The Relationship Between Retirement and Mental Health.

Our society is aging drastically and the sustainability of social security systems might be compromised. This development in age-pattern implies a potential change in retirement behavior and the impact of retirement on mental health therefore seems vital. This has been noticed in the literature and some effort has been made in understanding the (causal) impact on health. However, focus has been limited with respect to mental health. Studies find ambiguous impacts be it associations or causal effects. We carry out several approaches in order to answer the question on how retirement influences mental health. We focus on the causal relationship using retirement eligibility ages as instruments that vary in the observed countries. While previous work has focused only on average effects we add a relatively unexplored approach of quantile regression to test whether there are differences in returns depending on the conditional distribution of mental health, against the assumption that the effects are constant for all workers. Not much work has been born using quantile regression and to our knowledge this has not been done in the context given. We find limited causal, statistically relevant impact on clinically defined depression, but once examining the impact on the depression scale itself we find increases in severity of depression, that can be heterogeneous for different individuals. The evidence suggests that this impact is most prominent around the median of the distribution and less severe for those at the start and end of the distribution, i.e. those with extreme (lowest or highest) mental health problems. We find significant causal impact of retirement that increases severity of depression by 0.29 to 0.85 points on a scale from 0 to 12, depending on the quantile chosen. When retirement is an underlying source of depression this can have considerable implications for policies regarding retirement eligibility ages and occupational offerings after those ages.