This paper examines the causal effect of extended working lives on mortality among German pensioners. I exploit a German pension reform that raised statutory retirement ages and introduced actuarial deductions for early retirement for specific cohorts, using this exogenous variation in retirement timing through an instrumental variable strategy. The analysis draws on pension records covering nearly the universe of German public pensioners, with precise information on retirement dates and mortality outcomes, allowing me to track survival up to ages 78–81. The results reveal substantial heterogeneity: for women in the sample, later retirement has no discernible causal effect on age at death, while for men, delayed retirement significantly decreases longevity. Planned heterogeneity analyses will investigate the role of occupational strain in shaping post-retirement mortality. Robustness checks address concurrent labor market reforms and income effects arising from changes in benefit calculation. Overall, the findings indicate that extensions of working life do not necessarily reduce longevity but can have negative effects on mortality, highlighting important heterogeneity in the welfare implications of raising retirement ages.