Predicting Order Variability in Inventory Decisions : A Model of Forecast Anchoring

Problem Definition: In this paper, we develop a forecast anchoring model that explains and predicts order variability behavior in a multi-period newsvendor problem. Our model assumes that people anchor on random point forecast (hence the name) and insufficiently adjust toward the profit-maximizing quantity.Academic/Practical Relevance: We show that this model is capable of predicting both the mean and variance of order quantities observed in experimental studies, and thus extends the existing literature that focus on predicting the mean order quantities only. It also helps identify possible behavioral causes that lead to excessive order variability (or the bullwhip effect) in supply chains.Methodology: We apply the generalized method of moments to fit the forecast anchoring model to an experimental data set, and then use these estimates to generate predictions for order decisions in a separate data set, as an out-of-sample test. As a further robustness check, we conduct our own experiment, which considers a unique setting with asymmetric two-point demand, and fit the model to its data as well.Results: Our model fits both the mean and variance of order quantities well across the different data sets. We also find a consistent pattern of subjects anchoring heavily on their demand forecast when the product profit margin is high. Moreover, our results show that the main cause of order variability at the individual level is random point forecasts, whereas the order variability at the aggregate level is driven by both random point forecasts and variable adjustment levels.Managerial Implications: Our study provides a new perspective for explaining and predicting order variability in inventory decisions. By providing better predictions, our model can help the upstream supply chain parties better anticipate the order variability from the downstream buyers and thus improve overall profitability