Motivated by the emergence of the Internet‐enabled inventory sharing across firms, we investigate different commitment scenarios of decentralized inventory sharing platforms, through a behavioral lens. In particular, we consider two transfer price commitment settings (ex ante or ex post depending on whether or not the inventory transfer price is committed before demand realization) and two sharing commitment rules (automatic or voluntary depending on whether or not inventory sharing are pre‐committed). Our experimental results suggest that individuals set transfer prices much lower, and order much less, than what Nash equilibrium predicts. We also find substantial treatment effects that the rational model cannot explain. The magnitude of disparities relative to the Nash equilibrium prediction appears to be most substantial in the situation where transfer price is set ex ante (i.e., transfer price commitment) and inventory sharing is voluntary (i.e., no sharing commitment). Motivated by these observations, we develop a behavioral model that incorporates quantal response equilibrium and fairness concerns. Empirical analysis indicates that our model provides a compelling explanation of the behavior observed in the data. This study provides implication on the design of the commitment rules in decentralized inventory‐sharing platforms. Specifically, in order to gain the most benefit from inventory sharing, parties should either postpone the decision of transfer price until the need for sharing arises, or pre‐commit to sharing.
