Sharing Retailers’ Demand Information with Privacy Protection

Problem definition: This paper investigates the issue of sharing private demand information between two complementary retailers considering information protection. Academic/practical relevance: In the existing literature, firms decide whether to share their private information. However, they cannot determine the precision of shared information. Methodology: We develop a game-theoretic model with two retailers, endogenize the degree of information sharing, and solve for the game's equilibrium. We consider three scenarios: no sharing (NN), full sharing (SS), and partial sharing (SN). Results: We show that under ex-ante information sharing, scenario NN and scenario SS will be Nash equilibria, while under ex-ante information sharing, all three scenarios can be equilibrium information-sharing outcomes. What is counterintuitive is that consumer welfare is better when sharing information if complementary is strong and privacy cost is low. Managerial implications: Our results provide novel political implications that the application of federated learning can create a win-win situation, which not only benefits firms’ profit but also improves consumer welfare.