Understanding outcome bias

Disentangling effort and luck is critical when evaluating outcomes. In a principal-agent experiment, we demonstrate that principals' judgments of agents are biased by luck, despite perfectly observable effort. This erodes the power of incentives to stimulate effort. We explore two potential solutions to this “outcome bias”–information control, and outsourcing judgment to independent third parties. Both are ineffective. When principals control information about luck, they do not avoid it. When agents control information, they manipulate principals' outcome bias to minimize punishments. We also find that even independent third parties exhibit outcome bias. These findings suggest that outcome bias cannot be driven solely by disappointment nor distributional preferences. Instead, we hypothesize that luck directly affects principals' inference about agent type even though effort is observed. We elicit the beliefs of third parties and principals and find that lucky agents are believed to be harder workers than identical, unlucky agents.