Why is the Return to Training so High?

Cross-sectional results show that training increases wages by 5 per cent. This return is on a par with the return to 1 year of education. Considering that the average duration of training is very short, this result is strange and needs further examination. After leaving out the importance of measurement error, we control for accumulated stock of firm-specific skills, unobserved heterogeneity in wage levels, heterogeneity in training returns, and heterogeneity in wage growth. By this we manage to reduce the return considerably. Unobserved heterogeneity in wage levels is the most important contributor to the 'too high' returns to training