本文利用2002年和2013年CHIPS数据构成混合横截面数据，使用工具变量法估计了我国省级层面城镇地区的教育外部收益率。结果表明，劳动者的平均受教育年限每提高一年，将导致该省城市劳动者工资提高 45%，而大学学历劳动者占比每提高一个百分点将导致劳动者工资提高1.09%。另外，高学历劳动者获得的教育外部收益率高于低学历劳动者。研究结论证明我国的教育具有明显的正外部性，为政府在教育领域的投资提供了实证支持。

Base on the mixed cross-sectional data which is constituted by the 2002 and 2013 CHIPS data, this paper has used instrumental variables method to estimate the external returns to education in urban China. The IV results show that one year increase in prefecture-level average education could increase individual wage by 45%, and one percentage point increase in the supply of college graduates could increase individual wage by 1.09%. In addition, the external returns to education of highly educated workers is higher than that of low education workers. Our findings prove that education in urban China has significant positive externalities, which provides empirical support for government investment in education.