The causal effect of health on income: quasi-experimental evidence from commuting accidents

Martina Zweimüller, Martin Halla
Johannes Kepler University, Austria

This paper interprets accidents occurring on the journey to and from the place of work as negative health shocks. We argue that in our sample of (on their labor market history) exactly matched treated and control units, these negative health shocks are quasi-randomly assigned. To estimate the causal effects of this negative health shocks on labor market outcomes we apply a difference-in-differences (DiD) approach. This research design allows us the establish a clear causal effect. We find a negative and persistent causal effect of a negative health shock on employment and earnings. The size of these effects varies along the dimensions of sex, age and occupation. The effects on employment are higher for older than for younger workers, which is plausible since older workers recover less well and are less attached to the labor market. The most striking difference in the earnings effect is found between blue-collar and white-collar workers. It seems that if blue-collar workers manage to return to work after the negative health shock, they do not lose compared to untreated units. For white-collar workers, however, we find a substantial income loss almost throughout the whole post-treatment period. This finding is in line with theoretical arguments suggesting that labor market interruptions are more costly in the case of a job associated with firm-specific human capital compared to a job where more general human capital is decisive.

View full paper