An empirically-simulated investigation of the impact of demand forecasting on the bullwhip effect: Evidence from U.S. auto industry

This study empirically examines the impacts of three major aspects of demand forecasting on the magnitude of the bullwhip effect. Three research questions are addressed to investigate the association between (1) forecast accuracy, (2) aggregate forecasting, and (3) responsive forecasting and the bullwhip effect. Using forecasted demand generated from popular time-series forecasting models and real-life demand data, the study investigates the relationship between the forecasted results and the consequential bullwhip effect. The findings show that the forecasting methods used lead to the variation of the bullwhip effect. Moreover, the lead time reduction and the stable demand forecast are beneficial to reduce the bullwhip effect. However, our empirical results differ from previous findings in two ways: (i) improving forecast accuracy does not necessarily reduce the bullwhip effect and (ii) aggregate forecasting does not always reduce the bullwhip effect.