Perry's Model of Wage-Determination with Stochastic Parameters

Empirical studies of wage-determination, within the general framework of the Phillips curve and its variants, demonstrate remarkable consistency with respect to two particular specifications. They usually involve explanatory variables in the form of simple fourth-order moving averages and a dependent variable which is represented by the sequence of overlapping annual changes in an aggregate index of wages. Explicit justifications for these choices are rare, but George Perry does indicate some basis for them, and his arguments may have proved persuasive to other economists. These arguments, which are based on institutional features of the labour market, have been discussed. Two important implications of them within the context of distinct bargaining groups in the labour force are discussed in the paper. If the specifications are justified by reference to aggregation over those groups which bargain at different points in time, then a fixed-weight paradox can be illustrated, and a Yule-Slutsky effect is indicated for the linear statistical model, which is used as the basis for estimates in the empirical studies of wage-determination.