We analyze the eect of retirement and length of schooling on word recall at older ages using micro-data on 25,037 individuals and schooling and pension policy variation across 12 countries. Accounting for endogeneity of both schooling and retirement from the labour market, we show that they both have a positive eect on word recall. Mental and physical activities do not decrease due to retirement, and time use on non-passive activities signicantly increases. Dierent levels of schooling across cohorts, ages and retirement states drive the spurious negative association between retirement and recall in raw data and in previous causal work that has ignored the role of schooling.
Paul Bingley, SFI