Research Discussion Paper – RDP 31 The Short Run and Long Run Trade-offs between Inflation and Unemployment in Australia

Introduction

Does Australia have a long run trade-off between inflation and unemployment? In other words, if demand management policies were able to select and maintain a steady rate of inflation, would the level of unemployment that emerged be lower, the higher the chosen inflation rate? If such a trade-off exists, what are its most likely quantitative characteristics? Alternatively, is the Phelps-Friedman [8,1] accelerationist hypothesis a better characterisation of Australia's long run inflation-unemployment options. That is, is there some rate of unemployment, (usually termed the natural rate), not necessarily a constant rate over time but unrelated to the inflation rate, at which any steady rate of inflation is possible but below which inflation will persistently accelerate? If this accelerationist model better describes the Australian economy, what is (and has been)the natural unemployment rate? These are the questions dealt with in this paper.

Earlier empirical studies of inflation in the Australian economy have focussed on the short run trade-off between inflation and unemployment and have, with the exception of Nevile [7], left the implications of their findings for the long run trade-off largely unexplored. One of the major things done in this paper therefore is the exploration and comparison of the long run implications of these earlier studies. The studies examined are those by Hancock [2], Higgins [3] Jonson, Mahar and Thompson [4], Nevile [6,7] and Pitchford [9]. This is done in Part III of the paper. First however, (in Part II), it is necessary to provide some theoretical background so that the long run relationships can be derived and interpreted. In Part IV a simple model which draws heavily on the work of Jonson, Mahar and Thompson, and which is in effect an Australian version of the Lipsey-Phelps-Friedman model, is presented. This model, it is argued, is more compatible with the data than any of the earlier studies. Part V summarises the analysis and presents the conclusions.

The major conclusion is worth anticipating and is that Australia probably does not have a long run trade-off between inflation and unemployment and only has a weak short run trade-off. In other words, in the short run, inflation may be reduced by allowing unemployment to stand above its so called natural rate and it will accelerate if unemployment is held below that rate. There is only one rate of unemployment at which, inflation will be steady in the long run. What that rate of unemployment is is a more problematic question. Obtaining a firm answer to that question is hampered by the fact that the natural unemployment rate is probably not constant and is determined by a process which we do not at the moment understand. It is argued that the natural unemployment rate in Australia has probably (and paradoxically) decreased since the early 1960's and currently lies in the 1.5 to 2 per cent range. This is not, in any formal statistical sense, a maximum likelihood estimate and not based directly on regression equations. The latter give rise to higher estimates. Rather, it is an estimate based on a detailed consideration of all the available evidence and in particular, on the compatibility of alternative estimates with each of the post 1960 cycles. Because of the nature of the statistical analysis performed, it is not possible to place a formal confidence interval on the estimated natural unemployment rate. Further work, improving our understanding of the determinants of the (probably variable) natural unemployment rate and of the relationship between excess demand and unemployment will have to be undertaken before such a confidence interval can be established.