Compensating Differentials for Dangerous Work in a Labour Discipline Model

Abstract

Wage compensation for occupational risk in Walrasian models has three properties: it generates efficient levels and allocations of risk, it enables workers in dangerous jobs to be just as well off, ceteris paribus, as workers in safe jobs, and it permits the estimation of willingness to pay schedules for safety from labour market data. This paper tests the status of these three properties under a simple model in which variable labour effort is induced by wage–safety rents and the threat of dismissal. It finds that all three must be modified, and that these results are robust over alternative specifications of labour market structure, the cost of providing safety, and the specification of the production function. The implications of this analysis are considered for both public policy questions and the interpretation of wage–risk studies.