Abstract

This paper estimates the effect of disability on labor force participation by using symptoms or diseases as instruments in a simultaneous equations model of endogenous reported disability and labor force participation. The results show that each measure of disability explains a significant amount of variation in labor force participation, though the two are not perfect substitutes. There is only weak evidence of endogeneity of the disability variables. For cases where there is evidence of endogeneity, the bias it causes has the opposite effect of that hypothesized in the literatute (i.e., stress causes health to deteriorate with labor force participation). Furthermore, it has only insignificant effects on the coefficients for other variables.