Employer Learning and Schooling-Related Statistical Discrimination in Britain

This paper develops and tests a new model of asymmetric information in the labour market involving employer learning. In the model, I provide theoretical conditions for the identification -- based on the experience and tenure profiles of estimated returns to ability and education -- of employer learning about unobserved worker's productivity and statistical discrimination based on years of schooling. Using data from two British birth cohorts, estimates based on this model support the hypothesis that British employers have limited information about their workers, make inferences based on their education levels, and progressively learn about their true ability. Moreover, this learning process -- particularly among blue-collar workers-- favours incumbent employers relative to potential competitors (asymmetric learning). This informational advantage implies an additional distortion in the functioning of the labour market and policy evaluation rarely takes into account the informational impact of interventions and its implications for individual behaviour.