中国城乡居民养老保险制度的家庭收入效应——基于倾向得分匹配(PSM)的反事实估计

The Effect of Residential Pension Insurance Policy on Household Income in China——A Counterfactual Estimate Based on Propensity Score Matching(PSM)

基于2015年中国健康与养老追踪调查数据（CHARLS 2015）,本文采用倾向得分匹配方法（PSM）准确估计城乡居民养老保险制度实施对我国城乡居民家庭收入的影响效应。实证结果表明：（1）参加城乡居民养老保险制度确实产生了显著的家庭收入效应,给＂参保家庭＂带来的收入增加幅度为61. 39%,这为政府进一步扩大养老保险制度覆盖面的合理性提供了佐证;（2）经过倾向得分匹配方法处理后＂参保家庭＂的平均处理效应仅为0. 493,故而使用OLS模型高估了城乡居民养老保险制度因素对我国城乡居民家庭收入的增长作用,大致高估了12%左右。据此本文提出优化养老保险制度、推动收入分配改革的对策建议。

Using a merged data from the CHARLS（ 2015）,this paper studied the income effect of China＇s residential pension insurance policy by employing the propensity score matching（ PSM） method. The results show： first,the pension insurance policy has caused a 61.39% increase of household income under OLS regression,which provides strong evidence for further expanding of this policy; second,the average treatment effect was reduced to 49.3% when the PSM method was used,suggesting a 12% overestimate by OLS regression. Policy suggestions on improving the pension insurance policy and income distribution reform were put forward.