We investigate to what extent worker heterogeneity explains the well-known wage and productivity exporter premiums, employing a matched employer-employee data set for Norwegian manufacturing. The wage premium falls by roughly 50% after controlling for observed and unobserved worker...