The main point of the recent paper by Bernard Corry and David Laidlerl is that the shape of the Phillips curve depends on an arbitrary assumption regarding the level of frictional unemployment. Thus Lipsey's derivation of the negative slope of the Phillips curve under conditions of rising wages (that is positive excess demand for labour) amounts to no more than the assertion that of the two factors that can affect the level of frictional unemployment, the number of people changing jobs and the average time they spend between jobs, it is the latter which varies more with the excess demand for labour, hence causing the level of unemployment to fall as excess demand rises.2 The implication is that the frequently-used empirical specification of the Phillips curve with its negative slope throughout stands on shaky theoretical grounds. Another implication would be that the wide use of the unemployment rate as a cyclical indicator of excess capacity in the economy is subject to serious doubt since high unemployment might mean the existence either of slack or of a great deal of pressure of excess demand. This Comment is a two-fold attack on these conclusions. First, three general reasons are given for the negative slope of the Phillips curve, one of which relates to frictional unemployment. Second, I discuss the empirical validity of the arguments which suggest a negative relation between unemployment and excess demand. The issue in question does not concern the basic adjustment hypothesis, which states that the rate of wage change is a function of the level of excess demand; but it does concern the relation between excess demand and unemployment. If it can be shown that this latter relation is negative throughout, the Phillips relation will be presumed to have a negative slope throughout. There are three general reasons why the relation between unemployment and excess demand is negative.3 (1). If in an individual labour market unemployment were the perfect counterpart for excess supply proper, then the relation between unemployment (U) and excess demand (D - S) would be kinked as in Figure 1. But even in an individual labour market it cannot be presumed that all parts of that market (e.g. all the firms) are in exactly the same