The impact of retirement on cognitive function of urban women under China's age-50 retirement policy: a fuzzy regression discontinuity study

Existing evidence on retirement and cognitive function is inconclusive. Thus, using five national waves (2011-2020) of the China Health and Retirement Longitudinal Study (CHARLS), this study examines urban female employees aged 45-65 covered by China's age-50 statutory retirement policy. A fuzzy regression discontinuity design uses the age-50 as an instrument for retirement. Accordingly, the identified Local Average Treatment Effect (LATE) applies primarily to urban women subject to the age-50 retirement rule, mainly blue-collar or non-managerial workers, rather than all urban women in China. At the cutoff, no statistically significant immediate net effect of retirement is observed on either mental intactness or episodic memory. Donut regression excluding individuals aged 50 shows a significant negative effect of retirement on mental intactness. This is consistent with a potential delayed manifestation. Therefore, a zero immediate net effect does not mean that retirement is entirely unrelated to cognition. However, because this study employs a pooled cross-sectional regression discontinuity design and does not track changes in the same individuals over time, these results should be interpreted cautiously.