跨方言区流动、自选择与劳动力收入

本文基于2012和2014年中国劳动力动态调查（CLDS）数据，采用内生转换回归模型考察劳动力跨方言区流动对收入的影响。实证结果表明，与在同一方言区流动的个体相比，跨方言区流动收入要高31.9%；对于跨方言区流动的劳动力而言，跨方言区流动对收入的平均处理效应为30.5%，如果在同一方言区流动的劳动力选择跨方言区流动，其收入反而会减少3.9%，说明不同个体在流入地的选择上是基于其比较优势做出的决策。进一步，本文还发现跨方言区流动的劳动力无论选择在同一方言区流动还是跨方言区流动，均能获得较高的收入，在流动决策上具有绝对优势。在考虑了测量误差和极端值以及排除其他干扰因素后，上述结论依然稳健。

Cross-dialect Migration,Self-selection,and Income

Based on the data from CLDS (2012，2014), this paper employs endogenous switching regression model to control the self-selection effects and discusses the relationship between cross-dialect migration and income empirically. The results suggest that cross-dialect migration has positive effect on income. And the income of the cross-dialect migration group is 31.9% higher than the group of migration in the same dialect region. The average treatment effect on the treated is 30.5%, and the average treatment effect on the untreated is 3.9%. It implies that individual migration decision can be treated as a positive incentive types. Moreover, the results suggest that the cross-dialect migration group have an absolute advantage in migration decision. After taking the measurement error, outliers, and confounding variables into consideration, the conclusions remain robust.