Wage Differentials and Specification Bias in Estimates of Relative Labor Prices
OME empirical studies on developments of wage differentials over time use measurements of relative wage dispersion as indicators for the movement of relative labor prices. This is obviously the case when a reduction movement of relative wage dispersion is explained by the growth of education, and when an increase of relative wage dispersion is explained by unskilled immigration.' The purpose of this paper is to show (i) that measurements of relative dispersion overestimate the (average) change of relative wage differentials;( 2 ) that measurements of relative differentials tend to underestimate the (average) change of relative labor prices; and (3) to suggest a method by which a consistent estimate for the (average) change of relative labor prices can be computed.