With Booze, You Lose: The Mortality Effects of Early Retirement

Abstract

This study analyzes the effect of early retirement on male mortality. I exploit two reforms in Switzerland, which allowed men as of a certain cohort to retire one and two years before the statutory retirement age. This generates two sharp eligibility cutoff dates, which I use in a regression discontinuity design. I draw from two full sample administrative data sets: the mortality and the old age insurance register. Retiring two years before the statutory retirement age increases the absolute risk of death before the age 83 by 41 percentage points. Heterogeneity analysis reveals that the effect is driven by lifestyle diseases such as alcohol dependence and respiratory diseases related to smoking. The effect is largest for unmarried men and for men living in the German-speaking part of Switzerland — who show a higher social norm toward work than men living in the Latin part. Also, there is no effect heterogeneity regarding income, which suggests that the negative health effect is not caused by a loss in income due to retirement. The results support the lifestyle hypothesis suggesting that retirement increases mortality due to a loss of structure and a concomitant unhealthy lifestyle.