In this paper, we investigate the complex interrelationships among worktime, wages and health identified in the Grossman model of the demand for health. We specify a 3-equation simultaneous model designed to capture the tune dependent character of these interrelationships, and estimate the model using 8 years of panel data on 882 males aged 22 to 71. The model is estimated using Hansen's generalized methods of moments imposing a weak set of conditions on the error term covariance structure. Using our data, we estimate simpler models with more restrictive assumptions commonly found in the literature, and find substantial differences between these estimates and those from the simultaneous model. For example, the positive relationship between worktime and health found in other studies disappears when the relevant simultaneities are accounted for. Our simultaneous estimates also suggest that worktime spent in environmentally adverse conditions are inversely related to health status, while job related physical exercise retards health deterioration.
