Misaligned Agents

How should firms select and incentivize managers who allocate resources based on private information when their performance is imperfectly measured? We show that, contrary to common wisdom, firms may benefit from appointing managers with misaligned preferences. When performance can be measured more accurately along one dimension than another, firms optimally select managers who are biased toward the poorly measured dimension. This allows to provide strong performance-based incentives on the well-measured dimension while relying on managerial bias to counterbalance substitution effects. As a result, firms with worse performance measurement on some dimension will tend to hire more biased managers, such as managers with strong prosocial or environmental motivations. This framework can explain apparent inconsistencies between firms' stated objectives and the explicit incentives they provide, and it offers a novel perspective on the strategic roles of corporate culture and corporate social responsibility.