We use longitudinal linked employer-employee data and find that the probability of participating in firm-sponsored classroom training diminishes rapidly for workers aged 45 years and older. Although the standard human capital investment model predicts such a decline, we also consider the possibility that returns to training decline with age. Taking into account endogenous training decisions, we find that the training wage premium diminishes only slightly with age. However, estimates of the

more »

... t of training on productivity decrease dramatically with age, suggesting that incentives for firms to invest in classroom training are much lower for older workers. JEL Classification: C23, D24, J31 2 1 See Skirbekk (2004) for an overview of the literature and Göbel and Zwick (2009) and Cardoso, Guimaraes, and Varejão (2010) for recent examples. 2 Surveys generally distinguish between on-the-job and classroom training. However, studies comparing the productivity impact of classroom versus on-the-job training find low or negligible impact of on-thejob training on productivity. For example, Zwick (2005) finds that formal external and internal courses have a positive impact on productivity while on-the-job training has a negative impact. Barrett and O'Connell (2001) find positive impact for general training but not for specific training. Finally, Black and Lynch (1996) find a positive impact of training only in the case of formal off-the-job training in the manufacturing sector. Hence we focus on classroom training.