Using data from the 1982 Panel Study of Income Dynamics (PSID), the authors investigate the relationship between wages and the risk of work-related death or nonfatal injury. Including industry-level variables and using alternative risk measures dramatically affects measured wage compensation. The results cast doubt on the existence of compensating differentials for risk. Indeed, the strongest finding is the likely presence of negative compensation-relatively high risk and low wages-for nonunion workers. The role of rent-sharing or other forms of strategic bargaining behavior (captured by value-added per worker and other industry variables) and the gender distribution of both risk and wages demonstrate that noncompetitive elements in U.S. labor markets are sufficiently strong to overcome the competitive tendency toward equalizing differentials. This paper addresses two issues crucialto the interpretation of studies attempting to estimate wage compensation for dangerous work: the effect of noncompetitive aspects of labor markets, and potential error in the methods used to attribute risk to individual workers. We explore these issues using individual-level worker data from the 1982 wave of the Panel Study of Income Dynamics (PSID), data that enable us to replicate earlier findings and demonstrate the effect of changes in specification or choice of risk variables. Our findings are important apart from their significance for earlier work, because they shed additional light on the role that occupational risk plays in the wage structure. *James Madison College at Michigan State University and Hamilton College, respectively. The authors express their gratitude to Paul Leigh for his contributions at several stages of this project and to Jeffrey Pliskin for helpful comments.

No abstract

No abstract

No abstract