Bundling and Quality Assurance
Kathryn Spier (Harvard University, United States) and James Dana (Northeastern University)

We consider a repeated moral hazard model of product quality choice with multiproduct firms and imperfect private monitoring by consumers. When consumers are small, receive imperfect private signals of product quality, and have heterogeneous preferences, consuming multiple products from the same firm improves monitoring. But monitoring is a public good, and consumers free ride on each other’s monitoring. Product bundling improves product quality by constraining consumers to use better monitoring and punishment strategies. The social and private value of bundling is even larger if one of the two goods is a durable and the other is a complementary non-durable, or if the consumer cannot attribute negative signals to individual products. Joint with J. Dana (Northeastern University).

*********

Wald Deconstructed: The Organization of Persuasion
Marco Ottaviani (Bocconi University, Italy)

An agent sequentially collects information at a cost to influence a decision maker’s choice between rejection and approval. We model this situation through an organizational deconstruction of Wald’s classic model of sequential information acquisition. The payoff and control rights of the Wald statistician are split between the agent and the decision maker. The decision maker prefers approval in the good state and rejection in the bad state and thus benefits from information diffusion. The agent controls the information and pays for it, but benefits from approval regardless of the state and thus has no direct value for information. The agent’s desire to persuade the decision maker induces an indirect demand for information. We characterize the outcome achieved by different organizational structures from a positive and normative perspective, depending on the commitment power of the agent and the decision marker. We also examine situations where the agent can misrepresent information at a cost and we show that it is socially optimal to tolerate a positive amount of misrepresentation. Our results apply to the interaction between an upstream research division and a downstream sales division of a firm, pharmaceutical companies and the drug safety regulator, and merging firms and the competition authority. Joint with Emeric Henry (Sciences Po)