by Michael Dworsky, Stanford University
The onset of permanent disability leads to sharply reduced earnings and labor force participation. However, the role of health-related changes in productivity remain poorly understood. This project addresses two related questions: to what extent do declines in health status affect a worker’s productivity and work capacity, and can health-related productivity losses explain heterogeneity in earnings losses across and within industries? To answer these questions, I study job transitions and earnings changes following disability onset using a large (N > 100; 000) dataset of administrative earnings histories and medical information from Workers’ Compensation (WC) Partial Permanent Disability (PPD) claimants in California. PPD claimants suffer from permanent work-related impairments, but have high labor force attachment relative to other permanently disabled populations. The PPD population thus presents a valuable opportunity to examine the mechanisms that link poor health to reduced earnings. Among other important features, this dataset is the first dataset of administrative earnings histories for the partially disabled to contain information on the industry of the post-injury firm. Accordingly, I am able to study how permanent impairment affects the kinds of jobs impaired workers take, and how these career changes contribute to persistent earnings losses.
I show that a majority of newly disabled workers who remain in the labor force switch industries when moving to their first post-injury jobs, and that these disabled workers experience far greater earnings losses than do disabled workers who move to new jobs in the same industry as the at-injury employer. Furthermore, the additional losses associated with changing industries are larger for workers with higher job tenure. Both findings are robust to controls for impairment severity. To explain these findings, I build a simple human capital model in which good health is a complement in the firm’s labor input to other human capital, and the strength of this complementarity depends on the physical demands of the job. To test this hypothesis, I will assess whether impairments that interfere with the physical demands of the at-injury job, as measured in the Department of Labor’s O*Net survey, increase the probability that the worker exits the at-injury industry. To determine the effects of injury on job transitions, I will estimate transition rates among healthy workers in the Survey of Income and Program Participation. Finally, I will estimate a joint model of job transitions and earnings in order to quantify the industry-specific changes in productivity implied by injured workers’ job mobility and their subsequent earnings.