Medicated Optimism

Do people experiencing depression form accurate expectations about losing their jobs? Linking Estonian panel survey data to administrative employment, diagnosis, and prescription records, I track within-person changes in expected job loss and subsequent realized job loss as individuals move into and out of depression and depression medication. Using individual fixed effects and controls for depression severity and physical health, I find that as the same worker enters the medicated depression state from no depression, expected job loss falls by 3.6 percentage points while realized job loss rises by 5.1 percentage points. The resulting 8.7 percentage-point fall in the forecast error is split almost evenly between less over-prediction among workers who keep their jobs and more under-prediction among workers who lose them. This shift toward optimism is larger for medication classes more closely linked to reward-related pathways and is absent when prescribed medication is not purchased. The welfare implications are two-sided: more optimistic beliefs may reduce worry about job loss, but workers who understate rising risk may prepare less for a separation.