Publications des chercheurs de PSE

This paper studies the effects of analogy-based expectations in static two-player games of incomplete information. Players are assumed to be boundedly rational in the way they forecast their opponent's state-contingent strategy: they bundle states into analogy classes and play best-responses to their opponent's average strategy in those analogy classes. We provide general properties of analogy-based expectation equilibria and apply the model to a variety of well known games. We characterize conditions on the analogy partitions for successful coordination in coordination games under incomplete information [Rubinstein, A., 1989. The electronic mail game: Strategic behavior under 'almost common knowledge'. Amer. Econ. Rev. 79, 385-391], we show how analogy grouping of the receiver may facilitate information transmission in Crawford and Sobel's cheap talk games [Crawford, V.P., Sobel, J., 1982. Strategic information transmission. Econometrica 50, 1431-1451], and we show how analogy grouping may give rise to betting in zero-sum betting games such as those studied to illustrate the no trade theorem.

We study the interaction between multiple information designers who try to influence the behavior of a set of agents. When each designer can choose information policies from a compact set of statistical experiments with countable support, such games always admit subgame perfect equilibria. When designers produce public information, every equilibrium of the simple game in which the set of messages coincides with the set of states is robust in the sense that it is an equilibrium with larger and possibly infinite and uncountable message sets. The converse is true for a class of Markovian equilibria only. When designers produce information for their own corporation of agents, robust pure strategy equilibria exist and are characterized via an auxiliary normal form game in which the set of strategies of each designer is the set of outcomes induced by Bayes correlated equilibria in her corporation.

We use the analogy-based expectation equilibrium (Jehiel, 2005) to study cheap talk from a sender who does not perfectly understand all the messages available to him. The sender is endowed with a privately known language competence corresponding to the set of messages that he understands. For the messages that he does not understand, the sender has correct but only coarse expectations about the equilibrium response of the receiver. An analogy-based expectation equilibrium is always a Bayesian solution but usually differs from a standard communication equilibrium and from an equilibrium with language barriers (Blume and Board, 2013). We characterize conditions under which an outcome remains an equilibrium outcome when the sender's competence decreases. Partial language competence rationalizes information transmission and lies in pure persuasion problems, and can facilitate information transmission from a moderately biased sender.

Cet article étudie les mécanismes de communication optimaux sans transfert lorsque l'espace de décision est multidimensionnel. Contrairement au cas où la décision est unidimensionnelle, lorsqu'un principal contrôle les deux activités de son agent (cas coopératif), le mécanisme de communication optimal est séparateur et la politique optimale de l'agent n'est jamais choisie. Cependant, lorsque les conflits d'intérêts entre l'agent et le principal sont proches sur chaque dimension, des mécanismes plus simples, qui généralisent ceux du cas unidimensionnel, sont approximativement optimaux. Ces mécanismes simples ne sont plus séparateurs sur tout l'intervalle. Lorsque chaque activité de l'agent est contrôlée par un principal différent (cas non coopératif) et entre séparément dans la fonction d'utilité de l'agent, les mécanismes optimaux lorsque la communication est privée prennent la forme d'ensembles de délégation simples, comme dans le cas unidimensionnel. Au contraire, quand l'agent souhaite coordonner les deux activités, une externalité contractuelle unidirectionnelle apparaît entre les principaux. Lorsque la communication est publique, il n'existe plus d'équilibre de Nash en stratégies pures avec des mécanismes de communication continus et différentiables par morceaux. Un équilibre est cependant obtenu dans version cheap talk du modèle, sans engagement des principaux, et la comparaison entre la communication privée et publique dépend du profil des biais des principaux par rapport à la politique idéale de l'agent.

Cet article étudie des situations stratégiques dans lesquelles chaque individu cherche à choisir une action à la fois proche des actions choisies par les autres individus et proche d'un état de la nature, la proximité idéale avec cet état variant entre les agents. Avant que ce jeu de coordination soit joué, les joueurs aux préférences hétérogènes peuvent se transmettre les informations privées qu'ils détiennent sur l'état de la nature lors d'une étape de communication gratuite. La transmission stratégique d'information qui a lieu au cours de cette étape est caractérisée par un réseau de communication stratégique. Nous examinons les réseaux de communication stratégique qui peuvent émerger à l'équilibre en fonction des préférences des joueurs, de la structure d'information et de l'importance accordée à la coordination des actions et nous comparons les réseaux de communication en terme d'efficacité. En général, il n'existe pas de réseau de communication d'équilibre maximal et les réseaux de communication ne peuvent pas être ordonnés au sens de la Pareto optimalité. Cependant, le bien-être social espéré augmente toujours lorsque le réseau de communication s'élargit. Nous montrons également comment la transmission d'information est facilitée lorsque l'information privée des joueurs est certifiable et/ou lorsque la communication publique est autorisée.

This paper compares public and private information certification in a simple class of communication games with one sender and two receivers. It is shown that, contrary to the cheap talk setting of [Farrell, J., Gibbons, R., 1989. Cheap talk with two audiences. American Economic Review 79, 1214-1223], allowing certifiable statements excludes mutual discipline (i.e., full information revelation in public but not in private) but allows for mutual subversion (i.e., full information revelation in private but not in public). In the latter case, the sender is always better off with public communication, while in other situations he may prefer either private or public communication. Compared to the previous models of strategic information revelation the paper also emphasizes the role of the "common belief " consistency condition of the strong version of sequential equilibrium.

Cet article examine les incitations des firmes à s'engager à une conduite transparente (qui exclue la corruption) dans des procédures compétitives gérées par un agent corrompu. Ces procédures sont modélisées comme un concours de beauté avec information asymétrique. L'agent a un certain pouvoir discrétionnaire dans l'évaluation des offres. Il peut favoriser une firme en échange d'un pot-de-vin. Nous montrons que l'engagement conditionnel peut éliminer la corruption lorsque celle-ci est de l'extorsion pure. Sinon, lorsque la corruption peut affecter l'allocation du contrat et que le marché est peu profitable, une firme de basse qualité peut préférer ne pas s'engager. Dans cette situation l'existence d'un équilibre séparateur où seules les firmes de haute qualité s'engagent est garantie si les décisions d'engagement sont gardées secrètes. Par contre si les décisions sont publiques des conditions supplémentaires deviennent nécessaires. Généralement, un mécanisme d'engagement unilatéral qui récompense la décision d'engagement avec un bonus est moins efficace en termes de réduction de la corruption. Finalement nous montrons qu'un mécanisme qui combine conditionnalité et bonus a le potentiel d'éliminer complètement la corruption.

This paper examines the occurrence and fragility of information cascades in two laboratory experiments. One group of low informed participants sequentially guess which of two states has been randomly chosen. In a matched pairs design, another group of high informed participants make similar guesses after having observed the guesses of the low informed participants. In the second experiment, participants' beliefs about the chosen state are elicited. In equilibrium, low informed players who observe an established pattern of identical guesses herd without regard to their private information whereas high informed players always guess according to their private information. Equilibrium behavior implies that information cascades emerge in the group of low informed participants, the belief based solely on cascade guesses is stationary, and information cascades are systematically broken by high informed participants endowed with private information contradicting the cascade guesses. Experimental results show that the behavior of low informed participants is qualitatively in line with the equilibrium prediction. Information cascades often emerge in our experiments. The tendency of low informed participants to engage in cascade behavior increases with the number of identical guesses. Our main finding is that information cascades are not fragile. The behavior of high informed participants differs markedly from the equilibrium prediction. Only one-third of laboratory cascades are broken by high informed participants endowed with private information contradicting the cascade guesses. The relative frequency of cascade breaks is 15% for the situations where five or more identical guesses are observed. Participants' elicited beliefs are strongly consistent with their own behavior and show that, unlike in equilibrium, the more cascade guesses participants observe the more they believe in the state favored by those guesses.

We characterize sufficient conditions for full and decentralized disclosure of hard information in organizations with asymmetrically informed and self interested agents with quadratic loss functions. Incentive conflicts arise because agents have different (and possibly interdependent) ideal actions and different incentives to coordinate with each others. A fully revealing sequential equilibrium exists in the disclosure game if each player's ideal action is monotonic in types and types are independently distributed, but may fail to exist with non-monotonic ideal actions or correlated types. When biases between players' ideal actions are constant across states, complete information is the Pareto dominant information structure. In that case, there is a fully revealing sequential equilibrium in which informational incentive constraints are satisfied ex-post, so it exists for all possible prior beliefs, even when players' types are correlated. This existence result applies whether information disclosure is private or public, and is extended to partial certifiability of information.

We study the problem faced by firms that invest in a foreign country characterized by weak governance. Our focus is on extortion relying on the threat of expropriation and bureaucratic harassment. The bureaucrat's power is characterized by looking at a general extortion mechanism adapted from Myerson's (1981) optimal auction theory. This characterization is used to study the determinants of the quality of governance and whether and how political risk insurance of foreign direct investments improve upon it. We find that it does not always improve upon all governance indicators. It always decreases the bureaucrat's total revenue from corruption, but it may also increase the risk of expropriation and the extortion bribes paid by some firms.

This paper studies the impact of belief elicitation on informational efficiency and individual behavior in experimental parimutuel betting markets. In one treatment, groups of eight participants, who possess a private signal about the eventual outcome, play a sequential betting game. The second treatment is identical, except that bettors are observed by eight other participants who submit incentivized beliefs about the winning probabilities of each outcome. In the third treatment, the same individuals make bets and assess the winning probabilities of the outcomes. Market probabilities more accurately reflect objective probabilities in the third than in the other two treatments. Submitting beliefs reduces the favorite-longshot bias and making bets improves the accuracy of elicited beliefs. A level-k framework provides some insights about why belief elicitation improves the capacity of betting markets to aggregate information.

A firm chooses a price and the product information it discloses to a consumer whose tastes are privately known. We provide a necessary and sufficient condition on the match function for full disclosure to be the unique equilibrium outcome whatever the costs and prior beliefs about product and consumer types. It allows for products with different qualities as well as some horizontal match heterogeneity. With independently distributed product and consumer types, full disclosure is always an equilibrium and a necessary and sufficient equilibrium condition is that all firm types earn at least the full-disclosure profit.