The paper extends the theoretical approach in Lazear (1986, 1996) to show that jobs with performance related pay (PRP) attract workers of higher unobservable ability, and also induce workers to provide greater effort. We then test some of the predictions of this model against data from the British Household Panel Survey, using earnings as a proxy for productivity. We find that PRP raises wages by about 9% for men and 6% for women over the entire (union and non-union) sample. Our theoretical calculations show that the estimated earnings differentials represent average productivity differentials net of monitoring costs, but not of the disutility of additional effort expended by workers. But the productivity differential is not a true productivity gain, for it includes a non-productive sorting effect as well as the effort effect. For all these reasons, the estimated return to PRP of 9% for men and 6% for women represents upper bounds on the productivity gains.
