We investigate the causal effect of commuting on sickness absence from work using German panel data. To address reverse causation, we use changes in commuting distance for employees who stay with the same employer and who have the same residence during the period of observation. In contrast to...