A Predictive Model for Newsvendor Behavior and the Implications for Supply Chain Performance

Abstract

The mismatch between newsvendor theory and experimental observations has garnered significant attention in the literature; several theories have been developed to capture the observed behavior. Yet, there has been limited research on the implications of observed behavior, leaving both a theoretical and practical gap. We aim to close this gap by applying experimental findings to improve practical decisions. For this purpose, we first provide a comparative analysis of alternative behavioral theories using experimental data and show that a model of reference dependence and random errors organizes the data better at the population level than others, even under model selection criterion that penalizes complexity. We then develop a behavioral theory of reference dependence and random errors in newsvendor decision making and assess the impact of behavioral influences on the performance of a supply chain. Ignoring a retailer’s behavioral influences can result in significant profit loss for a supplier. In a wide range of conditions, the supplier benefits and the retailer is hurt by the retailer’s behavioral influences because they may increase the supplier’s pricing power. However, the retailer can benefit from small levels of random errors. Both behavioral influences can result in an improvement in supply chain performance, particularly when the supplier’s margin is fixed and high.