A Quasi-Experimental Approach to Indentifying Compensating Wage Differentials for Occupational Risks

This research employs the first quasi-experimental design within a labor market setting to provide causal evidence on the existence and magnitude of compensating wages for workplace fatality and accident risks. Federal programmed inspections conducted by the Occupational Safety and Health Administration that randomly select plants for inspection (conditioned on known and measured plant characteristics) are used to instrument for plant-level risks. Inspection records are combined with confidential U.S. Census data on plant-level wages and worksite characteristics to estimate the compensating wages associated with riskier working conditions. Results strongly suggest that compensating wage differentials for risky working conditions do indeed exist as suggested by theory and explored empirically in the hedonic wage literature for over 40 years. However, our results also suggest that the empirical challenges inherent in estimating these differentials via cross-sectional or panel-data hedonic wage models have not yet been fully addressed. Data limitations, especially with regards to the measurement of risk faced by workers at their worksite, and correlated unobservables are likely the key impediments to identification of wage/risk tradeoffs. Our results are important for environmental, health and public safety policies that rely on labor market studies to provide measures of the value of reducing fatality risks. Our point estimates suggest that prior studies may substantially overstate the value workers place on reducing workplace risks.