Do allocation mechanisms drive strategic ordering? The case of integrated distribution systems (joint work with Karen Donohue and Mustafa Cagri Gurbuz)
We study the impact of three well-known inventory allocation mechanisms, i.e. proportional, linear and uniform, on the ordering behavior of retailers serviced from a central distribution center. Based on the allocation scheme, retailers may have an incentive to strategically inflate/deflate their orders to gain a more favorable allocation, a behavior that may reduce allocation efficiency from a system’s perspective. We find that while all three allocation rules are system optimal under common knowledge of local demands, only the uniform rule incentivizes retailers to place an order equal to local demand. Consistent with theory, our experimental results show that the proportional and linear rules result in larger and more frequent order inflation/deflation, with the degree of strategic ordering being largest under the linear rule. Strategic ordering not only decreases system efficiency but also a retailer’s own profits. Motivated by these results, we conclude by sketching a practical, tailored, application of the uniform allocation principle. click here
Inventory rationing across sales channels (joint work with Nienke Hofstra)
When demand exceeds available inventory, suppliers ration their inventory among sales channels. How does inventory risk and payment schemes affect allocation decisions in such cases? We conduct lab experiments to study the role of risk aversion, loss aversion and mental accounting in inventory allocation decisions between two channels. On average, subjects allocate significantly less inventory than the expected profit maximizing quantity to the risky, yet more profitable, channel. Subjects with stronger risk appetite allocate larger quantities to the risky channel, but also exhibit higher loss aversion. Risk attitude moderates the effect of the timing of payments on allocation decisions.
Ordering decisions under supply uncertainty and inventory record inaccuracy (joint work with Nienke Hofstra and Sander de Leeuw)
How do decision makers react to inventory uncertainty when placing orders? In this paper, we investigate ordering decisions under two forms of uncertainty regarding total available inventory to satisfy demand: a) supply uncertainty (unreliability in the external supply system), and b) inventory record inaccuracy (internal inefficiencies that lead to a discrepancy between physical and recorded inventory levels). In incentivized laboratory experiments, we find that under both forms of inventory uncertainty, subjects overstock for low-margin products, while for high-margin products they place orders that are around the optimal quantity. We propose that this can be explained by a combination of biases in probability assessment: the availability heuristic and overconfidence. We also find that for low-margin products subjects overstock more under inventory record inaccuracy than under supply uncertainty. This can be explained by a stronger aversion to leftovers under supply uncertainty (i.e., when uncertainty is external) than under inventory record inaccuracy (i.e., when uncertainty is internal). Given that both forms of uncertainty are often simultaneously present in practice, but their causes are different, this study highlights the importance of reducing inventory record inaccuracy for products with low-profit margins.
Fairness ideals in resource allocation: the case of shared inventory
We study fairness ideals regarding inventory allocation among retailers who are serviced from the same inventory pool. In particular, we focus on (a) what is considered a fair allocation in this setting (e.g., equal profit, same fill rate, equal share of inventory shortage/surplus?) and (b) how much weight decision makers attach to fairness considerations. Experimental data shows that fairness considerations play a role in the allocations proposed, with participants proposing an own inventory allocation that is not exactly equal to their demand in about 40% of the instances. Furthermore, participants seem to base their proposals for inventory split on realized demands rather than on total profit comparisons. While there is considerable heterogeneity in fairness preferences, the most prevalent fairness ideal is that of equal split of total inventory-demand mismatch. Participants place, on average, 1.4 times more importance on their own income over fairness considerations. We also find that random pre-selection of participants to decide the inventory allocation induces more selfish behavior.