On-the-job training and productivity: Firm-level evidence from a large developing country

•We investigate the effects of on-the-job training on firm productivity and wages.•We use a large panel data set of manufacturing firms in China with national representativeness.•We combine the propensity score matching and the difference-in-differences techniques to estimate the treatment effect of training.•We find statistically and economically significant effect of training on productivity and wage.•These findings are not sensitive to industrial capital intensity or firm ownership structure.

We investigate the effects of on-the-job training on firm productivity and wages using a large panel data set constituted of all large and medium size manufacturing firms in China over 2003-2007. We estimate firm productivity carefully with a recent semiparametric method and combine the propensity score matching and the difference-in-differences techniques to estimate the treatment effect of training. We find consistent evidence that i) training helps boost firm productivity and wages; ii) the higher the training expenditure per capita, the higher the increase in productivity and wages; iii) firms benefit more from training than workers. These findings are not sensitive to industrial capital intensity or firm ownership structure.