Academic journal article History of Economics Review

Crank or Proto-Monetarist?: J.K. Gifford and the Cost-Push Inflation Fallacy

Academic journal article History of Economics Review

Crank or Proto-Monetarist?: J.K. Gifford and the Cost-Push Inflation Fallacy

Article excerpt

Abstract: This paper examines the career and contribution of J. K. Gifford (1899-1987), the Foundation Professor of Economics and first Head of the Department of Economics at the University of Queensland, and one of the first in Australia to write an introductory textbook. Gifford's publications were often poorly written and with few references. They focussed mainly on monetary theory and inflation and towards the end of his career concentrated on challenging the notion of a wage-price spiral. Much of his work on the 'cost-push fallacy' seems to have been based on a crude kind of monetarist thinking: governments were prone to allow monetary growth to sustain high profit levels that businesses enjoyed in an inflationary environment. However, his policy proposals were not those of the freemarket Right and focussed on safeguarding employees' interests by ensuring that their wages increased at the rate of inflation plus productivity growth, thereby limiting the scope for employers to benefit from inflation. Although he saw the money supply as exogenous and prone to be mismanaged by governments, be did not articulate a model of the demand for money or defend the stability of the velocity of circulation. His most important article, a brief paper in the Journal of Political Economy in 1968, came about from his objections to the original Phillips analysis, and argued that correlation does not establish causation. Precisely this argument could also be levelled against the monetarist thinking of Milton Friedman and it was not long before the paper's basic argument was used by Nicholas Kaldor in this way.

1 Introduction

Very few Australian economists have published in the Journal of Political Economy; down to 1968 there had been only four of them. (2) J.K. Gifford's 'Correlationism: A Virulent Disease in Economic Science', which appeared in the 'Miscellany' section of the September-October 1968 issue of the Chicago journal, was therefore a significant personal achievement. This five-page article, which was by far John King Gifford's best published work, was also his last, as the Research Professor and first Head of the Department of Economics at the University of Queensland retired in the following year. (3) 'A disease is spreading in economic writing', Gifford began, 'on the subject of inflation of the price level in times of rising output. The disease is the misuse of correlation measures to explain causal relations' (Gifford 1968, p. 1091). He cited three papers on aggregate wage determination as symptoms of the disease: one Australian (Hancock 1966), one using British data (the classic study by Phillips 1958) and one from the US (Eckstein and Wilson 1962). All three claimed that statistical evidence of correlation between money wage growth and unemployment (together with certain other variables) explained the rate of wage inflation in their respective three countries. This, Gifford objected, was fallacious. 'It would be quite inappropriate to suggest that a theory of causal relations could be confirmed by a single high coefficient of determination, obtained by means of a regression equation', which could establish only the 'degree of correspondence' between the variables under review (ibid., pp. 1091-2).

'Some statisticians and economists', Gifford continued, 'when they use the word "explain" in this technical sense, use apologetic quotation marks, but this does not excuse the misleading use of the word in economic science, where explanation usually means revealing relations of cause and effect' (ibid., p. 1093). Taking as an example the popular textbook by M.R. Spiegel (1961), Gifford criticised the way in which statisticians defined the 'coefficient of determination' as the ratio of the 'explained variation' to the 'total variation' in the dependent variable. It would be less misleading, he argued, if r were renamed the 'coefficient of quantitative relation', defined as the ratio of the 'related variation' to the total ('related' plus 'unrelated') variation. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.