Measuring job risks when hedonic wage models do not do the job

Hedonic wage regressions show little evidence that European workers facing larger job risks and other workplace disamenities receive higher wages. On the other hand, workers in more risky or unpleasant jobs are less satisfied with their jobs, ceteris paribus. If labor markets were perfectly competitive and workers fully informed of their working conditions ex ante, according to the theory of compensating differentials, there should be no relationship between on-the-job risk and job satisfaction because wages would fully adjust to compensate for differences in job characteristics. We show that when wages do not fully compensate for on-the-job risks, the willingness to pay to reduce mortality risks estimated from hedonic regressions needs to be complemented with a residual effect of job risks on utility which is not capitalized on wages. We explore the potential of job satisfaction regressions as an additional valuation approach to estimate the tradeoffs between wages and risks that keep job satisfaction constant.