Description:
Using harmonised data from the English Longitudinal Study of Ageing (for England), the Health and Retirement Study (for the US) and the Survey of Health, Ageing and Retirement in Europe (for 11 European countries), this paper estimates the impact of retirement on the self-assessed general health and physical health of individuals who retired between 2004 and 2007, being aged 50 or over (retirees), and individuals who retire upon reaching State Pension age (compliers). The impact on retirees is estimated using a difference-in-differences approach combined with propensity score matching while estimates of the impact on compliers are computed by means of an instrumental variable strategy. The results from both estimation approaches point towards a negative impact of retirement on older people’s health in that they suggest that, in the immediate post-retirement period, retiring increased (decreased) the proportion who had been diagnosed with a physical condition (reported good/better health). Retirement appears to have been more detrimental for the physical health of compliers than of retirees (the impact detected is almost eleven times larger for the former). In addition to providing new evidence of the causal effect of retirement on health, this paper contributes to the existing literature by estimating the (unobservable) characteristics of the compliers using Abadie’s kappa weighting scheme. Comparing these characteristics to those of retirees provides a possible interpretation to reconcile the different magnitudes of the estimates for compliers and retirees