Abstract
Depression is a common disorder that impacts on individuals' ability to perform daily activities, including those required for working. People with poor health tend to have problems needing medical care and therefore need time away from their work. This paper considers a structural model of labor absenteeism, considering the effect of depression. Our objective is to estimate the effects that depressive symptoms (among other factors) have on absenteeism while avoiding inconsistency in estimators due to sample selection and endogenous regressor. We are unwilling to impose strong assumptions, which are sometimes not required by theory, so our model is semiparametric. Based on microdata from the European Health Survey in Spain, our results indicate that depressive symptoms have a negative effect on working time and increase absenteeism. We conclude that depressed workers lose on average around 12 more days per year than non depressed ones. Levels of absenteeism are also estimated to be higher on average among obese people and among older people (the effect of age is positive). On the other hand, non-college education, being male and being self-employed are factors related to lower levels of absenteeism.