Predicting the Behavior of a Newsvendor Using Prospect Theory

Because inventory systems are managed or heavily influenced by decision makers, understanding human behaviors in inventory management has become an important research topic. Based on newsvendor decisions, in the literature, it has been argued that a few methods (e.g., ex-post inventory error, and anchoring-insufficient-adjustment) are able to predict the human behavior, and it is not clear whether prospect theory is capable of making the prediction. In this paper, using both an empirical study and an analytical model with newsvendor settings, we justify the validity and the usefulness of prospect theory. We first present theoretical results and then report on empirical findings from laboratory experiments. Based on experimental data, we estimate parameters of the prospect theory model and demonstrate its validity in predicting inventory decisions. We also compare the prediction power of prospect theory with other methods and show that prospect theory provides better prediction than other methods do, such as ex-post inventory error