Long-run Relationships without Money – LRWM

The application’s theme is mechanism design without transfers. This leaves two instruments to the designer: allocation of goods, and provision of information. It proposes to apply and develop tools in game theory and economic theory to understand how private information can be elicited and used to motivate economic agents. Relative to standard mechanism design, the focus is on information as an instrument and on dynamic environments. Consider:
- Report cards in health care: this is a tool used to make information about surgeons’ performance more widely available and help patients make informed choices. But its impact on the surgeons’ incentives is unclear: they might turn down severely ill patients to protect their records; use low-risk procedures rather than appropriate ones. Or indeed put in additional effort and care to avoid malpractice.
- Recommendation systems: Online sellers and search engines must provide information to users who are the main source of evaluation of the information’s quality. Hence, there is a tension between the sellers’ incentives to collect and their incentives to provide reliable information.
- Performance appraisal: firms provide feedback to their workers, as a way of motivating and directing their effort choices. Career concerns rely on feedback that is neither too noisy (so that past effort gets recognized) nor too precise (as otherwise past effort won’t affect expectations regarding future effort choices). How should the motivational feedback system be designed?
- Credit rating agencies must provide information to channel resources towards profitable projects. But bad ratings compound the difficulties of low-return projects, as the higher interest rates lead to more bankruptcies.
These examples illustrate the complexity of designing optimal information channels. Providing all information might be as detrimental as no information. The goal is to understand how the economic environment shapes this design. When should information be partitioned into grades? When should it be garbled? What kind of information should be disclosed, and what kind shouldn’t?

This application is divided into three topics:
1. Agency under “pure” private information, with or without money.
2. Markets with unknown information (Learning).
3. Markets with hidden actions (Moral hazard)

The application involves primarily game theory and mechanism design, but interactions with industrial organization, market design and information theory are called for.