Abstract: Based on the data of the national dynamic survey of migrant population in 2014 in China and considering the deviation and heterogeneity of sample selection, this paper estimates the impact of government training on hourly wage and monthly income of migrant population by using endogenous transformation regression model, and investigates the selection mechanism of employees and self-employed workers participating in training. The results show that the migrant population with women, longer education years and more social capital are more likely to choose training. Employees’ participation in training decision-making follows comparative advantages. There is no clear choice mechanism for self-employed workers to participate in training. The effect of government training on hourly wage and monthly income of employees was 1.04% and 0.39% respectively, which had no significant effect on hourly wage and monthly income of migrant workers. The influence of government training on the income of migrant population is very limited, so it is necessary to improve the management of government training.

Key words: Government Training, Employee Floating Population, Floating Population of Self-employed Workers, Labor Training, Employment Training for Migrant Workers, Education Years, Income Level

摘要： 本文基于2014年全国流动人口动态调查数据,在考虑样本选择偏差和异质性基础上,运用内生转换回归模型估计了政府培训对雇员和自营劳动者两类流动人口小时工资和月收入的影响,考察了雇员和自营劳动者两类流动人口参加培训的选择机制。结果表明：女性、受教育年限较长、社会资本较多的雇员流动人口更愿意参加政府培训,其参加培训的决策遵循比较优势;自营劳动者流动人口参加政府培训没有明确的选择机制。政府培训对雇员流动人口参加培训者的小时工资和月收入的处理效应分别为1.04%和0.39%,而对自营劳动者流动人口参加培训者的小时工资和月收入都没有显著影响。可见,政府培训对提高流动人口收入的作用非常有限,改善政府培训管理办法十分必要。

关键词: 政府培训, 雇员流动人口, 自营劳动者流动人口, 劳动力培训, 农民工就业培训, 受教育年限, 收入水平