Is sick absence related to commuting travel time? - Swedish Evidence Based on the Generalized Propensity Score Estimator∗

This paper focuses on the effects of commuting time on sickness insurance utilization by applying a generalized propensity score estimator to a large sample of Swedish employees. We analyse the effect of commuting time both on the probability of using sickness insurance at all and on the probability that an individual on sick leave is on so-called partial sick leave rather than being completely absent from work. Insurance utilization is in both cases defined as being ill for more than 14 days. The results indicate, in general, that individuals do not use sickness insurance because of their commuting time. However, commuting time seems to increase the risk of being on sick leave among females with relatively low annual wage earnings. The results indicate, furthermore, a relatively weak and negative relationship between the probability of being on partial sick leave and commuting time in the group of individuals who have utilized sickness insurance. The latter result applies to both men and women.