How do workers perceive the risks from automation and the opportunities to retrain? Evidence from a survey of truck drivers

Using a new survey of truck drivers, we find that those who are most concerned about automation are, counterintuitively, also most likely to say they intend to re-invest in driving. This arms race for remaining positions is socially inefficient. We find that the effect disappears among drivers familiar with the low costs of community college, consistent with a model of idiosyncratic noise in the perceived cost of retraining. However, we calibrate that even with correct information about the true costs of retraining, reasonable beliefs about other facets of the labor market may cause truckers to stick with their current occupation.