摘要：This paper assesses the effect of the U.S. Job Corps (JC), the nation's largest and most comprehensive job training program targeting disadvantaged youths, on wages. We employ partial identification techniques and construct informative nonparametric bounds for the causal effect of interest under weaker assumptions than those conventionally used for point identification of treatment effects in the presence of sample selection. In addition, we propose and estimate bounds on quantile treatment effects of the program on participants' wages. In general, we find convincing evidence of positive impacts of JC on participants' wages. Importantly, we find that estimated impacts on lower quantiles of the distribution are higher, with the highest impact being in the 5th percentile where a positive effect on wages is bounded between 8.4 and 16.1 percent. These bounds suggest that JC results in wage compression within eligible participants.
