Utilizing managerial beliefs for set identification of price elasticities of demand

Data-driven decision-making is increasingly prevalent but can clash with managerial beliefs, risking biased decisions. A prime example is pricing strategy optimization, where traditional methods for estimating price elasticities of demand often lead to counter-intuitive results due to model misspecification and the reliance on single-point estimates. To address this, we propose utilizing structural vector-autoregressions (SVARs) to generate identified sets of elasticities, integrating managerial beliefs into the analysis to improve decision-making processes. Using weak restrictions about the directional effects of supply and demand shocks on sales and prices, and assumptions about the functioning of in-store promotions effectively sharpens the identified sets. Specifically, we analyze the demand for beer at a large scale for 1,953 stores in the US. For many stores (i.e., at least 40%), both recent endogeneity-robust single-equation methods and alternative identification strategies for SVARs used in marketing studies yield positive price elasticity estimates that oppose behavioral fundamentals. Hence, these are hardly informative for designing pricing strategies. Instead, the suggested approach to set identification yields elasticity estimates that are sufficiently precise to improve the design of retail pricing strategies and offer insights into customer's distinct price sensitivities in grocery and drug stores. Overall, our approach emphasizes the importance of combining data-driven analysis with managerial insights for evidence-based decision-making.