Further Essays in Monetary Economics (Collected Works of Harry Johnson)

Harry G. Johnson

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781134623914
• Edition: 1
• Topic: Economics
• Subtopic: Economic Theory
• Index: Economics

A sequel to Essays in Monetary Economics, this book develops the ideas on domestic and international monetary issues, with reference to specific events and crises of the 1960s and 70s. These essays are distinguished by the author's expert grasp of the analytical techniques and contemporaneous policy problems of both domestic and international monetary economics.

Part 1

GENERAL MONETARY THEORY

Chapter 1

Recent Developments in Monetary Theory—a Commentary*

I. INTRODUCTION

The Radcliffe Committee initiated its investigations into the working of the British monetary system at a time when the intellectual environment could be characterized as the high tide of Keynesian scepticism about the importance of monetary policy and the relevance of monetary theory. By the time its Report was published, however, that tide had begun markedly to ebb, a fact which accounts for the unexpectedly harsh reception the Report received even from what might have been expected to be intellectually sympathetic quarters. In the ensuing ten years, the tide has set markedly in the opposite direction, towards emphasis on the importance of monetary policy and concern with the theory of money (as distinct from the theory of income and employment), to the point where contemporary controversy centres on the so-called ‘rise of monetarism’.

The purpose of this chapter is to survey developments in monetary theory since the Radcliffe Report, with particular emphasis on developments subsequent to two previous surveys of mine, written in 1961–62 and 1963.¹ The first of those surveys, designed to provide a comprehensive overview of the field for graduate students and non-specialist professional economists, organized the material presented within the broad analytical framework of demand for and supply of money. The second, an essay in personal interpretation unrestricted by the obligation of representation by population, took as its organizational focus six problems originating in Keynes’s General Theory, and emphasized the two themes of the application of capital theory to monetary theory and the trend towards dynamic analysis.

For the purposes of this chapter, the type of approach of the second survey referred to seems the more suitable. Accordingly, I will discuss recent developments under a series of topical headings chosen to represent themes considered to be of general interest to monetary economists. However, in comparison with the 1963 survey, emphasis is placed rather less on unifying strands of thought and rather more on current controversy. Also, certain themes are treated in exceptional detail, in view of their presumed interest to British readers – particularly the first two, the revival of monetarism and the rehabilitation of Keynes. The remaining topics are the fundamentals of monetary theory, the problems associated with financial intermediation, money in growth models, and the theory of inflation and economic policy.

II. THE REVIVAL OF THE QUANTITY THEORY AND THE RISE OF ‘MONETARISM’

As already remarked, the dominant feature of the post-Radcliffe era, in the American if not quite yet the British literature, has been the revival of the quantity theory of money and the rise of the associated ‘monetarist’ approach to economic policy – an approach which stresses the explanatory and controlling power of changes in the quantity of money, in contrast to the Keynesian emphasis on fiscal policy, and as regards monetary policy on credit and interest-rate policies. It may be useful, though it risks the accusation of putting an unwarranted motivational construction on a process of scientific development, to trace the stages in the intellectual revival of the quantity theory, which has been almost exclusively the work of Milton Friedman.

The Keynesian revolution left the quantity theory thoroughly discredited, on the grounds either that it was a mere tautology (the quantity equation), or that it ‘assumed full employment’ and that the velocity factor it emphasized was in fact highly unstable. The revival of a quantity theory that could claim to rival the Keynesian theory required a restatement of it that would free it from these objections and give it an empirical content. Such a restatement was provided by Milton Friedman’s classic article,¹ which redefined the quantity theory as a theory of the demand for money (or velocity) and not a theory of prices or output, and made the essence of the theory the existence of a stable functional relation between the quantity of real balances demanded and a limited number of independent variables, a relation deduced from capital theory. This version of the quantity theory, Friedman asserted, had been handed down through the ‘oral tradition’ of the University of Chicago. In fact, as Don Patinkin has recently shown conclusively,² it is to be found neither in the written tradition of Chicago – which on the contrary stressed the quantity equation and the cumulative instability of velocity – nor in the oral tradition of Chicago as Patinkin himself experienced it: ‘What Friedman has actually presented is an elegant exposition of the modern portfolio approach to the demand for money which … can only be seen as a continuation of the Keynesian theory of liquidity preference’ (p. 47). In recent writings, Friedman has ceased to refer to the Chicago oral tradition, and has admitted that his reformulation of the quantity theory was ‘much influenced by the Keynesian liquidity analysis’.³

Redefinition of the quantity theory as hypothesizing a stable demand function for money not only gave it an empirical content subject to testing,⁴ but facilitated the interpretation by researchers of good statistical results as evidence in favour of the quantity theory and against the rival ‘income–expenditure’ theory (the Chicago term for the prevalent version of Keynesian economics). The next stage was to devise a set of tests of the rival theories against one another; this was the subject of the Friedman-Meiselman study for the Commission on Money and Credit of ‘The Relative Stability of Monetary Velocity and the Investment Multiplier in the United States, 1898–1958’.¹

The tests in question rested on some fundamental – and debatable – methodological principles; and the failure of the critics to understand these principles, as well as to appreciate the depth of the intellectual effort put into the tests, made their criticisms and attempted refutations less powerful and persuasive than they might have been. The crucial principle is that the test of good theory is its ability to predict something large from something small, by means of a simple and stable theoretical relationship; hence the essence of the quantity theory was specified to be the velocity function relating income to money, and the essence of the income–expenditure theory was specified to be the multiplier relationship relating income to autonomous expenditure (for the purposes of the tests, these relationships were redefined in terms of consumption rather than income, to avoid pseudo-correlation). This principle is in sharp contrast to the more common view that the purpose of theory in this context is to lay out the full structure of a general-equilibrium model in the detail necessary to produce an adequately good statistical ‘fit’. A second principle is that behavioural relationships should be invariant to institutional and historical change; hence the Friedman-Meiselman emphasis on a long run of data. A third principle, whose practical application has given rise to legitimate criticism, is that since Keynesian theory does not specify exactly what is to be treated as ‘autonomous’ and what as ‘induced’ in an economy with governmental and foreign trade sectors, the classification must be effected by statistical tests of independence and interdependence.

According to the Friedman-Meiselman tests, the quantity theory consistently out-performed the Keynesian theory, with the exception of the 1930s sub-period. A conscientious, or nonchalant, Keynesian might well have interpreted these results as confirming the master’s insight, insofar as the tests could be considered relevant at all. Instead a number were provoked into attempting to disprove the findings; and, as mentioned, their efforts were generally vitiated in their impact by violation of one or another of the rules of the game as laid down by Friedman and Meiselman.¹

As restated by Friedman, the quantity theory still laboured under the handicap of two potentially powerful criticisms. The first was the long-standing traditional criticism of the quantity theory, that the theory is irrelevant because the quantity of money supplied responds passively to the demand for it – the ‘Banking School’ position which remains strong in popular thinking on monetary policy. This criticism was quelled by the publication of the long-awaited, monumental volume by Friedman and Schwartz on the monetary history of the United States,² and the companion volume by Cagan on the supply of money in the United States.³ These works demonstrated both the independent determination of the supply of money, and the significant influence of monetary changes on U.S. economic history. The second criticism stemmed from the strongly-held belief that the great depression of 1929 and after was the consequence of the collapse of the willingness to invest and proved conclusively the inability of monetary policy to remedy mass unemployment. The Friedman-Schwartz volume demonstrated as conclusively as possible the causal role played by rapid and substantial monetary contraction in the depression of the 1930s, and thus paved the way for a dismissal of the Keynesian analysis as based on a misinterpretation of the facts of experience.¹ There remains, however, acute controversy over the interpretation of the emergence of large holdings of excess reserves by the American banking system in the latter part of the 1930s.

While, as already mentioned, Friedman’s restatement of the quantity theory of money should probably be interpreted as an appropriation of portfolio-balance analysis on Keynesian lines for use against those Keynesians who have neglected the monetary side of Keynes’s theory in favour of the income–expenditure side, there is one important difference between the Friedman (quantity-theory) approach and the Keynesian approach to that analysis which is of considerable importance both theoretically and practically. This difference is that the restated quantity theory introduces explicitly, and emphasizes, expected changes in the price level as an element in the cost of holding money and other assets fixed as to both capital value and yield in money terms, whereas Keynesian portfolio-balance theory almost invariably starts from the assumption of an actual or expected stable price level (though this asumption may subsequently be modified).² The assumption in question has the great theoretical advantage of endowing money with an absolutely certain yield of zero per cent, and hence making it a fixed point of reference for portfolio choices; but this advantage is bought at the cost of giving money in the portfolio attributes of safety which in general it does not possess. Moreover, from the standpoint of application of monetary theory to the interpretation of actual events and policies, the assumption is likely to be consistently misleading, because it encourages practitioners of the approach to interpret changes in market interest rates on monetary assets as indicators of changes in monetary ease or tightness, without proper allowance for the effects on the relation between money and real rates of interest of changes in expected rates of inflation or deflation.

This difference is in an important sense the essence of the differentiation between the ‘monetarist’ and the alternative ‘Keynesian’ approach to problems of economic policy. The monetarist approach stresses the unreliability of money interest rate changes as economic indicators, owing to the influence on them of price expectations, and concentrates instead on changes in the money supply as a variable over which the monetary authority has control and whose meaning is theoretically clear. In addition, and more fundamentally, the monetarist approach rests on the assumption that velocity rather than the multiplier is the key relationship in the understanding of macro-economic developments in the economy. This was the point of the Friedman-Meiselman test already discussed. Subsequently, the focus of the controversy has shifted from autonomous expenditure versus money supply to fiscal policy versus monetary policy as the subject of empirical testing. In this connection, tests performed at the Federal Reserve Bank of St Louis at the initiative of the Bank’s staff have been advanced in support of the monetarist as against the Keynesian approach; but these tests too have been the subject of considerable criticism.¹

As mentioned above, Friedman’s restatement of the quantity theory obtained the immediate tactical advantage of freeing it from the Keynesian criticism of assuming an automatic tendency towards full employment in the economy, by making it a theory of the demand for money without commitment to the analysis of prices and employment. This advantage, however, has proved something of an embarrassment subsequently, given the success of the quantity-theory counter-attack on Keynesianism and the rise of the monetarist approach to economic policy, since it apparently leaves the quantity theorist with nothing to say about the relative impact of short-run variations in the money supply, and hence in aggregate demand, on money prices on the one hand and physical output on the other. (It may be noted that a similar problem arises for the Keynesian theory, under conditions of near full employment, which the Phillips curve analysis seeks to resolve but which it resolves rather unsatisfactorily – see below.) The obvious answer to this problem, in neo-quantity terms, lies in an application of expectations theory to the determination of the division of an increase in monetary demand between changes in money wages and prices and changes in employment and output; but thus far no satisfactory theory along these lines has been produced.¹

In concluding this section, a brief reference should be made to recent extensions of the monetarist approach to problems of balance-of-payments analysis and policy. As regards individual countries, prevailing theory emphasizes the balance of aggregate demand and aggregate supply capacity on the one hand, and the relation between domestic and foreign price levels on the other, as the key determinants of the balance of payments. A monetarist approach, on the other hand, as reflected in the new IMF-inspired emphasis of British economic policy on ‘Domestic Credit Expansion’, emphasizes the relation between the growth of domestic demand for money and the growth of supply intended by the monetary authority as the key determinant of international reserve gains or losses. As regards the international monetary system as a whole, the monetarist approach emphasizes the relation between the growth of total desired reserves and the growth of overall reserve supplies as determining the need for some countries to have deficits, and the relation between national growth of desired money balances and national expansion of domestic credit as determining which countries will have the necessary deficits.²

III. THE REHABILITATION OF KEYNES

The quantity-theory counter-revolution discussed in section II has been directed against the so-called ‘income–expenditure’ school, by which is meant those economists in the Keynesian tradition who have concentrated their analysis and policy prescriptions on the income–expenditure side of the Keynesian general-equilibrium apparatus. (This focus has been the dominant impact of the Keynesian revolution on governmental and other practical thinking on economic forecasting and policy-making.) There is, it should be remarked, nothing to prevent the absorption of the empirical evidence of a stable demand function for money into the corpus of the Keynesian general-equilibrium model – a stable demand function for money is in fact implicit in the liquidity-preference component of the standard Hicksian IS–LM diagram – other than the conditioned Keynesian reflex against the ‘quantity-theory’ label and the conditioned Keynesian belief that ‘money does not matter’ (or, at least, ‘does not matter much’).

‘Keynesian economics’ came very rapidly to be epitomized by the IS–LM diagram of the textbooks, in terms of which Keynesian underemployment equilibrium depends either on the rigidity of money wages or on the special case of the ‘liquidity trap’ (a perfectly interest-elastic liquidity-preference function) keeping the rate of interest above that consistent with full-employment equilibrium between saving and investment; and this special case was disposed...