Wages and Working Conditions

This paper presents another extension of the approach initiated by Brown. As in Brown's work, the wage change specification is used to control for bias due to omitted ability data. Then, as in Duncan and Hoimlund's study, working conditions are measured using subjective self?reported data. However, in this paper, working conditions are measured by a single comprehensive variable. This approach eliminates omitted working conditions as a source of bias. The working conditions measure is then treated as an unobserved variable which limits measurement error to an unknown scale factor. The model is estimated using a technique derived by memiya (1978).