Job Allocation Rules and Sorting Efficiency: Experimental Outcomes in a Peter Principle Environment

International audienceAn important issue in personnel economics is the design of efficient job allocation rules. Firms often use promotions both to sort workers across jobs and to provide them with incentives. However, the Peter Principle states that employees' output tends to fall after a promotion. Lazear (2004) suggests that self-selection may improve job allocation efficiency while preserving incentive effects. We reproduce this Peter Principle in the laboratory and compare the efficiency of a promotion standard with subjects self-selecting their task. We find no evidence of effort distortion, as predicted by theory. Furthermore, we find that when the Peter Principle is not severe, promotion rules often dominate self-selection efficiency of task assignment. Results are consistent with imperfect appraisal of transitory ability and a lack of strategic behavior