International audience

Despite the growing literature on behavioral inventory problems, there is a surprising lack of research in dynamic settings. Focusing on this gap, we consider a multi-period inventory system with a decision maker characterized by overestimation and overprecision, two key dimensions of overconfidence. The decision maker faces random demand, which follows an AR(1) process. The decision maker forecasts future demand using the minimum mean square error method and utilizes an order-up-to policy to determine inventory levels. Crucially, the replenishment lead time is also stochastic, following any possible discrete probability distribution, and the decision maker forecasts the stochastic lead time either with an expectation-based approach or a moving average. Analysis of our main model reveals that overestimation and overprecision have different impacts on the bullwhip effect, which depends on the degree of autocorrelation. Different lead time forecasting methods also further alter the influence of overconfidence on the bullwhip effect. Moreover, we show that expectation-based forecasts generally lead to a lower bullwhip effect than moving averages, but when demand exhibits autocorrelation, moving averages can yield lower bullwhip effects under specific conditions. Overall, our findings offer strategic guidance to decision makers in inventory management, highlighting how overconfidence and lead time forecasting choices interact to shape the bullwhip effect.