Estimates of the performance pay (PP) premium using household survey data are upwardly biased due to the use of PP contracts in higher paying workplaces. Using nationally representative linked employer‐employee data we estimate the PP premium at the mean and across the wage distribution, after accounting for workplace heterogeneity. We then present the effects of PP on the wage distribution. Failure to account for the use of PP in high paying workplaces overstates the PP premium by one‐third. The premium rises as one moves up the wage distribution, but the effect is more modest after having accounted for workplace heterogeneity.