The Phillips Relation: A Theoretical Explanation-A Reply

Abstract

If we use Professor R. G. Lipsey's derivation of the Phillips relationship, and we observe this relationship to be weak empirically, the reasons must be in either (a) the underlying adjustment mechanism, or (b) the assumptions about the behaviour of frictional unemployment as aggregate demand increases; or in both. Under (b) it is possible to hypothesize sets of assumptions that will, given (a), generate a negative relationship between the rate of money-wage change and the level of unemployment. It is also possible, as we suggested in our paper, to generate other results by changing the assumptions. As Dr. Vanderkamp points out, the question at bottom is an empirical one. However, we cannot agree with him that if one accepts the underlying adjustment hypothesis relating wage change to excess demand the non-linear, negatively-sloped specification of the Phillips curve has ample justification (p. 183). The empirical evidence that he refers to concerns the over-all correlation between vacancies and unemployment (used as surrogates for the excess demand and excess supply of labour). We, however, argued that it was crucial to distinguish periods of excess demand from periods of excess supply. Particularly, for reasons that we outlined in our paper, we suspect that the Phillips relationship would be weaker where V> U. That this is the case for a range of countries in the post-war period we hope to show in a forthcoming paper.