摘要
本文利用从人口抽样调查中生成的16—35岁的双胞胎数据,采用OLS和双胞胎组内差分法估计了教育的收益率及其异质性。主要得到两个结果:第一,在没有纠正测量误差的情况下,混合OLS估计的教育收益率为14%,而组内差分估计仅为4%左右。参照相关研究对结果进行一个简单的修正后,差分估计值也不到6%。这意味着用OLS会高估教育的收益率。第二,那些能力越高、家庭背景越好的个体,越有可能获得更多的教育,同时从教育中获得的收益也越多。
Using a sample of 16--35 year-old twin data generated from 1% Chinese Census data in 2005, we estimated return to education by pooled OLS and within-twin fixed effects model considering the heterogeneity problem. The estimated rate of return is 14% by pooled OLS, while it is reduced to about 4% by within-twin fixed effects model. We construct a method to correct measurement errors by referring to other similar research and the estimator is still less than 6%. In addition, we found individuals with higher abilities and better family backgrounds are more likely to get more education and higher returns to education.
出处
《经济学（季刊）》
CSSCI
北大核心
2014年第2期1001-1020,共20页
China Economic Quarterly
基金
北京师范大学"中央高校基本科研业务费专项资金"(2012CXQT11)
北京市哲学社会科学首都教育经济研究基地项目的资助
