Abstract

I explore the short term effects of mental health symptoms on employment outcomes. To address potential simultaneity bias, I propose a new instrument related to Seasonal Affective Disorder (SAD) that depends only on latitude and day of year. For unmarried individuals, an additional day of poor overall mental health (of the last thirty days) decreases the probability of employment by over five percentage points and increases the probability of unemployment by under five percentage points. Results from the first and second stage suggest that low-income, uninsured individuals experience more adverse mental health and employment effects. Relative to instrumental variable models, ordinary least squares models underestimate the transition from employment to unemployment and out of the labor force due to mental health problems.