Returns to schooling of male migrants in China using instrumental variable (IV) estimation

Abstract

In this paper, we aim to cover the gaps in the existing literature by providing an accurate estimation of returns to schooling of male migrants in urban China after the reform, using Instrumental Variation (IV). Previous literature have used the Ordinary Least Squares (OLS) approach and those that used IV faced some issues with the instruments chosen. We used data from the 2009 Rural Urban Migration in China (RUMiC) with firstborn and its interaction with age as instruments. The returns to education are 3.8% and 10.8% under OLS and IV regression respectively. This disparity is due to problems of endogeneity and measurement errors under the OLS method. The instruments proved to be relevant and valid after various testing. Endogeneity tests confirmed that the variable, yearsed, is not exogenous. Under the Generalized Methods of Moments (GMM), returns to education are 10.2%, verifying the robustness of the IV method. However, indeed, there are some limitations of using IV. There tends to be large standard errors and possible endogeneity for other variables. Despite this, we hope that this research paper contributes to the field of IV and provides some information about returns to education in urban China to relevant parties.