Wage Inflation in the United Kingdom

Abstract

In Britain there is little agreement on what determines the growth of wages, or indeed on whether a stable wage equation exists.' In this paper I try to clarify matters in two ways. First, I aim to represent the data generation process as faithfully and accurately as possible, while using (with a few modifications) the criteria for model selection advanced in Davidson et al. (1978). Second, I use five different measures of the pressure of demand, rather than just picking on one particular measure.2 It may be helpful at this stage to summarize very crudely the argument and results. In Section II, having first set out the framework for the rest of the analysis (which relates to the period 1961 (I)-1980 (IV)), I modify the standard real-wage model by proxying the level of real wages by a moving average of the past level of real wages. It is found that the coefficient on the prevailing real wage is equal (but of opposite sign) to that on the desired wage. Thus, in the long-run, inflation is independent of the level of real wages. The pressure of demand and incomes policy dummies are both found to be significant. One would have expected that, in the long run, the rate of growth of earnings at given price inflation and pressure of demand would respond to differences in the rate of growth of productivity. But over the sample period there was no evidence of this (once we allowed for incomes policy), which implies considerable rigidity in the target rate of growth of wages. The analysis so far includes no specifically expectational variables. In Section III I use some survey data on wage expectations which have the advantage of being based on quantitative responses. One rationale for having wage expectations in a wage equation might be that in a world of staggered contracts, where workers care about relative wages, they will take into account what they expect to happen to the wages of other workers when settling now. A two-equation model for wages and wage expectations is presented in the subsequent section, and in most respects the results confirm our conclusions from the previous sections. However, the preferred specification obtained here, which encompasses much of the earlier work and probably constitutes progress, is one that we have some doubts about, and will, it is hoped form the basis for future research in this area.