Trust,reputation and bidding behaviour in online mystery auctions

We use transactions from a distinctive online environment of ‘mystery’ auctions to examine the role that trust plays and how it impacts bidding behaviour when the exact characteristics of a good being auctioned are purposefully concealed from buyers. We show that buyers are generally trusting seller claims in online transactions and that seller reputation becomes significantly more important to buyers (as demonstrated by their bids) when the quality (or value) of the good is unspecified. Our findings can be extrapolated to consider broader economic implications of bidding behaviour impacted by trust, such as in financial markets, where over-bidding may lead to price bubbles.     

Dynamic market for lemons with endogenous quality choice by the seller

We analyze a dynamic market for lemons in which the quality of the good is endogenously determined by the seller. Potential buyers sequentially submit offers to one seller. The seller can make an investment that determines the quality of the item at the beginning of the game, which is unobservable to buyers. At the interim stage of the game, the information and payoff structures are the same as in the market for lemons. Our main result is that the possibility of trade does not create any efficiency gain if (i) the common discounting is low, and (ii) the static incentive constraints preclude the mutually agreeable ex-ante contract under which the trade happens with probability one. Our result does not depend on whether the offers by buyers are private or public.     

Communication in repeated games with costly monitoring

We study repeated games with discounting where perfect monitoring is possible, but costly. It is shown that if players can make public announcements, then every payoff vector which is an interior point in the set of feasible and individually rational payoffs can be implemented in a sequential equilibrium of the repeated game when the discount factor is high enough. Thus, efficiency can be approximated even when the cost of monitoring is high, provided that the discount factor is high enough.     

Reputation for quality and adverse selection

We analyze a dynamic market with a seller who can make a one-time investment that affects the returns of tradable assets. The potential buyers of the assets cannot observe the seller׳s investment prior to the trade or verify it in any way after the trade. The market faces two types of inefficiency: the ex-ante inefficiency, i.e., the seller׳s moral hazard problem, and the ex-post inefficiency, i.e., inefficient ex-post allocations due to the adverse selection problem. We analyze how the observability of information by future buyers, through which the seller builds a reputation, affects the two types of inefficiency as well as the interplay between them.     

Sequential selling mechanisms

Summary In this paper we attempt to formalize the idea that a mechanism that involves multilateral communication between buyers and sellers may be dominated by one that involves simple bilateral communication. To do this we consider the well known problem in which a seller tries to sell a single unit of output to a group ofN buyers who have independently distributed private valuations. Our arguments hinge on two considerations. First, buyers communicate their willingness to negotiate with the seller sequentially, and second, buyers have the option of purchasing the good from some alternative supplier. It is shown that the seller cannot improve upon a procedure in which she offers the good to each buyer in turn at a fixed price. The seller reverts to multilateral communication if possible, only when no buyer is willing to pay the fixed price. In reasonable environments buyers will be too impatient to wait for the outcome of a multilateral negotiation and all communications will be bilateral.In many problems in mechanism design, informed traders have no alternative to participation in the mechanism that is offered by its designer. The best mechanism from the designer's point of view is then the one that is most efficient at extracting informational rents, that is, a simple auction. In a competitive environment it is likely to be costly for buyers to participate in an auction or any other multilateral selling scheme in which the seller must process information from many different buyers because alternative trading opportunities will be disappearing during the time that the seller is collecting this information. Buyers might be willing to participate in an auction, but only if they could be guaranteed that the competition that they face will not eliminate too much of their surplus.At the other extreme to the auction is a simple fixed price selling scheme 1. The seller simply waits until he meets a buyer whose valuation is high enough, given the opportunities that exist in the rest of the market, for him to be willing to pay this price. The seller extracts the minimum of the buyer's informational rents since the price that a buyer pays is independent of his valuation. Yet the seller might like this scheme if adding a second bidder to the process makes it very difficult for him to find a buyer with a valuation high enough to want to participate.In the presence of opportunity costs, the seller faces a trade-off between his ability to extract buyers informational rents and his ability to find buyers who are willing to participate in any competitive process. In practice this trade-off will impose structure on the method that is used to determine a price. In markets where there are auctions, limits are put on buyer participation. In tobacco auctions bids are submitted at a distinct point in time from buyers who are present at that time. In real estate auctions time limits are put on the amount of time the seller will wait before making a decision. These restrictions on participation are presumably endogenously selected by the seller (possibly in competition with other mechanism designers) with this trade-off in mind.On the other hand, markets in which objects appear to trade at a fixed price are rarely so simple. A baker with a fixed supply of fresh bagels is unlikely to collect bids from buyers and award the bagels to the high bidder at the end of the day. Buyers are unlikely to be willing to participate in such a scheme since they can buy fresh bagels from a competitor down the street. Yet despite the fact that bagels sell at a fixed price throughout the day, most bakers are more than willing to let it be known that they will discount price at the end of the day on any bagels that they have not yet sold. Selling used cars presents a similar problem. Each potential buyer for the used car is likely to have inspected a number of alternatives, and is likely to know the prices at which these alternative can be obtained. A seller who suggests that buyers submit a bid, then wait until the seller is sure that no higher offer will be submitted is asking buyers to forgo these alternative opportunities with no gain to themselves. To avoid the rigidity of the pure fixed price scheme most used cars are sold for a fixed price or best offer. These examples suggest that the best selling mechanism may involve a complex interplay between participation and surplus extraction considerations.The purpose of this paper is to provide a simple formalism within which the factors that determine the best contract can be evaluated. We consider the best known environment from the point of view of auction design in which there are a large number of buyers with independent private valuations for a unit of an indivisible commodity that is being sold by a single supplier who acts as the mechanism designer. We modify this standard problem in two critical ways. First, we assume that the seller meets the potential buyers sequentially rather than all at once. Secondly we assume that buyers have a valuable alternative that yields them a sure surplus. This creates a simple bidding cost that is effectively the expected loss in surplus (created by the disappearance of outside alternatives) that the buyer faces during the time that he spends negotiating with the seller.These simple assumptions allow us to calculate the impact of competition and communication costs using completely standard arguments from the mechanism design literature. We are able to show that with these assumptions the seller's expected surplus will be highest if the object is sold according to the following modified fixed price scheme: the seller contacts each of the potential buyers in turn and either offers to negotiate or announces that he no longer wishes to trade. If he offers to negotiate and the buyer agrees, the buyer immediately has the option of trading for sure with the seller at a fixed price set ex ante. If the buyer does not wish to pay this fixed price, he may submit an alternative bid. The seller will then continue to contact new buyers, returning to trade with the buyer only if no buyer wishes to pay the fixed price and no higher bid is submitted.It will be clear that in our environment, both the simple fixed price scheme and the simple auction are feasible. The simple auction prevails when the fixed price is set equal to the maximum possible valuation, while the simple fixed price scheme occurs when the fixed price is set so that buyers are willing to participate if and only if they are willing to pay the fixed price. Our results will show that a simple auction in never optimal for the seller. The seller can always strictly improve his payoff by moving to a scheme in which there is some strictly positive probability that trade will occur at the fixed price. On the other hand, there are reasonable circumstances in which the seller cannot achieve a higher payoff than the one she gets by selling at a fixed price. It is shown that for any positive participation cost, there is a large, but finite, number of potential buyers so that the seller cannot achieve a higher payoff than what she gets by selling at a fixed price. Two simple, but important continuity results are also illustrated. As the cost of participation in the mechanism increases (decreases), the probability with which the seller's unit of output is sold at a fixed price goes to one (zero) in the best modified fixed price mechanism for the seller.Our paper is not the first to generate such a modified fixed price scheme. Both McAfee and McMillan (1988) and Riley and Zeckhauser (1983) come up with similar schemes for the case in which the seller must bear a fixed cost for each new buyer that she contacts. There are two essential differences between our model and theirs. First, as the cost is interpreted as the opportunity cost of participation in the mechanism, it is reasonable to imagine that the seller advertises the mechanism ex ante. Another way of putting this is that the seller pays a fixed rather than a variable cost to communicate the mechanism to buyers. This makes it possible to assume that the mechanism is common knowledge to the seller and all the buyers at the beginning of the communication process. For this reason we can make our case using completely standard arguments. Secondly, the mechanism in the opportunity cost case plays a different allocative role than it does in the case when the seller bears a cost. The mechanism must decide whether buyers should communicate with the seller or pursue their alternative activities, as well as who should trade and at what price. It is this allocative role that makes bilateral communication superior to multilateral communication in a competitive environment. These differences allow us to show, for example, that a simple fixed price scheme is undominated for the seller when the number of buyers is finite. As shown by McAfee and McMillan, this is only possible when the number of potential buyers is infinite when the seller bears the cost of communication.Remarkably, the existence of opportunity costs to buyer participation is not, by itself, sufficient to explain why sellers might prefer bilateral communications mechanisms. Samuelson (1983) and McAfee and McMillan (1987) show that when buyers must pay a fixed cost to submit a bid, which is equivalent to giving up a valuable alternative, a seller cannot expect to earn more than she does in a second price auction (though Samuelson shows that the reserve price may depend on the number of potential buyers). One of the contributions of this paper is to show that the assumption that buyers make their participation decisions simultaneously is critical to this result. Simultaneous entry decisions means that whether or not any particular buyer is assigned to the alternative activity is independent of any other buyer's valuation. With sequential communication the seller is able to relax this constraint. It is precisely the enlargement of the class of feasible mechanism that breaks down the optimality of the simple auction.The second author acknowledges the support of the Social Sciences and Humanities Research Council of Canada and the CRDE at the Université de Montreal.     

Word-Of-Mouth Communication and Community Enforcement

We study a repeated game where a seller, who has a short-term incentive to supply low quality, is periodically matched with a randomly selected buyer. Buyers observe only the outcomes of their neighbors' games and may receive signals from them. When the buyer population is large, the seller may sell high quality even when each buyer observes her action in any given period with an arbitrarily small probability. When networking among buyers is costly, low quality is always supplied with a positive probability. For this case, we characterize an equilibrium where the seller randomizes between high and low quality.     

Correlated Equilibrium in Stochastic Games

We study the existence of uniform correlated equilibrium payoffs in stochastic games. The correlation devices that we use are either autonomous (they base their choice of signal on previous signals, but not on previous states or actions) or stationary (their choice is independent of any data and is drawn according to the same probability distribution at every stage). We prove that any n-player stochastic game admits an autonomous correlated equilibrium payoff. When the game is positive and recursive, a stationary correlated equilibrium payoff exists. Journal of Economic Literature Classification Numbers: C72, C73.     

Mechanism Design with Interdependent Valuations: Surplus Extraction

If valuations are interdependent and agents observe their own allocation payoffs, then two-stage revelation mechanisms expand the set of implementable decision functions. In a two-stage revelation mechanism agents report twice. In the first stage - before the allocation is decided - they report their private signals. In the second stage - after the allocation has been made, but before final transfers are decided - they report their payoffs from the allocation. Conditions are provided under which an uninformed seller can extract (or virtually extract) the full surplus from a sale to privately informed buyers, in spite of the buyers’ signals being independent random variables. This research was started when I was visiting the Department of Applied Mathematics of the University of Venice, and continued while visiting the European University Institute in Florence. Their financial support is gratefully acknowledged.     

Learning to play games in extensive form by valuation

Game theoretic models of learning which are based on the strategic form of the game cannot explain learning in games with large extensive form. We study learning in such games by using valuation of moves. A valuation for a player is a numeric assessment of her moves that purports to reflect their desirability. We consider a myopic player, who chooses moves with the highest valuation. Each time the game is played, the player revises her valuation by assigning the payoff obtained in the play to each of the moves she has made. We show for a repeated win-lose game that if the player has a winning strategy in the stage game, there is almost surely a time after which she always wins. When a player has more than two payoffs, a more elaborate learning procedure is required. We consider one that associates with each move the average payoff in the rounds in which this move was made. When all players adopt this learning procedure, with some perturbations, then, with probability 1 there is a time after which strategies that are close to subgame perfect equilibrium are played. A single player who adopts this procedure can guarantee only her individually rational payoff.     

The Sale of Ideas: Strategic Disclosure, Property Rights, and Contracting

Ideas are difficult to sell when buyers cannot assess an idea's value before it is revealed and sellers cannot protect a revealed idea. These problems exist in a variety of intellectual property sales ranging from pure ideas to poorly protected inventions and reflect the nonverifiability of key elements of an intellectual property sale. An expropriable partial disclosure can be used as a signal, allowing the seller to obtain payment based on the value of the remaining (undisclosed) know-how. We examine contracting after the disclosure and find that seller wealth is pivotal in supporting a partial disclosure equilibrium and in determining the payoff size.     

On the folk theorem with one-dimensional payoffs and different discount factors

Proving the folk theorem in a game with three or more players usually requires imposing restrictions on the dimensionality of the stage-game payoffs. Fudenberg and Maskin (1986) assume full dimensionality of payoffs, while Abreu et al. (1994) assume the weaker NEU condition (“nonequivalent utilities”). In this note, we consider a class of n-player games where each player receives the same stage-game payoff, either zero or one. The stage-game payoffs therefore constitute a one-dimensional set, violating NEU. We show that if all players have different discount factors, then for discount factors sufficiently close to one, any strictly individually rational payoff profile can be obtained as the outcome of a subgame-perfect equilibrium with public correlation.     

Reputation with observed actions

Summary A single long-run player plays a fixed stage game (simultaneous orsequential move) against an infinite sequence of short-run opponents that play only once but can observe all past realized actions. Assuming that the probability distributions over types of long and short-run players have full support, we show that the long-run player can always establish a reputation for theStackelberg strategy and is therefore guaranteed almost his Stackelberg payoff in all Nash equilibria of the repeated game.The financial support of the National Science Foundation, Grant SES 90-7999, and of Consiglio Nazionale delle Ricerche is gratefully acknowledged. I wish to thank David Levine, Wolfgang Pesendorfer and Seminar Participants at UCLA, Universidad Carlos III de Madrid and University of Naples for useful discussions and suggestions.     

The effects of heterogeneous interaction and risk attitude adaptation on the evolution of cooperation

This paper addresses the evolution of cooperation in a multi-agent system with agents interacting heterogeneously with each other based on the iterated prisoner’s dilemma (IPD) game. The heterogeneity of interaction is defined in two models. First, agents in a network are restricted to interacting with only their neighbors (local interaction). Second, agents are allowed to adopt different IPD strategies against different opponents (discriminative interaction). These two heterogeneous interaction scenarios are different to the classical evolutionary game, in which each agent interacts with every other agent in the population by adopting the same strategy against all opponents. Moreover, agents adapt their risk attitudes while engaging in interactions. Agents with payoffs above (or below) their aspirations will become more risk averse (or risk seeking) in subsequent interactions, wherein risk is defined as the standard deviation of one-move payoffs in the IPD game. In simulation experiments with agents using only own historical payoffs as aspirations (historical comparison), we find that the whole population can achieve a high level of cooperation via the risk attitude adaptation mechanism, in the cases of either local or discriminative interaction models. Meanwhile, when agents use the population’s average payoff as aspirations (social comparison) for adapting risk attitudes, the high level of cooperation can only be sustained in a portion of the population (i.e., partial cooperation). This finding also holds true in both of the heterogeneous scenarios. Considering that payoffs cannot be precisely estimated in a realistic IPD game, simulation experiments are also conducted with a Gaussian disturbance added to the game payoffs. The results reveal that partial cooperation in the population under social comparison is more robust to the variation in payoffs than the global cooperation under historical comparison.     

Uncoupled Aspiration Adaptation Dynamics Into the Core

Dynamics for play of transferable‐utility cooperative games are proposed that require information regarding own payoff experiences and other players’ past actions, but not regarding other players’ payoffs. The proposed dynamics provide an evolutionary interpretation of the proto‐dynamic ‘blocking argument’ (Edgeworth, 1881) based on the behavioral principles of ‘aspiration adaptation’ (Sauermann and Selten, 1962) instead of best response. If the game has a non‐empty core, the dynamics are absorbed into the core in finite time with probability one. If the core is empty, the dynamics cycle infinitely through all coalitions.     

On uniquely implementing cooperation in the prisoners' dilemma

Summary We consider the problem of a principle who wishes to induce two agents playing a one shot prisoner's dilemma to behave cooperatively. We assume that the principal cannot observe the actions of the agents, and is not able to change the strategy sets or payoff functions in the underlying game. The only power the principle has is to randomly delay the arrival of payoffs. Specifically, agents choose their one shot strategies, and then the principle randomly determines whether these are cheap talk, or if payoffs should be distributed. If the round is cheap talk, then each agent observes the strategy choice of the other and play moves to a new round. This continues until payoffs are distributed. We establish conditions under which the probability of cheap talk can be chosen at the beginning of the induced game in such a way that full cooperation is the only equilibrium outcome. The sufficiency condition is met by a wide class of economic interpretations of the prisoners' dilemma, including those involving strategic complementarities among players.The authors wish to thank Dilip Abreu, Robert Aumann, Michael Baye, James Friedman, Richard McLean, Herve Moulin, Ariel Rubinstein, Rajiv Vohra and Simon Wilkie for their comments. Also, participants in seminars and conferences at Arizona, Bellcore, Brown, Illinois, Northwestern, Princeton, Rochester, Vanderbilt and West Virginia have provided stimulating comments. We also thank the referee for many detailed and useful suggestions. The third author's research was supported in part by NSF grant SES-9213145.     

Optimal Information Disclosure in Auctions and the Handicap Auction

We analyse a situation where a monopolist is selling an indivisible good to risk-neutral buyers who only have an estimate of their private valuations. The seller can release, without observing, certain additional signals that affect the buyers' valuations. Our main result is that in the expected revenue-maximizing mechanism, the seller makes available all the information that she can, and her expected revenue is the same as it would be if she could observe the part of the information that is "new" to the buyers. We also show that this mechanism can be implemented by what we call a handicap auction in interesting applications. In the first round of this auction, each buyer picks a price premium from a menu offered by the seller (a smaller premium costs more). Then the seller releases the additional signals. In the second round, the buyers bid in a second-price auction where the winner pays the sum of his premium and the second highest non-negative bid. In the case of a single buyer, this mechanism simplifies to a menu of European call options.     

Impermanent types and permanent reputations

We study the impact of unobservable stochastic replacements for the long-run player in the classical reputation model with a long-run player and a series of short-run players. We provide explicit lower bounds on the Nash equilibrium payoffs of a long-run player, both ex-ante and following any positive probability history. Under general conditions on the convergence rates of the discount factor to one and of the rate of replacement to zero, both bounds converge to the Stackelberg payoff if the type space is sufficiently rich. These limiting conditions hold in particular if the game is played very frequently.     

Dynamic common agency

A general model of dynamic common agency with symmetric information is considered. The set of truthful Markov perfect equilibrium payoffs is characterized and the efficiency properties of the equilibria are established. A condition for the uniqueness of equilibrium payoffs is derived for the static and the dynamic game. The payoff is unique if and only if the payoff of each principal coincides with his marginal contribution to the social value of the game. The dynamic model is applied to a game of agenda setting.     

Absorbing Team Games

A team is a group of people having the same motives but possibly different available actions. A team game is a game where two teams face each other. An absorbing game is a repeated game where some of the entries are absorbing, in the sense that once they are chosen the play terminates, and all future payoffs are equal to the payoff at the stage of termination. We prove that every absorbing team game has an equilibrium payoff and that there are -equilibrium profiles with cyclic structure. Journal of Economic Literature Classification Numbers: C72, C73.     

Coexistence of Strategies and Culturally-Specific Common Knowledge: An Evolutionary Analysis

We analyze social dynamics in a continuous population where randomly matched individuals have to choose between two pure strategies only ('cooperate' (C) and 'not cooperate' (NC)). Individual payoffs associated with the possible outcomes of each interaction may differ across groups, depending on the specific social and cultural context to which each agent belongs. In particular, it is assumed that three sub-populations are initially present, 'framing' the game according to the prisoner's dilemma (PD), assurance game (AG) and other regarding (OR) payoff configurations, respectively. In other words, we assume that common knowledge about the payoffs of the game is 'culturally-specific'. In this context, we examine both the adoption process of strategies C and NC within each sub-population and the diffusion process of 'types' (PD, AG and OR) within the overall community. On the basis of an evolutionary game-theoretic approach, the paper focuses on the problem of coexistence of PD, AG and OR groups as well as of 'nice' (C) and 'mean' (NC) strategies. We show that coexistence between C and NC is possible in the heterogeneous community under examination, even if it is ruled out in homogeneous communities where only one of the three types is present. This revised version was published online in July 2006 with corrections to the Cover Date.