Decentralized Decision-Making in Retail Chains: Evidence from Inventory Management

This paper investigates the impact of decentralizing inventory decision-making in multi-establishment firms using data from a large retail chain. Analyzing two years of daily data, we find significant heterogeneity among the inventory decisions made by 634 store managers. By estimating a dynamic structural model, we reveal substantial heterogeneity in managers' perceived costs. Moreover, we observe a correlation between the variance of these perceptions and managers' education and experience. Counterfactual experiments show that centralized inventory management reduces costs by eliminating the impact of managers' skill heterogeneity. However, these benefits are offset by the negative impact of delayed demand information.