The Lost Productivity: An Experimental Investigation of Human Versus Algorithm-based Discretion in Incomplete Contracts

• In Event: Plenary Session

Abstract

Incentive compensation is often characterized by incomplete contracts. In a setting where managers have discretion over the size of employee compensation pool, prior work indicates that managerial discretion deteriorates employee productivity. In this study, I experimentally investigate whether replacing human managers’ decision making with algorithm-generated bonus allocations that mimic managers’ decision making improves employee productivity. I find that discretionary bonus pools determined by algorithms generate higher employee productivity without sacrificing firm’s residual profits. Further, the productivity-inducing effect from algorithms is stronger when employees receive generous rewards or when the rewards are not contingent on the performance signal. These results are consistent with the idea that it is hard for managers to establish credibility for rewarding or punishing employees in incomplete contracts. Employee productivity is improved once the human element is removed from the responsibilities in determining the rewards, even when the reward strategy and outcomes are held constant. This study advances our understanding of the behavioral factors influencing employee productivity in incomplete contracts and the potential ways algorithm-based evaluations can be used to improve firm outcomes.

Author

• Wenqian Hu, Georgia Institute of Technology