We provide first evidence regarding the direct impact of educational mismatch on firm productivity. To do so, we rely on representative linked employer–employee panel data for Belgium covering the period 1999–2006. Controlling for simultaneity issues, time-invariant unobserved workplace...