Résumé
We study a mechanism design problem where the principal can also manipulate the agent’s information about a payoff-relevant state. Jointly designing information and allocation rule is proved equivalent to certain multi-dimensional screening problem. Based on this equivalence, when the agent’s types are positively-related, full disclosure is proved optimal under regularity conditions; while with negatively-related types, the optimal disclosure policy takes the form of a bad-state alert, which is in general a type-contingent disclosure policy. In a binary environment, we fully charac- terize the optimal mechanisms and discuss when type-contingent disclosure strictly benefits the principal and its welfare consequences.
Mots-clés
Information design; Bayesian persuasion; Mechanism design;
Codes JEL
- C72: Noncooperative Games
- D82: Asymmetric and Private Information • Mechanism Design
- D86: Economics of Contract: Theory
Référence
Takuro Yamashita et Shuguang Zhu, « Type-contingent Information Disclosure », TSE Working Paper, n° 21-1242, août 2021.
Voir aussi
Publié dans
TSE Working Paper, n° 21-1242, août 2021