China’s delayed-retirement reform—recently announced and scheduled to start in 2025—seeks to bolster public-pension sustainability, yet its implications for wealth distribution remain unclear. Using six waves of the China Family Panel Studies (CFPS 2010–2020), we estimated a micro-level wealth-accumulation equation and embedded it in an agent-based counterfactual simulation. We examined three policy scenarios: a gradual schedule calibrated to the officially announced plan but re-indexed to begin in 2021 for the counterfactual window, and two comparators—a no-delay baseline and a one-off delay that immediately meets the target statutory ages. Using the 2020 CFPS wave (the latest available) as the baseline micro-dataset, we conduct an ex-ante counterfactual experiment that applies the reform as if it started in 2021; under a short-run, partial-equilibrium, full-compliance closure over a common 2021–2025 window (under 2020 conditions), the simulation traces changes in the national wealth distribution and the structural channels through which they propagate. Within our ex-ante counterfactual setting (2021–2025 under 2020 baseline conditions) and a short-run, partial-equilibrium, full-compliance closure, delayed retirement is associated with a higher wealth-inequality trajectory over 2021–2025, chiefly via widening gaps linked to urban–rural residence and health status; gender-related contributions remain comparatively small. Urban residents and healthier individuals are more likely to convert longer careers into additional wealth and pension entitlements, whereas gains among rural and less-healthy groups appear more limited. The findings are consistent with social-stratification and cumulative-advantage perspectives in China and present a transparent, ex-ante evaluation framework that can help inform the design and calibration of complementary measures in forthcoming pension reforms. These results should be interpreted as counterfactual policy-path differences over a common five-year horizon, not as realised outcomes or long-run forecasts.
Wang et al. (Mon,) studied this question.