Decision aids and the design of incentive compensation contracts: An experimental study of self-selection and effort effects

Compensation contracts affect both the type of worker attracted to an organization and the amount of effort exerted by the worker once on the job. Prior research suggests that the design of contract incentives should consider the sensitivity of incentives to effort and the precision with which effort can be measured, and finds that the positive effects of certain contract incentives may be undermined or reversed with the introduction of a decision aid. This study suggests that automated decision aids may provide an organization the capability to design incentives that are both more sensitive to effort and more precisely measured than without the decision aid and, accordingly, tests self-selection and effort effects of alternative incentive compensation schemes in an automated decision aid environment.;Agency theory and behavioral decision theory are used together to develop new contract provisions based on the availability of behavior measures from the decision aid. Expected self-selection and effort effects are developed for both traditional outcome-based compensation contracts, and the new behavior-based contracts in the presence of an expert system. A laboratory experiment is conducted with a 2x2 between subjects factorial design obtained by crossing the presence or absence of an expert system decision aid with high and low programmable tasks.;In the absence of an expert system, results are consistent with prior agency research findings that lower skilled individuals tend to select fixed pay contracts while higher skilled individuals tend to select outcome-based contracts, and that performance is higher for outcome-based contracts. Additionally, the study finds that task programmability does not significantly affect contract selection. In the presence of an expert system, approximately one-fourth of the subjects select some form of the new behavior-based contracts regardless of skill or risk preference. Further, the performance of subjects with behavior-based contracts is equivalent to that of subjects with outcome-based contracts. Behavior-based contracts may provide incentives for high performance at a lower cost than outcome-based contracts.