Unexpected Costs of Performance Incentives with Production Uncertainty: Theory and Evidence from a Real-Effort Laboratory Experiment

Policy advocates often promote market-like incentives for publicly provided services like education or healthcare (e.g., Medicare). Evidence for performance incentives in these sectors is mixed, possibly due to production uncertainty represented as uncertainty about the marginal effect of inputs. Using a principal-agent model, I demonstrate that such uncertainty can lead to inefficiencies in output-based incentives. The model illustrates how employees favor inputs with lower uncertainty and reduce overall effort. Input-based incentives might be more effective in such cases. I conduct a real-effort lab experiment which validates these predictions: participants shift from efficient inputs as uncertainty grows, and reduce overall effort.