Bounds on Average and Quantile Treatment Effects of Job Corps Training on Wages
用非参数方法估计美国最大青年职业培训项目对工资的平均和分位数处理效应的上下界，发现不同种族和性别群体的效应在工资分布的不同位置有差异。
*We thank Oscar Mitnik for detailed comments, and participants at the 2011 Institute for Research We assess the effectiveness of Job Corps (JC), the largest job training program targeting disadvantaged youth in the United States, by constructing nonparametric bounds for the average and quantile treatment effects of the program on wages. Our preferred estimates point toward convincing evidence of positive effects of JC on wages both at the mean and throughout the wage distribution. For the different demographic groups analyzed, the statistically significant estimated average effects are bounded between 4.6 and 12 percent, while the quantile treatment effects are bounded between 2.7 and 11.7 percent. Furthermore, we find that the program’s effect on wages varies across quantiles and groups. Blacks likely experience larger impacts in the lower part of their wage distribution, while Whites likely experience larger impacts in the upper part of their distribution. Non-Hispanic Females show statistically significant impacts in the upper part of their distribution but not in the lower part. 1