Reference-dependent preferences in flat penalty service-level contracts

This paper investigates a supply chain governed by a flat penalty service-level contract in which missing the target fill rate can lead to costly operational disruption. We focus on near-miss bias: (1) the preference for near-miss events, that is, risky production quantities that reach the target but narrowly avoid disruption; and (2) riskier decision-making due to such preferences. We propose a reference-dependent behavioral model that explains the near-miss bias. The findings of a laboratory experiment show that production quantities are evaluated based on realized profits and are below the optimal model prediction. Contracts associated with lower perceived severity, that is, the ratio of flat penalty to wholesale price, result in lower production quantities than those with higher perceived severity, even though the standard model does not predict any effect. A structural estimation analysis indicates that the behavioral model performs better than the standard model in terms of predictive accuracy and goodness of fit. Our analysis provides insights for managers who design supply chain contracts in settings with considerable risk of disruption due to a shortage of critical parts.