Abstract
Although the theoretical inventory literature has studied the impact of transshipments on inventory, the common assumptions of a small number of products and fixed fill rates in these analytical studies may limit our understanding of transshipments’ value. To complement these studies, we empirically evaluate the extent to which transshipments contribute to inventory turnover and order fulfillment. We develop empirical models and estimate them using proprietary data from the distribution centers of a distribution company. To alleviate the potential selection bias, we apply Coarsened Exact Matching (CEM) to match distribution centers with transshipments with those with no transshipments. Our analyses jointly evaluate multiple performance outcomes and provide empirical evidence for the theoretical value of transshipments. We find that on average, transshipments raise inventory turnover by 12.8% and improve fill rate by 0.8%. The counterintuitive findings imply that unlike transshipments between retail locations, transshipments between distribution centers may gain inventory and order fulfillment benefits. Furthermore, we analyze these contributions differently across three transshipment decisions: timing (proactive vs. reactive), frequency, and quantity. To explore the mechanism behind these changes, we investigate the variations in the inventory level and find that, on average, the inventory level remains unchanged with transshipments. However, the inventory level at distribution centers during the time periods when reactive transshipments occurred is lower than that in the time periods when proactive transshipments happened. Moreover, a higher transshipment frequency reduces the inventory level, but a larger transshipment quantity raises it.
