Estimating compensating wage differentials from worker mobility

We estimate the effects of the implementation of a compulsory work injury insurance in Sweden in 1978 on compensating wage differentials. This involves two steps. First, we investigate if there are compensating wage differentials on the Swedish labor market and second, we assess if these were altered by the reform. We use panel data for the period 1970 to 1990 with annual information for a sample of blue collar workers. The econometric model departs from the worker's job mobility decision. Endogeneity, selection and measurement errors of risk exposure are considered in the estimation. The estimates show significant compensating wage differentials for work-related diseases in the female sub-sample. No significant effect of the reform was found. -- Work injury insurance ; Panel data