This paper examines the earnings effects of performance pay using linked employee-employer panel data from Finland's metal industry for 1990-2000. The authors estimate the effects of performance pay contracts in the presence of individual and firm unobserved heterogeneity as well as in tasks of different complexity. Unobservable firm characteristics explain about 40% of the variance in the use of performance pay. Performance pay workers earned substantially more than fixed rate workers, a finding that persists even in analyses that use for identification only those workers who changed firms (and contracts) due to an establishment closure. There is also evidence of a strong, negative relationship between job complexity and the incentive effects of performance pay. Finally, several "quasi-experiments" show that when one plant underwent a compensation regime change but other highly similar plants in the same firm did not, workers in the "treatment" plant gained substantial earnings premiums.