Optimal Incentive Contracts, Common Uncertainty, and Intrinsic Value

Abstract

This paper concerns the optimal use of incentive contracts in a multiple agent situation where moral hazard problems are issues. This paper proposes to investigate (1) the effect of common uncertainty on expected contract cost of different contract types (an Individual contract, a Team-based contract, a Relative performance contract), (2) the effect of common uncertainty on the principal’s preference over the different contract types under different levels of common uncertainty, and (3) the intrinsic value of using relative performance information on agents’ efforts. Taking the standard agency paradigm where a principal can improve his welfare by achieving better risk-sharing without altering agents’ incentives to take the desired action, this paper hypothesizes that (1) the expected contract cost of a Team-based contract would increase, while that of a Relative performance contract would decrease as the level of common uncertainty increases, (2) there exists a critical value of the common uncertainty such that a Team-based contract is preferred by the principal when the level of common uncertainty is below the critical value and a Relative performance contract is preferred by the principal when the level of the common uncertainty is above the critical value, and (3) neither the level of common uncertainty nor the contract type would change the agents’ effort levels, Hypotheses are to be tested in two computer-assisted experiments with 80 MBA students. In experiment I, 60 MBA students are required to act as managers and make production decisions. 20 subjects are assigned to one of three contract groups and matched as a pair. Subjects under an Individual contract are compensated only based on their individual performance. While subjects under a Team-based contract are compensated based on both performances of subjects, subjects under a Relative performance contract are compensated based on relative performance to other subject's performance. In experiment II, 20 MBA students are required to act as principals to choose an incentive contract among three incentive contracts given a common uncertainty. There are 10 levels of common uncertainty, and common uncertainty is manipulated by providing conditional contingency tables on their effort choice. Subjects’ risk and effort preferences are induced by utilizing Berg et al technique (1986).