank of England had been made independently accountable for decisions about monetary policy—but that is the subject for another lecture!’ (BEQB, 1997a: 103).16 15 The views expressed in the paper were attributed to the author personally, but King was an Executive Director of the Bank at the time. 16 On the other hand, there is no clear written statement by the Bank from the period after May 1997 which identifies political intervention in the setting of interest rates before that.

104 The Making of Monetary Policy in the UK, 1975–2000 A rather different perspective is given by the working paper by Briault, Haldane and King (1996) on ‘Independence and accountability’. This includes a wide-ranging analytical discussion, and an index of central bank accountability for fourteen countries, on which the UK turns out to score the highest. The index is then used to consider the relationship between goal independence and accountability, and between accountability and credibility as measured by bond yields for 1985–94, without any recognition that the UK’s high score is specific to the new arrangements (which were in full operation only from 1994) and would have been 0 or 1 rather than 31/2 if calculated for the years before 1993.17 It is difficult not to see this paper as ‘talking up’ those particular aspects of the new arrangements which gave the Bank more autonomy, without full regard to the demands of scholarly accuracy.18

6.3 THE CREDIBILITY OF POLICY UNDER THE NEW FRAMEWORK An interesting analysis of the credibility of monetary policy during this subperiod has been carried out by Cobham, Papadopoulos and Zis (CPZ) (2001), who make use of the fact that the Bank of England’s advice on interest rates was published and therefore known to the markets between January 1994 and May 1997. CPZ argue that the financial markets would naturally assume that the Bank of England had no objective other than the formal inflation target, whereas the Chancellor of the Exchequer might have conventional ‘political’ objectives. When the Chancellor disagreed with the Bank the financial markets would therefore interpret this as evidence that the Chancellor was pursuing such other objectives, so that disagreements between the two would lead to a decline in the credibility of monetary policy. They construct an index of what they call the ex ante credibility of policy, CREDt, on the interval [0,1] with an initial value of 0.5 and subsequent values CREDt = γ CREDt−1 + (1 − γ )Dt , 0 < γ < 1 where D takes the value of 0 for a disagreement and 1 for an agreement, and (1 − γ ) represents the speed with which credibility adjusts in the light of the latest information on agreements/disagreements. This is an ad hoc index, but CPZ argue that its gradual and asymptotic adjustment is plausible in the context of noisy and hard-to-interpret signals from the economy. They then examine whether this index is systematically related to an ex post measure of credibility in the form of the differential between the yield on 10-year UK government bonds and that on 10-year German, or US, bonds. They find a cointegrating vector relating the index inversely to the UK–German differential: as ex ante credibility deteriorates from accruing disagreements between Chancellor and Bank, the spread which has to be paid on UK bonds over German bonds rises. Moreover, the magnitude of this effect is striking: according to their simulation, if there had been no disagreements over the subperiod the yield differential would have been 1.5% lower in April 1997 (i.e. the UK bond yield would have been 6.11% instead of the observed 7.61%).19 17 The precise score depends on how the more limited Parliamentary monitoring in the earlier years is evaluated. 18 Interventions of the kind described in the last two paragraphs are consistent with the notion that there is a reverse

U-shaped relation between a central bank’s independence and its political activity: when the level of independence is very low or very high the bank has nothing to gain from political activity, but when it is at an intermediate level there is an incentive to engage in activity that may protect and increase it. 19 However, CPZ found no comparable relationship between the credibility index and the UK–US differential, a finding which they rationalised in terms of greater country-specific factors leading to more volatile yields in the US than in Germany.

The 1993–1997 ‘New Framework’ for Monetary Policy 105

6.4 THE POST-ERM REACTION FUNCTION The study by Adam, Cobham and Girardin (ACG) (2001), which has already been referred to in Chapters 4 and 5, reports the following preferred reaction function for the ‘post-ERM’ subperiod, October 1990 to April 1997: i t = 0.5‡ + 1.10‡πt + 0.39‡yt + 0.29‡igert + 0.27‡iust where πt is the inflation rate and yt is a measure of the output gap. The (short-run) parameter on the lagged dependent variable is 0.45‡ and the equation standard error is 0.16% (as against the standard deviation of the Treasury bill rate over this subperiod of 0.58%).20 Thus in this subperiod monetary policy was responding to domestic as well as international variables, which ACG interpret as implying improved credibility relative to the pre-ERM subperiod (despite the initial loss of credibility due to the exit from the ERM); however, the international variables still enter the function, partly because of official concerns about the likely pass-through from the exchange rate to prices but also because credibility was still limited. ACG’s estimated reaction functions for the pre-ERM and MPC subperiods and their conventional Taylor rule, as graphed in Figures 4.6 to 4.9, can be used to present a counterfactual analysis for the post-ERM subperiod. The pre-ERM rule predicts a rate consistently above the actual rate but converging on it, from around 7% above in early 1993 to around 2% above in early 1997; this can be understood as implying that policy on the pre-ERM basis with its lower credibility would have had to be much more cautious in promoting recovery (for fear of renewed depreciation). The MPC rule indicates a rate initially much lower than the actual but more or less converging on it by the end of the subperiod; this is consistent with the view that the MPC, whose rule according to ACG incorporates a much higher level of credibility, would have been able to act more strongly to promote recovery from the recession of the early 1990s. The Taylor rule rate is closer throughout to the actual rate than that suggested by either of the other rules. It is noticeable that neither the MPC nor the Taylor rule imply that rates should have risen markedly in mid-1995, when the Bank of England repeatedly clashed with the Chancellor over its call for a rise in rates; but that call was based on the likely effects of a depreciation of the exchange rate which is not taken into account by any of these rules (and which was later reversed rather than accentuated). On the other hand, the Taylor rule does imply a small rise in the rate in late 1996/early 1997, and the MPC rule a small rise in January and February 1997, at the time when the Bank of England was again calling unsuccessfully for policy to be tightened.

6.5 CONCLUSIONS The previous decade had seen experiments with monetary and exchange rate targeting, which had turned out to be unsuccessful for reasons connected both with external events and with the UK monetary authorities’ own behaviour. The crisis of September 1992 led to an explicit recognition for the first time of the time-inconsistency problem and of the authorities’ own loss of credibility. A new framework for monetary policy was put in place, which involved an inflation target and a new openness in monetary policy together with the granting of some 20 ‡ denotes significance at the 5% level. The coefficients on the German and US interest rates (iger and ins) are not significantly different from each other.

106 The Making of Monetary Policy in the UK, 1975–2000 limited and informal autonomy to the Bank of England, whose own reputation had been less seriously damaged by the events of Black Wednesday. Inflation remained within the initial wider target range of 1–4%, despite the depreciation of 1992, and hit the narrower target of 2.5% or less set for the spring of 1997 in April and May, but then rose again. The outcome of the new framework was also unimpressive in terms of expected inflation and credibility, and these findings are consistent with an analysis of the process by which the new arrangements operate. In particular, the focus on the inflation rate expected in two years’ time allowed a great deal of scope for discretion, and the monetary decision-making process did not preclude the Chancellor from seeking both to recoup the power lost from the Treasury to the Bank of England and to practice traditional forms of political manipulation of monetary policy. Moreover, while the results of ACG suggest that the credibility of policy had improved relative to the pre-ERM subperiod but was still imperfect, the analysis by CPZ is consistent with the view that the disagreements between Chancellor and Bank were damaging to credibility.21 Overall, the new arrangements introduced after the exit from the ERM represented an undoubted improvement, but one that was limited and insecure. Although they may well have been strong enough to preclude extreme monetary excesses such as the Heath–Barber boom in the early 1970s, whether they could have prevented another Lawson boom is less obvious. In addition, as will be discussed in more detail in Chapters 8 and 9, it is far from obvious that the new arrangements constituted a satisfactory solution to the vexed issue of the UK’s external relationship: while the exchange rate was not particularly volatile over this subperiod it began to appreciate towards the end of it and became seriously overvalued shortly afterwards. 21 See also the later statement on this subperiod by King (BEQB, 1999c: 297): ‘many, especially elsewhere in Europe, doubted the United Kingdom’s commitment to monetary stability, in the absence of a willingness to remove operational decisions on interest rates from the political arena. Long-term interest rates contained a risk premium to reflect the possibility that the timing and magnitude of interest rate changes might reflect political considerations.’

7 Monetary Policy under the MPC, 1997–2000 The Labour party’s election victory in May 1997 was followed within days by the announcement that the Bank of England was to be given control over interest rates and responsibility for achieving the government’s inflation target, an announcement that caught most observers by surprise. The new Monetary Policy Committee began to operate in June 1997, the details of the new arrangements were fleshed out within the first few months, and the Bank of England Act came into force in June 1998. This chapter starts with an outline of how this new regime is meant to operate. Section 7.2 examines the changes in formal central bank independence, including those associated with the Maastricht Treaty agreed in December 1991 and the 1992–97 new framework. Section 7.3 offers a preliminary assessment of the new regime. Section 7.4 considers an estimate of the monetary policy reaction function in this subperiod and contrasts it with the policy rules for earlier subperiods, and Section 7.5 concludes.

7.1 INFLATION TARGETING WITH INSTRUMENT INDEPENDENCE Under the new framework for monetary policy the Chancellor of the Exchequer continues to set the inflation target, while the Bank of England has operational responsibility for setting short-term interest rates in such a way as to achieve the inflation target.1 Formally, the Bank’s objective is ‘to maintain price stability and, subject to that objective, to support the government’s economic policy, including its objectives for growth and employment’ (BEQB, 1998b: 93). The government retains the responsibility for determining the exchange rate regime, but the Bank has its own pool of foreign exchange which it can use at its own discretion in support of its monetary policy; it is also expected to intervene on the government’s behalf, using the government’s own foreign exchange reserves, if the government so instructs it. At the same time the Bank’s role as the government’s agent for debt management was transferred to the new Debt Management Office, while the Bank’s responsibility for banking supervision was transferred to the new Financial Services Authority (FSA). A Memorandum of Understanding published in October 1997 set out the specific responsibilities of each of the Bank, the Treasury and the FSA in the area of financial stability.2 Interest rate decisions are taken by the new Monetary Policy Committee (MPC) of the Bank, which consists of the Governor and two Deputy Governors (one responsible for monetary 1 See 2 See

BEQB, 1997c: 241–7, 1998b: 93–9, 2001d: 434–41; Budd (1998). Goodhart and Schoenmaker (1995) and Goodhart (2001a) for analysis of the issue of whether banking supervision should be ‘combined’ with monetary policy within the central bank or undertaken by a ‘separate’ agency.

108 The Making of Monetary Policy in the UK, 1975–2000 policy, the other for financial stability), these three being appointed by the government; two other members appointed by the Bank after consultation with the Chancellor (one responsible for monetary analysis within the Bank, the other for monetary operations); and four external members, who must have relevant knowledge and experience, appointed by the Chancellor. The last six are appointed for renewable terms of three years, while the Governor and his deputies are appointed for renewable terms of five years. The Bank of England’s Court (governing body) was also reformed. It now consists of the Governor and two Deputy Governors, plus 16 Non-Executive Directors, and the latter are required to review the performance of the Bank and in particular of the MPC, and to ensure that the MPC collects adequate regional and sectoral information for monetary decision-making. Since June 1997 the inflation target has been set at 2.5% (rather than the previous 2.5% or less) on the RPIX. In the event of inflation deviating by more than 1% above or below the target, the Governor, as chairman of the MPC, is required to write an open letter to the Chancellor explaining why this has happened, how long the deviation is expected to last and what action has been taken to eliminate it. However, the inflation target is clearly 2.5%, not the range 1.5–3.5%. The MPC is accountable through a variety of channels: it is accountable for its procedures to the Court; it is required to produce a quarterly Inflation Report with an inflation forecast; the minutes of its meetings have to be published no more than six weeks after the meeting concerned (the time lag was originally around five weeks but was shortened to two weeks as from October 1998). In addition the Court is required to produce an annual report for the Chancellor to lay before Parliament which then holds a debate on it, and the Chancellor is accountable to Parliament for the inflation target (which must be confirmed or reset at least once a year, this being usually done in the Budget). The House of Commons Treasury Select Committee holds regular (non-statutory) meetings to interview new members of the MPC and to investigate the MPC’s procedures and decisions.3

7.2 CENTRAL BANK INDEPENDENCE Tables 7.1 to 7.3 provide an analysis of how much the degree of central bank independence (CBI) has changed, first from the 1993 changes and then in 1997. Table 7.1 shows Cukierman’s (1992) index of CBI as recalculated for the earlier period and then calculated for the later periods by Cobham, Cosci, Mattesini and Serre (CCMS) (2000). Between 1971–92 and 1993– 97 Cukierman’s overall unweighted index LVAU rises from 0.31 to 0.57 (the initial level was also 0.31 in Cukierman’s calculation for the 1980s, as compared with 0.68 for Switzerland, 0.66 for West Germany and 0.51 for the US), but the main differences between the periods are from the greater restrictions on Bank of England lending to the government (rows (9), (10) and (16)) due to the provisions of the UK Protocol to the Maastricht Treaty, which contribute 0.24 of the improvement, rather than from the increase in the Bank of England’s influence on policy (rows (5) and (6)) which contribute 0.02 of the increase. From 1997, however, there is a further increase in LVAU from 0.57 to 0.70, and this time the differences are entirely in the Bank’s influence on policy (rows (5) and (6)) and in the Bank’s objectives (row (8)).4 3 The House of Lords also has a committee examining the role of the MPC, now organised as the Economic Affairs Select Committee. 4 CCMS award 0.83 for the latest period in row (5) in recognition of the Bank’s acquisition of operational but not goal independence: this is equidistant between 0.67 (‘CB participates in formulation of monetary policy together with government’ in Cukierman) and 1.0 (‘CB alone has authority to formulate monetary policy’).

Monetary Policy under the MPC, 1997–2000 109 Table 7.1 Cukierman’s (1992) index of CBI for the UK Criterion 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Chief executive officer Term of office Who appoints Governor? Provisions for dismissal Is the Governor allowed to hold another office? Policy formulation Who formulates monetary policy? Government directives and resolution of conflicts Is CB given active role in formulation of Government budget? CB objectives Price stability? Limitations on lending Limitations on advances Limitations on securitised lending Who decides control of terms of lending ? How wide is the circle of potential borrowers from CB? Type of limit when such limit exists Maturity of loans Restrictions on interest rates Prohibition on lending in primary market Indices of legal independence LVAU (unweighted) LVAW (weighted)

1971–92

1993–97

1997–

0.5 0 0.83 1

0.5 0 0.83 1

0.5 0 0.83 1

0 0 0

0.33 0.2 0

0.83 0.4 0

0

0

0.8

0 0 0.33 1 1 1 0.25 0

0.67 1 0.33 1 1 1 0.25 1

0.67 1 0.33 1 1 1 0.25 1

0.31 0.26

0.57 0.51

0.70 0.66

Source: CCMS (2000).

Table 7.2 Grilli et al.’s (1991) index of political independence for the UK Appointments (1) 1971–93 1993–97 1997–

(2)

(3)

(4)

Relations with government

Constitution

(5)

(7)

* * 1/ 2

(6)

1/ 2

(8)

1/ 2

Key: (1) Governor of central bank not appointed by government; (2) Governor appointed for more than 5 years; (3) none of Board of central bank appointed by government; (4) Board appointed for more than 5 years; (5) no mandatory government representative in Board; (6) no government approval of monetary policy needed; (7) statutory requirement that central bank pursues monetary stability; (8) legal provisions protect central bank against government; (9) overall index = sum of columns (1) to (8). Source: CCMS (2000).

Index (9) 1 1 11/2

110 The Making of Monetary Policy in the UK, 1975–2000 Table 7.3 Grilli et al.’s (1991) index of economic independence for the UK Monetary financing of deficits

1971–93 1993–97 1997–

(1)

(2)

(3)

(4)

(5)

* * *

* * *

* * *

* * *

* *

Monetary instruments (6)

*

Index

(7)

(8)

**

4 5 8

Key: (1) direct credit facility: not automatic; (2) direct credit facility: at market interest rate; (3) direct credit facility: temporary; (4) direct credit facility: limited amount; (5) central bank does not participate in primary market for public debt; (6) discount rate set by central bank; (7) banking supervision not (**) or only partly (*) entrusted to central bank; (8) overall index = sum of columns (1) to (7). Source: CCMS (2000).

Table 7.2 shows the Grilli, Masciandaro and Tabellini (GMT) (1991) index of political independence as given by CCMS. This is 1 for the earlier period, 1 again for 1993–97, and 1 12 for 1997 and after.5 Their index of economic independence in Table 7.3 rises from 4 in the earlier period6 to 5 in 1993–97, due to the ending of Bank of England participation in the primary market for public debt, and to 8, the maximum possible, from 1997; two-thirds of the latter change is due to the reallocation of banking supervision from the Bank of England to the FSA, and one-third to the Bank’s acquisition of control of the discount rate. These analyses imply that there were substantial increases in CBI in the UK, most of which resulted from the (UK Protocol to the) Maastricht Treaty and later from the 1997 changes. After those changes the GMT economic index is at its maximum, but the political index is still relatively low, mainly because the government continues to make the key appointments at the Bank of England, and they have terms of only five years. In addition, the GMT indices are constructed in a way that cannot deal easily with the condition of instrument independence without goal independence. The same factors are also responsible for most of the shortfall from the maximum level in the Cukierman index. While under the UK’s (formal and informal) constitutional arrangements it is difficult to see how anybody other than the government could appoint the Governor and Deputy Governors of the Bank, the government could easily enough choose to take advice from independent experts over these and the external members of the MPC, and this would be consistent with the concept of the Bank as the technical implementer of the government-set inflation target.7 At the same time the government could easily make the terms longer. However, Debelle and Fischer (1994, cited in Fischer, 1994) have provided evidence that the aspects of CBI which are important determinants of inflation are the nature of the central bank’s statutory objectives and the nature of its lending to the government, together with its right to set the discount rate, 5 CCMS award 1 for the latest period in column (5) in recognition of the attendance at MPC meetings of a non2 voting Treasury representative, and 12 in each of columns (6) and (7) in recognition of the Bank’s operational but not goal independence. 6 Grilli et al. gave the UK 5, but CCMS correct Grilli et al.’s mistaken view that the discount rate was set by the central bank. 7 A specific proposal for the involvement of recognised outside experts in the appointment of external members of the MPC was made in David Cobham, ‘Restoring trust at the Treasury’, Financial Times, 10 August 2000. See also House of Lords (2001, chapter 5).

Monetary Policy under the MPC, 1997–2000 111 rather than the legal provisions relating to appointments and the central bank’s relationship with the government. In addition, Masciandaro and Spinelli (1994) found that (developed country) central bankers themselves attached most importance to the statutory requirement to pursue monetary stability, the absence of a requirement that the government should approve the central bank’s monetary policy, and the ability of the central bank to set the discount rate (columns (6) and (7) in Table 7.2 and column (6) in Table 7.3). On this basis the degree of independence of the Bank of England since 1997 must be considered substantial, and not a cause for concern.8 The reasons for changes in formal or statutory CBI have also been investigated by CCMS, in a comparative study of France, Italy and the UK which relates such changes to (perceptions of) the benefits (mainly lower inflation) and costs (e.g. greater difficulties in coordinating monetary and fiscal policy, political costs of making the change) over time together with the discount rate used to evaluate them. For the UK they note that Chancellors Lawson and Lamont had both raised the issue but received no sympathy for it from their respective prime ministers, while the Roll Committee (1993) had produced a report in favour of CBI which was cogent, non-technical and accessible. They also note that the main argument used against CBI was the problem of ensuring accountability, yet there were significant improvements in accountability in 1993–4 without, and then in 1997 with, independence. And with respect to the introduction of independence in 1997 they appeal to Balls (1998) to support their claim that ‘Labour had indeed come to terms with the intellectual case for CBI . . . In addition, the Labour government appears to have been strongly aware of the possible credibility problem it faced . . . ’ (CCMS, 2000: 34).

7.3 THE WORKINGS OF THE MPC REGIME The experience of the new framework up to the end of 2000 is short, and the subperiod does not contain any major recession or crisis. It is not therefore possible to make a full evaluation of its performance, but a preliminary assessment can be made. This section looks in turn at the actual macro outcomes—inflation, inflation expectations, income and unemployment, and the sterling exchange rate—and then goes on to discuss the workings of the new framework with particular reference to the Kohn (2001) report and papers by current or former members of the MPC.9 7.3.1 Macroeconomic Outcomes RPIX inflation hit the 2.5% or less target set for the latter part of the 1992–97 Parliament only in April and May 1997, and then immediately rose, reaching 3.0% in July 1997 and then a higher peak of 3.2% in May 1998 (see Figure 7.1); for the second half of 1998 it was close to the post-May 1997 target of 2.5%, and in 1999 and 2000 it was mostly closer to 2%.10 Headline inflation (the RPI) fluctuated more widely as expected (since it reflects movements in interest rates as well as goods prices), reaching a peak of 4.2% in May 1998 and a trough of 1.1% in August–September 1999, but hovering around 3% for the latter part of 2000. Market expectations of inflation, as tracked by the Bank of England’s calculations from conventional and index-linked yields, were above the 2.5% target in 1997 but had come down 8 See also Bean (1998: 1799) for the argument that in the context of goal dependence (but instrument independence) the length of the term is less important. In addition, Bean argues convincingly that, when the loss function for the monetary authorities is properly specified, ‘it is the act of delegation itself [i.e. the granting of central bank independence] that solves the time inconsistency problem’ (1998: 1799, emphasis in the original). 9 A detailed narrative of policy decisions is given in the final part of Section 10.5 of Chapter 10. 10 By late 2001, Wadhwani (2001a) was worrying about a possible tendency to undershoot the inflation target, but there seems to be less cause for concern over the subperiod considered here (which ends in December 2000).

112 The Making of Monetary Policy in the UK, 1975–2000 4.5

2.5

RPIX

0.5 June 97

Dec 97

RPIX target

June 98

RPI

Dec 98

June 99

Dec 99

June 00

Dec 00

Figure 7.1 Inflation, 1997–2000 (Source: Economic Trends)

below it by the end of 1998; they rose again in late 1999 but returned to the target level by the end of 2000.11 Similarly the Consensus Economics index of independent forecasts of inflation, which had remained stubbornly above 2.5% in 1997 and most of 1998, was below that level in 1999 and 2000 (Wadhwani, 2001a: 352). At the same time, as shown in Figure 7.2, UK 4

2

0

−2 1994M1

1995M1

1996M1

1997M1 UK-Germany

Figure 7.2 Bond yield spreads, 1994–2000 (Source: International Financial Statistics) 11 Data

supplied by the Bank of England.

1998M1 France-Germany

1999M1

2000M1

Monetary Policy under the MPC, 1997–2000 113 government bond yields went below French yields in the third quarter of 1999, and below German yields from late 1999 (they had been below US yields since late 1998). Thus, while inflation performance can be considered satisfactory from about 15 months after operational independence, there were clear improvements in the credibility of monetary policy within roughly two and a half years. It also seems that the MPC’s policy announcements, which had been associated with rather larger changes in implied interest rates on the same day in the period from June 1997 to May 1999 than those of the US and eurozone central banks, were no more surprising to the markets than the announcements of the latter in the following two years (Wadhwani, 2001a: 355). Moreover, these improvements did not appear to have been obtained at the expense of poor economic growth or high unemployment: GDP growth averaged 2.6% over 1998–2000, while unemployment continued to fall, from 7% in 1997 to 5.5% in 2000 (Table 2.1).12 On the other hand, the sterling exchange rate had appreciated strongly in late 1996 and the first half of 1997 but not readjusted, so that from late 1998 sterling was overvalued to an extent that caused substantial imbalances between different sectors of the economy.13 7.3.2 The Operating Procedures of the MPC The way the MPC operates has been described in some detail elsewhere, notably in Goodhart (2000) and Bean and Jenkinson (2001), while the Kohn report (Kohn, 2001) provides a comprehensive assessment of various aspects of these procedures and of the work done by Bank of England staff in support of them. Kohn was generally praising of the flow of information to the MPC, including the research and analysis made available in the pre-MPC meetings14 and in other ways, and the contributions of the Bank’s regional Agents. He also gave a positive overall evaluation of the work of the Monetary Analysis section in the Bank, though he drew attention to several aspects of the recruitment and retention of staff and to the relationships between staff and MPC members. An issue which Kohn does not address is the tensions that developed in late 1999 over the access of external members of the MPC to research resources, tensions that were eventually defused by the establishment of an MPC Unit consisting of a small number of staff working under each of the external members on the matters which those members thought important.15 7.3.3 The Decision-making of the MPC It is clear, not least from Minutes of MPC meetings, that the way in which decisions are arrived at is different from that in most other central banks. The main reasons for this are the principle enshrined in the legislation that the individual members of the MPC are individually accountable for their decisions, and the fact that many of the members of the MPC are academics or former academics who value their own reputations as independent experts: they wish not merely to get the decision right but to be seen to get it right. While the Minutes do not identify the proponents of every argument, they record the way each member votes and in practice it is 12 The numbers on their own suggest that economic performance was less good than in the previous phase, but since the economy was then starting from a state of severe recession, performance in this phase should be regarded as superior. 13 See, for example, Allsopp (2001). Exchange rate issues are discussed in more detail in Chapters 8 and 9. 14 These are half-day (previously full-day) meetings at which the Bank staff present and analyse the latest data for the MPC, usually on the Friday before the MPC meetings. 15 See the Bank’s response to the Kohn report, BEQB, 2001a: 50–52.

114 The Making of Monetary Policy in the UK, 1975–2000 usually possible to relate particular arguments to particular voting positions. In a speech made while he was Chief Economist at the Bank, Vickers argued that the MPC had no desire to spring inflation surprises: ‘Quite apart from the obligation to fulfil our statutory duty, we have the strongest professional and reputational incentives, which in my opinion are incapable of being enhanced by financial incentives, to get as close as we can to our inflation target’ (BEQB, 1998d: 369–70).16 King, as Deputy Governor, argued that ‘the voting records of individual members are in the public domain. That disclosure improves the quality of decisions, as well as the accountability of Committee members, because there is no better incentive to cast one’s vote for the policy most likely to hit the inflation target than the prospect of having to defend that voting record in public’ (BEQB, 1999c: 300). Actual decisions are taken in the form of a vote for or against a particular proposition (rather than between two alternatives). The proposition is identified by the Governor, as chairman of the committee, but the vote follows a discussion in which each member has made clear his or her precise preference, so that the ability to select the proposition for voting does not confer power over the outcome. And while the MPC has had plenty of non-unanimous decisions, including two (February and March 1998) in which the eight members present split equally and the decision was taken by the casting vote of the Governor as chairman, it has never been near to an equal three-way split in which members might have had an incentive to vote for their second rather than their first preference.17

7.3.4 The Specification of the MPC’s Objectives A potential weakness in the post-1997 framework lies in the specification of the MPC’s objectives as ‘to maintain price stability and subject to that objective, to support the government’s economic policy, including its objectives for growth and employment’ (BEQB, 1998b: 93). The inflation target of 2.5% is symmetric, so that deviations in either direction are equally undesirable and the committee has just as much of a duty to expand demand if inflation is in danger of falling below 2.5% as to restrain it if inflation threatens to exceed 2.5%.18 But the remit does not lay down the time horizon over which inflation must be returned to the target level after a deviation occurs, e.g. if some external shock or policy mistake were to push inflation more than 1% away from the 2.5% target and so trigger an open letter from the Governor to the Chancellor which would explain, among other things, how quickly the MPC proposed to return inflation to 2.5%. However, it is generally assumed that there is a trade-off between the variability of inflation and the variability of output, so that a quicker return of inflation to target (to ensure lower inflation variability) would involve a higher variability of output. Nevertheless, the evidence seems to suggest that this is not a serious weakness. Both Bean (1998) and Batini and Haldane (1999), the former using a simple backward-looking model estimated for the UK and the latter a forward-looking calibrated model, find that the trade-off is nearly L-shaped. This means that a wide range of preferences as between inflation and output 16 Goodhart (2000: 230) is more sympathetic to the idea of pecuniary incentives, but notes that ‘most central bankers and members of monetary policy committees’ take a different view. 17 Table 10.8 gives the number of votes for and against the proposition that was agreed, together with the direction(s) of change preferred by the minority, for each meeting from June 1997 to December 2000. 18 The symmetric nature of the target has served as a useful defence for the MPC against the suspicion that it is biased towards undershooting the inflation target (Wadhwani, 2001a: 354).

Monetary Policy under the MPC, 1997–2000 115 variability would lead the monetary authorities to pick the same point on the trade-off, in which case the fact that the remit given to the MPC is an ‘incomplete contract’ is not important.19 It should also be noted that, although in principle the government could decide to set a higher (looser) inflation target, such a move would be so visible and would affect financial markets so strongly that the government’s power in this respect is in practice heavily restricted. 7.3.5 The Individual versus the Collective and the MPC’s Forecast One of the innovations of the new framework of 1993–97 was that the Bank of England was required to publish its own forecast of inflation, within the quarterly Inflation Report. At the beginning the Bank published its central projection together with a shaded area in the relevant chart showing the central projection plus or minus the average error on such forecasts in the past. But from February 1996 it began to publish its inflation forecasts in the form of the ‘fan chart’ which uses different-coloured bands to convey information about the variance and the skew of the forecast (see Chapter 6, page 98).20 When the Bank became operationally independent in May 1997 the forecast became the property, or output, of the MPC, but the basic procedure was much the same as it had been before. The immediate difference was that instead of being produced by a hierarchy the forecast—referred to as the MPC’s ‘best collective judgement’ or ‘projection’—was now produced, or at least had to be agreed, by a committee. A relatively early discussion of the procedure by one of the external members (Goodhart, 1999b) noted the heavy demand made on members’ time by the forecasting procedures, but identified five benefits: the MPC should be better informed as a result (if only about the limits to their knowledge); the quality of the forecasts should be better as a result of the pooling of expertise; the MPC should be more disciplined, that is more likely to make the necessary changes in interest rates because it would feel it had to produce a forecast showing future inflation on target; the accountability of policy should be greater since the forecasts can later be compared with the outturns; and transparency should also be better as the publication of the forecast aids both explanation by the MPC and assessment by outside observers. Goodhart himself ranked these benefits, from most to least important, in the following order: transparency—discipline—informedness—forecast quality—accountability. However, the task of producing joint forecasts from a committee composed of individually accountable experts—despite the absence of ‘ideological’ disagreement about the underlying macro model—became more difficult. Kohn (2001: 40) has suggested that members of the MPC may on occasion argue for particular assumptions to be built into the model, ‘not out of conviction on those assumptions, but rather to shape the overall outcome in a direction they are most comfortable with’.21 In August 1999 the Inflation Report included for the first time a Table 6.B which showed the possible effects on inflation and GDP growth of different assumptions about the exchange rate, earnings growth, margins and oil prices; and a table of this sort became standard. This development makes it more difficult for the outsider to understand the precise balance of opinion on the committee, which implies that the publication of a forecast provides a smaller degree of benefits, notably in terms of transparency, and this has led to some discussion of possible alternatives. 19 But see Henry, Satchi and Vines (2001) for an analytical argument that when the central bank discounts the future its preferences do affect the choice of point on the trade-off. 20 Similar fan charts have been produced for the GDP forecasts since November 1997. See the article in BEQB, 1998a: 30–7, for a thorough discussion of the meaning and construction of the fans. 21 See also Goodhart (2001b: 63).

116 The Making of Monetary Policy in the UK, 1975–2000 The most comprehensive discussion can be found in the Kohn report which considers five alternatives: (a) the forecast to be replaced by a more comprehensive discussion of trends and concerns; (b) the forecast to be produced by the Bank staff (as is done in the Federal Reserve) and published as a benchmark around which MPC members’ views would be arrayed; (c) the forecast to be produced by/for the Governor or the Bank, and submitted to but not necessarily approved by the MPC—more or less the alternative proposed by Goodhart (2001b);22 (d) the published forecast to be the median of the MPC members’ forecasts, perhaps with some indication of the dispersion around that median;23 and (e) the forecast to be identified as the view of the majority of the MPC, with the views of dissenters presented only briefly in order to encourage them to remain in and try to influence the majority. Kohn makes no clear judgement between these five but appears to favour the last. It seems probable that some evolution in the MPC’s practice will occur, but what that evolution will be is difficult to forecast. 7.3.6 The Constant Interest Rate Assumption One feature of the MPC’s forecasting procedures which has attracted particular attention is the use of the assumption that interest rates remain unchanged. This assumption, like the forecast itself, dates back to the 1993–97 period, when it could be justified on the grounds that the Bank of England should not be expected to ‘second-guess’ the Chancellor’s decisions on interest rates (Goodhart, 2001c: 171). However, the MPC has no such constraint, and the assumption is clearly problematic: it will not typically be optimal to keep interest rates constant for the next eight (or more) quarters, such an assumption will almost always be at odds with market expectations and therefore not be credible, the procedure may adversely affect the credibility of policy more widely, and it makes forecasting more complicated (since the market typically has different expectations).24 However, it is far from obvious what satisfactory alternative could replace the constant interest rate assumption, which needs to be understood as a conditioning assumption that the MPC knows does not correspond to market expectations rather than as a supposedly ‘realistic’ assumption. It would, for example, be extraordinarily complicated for the committee to agree a time path for interest rates up to some horizon, to agree on applying optimal control procedures to the task, or even to agree to apply a simple Taylor-type rule.25,26 In addition, as Goodhart (2001c) also argues, the obligation to publish an inflation forecast based on a constant interest rate assumption may usefully offset a tendency towards late and slow adjustment of the interest rate. In the light of these points it is notable that the Kohn (2001) report recommends improving the existing forecasts rather than changing the constant interest rate assumption. Kohn also emphasises the danger that an alternative interest rate assumption might be misunderstood 22 Goodhart favours the forecast being the Governor’s personal responsibility, and argues that the MPC could then ‘approve’ it as required in the Bank of England Act (1998) without having to agree with every detail. The provisions of the Act would appear to rule out Kohn’s alternatives (a) and possibly (b). 23 Svensson (2001) proposes the choice of the median forecast in a slightly different context (see below). 24 See also Meyer’s (2001) comments on Goodhart (2001c). 25 See Goodhart (2001c). However, he suggests that the MPC should publish, in addition to the constant interest rate forecast, two simple variants involving lower (higher) then higher (lower) interest rates. 26 Svensson (2001) argues that a median voting procedure could be applied to members’ forecast paths for interest rates in order to produce a path from which the inflation and output forecasts can be derived. He also argues for median voting procedures to be applied to identify, and then publish, the MPC’s loss function, but the fact that Goodhart (2001c: 179) rejects his preferred quadratic functional form for the loss function should surely indicate that exercises of this kind are likely to be much more difficult than Svensson imagines.

Monetary Policy under the MPC, 1997–2000 117 in the markets, while Wadhwani (2001a: 355) argues that the inability of a committee with individual accountability to give advance ‘guidance’ to the markets in the way that the Federal Reserve Board tries to do has to be accepted as a cost of the system (but may be outweighed by its ability to produce better policy decisions on average). 7.3.7 A Preliminary Assessment The macroeconomic outcomes under the MPC regime—in terms of inflation and overall GDP— have been clearly satisfactory, and the improvement in the regime’s credibility has yielded benefits in terms of lower long-term interest rates. There are some unresolved problems in the operating procedures but these are likely to be containable at least so long as inflation remains near target and serious recession is avoided.

7.4 THE MPC’S REACTION FUNCTION Adam, Cobham and Girardin (ACG) (2001) report the following preferred reaction function for the subperiod June 1997 to August 2000: i t = 0.2 + 2.04‡πt + 1.15‡yt where πt is the inflation rate and yt is a measure of the output gap. The (short-run) parameter on the lagged dependent variable is 0.76‡ and the equation standard error is 0.20% (as against the standard deviation of the Treasury bill rate over the subperiod of 0.82%).27 Thus in this subperiod they find that monetary policy was responding entirely to the domestic variables, inflation and the output gap, and the external variables enter only as instruments for the domestic variables. ACG interpret this as implying that the acquisition of instrument independence had improved credibility to the extent that the monetary authorities were able to concentrate on the domestic variables (including the effects of the external environment on the latter) rather than looking over their shoulders at the actions of foreign central banks as they had felt obliged to do in the previous subperiods. ACG’s estimated reaction functions for the pre-ERM and post-ERM subperiods and their conventional Taylor rule, as graphed in Figures 4.6 to 4.9, provide some interesting comparisons for the MPC subperiod. The pre-ERM rule predicts a rate consistently and increasingly higher than the actual; this divergence probably reflects the fact that the UK was now no longer a high-inflation country, so that interest rates well above German and US levels were no longer appropriate. The post-ERM and Taylor rules, on the other hand, imply rates that are somewhat lower through 1997 and 1998, by as much as 1% (relative to the actual level of around 7%), but slightly higher in 1999.28 These contrasts also suggest that the MPC reaction function was particularly activist, since 1997–98 was a phase in which the MPC raised the interest rate to pre-empt a resurgence of inflation while in 1999 they lowered it to head off a recession.29 27 ‡denotes significance at the 5% level. 28 The implication that the interest rate

would not have been raised so much in 1997–98 on the post-ERM rule suggests that the claim in Chapter 6 that Chancellor Clarke was taking risks on inflation was not entirely misplaced. 29 ACG also note that RPIX inflation was lower and a bit less variable in the MPC than in the post-ERM subperiod, but the Treasury bill rate was more variable, and they interpret this as evidence that policy was more ‘activist’ under the MPC.

118 The Making of Monetary Policy in the UK, 1975–2000

7.5 CONCLUSIONS Up to the end of the period covered in this book there seems little doubt that the MPC regime has provided the most successful and satisfactory framework for monetary policy over the last quarter of a century. The Bank of England now has substantial independence, macro outcomes have been highly satisfactory compared to previous regimes, and the operating and decisionmaking procedures of the MPC are generally satisfactory. In particular, the combination of the individual accountability of the members of the MPC with the publication of detailed minutes has produced a remarkable level of transparency. There are a number of lower-level areas where tensions and problems remain, and some evolution in the MPC’s practices and procedures can be expected if the UK does not enter EMU (if or when it does there will be more fundamental changes). However, there is one aspect of the experience of this regime which has not been satisfactory: the persistent overvaluation of sterling. This issue is discussed, in a longer historical context, in the next two chapters.

8 Interest Rate Decisions and the Exchange Rate This and the next chapter examine in more detail the decisions to change interest rates over the whole of the period covered in the book, with particular reference to the importance of the external dimension in policy decisions. Section 8.1 of this chapter reports the results of a detailed classification of the various concerns underlying each interest rate change as given in the official sources. Section 8.2 provides a criterion for identifying ‘large’ misalignments of the exchange rate, and takes a preliminary look at those that have occurred over the period. Section 8.3 presents a criterion for identifying ‘large’ exchange rate changes, and takes a preliminary look at the large changes experienced. The main component of Chapter 9 is an examination of how interest rates were changed through each phase of monetary policy, with particular reference to the monetary authorities’ attitudes to the exchange rate; that examination then provides the basis for a more comprehensive analysis of misalignments, large exchange rate changes and sterling crises.

8.1 THE OFFICIAL CONCERNS UNDERLYING INTEREST RATE CHANGES Table 8A.1 in the Appendix to this chapter reports the concerns expressed by the monetary authorities themselves when they reported their decisions to change interest rates in the Bank of England Quarterly Bulletin or, from 1994 to 1997, in the minutes of the monthly Monetary Meetings, or, since 1997, in the minutes of the Monetary Policy Committee meetings and the Inflation Report.1 The movement of the policy rate over time is shown in Figure 8.1. The table notes whether the authorities referred specifically to recent or current interest rates in Germany and the US (and if so, whether they had changed in one direction or another), and gives a rough quantification of the main concerns mentioned for each UK change—the exchange rate vis-`a-vis the Deutschemark and vis-`a-vis the dollar, domestic growth or inflation, domestic monetary growth, and domestic fiscal policy. A tick indicates that the relevant concern ‘favours’ the interest rate change actually made, while a cross indicates that it opposes the change; the scale of concerns is indicated by the number of ticks or crosses (between one and three). Table 8A.1 was then used to produce the summaries in Tables 8.1 and 8.2. The former groups interest rate changes in ‘cells’ where there were several changes that were closely related, notes key features of the context in which each change or group of changes was made, and identifies the main reason for the change. The latter shows the number of cells in each 1 It will be obvious from the table that the information provided by the authorities in this way was initially quite limited, particularly in the first year or so covered; but it became much more substantial in later years.

9.75 10 11 12 9 10.5 11.5 13 15 12 8 5 7 6.5 10 12.5 14 12 14 17 16 14 12 16 15 13 9 10.125 11 10 9.5 9 8.625 9.25 12 9.625 14

−13/4

20.1.75–21.4.75 5.5.75 28.7.75 6.10.75 17.11.75–8.3.76 26.4.76 24.5.76 13.9.76 7.10.76 22.11.76–3.2.77 10.3.77–16.5.77 8.8.77–17.10.77 28.11.77 9.1.78 12.4.78–8.6.78 9.11.78 7.2.79 1.3.79–5.4.79 13.6.79 15.11.79 3.7.80 25.11.80 11.3.81 16.9.81–1.10.81 14.10.81–10.11.81 4.12.81–12.3.82 8.6.82–5.11.82 29.11.82 12.1.83 16.3.83–15.4.83 15.6.83 4.10.83 7.3.84–15.3.84 10.5.84–27.6.84 9.7.84–12.7.84 9.8.84–23.11.84 11.1.85–28.1.85 +1/4 +1 +1 −3 +11/2 +1 +11/2 +2 −3 −4 −3 +2 −1/2 +31/2 +21/2 +11/2 −2 +2 +3 −1 −2 −2 +4 −1 −2 −4 +11/8 +7/8 −1 −1/2 −1/2 −3/8 +5/8 +23/4 −23/8 +43/8

New level

Size of change

Dates of changes

Table 8.1 UK interest rate changes, 1975–2000

falling world interest rates, £ depreciation large Treasury bill issues tougher anti-inflationary policies, rising US rates rising US rates declining US rates sterling depreciation sterling depreciation further sterling depreciation further sterling depreciation sterling strength sterling strength sterling strength sterling uncapped end-October, appreciation market pressure for reduction monetary overshoot, dollar recovery incomes policy problems, rising US rates incomes policy problems, rising US rates strengthening of confidence new government sterling weakness sterling strength, gathering recession sterling strength, gathering recession sterling strength, recession, tight Budget sharp sterling depreciation sterling stabilised sterling stable, slower monetary growth declining US rates, declining inflation sharp depreciation of sterling sharp depreciation of sterling sterling appreciation, declining inflation post-election sterling strength £M3 growth within target sterling stable, £M3 growth rising US rates, weaker confidence rising US rates, sterling weak sterling stabilised dollar strong, sterling very weak

Context

to follow US rates? (technical) to maintain differential with US rates? to keep in line with US rates? to keep in line with US rates? to stabilise sterling to stabilise sterling to stabilise sterling, control monetary growth and inflation to stabilise sterling, control monetary growth and inflation to unwind previous rises to unwind previous rises, prevent appreciation to prevent appreciation to control monetary growth (tactical) to control monetary growth and stabilise sterling to restore confidence and monetary control to restore confidence and monetary control to unwind previous rises to control monetary growth and inflation to control monetary growth and inflation to relax financial squeeze on companies to relax financial squeeze on companies to relax financial squeeze on companies to avoid upwards pressure on inflation to unwind preceding rises to unwind preceding rises to unwind preceding rises, encourage growth to stabilise sterling to stabilise sterling to unwind preceding rises to unwind preceding rises, encourage growth to unwind preceding rises, encourage growth to unwind preceding rises, encourage growth to keep in line with US and maintain confidence, plus technical to stabilise sterling to unwind preceding rises to control sterling crisis

Main reason(s) for change given in official sources

−11/4/–11/2 −1/–11/4 +1 −21/2 +1 −2 +1 −1 −1/2 +1/2 −11/2 +2 +21/2 +1 +1 +1 −1 −3 −1/2 −1/2 +2 (+3) −3 −2 −1 −1/2 −1/4 +1/2 +1/2 +1/2 −1/4 −1/4 −1/4 −1/4 +1/4 +11/4 +1/4 −2.5 +1 12.625 11.5 12.5 10 11 9 10 9 8.5 9 7.5 9.5 12 13 14 15 14 11 10.5 10 12 (15) 9 7 6 5.5 5.25 5.75 6.25 6.75 6.5 6.25 6 5.75 6 7.25 7.5 5 6

dollar weaker US interest rate cut oil-price fall overseas interest rate cuts March sterling very weak Louvre Accord February sterling weak concerted overseas interest rate cuts German interest rate cuts stronger evidence of domestic boom abandonment of DM-shadowing rising overseas interest rates rising overseas interest rates further evidence of boom conditions slow response to policy tightening other European interest rate rises UK enters ERM sterling stable within ERM sterling stable within ERM narrowing UK–German interest gap severe ERM turbulence sterling now suspended from ERM new monetary policy framework strong disinflationary pressures improved inflation prospects likely impact of November Budget above trend growth Government defeat on VAT evidence of strong growth improving inflation prospects overseas interest rate cuts December good inflation prospects sterling stronger evidence of strong upswing evidence of strong upswing evidence of strong upswing economic downturn economic recovery

Source: compiled from Table 8A.1 and from the underlying official sources.

21.3.85–19.4.85 12.6.85–29.7.85 9.1.86 19.3.86–23.5.86 14.10.86 10.3.87–11.5.87 6.8.87 23.10.87–4.11.87 4.12.87 2.2.88 17.3.88–17.5.88 2.6.88–28.6.88 4.7.88–25.8.88 25.11.88 24.5.89 5.10.89 8.10.90 13.2.91–12.7.91 4.9.91 5.5.92 16.9.92 17.9.92–22.9.92 16.10.92–13.11.92 26.1.93 23.11.93 8.2.94 12.9.94 7.12.94 2.2.95 13.12.95 18.1.96 8.3.96 6.6.96 30.10.96 6.5.97–6.11.97 4.6.98 8.10.98–10.6.99 8.9.99–10.2.00

to unwind preceding rises as before and to maintain domestic activity to counter sterling fall in response to oil-price fall to undo last rise and maintain domestic activity to counter ‘excessive’ depreciation of sterling to moderate sterling appreciation to reverse cuts unjustified on domestic grounds to counter equity market crash as before and to prevent rise of sterling against DM to counter domestic inflation to moderate sterling appreciation to undo previous falls and to counter domestic inflation to counter domestic inflation to counter domestic inflation to counter domestic inflation and depreciation to counter domestic inflation to relax monetary tightening which now taking effect to stimulate domestic recovery to stimulate domestic recovery to stimulate domestic recovery to keep sterling within ERM to undo last rise to stimulate domestic recovery to stimulate domestic recovery to stimulate domestic recovery to stimulate domestic recovery to forestall inflation to forestall inflation to forestall inflation to allow growth to allow growth to allow growth to allow growth to forestall inflation to forestall inflation to forestall inflation to allow growth and prevent inflation undershoot to prevent inflation

122 The Making of Monetary Policy in the UK, 1975–2000 Table 8.2 Frequency of particular concerns

Phase in months

Cells in which one Cells in which Cells in which Cells in which Total number or other exch rate growth/inflation monetary growth fiscal policy has of cells has

or


has

or


has

or




or




January 1975 to February 1977 March 1977 to May 1979 June 1979 to March 1981 April 1981 to February 1985 March 1985 to February 1987 March 1987 to February 1988 April 1988 to September 1990 October 1990 to 16 September 1992 17 September 1992 to April 1997 May 1997– (December 2000) January 1975 to September 1990 October 1990 to December 2000

11

6

3

4

2

8

3

0

4

4

5

2

4

2

1

14

7

1

3

1

5

3

0

0

0

5

3

1

0

0

6

4

5

1

0

4

1

2

0

0

13

0

9

0

0

7

0

7

0

0

54

28

13

14

8

24

1

18

0

0

1985

1990

1995

Source: compiled from Table 8A.1.

18 16 14 12 10 8 6 4 2 0 1975

1980

Figure 8.1 Policy rate (daily data), 1975–2000 (Source: Bank of England)

2000

Interest Rate Decisions and the Exchange Rate 123 phase of monetary policy for which the official sources indicate that the interest rate change was motivated by a major concern (i.e. two or three ticks) with the exchange rate, with domestic growth or inflation, with monetary growth or with fiscal policy.2 It is immediately clear from these summary tables that the sterling exchange rate was a frequent concern in interest rate changes between 1975 and 1990, but much less so thereafter. Table 8.1 gives the exchange rate as a ‘main reason’ for change in 15 out of 54 cells between January 1975 and September 1990, but in only one (the quickly reversed rise of Black Wednesday) between October 1990 and December 2000.3 Table 8.2 shows that the exchange rate was a major concern in 28 out of 54 cells in the earlier period, but only one out of 24 in the later. In addition, Table 8A.1 contains four cells in the pre-ERM period where the exchange rate ‘favours’ the interest rate change but domestic growth or inflation is against it; while between 1990 and 2000 there is one case of this kind but five cases where domestic growth or inflation ‘favours’ the interest rate change but the exchange rate is against it. The broad difference between these subperiods is not unexpected, but the location of the dividing line is striking: the exchange rate apparently ceased to be of major importance for interest rate decisions before entry into, rather than after exit from the ERM. Standard domestic concerns, classified in Table 8A.1 as ‘domestic growth or inflation’, were important in some of the earlier phases though not all, but unequivocally important after 1988 and particularly under inflation targeting from 1992. Table 8.1 gives domestic concerns of some kind (including entries such as ‘to relax financial squeeze on companies’) as the main reason for an interest rate change in 21 out of 54 cells for 1975–90, but 21 out of 24 in 1990–2000. Table 8.2 shows that, while domestic growth/inflation was a frequent major concern between January 1975 and February 1977 and between June 1979 and March 1981, from April 1988 onwards it was the dominant concern. The 1975–77 and 1979–81 phases cover the two highest peaks in inflation and the efforts to control it, so the emphasis on inflation is unsurprising. The priority to inflation under inflation targeting since 1992 is also to be expected, but the timing of the shift to domestic concerns, which roughly parallels the shift away from external concerns discussed in the previous paragraph, is perhaps more striking. Monetary growth appears to have been important only in the first decade. Table 8.1 gives monetary control as a main reason for interest rate changes only in eight cells, all of them between September 1976 and November 1979. Table 8.2 shows monetary growth as a major concern in 13 out of 38 cells up to February 1985, but there is only one such entry later (July– August 1988). In Table 8A.1 domestic monetary growth refers (by default) almost exclusively to broad money growth. A separate but parallel analysis of the contribution of concerns about M0 growth to interest rate decisions in the official sources, between the 1984 Budget when M0 targets were first set and May 1997 when the last ‘monitoring range’ for M0 lapsed, found no cases where the level or rate of growth of M0 either absolutely or in relation to its target was the driving force behind an interest rate change.4 Finally, domestic fiscal policy was never a main reason for a change in interest rates but it figures as a major concern in a few cases in the late 1970s and early 1980s. These cases were all ones where the PSBR was both large and erratic (that is, it turned out much higher 2 There are minor differences between the months of the phases in this table and those used elsewhere in the book, including Chapter 9, in order to avoid splitting the cells between phases. 3 These numbers do not include the interest rate changes designed to unwind preceding rises which were associated with exchange rate concerns—at least another six cells in the earlier but only one in the later years. 4 The nearest cases are March 1986 (BEQB, 1986b: 185), February 1988 (1988b: 180–81) and July 1988 (1988d: 485).

124 The Making of Monetary Policy in the UK, 1975–2000 or—less often—lower than forecast), and in a period when the degree of funding of the PSBR was variable, and an instrument of policy. Two points should be noted about Table 8A.1 and its derivatives. First, in compiling such a table it is impossible to avoid the exercise of judgement in the process of interpreting the official sources.5 It is possible that some of these judgements could be challenged. However, the points made subsequently from the table are typically based on findings which are clear-cut rather than marginal, so that a high proportion of the interpretations would have to be wrong for a different conclusion to be drawn. Second, while the official sources may naturally tend to justify official actions in terms of what was meant to be the monetary framework at the time, the differences discovered between different phases of policy are sufficiently clear and consistent to suggest that they depict real differences in the authorities’ concerns. In particular, the exchange rate was not formally specified in the monetary framework except for 1987–88 and 1990–92, so the findings on the exchange rate cannot be a reflection of how the authorities thought they should justify themselves rather than how they really thought about their decisions. And the findings on monetary growth are in any case inconsistent with such self-justifying behaviour.6

8.2 A PRELIMINARY LOOK AT MISALIGNMENTS The analysis of the external dimension in this and the next chapter refers primarily to the nominal effective exchange rate, which averages the effects of varying movements in other currencies with weights reflecting the importance of each country for UK trade.7 Most attention is paid to quarterly average data which excludes higher-frequency variations whose impact on the economy or on policy is likely to be limited, but some references are made to monthly data. The real exchange rate is measured by the IMF index of relative unit labour costs; this is one of a number of broadly similar measures of competitiveness or the real exchange rate but one that focuses on costs rather than actual prices charged in export or wholesale markets for goods actually traded, which may reflect variations in mark-ups and in goods traded. Figure 8.2 shows the movement of these two series (indexed on 1995 = 100) for the whole of the period for which the relative unit labour costs series is available, 1975–98.8 It is clear that the nominal and real rates move closely together, particularly after the first few years,9 and this closeness suggests that the predominant influence in the short run is from the nominal exchange rate to the real, rather than the other way round: it is easy to see how, given sticky prices, changes in the nominal rate can cause changes in the real rate, but more difficult to 5 This is, of course, judgement on what the authorities claimed to be concerned about, not what they were ‘really’ concerned about. 6 This suggests that the absence of concern with monetary growth after 1985 reflects a real change in official behaviour (despite the argument in the Loughborough Lecture that even if monetary targets were abandoned liquidity and credit would still need to be monitored—BEQB, 1986d: 507), rather than the fact that policy no longer needed to be justified by reference to the monetary targets. 7 The effective rate moves closely with the DM/£ rate, since the Deutschemark (DM) and DM-related European currencies have large weights in the effective rate. However, the DM/$ rate is relatively stable over most of the period, with the exception of the rise in the dollar in the mid-1980s, so on the whole, and in the medium term, the pound has varied in a similar way against both DM and dollar. 8 Data from International Financial Statistics. 9 The correlation between the percentage changes in each quarter from the previous quarter for the two series is 0.91 for the whole period, and 0.95 for 1980–98. See also Mussa (1986) for a more general discussion of the relationship between nominal and real exchange rates.

Interest Rate Decisions and the Exchange Rate 125 180.00 160.00 140.00 120.00 100.00 80.00

UK real effective exch rate (relative unit labour costs)

1998Q1

1997Q1

1996Q1

1995Q1

1994Q1

1993Q1

1992Q1

1991Q1

1990Q1

1989Q1

1988Q1

1987Q1

1986Q1

1985Q1

1984Q1

1983Q1

1982Q1

1981Q1

1980Q1

1979Q1

1978Q1

1977Q1

1976Q1

1975Q1

60.00

UK nominal effective exch rate

Figure 8.2 UK exchange rates, 1975–1998 (Source: International Financial Statistics)

see by what mechanism changes in the latter could (within the quarter) lead to corresponding changes in the former.10 The general trend in the nominal rate is a strong depreciation over the period as a whole, with marked appreciations in 1979–81 and 1996–98. The general trend in the real rate is less clear: for much of the 1980s and 1990s the rate varied within a range of about 20%, but this was preceded by periods when the rate was much lower and then much higher, and in the last years of the 1990s the rate again rose well outside its previous range. For what follows it will be necessary to have a rough idea (but not more than that) of the extent of misalignment in different periods. Attempts to do this precisely by using purchasing power parity or measures such as the fundamental equilibrium exchange rate (FEER) are inevitably open to objections related to the concepts as well as their measures. The alternative approach followed here is to identify as misalignments substantial deviations from the average value for the real rate through some period over which it seems unlikely that the exchange rate was on average seriously overvalued or undervalued. The average for the whole period of 1975–98 is 107.16, but that period includes several years at the beginning when it is widely agreed that sterling was heavily undervalued (by an amount that is without later parallel). The average for 1980–98, which excludes the 1970s but includes more recognised overvaluation than undervaluation, turns out to be 112.99.11 These two values (which are roughly 5% apart) are taken to indicate a ‘central range’ within which the level of zero misalignment should probably be located. Values 5% above the higher and 5% below the lower of these two values, i.e. 118.64 and 101.80, are then taken as the ‘wide band’ cut-off points for identifying ‘large’ 10 See, for example, Obstfeld (2001) for a discussion of pricing to market and local currency pricing as mechanisms of sticky prices, and their effects on the relationship between real and nominal exchange rates. 11 The average for 1980–89, which was used to ‘justify’ the rate at which sterling entered the ERM in 1990 (e.g. BEQB, 1991a: 54) is 114.26, but that decade contains a larger period at the start when sterling is agreed to have been overvalued but only much shorter periods of likely undervaluation.

126 The Making of Monetary Policy in the UK, 1975–2000 150.00 140.00 130.00 120.00 110.00 100.00 90.00 80.00 70.00

relative unit labour costs (IMF index) rulc average 1975-98 = 107.16 rulc average 75-98-5% = 101.80

1999Q1

1998Q1

1997Q1

1996Q1

1995Q1

1994Q1

1993Q1

1992Q1

1991Q1

1990Q1

1989Q1

1988Q1

1987Q1

1986Q1

1985Q1

1984Q1

1983Q1

1982Q1

1981Q1

1980Q1

1979Q1

1978Q1

1977Q1

1976Q1

1975Q1

60.00

rulc average 1980-98 = 112.99 rulc average 80-98+5% = 118.64

Figure 8.3 UK real exchange rate, 1975–1998

misalignments. Figure 8.3 plots the relative unit labour costs series against these four values. It suggests a period of very large undervaluation in the second half of the 1970s; periods of very large overvaluation in the early 1980s and the late 1990s; a period of large undervaluation after the exit from the ERM; and some shorter episodes of misalignment during the 1980s.12 The ‘central range’ should be thought of as a range for the long-run equilibrium rate, a range which is probably wide enough to cover possible movements in the long-run equilibrium as the result of, for example, changes in national debt positions or in consumer preferences.13 The actual real rate can be expected to vary around this long-run equilibrium, as the nominal, and hence the real, rate adjusts to absorb exogenous shocks and acts as part of the transmission mechanism for changes in monetary policy: a domestic monetary contraction, for example, causes the nominal rate to appreciate, which puts downward pressure on output (in the short term) and prices (in the short and longer term), and the latter causes the real rate to depreciate back towards the long-run equilibrium. Thus divergences from the long-run equilibrium due to divergent monetary policies, or indeed due to exogenous shocks, should be limited in scope and duration.14 Table 8.3 lists the ‘large’ misalignments over the period. Columns (3) and (4) note how each misalignment came about, in terms of the pattern of changes that led to it, and how it came to an end (see also Figure 8.3). It is clear that there are substantial differences in extent and duration between the major and minor misalignments and that, although higher UK inflation 12 It is worth noting that the period of ERM membership does not qualify as a period of large overvaluation on this basis: the real rate was above the central range but within the wide band. 13 It should also be emphasised that misalignments are identified here with respect to the broader ‘wide band’ rather than this central range. 14 The broad picture of misalignments given here is not in fact very dissimilar to that which comes out of analytically based exercises such as Barisone, Driver and Wren-Lewis (2000).

Interest Rate Decisions and the Exchange Rate 127 Table 8.3 Large misalignments How the misalignment came about (3)

How the misalignment was ended (4)

major undervaluation

repeated nominal depreciations 1972–76

1980Q2–1982Q4 (11 quarters)

major overvaluation

1985Q1 (1 quarter)

marginal undervaluation

1985Q3 (1 quarter)

near overvaluation

1986Q4–1987Q1 (2 quarters)

undervaluation

1992Q4–1994Q3 (8 quarters) 1995Q2–1996Q2 (5 quarters)

undervaluation

sustained nominal appreciations from 1979Q2 plus higher UK inflation sustained nominal depreciations from 1983Q3, plus January 1985 sterling crisis (against strong $) sharp nominal rebound from January 1985 crisis nominal depreciation after oil-price fall, especially 1986Q3, Q4 Black Wednesday nominal depreciation further nominal depreciation

faster UK inflation, nominal appreciations especially 1978Q1, 1979Q1 nominal depreciations especially late 1981

1997Q2–2000Q2 (13 quarters)

major overvaluation

Quarters (1)

Nature of misalignment (2)

1975Q1–1979Q1 (17 quarters)

undervaluation

sustained nominal appreciations from late 1996

nominal rebound from crisis

nominal depreciation from unwinding some nominal appreciation early 1987 UK faster inflation minor nominal appreciation, faster UK inflation [not yet]

Sources: see text.

was important in some episodes, nominal exchange rate movements have been the dominant factor in both the generation and the ending of misalignments.

8.3 A PRELIMINARY LOOK AT LARGE EXCHANGE RATE CHANGES It is also useful to identify ‘large’ exchange rate changes, which will be defined here as changes of 4% or more between one quarter and the next, and changes of 3% per quarter where the movement is sustained over two successive quarters.15 Table 8.4 lists the 20 quarters in which the exchange rate changed by 4% or more and the two additional two-quarter periods in which 15 Initially a 5% change (quarter to quarter) seemed an appropriate cut-off, but closer examination of the data suggested that 4% was more of a ‘natural break’ than 5%. This 4% may be compared with the following benchmarks: width of normal band under original ERM = 4.5%; average size of realignment of central parities in the ERM = 4.8%; sterling depreciation between 11.9.92 and 25.9.92 = 9.4%.

128 The Making of Monetary Policy in the UK, 1975–2000 Table 8.4 Large exchange rate changes Quarter (1)

1975 Q2 + Q3 1976 Q2 1976 Q4 1978 Q2 1979 Q2 1979 Q3 1980 Q1 1980 Q4 1981 Q3 1983 Q1 1983 Q2 1984 Q4 + 85 Q1 1985 Q2 1986 Q1 1986 Q3 1986 Q4 1987 Q2 1989 Q4 1990 Q3 1992 Q4 1996 Q4 1997 Q1

Nominal effective exchange rate change (e) (%) (2)

−6.29 −7.82 −7.93 −5.37 +5.10 +4.52 +4.9 +5.04 −6.78 −9.28 +5.33 −6.50 +8.65 −6.70 −6.15 −5.83 +4.49 −4.34 +6.28 −11.80 +6.87 +6.08

Prior over-/undervaluation? (3)

Did e improve/worsen over-/under-valuation? (4)

under under under under under no no over over over no no under no no no under no no no no no

worsen worsen worsen worsen improve — — worsen improve improve — led to marginal undervaluation improve — — led to undervaluation improve — — led to undervaluation — —

Sources: see text.

it changed by 6% or more.16 It should be noted that these large changes are sufficient to describe the broad swings in the exchange rate over the period as a whole. These large exchange rate changes can now be considered in relation to the extent of misalignment as identified in the preceding section. Column (3) of Table 8.4 identifies the existence of prior over-/undervaluation in terms of whether the real exchange rate was previously outside the wide band as defined above, i.e. above 118.64 or below 101.80. Column (4) then indicates whether the change improved or worsened the prior misalignment, that is, whether it moved the real rate towards or away from the wide band. On this basis it appears that only five out of 22 of the large exchange rate changes had the effect of improving a pre-existing misalignment; five made pre-existing misalignments worse; and three led to undervaluations where no misalignment previously existed. The other nine large changes involved movements entirely within the wide band. Moreover, apart from the fact that four of the five cases where the change worsens a pre-existing misalignment occurred in the second half of the 1970s, it is difficult to read the table as implying any substantial improvement over time (that is, any tendency for exchange rate changes to become more misalignment-reducing). Other aspects of these large exchange rate changes are considered further in Chapter 9. 16 Applying the same criteria for Germany, Japan and the US identifies 4, 40 and 28 quarters respectively where the quarter to quarter change in the nominal effective rate exceeded 4%, and another 0, 3 and 1 additional episodes respectively where the two-quarter change exceeded 6%.

Interest Rate Decisions and the Exchange Rate 129

8.4 INTERIM CONCLUSIONS This chapter was designed to lay the foundations for the more detailed analysis of monetary policy in Chapter 9. But it has already established some points of interest. First, the identification of the main concerns expressed in the official sources with respect to each interest rate change has highlighted the importance of the exchange rate between 1975 and 1990 but not thereafter; the greater importance of standard domestic concerns (economic growth and inflation) from 1988 onwards; and the importance of monetary growth between 1975 and 1985 (but not thereafter). Second, a simple broad-brush criterion has been introduced to identify large misalignments of the exchange rate. Most of these misalignments were brought about, and most were ended, mainly by changes in the nominal exchange rate rather than in price or cost inflation. Third, a simple criterion has been used to identify large exchange rate changes, and the 22 such changes are sufficient to describe the broad movements in the exchange rate over the period. Most of these large exchange rate changes did not have the effect of reducing prior misalignments, and some of them made sterling more misaligned than it had been before.

New level

11.25 11 10.75 10.5 10.25 10 9.75 10 11 12 11.75 11.5 11.25 11 10.75 10.5 10 9.5 9.25 9 10.5 11.5

Size of change

−1/4 −1/4 −1/4 −1/4 −1/4 −1/4 −1/4 +1/4 +1 +1 −1/4 −1/4 −1/4 −1/4 −1/4 −1/4 −1/2 −1/2 −1/4 −1/4 +11/2 +1

Date(s) of change

20.1.75 27.1.75 10.2.75 17.2.75 10.3.75 24.3.75 21.4.75 5.5.75 28.7.75 6.10.75 17.11.75 1.12.75 29.12.75 5.1.76 19.1.76 26.1.76 2.2.76 9.2.76 1.3.76 8.3.76 26.4.76 24.5.76

yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ no no no no no no no no no no no no no no no no

German interest rates

yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / 0 no yes / ↑ yes / ↑ no no no no yes / ↓ yes / ↓ yes / ↓ yes / ↓ no no no yes / ↑

US interest rates

Explicit mention of / current or recent change in (↑, ↓ , 0):

Table 8A.1 Explicit concerns underlying interest rate changes

APPENDIX 8.1





















Exchange rate vs dollar




Exchange rate vs DM











Domestic growth or inflation











Domestic monetary growth

Main concerns mentioned behind interest rate change:
indicates factor in favour of change, X factor against, number of ticks or crosses indicates strength of concern

INTEREST RATE CHANGES




Domestic fiscal policy

13 15 14.75 14.5 14.25 14 13.25 12.25 12 11 10.5 9.5 9.25 9 8.75 8.25 8 7.5 7 6.5 6 5.5 5 7 6.5 7.5 8.75 9 10 12.5 14

+11/2 +2 −1/4 −1/4 −1/4 −1/4 −3/4 −1 −1/4

−1 −1/2 −1 −1/4 −1/4 −1/4 −1/2 −1/4 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 +2 −1/2

+1 +11/4 +1/4 +1 +21/2 +11/2

13.9.76 7.10.76 22.11.76 20.12.76 29.12.76 10.1.77 24.1.77 31.1.77 3.2.77

10.3.77 21.3.77 31.3.77 12.4.77 18.4.77 25.4.77 2.5.77 16.5.77 8.8.77 15.8.77 12.9.77 19.9.77 10.10.77 17.10.77 28.11.77 9.1.78

12.4.78 8.5.78 15.5.78 8.6.78 9.11.78 7.2.79

no no no no no no

no no no no no no no no no no no no no no no no

no no no no no no no no no

yes / 0 yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑

no no no no no no no yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ no yes / ↑

no no yes / ↓ yes / ↓ no no yes / ↑ yes / ↑ yes / ↑






















































































































































continues overleaf

























New level

13 12 14 17 16 14 12 14 16 15.5 15 14.5 14 13.5 13 12.5 12 11.5 11 10.5 10 9.5 9 10.125 11

Size of change

−1 −1 +2 +3 −1 −2 −2 +2 +2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 +11/8 +7/8

Date(s) of change

1.3.79 5.4.79 13.6.79 15.11.79 3.7.80 25.11.80 11.3.81 16.9.81 1.10.81 14.10.81 10.11.81 4.12.81 25.1.82 25.2.82 12.3.82 8.6.82 14.7.82 2.8.82 18.8.82 31.8.82 7.10.82 14.10.82 5.11.82 29.11.82 12.1.83

Table 8A.1 (continued )

no no no no no yes / ↓ yes / ↑ no no yes / ↓ yes / ↓ no yes / ↓ no no no no no no yes / ↓ yes / ↓ yes / ↓ yes / ↓ no no

German interest rates

no no no yes / ↑ yes / ↓ yes / ↑ yes / ↓ yes / 0 yes / ↓ yes / ↓ yes / ↓ yes / 0 yes / ↑ yes / ↓ yes / ↓ no yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / ↓ yes / 0

US interest rates

Explicit mention of / current or recent change in (↑, ↓ , 0):


































































Exchange rate vs dollar





Exchange rate vs DM





































Domestic growth or inflation






















X

X

























X

Domestic fiscal policy

Domestic monetary growth

Main concerns mentioned behind interest rate change:
indicates factor in favour of change, X factor against, number of ticks or crosses indicates strength of concern

16.3.83 15.4.83 15.6.83 4.10.83 7.3.84 15.3.84 10.5.84 27.6.84 9.7.84 12.7.84 9.8.84 10.8.84 20.8.84 7.11.84 20.11.84 23.11.84 11.1.85 14.1.85 28.1.85 21.3.85 29.3.85 12.4.85 19.4.85 12.6.85 15.7.85 29.7.85 9.1.86 19.3.86 8.4.86 21.4.86 23.5.86 14.10.86

10.5 10 9.5 9 8.875 8.625 9.125 9.25 10 12 11.5 11 10.5 10 9.875 9.625 10.5 12 14 13.5 13.125 12.875 12.625 12.5 12 11.5 12.5 11.5 11 10.5 10 11

−1/2 −1/2 −1/2 −1/2 −1/8 −1/4 +1/2 +1/8 +3/4 +2 −1/2 −1/2 −1/2 −1/2 −1/8 −1/4 +7/8 +11/2 +2

−1/2 −3/8 −1/4 −1/4 −1/8 −1/2 −1/2 +1 −1 −1/2 −1/2 −1/2 +1 no no no no no no no no yes / ↓ yes / ↓ no no yes / 0

no no no no no no no no yes / ↑ yes / ↑ no no no no no no no no no no no no no yes / ↓ no no yes / 0 yes / ↓ yes / ↓ yes / ↓ no no

no no no yes / 0 yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / 0 yes / 0 yes / 0 yes / ↓ yes / ↓ yes / ↓ no no no









































































































X X X X








X X X X




X X
















X X







X

X X

continues overleaf







X X X X








X X

New level

10.5 10 9.5 9 10 9.75 9.5 9 8.5 9 8.5 8 7.5 8 8.5 9 9.5 10 10.5 11 12 13 14 15 14

Size of change

−1/2 −1/2 −1/2 −1/2 +1 −1/4 −1/4 −1/2 −1/2 +1/2 −1/2 −1/2 −1/2 +1/2 +1/2 +1/2 +1/2 +1/2 +1/2 +1/2 +1 +1 +1 +1 −1

Date(s) of change

10.3.87 18.3.87 28.4.87 11.5.87 6.8.87 23.10.87 29.10.87 4.11.87 4.12.87 2.2.88 17.3.88 11.4.88 17.5.88 2.6.88 6.6.88 22.6.88 28.6.88 4.7.88 18.7.88 8.8.88 25.8.88 25.11.88 24.5.89 5.10.89 8.10.90

Table 8A.1 (continued )

no no no no no no no yes / ↓ yes / ↓ no no no no no no yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / ↑ no yes / 0 yes / ↑ no

German interest rates

no no no no no yes / ↓ yes / ↓ yes / ↓ no yes / ↓ no no no yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes/ ↑ yes/ ↑ yes/ ↑ yes/ ↑ no no no no

US interest rates

Explicit mention of / current or recent change in (↑, ↓ , 0): Exchange rate vs dollar








































Exchange rate vs DM













































X


X X X


















X

XX XX XX
































X X X









X X X

Domestic monetary growth




X X X X

Domestic growth or inflation

Main concerns mentioned behind interest rate change:
indicates factor in favour of change, X factor against, number of ticks or crosses indicates strength of concern







Domestic fiscal policy

13.5 13 12.5 12 11.5 11 10.5 10 12 10 9 8 7 6 5.5 5.25 5.75 6.25 6.75 6.5 6.25 6 5.75 6 6.25 6.5 6.75 7 7.25 7.5

−1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 −1/2 +2 −2 −1 −1 −1 −1 −1/2 −1/4 +1/2 +1/2 +1/2 −1/4 −1/4 −1/4 −1/4 +1/4 +1/4 +1/4 +1/4 +1/4 +1/4

+1/4

13.2.91 27.2.91 22.3.91 12.4.91 24.5.91 12.7.91 4.9.91 5.5.92 16.9.92 17.9.92 22.9.92 16.10.92 13.11.92 26.1.93 23.11.93 8.2.94 12.9.94 7.12.94 2.2.95 13.12.95 18.1.96 8.3.96 6.6.96 30.10.96 6.5.97 6.6.97 10.7.97 7.8.97 6.11.97

4.6.98

no

no no no no no no yes / ↑ yes / 0 yes / ↓ no no yes / ↓ no no yes / ↓ yes / 0 yes / 0 yes / 0 yes / 0 yes / ↓ yes / ↓ yes / 0 yes / ↓ yes / 0 yes / 0 no no yes / 0 yes / ↑ no

no no no no no no yes / ↓ no no no no no no no no yes / ↑ yes / ↑ yes / ↑ yes / ↑ yes / 0 yes / ↓ yes / 0 yes / 0 yes / 0 yes / 0 yes / 0 no yes / 0 yes / 0 X

X

X X XX XXX XXX XX





X

X

X

XX XX XXX XXX XX

X





X


















X

X

X

X

X











XXX























































































continues overleaf





7.25 6.75 6.25 6 5.5 5.25 5 5.25 5.5 5.75 6

−1/4 −1/2 −1/2 −1/4 −1/2 −1/4 −1/4 +1/4 +1/4 +1/4 +1/4

8.10.98 5.11.98 10.12.98 7.1.99 4.2.99 8.4.99 10.6.99 8.9.99 4.11.99 13.1.00 10.2.00

no no yes / ↓ yes / no yes / no yes / ↓ yes / no yes / no yes / ↑ yes / no yes / ↑

German interest rates

yes / ↓ yes / ↓ yes / ↓ yes / no yes / no yes / no yes / no yes / ↑ yes / no yes / no yes / ↑

US interest rates

Explicit mention of / current or recent change in (↑, ↓ , 0):

X X












Exchange rate vs dollar

X X

Exchange rate vs DM




































Domestic growth or inflation

Domestic monetary growth

Main concerns mentioned behind interest rate change:
indicates factor in favour of change, X factor against, number of ticks or crosses indicates strength of concern Domestic fiscal policy

Notes: 1. Interest rates refer to banks’ base rates in the UK and official short term and/or prime rates in Germany and the US. Current or recent changes cover the month or so before and the day or so after the UK change. 2. When the clearing banks’ base rates were adjusted by differing amounts, the (unweighted) average of the two changes is taken.

New level

Size of change

Date(s) of change

Table 8A.1 (continued )

9 Monetary Policy and the Exchange Rate In the standard case for separate currencies with variable exchange rates the exchange rate is seen as an important adjustment mechanism or an important policy instrument. It is argued that fixing the exchange rate, particularly in the form of entry into a monetary union, involves giving up an adjustment mechanism which can deal efficiently and more or less automatically with asymmetric exogenous shocks, and/or which plays an important role in the transmission mechanism; alternatively it involves giving up a policy instrument which the monetary authorities can use in a more deliberate fashion to deal with such shocks. By contrast, advocates of monetary union tend to see the exchange rate more as a ‘source of extraneous asymmetric shocks’ (Buiter, 2000: 236), whose flexibility is not worth preserving.1 This and the preceding chapter are designed to provide evidence on the extent to which the possibility for the exchange rate to vary has been useful or unhelpful for UK monetary policy over the last two and a half decades. The aim is to answer through a limited and structured analysis the following questions. Did the exchange rate adjust automatically towards longrun equilibrium? Did the exchange rate perform a useful role as part of the transmission mechanism of monetary policy? Did the monetary authorities have defined preferences for the level and change of the exchange rate, and did the exchange rate move in accordance with those preferences? Did exchange rate movements, or exchange market pressures, present themselves to the monetary authorities as unexpected and unwelcome exogenous shocks? Chapter 8 presented an analysis of each interest rate change, or group of changes, with particular reference to the concerns expressed by the monetary authorities at the time of the change, and then made a preliminary examination of ‘large’ misalignments and ‘large’ exchange rate changes. The first, and longest, section of this chapter examines the development of monetary policy more broadly in terms of the sequence of interest rate decisions in each phase, with particular reference to the authorities’ attitudes to the level and rate of change of the exchange rate. Sections 9.2 to 9.4 use that broader perspective to re-examine the misalignments and large exchange rate changes discussed in Chapter 8, and to review the sterling crises experienced over the period. Section 9.5 concludes.

9.1 MONETARY POLICY AND THE EXCHANGE RATE, BY PHASES This section discusses the reaction of the monetary authorities to the level of and changes in the exchange rate in the ten different phases of policy identified in Chapter 2. Table 9.1 1 See also Flood and Rose (1999) who show that exchange rates are more volatile when they are flexible than when they are fixed, for a given level of underlying macroeconomic volatility.

138 The Making of Monetary Policy in the UK, 1975–2000 Table 9.1 Phases of monetary policy Phase [number of months] (1)

Monetary policy framework (2)

Exchange rate movement (nominal effective) (3)

January 1975 to March 1977 [25] April 1977 to May 1979 [28] June 1979 to March 1981 [22] April 1981 to March 1985 [48]

informal monetary targets monetary targets (with discretion) monetary targets (MTFS) with less discretion monetary targets (MTFS) with more discretion

persistent depreciation

April 1985 to February 1987 [23]

no binding targets

depreciation on balance, with appreciations late 1981, mid-1983 depreciation on balance

March 1987 to March 1988 [13] April 1988 to September 1990 [30] October 1990 to September 1992 [24] October 1992 to April 1997 [55] May 1997 to July 2000 [39]

DM-shadowing

minor swings

no explicit target

swings but no clear trend

fixed exchange rate (ERM) inflation target, Bank advising Chancellor inflation target, decisions by MPC

very minor swings then sharp fall no clear trend then rise

swings but no clear trend persistent appreciation

continuing rise then no clear trend

Misalignment? (4) undervaluation throughout undervaluation throughout overvaluation from 80Q2 overvaluation till 82Q4, near-undervaluation 85Q1 85Q3 overvaluation, 86Q4–87Q1 undervaluation none undervaluation 86Q4–87Q1 none undervaluation from 92Q4 to 94Q3 overvaluation from 97Q2

summarises the monetary framework, the movement of the nominal exchange rate and the extent of misalignment in each phase. Table 9.2 provides some basic data for each phase, on nominal and real exchange rates, inflation, and nominal and real interest rates2 (see also Figure 8.1 for the movement in the policy interest rate in each phase, and Figures 8.2 and 8.3 for the movement in nominal and real exchange rates). Each subsection identifies so far as is possible the authorities’ attitudes towards the level of and changes in the exchange rate, and how this affected the policy actions which they took; these parts typically contain a large number of references to, and a few quotes from, the official sources. The final subsection of Section 9.1 brings this material together in summary form. 9.1.1 January 1975 to March 1977 In this phase official macroeconomic policy thinking was still largely Keynesian, with a heavy emphasis on incomes rather than monetary policies and a reluctance to manipulate excess demand in order to control inflation at a time when, it was believed, there was much excess capacity.3 However, the Bank of England had had some internal targets for broad money since 1974, and a monetary target was first announced in July 1976. The nominal exchange rate depreciated in every quarter except the last, and the period from March to November 1976 was one of more or less continuous sterling crisis. The real exchange rate was heavily undervalued 2 Real interest rates are measured ex post as nominal rates minus consumer price inflation over the preceding 12 months. 3 See Artis and Cobham (1991, especially chapters 1 and 16) on the overall strategy with its preference for a long and gradual disinflation.

Monetary Policy and the Exchange Rate 139 Table 9.2 Exchange rates, inflation and interest rates by phases (monthly) Nominal exchange rate Phase (1) January 1975 to March 1977 April 1977 to May 1979 June 1979 to March 1981 April 1981 to March 1985 April 1985 to February 1987 March 1987 to March 1988 April 1988 to September 1990 October 1990 to September 1992 October 1992 to April 1997 May 1997 to July 2000

Real exchange rate

Inflation

Interest rate

average s.d. average (6) (7) (8)

s.d. (9)

Ex post real interest rate

average (2)

s.d. (3)

average (4)

s.d. (5)

average s.d. (10) (11)

140.8

13.2

[80.5]

[4.9]

20.0

4.8

10.7

1.5

−9.3

5.7

126.8

2.9

[87.3]

[9.0]

11.2

3.6

8.4

2.4

−2.7

5.4

145.2

8.8

119.8

12.1

16.6

2.8

14.3

1.4

−2.3

1.7

138.5

8.6

119.2

8.4

7.1

2.9

10.8

1.9

3.7

1.7

124.6

9.3

107.5

6.6

4.6

1.6

10.8

0.9

6.2

0.9

119.0

2.1

108.8

2.1

4.0

0.4

8.9

0. 6

4.9

0.5

120.8

4.9

110.3

4.1

7.5

1.7

12.7

2.0

5.2

1.1

118.8

1.8

115.1

2.3

5.7

2.3

10.7

1.3

5.0

1.0

103.8

4.4

104.2

5.3

2.5

0.7

5.7

0.6

3.2

0.5

122.8

3.2

140.0

9.1

2.7

0.9

6.1

0.8

3.4

0.4

Source: International Financial Statistics. Note: real exchange rate is relative unit labour costs normalised; numbers in square brackets for real exchange rate in first two phases are calculated from quarterly data (monthly data not available); inflation is consumer price inflation; interest rate is Treasury bill rate.

throughout. Nominal interest rates fluctuated within a relatively narrow range until the last few months, while ex post real rates were well below zero. The evidence is that the authorities probably did not want a major depreciation beyond that which had already occurred since the pound was floated in June 1972, but they wanted to maintain some competitive advantage.The BEQB contains numerous references to improvements or deteriorations in competitiveness (e.g. 1975b: 127, 1975c: 213, 1976c: 298, 1976d: 411). The ‘Assessments’ of September and December 1975 appear almost indifferent to the depreciation in 1975 Q2 and Q3, that of June 1976 shows some concern at that of March 1976, but the first evidence of strong concern about depreciation comes at the end of the crisis in December 1976 (1975c: 220–2, 1975d: 334–7; 1976b: 166; 1976d: 419–21).4 For most of 1975 UK interest rates moved—first down, then up—largely in line with US rates, and the BEQB refers specifically to the authorities’ wish to maintain the existing differential between them (about 4%) (1975c: 219). Between November 1975 and March 1976 there was a 4 According to Allsopp (1991: 32), with reference to the period up to the summer of 1976, ‘The exchange rate was regarded as too high in 1975: there was, it seems, a strategy of depreciation, though it is not clear what degree of devaluation was contemplated, or even whether it was formalised’. Britton (1991: 31) says that, after the fall in March 1976, ‘There was a suspicion in the markets that the British authorities would welcome a similar depreciation of the pound [to that of the French franc after it left the Snake]. That suspicion was indeed well-founded, although it would be wrong to say that a definite decision had been taken in the Treasury that the pound should be devalued.’ See also Smith (1987: 64).

140 The Making of Monetary Policy in the UK, 1975–2000 succession of ten small cuts that reduced rates by 3%, and these were justified in terms of both external factors (a decline in US rates) and domestic factors (recession, improving inflation expectations, low demand for bank advances) (BEQB, 1976a: 13, 17, 19). The last 0.25% cut occurred the day after a big slide in sterling, when sales of sterling by the Bank of England to prevent appreciation had been ‘misinterpreted by the market’ (BEQB, 1976b: 170–1). The Bank intervened heavily in the foreign exchange market,5 but did not raise interest rates until late April (by 1.5%) and late May (another 1%). Both of these rises were justified largely in terms of the exchange rate (BEQB, 1976b: 170, 1976c: 300). Sterling then became more stable, but did not reverse its previous fall. It depreciated again sharply in September and October, and interest rates were raised by 1.5% in September and 2% in October. These rises ( justified by the authorities in terms of domestic—monetary acceleration—as well as external factors, BEQB, 1976d: 409) took the UK–US differential, which had stood at 5–6% since late April, to about 10% and, together with other measures (the reintroduction of the corset in November and, in particular, the agreement with the IMF in early December which covered both fiscal policy and targets for domestic credit expansion) led to the ending of the long-drawn-out sterling crisis. Within weeks a partial unwinding of the process began, with the pound strengthening (despite foreign exchange purchases to rebuild the reserves) and interest rates declining by 3% in steps between late November 1976 and early February 1977 (which took the UK–US differential back to about 6%), and then embarking on a further decline in March. Overall, in this phase, it is probable that the authorities did not actively seek the depreciation that occurred, but they did not always resist it. Although they supported the pound repeatedly and heavily in the foreign exchange markets they were reluctant to use monetary policy to prevent or reverse the depreciation, and that reluctance was one of the reasons why the crisis of 1976 went on for as long as it did (compared, for example, to sterling crises in the 1980s).6 The repeated depreciations of sterling in 1975 and especially 1976 were seen by the authorities largely as unwanted and unexpected; they were regarded by them not as a useful way of dealing with the ‘shocks’ of domestic industrial and social tensions, but as an aggravating factor for those tensions (and a source of unwelcome pressure for the government). 9.1.2 April 1977 to May 1979 In this phase the authorities had accepted the need to control excess demand and monetary growth, and had introduced more serious targets for broad money which they eventually chose over an informal exchange rate target. The nominal exchange rate was relatively constant for the first three quarters, underwent a short-lived appreciation of 6–7% in late 1977/early 1978, and returned by April 1978 to its earlier level where it remained for the rest of the phase. However, as UK inflation continued to exceed that of other countries, sterling became progressively less undervalued in each quarter (with the exception of 1978 Q2 when there was a large nominal depreciation). Nominal interest rates continued to be reduced until a low of 5% in October 1977; they then rose more or less continuously until the two cuts of 1% in 5 No good series for official foreign exchange market intervention is available except for the period since 1997, so references to the scale of intervention in this and other episodes are based on statements in the official sources and on the data which are available for the change in official reserves and the ‘underlying’ change in reserves (excluding the effects of valuation changes and certain public sector borrowings). Neither of these series include the forward transactions which it is widely believed that the UK authorities undertook in certain episodes. 6 For a wider discussion of the causes of the prolonged sterling crisis of 1976 see also Allsopp (1991), Cobham (1982), and Artis and Cobham (1991: chapters 1 and 16).

Monetary Policy and the Exchange Rate 141 March and April 1979. Real rates were negative on average, but positive for the last year of the phase. There is little doubt that the stability of the (nominal) exchange rate was actively desired by the authorities. In September 1977 the BEQB stated that ‘[the] authorities’ aim has been to maintain stability in the exchange rate, taking the opportunity to rebuild the much-depleted reserves’, and at this point they argued that conflict between an exchange rate target and a monetary target was ‘not inevitable’ (1977c: 283; 299).7 The pound was first allowed to rise against a weak dollar (with policy putting more emphasis on the effective rate—BEQB, 1977c: 283, 305), and then, at the end of October 1977 when large capital inflows forced them to choose between the exchange rate and the monetary target, they chose the latter. In the event sterling appreciated only a little in the first month, but it went higher in January and February 1978 before falling back. In October 1978 the Governor of the Bank was able to report that: ‘In a world where exchange rates have fluctuated very widely, ours has been remarkably stable.’ He also referred to ‘policies of prudence in both fiscal and monetary policy—policies aimed at the progressive reduction of inflation, policies in short which not only underpin the domestic value of our currency but also the stability of its external value’ (BEQB, 1978d: 534). At the beginning of the phase interest rates were still declining in a further ‘unwind’ from the crisis level of 15% in November 1976, and this process took rates down to 8% by mid-May; between August and October another 6 cuts of 0.5% each brought MLR to 5%. For much of 1977 the authorities were resisting the upward pressure on sterling (which was partly a reflection of dollar weakness) by buying foreign currency and selling gilts to sterilise a large capital inflow, but they were also slowing the fall in interest rates (BEQB, 1977c: 301–2, 1977d: 434). The uncapping of sterling at the end of October was followed a month later by a 2% rise in MLR, associated with fears that the monetary target would still be overrun, but 0.5% of that was undone in January. However, between mid-April and mid-May 1978 rates were raised in four steps by 3.5% in the face of stronger evidence of monetary overshoot and changed external conditions (the recovery of the dollar and an upward move in US interest rates). Interest rates were raised in one step by 2.5% in November, and again by 1.5% in February 1979, for similar reasons. There was strong upward pressure on the exchange rate in March and April, and the Bank of England, after buying foreign exchange heavily in February and March, allowed sterling to appreciate (BEQB, 1979b: 126); it rose to the level it had attained briefly in January 1978 before falling back. The authorities also cut interest rates in two steps of 1% each in early March and early April (in the run-up to the general election of 3 May 1979). Overall in this phase, the authorities pursued with limited success an informal objective of nominal exchange rate stability. The conflict with their monetary target was resolved easily enough in 1977 (by a temporary appreciation), but by the end of the phase the monetary target was again at risk of being overshot and sterling had appreciated again (in a movement that was to turn out to be far from temporary). Meanwhile interest rates had varied unusually widely, going from 12% to 5% and back to 14% within 24 months, and there had been a high level of intervention in the foreign exchange market with heavy purchases of foreign currency in most months of 1977 and early 1979 but some heavy sales in 1978. The stability of the exchange rate up to early 1979 may have assisted the ‘industrial health’ of the UK economy and helped to reduce inflation (BEQB, 1978d: 534), but it had been bought at a high price and the policy was not clearly sustainable. 7 They also continued to show awareness of movements in competitiveness—e.g. 1977b: 140, 1978a: 5, 1978c: 333.

142 The Making of Monetary Policy in the UK, 1975–2000 9.1.3 June 1979 to March 1981 In this phase the (new) government regarded monetary targets as the main instrument of macroeconomic policy and the control of inflation as the primary objective. The nominal exchange rate continued to rise strongly with the exception of a brief dip in the autumn of 1979. With UK inflation exceeding that in most other industrialised countries, the real exchange rate appreciated in every quarter, and over the phase the UK moved from being substantially undervalued to being substantially overvalued. Nominal interest rates were raised in large steps in 1979, and then cut in 1980 and 1981; real rates were negative on average, and for the latter half of the phase. It is unlikely that the authorities actively favoured the appreciation that occurred. At the beginning they may have been somewhat indifferent—sterling was still marginally undervalued on the criterion used here in 1979 Q2, and the new government focused heavily on the domestic economy and its monetary targets. But the BEQB noted the deterioration in competitiveness both before and after sterling became overvalued in 1980 Q2, and by March 1981 it was arguing that ‘prospects of expansion stand to be largely dependent on how far industry can improve its competitive position’ (1981a: 20).8 The new government reversed the March and April 1979 interest rate cuts in its Budget in June, with the aim of bringing inflation under control and ensuring that monetary growth was within the target (BEQB, 1979c: 261, 265). This was also the main motivation for the 3% rise in November, although by then there had also been a significant dip in the exchange rate.9 The reductions of 1% in July 1980, 2% in November 1980 and another 2% in March 1981 seem to have been largely a response to a recognition of the very severe financial squeeze that companies were experiencing, in part because of the appreciation of sterling;10 they were made despite the overshooting of the monetary target (BEQB, 1980c: 264, 279; 1980d: 406–7; 1981b: 169).11 Overall in this phase, the authorities were concerned mainly with monetary control and the reduction of inflation. The appreciation of sterling was a largely unintended by-product of monetary tightening, together with the impact of the second oil-price shock on a country that was self-sufficient in oil, and the BEQB typically mentioned these two causes of appreciation together. Later academic attempts to distinguish and evaluate the causes of the appreciation, including Spencer (1986), Bean (1987) and Britton (1991), have found it difficult to allocate the blame, but all agree that monetary policy was an important cause.12 In any event, there can be no doubt that for the authorities at the time the appreciation constituted an unexpected and unwelcome development (and one which they tried to counter in part by intervening in the foreign currency market13 ). The depth and severity of the recession then led the authorities to lower interest rates despite an overshoot of the monetary target, in the first year of the MTFS. 8 The BEQB typically attributed the loss of competitiveness to both nominal appreciation and relatively high domestic cost inflation, and drew policy conclusions in terms of the latter rather than the former. For earlier references see, for example, BEQB, 1979c: 249,253, 264; 1980b: 127. 9 This dip, of 5.9% between July and November, is, however, barely mentioned in the BEQB. 10 The decision to reduce the rate was preceded by a lengthy debate in policymaking and governmental circles, in which the study by Niehans (1981) played an important role: see Smith (1987: 96–8) and Keegan (1984: chapter 5). 11 Some of the monetary overshoot could be explained in terms of the unwinding of distortions after the abolition of the corset—see BEQB, 1980c: 264–5. 12 See also Buiter and Miller (1981) for a model designed to explain how monetary tightening could have such effects on the exchange rate and the economy, and Forsyth and Kay (1980) for an influential contemporary paper which attributed the appreciation to the oil-price shock. 13 They also sold foreign currency to support sterling when it dipped in late 1979.

Monetary Policy and the Exchange Rate 143 9.1.4 April 1981 to March 1985 In this phase the government continued to pursue targets for broad money, supplemented in the later years by targets for other aggregates, but the exchange rate was taken into account as an indicator of the monetary stance, and in 1982 the original MTFS target ranges were adjusted upwards, in recognition of the fact that the trend in the velocity of broad money had changed. Sterling depreciated overall, despite more or less short-lived appreciations in late 1981 and mid-1983; the real exchange rate was initially overvalued but moved inside the wide band. Nominal interest rates followed a partly similar pattern: a continuation of the trend decline which had started in 1980, punctuated by rises in late 1981, early 1983 and mid-1984. Initially there seems little doubt that the authorities would have liked to eliminate the misalignment that had developed in the preceding phase. While for obvious reasons this intention was never made explicit, there are repeated discussions in the BEQB of the improvement in competitiveness as sterling depreciates and inflation falls (e.g. 1981b: 160, 1981d: 459, 1982a: 5), and early in the phase there were some sharp downward movements of the exchange rate about which the authorities seemed quite sanguine (BEQB, 1981a: 21, 1981c: 309–10). In addition, there is evidence from elsewhere that the tight Budget of 1981 was designed primarily to facilitate a reduction in interest rates; this would in turn allow the exchange rate to decline.14 On the other hand the depreciation seems at times to have gone too fast or too far for official comfort, and this led to sharp rises in interest rates in September/October 1981 and November 1982/January 1983. In the first of these cases the authorities believed that a further depreciation would have had serious consequences for inflation (BEQB, 1981d: 464), and interest rates were raised by 2% in mid-September and another 2% at the start of October, to reach 16%. They were then lowered in 14 steps of 0.5% to reach 9% in early November 1982. But in the middle of that month sterling fell sharply; in this case the authorities ‘did not seek the depreciation of the exchange rate which occurred; but nor did they try to prevent the underlying market movement . . . The fall in sterling clearly had implications for the assessment of monetary conditions . . . In the most recent three months, with the exchange rate weak, the authorities did not ultimately stand against a rise in interest rates’ (BEQB, 1983a: 19), and rates were raised by 1% in November 1982 and 1% in January 1983. In fact the real exchange rate first ceased to be seriously overvalued (on the criterion used here) in 1983 Q1; it rose again in response to the rise in interest rates, but it remained within the wide band. It is less obvious that the authorities favoured the depreciation which took place over much of the rest of the phase, a part of which reflected the strengthening of the dollar. In December 1983 the Bank considered that although the depreciation since the autumn of 1982 had contributed to an improvement in competitiveness, UK costs were ‘still relatively high when measured in common currency’ (BEQB, 1983d: 457).15 In June 1984 when upward pressure developed on interest rates (which had returned to 9% in October 1983 and gone below that in March 1984), the Bank of England stated publicly that ‘it saw no need on domestic monetary policy grounds for any general increase in the level of domestic interest rates’. However, in early July when sterling weakened against other currencies as well as the dollar, the Bank ‘endorsed’ temporary rises in banks’ base rates of 2.75% (BEQB, 1984c: 322–3). The exchange rate stabilised and interest rates began to fall within a month (they were almost back to 9% by the end of the year). 14 Walters (1986: 145); see also Smith (1987: 100–2). 15 Relative unit labour costs stood at 114.60 in 1983 Q4,

on the criteria used here.

which is above the central range but within the wider band

144 The Making of Monetary Policy in the UK, 1975–2000 January 1985 saw a full-fledged sterling crisis, in which the pound nearly went to parity with a dollar that was at a historic peak (in real as well as nominal terms). The authorities took the view that the weakness of sterling was ‘plainly reflecting extraneous developments—the strength of the dollar and market uncertainties about oil prices’, but whatever the cause of the depreciation, it ‘could, if it persisted, reach a point at which its possible inflationary consequences would call for an offsetting tightening of domestic monetary policy’ (BEQB, 1985a: 24), so when sterling weakened against non-dollar currencies as well the authorities acted ‘to demonstrate as clearly as possible that there had been no weakening in the Government’s resolve’ by raising interest rates by 2.5% and then by another 2% two weeks later (1985a: 5).16 Overall in this phase, the authorities succeeded in ending the overvaluation of sterling without incurring any lasting new misalignment, while at the same time inflation was substantially reduced. However, the gradual declines in interest rates were repeatedly interrupted by sharp depreciations to which the authorities felt they had to respond, and both the exchange rate and the interest rate were relatively unstable (Table 9.2). In addition, the strength of the dollar was clearly an unexpected and unwelcome exogenous factor towards the end of the phase. 9.1.5 April 1985 to February 1987 In this phase there were no binding targets. In principle there were still targets for broad money, but the targets for 1985/86 and 1986/87 were both overshot by wide margins (see Chapter 3). The exchange rate rose for the first few months but then depreciated by more, while the real exchange rate varied widely. Nominal interest rates were reduced during 1985, raised briefly again in January 1986 and then again in October 1986, with successive peaks being lower. Real rates were very high throughout. The exchange rate was a continuing influence on market sentiment and hence on policy, but fluctuations in oil prices were also important. At the beginning of the phase the authorities were dealing with the after-effects of the sterling crisis of January 1985, and interest rates were initially reduced in an unwinding from the crisis level, in the light of a weakening of the dollar. However, the Bank acted to brake a faster downward trend in interest rates because of concerns over monetary growth and inflationary pressures (BEQB, 1985b: 188, 190; 1985c: 361, 365), and the relatively high level of UK interest rates may have contributed to the strong rebound of the exchange rate in effective terms as well as against the dollar (BEQB, 1985c: 377). When the price of oil dropped sharply in December 1985, the UK authorities believed that some depreciation of sterling was appropriate (BEQB, 1986a: 6, 1986c: 331–2). They raised interest rates by 1% in early January, but resisted pressure for further rises and allowed an ‘orderly’ decline in sterling, which amounted to 8.1% as between November 1985 and February 1986 (BEQB, 1986a: 27–8, 32–4). Interest rates had fallen back almost to the preJanuary 1985 level by May, but sterling began to slide again from July, and this led to a 1% rise in October: ‘The authorities accepted that some rise in interest rates was required to maintain sufficiently firm monetary conditions, but judged that the market pressures were exaggerated by short-run influences. As a tactical matter, therefore, they delayed making any move until 16 The Bank noted that ‘the impression [had] gained ground, fuelled by conflicting press reports on the Government’s attitude to the exchange rate, that its commitment to the further reduction of inflation had been weakened, and that it was, instead, intent on lowering interest rates, being prepared to take a risk on the exchange rate and any inflationary consequences’ (BEQB, 1985a: 5). See also Lawson (1992: 467–71).

Monetary Policy and the Exchange Rate 145 those influences had subsided and there was a prospect that the market would stabilise around a 1% rise in the structure of interest rates’ (BEQB, 1986d: 474).17 However, this manoeuvre left sterling significantly undervalued in 1986 Q4 and 1987 Q1.18 Overall in this phase the monetary authorities clearly had difficulty in managing the exchange rate, which swung widely in nominal, and hence also in real, terms. Initially they were concerned to deal with the effects of the slide of January 1985, but the fluctuations in sterling were probably larger than they would have wished. Then when the oil price fell sharply the decline in sterling (which they favoured) eventually went too far and corrective action had to be taken. On the other hand, interest rates (and inflation) fluctuated less widely in this than in preceding phases. The ex post real rate continued to be highly positive. 9.1.6 March 1987 to March 1988 This is the phase in which Chancellor Lawson ‘shadowed’ the DM within the framework of the Louvre Accord, keeping the DM/£ rate within the range 2.90–3.00.19 In January and February 1987 the rate had been around 2.80, and Lawson is thought to have believed that sterling was undervalued at the time of the Louvre meeting on 21–22 February 1987, when international decisions were taken to try to stabilise the exchange rates between the dollar, the DM and the yen.20 However, a ‘large’ rise in March took the rate to just over 2.90 and it is from then that the shadowing experiment began.21 The real exchange rate, which had been initially undervalued, appreciated in 1987 Q2 and from 1987 Q4 moved into the central range. Nominal interest rates over this phase fluctuated from 11% to 9% to 10% to 9% to 8.5% to 9% to 8.5%.22 Real interest rates continued to be strongly positive. Throughout this phase the exchange rate was a decisive influence on monetary policy, with the authorities monitoring competitiveness and apparently hoping to preserve it (e.g. BEQB 1987b: 165, 1987d: 490). The authorities had considered the level of sterling too low, relative to the oil price, to prevent a rise in inflation (BEQB, 1987a: 27),23 but by the end of March 1987 they believed this had corrected itself (BEQB, 1987b: 183–4, 189). However, the strengthening of sterling created a policy dilemma: for domestic reasons they did not want interest rates to fall ‘too far or too fast: at the same time a substantial rise in the exchange rate would have had damaging consequences for industrial confidence’ (BEQB, 1987b: 189; see also 1987c: 344). This perception led to the interest rate cuts of March–April 1987 which were partly reversed in August when a weakening of sterling provided ‘the opportunity to raise interest rates without 17 Lawson’s own account (1992: chapter 52) suggests that the intention was also to delay the interest rate rise until after he had returned from the IMF–World Bank meetings in Washington, at which the Bundesbank was eventually persuaded to support the pound through a standby swap agreement with the Bank of England, and then to delay the rise until after the Conservative Party Conference. 18 In September 1986 the BEQB published an analysis of the various combinations of sterling oil price and effective exchange rate at which output and prices (and the PSBR and the current balance) would be unchanged (1986c: 331–2). At that time the economy was at a point giving no change in prices but an increase in output (from the stimulating effect of the depreciation), but the economy moved sharply into the zone of price increases in the following months (see Chapter 4). 19 Journalistic accounts of the DM-shadowing can be found in Keegan (1989: chapters 8 and 9) and Smith (1987: chapter 5); see also Lawson (1992: chapters 59, 63 and 64). 20 Funabashi (1988: 186). 21 The nominal effective rate was around 111 in January and February and went up to 116 in March. 22 These figures give successive peaks and troughs within the period; most of the falls were in steps of 0.5% but some of the rises were in larger steps. 23 The implication was that the economy was now to the south-east of the constant inflation line in their September 1986 analysis.

146 The Making of Monetary Policy in the UK, 1975–2000 prompting renewed upward pressure on sterling’, an opportunity that was accepted in the light of the buoyancy of domestic demand (BEQB, 1987d: 499). Interest rates were cut in October– December in the aftermath of the stock market crash along with those in other countries, though the BEQB noted that in these circumstances ‘it was important that the exchange rate should remain firm to counter any impression of monetary ease’ (1988a: 8). A part of the interest rate reduction was recouped in February, but the policy dilemma due to ‘conflicting indications from the domestic economy and the exchange rate . . . became particularly acute in March’ (BEQB, 1988b: 180); the policy of shadowing the DM was abandoned in March 1988 and sterling was allowed to appreciate significantly. The monetary tightening due to appreciation was ‘offset in part by a 12 % reduction in interest rates’ in mid-March (which was repeated in April and May) (BEQB, 1988b: 181). In this phase the authorities tried actively to stabilise the (DM/£) exchange rate, partly by large purchases of foreign currency but mainly through interest rate changes. They managed to do this for some twelve months, but interest rates oscillated continuously (though not as widely as in some other phases) and by the end of the phase the policy was in disarray, partly but not only because of political tensions between Chancellor and Prime Minister. It seems likely that a major reason for the growing difficulties was that the authorities had tried to stabilise the rate at too low a level, and it was the periodic upward pressures on the exchange rate (at a time of growing inflationary pressures) that caused the conflict between domestic and external indications for interest rates. Moreover, the undervaluation of late 1986/early 1987 was an obvious trigger for the inflationary boom that developed (see Chapter 4). But from the authorities’ point of view the upward pressure on the exchange rate was an unexpected and unwelcome development which eventually forced the abandonment of the DM-shadowing strategy. On the other hand real interest rates were firmly positive, at around 5%, and showed no significant downward trend, so it is at least not obvious a priori that interest rates were set too low in order to stabilise the exchange rate. 9.1.7 April 1988 to September 1990 In this phase the informal peg to the DM had been abandoned, and no other target had been put in its place. The nominal effective exchange rate appreciated by some 6% from its level around the turn of the year, and stayed near that level until May 1989; by the end of 1989 it had declined to the level of early 1987, but a ‘large’ rise in 1990 Q3 took it back to the level of late 1987/early 1988. The real exchange rate followed a similar pattern. Nominal interest rates were initially reduced to a low of 7.5%, then raised in small steps to 9.5% by the end of June 1988 and 12% by the end of August. There were further rises of 1% each in November 1988, May 1989 and October 1989. Real rates continued to average around 5%. Despite the abandonment of the informal DM-peg the authorities noted in May that ‘exchange rate stability remains desirable in its own right, with benefit to industry, and will continue to be sought so far as is possible without jeopardising the counter-inflationary thrust of policy’ (BEQB 1988b: 162), and the BEQB continued to show an interest in movements in competitiveness (e.g. 1988d: 480, 1989d: 489, 1990b: 170). The interest rate cuts of 0.5% in April and May were presented as a response to upward pressure on the exchange rate. ‘However, it became increasingly apparent that there was a need to tighten monetary policy so as to exert downward pressure on inflation. During the second half of [1988 Q2], underlying sentiment towards sterling weakened and interest rates abroad rose. Domestic interest rates were therefore increased in a series of 12 % steps without creating undue upward pressure on

Monetary Policy and the Exchange Rate 147 the exchange rate’ (BEQB, 1988c: 326).24 The rise of November 1988 was aimed primarily at domestic inflation, in the light of evidence that GDP had grown more rapidly than previously thought (BEQB, 1989a: 5–6), while the rises of May 1989 and October 1989 were also responding to further falls in sterling which were seen as likely to push up domestic inflation (BEQB, 1989c: 317, 1989d: 476). These falls were halted but not immediately reversed; however, the pound rose significantly in the summer of 1990, in a move attributed to speculation that the UK would enter the ERM and to oil-price effects resulting from the Iraqi invasion of Kuwait in August, but welcomed by the authorities for its counter-inflationary impact (BEQB, 1990c: 326–7, 1990d: 442, 465).25 Overall for most of this phase the authorities wanted to raise interest rates substantially to counter the inflationary boom that had developed, but they wanted to avoid a large sterling appreciation (and therefore raised rates in a series of small steps); in 1990 they may have favoured (limited) appreciation rather than a further rise in the already high interest rates. On balance, the nominal exchange rate was more stable than in most phases (though less so than in the DM-shadowing period, see Table 9.2), but there were a number of smaller movements and pressures that appeared to the authorities as exogenous. Foreign exchange market intervention over the phase was small in net terms, but there were periods of significant purchases (mainly 1988) and sales (mainly 1989) of foreign currency. With hindsight it can be argued that interest rates should have been raised sooner so that inflation would have risen by less, but in this phase at least they were raised rapidly and to high levels. Moreover, real interest rates remained positive at around 5%.26 9.1.8 October 1990 to September 1992 In this phase the UK was a member of the ERM, with sterling pegged to the other currencies with a ±6% band. The nominal effective exchange rate fluctuated within a narrow range for most of the period. The DM/£ rate varied between 2.90 and 2.98 for most of the period, dipped below 2.90 in December 1990, from December 1991 to March 1992, and again from July 1992, but fell much further when the UK left the ERM in September 1992 (to 2.51 by the end of that month). The real exchange rate fluctuated throughout within, but near the top of, the wide band. Nominal interest rates were cut on successive occasions up to May 1992; real rates continued to average around 5%. Interest rates were initially reduced from the peak of 15% on the day of the UK’s entry into the ERM in October 1990, by which time the domestic economy was in recession but inflation had not clearly peaked; the reduction was regarded at the time as premature (see Chapter 5, Section 5.1). A more sustained easing of policy began in February 1991, and took interest rates down in 0.5% steps to 11% by mid-July. The BEQB described the initial cuts as ‘made in the light of developments in the domestic monetary and economic situation and taking full account of the need to maintain market confidence in the authorities’ commitment to the counterinflationary 24 See also BEQB, 1988d: 485: ‘Domestic interest rates had been increased four times during June, on each occasion following a weakening of sterling, thus tightening monetary conditions without provoking an over-reaction in the foreign exchange market. At the beginning of July a continuation of this process was necessary . . . ’. According to Lawson (1992: 838–9), ‘The purpose of doing it by half point steps was to demonstrate that this was a tightening of monetary policy for its own sake, and not a macho move designed to impress the foreign exchange market in order to protect the pound.’ 25 The authorities may have actively wanted some rise in the exchange rate, to prevent downwards pressure on interest rates within the ERM compromising the counter-inflationary stance of policy (see Chapter 5). 26 They dropped to below 4% in the final months as inflation rose to its peak.

148 The Making of Monetary Policy in the UK, 1975–2000 stance of policy and to the ERM discipline’ (1991b: 203). It noted in August that ‘[w]hile the ERM represents a potential constraint on policy, its requirements have not diverged from what has been seen as appropriate for domestic objectives’ (1991c: 344–5). A further cut of 0.5% was made in September (in the context of a German rate rise and a US rate fall), and the BEQB noted that the narrowing of the UK–German interest differential ‘reduced the scope for a further fall in domestic interest rates. However, the substantial reduction in interest rates which had occurred since sterling joined the ERM had already provided a significant stimulus to demand’ (1991d: 478). The UK declined to follow the rise in German rates in mid-December 1991 (or the contemporary fall in US rates), but allowed sterling to make fuller use of its wide band in the ERM, while ‘[a]t no time during [1991 Q4] did the market seriously question the commitment to sterling’s ERM parity’ (BEQB, 1992a: 30). In May 1992 ‘the case on domestic grounds for a further reduction of interest rates was clearly established, and, as sterling maintained its post-election gains in the ERM’, interest rates were cut by a further 0.5% (BEQB, 1992c: 269). In mid-July when the Bundesbank raised its discount rate but not its Lombard rate this decision ‘was greeted with relief by markets and the United Kingdom did not need to follow the rise in rates in some other ERM countries. Fears soon emerged, however, that a rise in the Lombard rate might still be in prospect and sterling’s weakening suggested that an increase in UK rates might have to follow such a move’ (BEQB, 1992d: 388). The UK authorities allowed sterling to move towards the bottom of its wide band in the ERM, which produced ‘[s]ome slight easing of monetary conditions’ (BEQB, 1992d: 388). In late July and early August the government cut the interest rates on some National Savings products, apparently in order to avoid rises in building society deposit and mortgage rates; arguably this move made clear to the markets the authorities’ reluctance to raise UK rates and thereby undermined the credibility of the government’s commitment to the ERM.27 When the turbulence in the financial markets built up in the first half of September, the authorities delayed raising interest rates until 11 a.m. on Wednesday 16 September, when they were raised by 2%, while another 3% rise was announced at 2.15 p.m. for the following day. However, the second rise was never implemented and the 2% rise was reversed instead, after sterling had been suspended from the ERM. There was also a further 1% cut a week later as part of a ‘rebalancing’ of the policy mix (BEQB, 1992d: 394). Overall in this phase, policy appears to have been focused primarily on the domestic economy, with the exchange rate peg being regarded as a constraint rather than a target. The strategy worked well in the early stages, and there was an unprecedented reduction in the short-term interest rate differential between the UK and Germany. But it became more and more risky as time went on and that differential tended to zero. In the end, despite unequivocal statements of the government’s commitment to the ERM parity (notably by Prime Minister Major in a speech in Glasgow on 10 September), the markets lost confidence and sterling was unable to remain within the ERM. 9.1.9 October 1992 to April 1997 In the aftermath of Black Wednesday the authorities constructed a new monetary policy framework consisting of inflation targeting together with a revised relationship between the Treasury and the Bank of England. The exchange rate was initially volatile then settled at a rate roughly 10% below that of the ERM period, dipped further below in 1995 Q2, but began to appreciate 27 See

Chapter 5, Section 5.2, and Stephens (1996: 218, 221). Curiously, these cuts are not mentioned in the BEQB.

Monetary Policy and the Exchange Rate 149 strongly from 1996 Q4 and ended the phase back at the level of the ERM period. The initial fall in the nominal rate led to a major undervaluation which, apart from a short interlude in late 1994 and early 1995, disappeared only in 1996 Q3. Nominal interest rates continued to decline up to February 1994 and then moved within a narrow range; real rates were positive but lower than in the three preceding phases. There is little evidence in this phase of the authorities’ seeking to keep the exchange rate at any particular level, or of their seeking to preserve the significant gain in competitiveness which resulted from Black Wednesday. The initial interest rate cuts (of 3% between midOctober 1992 and the end of January 1993) were part of a ‘rebalancing’ of policy designed to speed up the recovery which some observers felt had been hampered by the ERM exchange rate commitment (given the context of German policy following reunification), but the authorities were concerned at the likely extent of the pass-through of the Black Wednesday depreciation into import prices and then retail prices more generally (BEQB, 1993a: 17–18, 26, 44; 1993b: 178, 189). Gradually, however, it began to seem that while the pass-through into import prices was of the order of 100%, that into the general price level was much smaller (BEQB, 1993c: 325, 326–7; 1993d: 439, 462; BEIR, 1994a: 30–1, 35), and this was part of the background to the interest rate cuts of 0.5% in November 1993 and 0.25% in February 1994.28 The rises of September 1994 to February 1995 were pre-emptive responses to evidence of domestic inflationary pressures. The depreciation of March–April 1995 (which made sterling undervalued again) clearly worried the Bank of England and led it to recommend a rise in interest rates in the Monetary Meetings of May, June and July 1995; in rejecting those recommendations Chancellor Clarke barely addressed the Bank’s argument, but he did claim that it was hard to judge whether it would feed through to retail price inflation given the weakness of economic activity (MMM, June 1995, §26). In the rest of 1995 the minutes show the Bank, but much less often the Chancellor, continuing to mention the exchange rate, but by January 1996 the Governor (who had agreed a 0.25% interest rate cut in December 1995 despite the weakness of sterling but was less happy about a further cut in January 1996) conceded that ‘it did now seem that the period of externally-originated cost-push pressure—aggravated by last year’s fall in the exchange rate—had largely passed, with encouragingly little damage in terms of the knock-on effect onto domestic inflation’ (MMM, January 1996, §14). The agreed cut of March 1996 was similarly focused mainly on domestic developments, but the June 1996 cut, opposed by the Bank, was justified by the Chancellor partly by reference to a limited appreciation of sterling (MMM, June 1996, §26). In every month from September 1996 to April 1997 the Bank of England argued for a rise in interest rates, but the Chancellor agreed to a rise (of 0.25%) only in October 1996. The Bank’s case, as set out in the August 1996 Inflation Report, was based on the acceleration of domestic demand reinforced by continuing rapid monetary growth and large fiscal deficits. As sterling now appreciated more strongly and consistently, the Bank seems to have devoted much effort to analysing the likely causes, and hence the likely consequences, and to developing the argument that appreciation would have only a short-term impact in reducing inflation without having much effect on the excess demand that would fuel inflation over the medium term.29 The Chancellor, however, seems 28 At the MMM of February 1994, the Chancellor, who was ‘concerned that advice was erring excessively on the side of caution’, used the recent small appreciation as one argument for a cut (which the Governor opposed) (§§19, 21). 29 For example, ‘In the short run, the rise in sterling will lead to a fall in inflation as import prices fall. But that is primarily a one-off impact on the domestic price level, rather than a continuing reduction in the underlying rate of inflation’ (BEIR, 1997a: 53). See also BEIR, 1996d: 16–17, 46; 1997b: 43–6, 51.

150 The Making of Monetary Policy in the UK, 1975–2000 not to have addressed the issue of timing and insisted that the Bank was not taking enough account of ‘the strength of the exchange rate. Most economic models suggested an exchange rate change of this magnitude would have a significant effect on the prospects for inflation. The strength of the pound would also have a marked effect on activity, where the full effect had not been felt yet’ (MMM, December 1996, §33). Overall in this phase, monetary policy was clearly focused on domestic variables, as would be expected under an inflation target, but the exchange rate remained an important issue in the discussions between Chancellor and Governor because of its possible, though not certain, effect on domestic inflation. After the depreciation of Black Wednesday (and the minor rebound which followed) the nominal exchange rate was relatively stable for several years, while the real exchange rate remained undervalued, but sterling began to rise strongly towards the end of 1996, transforming the undervaluation into overvaluation in less than a year. This appreciation was as unexpected and unwelcome for the authorities as the depreciation of Black Wednesday, but by now the focus of policy was on domestic inflation and output with the exchange rate seen as important only for its implications for these variables. 9.1.10 May 1997 to July 2000 In this phase interest rate decisions were taken by the newly created Monetary Policy Committee of the Bank of England, on the basis of an inflation target (of 2.5% for the RPIX) set by the new Chancellor. The nominal exchange rate continued for several quarters to appreciate strongly, except for a small dip in 1998, and reached a peak in May 2000 that was some 28% above the August 1996 level. Nominal interest rates moved within a relatively narrow range; real interest rates were marginally higher on average than in the preceding phase. The data for relative unit labour costs used in the rest of the chapter are available only to 1998 Q4. However, data for the ‘normalised’ relative unit labour costs series are available on a monthly basis up to July 2000. Figure 9.1 shows the normalised relative unit labour costs series as well as the raw series for 1991–2000 Q2 on a quarterly basis (quarterly data for the normalised series calculated from the published monthly data). It is clear that the two series for relative unit labour costs, raw and normalised, move very closely over the years 1990–98; this suggests that no serious distortions are introduced by focusing on the normalised series for the next six quarters in 1999–2000. On this basis the real exchange rate ceased to be undervalued in 1996 Q3, first became seriously overvalued in 1997 Q2 and remained so for the rest of the phase, appreciating further up to 1998 Q1, falling back slightly in 1998 Q4 but then appreciating monotonically for the rest of the phase to reach an apparently unprecedented level.30 The change of Government in May 1997 brought agreement to an interest rate rise by the new Chancellor at the May meeting, at which he also announced that in future the Bank would be assigned responsibility for using interest rates to pursue the government’s inflation target. The Monetary Policy Committee continued to raise interest rates in June, July and August. It was well aware of the continuing appreciation of sterling, which it saw as creating a dilemma for policy: at its August meeting the MPC agreed that sterling was probably overvalued, analysed the policy dilemma in terms of both the differential impact on activity in different parts of 30 The 2000 Q2 level of 154.6 contrasts with the previous peak for the normalised series of 139.6 in 1981 Q1 (when the peak for the raw series was 142.0).

Monetary Policy and the Exchange Rate 151 180

150

120

90

rulc

1

20

00

Q

3

1 19

99

Q

3

Q 99

19

19

98

Q

1

3 19

98

Q

1

Q 97

19

19

97

Q

3

1

Q

19

96

Q

3 96 19

19

95

Q

1

3

Q 95

19

19

94

Q

1

3 19

94

Q

1

Q 93

19

19

93

Q

3

1

Q

19

92

Q

3 19

92

Q 91

19

19

91

Q

1

60

rulc normalised

Figure 9.1 Relative unit labour costs, 1991–2000 (Source: International Financial Statistics)

the economy and the problem for policy, and considered (but dismissed) alternative policy instruments that might help to resolve the dilemma (MPC, August 1997, §§58–64).31 Interest rates were raised again in November 1997 and in June 1998, in both cases largely on grounds of the strength of domestic demand, though the MPC also noted that the downward impact of sterling’s appreciation on net exports (and also that on import prices) was taking much longer to come through than expected (e.g. BEIR, 1997d: 26, 29, 36). In October 1998 the MPC started lowering interest rates as evidence accumulated of domestic and world economic slowdown. By now it was the depreciation of sterling that was difficult to explain in full (BEIR, 1998d: 9–10), but the possibility of a larger depreciation in response to an interest rate cut was given as one argument against cutting rates in the September MPC meeting (MPC, September 1998, §36). The initial cuts occurred at a time when sterling was weaker, but the cuts continued through to the first half of 1999 when sterling was rising; in March the MPC thought ‘it was too soon to judge whether the rise would persist . . . In any case, it would not be sensible for policy to react to high frequency movements in the exchange rate, as this could lead to a volatile path for interest rates from month to month, and might make it more difficult for others to understand the motives for interest rate changes’ (MPC, March 1999, §28).32 Interest rates were then raised again in four steps of 0.25% each between September 1999 and February 2000, while sterling rose further (mainly against the euro) to the peak of May 2000. These rises were again undertaken primarily for domestic reasons (the likely impact on later inflation of the strengthening of domestic demand), but the movement and level of the exchange rate continued to preoccupy the MPC. In particular the committee discussed and eventually changed the exchange rate assumption underlying its inflation forecast (e.g. MPC, 31 See also BEIR, 1997b: 51; 1997c: 49. 32 On the latter point see also MPC, May

1999, §31.

preference for stability

no (focus on domestic objectives)

implicit preference initially for depreciation to eliminate overvaluation, later for stability (?) preference initially for stabilisation, then for controlled depreciation (in response to oil-price fall) strong preference for stability (DM-shadowing) some preference for stability, but priority to inflation yes (commitment to ERM)

Apr 77–May 79

Jun 79–Mar 81

Apr 81–Mar 85

only indirectly (focus on domestic objectives) May 97–July 00 no (focus on domestic objectives)

Oct 92–Apr 97

Oct 90–Sept 92

Apr 88–Sep 90

Mar 87–Mar 88

Apr 85–Feb 87

some preference for limited depreciation at most

Jan 75–Mar 77

Phase (1)

Did authorities have clear preferences on nominal exchange rate? (2)

Table 9.3 Summary of Section 9.1

yes on balance (but depreciation went too far)

(appreciation increasingly unwelcome) yes on balance

mostly yes, but appreciating by end of phase

no, it depreciated for much of the phase

Did exchange rate move in line with their preferences? (4)

strongly, leads to interest yes (until the end of rate oscillations the phase, only) they strongly affected yes on balance the timing of interest rate changes less strongly than more or less until the might have been suspension from expected the ERM on Black Wednesday weakly (but underlay (later appreciation some of disputes) unwelcome) no, but recurring (appreciation concern re effects unwelcome) of appreciation

strongly, leads to interest rates (and exchange rate) varying widely strongly

(no)

strongly, leads to interest rates varying widely

weakly, until height of £ crisis 1976

How did such preferences affect interest rate policy? (3)

possibly, but serious side-effects: interest and exchange rate instability possibly, but considerable instability

no (policy had already changed by the end of the phase) no, worsening conflict between monetary and exchange rate targets no, cost (recession) already too high

Was policy sustainable beyond this phase? (6)

partly

sharp changes yes, smaller movements no

yes

(there was little change)

yes

(not very smoothly)

(smoothly)

possibly, but overvaluation becoming undesirable possibly, but overvaluation clearly undesirable

yes, for upward pressure from late 1996 yes for recurring upward pressure

no (abandoned by end (yes, for upward pressures of phase) on sterling) (not very smoothly) possibly, but later stability yes, for minor depended partly on ERM movements and entry expectations pressures smoothly until no (abandoned by end yes, for strong Black Wednesday of phase) downward pressure in September 1992

(smoothly)

not smoothly

very non-smoothly

(not smoothly)

(fairly smoothly)

not smoothly

How smoothly did it move? (5)

Did exchange rate changes appear to authorities as unexpected shocks? (7)

Monetary Policy and the Exchange Rate 153 August 1999, §14; BEIR, 1999d: 47–9); it considered whether the equilibrium level of sterling might have risen, and why (MPC, February 2000, §§3–5); and it continued to be concerned about the sectoral imbalance created by the overvaluation (e.g. MPC, February 2000, §§33–5). Moreover, one member of the MPC explored the possibility that the MPC could respond to asset price (especially exchange rate) misalignments in a way that would reduce the amplitude of exchange rate fluctuations without compromising the objective of price stability (Wadhwani, 2000).33 Overall in this phase, policy-makers were well aware of exchange rate movements but took decisions which were based essentially on domestic rather than external factors, even to the extent of allowing severe pressures to fall disproportionately on the traded goods sector as the result of a significant appreciation. Policy-makers had not sought this appreciation, but were unsure of its causes and felt that under inflation targeting they could do nothing to prevent or eliminate it. 9.1.11 Summary The main points to come out of the preceding sections are brought together in Table 9.3. Column (2) summarises the findings on the authorities’ preferences with respect to the exchange rate: on this analysis they had specific preferences of some kind in seven out of the ten phases, but not in June 1979 – March 1981 or in the final two phases (from October 1992 onwards). Column (3) summarises the effect of these preferences on interest rate policy: in the seven phases where they had such preferences interest rate policy was affected by these preferences, strongly in five cases but less so in the ERM period and less again in 1975–76. Column (4) summarises the movement of the exchange rate: in six of these seven phases (the exception is 1975–76) the exchange rate moved more or less in line with the authorities’ preferences for most of the phase, though in four its movements were not smooth (column (5)). However, it is important to consider the extent to which the policy applied was sustainable beyond the end of the phase (the phases are defined in terms of the monetary policy regime in which the exchange rate is an important element). Column (6) shows that in four cases the policy was not sustainable beyond the end of the phase (or had already been abandoned), while in another two there were considerable undesirable side-effects. Finally, column (7) summarises the findings on how the actual changes in exchange rates were perceived by the policy-makers: in nearly all phases there were some movements of the exchange rate, or tangible pressures on it in one direction or another, which presented themselves to the authorities as unexpected and unwelcome developments, and more often than not as discrete jumps rather than smooth adjustments. The following sections make direct comparisons between comparable developments occurring at different times, as an alternative way of analysing the historical experience, and in order to identify the regularity of the relationships involved.

9.2 THE MAJOR MISALIGNMENTS The two major overvaluations of the early 1980s and the late 1990s both followed periods of sustained undervaluation, with the undervaluation turning rapidly into its opposite, and both occurred at times when monetary policy was both disinflationary and strongly focused on 33 See

also Wadhwani (1999) and Cecchetti et al. (2000).

154 The Making of Monetary Policy in the UK, 1975–2000 domestic variables—in the earlier case at the height of official enthusiasm for monetary targets, in the latter case under an increasingly rigorous concern with inflation targets (plus instrument independence). In the earlier case there is an obvious potential exogenous cause, in the form of the effects of the second oil-price shock on an oil-producing economy, while the appreciation can also be regarded as an automatic transmission-mechanism adjustment to the tightening of monetary policy. Attempts to discriminate between the oil-price shock and monetary policy as causes of the appreciation of sterling have generally attributed a large role to monetary policy, including an element of overshooting.34 In the late 1990s, on the other hand, the only potential exogenous factor was the unexpected weakness of the euro; and monetary policy was not seriously tightened until well after the appreciation (and overvaluation) had begun. The earlier overvaluation reached a peak in its fourth quarter and then began to diminish. It disappeared (on the criterion used here) after another seven quarters, largely as the result of nominal depreciation. As shown above this was the partly intended by-product of a relaxation of monetary policy mainly driven by domestic concerns, which was made possible in part by the fiscal tightening of the 1981 Budget. That relaxation was in turn a response to the very severe recession of 1980–81 to which the overvaluation made an important contribution. The overvaluation of the late 1990s, by contrast, reached a temporary peak in its fourth quarter (1998 Q1), but then (on the basis of Figure 9.1) rose again to a higher peak sometime in 2000.35 It occurred in the context of a much higher level of economic activity, so that while it contributed to a sectoral imbalance (which was clearly of concern to policy-makers) it did not threaten overall income, and perhaps for this reason there had been no attempt even by the end of 2001 to bring the overvaluation to an end through depreciation.36 It should be noted, however, that the overall period covers not two but three episodes in which there were large and sustained rises in interest rates designed to reduce inflation (rather than to prevent or reverse a slide in the exchange rate): June to November 1979 (a total rise in the policy rate of 5%); June 1988 to October 1989 (7.5%, or 6% from the February 1988 level); and May 1997 to June 1998 (1.5%). Overvaluation occurred only in the first and third of these cases. In the second, despite the fact that the rise in interest rates was much larger (while inflation had not risen as much as in the first case), sterling was weak (and the last two 1% rises in interest rates of May and October 1989 were designed in part to counter that weakness). One very obvious difference between the second and first cases is that, in the second, interest rates were raised in small steps, and particularly at times when the exchange rate was weak;37 another is that in the second case the monetary authorities were known to prefer that the exchange rate should remain stable. The period also contains one very large and long-lasting undervaluation in the second half of the 1970s, a few quarters of undervaluation in the 1980s, and the more significant undervaluation which followed the depreciation of Black Wednesday. That of the 1970s was perhaps sui generis. It arose initially from the highly expansionary monetary policies followed by the Heath government in 1972–73, exacerbated by the first oil-price shock and the very gradualist 34 See, for example, Niehans (1981), Spencer (1986), Bean (1987) and Britton (1991). 35 The data used here show the overvaluation continuing up to 2000 Q2, and it seems highly

unlikely (on the basis of movements in the nominal effective rate and relative inflation rates) that the position changed significantly in the rest of 2000 (or 2001). 36 The MPC was naturally concerned about the inflationary consequences of any depreciation, but the evidence from the first half of the 1990s was that the pass-through could be much smaller than had previously been thought. 37 The rises in the third case were also in small steps, here of 0.25%, but the average levels of interest and inflation rates were much lower.

Monetary Policy and the Exchange Rate 155 response to it by the succeeding Labour government, which led to recurring depreciation and delayed disinflation. The marginal undervaluation of 1985 Q1 was the temporary result of the January 1985 sterling crisis which, although it came towards the end of a long trend decline in interest rates and depreciation of the exchange rate, was heavily influenced by the international environment (the strong dollar); sterling rebounded strongly, nearly becoming overvalued two quarters later. The undervaluation of 1986 Q4 and 1987 Q1 occurred when an intended depreciation (in response to the fall in oil prices) got out of hand; the real exchange rate appreciated back into the wide band but remained below the central range for another two quarters. The undervaluation of the mid-1990s, which lasted for nearly four years (if the short interlude in 1994 Q4 and 1995 Q1 is ignored), was the consequence of the UK’s exit from the ERM on Black Wednesday. It occurred when the economy was only just beginning to recover from recession, and this must have been an important factor in the low pass-through to the general price level. By contrast there seems little doubt that the repeated depreciations of 1975–76 (when the boom had not yet been brought under control) contributed to high inflation, and it has been argued in Chapter 4 that the 1986–87 depreciation (at a time of above average growth but excess capacity) was the trigger for the Lawson boom.

9.3 LARGE EXCHANGE RATE CHANGES The discussion of Section 9.1 makes it possible to return to the large exchange rate changes identified in the previous chapter and to consider the additional information in Table 9.4. Columns (3) and (4) present judgements on whether each exchange rate change was the ‘desired’ and ‘intended’ result of a deliberate act of policy by the monetary authorities, where the desire and intention are assessed on the basis of the official sources. Desire here relates purely to preference, but intention requires that the authorities should have taken some specific policy action designed to bring about the exchange rate change. This evidence suggests that, in their statements at least, the authorities did not favour many changes; indeed there are only four equivocal ‘yes?’ entries in column (3) and no unequivocal ‘yes’ entries, but eight outright ‘no’ and three ‘no?’ or ‘no??’ entries. As for their intentions, there are only two ‘yes?’ entries but 13 unequivocal ‘no’ entries in column (4). In that sense, at least, the majority of these large exchange rate changes must have been unexpected shocks to the monetary authorities. Columns (5) and (6) offer assessments of whether in these large changes the exchange rate was performing a useful role as a shock absorber or as a transmission mechanism. These assessments rely on a wide background of material, not all of it covered in Section 9.1, and need to be discussed. The first three large changes, the depreciations of 1975–76, were part of a spiral of depreciation and inflation in which the ultimate cause is not obvious: they can be seen as helping to absorb the shock of relatively high UK inflation but also as contributing to that higher inflation, and they were not clearly acting as a transmission mechanism since monetary growth was both well below the 1972–73 rate and to a large extent accommodating rather than active. The 1978 depreciation represented the unwinding of the appreciation which followed the uncapping of sterling in late 1977. The 1979 appreciations can be seen as a response to the tightening of monetary policy by the new government, i.e. as part of a transmission mechanism. The 1980 appreciations can also be regarded partly as a response to tight monetary policy (though interest rates were reduced

156 The Making of Monetary Policy in the UK, 1975–2000 Table 9.4 Large exchange rate changes

Quarter (1) 1975 Q2 + Q3 1976 Q2 1976 Q4 1978 Q2 1979 Q2 1979 Q3 1980 Q1 1980 Q4 1981 Q3 1983 Q1 1983 Q2 1984 Q4 + 85 Q1 1985 Q2 1986 Q1 1986 Q3 1986 Q4 1987 Q2 1989 Q4 1990 Q3 1992 Q4 1996 Q4 1997 Q1

Nominal effective Was e Was e exchange rate desired by ‘intended’ by change (e) (%) authorities? authorities? (2) (3) (4) −6.29 −7.82 −7.93 −5.37 +5.10 +4.52 +4.9 +5.04 −6.78 −9.28 +5.33 −6.50 +8.65 −6.70 −6.15 −5.83 +4.49 −4.34 +6.28 −11.80 +6.87 +6.08

? no no no? ? ? no no no no?? ? no yes? yes? yes? no ? no? yes? no ? ?

no? no no no no? no? no no no no no? no no? yes? no? no no no? yes? no no no

Did e perform a useful role as: shock absorber? (5) ? ? ? no no no yes no no no no no no yes no no no no yes? yes? no no

transmission mechanism? (6) no no no no yes yes yes no no no no no no no no no no no no no no no

Sources: column (2) from International Financial Statistics; columns (3) and (4) compiled from Section 9.1 and underlying official sources; for columns (5) and (6) see discussion in text.

in Q3 and Q4), but they could also be seen as absorbing the second oil-price shock. Most observers would agree that, by 1980 Q4 at least, there was also a strong element of overshoot. The 1981, 1983 and 1984–85 depreciations represent the effect of policy easing that went too far, while the 1983 and 1985 appreciations may have been desirable rebounds; but in none of these cases was the exchange rate usefully absorbing shocks or transmitting monetary policy. The first 1986 depreciation can be seen as a useful shock-absorbing response to the oil-price fall, but the second and third reflect that process going beyond what was required or desired. The 1987 appreciation was a useful rebound. The 1989 depreciation was an undesired development which the authorities resisted. The 1990 appreciation may have been absorbing the oil-price shock that followed the Iraqi invasion of Kuwait, but it also reflected growing expectations of UK entry into the ERM. The 1992 depreciation can be seen as absorbing the shock of German reunification, but there was a strong domestic policy element involved too. Finally the 1996 and 1997 appreciations have no obvious interpretation as shock absorbers or transmission mechanisms. Thus most large exchange rate changes were undesired and unintended, and were in that sense shocks to the authorities; and in most large changes the exchange rate was not performing a useful shock-absorbing or policy-transmitting role. To a large extent, of course, both these points simply reflect that fact that the determination of exchange rates is poorly understood

Monetary Policy and the Exchange Rate 157 (by economists or policy-makers). But they also emphasise that it is a mistake to assume that the exchange rate is likely to move in a benign way.

9.4 STERLING CRISES For present purposes sterling crises can be identified on the basis of sharp downward movements in the exchange rate associated with heavy sales of foreign exchange and followed by rises in interest rates introduced under pressure and designed to limit and/or reverse the depreciation. Such crises occurred during this period in March–November 1976, September–October 1981, November 1982 – January 1983, January 1985, July–October 1986, and September 1992. Of these, the most serious were probably those of 1976 and 1992, followed by 1985, and then by the others. Table 9.5 shows how the UK policy rate changed, both absolutely and relative to US and German policy rates, in the months preceding the crises. It is clear that there is a general tendency for crises to have been preceded by absolute and relative interest rate cuts, but January 1985 (when the sterling crisis arose on the back of a peaking dollar in a highly turbulent international environment) was an exception, and even apart from January 1985 the pattern is far from regular. There have also been other cases when absolute and relative interest rate cuts have not led to crisis, notably in the summer of 1977 and in the first half of 1988. Table 9.6 shows how interest and exchange rates changed over the crises. In most cases the exchange rate rebounded to some extent from the crisis level,38 but even apart from the exception of January 1985, when the exchange rate rose sharply beyond the pre-crisis level, the pattern is far from regular. Thus sterling crises occurred in the wake of quite different patterns of interest rate development, and there was considerable variation in the way the exchange rate responded to interest rate changes within sterling crises. What this meant for the policy-maker was that it was difficult to predict the effects on the exchange rate of given interest rate changes in either direction.39 In the words of the MPC towards the end of the period, ‘as always, the effect on the exchange rate of a possible reduction in interest rates was uncertain’ (MPC June 1999, §7).

9.5 CONCLUSIONS The aim of Chapters 8 and 9 was to provide evidence on the extent to which the possibility for the exchange rate to vary has been useful or unhelpful for UK monetary policy; in particular, has the exchange rate functioned as a useful automatic equilibrating mechanism, has it been a useful policy instrument, or has it been a source of extraneous shocks? The conclusions to emerge from the analysis are surprisingly clear. First, the exchange rate has not generally functioned as a useful automatic equilibrating mechanism. Large exchange rate changes have typically not moved the real rate back towards long-run equilibrium (Chapter 8, Section 8.3). While the real exchange rate has probably not followed a random 38 The exchange rate data are for monthly averages, so the crisis level is the average for the month containing the crisis rather than the lowest point reached on daily data. 39 The unpredictability of interest rate – exchange rate relationships may also be one reason why the authorities, who frequently referred in the 1980s to ‘monetary conditions’ in the sense of both interest and exchange rates, never introduced a monetary conditions index of the kind used by some other open economy central banks.

−2.25 0 −2.5 0 −1.5 0

−1.5 −2 −3.5 +1.25 −1.5 −0.5

−1 0 −1 +0.875 0 0

1 month −1 −3 −1 +2.25 −0.5 0

6 months −1.75 0 −0.5 +1 −1 −0.5

3 months

−1 0 −0.5 +1.375 0 0

1 month

Change in UK–US interest rate differential

−1 −2 −1 +1.25 −1 −1.25

6 months

−2.25 0 −0.5 0 −1.5 −0.75

3 months

−1 0 −0.5 +0.875 0 0

1 month

Change in UK–German interest rate differential

Sources: International Financial Statistics, BEQB. Note: policy rates are central bank discount rates (end-month); the data refer to changes between the month preceding the crisis (i.e. the month before the first month given in the first column) and six/three/one month(s) before.

March–November 1976 September–October 1981 November 1982 – January 1983 January 1985 July-October 1986 September 1992

3 months

6 months

Change in UK policy rate over preceding:

Table 9.5 Interest rates in the months before sterling crises

Monetary Policy and the Exchange Rate 159 Table 9.6 Interest and exchange rates over and after sterling crises

March–November 1976 September–October 1981 November 1982 – January 1983 January 1985 July–October 1986 September 1992

Months for exchange rate to return to pre-crisis level

peak month

3 months after

6 months after

Months for policy rate to return to pre-crisis level

+6

−18.6

−15.0

−15.1

4

45

+4

−4.2

−1.2

−0.8

9

9

+2

−11.0

−9.7

−6.5

5

(never)

+41/2 +1 +2 (+3)

−3.1 −11.8 −3.7

+5.0 −11.3 −12.3

+11.6 −6.4 −14.0

(27) 5 0

2 13 57

Change in policy rate over crisis

Change in exchange rate from pre-crisis level to:

Sources: International Financial Statistics, BEQB. Note: the data for the change in the exchange rate refer to changes between the month preceding the crisis and the peak month (the single month or the second of the two months identified in the first column) and three/six months after, and the last two columns refer to the number of months from the peak month.

walk and has not diverged without limit from the central range, policy has played an important part in this, notably in the return to the central range in the early 1980s (Chapter 9, Section 9.1.4). There are few cases when the movement of the exchange rate could possibly be seen as a useful shock absorber or policy transmission mechanism (Chapter 9, Section 9.3). While it could be argued that some appreciation was appropriate in 1979–80, for example, in the light of the second oil-price shock and North Sea oil, the rise in sterling at this time went well beyond what was required; and the same can be said for the depreciation following the oil-price fall in late 1985/1986. At other times the exchange rate has moved ‘automatically’ (in the sense that policy had not intended or, in any obvious sense, caused the change) but in non-useful ways, for example in the various sterling crises or in the appreciation of the late 1990s. And at other times the exchange rate has not moved when it would have been useful, for example in eliminating more rapidly the overvaluations of the early 1980s and the late 1990s. Second, the exchange rate has not functioned as a useful policy instrument (Chapter 9, Section 9.1). The UK authorities have, in seven out of the ten phases considered, had preferences of some kind regarding the level of the exchange rate, ranging from a weak preference for no change through preference for controlled depreciation or (less often) appreciation to stronger preferences for stability. But their ability to make the exchange rate move (or not move) in the way desired has been limited; fine tuning, particularly in the form of attempts to stabilise the exchange rate, has sometimes been possible for up to two or three years, but has typically failed badly after such a period. And the authorities have had to put up with long periods in which the exchange rate was significantly different from their preferred level, so that even coarse tuning has not always been viable. Third, the evidence suggests that in nearly every phase there were movements of the exchange rate, or pressures on it, which for the monetary authorities were unexpected and unwelcome. There was plenty of short-term ‘noise’ of this sort, documented in the BEQB for the earlier years in particular (but largely ignored in Section 9.1 of this chapter precisely because it was short-term noise), but there were also a number of large movements of the exchange rate

160 The Making of Monetary Policy in the UK, 1975–2000 which fell into this category, notably those which led to the two major overvaluations and those which led to sterling crises. While for the 1970s it would be possible to argue that the monetary authorities were operating with a poor understanding of some of the key relationships involved and that it was this which made exchange rate movements unexpected, it would be much more difficult to argue this for the 1980s or, particularly, the 1990s. In that case it can be concluded that the exchange rate has indeed often been a source of extraneous shocks.40 Finally, it should be stressed that the main problem has not been short-term exchange rate volatility (much of which has been ignored here through the emphasis on quarterly and monthly averages), but longer term misalignments which (for a given degree of pressure on domestic demand and inflation) create sectoral imbalances that distort the structure of the economy and may via hysteresis effects lower its long-run growth performance. However, the evidence of the two great overvaluations is consistent with the proposition that ignoring medium-term movements of the exchange rate (and giving no ‘steer’ to the foreign exchange market) is dangerous because it is likely to lead to long-term misalignment. 40 This finding is complementary to the statistical evidence of Flood and Rose (1999), which suggests that, for the same level of underlying macroeconomic volatility, allowing the exchange rate to float produces an increase in overall volatility. Flood and Rose explain their findings on the basis of nonlinearities associated with a time-varying (and regime-varying) risk premium.

10 Interest Rate Smoothing In the last few years a number of papers have explored the previously unrecognised phenomenon of ‘interest rate smoothing’ by the monetary authorities. This chapter examines smoothing for the UK, with particular reference to direct evidence on the intentions and motivations of the authorities. Section 10.1 summarises the international evidence on smoothing and locates the UK with respect to other countries. Section 10.2 investigates, and largely confirms, Goodhart’s (1999a: 112) suggestion that the frequency of large interest rate reversals in the UK was mainly the result of changes in monetary policy regime and of belated corrections of recognised policy mistakes; the analysis demonstrates the importance in this connection of official concerns with the exchange rate. Section 10.3 reviews the literature on the reasons for smoothing. Section 10.4 presents data for the UK disaggregated into subperiods, by change and by quarter, and outlines the kinds of direct evidence on smoothing that can be found in the official sources, especially for recent years. Section 10.5 uses the disaggregated data and direct evidence to try to evaluate the relevance of the different explanations for interest rate smoothing, focusing on the 1997– 2000 phase in which policy was smoothest and for which much more information is available. Section 10.6 concludes.

10.1 INTERNATIONAL EVIDENCE Interest rate ‘smoothing’ refers to the tendency for monetary authorities to adjust official interest rates mainly in sequences of small steps in the same direction.1 The phenomenon is now widely accepted as a characteristic of monetary policy in most industrial countries. Detailed evidence on interest rate smoothing was first provided for the US by Rudebusch (1995), who examined the patterns of change over 1974–79 and 1984–92. He characterised the Federal Reserve’s interest rate targeting behaviour in the short term as involving a series of gradual increases or decreases in the target rate over the course of several weeks, with infrequent reversals of the direction of change.2 Goodhart (1997) provided a broadly comparable analysis of interest rate changes in Australia, Germany, Japan and the UK as well as the US. He selected the ratio of reversals (changes in direction where an increase follows a decrease, −+, or a decrease follows an increase, +−) to total changes as the best single summary statistic of smoothing. For the period 1974–79 plus 1984–92 used by Rudebusch, Goodhart calculated this ratio for the policy interest rates as varying from 1:10.84 for Australia through 1:9.05 (Federal funds target rate) or 1:7.40 (discount rate) for the US, 1:6.90 for Germany and 1:6.25 1 The term smoothing is also used to refer, e.g. in Clarida, Gal´ı and Gertler (1998) or in Sack and Wieland (1999), to the finding of large and significant coefficients on the lagged dependent variable in Taylor-rule reaction functions, but the focus here will be on the nature of actual movements in policy interest rates. 2 By contrast he identified large deviations of the spot rate from the target on a daily, transitory, basis, and mediumterm persistence in the target rate.

162 The Making of Monetary Policy in the UK, 1975–2000 for Japan, to 1:3.88 for the UK.3 Lowe and Ellis (1997) looked at these same countries from 1985 to 1997, with an emphasis on the ratio of continuations (cases of an increase following an increase, ++, or a decrease following a decrease, −−) to reversals, which varied from 8.5 for Japan and 8.3 for Germany through 5.3 for the US to 3.3 for Australia and 3.2 for the UK.4 The Bank for International Settlements’ 1997–98 Annual Report (quoted in Goodhart, 1999a) gives summary data for 12 countries (including four of the five above) over varying periods. The implied ratio of continuations to reversals ranges from 31.5 for Austria, 10.3 for Canada and 9.3 for the US, to 3.2 for the UK and 2.8 for each of Germany5 and Italy.6 Emphasis has also been placed on the ‘average duration’, that is the number of days since the previous interest rate change, and the average size of changes. For countries with higher ratios of continuations to reversals the duration of reversals tends to be higher than that of continuations. On the other hand for these countries the average size of changes does not seem to vary consistently as between reversals and continuations, or as between increases and decreases. While the general phenomenon of interest rate smoothing seems universal, the UK appears to differ from other countries in two respects. First, as already noted, the UK has had a higher frequency of reversals than many other countries, notably the US. Second, in the UK the average duration of changes tends to be higher for increases than for decreases (rather than as between reversals and continuations), and the average size of changes tends also to be higher for increases than for decreases. Goodhart (1999a: 112) speculates that ‘these really large reversals in the UK were mainly due to regime changes and recognition of prior policy errors’,7 in which case ‘we are left, as in the US case, with a problem of accounting for a general, apparent reluctance to reverse the direction of change’.

10.2 LARGE CHANGES AND REVERSALS IN THE UK This section considers the large changes and the reversals that occurred in the UK over the period. Table 10.1 gives details of the 13 policy rate changes equal to or greater than 2%. These consist of 10 increases and 3 decreases, and of 5 reversals and 8 continuations; the proportions of increases and reversals are notably higher than for all changes in the period, where the proportion of increases is 32% and that of reversals is 20% (see Table 10.2). The durations of the positive continuations and reversals are in line with those for the whole sample, but the durations of the 2 negative continuations are higher, and that for the single negative reversal is only one day. Table 10.1 also contains columns giving the context of the interest rate changes and an assessment of whether each one was the result of a regime change (RC) or a recognition of prior policy error (RPPE). There are 3 cases which are neither RC nor RPPE, 4 cases of RC and 6 of RPPE. Four of the latter are cases where sterling had begun to slide, typically after a period of interest rate cuts, and the authorities felt they had no alternative but to raise interest rates sharply to stabilise the exchange rate; two are cases where the high interest rates and large 3 Goodhart also reported ratios for longer and more inclusive periods, which were somewhat lower for the US and Australia, but higher for Germany, Japan and the UK, and for some money market rates. 4 They also report ratios for 1992–97, in which Australia, the US and the UK are closer to each other. 5 Goodhart (1997: 124) notes that the German money market rate ‘vibrated gently’ between 1983 and 1987, which made for a higher frequency of reversals (albeit very minor ones). 6 The corresponding ratios for the other countries are: Sweden 12.7, Australia 9.5, France 8.5, Belgium 6.1, Spain 4.1 and the Netherlands 3.0. For the UK the period is 1978–98 and the rate is that for Band 1 bank bills. 7 This comment was in connection with data collected on a quarter-by-quarter basis, as presented in Section 10.4 below, but the point is also valid for data in terms of changes.

Interest Rate Smoothing 163 Table 10.1 Very large policy rate changes (i ≥ 2%) Date 7.10.76 28.11.77 9.11.78 13.6.79 15.11.79 25.11.80 11.3.81 16.9.81 1.10.81 12.7.84 28.1.85 16.9.92 17.9.92

Sign/size of i

New level

Nature and duration of i

+2 +2 +2.5 +2 +3 −2 −2 +2 +2 +2 +2 +2 −2

15 7 12.5 14 17 14 12 14 16 12 14 12 10

++18 −+30 ++110 −+49 ++111 −−103 −−81 −+135 ++11 ++3 ++10 −+86 +−1

Context of i

RC or RPPE

final stage of £ crisis follows uncapping of £ monetary overshoot new government monetary overshoot, £ weak £ strong, domestic squeeze £ strong, domestic squeeze £ weak £ weak £ weak £ crisis ERM crisis undoing crisis measures

RC RC — RC — RPPE RPPE RPPE RPPE RPPE RPPE — RC

Sources: Tables 8A.1 and 8.1, and the underlying official sources. Notes: RC = regime change; RPPE = recognition of prior policy error.

appreciation of 1979–80 had created intense deflationary pressure on the economy, notably on the traded goods sector. Table 10.3 brings together the 30 policy rate changes which were equal to or greater than 1% but less than 2%. Here there are 18 increases and 10 reversals, proportions which are again above those for the whole sample (Table 10.2). The durations of the positive continuations and, to a lesser extent, the positive reversals are above those for the whole sample, those for the negative continuations are a bit below, and those for the negative reversals are much higher (but they are relatively few). As for regime changes and recognition of prior policy errors, Table 10.3 classifies 20 cases out of 30 as neither but 10 as RPPE. The latter include nearly all of the changes greater than 1%; nine are cases where sterling had weakened and the authorities felt they had to respond by raising interest rates, the other constitutes the first interest rate response to the domestic squeeze and misalignment of 1980–81. Table 10.4 lists the reversals (of all sizes), and provides a double classification of them. The first distinction is between sustained, temporary and very temporary reversals, where ‘very temporary’ are reversals such that the policy rate returns to its previous level within 6 months; ‘temporary’ are those where it returns to its previous level within 12 months; and ‘sustained’ are those where the policy rate returns to its previous level, if at all, only after at least 12 months. On this basis there are 20 sustained reversals, 6 temporary and 9 very temporary. Table 10.2 Different types of policy rate change Nature of change (% of total)

i ≥ 2% 1% ≤ i < 2% All i Source: Table 8A.1.

Average duration (days)

++

+−

−+

−−

++

+−

−+

−−

46.1 40.0 22.2

7.7 13.3 9.7

30.8 20.0 10.2

15.4 26.7 58.0

43.8 51.0 41.7

1.0 123.0 70.4

58.5 66.7 64.6

92 18.3 25.3

164 The Making of Monetary Policy in the UK, 1975–2000 Table 10.3 Large policy rate changes (1% ≤ i < 2%) Date 28.7.75 6.10.75 26.4.76 24.5.76 13.9.76 31.1.77 10.3.77 31.3.77 12.4.78 8.5.78 8.6.78 7.2.79 1.3.79 5.4.79 3.7.80 29.11.82 14.1.85 9.1.86 19.3.86 14.10.86 6.8.87 25.8.88 25.11.88 24.5.89 5.10.89 8.10.90 22.9.92 16.10.92 13.11.92 26.1.93

Size/sign of i

New level

Nature and duration of i

+1 +1 +1.5 +1 +1.5 −1 −1 −1 +1 +1.25 +1 +1.5 −1 −1 −1 +1.125 +1.5 +1 −1 +1 +1 +1 +1 +1 +1 −1 −1 −1 −1 −1

11 12 10.5 11.5 13 12.25 11 9.5 7.5 8.75 10 14 13 12 16 10.125 12 12.5 11.5 12.5 10 12 13 14 15 14 9 8 7 6

++60 ++50 −+35 ++20 ++80 −−5 −−25 −−7 −+68 ++23 ++18 ++60 +−16 −−20 +−165 −+18 ++1 −+114 +−49 −+102 −+63 ++13 ++66 ++127 ++94 +−262 −−3 −−18 −−20 −−48

Context of i

RC or RPPE

rising US interest rates rising US interest rates £ weak £ weak £ weak unwinding of 1976 £ crisis unwinding of 1976 £ crisis unwinding of 1976 £ crisis monetary overshoot, £ weak monetary overshoot, £ weak monetary overshoot, £ weak domestic expansion £ strong, pre-election £ strong, pre-election £ strong, domestic squeeze £ weak £ weak ($ strong) oil-price fall oil-price fall £ weak DM-shadowing inflationary boom inflationary boom inflationary boom, £ weak inflationary boom, £ weak entry to ERM exit from ERM, recession recession recession recession

— — RPPE RPPE RPPE — — — RPPE RPPE RPPE — — — RPPE RPPE RPPE — — RPPE — — — — — — — — — —

Sources: Tables 8A.1 and 8.1, and the underlying official sources.

The second distinction is between strategic, tactical and regular, where ‘strategic’ reversals are ones that result from a major change in the monetary authorities’ perspective on the conjunctural position of the economy or in their underlying objectives as the result, perhaps, of foreign exchange market pressures; ‘tactical’ reversals are those where the authorities feel they have to deviate in the short term from unchanged objectives, while ‘tactical unwind’ reversals are those where they return to those underlying objectives when the pressure has passed; and ‘regular’ reversals are those where the authorities expect a priori to be varying rates regularly in both directions. On this basis 13 reversals (out of 35) were strategic; these include, for example, the first of the repeated rises from May 1988 designed to bring the boom under control, and the first rise under the Thatcher government in 1979. There were 16 tactical reversals (including 6 ‘tactical unwind’), many of them connected to sharp movements in sterling, and 6 regular (the latter mostly in the DM-shadowing phase). It is also worth noting that there are 13 cases in which the change is classified as (a) temporary or very temporary and (b) tactical or regular; and 11 cases where the change is (a) sustained and (b) strategic; for the latter cases the average duration is 107.9, whereas for the former it is

Interest Rate Smoothing 165 Table 10.4 Policy rate reversals Size/sign of i

Date 5.5.75 17.11.75 26.4.76 22.11.76 28.11.77 9.1.78 12.4.78 1.3.79 13.6.79 3.7.80 16.9.81 14.10.81 29.11.82 16.3.83 10.5.84 9.8.84 11.1.85 21.3.85 9.1.86 19.3.86 14.10.86 10.3.87 6.8.87 23.10.87 2.2.88 17.3.88 2.6.88 8.10.90 16.9.92 17.9.92 12.9.94 13.12.95 31.10.96 8.10.98 8.9.99

0.25 −0.25 1.5 −0.25 2 −0.5 1 −1 2 −1 2 −0.5 1.125 −0.5 0.5 −0.5 0.875 −0.5 1 −1 1 −0.5 1 −0.25 0.5 −0.5 0.5 −1 2 −2 0.5 −0.25 0.25 −0.25 0.25

Background to i

Sustained, temporary or very temporary

Strategic, tactical or regular

monetary policy passive recession, falling inflation £ weak end of £ crisis uncapping of £ real economy weak monetary overshoot, £ weak £ strong, pre-election new government £ strong, domestic squeeze £ weak £ stabilised £ weak £ strong £ weaker, US rates rising £ stabilised £ weak, $ very strong £ stabilised oil-price fall oil-price fall £ weak DM-shadowing DM-shadowing DM-shadowing DM-shadowing £ appreciating inflationary boom ERM entry ERM crisis undoing crisis measures inflationary pressure inflation controlled inflation pressure, pre-election domestic + world slowdown domestic demand growth

temporary temporary sustained sustained sustained very temporary sustained very temporary temporary sustained temporary sustained temporary sustained sustained very temporary sustained sustained very temporary sustained very temporary sustained very temporary temporary very temporary very temporary sustained sustained very temporary sustained sustained sustained sustained sustained sustained

regular strategic tactical strategic strategic tactical tactical unwind tactical strategic strategic tactical tactical unwind tactical tactical unwind tactical tactical unwind tactical tactical unwind tactical strategic tactical regular regular regular regular regular strategic strategic tactical tactical unwind strategic strategic strategic strategic strategic

Sources: Tables 8A.1 and 8.1, and the underlying official sources. Notes: sustained reversal = policy rate does not return to its previous level within 12 months; temporary = rate returns within 12 months; very temporary = rate returns within 6 months; strategic reversal = related to major change in authorities’ perspective of conjuncture and/or authorities’ underlying objectives; tactical = authorities feel obliged to deviate in short term from unchanged objectives; tactical unwind = unwind of previous tactical reversal; regular = authorities expecting a priori to adjust interest rates in both directions.

55.4,8 so that the sustained and strategic reversals look more like the typical reversals found in other countries. Moreover, it is clear from the background column in Table 10.4 that many of the reversals in the 1970s and 1980s, and a particularly high proportion of those that were tactical or regular, were connected to the authorities’ efforts to manage the exchange rate. 8 The

corresponding number for the other 11 reversals is 41.

166 The Making of Monetary Policy in the UK, 1975–2000 There are three points to be drawn from this discussion. First, it is indeed true that a relatively high proportion of the large changes and the large reversals reflected regime changes or the recognition of prior policy errors, as Goodhart (1999a) suggested. Second, a considerable number of the reversals can be classified as temporary or very temporary, and/or as tactical or regular; and a high proportion of these were connected to efforts to manage the exchange rate. Finally, there appear to be some important differences between 1975–89 and the 1990s: in the 1990s there are fewer large changes, fewer reversals, and no temporary or tactical/regular reversals after 1992, so that interest rates are smoother, especially under inflation targeting.9 Thus the impression that the UK has been less of an interest rate smoother than other countries can be explained largely by reference to its preoccupation with the exchange rate in the earlier years, and the impression does not hold in any case for the 1990s.

10.3 EXPLANATIONS OF SMOOTHING There are five main kinds of explanation in the literature for the existence of interest rate smoothing. First, it has been suggested that central banks (monetary authorities) intrinsically dislike financial market volatility. Thus Goodfriend (1991: 10) reports that ‘The Fed is understood to dislike “whipsawing the market”, i.e., following a target change too closely with a change in the opposite direction’, while Rudebusch (1995: 272) refers to a presumed desire on the part of the Federal Reserve to avoid ‘undue stress’ on financial markets. Some analytical models (e.g. Svensson, 2000) have included an interest rate stabilisation element in the central bank’s loss functions. However, no convincing explanation has been offered of the costs that volatility might impose (Bean, 1999), and Lowe and Ellis (1997) have questioned whether official interest rate reversals could cause bond market volatility in any case. Second, various authors have pointed to aspects of the decision-making process as encouraging central banks to smooth. On the one hand (a) Goodhart (1997) suggested that politicians in particular are reluctant to raise interest rates on the basis of forecasts rather than current evidence of inflationary pressures—so central banks then become reluctant to lower them until current evidence of inflation being under control has accumulated. On the other hand (b) Caplin and Leahy (1997) note that frequent reversals make central bankers look badly informed and may lower their credibility, while Goodhart (1999a) emphasises the reluctance of central bankers to risk having to make reversals, which he relates to the difficulty of explaining to outsiders why interest rates should follow something approaching a random walk. Third, a number of papers have argued that interest rate smoothing as it exists may be justified if there is significant uncertainty over the relevant economic data and/or the key relationships involved. Sack (2000) showed that when account is taken of both additive (data) uncertainty and multiplicative (parameter) uncertainty optimal monetary policy implies a path for the Federal funds rate which is closer (than the optimal path without uncertainty) to that historically observed, although this path still has fewer reversals and more no-change observations than the historical series.10 Goodhart (1999a) reported some similar analysis for the UK by Martin and Salmon (see also their 1999 paper). These results were all obtained with VAR models of the 9 These differences are discussed in more detail in Section 10.4 below. A part, but only a part, of the difference may be due to the lower average level of interest rates (and inflation) in the 1990s. See Table 10.5 below. 10 See also Sack and Wieland (1999) for a review of empirical evidence for the US on both uncertainty and forwardlooking behaviour.

Interest Rate Smoothing 167 US and UK economies, but similar exercises involving small structural models with Taylorrule type reaction functions have indicated that data uncertainty may be important but not parameter uncertainty (Rudebusch, 2001a; see also Estrella and Mishkin, 1999, and Peersman and Smets, 1999). Aoki (2000) provides a rigorous analytical demonstration of why the central bank should be more cautious in the presence of data uncertainty; in his model the demand and supply shocks are serially correlated, and this means that the central bank needs to use the lagged interest rate in its estimation of the true values (as opposed to the erratic first published estimates) of the output gap and inflation. Fourth, Woodford (1999) has developed an insight first put forward in Goodfriend (1991), who notes that what is important for real expenditure is not the current very short-term interest rate which the central bank can influence most strongly, but short- and medium-term interest rates. Woodford’s model incorporates forward-looking behaviour by the private sector. In this case it may be efficient for the central bank (if it dislikes financial market volatility) to change interest rates in a highly predictable manner, that is with few reversals, in order to maximise the impact of its actions on the short- and medium-term interest rates that are arguably more important for economic activity. Woodford also argues that this explanation sheds light on ‘why those central banks that exhibit the greatest degree of inertia in their behaviour . . . are those with the best records with respect to inflation control . . . it makes a great deal of sense that those central banks that have most clearly learned the benefits of commitment with regard to the average rate of inflation should also be the ones that are also [sic] most able to benefit from a perceived commitment to predictable responses to shocks as well—a commitment that should manifest itself in inertial behaviour’ (1999: 8–9).11 Fifth, Sack and Wieland (1999: 4) and Woodford (1999: 4–5) among others have noted in passing that, if there is persistence in inflation and output, monetary policy which reacts systematically to those variables may be smoothing simply because of that persistence. This idea is given more prominence by Rudebusch (2001b), who rejects the partial adjustment explanation of the importance of the lagged dependent variable in estimated Taylor-rule reaction functions. Such adjustment should be predictable, but the term structure appears to contain little predictive information about the future path of short-term interest rates beyond one or two months. Rudebusch considers serially correlated output gap revisions as an explanation for the apparent existence of partial adjustment, but argues they cannot provide a complete explanation. Instead he emphasises ‘serially correlated or persistent special factors or shocks that cause the central bank to deviate from the policy rule’ (2001b: 20). Aoki’s (2000) model, discussed above, explains smoothing as the result of serially correlated shocks in the presence of noisy indicators (of inflation and output). Goodhart (2001c) suggests that smoothing is due to serial correlation in the forecasting process, with forecasters underestimating the extent of upward/downward pressure on inflation at the start of upswings/downswings. Of these explanations, the first, second (b) and fourth predict fewer reversals, while the first, second (a), third and fourth predict more gradual adjustment of the policy rate. More gradual adjustment implies that the monetary authorities move the policy rate only part of the way towards the level they consider appropriate. By contrast, the fifth explanation implies that the authorities always adjust the rate fully, but they repeatedly find that they have to go 11 However, he does not try to distinguish between the proposition that banks who understand inertial behaviour have been best able to control inflation, and the proposition that the advantages of inertial behaviour become apparent only when central banks have established reasonable credibility.

168 The Making of Monetary Policy in the UK, 1975–2000 further in the same direction in succeeding periods, and it is this that leads to serially correlated movements of the policy rate.

10.4 EVIDENCE ON SMOOTHING FOR THE UK This section first presents disaggregated data on the degree of smoothing in the UK since 1975, which indicates that policy has been more smooth in the 1990s and particularly in the phase of inflation targeting with instrument independence since 1997. It then discusses the direct evidence that can be brought to bear on the reasons for smoothing in the latter phase. Table 10.5 provides the standard array of data on policy interest rate changes for seven different phases and two larger subperiods, differentiated in terms of the underlying monetary policy regimes, within the overall sample.12 Excessive disaggregation is unlikely to be helpful Table 10.5 Disaggregated data on smoothing by change

Period or phase 1975M1–2000M12 1975M1–1990M9 1990M10–2000M12 1975M1–1979M5 1979M6–1981M3 1981M4–1985M3 1985M4–1990M9 1990M10–1992M9 1992M10–1997M5 1997M6–2000M12

Months

Total changes

Average change

Average level

Average duration

Ratio of continuations to reversals

312 189 123 53 22 48 66 24 56 43

176 134 42 55 5 39 35 12 14 16

0.64 0.68 0.51 0.64 2 0.68 0.58 0.83 0.48 0.30

9.9 11.6 7.2 10.2 15.3 11.1 11.8 11.5 6.1 6.3

37.3 29.1 63.7 21.2 101.8 27.0 33.4 63 85.8 44.9

4.0 4.0 4.3 5.9 1.5 3.9 2.9 3.0 3.7 7.0

Number of changes Period or phase

++ +− −+ −−

1975M1–2000M12 1975M1–1990M9 1990M10–2000M12 1975M1–1979M5 1979M6–1981M3 1981M4–1985M3 1985M4–1990M9 1990M10–1992M9 1992M10–1997M5 1997M6–2000M12

39 28 11 10 1 7 10 0 3 8

17 13 4 4 1 4 4 2 1 1

18 14 4 4 1 4 5 1 2 1

Average duration ++

102 41.7 79 35.8 23 56.8 37 44.4 2 111 24 14.1 16 34.7 9 — 8 77.3 6 49.1

+−

−+

Average size −−

++

+−

−+

−−

70.4 64.6 25.3 0.91 47.6 53.8 19.3 1.15 144.3 102.3 46.3 0.30 26 35.8 12.8 1.3 165 49 92 3 28.0 57.0 25.5 1.32 59.5 66.6 15.6 0.7 131.5 86 45.2 — 224 129.5 60.8 0.42 90 64 28.5 0.25

0.63 0.56 0.88 0.5 1 0.5 0.56 1.5 0.25 0.25

1.01 1.09 0.75 1.19 2 1.13 0.8 2 0.38 0.25

0.48 0.47 0.51 0.41 2 0.44 0.43 0.56 0.56 0.38

Sources: compiled from Bank of England Statistical Abstract, and Bank of England Quarterly Bulletin, various issues. 12 The raw data on policy rate changes can be found in Table 8A.1. In calculating durations Goodhart’s (1997) procedure has been followed: weekends are omitted and holidays other than 24–26 December and 1 January are ignored.

Interest Rate Smoothing 169 in an exercise of this kind, and some of the phases previously used in this study for the 1970s and 1980s are too short to be used here. However, it seems sensible to separate out the 1979–81 phase of hard monetary targets and the 1990–92 ERM phase which were very different from what came before and after them; to divide the 1980s as between the soft monetary targeting of 1981–85 and the lack of binding targets of 1985–90; and to divide the rest of the 1990s as between before and after the granting of instrument independence to the Bank of England in May 1997.13 It is also useful to contrast the subperiod 1975–90 with that of 1990–2000. The daily data is graphed in Figure 8.1. In terms of the ratio of continuations to reversals there is little difference between the 1975–90 and 1990–2000 subperiods. However, the average duration data qualify this picture somewhat: the second subperiod shows much more clearly the pattern characteristic of the more smoothing countries, namely that the durations are higher for reversals than for continuations. In addition the average sizes of changes are closer, as between the four different categories, for the second than for the first subperiod. Of the four phases which make up the first subperiod, 1975–79 has a particularly high, and 1979–81 a particularly low, ratio of continuations to reversals, while the first three phases show varying ‘non-smooth’ patterns of durations and sizes as between increases and decreases, and continuations and reversals. Of the three phases which make up the second subperiod, 1997–2000 has a much higher ratio of continuations to reversals, while all three (and the fourth phase of the second subperiod, 1985–90) have the typical ‘smooth’ pattern of durations. The various types of change are closer in size for the 1997–2000 phase and to a lesser extent that of 1992–97 (and 1985–90).14 Data collected on an ‘event’ basis, such as that in Table 10.5, suffers from two disadvantages: it takes no explicit account of the times for which there is no change in interest rates, and in counting the changes it pays as much attention to changes that were superseded within two weeks as those which stood for six months. An alternative approach, which has been used by Sack (2000), Goodhart (1999a) and Martin and Salmon (1999) mainly in order to provide a basis for comparison with model predictions, is to aggregate changes on a quarter-by-quarter (or month-by-month) basis, using ‘bins’ to collect together the interest rate changes in each quarter (month). Table 10.6 presents data on this basis, where the 0.5↑ columns, for example, gives the number of quarters in which (for continuations or for reversals) the interest rate rise was equal to or greater than 0.375 but less than 0.625; the table also reports the quarters with no change. The phases and subperiods used correspond to those used in Table 10.5, but there are some minor differences in the phases due to the use of quarters rather than months. Figure 10.1 graphs the end-quarter data; comparison with Figure 8.1 makes clear that some of the very short-term oscillations in the 1970s and 1980s on the daily data are not visible in the end-quarter data. Table 10.7 then arranges the data in a more tractable form comparable to Tables 5.5 and 5.6 of Martin and Salmon (1999). In terms of either the proportion of reversals (the reciprocal of Goodhart’s ratio of reversals to total changes) or the ratio of continuations to reversals given in the final column of Table 10.7, it is clear that the first subperiod is considerably less smooth than the second. The data for the phases show, however, that in the first subperiod 1979–81 is smooth15 but the other three phases very non-smooth. In the second subperiod, 1990–92 loses 13 With respect to the ten phases used in Chapter 9, Table 10.5 aggregates the first two (1975–77 and 1977–79) and the fifth to seventh (1985–87, 1987–88 and 1988–90). 14 Policy was also smoother in the final phase in the sense that the ratio of the average size of change to the average level was lower than in other phases or subperiods. 15 This partly reflects the fact that 1979 Q2 – 1981 Q1 has one less reversal than June 1979 – March 1981; given the short length of the phase and the small number of changes, this makes a big difference.

3 2 1 4

10 10

15 9 6 3 2 1 3 4 2

↓ >1

↑ 0.75

1

1

1

4 1 2 1

2 3

8 4 4

↓ 0.5

1

7 2 5 1

Continuations ↓ ↑ 0.75 0.5

3

4 1 3 1

↑ 0.25

2

2

2

↓ 0.25 30 16 14 2 3 2 9 2 7 5

No change

1 1

2

2

↑ 0.25

1

1

1

↓ 0.25

1

1

2 1 1

↑ 0.5

1

1 1

1

1 1

Reversals ↓ ↑ 0.5 0.75

1

1 1

↓ 0.75

2 3

4

9 9

↑ >1

1

10 8 2 2 1 3 2 1

↓ >1

Sources: compiled from Bank of England Statistical Abstract, and Bank of England Quarterly Bulletin, various issues. Note: the ‘bins’ for 0.25 include all changes where 0.125 ≤ i < 0.375, and similarly those for 0.5 and 0.75; the bins for 1 include all changes above 1 as well as those ≥ 0.875.

1975Q1–2000Q4 1975Q1–1990Q3 1990Q4–2000Q4 1975Q1–1979Q1 1979Q2–1981Q1 1981Q2–1985Q1 1985Q2–1990Q3 1990Q4–1992Q3 1992Q4–1997Q1 1997Q2–2000Q4

Period

↑ >1

Table 10.6 Disaggregated data on smoothing by quarter

Interest Rate Smoothing 171 18 16 14 12 10 8 6 4 2 0 1975Q1

1980Q1

1985Q1

1990Q1

1995Q1

2000Q1

Figure 10.1 Policy rate (end quarter), 1975–2000

a reversal (that of Black Wednesday which lasted for a day only) from the move to quarterly data, and looks much smoother on the latter. 1997–2000 is smoother on both quarterly and event data than 1992–97. Overall, the evidence from Tables 10.5 and 10.7 suggests that the smoothest phase was that of 1997–2000: this scores high on the ratio of continuations to reversals (Tables 10.5 and 10.7) and low on the proportion of reversals (Table 10.7), and it has the standard smoothing pattern of higher durations for reversals than for continuations (Table 10.5). It also happens to have a profile, in terms of Table 10.7, which is remarkably close to that reported for the US (over the

Table 10.7 Data on smoothing by quarters in percentages

Period 1975Q1–2000Q4 1975Q1–1990Q3 1990Q4–2000Q4 1975Q1–1979Q1 1979Q2–1981Q1 1981Q2–1985Q1 1985Q2–1990Q3 1990Q4–1992Q3 1992Q4–1997Q1 1997Q2–2000Q4 Source: Table 10.6.

Ratio of continuations No change 0.75 and to reversals and 0.25 0.5 above Continuations No change Reversals (not %) 38 27 53 18 38 13 41 25 61 60

17 13 24 12 0 38 0 13 28 27

45 60 22 71 63 50 59 63 11 13

45 41 51 47 50 44 32 63 44 53

29 25 34 12 38 13 41 25 39 33

26 33 15 41 13 44 27 13 17 13

1.7 1.2 3.5 1.1 4.0 1.0 1.2 5.0 2.7 4.0

172 The Making of Monetary Policy in the UK, 1975–2000 8

7

6

5

4 Jan 1997

Jan 1998

Jan 1999

Jan 2000

Figure 10.2 Policy rate (daily data), 1997–2000

period 1984–96) in Martin and Salmon (1999).16 1992–97 is very smooth on quarterly data, but less so on the change basis (Table 10.5). 1975–79, 1981–85 and 1985–90 are the least smooth on the quarterly data, but the first two of these include a lot of within-quarter smoothing and look smoother on the change basis; however, 1975–79 and 1981–85 do not have a clear pattern of average durations being larger for reversals than for continuations. These distinctions are broadly consistent with the point made in Section 10.2 above that, insofar as UK policy has been less smooth than that in other countries, this is largely because for periods in the 1970s and 1980s policy was determined in part by the aim of managing the exchange rate. It follows that, in order to explain smoothing in more general terms, more attention should be paid to the 1990s, and particularly to the final phase considered here, 1997–2000. However, the contrasts between different phases or subperiods will sometimes be brought to bear on the different explanations. Concentration on 1997–2000 is convenient because the main type of evidence that can be adduced here is much stronger. For this phase there is surprisingly good direct evidence in the form of the minutes of the Monetary Policy Committee (MPC) meetings (including the Annexes which give the summary of data presented to the MPC by the staff of the Bank). The framework of individual accountability, together perhaps with the presence on the committee of external members (mainly academics) as well as central bankers, means that individual members have incentives not merely to make the ‘right’ decisions, but to explain and justify those decisions. The result is that the minutes provide significant detail not just on the decisions taken or the immediate reasons for them, but on the underlying thinking of members of the committee (including differences of view on, for example, the issues of reversals and uncertainty). The quarterly Inflation Reports are also useful insofar as they contain in the forecasts information 16 Martin and Salmon report that actual US changes were as follows: no change or 25bps: 55%, 50bps: 18%, 75bps or more: 27%; and continuations: 55%, no change: 31%, reversals: 14%. The profile in Table 10.7 for 1990–2000 is also close to this.

Interest Rate Smoothing 173 on the MPC’s expectations regarding output and inflation which can be brought to bear on the serial correlated shocks explanation. For the earlier years the relevant sources are the Bank of England Quarterly Bulletin, the Inflation Report from its inception in 1993, and for the period from January 1994 to May 1997 the published minutes of the Monthly Monetary Meetings. However, the Bulletin contains relatively few relevant remarks. Its discussions of interest rate decisions are largely about the appropriate level for rates, but there are occasional remarks about ‘tactics’, about uncertainty and about the exchange rate effects of interest rate changes. The main tactical idea to be found in more general discussions of monetary policy, such as lectures by the Governor, is that an earlier change in the interest rate may avoid the necessity of a larger change later;17 this implies some preference for a path of interest rates with less variance, but not necessarily for one with less reversals. The MMM minutes contain a little more discussion of tactical issues, and some other relevant information, but do not come near the detail and openness of the MPC minutes. The relevant information for each MPC meeting is summarised in Table 10.8. It is designed not to cover everything contained in the minutes, but to give an assessment of the role of ‘fundamental’ factors, i.e. the latest news on the factors regarded by the MPC as driving inflation—domestic demand and output, the world economy, the labour market, the exchange rate, and so on—relative to non-fundamental factors; and to note the various non-fundamental issues discussed such as those of timing or uncertainty or the cost of reversals which are relevant for evaluating the various explanations of smoothing. It notes for each meeting the change if any in the policy rate, and the vote that led to the change (with arrows to indicate the choice(s) preferred by the outvoted minority); it lists the main non-fundamental issues discussed by the MPC, including those referred to in the minutes as ‘tactical’ (indicated by T) and ‘other’ (O) considerations and any other issues relevant to smoothing; and in the final column it gives a summary assessment of the relative importance of fundamental factors to other considerations in the actual decisions reached.

10.5 EVALUATING THE EXPLANATIONS OF INTEREST RATE SMOOTHING 10.5.1 The Dislike of Financial Market Volatility The equity market has not in general been a major concern of the UK monetary authorities, because of the lower importance of direct holdings of equity in household portfolios compared, for example, with the US. Interest rates were temporarily reduced in line with other countries in the aftermath of the October 1987 crash, but this is the only obvious example within the period. The bond market was more of a concern in the 1950s and 1960s, when the Bank of England preferred to ‘lean against the wind’ in order to stabilise gilt-edged prices, on the grounds that the demand for bonds depended on the movement in prices rather than their level. By the midor late 1970s these ideas had given way to a greater readiness to manipulate interest rates in order to ensure sales of public sector debt when required.

17 See, for example, BEQB, 1987c: 370, and, for later echoes in monetary policy discussions, the minutes of the MMMs for June 1994 (§16) and 23 September 1996 (§27).

0

+0.25

0

0

0

October 1997

November 1997

December 1997

January 1998

February 1998

7

+0.25

August 1997

7.25

7.25

7.25

7.25

7

7

6.75

+0.25

July 1997

0

6.5

+0.25

June 1997

September 1997

Level (after)

Rise/ cut

4c:4↑

5:3↑

8:0

7:0

7:0

7:0

5:0

6:0

6:0

Pro: anti

various arguments for and against delaying rise: delay could help smooth output without risk, February IR forecast unusually uncertain, rise now might need to be reversed with damage to credibility; but uncertainty normal, cost to

possible global impact of UK change on Asian crisis; uncertainties about output gap and its impact; case for waiting for February IR forecast

uncertainty about output gap, case for waiting for more news; exchange rate could be especially sensitive in current global financial market uncertainty

(T) effect of unexpected rise will be alleviated by new IR; (T) case for raising now rather than later when MPC could be wrongly perceived as reacting to some bad data

August IR statement; uncertainty about effect of monetary and fiscal tightening so far; gains from waiting for new data expected November and new IR forecast

August IR statement that it should now be possible to pause to assess direction in which risks are likely

gains from waiting for more data and avoiding upward pressure on exchange rate versus importance for credibility of raising now to ensure forecast on-target

danger of raising too little too late; advantages of early tightening to smooth output; uncertainty about size of rise needed or effects of rise on exchange rate; more comprehensive analysis in August IR will clarify size of rise needed

timing of Budget; likely effect of rise on exchange rate, which also susceptible to EMU uncertainties: no reason to think delaying rise would lessen ultimate effect on £

Non-fundamental issues: ‘tactical’ (T) considerations, ‘other’ (O) considerations, and other timing or smoothing-relevant issues or arguments

Table 10.8 MPC discussions of interest rate decisions, 1997–2000

fundamentals and other considerations both balanced

fundamentals, uncertainty (both)

uncertainty (data): wait for more news

fundamentals, to ensure IR projection on target

fundamentals balanced, uncertainty (parameter)

fundamentals balanced, pausing to take stock

fundamentals, to ensure IR projection on target

mainly fundamentals; also uncertainty (parameter)

fundamentals (existing, not new)

Basis for decision: fundamentals/other

7.25 7.5

7.5 7.25

6.75

0 +0.25 0 0 0 −0.25

−0.5

May 1998

June 1998

July 1998

August 1998

September 1998

October 1998

November 1998

7.5

7.5

7.25

0

April 1998

7.25

0

March 1998

8:1↓

7:2↓

7:2↓

7:2↑↓

9:0

8:1↓

6:2↑↓

5:3↑

4c:4↑

(T) uncertainty about earnings data but no case for delay; (T) market impact of cut larger than expected 0.25% would be mitigated by publication of IR shortly; (T) larger cut could give useful information about MPC’s behaviour, but that can’t justify it; likely that further cuts will be needed, but lags imply those cuts not optimal now

only small monetary easing needed, but must avoid blow to confidence given market expectations of cut; risk that large reduction might have to be partly reversed

agreement to state that balance of risks had changed since August as result of international financial crisis

discussion of whether policy should move to reduce the hump in inflation forecast for first half of 1999

large amount of recent news, trends unclear; case for waiting for August IR to evaluate

despite market expectation of no further rise, impact of rise on equity market and exchange rate likely to be small

concern to signal that no change does not mean interest rate has peaked

differing views on impact of rise on exchange rate, given market expectation of no change; cost of waiting a few months, in terms of any additional tightening necessary, would be small; May IR as opportunity to reassess

discussion on how uncertainty enters forecast and implications for interest rate decisions, including that delay was justified by uncertainty when only small change was in question, cost of short delay would be small and new information was due soon; discussion of costs of reversals: could be misunderstood (create uncertainty, damage credibility), versus can be explained by new developments or improved analysis in which case there was no case for delay

credibility of not raising when forecast above target, later rise might need to be larger, exchange rate effects

continues overleaf

fundamentals, to ensure IR projection on target

mainly fundamentals plus uncertainty (both)

fundamentals

mainly fundamentals

fundamentals plus uncertainty (parameter)

fundamentals

fundamentals

mainly fundamentals

mainly fundamentals, also uncertainty (data)

0

May 1999

5.25

5.25

−0.25

April 1999

5.5

−0.5

February 1999

5.5

6

−0.25

January 1999

0

6.25

−0.5

December 1998

March 1999

Level (after)

(continued )

Rise/ cut

Table 10.8

5:4↓

8:1↓

8:1↓

8:1↓

7:2↓→

8:1↓

Pro: anti

(O) discussion of weight to put on current saucer-shaped inflation projection; (O) implications of renewed appreciation; dangers in appearing to change interest rates in response to high-frequency exchange rate movements; reversal might be needed if cut now which, given still incomplete understanding of MPC’s monetary policy reaction function, might adversely affect sentiment

(T) MPC agrees not to take into account possibility that, given market expectation of cut, no change or larger cut would be interpreted as signal that rates at trough, and not to take into account likely imminent cut by ECB

(O) some signs of faster response to interest rate changes, which could facilitate action to offset short-term weakness; (O) it would not be sensible for policy to react to high-frequency movements in exchange rate; (O) argument that changing rates soon after IR needs to be justified on basis of clear news; discussion of case for cut 0.5% which would be seen as over-reaction and might shock financial markets

Non-fundamental issues: ‘tactical’ (T) considerations, ‘other’ (O) considerations, and other timing or smoothing-relevant issues or arguments

fundamentals finely balanced, so some tactics

fundamentals

fundamentals, also uncertainty (mainly parameter)

fundamentals, to ensure forecast on target, some uncertainty (both)

fundamentals, uncertainty (data)

mainly fundamentals

Basis for decision: fundamentals/other

0

August 1999

5.5

5.5

5.75 6

+0.25

0

+0.25 + 0.25

0

November 1999

December 1999

January 2000

February 2000

March 2000

6

5.25

0

5.25

5

5

5

October 1999

+0.25

0

July 1999

September 1999

−0.25

June 1999

9:0

8:1→

8:1↑

6:3↑

8:1↑

9:0

7:2→

9:0

9:0

8:1↑

(O) MPC agrees timing of Budget is not a constraint; (O) markets expect no change: agreed this is not a constraint but change would need careful explanation; case for waiting for more data and analysis of Budget

(O) discussion of further sterling appreciation, sectoral imbalances, possible policy responses, likely impact of no change given market expectation of rise

(O) Y2K etc.; (O) new IR + forecast due February, will allow analysis of puzzles, but news should lead to change now

(T) discussion of Y2K, end-year and end-millennium effects; little risk in waiting a month or two for more data on effects of previous rises and competitive pressures

MPC discusses (T) steepening of sterling yield curve, (T) relevance of imminent ECB and FOMC decisions, (T) deeper saucer shape of inflation projection

some uncertainties, but no pressure for change given September rise

Y2K problems should not constrain policy; given market expectation of no change, discussion of changes 2.5% 2 years out might cause modest money market tightening, welcome

(T) markets expect no change in UK, or by FOMC or ECB: any UK change would need to be well explained; (T) difficult to calibrate recent effects, but forthcoming forecast round, plus IR to explain

even if further rise needed later, costs of waiting are low compared to possible costs of rise now (worsening sterling overvaluation)

challenge posed by mix of domestic inflationary pressures and external disinflationary pressures from sterling’s further appreciation; (T) FOMC expected to raise, if MPC does not UK might decouple from dollar relative to euro

(O) markets expect rise sometime in Q2 but not this month—agreed this not binding; case for waiting for more analysis, in May IR; volatility of equity markets means rise might soon need to be reversed; no rise now might create expectation of larger rise later

Non-fundamental issues: ‘tactical’ (T) considerations, ‘other’ (O) considerations, and other timing or smoothing-relevant issues or arguments

Source: Minutes of MPC meetings. Note: c under pro:anti for February and March 1998 indicates that the Lorman used his casting vote to keep the policy rate unchanged.

0

Rise/ cut

(continued )

April 2000

Table 10.8

fundamentals

fundamentals

mainly fundamentals

mainly fundamentals

fundamentals

mainly fundamentals, some uncertainty (parameter)

mainly fundamentals

mainly fundamentals

fundamentals, uncertainty (parameter)

Basis for decision: fundamentals/other

Interest Rate Smoothing 179 The financial market which the UK monetary authorities were most concerned to stabilise in the 1970s and 1980s was the foreign exchange market (see Chapters 8 and 9). In 1977 they lowered interest rates in a long series of small steps in an attempt to stabilise the existing nominal exchange rate (but eventually had to choose their monetary target over the exchange rate target, and then restored interest rates to a more appropriate level). In the 1980s there were a number of episodes in which the authorities seem to have lowered interest rates in small steps partly in order not to disturb the foreign exchange markets: in March 1985, for example, ‘The Bank’s cautious attitude [in preventing a larger fall in interest rates] at this time reflected in particular continuing volatility of the foreign exchange markets’ (BEQB, 1985b: 189). In fact there were four episodes in the 1980s when interest rates were reduced in a series of at least four steps by a total of at least 2%: November 1981 – November 1982 (14 steps, total reduction 7%), August–November 1984 (6 steps, 2.375%), March–July 1985 (7 steps, 2.5%) and March–May 1987 (4 steps, 2%). In each of these episodes the authorities were concerned either to limit a depreciation or (in the last case) to prevent an appreciation. A plausible interpretation of these episodes is that the authorities reduced interest rates gradually both so as to avoid disturbing the foreign exchange market into a significant depreciation, and because they did not know how far rates could fall without an effect on sterling. Thus they were not setting rates at each moment at the level they thought appropriate given the fundamentals of the situation. This behaviour led in each case to a series of small steps in an obvious example of smoothing. However, in three of these episodes the gradual reduction was in fact followed before long by a depreciation which led to a sharp rise in rates (see the reversals in Table 10.4 of 29.11.82, 11.1.85 and 6.8.87). Another episode of a partly comparable nature is the reduction in interest rates in early 1991, which the authorities described as ‘deliberately cautious’ and as ‘taking full account of the need to maintain market confidence in the authorities’ commitment to the counterinflationary stance of policy and to the ERM discipline’ (BEQB, 1991b: 203). This series of reductions was also followed, with a lag, by the sharp rise of 16 September 1992 (but that was reversed the next day so it disappears from the quarterly data). Thus while attempts to stabilise the exchange rate often produced smooth movements in interest rates they also led to sharp reversals, and the phases in which such attempts occurred are among the least smooth, particularly on quarterly data (1981–85 and 1985–89). By contrast 1979–81, when policy largely neglected the exchange rate, was much smoother on the quarterly data. There was also one episode in which interest rates were raised in a long series of small steps: June–August 1988 (8 steps, 4.5%). With respect to that occasion the authorities remarked that ‘Domestic interest rates had been increased four times during June, on each occasion following a weakening of sterling, thus tightening monetary conditions without provoking an over-reaction in the foreign exchange market’ (1988d: 485). The authorities did indeed succeed in avoiding a sharp appreciation and overvaluation of the kind that occurred in 1979–80 or later in 1997–8, and there was no quick reversal. But there are no other obvious examples of smoothing being successful in this way. There are one or two cases in the MMM minutes where decisions were affected by concern about financial market volatility. In March 1994, for example, the Governor opposed an interest rate cut for a variety of reasons, but one was that ‘financial markets internationally remained very unsettled . . . it was important to build up credibility in monetary policy . . . ’ (MMM, 30 March 1994 §18). In February 1995 the Governor argued (and the Chancellor accepted) that the fragility of financial markets made it particularly important to raise interest rates at once,

180 The Making of Monetary Policy in the UK, 1975–2000 as expected in the markets, rather than delaying until the next month (§§23–25, 35).18 But the contribution of such concerns to actual decisions must have been minor, occasional and short-term. For the MPC period there are a few statements in the minutes consistent with the MPC ‘caring’ about financial market volatility. In March 1999, for example, the MPC took the view that ‘it would not be sensible for policy to react to high frequency movements in the exchange rate, as this could lead to a volatile path for interest rates from month to month, and might make it more difficult for others to understand the motives for interest rate changes’ (§28) which (given the word ‘and’) implies that it cared at least to some degree about interest rate volatility. But it seems rare for decisions to have been motivated in this way. The summary of data presented to the MPC by the Bank’s staff reports, more or less clearly and more or less consistently, the financial market’s expectations of the decision the MPC is about to make,19 and in 20 out of the 43 months considered here those expectations are explicitly acknowledged in the minutes by the MPC itself. Out of the 43 there are 7 cases where the MPC made decisions that were clearly unexpected: three unexpected rises (November 1997, June 1998 and September 1999); two unexpected cuts (February and June 1999); and two larger than expected cuts (November and December 1999). If the MPC always felt constrained by the market’s expectations, or alternatively if policy was always predictable and ‘boring’ (King, 1997b: 440), there would have been no such cases, so the actual proportion—one-sixth of the total number of meetings and nearly one-half of the meetings at which the policy rate was changed—would seem to indicate some willingness to confound the market’s expectations. Early on the MPC seems to have taken the view that the impact of an unexpected interest rate decision would be small or unclear. In November 1997, for example, it ‘noted that, although a rise would probably come as a surprise to some, any initial impact on financial markets might be reassessed once the Inflation Report was published’ (§32).20 In December 1998 the MPC took the view that it ‘was difficult to judge the market consequences of a different decision’ from that expected, since the reaction of sterling could go either way (§33).21 But by March 2000 the MPC was arguing that ‘a change in interest rates this month would come as a surprise to the market and would need to be explained carefully. That of itself should not be a constraint on individual members’ decisions, but the Committee would need to take into account the prospective effects of any surprise to the markets’ (§§31).22 In the light of these considerations it is difficult to believe that a desire to avoid financial market volatility was a major determinant of the MPC’s interest rate smoothing. 10.5.2 The Decision-making Process (a) Tendency to Adjust Rates in Response to Current Evidence rather than Forecasts It is certainly possible to find examples from the 1970s and 1980s of reluctant increases consistent with this idea. For example, the authorities may have been slow to raise rates in the late 1970s and in the Lawson boom. It is perhaps less easy to find examples of reluctant 18 See also MMM, May 1995 §24; May 1997 §30. 19 More clearly and more consistently as time went

on. The summary originally formed part of the Minutes, but from September 1997 it was attached to the Minutes as an Annex. From April 1999 these expectations are presented in the Market Intelligence section of the Annex. 20 See also MPC, June 1998, §36. 21 See also MPC, April 1998, §46; April 1999, §21–22; February 2000, §36. 22 See §33 for discussion of the latter in this instance. Similar general statements can be found in MPC, July 2000, §39; December 2000, §36.

Interest Rate Smoothing 181 decreases, but 1977 (when interest rates were reduced by some 9% in 20 steps between January and October) may have an element of this, although the main factor must have been the authorities’ desire to preserve the competitive advantage gained from the 1976 depreciation in the face of capital inflows, but to keep to their new monetary target. Other long stepped declines in the 1980s were more a matter of the authorities feeling their way as to how far they could cut rates without the exchange rate sliding, rather than, say, a reluctance to cut on the basis of forecasts of declining inflation. For the 1990s, on the other hand, it is more difficult to find examples of increases or decreases consistent with the idea. If rates were reduced more slowly than they might have been in 1990–92 that was largely due to the UK’s membership of the ERM. It is arguable that rates should have been raised faster and further in 1996 and early 1997, but the reasons for this lie partly in the political context (a forthcoming election) and partly in the appreciation of sterling (used by the Chancellor as an argument for resisting the Bank of England’s call for rates to be raised), as well as in any reluctance to accept forecasts (as opposed to current evidence) of rising inflation as a reason for action. In the 1997–2000 phase, it is difficult to think of any examples. In any case the discipline imposed by the MPC procedures, with their forecasts published at quarterly intervals, would militate against any such smoothing other than very short-term, between forecasts. However, as in the dislike of volatility case, it is also necessary to consider carefully what would be the overall consequences of this kind of reluctance to adjust interest rates. Insofar as sluggish adjustments lead to worse outcomes, in terms of higher inflation in booms and deeper recessions, they are likely to lead in the end to sharp and large reversals, of the kind discussed above in connection with attempts to stabilise the exchange rate. The overall effect on smoothing is not clear-cut. But there was less smoothing in the 1970s and 1980s, when the authorities may perhaps have acted in this way, than in the 1990s when they do not seem to have done so, particularly in 1997–2000 when policy was smoothest. This view of the decision-making process does not therefore offer a convincing explanation of smoothing. (b) Central Bankers’ Dislike of Reversals It is difficult to find any references to interest rate reversals or their cost in the 1970s and 1980s. However, there are two references in the MMM minutes, for December 1995 and 3 July 1996. In the first case, the Governor recommended a cut of 0.25% (rather than 0.5%), saying, ‘Such a move might also help to encourage public acceptance of smaller interest rate changes, entailing lower costs if they had to be reversed subsequently’ (MMM, December 1995, §27). In the second case, the Governor said there was no strong new evidence yet that could justify a reversal of last month’s cut, and to raise now ‘would send confusing signals to both financial markets and the wider economy’ (MMM, 3 July 1996, §21). These statements are not inconsistent with, but certainly do not establish, the proposition that central bankers dislike reversals. However, the MPC minutes contain a number of references of this sort. In February 1998, when a tied vote was decided by the Governor’s casting vote in favour of no change, one of the arguments made on the winning side was that [i]f the downturn proved to be much sharper than currently expected, an immediate increase in interest rates might have a larger negative effect on output than in other circumstances and would have to be quickly reversed. Such a reversal could impair confidence in the economy and damage the credibility of the MPC process at this stage of its development by creating confusion about monetary policy and by encouraging the mistaken impression that the Committee had been

182 The Making of Monetary Policy in the UK, 1975–2000 indifferent to the course of output. Such an outcome would be more damaging than if an increase in interest rates which proved to be necessary were delayed, so that a larger increase was needed later. (MPC, February 1998, §27)

The next month the minutes contain a section on ‘general considerations bearing on interest rate setting’ which includes a discussion of the optimal frequency of interest rate changes with particular reference to the costs of rapid reversals. On one view such reversals could be misunderstood, creating uncertainty both in financial markets and in the wider economy and damaging the credibility of the MPC process. An alternative view was that this position was irrational. So long as any policy reversals could be properly explained by new developments or improved analysis of the outlook, they need not create confusion about policy goals or the MPC’s approach. On that view, there were no benefits in delaying changes in rates to reduce the risk of reversals, and it was better to make changes as soon as they appeared necessary. Also the desire to minimise the risk of policy reversals was likely to mean that interest rate changes would, on average, be made too late. (MPC, March 1998, §44)

These two views both surface again later, in October 1998, January 1999, May 1999 and July 1999.23 It seems that the MPC was divided between those who thought that reversals had a significant cost at least in some situations in terms of credibility, and those who thought that if properly explained there should be no such cost. The availability of individual voting records for the MPC members makes it possible to examine whether there is something systematic at work here, whether, for example, career central bankers are more wary of reversals than external academic members. Table 10.9 classifies the dissenting votes of the various members in categories designed to show their attitudes to reversals. Dissents are defined with reference to actual policy which over this period was very smooth. Thus voting for no or less change (against a majority favouring change) (column (1)) can be thought of as highly smoothing; voting (at a turning point) for a reversal sooner than the majority (column (2)) is also smoothing, in the sense that it would decrease the variance of interest rates; voting for a further or larger change in the course of an upwards or downwards movement (i.e. not at a turning point) (column (3)) is anti-smoothing in the sense that it speeds up the adjustment of rates; voting for a larger or further change than the majority at a turning point (column (4)) is (more) anti-smoothing in the sense that it is likely to widen the variance of interest rates and increase the chance of a reversal being necessary; and voting for a reversal against a continuing trend (column (5)) is highly anti-smoothing since it would be likely to add directly to the number of reversals. Table 10.9 also lists the number of meetings attended, the percentages of overall dissent (column (7)) and of anti-smoothing dissent (column (8)), and the degree of symmetry or asymmetry, in the sense of the number of times when members voted for the policy rate to be above or below the majority: asymmetry might indicate that the dissents were due to a difference in overall preferences (from the median) rather than in the degree of smoothing espoused. Table 10.9 classifies the members of the MPC in four different groups: Internals I are members with a central banking or at least non-academic background; Internals II are people with 23 The minutes for July 1999 also contain an interesting statement, associated with the voting position of DeAnne Julius, which gives three reasons for moving in steps rather than one large cut: ‘First, a series of reductions might have a greater impact on consumer and business sentiment. Second, it might allow the Committee to learn more about the way the economy was responding to past shocks and policy changes. Third, it was important not to out-pace significantly the expectations of financial markets, as this could have unwanted consequences for asset prices and the exchange rate’ (§32). The first of these arguments seems to echo a discussion in Lowe and Ellis (1997), the second to draw on some parts of the uncertainty explanation for smoothing, and the third to assume that financial market volatility is undesirable.

1 1 0 4

No or less change (not at turn) (1) 0 0 0 5

Sooner reversal (2) 2 5 7 17

Larger/further change not at turn (3) 2 8 0 5

Larger/further change at turn (4) 0 0 0 2

Reversal against trend (5) 128 74 54 109

Number of meetings attended (6) 3.9% 18.9% 13.0% 30.3%

Degree of dissent = (cols (1)–(5)) ÷ (col. (6)) (7)

3.1% 17.6% 13.0% 22.0%

Degree of dissent = (cols (3–5)) ÷ (col. (6)) (8)

Source: MPC minutes. Note: ‘symmetry of dissent’ shows the number of decisions on which members preferred a level of the policy rate above (up) or below (down) the majority.

Internals I Internals II Externals I Externals II

MPC members:

Votes for:

Table 10.9 MPC members’ dissenting votes, June 1997 – December 2000

3 up, 1 down 14 up, 0 down 7 up, 0 down 12 up, 21 down

Symmetry of dissent (9)

184 The Making of Monetary Policy in the UK, 1975–2000 backgrounds in academic economics who became full-time employees of the Bank (and internal members of the MPC); Externals I are external members of the MPC who had significant prior experience within the policymaking process, either in the Treasury or at the Bank itself; and Externals II are members without substantial previous official policy involvement. The most important point that emerges is that Internals I (who have typically constituted three out of the nine members of the MPC) are the least dissenting and Externals II the most; and, curiously perhaps, Externals I are less dissenting than Internals II. It is also worth noting that much the most common form of dissent is that of column (3), voting for a larger or further rise when (as we know from hindsight) the rate is in the middle of an upswing or downswing. The fact that this is common implies that many members have felt at different times that the majority is dragging its feet and adjusting less quickly than would be desirable; even so, there were only 14 quarters out of 43 in which at least one member voted in this way. Finally, with respect to symmetry, it is important to note that the period as a whole contains two upswings in interest rates and only one downswing, so it might be expected that a member whose preferences over inflation were equal to the median but who consistently favoured faster adjustment (column (3)) would have more ‘ups’ here than ‘downs’. On this basis some of the apparent anti-smoothing tendency of Internals II should be regarded as the result of these members being ‘tough’ relative to the rest of the MPC with respect to possible overshoots of the inflation target but not on undershoots, and some of the apparent anti-smoothing tendency of Externals II as a group (there are very large variations within the group) should be regarded as the result of their being relatively less tough on undershoots but not on overshoots. Overall, however, the table indicates some support for the proposition that the career central bankers are more anti-reversal and pro-smoothing than the external and/or academic members of the MPC. 10.5.3 Uncertainty For the earlier years there are some statements in the official sources which indicate that policy-makers changed interest rates in smaller steps as the result of uncertainties. In May 1977, for example, ‘continuing uncertainties about the course of inflation, and the need to finance the prospective public sector borrowing requirement for 1977/78, indicated that the decline [in interest rates] should not go too far or too fast’ (BEQB, 1977c: p. 301–2). And in March 1983, ‘the authorities took the view that the underlying monetary and fiscal situation called for caution, so that any immediate fall in rates should be limited in size’ (BEQB, 1983b: 171).24 These are cases of parameter uncertainty, which seems to be more common, but there are examples of data uncertainty also. For example, in January 1996 the Bank of England argued against a further cut in interest rates partly on the grounds that some important new data would soon become available (MMM, January 1996, §20). However, in order to assess its contribution to smoothing it is also necessary to consider the outcome of these gradual declines. In most cases they were followed before long by sharp and large rises. But if the policy rate had been reduced less gradually but less far (the obvious counterfactual here), there might have been fewer negative continuations (particularly on the change basis rather than on quarterly data), but there might also have been fewer reversals. Thus, insofar as uncertainty was an important factor in policy, its effect on smoothing is not clear. For 1997–2000 it is possible to analyse the issue in more detail. Table 10.8 includes 15 meetings (out of a total of 43) for which ‘uncertainty’ appears in the final column as an 24 See

also BEQB, 1985c: 344.

Interest Rate Smoothing 185 important element in the decision, mostly in conjunction with ‘fundamentals’; in four of the 15 it is mainly data uncertainty, in seven it is mainly parameter uncertainty, and in four it is both.25 Uncertainty is also mentioned among the non-fundamental issues discussed in a number of other cases. The MPC opted for no change in 11 of those 15 meetings or 73%, while it opted for no change in 63% of all meetings; at the same time uncertainty featured in 41% of the 27 meetings where no change was made, but in 35% of all meetings. Thus uncertainty seems to have been only slightly more influential in cases where the decision was for no change. In 3 out of the 4 other cases where uncertainty appears in the final column it contributed to a decision to change the policy rate by 0.25% rather than a larger amount; but these 3 are only 23% of the 13 cases in the whole sample where a change of 0.25% was made.26 So uncertainty does not seem to have been a major contributor to decisions for more gradual change. In addition, in 8 cases out of the 15 the minutes mention the benefits of waiting for the next (quarterly) Inflation Report and its forecast as a way of reducing the uncertainty; this implies that if uncertainty did contribute to smoothing, it was only to smoothing over one or two months.27 That conclusion is strengthened by consideration of the MPC’s explicit discussion of how uncertainty entered the forecast, and the implications for the determination of interest rates, in its March 1998 meeting. Two views were expressed on the issue of whether uncertainty was a reason for delay in taking action: ‘The first was that policy should reflect the latest news and that uncertainty in itself was no reason for delay. The second was that there might be particular cases, where the implied need was for only a small interest rate change, when the cost of a short delay would be small and new information might give confidence about the need for the change, or not. In those circumstances a delay could be warranted to reduce the risk of unnecessary reversals of policy’ (§43). These points suggest that, although uncertainty was undoubtedly an important element in policymaking, it was not a major cause of smoothing. 10.5.4 Forward-looking Behaviour For the 1970s and 1980s it is difficult to believe that the authorities were restricting themselves to changing their very short-term policy rate in small and predictable steps, with few reversals, in order to maximise the effect of their actions on short-term and medium-term interest rates. Policy at that time was geared in large part to the exchange rate: since the latter is affected precisely by very short-term interest rates the authorities are likely to have been thinking in large part of the very short rather than the longer maturities. In addition, ideas of rational expectations and forward-looking behaviour were less well-developed and less prevalent in policymaking circles (notably in the 1970s).28 In the 1990s, on the other hand, these concepts were typically incorporated in some form into many of the models used by the Bank of England, and policy was more inclined to neglect, than to target, the exchange rate. Policy in the 1990s was smoother, which would be consistent with the notion that the authorities were behaving ‘inertially’ in the later subperiod. The difference between the subperiods would also be consistent with Woodford’s (1999) link between inertial 25 As indicated by the parentheses after ‘uncertainty’ in the final column of Table 10.8. 26 There were three changes of 0.5%, one of them at a meeting where uncertainty featured. 27 In fact there is no concentration of change decisions in Inflation Report months: of the 16

change decisions six were in IR months, five in ‘IR + 1’ (or ‘IR − 2’) months, and five in ‘IR + 2’ (or ‘IR − 1’) months. 28 See for example Lane (1985).

186 The Making of Monetary Policy in the UK, 1975–2000 behaviour and successful anti-inflationary policies, since the rate of inflation was much lower in the 1990s: the UK authorities should be expected on those grounds to smooth interest rates more in the 1990s, and perhaps especially in 1997–2000, than in the earlier years.29 It is difficult to find references in the official sources that shed light on the relevance of this explanation; indeed before 1997 there is nothing. The MPC minutes contain two statements that appear to be relevant. In June 1999 the rate was reduced on a 8:1 vote (the reduction was opposed by Mervyn King), and in July 1999 the minutes record the view that, even though ‘it might have been preferable last month to have left the repo rate unchanged . . . there were two reasons for not reversing last month’s decision . . . Second, a reversal might be wrongly interpreted as a stronger signal of the future path of interest rates than was warranted’ (§29). In November 1999 it was noted that, following the reversal of September 1999, short- and mediummaturity interest rates had risen so that ‘the market was doing some of the Committee’s work for it’; though it was argued against this that ‘the Committee would take material risks with its credibility if it failed to tighten in line with market expectations . . . ’ (§33). These statements both have echoes of Woodford’s (1999) analysis, but they were not important to the decisions that were taken: other arguments in the minutes suggest the MPC would have kept the policy rate unchanged in July 1999 anyway, and in November 1999 it was raised. Attitudes to reversals are also relevant here: if the authorities were behaving inertially they would want reversals to occur only at strategic turning-points so that they would be perceived as signalling a sustained move in interest rates in the opposite direction. The discussion above suggested that the external and/or academic members of the MPC had less inhibitions about reversals, while the internal members may have been inclined to avoid reversals—but this was for credibility reasons rather than for the benefits to be gained from Woodford’s inertial behaviour. Inertial central bankers might also be inclined to signal when interest rates had peaked or bottomed, but the MPC has not done this very clearly. In May 1998 the minutes include an expression of concern from the majority position to signal that a lack of change in that month did not mean that the rate had peaked. In September 1998 the MPC kept the rate at 7.5% but issued a statement that ‘it recognises that deterioration in the international economy could increase the risks of inflation falling below the target. The Committee will continue to monitor these risks.’ The statement the following month, when the policy rate was first cut, refers strictly to what ‘is now appropriate’. It would certainly have been possible to convey much more clearly, if the MPC had so wished, that interest rates were now firmly heading downwards. On the other hand, the statement accompanying the other reversal in this period, in September 1999, justifies the rise in part by saying that ‘an early move could lower the level at which interest rates might otherwise need to be set’, which could be taken to imply an expectation of some further increase.30 However, the clearest argument against this explanation of smoothing under the MPC relates to the framework of policy. In this phase each Inflation Report contains a forecast of inflation made on a constant interest rate assumption. Moreover, there are a number of occasions on which it has been argued in the MPC that the policy rate should be adjusted so that the forecast 29 However, this difference is not consistent with the view expressed by Freeman (1999: 120): ‘I think that the growing credibility of inflation-targeting regimes and the increased attention being paid by financial markets to the need for central banks to get “ahead of the curve” bode well for future monetary policy actions being closer to the optimal path, with more reversals in response to changes in view or new information than we have seen in the past.’ 30 Bank of England press notices of 10 September and 8 October 1998 and 8 September 1999.

Interest Rate Smoothing 187 comes into line with the target: August 1997, November 1997, November 1998 and, less clearly, November 1999 and February 2000. But this implies that, at least in each forecasting month, the policy rate is fully adjusted to what the MPC thinks will be required to hit the target in two years’ time. Inertial behaviour, on the other hand, would typically involve less than complete adjustment by a central bank that knows it will raise or lower interest rates further but chooses to do this gradually and predictably.31 On the basis of this evidence, it seems difficult to accept the forward-looking behaviour explanation of smoothing. 10.5.5 Serially Correlated Shocks In principle it would be possible to examine the serially correlated shocks explanation for the whole period 1975–2000 by looking at (a) official forecasts of GDP, inflation and so on, (b) outturns for these variables, (c) potential exogenous shocks, and (d) indications of awareness of the shocks and deviations from forecast on the part of the authorities and their responses to them. However, such an exercise would be fraught with difficulties, notably with respect to (c) and (d).32 Instead, it seems better to concentrate here again on the 1997–2000 phase and to draw on the direct evidence, including the Inflation Report with its forecasts as well as the MPC minutes. The question at issue is whether it is possible to tell a story of the changes in the policy rate in terms of (i) the rate being (more or less) completely adjusted each month, and (ii) news leading to changes in subsequent months. Table 10.8 sheds some indirect light on (i): the assessment in the final column gives the fundamentals as the main determinant of policy rate decisions in most months, and in all months in which there was a change in the rate. This suggests that the MPC was adjusting the policy rate fully rather than partially. The discipline of having to publish a quarterly inflation forecast on a constant interest rate assumption would also have pushed the MPC to adjust fully, as already discussed. However, a closer look at the news involved (ii) may strengthen the argument.33 The phase starts in June 1997 after a long period in which the Bank of England called for higher interest rates but the Chancellor declined. In the final meeting under the old framework but with a new Chancellor, the policy rate was raised by 0.25% but the Chancellor clearly accepted that rates might need to rise again shortly (MMM, May 1997, §39). The newly formed MPC raised the rate by 0.25% in each of the first three meetings (June–August 1997), and concluded in its August Inflation Report that ‘monetary policy has now reached a position at which it should be possible to pause in order to assess the direction in which the risks are likely to materialise’ (BEIR, 1997c: 5). The implication is that the June–August rises were a gradual adjustment from a starting point at which it was accepted that the rate was well short of where it needed to be. The gradualism should probably be ascribed mainly to the fact that the 31 Kohn, in his (2001) report on MPC procedures takes a similar view of the implications of the constant interest rate assumption for the forecast: ‘One risk of the emphasis on the two-year-out inflation target, taken together with the use of the unchanged policy assumption to present that target, may be a more active policy—one with greater movements in the policy rate—than might be optimal’ (Kohn, 2001: 45). See also Goodhart (2001c: 176). 32 Goodhart (2001c) also stresses the issue of interim policy changes which would have affected the outturns, especially over non-short forecast horizons. 33 The next four paragraphs draw on the ‘Overview’ and ‘Monetary policy since . . . ’ sections in the relevant Inflation Reports, as well as the various MPC minutes. Table 10.2 on page 172 shows the movements of the policy rate from January 1997.

188 The Making of Monetary Policy in the UK, 1975–2000 MPC was feeling its way as a new body, with its credibility not yet established or its reaction function understood.34 By November, both inflation and output growth had turned out to be about 0.5% higher than expected in August, while the appreciation of the exchange rate had had less effect so far than expected (BEIR, 1997d: 5), and the MPC decided to raise the policy rate another 0.25%. For the next six months the fundamentals seem to have been fairly balanced and/or uncertain, and a majority of the committee35 voted to hold the rate constant. But in June 1998 a sharp fall in the exchange rate and some new earnings data persuaded a majority to vote for a further 0.25% rise.36 Policy was unchanged for the next three months, but in September the deterioration in the world economy led the MPC to issue the statement quoted in the previous subsection, which was an indication that the ‘bias’ of policy had now shifted. The following month the MPC reduced the policy rate for the first time, in response to the international situation and domestic survey evidence, having ‘concluded that the slowdown in aggregate demand was likely to be greater, and more protracted, than thought at the time of the August Inflation Report’ (MPC, October 1998 §23). There was a succession of further cuts in the next four months, and the policy rate was reduced in all from 7.5% to 5.5%. The main items of news that convinced the MPC to cut were the further weakening of the world economy, falls in business and consumer confidence in the UK, and slower than expected consumption growth. Rates were cut again by 0.25% in each of April and June 1999; in the former case the continued strength of the sterling exchange rate, lower than expected growth in earnings and revisions to GDP data were the most important factors, and in the latter case the first two of these, together with declining retail price inflation, were the main influences. In July and August 1999 the MPC voted unanimously for no change in rates, but in September a majority voted for an increase of 0.25%, and this was followed by similar rises in November and in January and February 2000. In September and November 1999 the (majority of the) MPC were responding mainly to new data showing stronger growth in final domestic demand, notably consumption, than expected in the August Inflation Report, together with evidence of recovery from the housing and labour markets. By January there had been a tangible improvement in the world economy (relative to what had been expected in the November Inflation Report), and by February the growth rates of domestic money, credit and consumption were causing concern. This was the last change in the period considered here; it was followed by ten months in which the main choice was between no change and a further rise, but there was never quite a majority for the latter. The implication of the above account is straightforward: it is possible to tell a story in which the main changes in the policy rate are responses to new information becoming available in the month concerned.37 Insofar as the rate moves in a smooth way, this therefore implies 34 Moreover, the MPC was dominated by internal members at this point since the third external joined only in September and the fourth in December, and in the light of the discussion on the decision-making process the internals are likely to have favoured gradualism; all decisions were, however, unanimous. 35 In February and March 1998 the MPC split 4:4, and it was the Governor’s casting vote that kept the policy rate unchanged. 36 Some doubt was cast on the accuracy of the earnings data and major revisions were made twice in October 1998 by the Office of National Statistics; this led to an enquiry carried out by Peter Sedgwick and Martin Weale (Turnbull and King, 1999). 37 The number of members voting in change decisions for something different from what they had voted for in the previous month was 5 in one case, 6 in four cases, 7 in five cases and 8 in six cases. The fact that change decisions involved such widespread shifts strengthens the probability that they were responding to a common exogenous factor in the shape of news.

Interest Rate Smoothing 189 that the domestic and world economy change (relative to the MPC’s expectations) in serially correlated ways. That in turn implies that the MPC’s forecasts respond only partially to new information38 and/or that the underlying economies have stronger serially correlated processes than the MPC’s models of them.

10.6 CONCLUSIONS This chapter has examined the nature of interest rate smoothing in the UK, distinguishing between different phases and subperiods, and reviewed the various explanations of smoothing found in the literature. It then evaluated those explanations by making use of the differences between phases and subperiods and a range of direct evidence, particularly the minutes for the MPC period when the policy rate was smoother and for which much more information is available. The principal findings are the following. First, there was on the whole less smoothing in 1975–90 but more in 1990–2000, particularly in the years from 1997 when on the standard indicators of smoothing the UK looks very similar to the US. In addition, the relatively low degree of smoothing in 1975–90 was largely the consequence of regime changes and the recognition at intervals of prior policy errors. The latter in turn were mainly related to (often unsuccessful) attempts to manage the sterling exchange rate. Thus the relative lack of smoothing in this subperiod was clearly a sign of the poor performance of monetary policy rather than its success (as might be suggested by arguments that optimal policy would be less smooth). Second, the monetary authorities may have preferred to avoid financial market volatility, but such preferences do not appear to have had much influence on their decisions. Third, smoothing does not seem to be due to a reluctance on the part of the authorities to adjust rates on the basis of forecasts rather than current data, but there is some evidence of a tendency for career central bankers to be more pro-smoothing than other MPC members. Fourth, uncertainty was often present, but it was not a major cause of smoothing. Fifth, smoothing cannot be explained on the basis of forward-looking behaviour by the private sector and inertial behaviour by the central bank. Sixth, serially correlated shocks to the monetary authorities’ expectations do seem to be able to explain the major movements in the policy rate in the MPC period, though the cause of the serial correlation in the shocks is not obvious. Overall, then, the chapter suggests that the explanation of interest rate smoothing should be sought primarily in the serial correlation of shocks, with some minor and short-term influences from the apparent aversion of career central bankers to reversals and from uncertainty. 38 Goodhart (2001c) provides some suggestive evidence (in a table of two quarter ahead forecast errors for output and inflation from November 1997 to November 1999) that forecast errors have been serially correlated.

11 Conclusion: Competence, Commitment and Monetary Policy Policy Performance Over the period covered in this book the UK experienced an extraordinary variety of different monetary frameworks, including monetary targets with more or less discretion, formal and informal exchange rate targets, inflation targets, and no targets or framework at all. All of these frameworks were operated in a context where the central bank did not have instrument independence, but as a result of the changes of 1997 they were succeeded by a phase of inflation targets with instrument independence. On basic macro outcomes (inflation, economic growth and unemployment) there is little doubt that the final (and at this point continuing) phase of inflation targets with instrument independence has seen the best performance. However, the external environment was much more propitious to low inflation in the 1990s than in the 1970s or 1980s; and a comprehensive assessment that takes full account of the external environment would be both complex (notably in the light of the short periods involved) and beyond the scope of this book. What can be done here, however, is to offer some overall judgements by reference to the ending of the different phases of policy as well as their performance. The informal monetary targeting of 1975–76, which had coincided with the first and highest peak for inflation in the period, was quickly recognised as unsatisfactory. However, the more formal monetary targeting of 1977–86 was hardly an example of an effective policy regime: it is true that inflation came down (from its second peak in 1980) into line with that in other developed countries, but growth and unemployment were poor. More relevant here is that the monetary targets were repeatedly overshot to the extent that they can no longer have been (and were not believed by policy-makers to be) fulfilling their underlying objectives. The years of no binding or explicit targets which followed (and the short-lived phase of informal exchange rate targeting which came between these phases) were also an unhappy experience, with growth turning into an inflationary boom and credibility finally improving only on the back of tough counter-inflationary measures and the expectation that the UK would enter the ERM. The formal exchange rate targets of 1990–92 within the ERM saw inflation brought down from the peak of 1990, but the framework was abandoned after less than two years. The first phase of inflation targeting in 1992–97 seemed to offer some clear improvement in terms of macro outcomes as well as the monetary framework, but by the end of the phase the improvements were looking increasingly threatened. Finally the phase of inflation targeting with instrument independence since 1997 was accompanied by generally favourable macro outcomes and appeared to represent a satisfactory solution to earlier problems of political intervention in monetary policy. This chapter first reviews the operation of the different frameworks by reference to the strengths and weaknesses identified in the analytical literature surveyed in Chapter 1. This

192 The Making of Monetary Policy in the UK, 1975–2000 review, particularly the discussion of the issue of control, makes clear that a more comprehensive evaluation needs also to take into account the competence and commitment shown by the monetary authorities at different times, and these are considered in Section 11.2. Section 11.3 uses the analysis of interest rate smoothing to comment on the performance of monetary policy from a different perspective. Section 11.4 returns to the Achilles’ heel of UK monetary policy, the external dimension. Section 11.5 concludes.

11.1 DEMAND AND SUPPLY SHOCKS, INFLATION EXPECTATIONS AND CONTROL The analytical comparison of different monetary frameworks in Chapter 1 first emphasised the issue of stabilisation properties in the face of shocks, following on from the Poole (1970) analysis. The question that can now be posed is whether any part of the relative success and failure of the different frameworks as they have been operated in the UK was due to their different performance in the face of shocks; that is, whether substantial shocks—large enough to be noticeable to policy-makers—were inevitably handled in more or less satisfactory ways under the different frameworks. Domestic expenditure shocks are a continuing feature of the environment facing monetary policy-makers, and the monetary framework which cannot easily deal with them is that of exchange rate targets. As seen in Chapter 5 it is possible to argue that policy could have responded more effectively to the deeper than expected recession of 1990–91 under some other framework. But against that it needs to be stressed that the recession was a product of the policy required to control an inflationary boom which had been unleashed before the introduction of exchange rate targets, rather than an exogenous shock occurring within that framework; and there was a large, even if not large enough, reduction in interest rates during the ERM phase. The other frameworks can in principle adjust satisfactorily to domestic expenditure shocks, and there is no obvious example of a difficulty of this kind. The framework which cannot easily handle money demand shocks is that of monetary targeting, and such shocks in the form of financial innovation have long been identified, not least by the authorities themselves, as the crucial reason for the problems in and eventual abandonment of monetary targets in the UK. However, this account is open to challenge: Chapter 3 has argued that although its trend changed velocity did not become more unpredictable in the targeting period than it had been before, and that the monetary authorities themselves made a substantial contribution to the repeated overshoots of the monetary targets by the way in which they set the target ranges. In later years velocity was generally more stable, but in any case under the other frameworks money demand shocks are not in principle a source of difficulty for policy. The monetary framework for which foreign expenditure shocks are troublesome is that of exchange rate targeting. German reunification initially generated a positive foreign expenditure shock for other ERM countries, and for the UK this would have acted to soften the 1990–91 recession. Later, the relaxation of German fiscal policy together with the Bundesbank’s response to it brought about a negative foreign expenditure shock, but the impact on the UK came mostly after the exit from the ERM.1 So for the period of the UK’s membership of the ERM foreign expenditure shocks were not a major source of difficulty. Developments in 1 German GDP grew by over 5% in both 1990 and 1991, and by some 2% in 1992 (but –1.1% in 1993) (data from OECD Economic Outlook).

Conclusion: Competence, Commitment and Monetary Policy 193 the world economy are always difficult for policy-makers to predict, but foreign expenditure shocks can be, and appear to have been, taken care of more or less satisfactorily in the other frameworks. From the analysis of Chapter 1 shocks to foreign asset prices or the foreign exchange risk premium pose difficulties for monetary targets and larger difficulties for exchange rate targets. Chapters 8 and 9 have argued that there were throughout the period recurring exogenous shocks to the exchange rate, and these can be thought of as shocks to the risk premium. For monetary targeting the appreciation of 1979–81 was at least partly a shock of this sort, and policy-makers certainly found it difficult to deal with. But the appreciation also reflected a tightening of monetary policy which was partly a matter of poor design (that is, the monetary targets for 1979/80 and 1980/81 were non-reasonable and/or non-feasible); and there was an effect on the exchange rate whose direction, if not size, was predictable, from the rise in the world oil price given the UK’s new status as an oil producer. Under exchange rate targeting there was a major problem for policy in the form of a rise in foreign (German) interest rates, but it was argued in Chapter 5 that this rise could have been better handled by UK policy-makers and was not responsible in itself for the UK’s exit in crisis from the ERM. While inflation targeting can in principle deal more easily with shocks of this kind, the appreciation of the late 1990s (which started in the 1992–97 phase) was arguably not well handled: policy did not ‘lean into’ the appreciation enough to prevent it from being substantial, and policy-makers then became concerned about the impact on inflation of an unwinding of the appreciation which was repeatedly expected but had not (even by the end of 2001) materialised. Supply shocks are not an obvious candidate for explaining relative success. All of the monetary frameworks are frameworks for controlling demand, so that none of them can deal easily with supply shocks, but analysis suggests inflation targeting would find them harder to deal with. There is no obvious example of a major adverse supply shock during either of the phases of inflation targeting which did or did not cause difficulties for policy. However, the UK may have experienced a major positive supply shock over the 1990s in the form of a fall in the NAIRU,2 and this appears to have been handled well by monetary policy. The analysis of Chapter 1 went on to discuss the influence that different forms of target might exert on inflation expectations, as the result of varying degrees of interpretability and verifiability. The historical evidence gives no reason to question the argument that monetary targets are difficult for many agents to interpret and therefore have less ability to produce a useful effect on inflation expectations. But the evidence also suggests that the failure of monetary targets in this respect was partly attributable to the way in which the authorities themselves set the target ranges, and then missed them. In any case, the lack of transparency was not accidental: as Fforde (1983) has argued, the use of monetary targets was partly designed to deflect attention away from output and employment,3 whereas to affect inflation expectations strongly it would have been necessary to make explicit the expected trend in velocity and the implications of the target for output as well as inflation. 2 See, for example, Wadhwani (2001b), and Nickell (2001). 3 See Fforde (1983: 207): ‘it would have been possible to initiate

[a determined counter-inflationary strategy] with a familiar ‘Keynesian’ exposition about managing demand downwards, and with greater concentration on ultimate objectives than on intermediate targets. But this would have meant disclosing objectives for, inter alia, output and employment. This would have been a very hazardous exercise, and the objectives would either have been unacceptable to public opinion or else inadequate to secure a substantial reduction in the rate of inflation, or both. Use of strong intermediate targets, for money supply and government borrowing, enabled the authorities to stand back from output and employment as such and to stress the vital part to be played in respect of these by the trend of industrial costs.’

194 The Making of Monetary Policy in the UK, 1975–2000 The experience of exchange rate targets was too short for any judgement on their ability to influence expectations. Under inflation targeting there was a significant reduction in inflation expectations and increase in credibility but only well into the 1997–2000 phase, and the improvements must have reflected other factors besides the interpretability and verifiability of the particular target being operated. In Chapter 1 it was also suggested that there were problems of precision and of lags with respect to the control of monetary and inflation targets, and with respect to precision in crisis periods for exchange rate targets. The problems for monetary targets appear to have been strongly confirmed by experience, with six out of ten monetary targets being overshot, and there is no doubt that the repeated overshoots contributed to a loss of credibility which in turn contributed to the abandonment of monetary targeting. However, as argued in Chapter 3, a considerable part of the control problem related to the way in which the authorities set the target ranges and the way in which they chose to react to divergences from target. The control of the exchange rate during the formal targeting period4 was, as expected, straightforward in tranquil times but very difficult in the crisis which culminated in Black Wednesday. However, as argued in Chapter 5, the causes of the UK’s exit from the ERM were not ‘technical’ problems of control; the exit was the result partly of a difficult external environment but mainly of a fundamental weakness of policymaking in which the implications of a fixed exchange rate had not been worked out and agreed in advance. The control of inflation has generally been more successful. If the 1992–97 phase with decisions made by the Chancellor had continued it seems likely that the target would have been significantly overshot, but this would have occurred because of deliberately risky decisions rather than technical control problems. Under instrument independence inflation has been kept within 1% of the 2.5% target, so that (despite earlier expectations that this would be frequent) the Governor of the Bank has not yet been obliged to write an open letter to the Chancellor explaining the MPC’s past and future actions. One general reason for this relative success is the conducive international environment, but that can have been only one among several factors.

11.2 COMPETENCE AND COMMITMENT It is obvious from the previous section that control is not just a question of objective, exogenous difficulties in manipulating the relevant levers and switches. The ability of the monetary authorities to control a chosen target also depends on their own competence in and commitment to operating a particular monetary framework. Competence here refers to a wide range of matters, from the way in which the targets are set, through the understanding of exogenous factors and their impact, to the ability to forecast the development of the target under alternative values of the instruments which they control directly. Commitment refers to the authorities’ determination to hit their targets, which encompasses their willingness to react decisively to divergences, to face down criticism of their actions and to accept uncomfortable short-run trade-offs; but it also reflects the institutional constraints under which they are operating. The problems of control under monetary targeting can be understood only by appealing to these features: the targets were repeatedly missed not just because the authorities did not have adequate levers, but because they did not always set the targets competently and because—in the 4 Control under DM-shadowing was more difficult, no doubt in part because the informality of the target did not stimulate the equilibrating market movements that could be expected under a regime of explicit targets.

Conclusion: Competence, Commitment and Monetary Policy 195 face of unexpected pressures elsewhere in the economy—they were not ultimately committed to them. The failure to keep to the exchange rate target in 1992 reflects even more obviously a lack of both competence and commitment: the authorities did not understand clearly enough the choices with which they were faced and were unwilling to accept the short-run costs that adhering to the target would have involved, and it was these factors rather than the objective difficulties of an exchange rate target in that particular conjuncture that led to failure. Finally, the incipient problems in inflation targeting towards the end of the first phase in 1996–97 can be seen as the result of a lack of commitment to the announced inflation target on the part of the decision-maker. However, competence and commitment also influence the way in which announced targets affect inflation expectations. Poorly selected ranges for the monetary targets led in conditions of imperfect commitment to repeated overshoots which must have detracted from the credibility of subsequent target announcements. The credibility of the ERM exchange rate target was undermined by the visibly incomplete commitment of the authorities, together with the large role played by political rather than technical considerations in the details of entry. There is evidence that the repeated disagreements between Chancellor and Governor in the 1994–97 period, reflecting an incomplete commitment to the inflation target by the Chancellor, diminished the credibility of monetary policy in those years. In the 1997–2000 phase, on the other hand, the greater perceived commitment of policy-makers to the inflation target must have played a role in the reduction of inflation expectations and the improvement in credibility. On the other hand, that perceived commitment may have encouraged the foreign exchange market expectations which underlay the persistent overvaluation of sterling. Finally, competence and commitment may influence in rather different ways the reactions of policy-makers to shocks. For example, the impact of monetary targets on the real economy might well have been worse in the early 1980s if the authorities had not in the end compromised on their commitment to their monetary target (though the impact might have been even better if they had compromised sooner). Had the choice of target ranges been based on a higher level of technical competence, on the other hand, this might not have been necessary. In the ERM period a clearer understanding of the implications of German policy might have enabled the authorities to respond more effectively to the rise in German interest rates. And under inflation targeting greater accuracy in predicting the movement of the exchange rate might have enabled the authorities to reduce the scale of the appreciation without compromising on inflation.

11.3 INTEREST RATE SMOOTHING AND THE PERFORMANCE OF MONETARY POLICY It is clear that interest rates have been adjusted more smoothly in the 1990s, and particularly the late 1990s, than in the 1970s or 1980s. The literature on interest rate smoothing tends to assume that smoothing is inefficient, but in the UK case the lack of smoothing in the 1970s and 1980s (other than 1979–81) has been shown to be a sign of poor performance aimed mainly (and without success) at stabilising the exchange rate. In the 1990s, on the other hand, the degree of smoothing in the UK is much closer to that in the US. It is not obvious that US monetary policy is a good model, involving as it does a ‘just do it’ approach in place of a formal framework, and it has been criticised by, for example, Bernanke et al. (1999, chapter 12). Nevertheless, after the unproductive gyrations of UK interest rates in the earlier years it seems appropriate to count the movement towards the US model as evidence of progress.

196 The Making of Monetary Policy in the UK, 1975–2000 That judgement then opens the way for a search for an explanation of the apparently higher than optimal degree of smoothing, and that search focused on the serial correlation of shocks to the monetary authorities’ expectations which pushed them to move interest rates repeatedly in the same direction, even though at each point they believed that they had moved them far enough. The serial correlation of shocks is in some sense a sign of technical failure, though there is room for dispute over the extent of incompetence and the degree of objective difficulty in forecasting world economic activity, in particular. In the UK case, however, the existence of such a discussion is a clear sign that many of the earlier weaknesses of monetary policy have been resolved so that what are secondary matters by comparison come to the fore.

11.4 THE EXTERNAL DIMENSION Over the quarter of a century covered in the book there is no doubt that enormous progress has been made in monetary policy, in terms of price stability and credibility, and without detriment to macroeconomic outcomes in other aspects. The major continuing and unresolved problem is the exchange rate. During the period policy has moved sharply away from the deep concern with and attempts to control the exchange rate that characterised policy for most of 1975–89 (but not 1979–81). There is some evidence of a decline, rather than a rise, in the variability of the nominal exchange rate: according to Table 9.2 its standard deviation for 1988–90, 1992–97 and 1997–2000 was well below the levels of 1975–77, 1979–81, 1981–85 or 1985–87, though not as low as the levels of 1977–79, 1987–88 or 1990–92. But there is no clear trend of this kind for the real exchange rate. More importantly, the late 1990s saw a recurrence of a major misalignment, larger and more prolonged than that of the early 1980s, which policy has not so far been able to affect. UK exchange rate experience over the period can be summarised in three points: fixing the exchange rate has not been successful for more than short periods; managing the exchange rate has been difficult, costly and ultimately unsuccessful; but letting the exchange rate float has led to recurring problems of misalignment. The only ‘solution’ which has not been tried is that of joining a monetary union, which for the UK would mean EMU. Here there are many other factors to be taken into consideration— the framework, the competence and the commitment of the ECB, for example, the nature of the asymmetric shocks that the UK might experience in the future if it were a member of the eurozone, the extent of prior cyclical convergence, the appropriate entry rate, and so on. What has been said in this book does not lead to the conclusion that this ‘solution’ should be chosen. But it does indicate that keeping a separate currency with its own exchange rate involves some costs. It may well be that those costs are outweighed by the (economic and political) benefits. But it would be foolish to argue that they are insignificant.

11.5 CONCLUSIONS This book has examined the making of monetary policy in the UK over a quarter of a century which has covered a range of monetary frameworks and a wide variation in the external environment facing policy-makers. It started by reviewing the factors which are usually prominent in a priori comparisons of alternative frameworks. There is no question that those factors are important for the relative success or failure of different regimes. But the investigation of how

Conclusion: Competence, Commitment and Monetary Policy 197 individual frameworks operated, and how and why particular decisions within them were taken, has shifted the emphasis towards other factors, identified here as competence and commitment. Three further points may be made by way of conclusion. First, as between competence and commitment, priority should be accorded to the latter: committed policy-makers have an incentive to develop their technical competence so as to fulfil most completely their commitment (and there is reason to think that some such development has occurred at the Bank of England), whereas technically competent policy-makers cannot change their commitment insofar as that depends on the institutional environment within which they operate. Second, commitment and competence can override the choice of monetary framework. On the one hand, a committed and competent monetary authority may be able to make an intrinsically inferior framework function better than an intrinsically superior framework under an uncommitted and incompetent authority. On the other hand, in surveying the evolution of monetary frameworks in the UK it is the granting of instrument independence to the Bank of England in 1997 which must be seen as the most important advance, rather than the adoption of an inflation target in place of an exchange rate target in 1992–93. Finally, while the changes of the last 25 years have brought enormous improvements in the making, the operation and the outcomes of monetary policy, there are inevitably some problems of technical competence (notably with respect to forecasting) still to be resolved, and there remains one outstanding strategic issue: the exchange rate.

References Adam, C., Cobham, D. and Girardin, E. (2001), External influences and institutional constraints on UK monetary policy, 1985–2000, Discussion Paper 0115, Department of Economics, University of St Andrews. Alesina, A. and Grilli, V. (1993), On the feasibility of a one-speed or multi-speed European Monetary Union, Economics and Politics, 5, 145–65. Allsopp, C. (1991), Macroeconomic policy: design and performance, in M. Artis and D. Cobham (eds), Labour’s Economic Policies, 1974–79, Manchester: Manchester University Press. Allsopp, C. (1994), Discussion of Barrell, R., Britton, A., and Pain, N. (1994), in D. Cobham (ed.), European Monetary Upheavals, Manchester: Manchester University Press. Allsopp, C. (2001), Economic imbalances and UK monetary policy, Bank of England Quarterly Bulletin, 41, 484–94. Almeida, A. and Goodhart, C. (1998), Does the adoption of inflation targets affect central bank behaviour?, in J.L. Malo de Molina, J. Vi˜nals and F. Guti´errez (eds), Monetary Policy and Inflation in Spain, London: Macmillan. Aoki, K. (2000), On the optimal monetary policy response to noisy indicators, forthcoming in Journal of Monetary Economics. Aoki, K., Proudman, J. and Vlieghe, G. (2001), Why house prices matter, Bank of England Quarterly Bulletin, 41, 460–8. Artis, M. (1994), The EMS as a counter-inflationary framework, in J. Round (ed.), The European Economy in Perspective, Cardiff: University of Wales Press. Artis, M. and Cobham, D. (1991), Labour’s Economic Policies, 1974–79, Manchester: Manchester University Press. Artis, M. and Currie, D. (1981), Monetary targets and the exchange rate: a case for conditional targets, in W. Eltis and P. Sinclair (eds), The Money Supply and the Exchange Rate, Oxford: Oxford University Press. Artis, M. and Lewis, M. (1976), The demand for money in the UK, 1963–73, Manchester School, 44, 147–81. Artis, M. and Lewis, M. (1981), Monetary Control in the United Kingdom, Deddington, Oxfordshire: Philip Allan. Artis, M. and Lewis, M. (1984), How unstable is the demand for money in the United Kingdom? Economica, 51, 473–6. Artis, M. and Lewis, M. (1991), Money in Britain, Hemel Hempstead: Philip Allan. Atkinson, F. and Hall, S. (1983), Oil and the British Economy, London: Croom Helm. Avesani, R., Gallo, G. and Salmon, M. (1995), On the nature of commitment in flexible target zones and the measurement of credibility: the 1993 ERM crisis, in C. Bordes, E. Girardin and J. M´elitz (eds), European Currency Crises and After, Manchester: Manchester University Press. Bacchetta, P. (1994), Spain and the real exchange rate problem, in C. Johnson and S. Collignon (eds), The Monetary Economics of Europe: Causes of the EMS Crisis, London: Pinter. Ball, L. (1999a), Efficient rules for monetary policy, International Finance, 2, 63–83. Ball, L. (1999b), Policy rules for open economies, in J. Taylor (ed.), Monetary Policy Rules, Chicago: University of Chicago Press. Balls, E. (1998), Open macroeconomics in an open economy, Scottish Journal of Political Economy, 45, 113–32.

200 References Bank for International Settlements (1986), Recent Innovations in International Banking (Cross Report), Basle: BIS. Bank for International Settlements (1998), Annual Report 68 (1997–98), Basle: BIS. Bank of England (1984), The Development and Operation of Monetary Policy, 1960–83, Oxford: Oxford University Press. Barisone, G., Driver, R. and Wren-Lewis, S. (2000), Are our FEERs justified?, mimeo, University of Exeter. Barrell, R., Britton, A. and Pain, N. (1994), When the time was right? The UK experience of the ERM, in D. Cobham (ed.), European Monetary Upheavals, Manchester: Manchester University Press. Batini, N. and Haldane, A.G. (1999), Forward-looking rules for monetary policy, in J. Taylor (ed.), Monetary Policy Rules, Chicago: University of Chicago Press. Bayoumi, T. (1993), Financial deregulation and household saving, Economic Journal, 103, 1432–41. Bean, C. (1983), Targeting nominal income: an appraisal, Economic Journal, 93, 806–19. Bean, C. (1987), The impact of North Sea oil, in R. Dornbusch and R. Layard (eds), The Performance of the British Economy, Oxford: Oxford University Press. Bean, C. (1998), The new UK monetary arrangement: a view from the literature, Economic Journal, 108, 1795–809. Bean, C. (1999), Discussion of Goodhart (1999a), Bank of England Quarterly Bulletin, 39, 115–116. Bean, C. and Jenkinson, N. (2001), The formulation of monetary policy at the Bank of England, Bank of England Quarterly Bulletin, 41, 434–41. Begg, D., Danthine, J.-P., Giavazzi, F., Hamilton, C., Neven, D., Norman, V., Sapir, A., Smith, A., Winters, A. and Wyplosz, C. (1990), Monitoring European Integration: The Impact of Eastern Europe, London: Centre for Economic Policy Research. Bell, G. and Berman, L. (1966), Changes in the money supply in the United Kingdom, 1954 to 1964, Economica, 33, 148–65. Bell, M. (1995), The ERM and the route to EMU, in D. Currie and J. Whitley (eds), EMU after Maastricht: Transition or Revaluation?, London: Lothian Foundation. Bernanke, B., Laubach, T., Mishkin, F. and Posen, A. (1999), Inflation Targeting, Princeton: Princeton University Press. Bowen, A. (1995), British experience with inflation targetry, in L. Leiderman and L. Svensson (eds), Inflation Targets, London: Centre for Economic Policy Research; also in A. Haldane (ed.), Targeting Inflation, London: Bank of England Branson, W. (1993), Discussion of Eichengreen and Wyplosz (1993), Brookings Papers on Economic Activity, no. 1, 125–9. Branson, W. (1994), German reunification, the breakdown of the EMS and the path to Stage Three, in D. Cobham (ed.), European Monetary Upheavals, Manchester: Manchester University Press. Breedon, F. (1995), Bond prices and market expectations of inflation, Bank of England Quarterly Bulletin, 35, 160–5. Briault, C., Haldane, A. and King, M. (1996), Independence and Accountability, Bank of England Working Paper, no. 49. Brittan, S. (1982), How to End the ‘Monetarist’ Controversy, 2nd edn, London: Institute for Economic Affairs. Brittan, S. (1989), The Thatcher government’s economic policy, in D. Kavanagh and A. Seldon (eds), The Thatcher Effect, Oxford: Oxford University Press. Britton, A. (1991), Macroeconomic Policy in Britain 1974–87, Cambridge: Cambridge University Press. Bruce-Gardyne, J. (1986), Ministers and Mandarins, London: Sidgwick & Jackson. Brunner, K. and Meltzer, A. (1964), Some further investigations of demand and supply functions for money, Journal of Finance, 19, 240–83. Buckle, M. and Thompson, J. (1995), The United Kingdom Financial System in Transition, 2nd edn, Manchester: Manchester University Press. Budd, A. (1998), The role and operations of the Bank of England Monetary Policy Committee, Economic Journal, 108, 1783–94. Buiter, W. (2000), Optimum currency areas, Scottish Journal of Political Economy, 47, 213–50. Buiter, W. and Miller, M. (1981), Monetary policy and international competitiveness: the problems of adjustment, Oxford Economic Papers, 33, 143–75.

References 201 Buiter, W., Corsetti, G. and Pesenti, P. (1998a), Interpreting the ERM crisis: country-specific and systemic issues, Princeton Studies in International Finance, no. 84. Buiter, W., Corsetti, G. and Pesenti, P. (1998b), Financial Markets and European Monetary Cooperation, Cambridge: Cambridge University Press. Burns, T. (1988), The UK government’s financial strategy, in W. Eltis and P. Sinclair (eds), Keynes and Economic Policy, London: Macmillan. Cagan, P. (1965), Determinants and Effects of Changes in the Stock of Money, 1870–1960, New York: National Bureau of Economic Research. Caplin, A. and Leahy, J. (1997), The money game, New Economy, 4(1), 26–9. Carruth, A. and Henley, A. (1992), Consumer durables spending and housing market activity, Scottish Journal of Political Economy, 39, 261–71. Cecchetti, S., Genberg, H., Lipsky, J. and Wadhwani, S. (2000), Asset Prices and Central Bank Policy, Geneva Report on the World Economy no. 2, London: Centre for Economic Policy Research. Chick, V. (1973), The Theory of Monetary Policy, Oxford: Blackwell. Clarida, R., Gal´ı, J. and Gertler, M. (1998), Monetary rules in practice: some international evidence, European Economic Review, 42, 1033–67. Clarida, R., Gal´ı, J. and Gertler, M. (1999), The science of monetary policy: a New Keynesian perspective, Journal of Economic Literature, 35, 1661–707. Cobham, D. (1982), Domestic credit expansion, confidence and the foreign exchange market: sterling in 1976, Kredit und Kapital, 15, 434–56. Cobham, D. (1991a), The money supply process, in C. Green and D. Llewellyn (eds), Surveys in Monetary Economics, Vol. 2: Financial Markets and Institutions, Oxford: Blackwell. Cobham, D. (1991b), Monetary policy, in M. Artis and D. Cobham (eds), Labour’s Economic Policies, 1974–79, Manchester: Manchester University Press. Cobham, D. (1997), The German currency union and the crises in the European Monetary System, in S. Frowen and J. H¨olscher (eds), The German Currency Union of 1990, London: Macmillan. Cobham, D. (1998), Macroeconomic Analysis, 2nd edn, Harlow: Longman. Cobham, D. and Serre, J.-M. (1986) Monetary targeting: a comparison of French and UK experience, Royal Bank of Scotland Review, no. 149, 24–42. Cobham, D. and Serre, J.-M. (1987), The variability of monetary growth in France and the UK 1970–84, in C. Goodhart, D. Currie and D. Llewellyn (eds), The Operation and Regulation of Financial Markets, London: Macmillan. Cobham, D., Cosci, S., Mattesini, F. and Serre, J.-M. (2000), The nature and relevance of central bank independence: an analysis of three European countries, Discussion Paper 0017, University of St Andrews Department of Economics. Cobham, D., Papadopoulos, A. and Zis, G. (2001), The cost of political intervention in monetary policy, Discussion Paper 0114, University of St Andrews Department of Economics. Coleby, A. (1983), The Bank’s operational procedures for meeting monetary objectives, BEQB, 23, 209–15; also published in Meek (1983). Congdon, T. (1992), Reflections on Monetarism, Aldershot: Elgar. Coppock, D. and Gibson, N. (1963), The volume of deposits and the cash and liquid assets ratios, Manchester School, 31, 203–22. Courakis, A. (1981), Monetary targets: conceptual antecedents and recent policies in the US, UK and West Germany, in A. Courakis (ed.), Inflation, Depression and Economic Policy in the West, Oxford: Mansell. Cramp, A. (1966), The control of the money supply, Economic Journal, 76, 278–87. Cramp, A. (1971), Monetary Management, London: Allen & Unwin. Cukierman, A. (1992), Central Bank Strategy, Credibility and Independence, Cambridge, MA: MIT Press. Currie, D. and Dicks, G. (1990), Options for ERM entry, Economic Outlook (London Business School), June. Davies, S. (1991–92), UK interest rates and the level of sterling, Treasury Bulletin, 3(1), 27–36. Davis, R. and Meek, P. (1983), Monetary targeting—variations on a common theme, in P. Meek (ed.), Central Bank Views on Monetary Targeting, New York: Federal Reserve Bank of New York. De Grauwe, P. (1994), The Economics of Monetary Integration, 2nd edn, Oxford: Oxford University Press.

202 References De Grauwe, P. (2000), The Economics of Monetary Union, 4th edn, Oxford: Oxford University Press. Deacon, M. and Derry, A. (1994), Estimating market interest rate and inflation expectations from the prices of UK government bonds, Bank of England Quarterly Bulletin, 34, 232–40. Debelle, G. and Fischer, S. (1994), How independent should a central bank be?, mimeo, MIT (cited in Fischer, 1994). Dow, C. and Saville, I. (1988), A Critique of Monetary Policy, Oxford: Oxford University Press. Drazen, A. and Masson, P. (1993), Credibility of policies versus credibility of policymakers, Quarterly Journal of Economics, 109, 735–54. Eichengreen, B. (1993), The crisis in the EMS and the transition to EMU: an interim assessment, Working Paper no. C93-022, Centre for International and Development Economics Research, University of California at Berkeley. Eichengreen, B. and Wyplosz, C. (1993), The unstable EMS, Brookings Papers on Economic Activity, no. 1, 51–143. Estrella, A. and Mishkin, F. (1999), Rethinking the role of the NAIRU in monetary policy: implications of model formulation and uncertainty, in J. Taylor (ed.), Monetary Policy Rules, Chicago: Chicago University Press. Fforde, J.S. (1983), Setting monetary objectives, Bank of England Quarterly Bulletin, 23, 200–8. Filc, W. (1994), Credibility of German monetary policy on the road to EMU, in A. Steinherr (ed.), 30 Years of European Monetary Integration from the Werner Plan to EMU, London: Longman. Fischer, S. (1990), Rules versus discretion in monetary policy, in B. Friedman and F. Hahn (eds), A Handbook of Monetary Economics, Vol. 2, Amsterdam: North Holland. Fischer, S. (1994), Modern central banking, in F. Capie, C. Goodhart, S. Fischer and N. Schnadt, The Future of Central Banking, Cambridge: Cambridge University Press. Fisher, P. and Vega, J. (1993), An empirical analysis of M4 in the United Kingdom, Bank of England Working Paper no. 21. Flood, R. and Garber, P. (1984), Collapsing exchange-rate regimes: some linear examples, Journal of International Economics, 17, 1–13. Flood, R. and Rose, A. (1999), Understanding exchange rate volatility without the contrivance of macroeconomics, Economic Journal, 109, F660–F672. Foot, M. (1981), Monetary targets: their nature and record in the major economies, in B. Griffiths and G. Wood (eds), Monetary Targets, London: Macmillan. Forsyth, P. and Kay, J. (1980), The economic implications of North Sea oil revenues, Fiscal Studies, 1(3), 1–28. Freeman, C. (1999), Discussion of Goodhart (1999a), Bank of England Quarterly Bulletin, 39, 117–121. Friedman, B. (1975), Targets, instruments and indicators of monetary policy, Journal of Monetary Economics, 1, 443–73. Friedman, B. (1990), Targets and instruments of monetary policy, in B. Friedman and F. Hahn (eds), A Handbook of Monetary Economics, Vol. 2, Amsterdam: North Holland. Friedman, M. (1948), A monetary and fiscal framework for economic stability, American Economic Review, 38, 245–64. Friedman, M. (1953), The effects of a full-employment policy on economic stability: a formal analysis, in M. Friedman, Essays in Positive Economics, Chicago: University of Chicago Press. Friedman, M. (1960), A Program for Monetary Stability, New York: Fordham University Press. Friedman, M. (1968), The role of monetary policy, American Economic Review, 58, 1–17. Friedman, M. and Schwartz, A. (1963), A Monetary History of the United States, 1867–1960, Princeton: Princeton University Press. Funabashi, Y. (1988), Managing the Dollar: From the Plaza to the Louvre, Washington, DC: Institute for International Economics. Giavazzi, F. and Spaventa, L. (1990), The ‘new’ EMS, in P. De Grauwe and L. Papademos (eds), The European Monetary System in the 1990s, London: Longman. Girardin, E. (1994), Discussion of M´elitz (1994), in D. Cobham (ed.), (1994), European Monetary Upheavals, Manchester: Manchester University Press. Goodfriend, M. (1991), Interest rates and the conduct of monetary policy, Carnegie–Rochester Conference Series on Public Policy, no. 34, 7–30. Goodhart, C. (1973a), Monetary policy in the United Kingdom, in K. Holbik (ed.), Monetary Policy in Twelve Industrial Countries, Boston: Federal Reserve Bank of Boston.

References 203 Goodhart, C. (1973b), Analysis of the determination of the stock of money, in M. Parkin and A. Nobay (eds), Essays in Modern Economics, London: Longman (reprinted in Goodhart (1984), to which version page numbers refer). Goodhart, C. (1984), Monetary Theory and Practice, London: Macmillan. Goodhart, C. (1991), The conduct of monetary policy, in C. Green and D. Llewellyn (eds), Surveys in Monetary Economics, Vol. 1: Monetary Theory and Policy, Oxford: Blackwell. Goodhart, C. (1993), ERM and EMU, LSE Financial Markets Group Special Paper 58, London, LSE, November. Goodhart, C. (1994), What should central banks do? What should be their macroeconomic objectives and operations?, Economic Journal, 104, 1424–36. Goodhart, C. (1997), Why do the monetary authorities smooth interest rates?, in S. Collignon (ed.), European Monetary Policy, London: Pinter. Goodhart, C. (1999a), Central bankers and uncertainty, Bank of England Quarterly Bulletin, 39, 102–114. Goodhart, C. (1999b), Recent developments in central banking: some special features of the Monetary Policy Committee and of the European System of Central Banks, LSE Financial Markets Group Special Paper 0118. Goodhart, C. (2000), The role of the Monetary Policy Committee, in L. Mahadeva and G. Sterne (eds), Monetary Policy Frameworks in a Global Context, London: Routledge. Goodhart, C. (2001a), The organisational structure of banking supervision, in Brealey, R., Clark, A., Goodhart, C., Healey. J., Hoggarth, G., Llewellyn, D., Shu, C., Sinclair, P. and Soussa, F., Financial Stability and Central Banks, London: Routledge. Goodhart, C. (2001b), The inflation forecast, National Institute Economic Review, no. 175, 59–66. Goodhart, C. (2001c), Monetary transmission lags and the formulation of the policy decisions on interest rates, Federal Reserve Bank of St Louis Review, 83(4), 165–81. Goodhart, C. and Schoenmaker, D. (1995), Should the functions of monetary policy and bank supervision be separated?, Oxford Economic Papers, 39, 75–89. Gowland, D. (1982), Controlling the Money Supply, London: Croom Helm. Grilli, V., Masciandaro, D. and Tabellini, G. (1991), Political and monetary institutions and public financial policies in the industrial countries, Economic Policy, no. 13, 341–92. Gros, D. and Thygesen, N. (1992), European Monetary Integration, London: Longman. Haldane, A. (ed.) (1995), Targeting Inflation. London: Bank of England. Hasse, R. (1993), German–German monetary union: main options, costs and repercussions, in A. Ghanie Ghaussy and W. Sch¨afer (eds), The Economics of German Unification, London: Routledge. Henry, B., Satchi, M. and Vines, D. (2001), How square is the policy frontier?, Discussion Paper no. 2001.1, Centre for International Macroeconomics, Oxford University. Hester, D. and Tobin, J. (eds) (1967), Financial Markets and Economic Activity, New York: John Wiley. House of Lords (2001), Report of the Select Committee on the Monetary Policy Committee of the Bank of England, London: House of Lords. Jackson, P. (1991), Public expenditure, in M. Artis and D. Cobham (eds), Labour’s Economic Policies, 1974–79, Manchester: Manchester University Press. Johnson, H. (1970), Recent developments in monetary theory—a commentary, in D. Croome and H. Johnson (eds), Money in Britain, 1959–69, Oxford: Oxford University Press. Keegan, W. (1984), Mrs Thatcher’s Economic Experiment, London: Penguin. Keegan, W. (1989), Mr Lawson’s Gamble, London: Hodder & Stoughton. King, M. (1994), Monetary policy in the UK, Fiscal Studies, 15(3), 109–28. King, M. (1995), Credibility and monetary policy: theory and evidence, Bank of England Quarterly Bulletin, 35, 84–91. King, M. (1997a), Changes in UK monetary policy: rules and discretion in practice, Journal of Monetary Economics, 39, 81–97. King, M. (1997b), The inflation target five years on, Bank of England Quarterly Bulletin, 37, 434–42. Kohn, D. (2001), The Kohn report on MPC procedures, Bank of England Quarterly Bulletin, 41, 35–49. Krugman, P. (1979), A model of balance-of-payments crises, Journal of Money, Credit and Banking, 11, 311–25. Kydland, F. and Prescott, E. (1977), Rules rather than discretion: the inconsistency of optimal plans, Journal of Political Economy, 85, 473–92.

204 References Kynaston, D. (1995), The Bank of England and the Government, in R. Roberts and D. Kynaston (eds), The Bank of England: Money, Power and Influence 1694–1994, Oxford: Oxford University Press. Lane, T. (1985), The rationale for money-supply targets: a survey, Manchester School, 53, 179–207. Lane, T., Griffiths, M. and Prati, A. (1995), Can inflation targets help make monetary policy credible?, Finance and Development, 32(4), 20–23. Lawson, N. (1992), The View from No. 11, London: Bantam Press. Leiderman, L. and Svensson, L. (eds) (1995), Inflation Targets, London: Centre for Economic Policy Research. Llewellyn, D. (1991), Structural change in the British financial system, in C. Green and D. Llewellyn (eds), Surveys in Monetary Economics, Vol. 2: Financial Markets and Institutions, Oxford: Blackwell. Lowe, P. and Ellis, L. (1997), The smoothing of official interest rates, in P. Lowe (ed.), Monetary Policy and Inflation Targeting, Sydney: Reserve Bank of Australia. Lucas, R. (1972), Expectations and the neutrality of money, Journal of Economic Theory, 4, 103–24. Marsh, D. (1993), Faultlines show in Franco-German unity, Financial Times, 23 December. Marsh, D. (1994), Germany and Europe: The Crisis of Unity, London: Heinemann. Martin, B. and Salmon, C. (1999), Should uncertain monetary policy-makers do less?, Bank of England Working Paper no. 99. Masciandaro, D. and Spinelli, F. (1994), Central bank independence: institutional determinants, rankings and central bankers’ views, Scottish Journal of Political Economy, 41, 434–43. Masera, R. (1994), Single market, exchange rates and monetary unification, in A. Steinherr (ed.), 30 Years of European Monetary Integration from the Werner Plan to EMU, Harlow: Longman. Masson, P. (1995a), Modelling credibility: an application to the United Kingdom, in C. Bordes, E. Girardin, and J. M´elitz (eds), European Currency Crises and After, Manchester: Manchester University Press. Masson, P. (1995b), Gaining and losing credibility: the case of the United Kingdom, Economic Journal, 105, 571–82. Mayer, C. (1986), The assessment: financial innovation: curse or blessing?, Oxford Review of Economic Policy, 2(4), i–xix. McCallum, B. (1988). Robustness properties of a rule for monetary policy, Carnegie–Rochester Conference Series on Public Policy, no. 29, 173–204. Meade, J. (1978), The meaning of ‘internal balance’, Economic Journal, 88, 423–35. Meek, P. (ed.) (1983), Central Bank Views on Monetary Targeting, New York: Federal Reserve Bank of New York. M´elitz, J. (1991), German reunification and exchange rate policy in the EMS, Discussion Paper no. 520, Centre for Economic Policy Research, February. M´elitz, J. (1994), French monetary policy and recent speculative attacks on the franc, in D. Cobham (ed.), European Monetary Upheavals, Manchester: Manchester University Press. Meyer, L. (2001), Commentary on Goodhart (2001c), Federal Reserve Bank of St Louis Review, 83(4), 183–6. Micossi, S. and Padoan, P.C. (1994), Italy in the EMS: after crisis, salvation?, in C. Johnson and S. Collignon (eds), The Monetary Economics of Europe: Causes of the EMS Crisis, London: Pinter. Miles, D. (1992), Housing and the wider economy in the short and long run, National Institute Economic Review, no. 139, 64–78. Miles, D. (1994), Assessing the impact of financial liberalisation and income shocks on consumption, chapter 7 of Housing, Financial Markets and the Wider Economy, Chichester: John Wiley. Muellbauer, J. and Murphy, A. (1989), Why has UK personal saving collapsed?, London: Credit Suisse First Boston. Mussa, M. (1986), Nominal exchange rate regimes and the behaviour of the real exchange rate, Carnegie– Rochester Conference Series on Public Policy, no. 25, 117–24. Nelson, E. (2000), UK monetary policy 1972–97: a guide using Taylor rules, Bank of England Working Paper no. 120. Nelson, E. and Nikolov, K. (2001), UK inflation in the 1970s and 1980s: the role of output gap mismeasurement, Bank of England Working Paper no. 148. Newlyn, W. (1962), Theory of Money, Oxford: Clarendon Press. Nickell, S. (2001), Has UK labour market performance changed?, Bank of England Quarterly Bulletin, 41, 340–50.

References 205 Niehans, J. (1981), The appreciation of sterling—causes, effects, policies, SSRC Money Study Group Discussion Paper (February 1981). Norman, P. and Barber, L. (1992), The monetary tragedy that led to currency chaos, Financial Times, 11 December. Obstfeld, M. (1986), Rational and self-fulfilling balance-of-payments crises, American Economic Review, 76, 72–81. Obstfeld, M. (2001), International macroeconomics: beyond the Mundell–Fleming model, National Bureau of Economic Research Working Paper no. 8369. OECD (1991), Economic Survey: the United Kingdom, 1990/91, Paris: OECD. OECD (1993), Economic Survey: the United Kingdom, 1992/93. Paris: OECD. Ozkan, G. and Sutherland, A. (1994), A model of the ERM crisis, Discussion Paper no. 879, Centre for Economic Policy Research, London, January. Peersman, G. and Smets, F. (1999), The Taylor rule: a useful monetary policy benchmark for the euro area?, International Finance, 2(1), 85–116. Phelps, E. (1967), Phillips curves, expectations of inflation and optimal unemployment over time, Economica, 76, 678–711. Plender, J. and Wallace, P. (1985), The Square Mile, London: Hutchinson. Poole, W. (1970), Optimal choice of monetary policy instruments in a simple stochastic macro model, Quarterly Journal of Economics, 84, 197–216. Rogoff, K. (1985), The optimal degree of commitment to an intermediate monetary target, Quarterly Journal of Economics, 100, 1169–89. Roll, E., Begg, D., Corby, B., Daintith, T., Gleske, L., Goodhart, C., Lagayette, P., Middleton, P., Monti, M., Portes, R., Walker, D. and Wyplosz, C. (1993), Independent and Accountable: A New Mandate for the Bank of England. London: Centre for Economic Policy Research. Rose, A. and Svensson, L. (1994), European exchange rate credibility before the fall, European Economic Review, 38, 1185–216. Rowan, D. (1968), Output, Inflation and Growth, London: Macmillan. Rudebusch, G. (1995), Federal Reserve interest rate targeting, rational expectations and the term structure, Journal of Monetary Economics, 35, 245–74. Rudebusch, G. (2001a), Is the Fed too timid? Monetary policy in an uncertain world, Review of Economics and Statistics, 83, 203–17. Rudebusch, G. (2001b), Term structure evidence on interest rate smoothing and monetary policy inertia, Working Papers in Applied Economic Theory 2001–02, Federal Reserve Bank of San Francisco. Sack, B. (2000), Does the Fed act gradually? a VAR analysis, Journal of Monetary Economics, 46, 229–56. Sack, B. and Wieland, V. (1999), Interest-rate smoothing and optimal monetary policy: a review of recent empirical evidence, Finance and Economics Discussion Paper 1999–39, Federal Reserve Board. Sargent, J. R. (1991) Deregulation, debt and downturn in the UK economy, National Institute Economic Review, no. 137, 75–88. Sayers, R. (1964), Modern Banking, 6th edn, Oxford: Clarendon Press. Sheppard, D. (1968), Changes in the money supply in the United Kingdom, 1954 to 1964: a comment, Economica, 35, 279–302. Sheppard, D. (1971), The Growth and Role of UK Financial Institutions 1880–1962, London: Methuen. Sherman, H. and Kaen, F. (1994), The behaviour and thinking of the Bundesbank, in D. Cobham (ed.), (1994), European Monetary Upheavals, Manchester: Manchester University Press. Simons, H. (1936), Rules versus authorities in monetary policy, Journal of Political Economy, 44, 1–30. Smith, D. (1987), The Rise and Fall of Monetarism, London: Penguin. Smith, D. (1992), From Boom to Bust, London: Penguin. Spencer, P. (1986), Financial Innovation, Efficiency and Disequilibrium, Oxford: Clarendon Press. Steinherr, A. (1994), Has the case for EMU weakened since September 1992?, in D. Cobham (ed.), (1994), European Monetary Upheavals, Manchester: Manchester University Press. Stephens, P. (1996), Politics and the Pound. London: Macmillan. Sumner, M. (1980), The operation of monetary targets, Carnegie–Rochester Conference on Public Policy, no. 13, 91–130.

206 References Sutherland, A. (1995), Currency crisis models: bridging the gap between old and new approaches, in C. Bordes, E. Girardin, and J. M´elitz (eds), European Currency Crises and After, Manchester: Manchester University Press. Svensson, L. (1991), The simplest test of target zone credibility, IMF Staff Papers, 38, 655–65. Svensson, L. (1997), Inflation forecast targeting: implementing and monitoring inflation targets, European Economic Review, 41, 1111–46. Svensson, L. (1999), Inflation targeting as a monetary policy rule, Journal of Monetary Economics, 43, 607–54. Svensson, L. (2000), Open economy inflation targeting, Journal of International Economics, 50, 155–83. Svensson, L. (2001), The inflation forecast and the loss function, Princeton University and Stockholm University, November 2001. Taylor, J. (1993), Discretion versus policy rules in practice, Carnegie–Rochester Conference Series on Public Policy, no. 39, 195–214. Taylor, J. (1998), Applying academic research on monetary policy rules: an exercise in translational economics, Manchester School, 66, conference supplement, 1–16. Taylor, J. (ed.) (1999a), Monetary Policy Rules, Chicago: University of Chicago Press. Taylor, J. (1999b), The robustness and efficiency of monetary policy rules as guidelines for interest rate setting by the European central bank, Journal of Monetary Economics, 43, 655–79. Thatcher, M. (1993), The Downing Street Years, London: HarperCollins. Tietmeyer, H. (1994), On the architecture of EMU, in A. Steinherr (ed.), 30 Years of European Monetary Integration from the Werner Plan to EMU, London: Longman. Tobin, J. (1963), Commercial banks as creators of money, in D. Carson (ed.), Banking and Monetary Studies, Homewood, IL: Irwin; reprinted in Hester and Tobin (1967). Treasury (1990–91), Measures of real exchange rates and competitiveness, Treasury Bulletin, 1(1). Treasury (1992), Recent developments in UK economic policy, Treasury Bulletin, 3(3), 1–10. Treasury and Civil Service Committee (1979–80), Monetary Policy: Minutes of Evidence, House of Commons paper HC 679-vi, London: HMSO. Turnbull, A. and King, M. (1999), Review of the Revisions to the Average Earnings Index, London: House of Commons. Vaubel, R. (1994), The breakdown of the ERM and the future of the EMU: explanations, predictions and simulations from a public choice perspective, in D. Cobham (ed.), (1994), European Monetary Upheavals, Manchester: Manchester University Press. Vickers, J. (1998), Inflation targeting in practice: the UK experience, Bank of England Quarterly Bulletin, 38, 368–75. Vines, D. (1994), Discussion of Steinherr (1994), in D. Cobham (ed.), (1994), European Monetary Upheavals, Manchester: Manchester University Press. von Hagen, J. and Strauch, R. (1999), Tumbling giant: Germany’s experience with the Maastricht fiscal criteria, in D. Cobham and G. Zis (eds), From EMS to EMU, London: Macmillan. Wadhwani, S. (1999), Sterling’s puzzling behaviour, Bank of England Quarterly Bulletin, 39, 416–27. Wadhwani, S. (2000), The exchange rate and the MPC: what can we do?, Bank of England Quarterly Bulletin, 40, 297–306. Wadhwani, S. (2001a), Some reflections on the MPC, Bank of England Quarterly Bulletin, 41, 351–6. Wadhwani, S. (2001b), Do we have a new economy?, Bank of England Quarterly Bulletin, 41, 495–510. Walters, A. (1986), Britain’s Economic Renaissance, Oxford: Oxford University Press. Walters, A. (1990), Sterling in Danger, London: Fontana. Williams, N. (1991), Is sterling overvalued within the ERM?, CBI Economic Situation Report, January. Williamson, J. (1985), The Exchange Rate System, 2nd edn, Washington, DC: Institute for International Economics. Woodford, M. (1999), Optimal monetary policy inertia, Manchester School, 67 (supplement), 1–35. Wren-Lewis, S., Westaway, P., Soteri, S. and Barrell, R. (1991), Evaluating the UK’s choice of entry rate into the ERM, Manchester School, 59 (supplement), 1–22. Wright, S. (1992), Equilibrium real exchange rates, Manchester School, 60 (supplement), 63–84. Wyplosz, C. (1991), On the real exchange rate effect of German unification, Weltwirtschaftliches Archiv, 127, 1–17.

Index accountability, 103, 104, 108, 115, 172 Adam, C., 10, 67–70, 91–2, 105, 106, 117 Alesina, A., 84 Allsopp, C., 12, 75, 76, 113, 139, 140 Almeida, A., 94 Aoki, K., 57, 167 Artis, M., 4, 12, 40, 42, 55, 58, 76, 138, 140 asset price shocks, 2, 6, 8, 193 Atkinson, F., 44 Avesani, R., 88 Bacchetta, P., 82 balance of payments crises, see exchange rate crises Ball, L., 9, 10 Balls, E., 111 bank lending to private sector, 21, 22, 23–4, 25, 32–6, 39, 42–7, 54–6, 65, 66 Bank of England fan chart, 98 Governor, 15, 16, 19, 27, 39–40, 56–7, 65, 66, 73–4, 75, 76, 94, 95, 99–103, 107, 108, 110, 114, 116, 179, 181, 188 inflation forecasts, 94–5, 98, 103, 115–16, 185, 186–7, 188 Loughborough Lecture, 39, 40, 55, 124 relationship with Treasury, 15, 25, 93, 94–5, 107, 148 transparency, 94–5 Bank for International Settlements, 37, 162 Barber, L., 79 Barisone, G., 126 Barrell, R., 75 Batini, N., 114 Bayoumi, T., 57 Bean, C., 3, 4, 111, 113, 114, 142, 154, 166 Begg, D., 81, 87 Bell, G., 21 Bell, M., 84 Berman, L., 21 Bernanke, B., 94, 195 Black Wednesday, 15, 19, 73, 78, 90, 92, 93, 94, 148, 149, 150, 154, 171 bond yield differentials, 83, 97, 104, 112–13

Bowen, A., 93 Branson, W., 81–2, 84, 86, 89 Breedon, F., 96, 97 Briault, C., 104 Brittan, S., 4, 61, 62 Britton, A., 75, 139, 142, 154 Bruce-Gardyne, J., 45, 48 Brunner, K., 22 Buckle, M., 55 Budd, A., 107 building societies, 24, 28, 37, 38, 39, 45, 46, 47, 55, 56, 78 Buiter, W., 85, 88, 89, 90, 137, 142 Bundesbank, 62, 78, 79, 84–5, 87, 89, 90, 148 Burns, T., 40, 44, 45 Cagan, P., 20 Caplin, A., 166 Carruth, A., 57 Cecchetti, S., 2, 153 central bank independence, 3, 20, 107, 108–11, 191, 194–5, 197 Chick, V., 21 Clarida, R., 9, 10, 67, 70, 161 Clarke, K., 19, 93, 94, 99–106, 117 Cobham, D., 10, 12, 23, 24, 31, 67–70, 82, 91–2, 104, 105, 106, 108–11, 117, 138, 140 Coleby, A., 32, 33 competitiveness, 59, 66, 80, 82–3, 86–7, 139, 141, 142, 143 Congdon, T., 54, 55 consumers’ expenditure, 56, 57, 64, 103, 182, 188 Coppock, D., 21 Corsetti, G., 85, 88, 89, 90 Cosci, S., 108–11 Courakis, A., 29 corset, see Supplementary Special Deposit Scheme Cramp, A., 21 credibility, 3, 4, 6–7, 49, 74, 77, 88, 94–5, 96–8, 102, 103, 104–5, 166, 174, 175, 176, 179, 181–2, 191, 195 credit counterparts framework, 7, 22–3, 32–7, 42–7

208

Index

Cross Report, 37 Cukierman, A., 108–10 Currie, D., 4, 76 Davies, S., 77, 78 Davis, R., 16 De Grauwe, P., 82, 87 Deacon, M., 96 Debelle, G., 110 Debt Management Office, 107 Derry, A., 96 Deutschemark- (DM-) shadowing, 14–15, 18, 53, 58–60, 63–5, 67, 70–71, 145–6 Dicks, G., 76 disintermediation, 24 domestic credit expansion (DCE), 13, 14, 21, 27, 43, 50 Dornbusch model, 82 Dow, C., 24 Drazen, A., 83, 86 Driver, R., 126 Eichengreen, B., 82, 83, 84, 87 Ellis, L., 162, 166, 182 Estrella, A., 167 European Monetary System (EMS), see Exchange Rate Mechanism European Monetary Union (EMU), 74, 84, 87, 196 Exchange Rate Mechanism (ERM), 4, 15, 18–19, 53, 73–92, 125, 126, 191–4, 195 1992–93 crises in, 80–90 exchange controls, 24, 73 exchange rate as shock absorber, 137, 155–6, 159 exchange rate as transmission mechanism, 137, 155–6, 159 exchange rate crises, 8, 13 1992–93 ERM crises, 80–90 theories of, 81–3, 86 exchange rate determination, 5–6, 81–3 exchange rate targets, 1, 3–8, 19–20, 71, 77, 92, 141, 145–6, 148, 191, 192–4, 195 exchange rate volatility, 8, 144, 145, 147, 150, 160 expenditure shocks, 2, 5, 8, 192–3 Federal Reserve, 9, 57, 64, 161 FEER, see fundamental equilibrium exchange rate Fforde, J., 16, 22–3, 27, 193 Filc, W., 89 financial innovation, 17, 37–49, 192 financial liberalisation, 56–7, 66 see also financial innovation Financial Services Authority, 107, 110

fiscal policy, 57–8, 81–2, 85, 119, 120, 122, 123–4, 140, 142 see also public sector borrowing requirement Fischer, S., 1, 110 Fisher, P., 23 Flood, R., 83, 137, 160 Foot, M., 16 foreign asset market shocks, 6, 8, 193 foreign exchange market intervention, 4, 7, 140, 141, 142, 145, 146, 147, 151 foreign exchange risk premium, 4, 84, 193 foreign expenditure shocks, 6, 8, 192–3 Forsyth, P., 62, 142 France, 31, 79, 82, 84, 85, 86, 87, 89, 90, 97, 112–13, 162 Freeman, C., 186 Friedman, B., 2, 4 Friedman, M., 1, 2, 20 monetary rule of, 1, 2, 16 Funabashi, Y., 64, 145 fundamental equilibrium exchange rate (FEER), 59, 125 Gal´ı, J., 9, 10, 67, 70, 161 Gallo, G., 88 Garber, P., 83 German Economic, Monetary and Social Union (GEMSU), 15, 80, 81–2, 85, 86, 87, 89 Germany, 83, 84–5, 97, 108, 112–13, 128, 148, 157, 162, 163 see also Bundesbank; German Economic, Monetary and Social Union Gertler, M., 9, 10, 67, 70, 161 Giavazzi, F., 58, 81, 87 Gibson, N., 21 gilt-edged securities, 21, 22, 24–5, 32–6, 44–7, 173 index-linked, 24 Girardin, E., 10, 67–70, 84, 91–2, 105, 106, 117 Goodfriend, M., 166 Goodhart, C., 7, 9, 21, 22–3, 25, 55, 88, 94, 107, 113, 115, 116, 161, 162, 166, 167, 168, 169, 187, 189 Gowland, D., 24 Grilli, V., 84, 109–10 Gros, D., 81 Haldane, A., 94, 104, 114 Hall, S., 44 Hasse, R., 81 Henley, A., 57 Henry, B., 115 Hester, D., 20 Howe, G., 73, 95

Index inflation expectations, 3, 4, 6–8, 16, 96–8, 111–13, 140, 193–4 inflation targets, 1, 3–8, 19, 20, 93–4, 101, 102, 107–8, 111–18, 148, 150, 153, 191, 193, 194, 195, 197 interest rate differentials, 77, 83, 97, 139–40, 158 see also bond yield differentials interest rate smoothing, 161–89, 195–6 intermediate targets, 2, 3–8, 191–5 Jackson, P., 13 Japan, 128, 162, 163 Jenkinson, N., 113 Johnson, H., 20 Julius, D., 182 Kaen, F., 88 Kay, J., 62, 142 Keegan, W., 18, 58, 61, 62, 63–4, 65, 142, 145 King, M., 93, 94, 95, 96, 103, 104, 106, 114, 186, 188 Kohn, D., 111, 113, 115, 116, 117, 187 Krugman, P., 80, 88 Kydland, F., 1, 2, 16 Kynaston, D., 16 Lamont, N., 19, 76, 78, 79, 90, 95, 111 Lane, T., 16, 94, 185 Laubach, T., 94, 195 Lawson, N., 14, 18, 56, 58, 61, 62, 63, 64, 65, 66, 68, 73, 74, 76, 77, 95, 111, 144, 145, 146, 147 Leahy, J., 166 Leiderman, L., 94 Lewis, M., 12, 40, 42, 55 Llewellyn, D., 37 Louvre Accord, 18, 62, 63–4, 65, 145 Lowe, P., 162, 166, 182 Lucas, R., 2 Maastricht Treaty, 78, 79, 84, 87, 107, 108, 110 Major, J., 18, 73, 74, 75, 79, 95, 148 Marsh, D., 75, 88 Martin, B., 166, 169, 172 Masciandaro, D., 109–11 Masera, R., 82 Masson, P., 83, 86, 88 Mattesini, F., 108–11 Mayer, C., 37 McCallum, B., 9 Meade, J., 4 Medium Term Financial Strategy (MTFS), 14, 15, 16, 17, 29, 35, 44–6, 142, 143 Meek, P., 16, 29 M´elitz, J., 81, 84–5, 86, 88, 89, 90 Meltzer, A., 22

209

Meyer, L., 116 Micossi, S., 82 Miles, D., 56, 57 Miller, M., 142 Mishkin, F., 94, 167, 195 monetary base, 7, 8, 9, 18, 51, 65, 123 monetary conditions index, 10, 157 monetary instruments, 23–5 Monetary Policy Committee, 19, 25, 27, 70, 91–2, 105, 107–118, 150–3, 172–3, 174–8, 180, 181–9 monetary targets, 1, 2, 3–8, 15–18, 19, 25, 27–51, 53, 54, 61, 73, 138–9, 140–142, 143, 144, 191, 192–5 money, demand for, 1 money demand shocks, 2, 5, 8, 192 money supply, 2 determination of, 20–3 endogeneity/exogeneity, 28, 42, 55, 71 growth of, 54–5, 101, 119, 120, 122, 123, 124 see also monetary targets measures of, 7, 8, 14, 17, 27–8 monthly monetary meetings, 19, 94, 99–101, 103, 149–50, 173, 179–80, 181, 184 Muellbauer, J., 56 Mundell-Fleming model, 82 Murphy, A., 56 Mussa, M., 124 National Savings, 24, 32, 33, 34, 35, 78, 79, 148 Nelson, E., 10, 56, 67 Newlyn, W., 21 Nickell, S., 193 Niehans, J., 142, 154 Nikolov, K., 56 nominal income targets, 1, 3–8 Norman, P., 79 North Sea oil, 14, 62, 159 Obstfeld, M., 80, 83, 125 OECD, 75, 95 oil price, 11, 14, 58, 62–3, 142, 144, 145, 159 overfunding, 18, 25, 35, 36, 45, 46, 47, 49, 54, 55, 61 Ozkan, G., 83, 84, 89 Padoan, P., 82 Pain, N., 75 Papadopoulos, A., 104, 106 Peersman, G., 167 Pesenti, P., 85, 88, 89, 90 Phelps, E., 2 Plender, J., 38 P¨ohl, K., 75 Poole, W., 1–2, 4, 192

210

Index

portfolio shocks, 2, 6, 193 Posen, A., 94, 195 private sector liquidity (PSL), 17, 28, 50 Prescott, E., 1, 2, 16 Proudman, J., 57 public sector borrowing requirement (PSBR), 16, 17, 21, 22, 24, 25, 32–6, 43–7, 57, 62, 123–4 real exchange rates, 124–7, 138–9, 140, 142, 143, 144, 145, 146, 147, 149, 150 real interest rates, 138, 139, 141, 142, 144, 145, 146, 147, 149, 150 re-intermediation, 31, 44 relative unit labour costs, 59, 124–6, 150–1 Ridley, N., 90 Rogoff, K., 103 Roll, E., 95, 111 Rose, A., 87, 137, 160 Rowan, D., 21 Rudebusch, G., 161, 166, 167 rules versus discretion, 1, 2–3 Sack, B., 161, 166, 167, 169 Salmon, C., 166, 169, 172 Salmon, M., 88 Sargent, J., 56, 57 Satchi, M., 115 Saville, I., 24 saving ratio (household, personal), 57, 66 Sayers, R., 21 Schoenmaker, D., 107 Schwartz, A., 20 Sedgwick, P., 188 Serre, J-M., 31, 108–11 Sheppard, D., 21 Sherman, H., 88 Simons, H., 2 Smets, F., 167 Smith, D., 18, 61, 64, 75, 139, 142, 143, 145 Spaventa, L., 58 speculative attacks, 80–81, 83–5, 87–8 Spencer, P., 24, 142, 154 Spinelli, F., 111 stabilisation, 2, 4–6, 192–3 Steinherr, A., 86 Stephens, P., 18, 75, 77, 78, 79, 80, 95, 100, 148 sterling crises, 13, 32, 46, 138–40, 143, 144, 157, 158–9 sterling misalignments, 14, 49, 59, 75–6, 106, 118, 124–7, 138, 140, 142, 144, 145–6, 149, 150–5

stock market, 57, 64, 173 Strauch, R., 85 structural change, see financial innovation Sumner, M., 16 Supplementary Special Deposit Scheme, 11, 23–4, 25, 31, 32, 34, 44 supply shock, 6, 8, 193 Sutherland, A., 80, 83, 84, 89 Svensson, L., 9, 87, 88, 94, 116, 166 Tabellini, G. 109–10 targets and instruments, 1–2 Taylor, J., 1, 9 Taylor rules, 1, 9–10, 67–70, 91–2, 105, 117, 161 Thatcher, M., 14, 18, 58, 61, 62, 64, 65, 66, 73, 74, 75, 76, 95, 146 Thompson, J., 55 Thygesen, N., 81 Tietmeyer, H., 82 time-inconsistency, 2–3, 4, 77, 95, 103 Tobin, J., 20 Treasury, 11, 15, 32, 75, 76, 77, 79, 90, 93, 94–5 Treasury and Civil Service Committee, 29 Turnbull, A., 188 United States, 104, 128, 157, 162, 163, 195 Vaubel, R., 82 Vega, J., 23 velocity of money, 9, 17–18, 31, 39–42, 44, 47, 48, 49, 55 Vickers, J., 114 Vines, D., 86, 115 Vlieghe, G., 57 von Hagen, J., 85 Wadhwani, S., 111, 113, 114, 117, 153, 193 Wallace, P., 38 Walters, A., 57, 58, 61, 76, 143 Weale, M., 188 Wieland, V., 161, 167 Williams, N., 75 Williamson, J., 59, 75 Woodford, M., 167, 185, 186 Wren-Lewis, S., 59, 126 Wright, S., 75 Wyplosz, C., 81, 82, 83, 84, 87 Zis, G., 104, 106