Estimating the eect of retirement on mental health via panel discontinuity designs

This article explores the potential effects of retirement on mental health and health care utilisation isolating sources of potential heterogeneity in treatment effect. To estimate the effects of retirement, we devise a new identifying strategy that combines kink and regression discontinuity designs with panel data methods. Our method is then applied to the British Household Panel Survey, a rich representative longitudinal survey. It is found that retirement has a small impact on primary care use, but overall has little effect on mental health.