Rate of return to education of compliers: estimation based on Rubin causal models

Randomized experiments cannot determine the impact of migrant work on the rate of return to education due to the inability to randomly assign rural populations to rural and urban households. Under Rubin causal models, we evaluate the effect of migrant work on the rate of return to education of compliers by estimating the complier average causal effect (CACE) parameter. Our analysis focuses on data from rural and migrant residents in China in 2013, and we construct estimators of unknown parameters and the Mincer earnings function using a linear combination of polynomial spline functions. The empirical results indicate that migrant work increases the rate of return to education of compliers by 3.57\%. These results provide a scientific evaluation of the social and economic value of migrant work during the economic transition period, from a human capital perspective.