The age-productivity gradient: evidence from a sample of F1 driver

Estimating the effect of aging on productivity requires clean measures of productivity. Additionally, one needs to control for unobserved heterogeneity at the worker, firm and worker/firm level, to account for the role of experience and to correct for selection bias. We tackle these issues exploiting a panel of Gran Prix Formula One drivers, which provides a unique setting to single out the data requirements needed to credibly estimate the effect of age on productivity. Results robust to the inclusion of worker, firm and match effects show that the age-productivity link has an inverted U-shape profile with a peak at the age of 30-32. The use of repeated cross-sections of individuals also produces consistent results provided that cohort effects are properly accounted for. Relying on team-average measures of productivity makes instead inference harder. © 2010 Elsevier B.V.