Newsvendor models and biases under ambiguity

Abstract

In this study, we model the classical newsvendor ordering preferences under ambiguity. The extant literature on normative models in the newsvendor setting assumes decision-making under risk, where decision-maker has exact knowledge of the probabilities associated with the outcomes. In several business situations, the demand distribution is often incomplete or unknown. This results in decision-making under ambiguous situations. Decision theory recognizes the difference between exact probabilities and more realistic ambiguous probabilities. In his seminal paper, Scarf (1958) develops a max-min approach for the newsvendor with incomplete demand information. In the Scarf model, the newsvendor is assumed to be risk-neutral and ambiguity averse. In the recent experimental literature, it has been observed that the newsvendor behavior is not consistent with the Scarf model, and exhibits pull-to- center bias and other biases. This motivates our research to develop quantitative models under ambiguity to describe the observed biases in the literature. © IEOM Society International.