We present evidence from a firm level experiment in which we engineered an exogenous change in managerial compensation from fixed wages to performance pay based on the average productivity of lower-tier workers. Theory suggests that managerial incentives affect both the mean and dispersion of workers' productivity through two channels. First, managers respond to incentives by targeting their efforts towards more able workers, implying that both the mean and the dispersion increase. Second, managers select out the least able workers, implying that the mean increases but the dispersion may decrease. In our field experiment we find that the introduction of managerial performance pay raises both the mean and dispersion of worker productivity. Analysis of individual level productivity data shows that managers target their effort towards high ability workers, and the least able workers are less likely to be selected into employment. These results highlight the interplay between the provision of managerial incentives and earnings inequality among lower-tier workers.

We present evidence on whether workers have social preferences by comparing workers' productivity under relative incentives, where individual effort imposes a negative externality on others, with their productivity under piece rates, where it does not. We find that the productivity of the average worker is at least 50 percent higher under piece rates than under relative incentives. We show that this is due to workers partially internalizing the negative externality their effort imposes on others under relative incentives, especially when working alongside their friends. Under piece rates, the relationship among workers does not affect productivity. Further analysis reveals that workers internalize the externality only when they can monitor others and be monitored. This rules out pure altruism as the underlying motive of workers' behavior.