Mechanism design with multidimensional, continuous types and interdependent valuations

We consider the mechanism design problem when agents’ types are multidimensional and continuous, and their valuations are interdependent. If there are at least three agents whose types satisfy a weak correlation condition, then for any decision rule and any ?>0 there exist balanced transfers that render truthful revelation a Bayesian ?-equilibrium. A slightly stronger correlation condition ensures that there exist balanced transfers that induce a Bayesian Nash equilibrium in which agents’ strategies are nearly truthful.     

Rent-seeking with uncertain discriminatory power

We consider a two-player Tullock rent-seeking contest with uncertain discriminatory power in the contest success function. We examine the cases where both players are either informed or uninformed about the size of the discriminatory power, as well as the case where only one player has private information about it. We show that in all three cases the contest has a unique (Bayesian) Nash equilibrium. In each case we characterize key properties of the equilibrium.     

The robustness of the independent private value model in Bayesian mechanism design

Summary. In a Bayesian model of group decision-making, dependence among the agents' types has been shown to have a beneficial effect on the design of incentive compatible mechanisms that achieve the efficient choice associated with complete information. This effect is shown here to depend as much upon the use of large monetary transfers among the agents as it does upon dependence: if the transfers are bounded in magnitude, then nonexistence in the case of independence of an efficient, incentive compatible, ex ante budget-balanced and interim individually rational mechanism is robust to the introduction of a small amount of dependence among types. This robustness result supports the use of the simplifying assumption of independence in mechanism design. Received: October 28, 1996; revised version: May 28, 1997     

Collective decisions with interdependent valuations

Many collective decision problems have in common that individuals’ desired outcomes are correlated but not identical. This paper studies collective decisions with private information about desired policies. Each agent holds private information which mainly concerns his own bliss point, but private information also affects all other agents. We concentrate on two specific mechanism, the median and mean mechanisms establish the existence of symmetric Bayesian Nash equilibria of the corresponding games and compare the performance of the mechanisms for different degrees of interdependencies. Applications of our framework include the provision of public goods and the design of decision processes in international organizations.     

An experimental comparison of collective choice procedures for excludable public goods

This paper compares three collective choice procedures for the provision of excludable public goods under incomplete information. One, serial cost sharing (SCS), is budget balanced, individually rational, anonymous and strategy proof. The other two are “hybrid” procedures: voluntary cost sharing with proportional rebates (PCS) and with no rebates (NR). PCS satisfies all these properties except strategy proofness, and NR satisfies all the properties except for strategy proofness and budget balance. However, PCS and NR do not exclude any potential users, and they do not require equal cost shares, thereby overcoming the two main sources of inefficiency with SCS. We characterize the Bayesian Nash equilibria (BNE) of the hybrid mechanisms and conduct laboratory experiments to compare the performance of the three mechanisms. We find that PCS produces significantly more efficient allocations than either SCS or NR.     

Individually rational, budget-balanced mechanisms and allocation of surplus

We investigate the issue of implementation via individually rational ex-post budget-balanced Bayesian mechanisms. We show that all decision rules generating a nonnegative expected social surplus are implementable via such mechanisms if and only if the probability distribution of the agents’ type profiles satisfies two conditions: the well-known condition of Crémer and McLean [1988. Full extraction of the surplus in Bayesian and dominant strategy auctions, Econometrica 56, 1247–1257] and the Identifiability condition introduced in this paper. We also show that these conditions are necessary for ex-post efficiency to be attainable with budget balance and individual rationality, and that the expected social surplus in these mechanisms can be distributed in any desirable way. Lastly, we demonstrate that, like Crémer–McLean condition, the Identifiability condition is generic if there are at least three agents.     

Incomplete information and singleton cores in matching markets

We study ordinal Bayesian Nash equilibria of stable mechanisms in centralized matching markets under incomplete information. We show that truth-telling is an ordinal Bayesian Nash equilibrium of the revelation game induced by a common belief and a stable mechanism if and only if all the profiles in the support of the common belief have singleton cores. Our result matches the observations of Roth and Peranson [The redesign of the matching market for American physicians: some engineering aspects of economic design, Amer. Econ. Rev. 89 (1999) 748-780] in the National Resident Matching Program (NRMP) in the United States: (i) the cores of the profiles submitted to the clearinghouse are small and (ii) while truth-telling is not a dominant strategy most participants in the NRMP truthfully reveal their preferences.     

On the size and structure of group cooperation

When time preferences are heterogeneous and bounded away from one, how “much” cooperation can be achieved by an ongoing group? How does group cooperation vary with the group's size and structure? This paper examines characteristics of cooperative behavior in the class of symmetric, repeated games of collective action. These are games characterized by “free rider problems” in the level of cooperation achieved. The Repeated Prisoner's Dilemma games is a special case.We characterize the level of maximal average cooperation (MAC), the highest average level of cooperation, over all stationary subgame perfect equilibrium paths, that the group can achieve. The MAC is shown to be increasing in monotone shifts, and decreasing in mean preserving spreads of the distribution of discount factors. The latter suggests that more heterogeneous groups are less cooperative on average. Finally, in a class of Prisoner's Dilemma games, we show under weak conditions that the MAC exhibits increasing returns to scale in a range of heterogeneous discount factors. That is, larger groups are more cooperative, on average, than smaller ones. By contrast, when the group has a common discount factor, the MAC is invariant to group size.     

Payoff equivalence between Bayesian and ex post individually rational dominant strategy mechanisms

Summary. An efficient, interim individually rational, ex post budget balanced Bayesian mechanism is shown to be payoff equivalent to an ex post individually rational and ex ante budget balanced dominant strategy mechanism. This result simplifies the search for mechanisms that implement efficient allocation rules by pointing to a class of Groves mechanisms. It eliminates the strict requirement of common knowledge of priors and can be applied to many problems of incomplete information. Received: October 22 1996; revised version: November 25, 1997     

Quality cut-offs in procurement auctions

Recently the Government of India used procurement auction mechanisms with endogenously determined minimum quality. We find these auctions have no equilibrium in continuous symmetric monotonic pure strategies. This may substantiate the use of auction mechanisms with exogenously determined minimum quality.     

Dynamics of inductive inference in a unified framework

We present a model of inductive inference that includes, as special cases, Bayesian reasoning, case-based reasoning, and rule-based reasoning. This unified framework allows us to examine how the various modes of inductive inference can be combined and how their relative weights change endogenously. For example, we establish conditions under which an agent who does not know the structure of the data generating process will decrease, over the course of her reasoning, the weight of credence put on Bayesian vs. non-Bayesian reasoning. We illustrate circumstances under which probabilistic models are used until an unexpected outcome occurs, whereupon the agent resorts to more basic reasoning techniques, such as case-based and rule-based reasoning, until enough data are gathered to formulate a new probabilistic model.     

Bundling as an optimal selling mechanism for a multiple-good monopolist

Multiple objects may be sold by posting a schedule consisting of one price for each possible bundle and permitting the buyer to select the price-bundle pair of his choice. We identify conditions that must be satisfied by any price schedule that maximizes revenue within the class of all such schedules. We then provide conditions under which a price schedule maximizes expected revenue within the class of all incentive compatible and individually rational mechanisms in the n-object case. We use these results to characterize environments, mainly distributions of valuations, where bundling is the optimal mechanism in the two and three good cases.     

A two-person game of information transmission

We consider a statistical decision problem faced by a two player organization whose members may not agree on outcome evaluations and prior probabilities. One player is specialized in gathering information and transmitting it to the other, who takes the decision. This process is modeled as a game. Qualitative properties of the equilibria are analyzed. The impact of improving the quality of available information on the equilibrium welfares of the two individuals is studied. Better information generally may not improve welfare. We give conditions under which it will.     

Fair imposition

We introduce a new mechanism-design problem called fair imposition. In this setting a center wishes to fairly allocate tasks among a set of agents whose cost structures are known only to them, and thus will not reveal their true costs without appropriate incentives. The center, with the power to impose arbitrary tasks and payments on the agents, has the additional goal that his net payment to these agents is never positive (or, that it is tightly bounded if a loss is unavoidable). We consider two different notions of fairness that the center may wish to achieve. The central notion, which we call k-fairness, is in the spirit of max-min fairness. We present both positive results (in the form of concrete mechanisms) and negative results (in the form of impossibility theorems) concerning these criteria. We also briefly discuss an alternative, more traditional interpretation of our setting and results, in the context of auctions.     

Stochastic mechanisms in settings without monetary transfers: The regular case

We analyze relative performance of stochastic and deterministic mechanisms in an environment that has been extensively studied in the literature on communication (e.g., [Vincent P. Crawford, Joel Sobel, Strategic information transmission, Econometrica 50 (6) (1982) 1431-1451]) and optimal delegation (e.g., [Bengt Holmström, On the theory of delegation, in: M. Boyer, R.E. Kihlstrom (Eds.), Bayesian Models in Economic Theory, North-Holland, 1984, pp. 115-141]): a principal-agent model with hidden information, no monetary transfers, and single-peaked preferences. We demonstrate that under the common assumption of quadratic payoffs and a certain regularity condition on the distribution of private information and the agent's bias, the optimal mechanism is deterministic. We also provide an explicit characterization of this mechanism.     

A new necessary condition for implementation in iteratively undominated strategies

We uncover a new necessary condition for implementation in iteratively undominated strategies by mechanisms that satisfy the “best element property” where for each agent, there exists a strategy profile that gives him the highest payoff in the mechanism. This class includes finite and regular mechanisms. We conclude that either the quasilinearity-like assumptions of available sufficiency results cannot be completely dispensed with or some mechanisms that do not satisfy the best element property must be employed. We term the condition “restricted deception-proofness.” It requires that, in environments with identical preferences, the social choice function be immune to all deceptions, making it then stronger than incentive compatibility. In some environments the conditions for (exact or approximate) implementation by mechanisms satisfying the best element property are more restrictive than previously thought.     

Finite supermodular design with interdependent valuations

This paper studies supermodular mechanism design in environments with arbitrary (finite) type spaces and interdependent valuations. In these environments, the designer may have to use Bayesian equilibrium as a solution concept, because ex-post implementation may not be possible. We propose direct (Bayesian) mechanisms that are robust to certain forms of bounded rationality while controlling for equilibrium multiplicity. In quasi-linear environments with informational and allocative externalities, we show that any Bayesian mechanism that implements a social choice function can be converted into a supermodular mechanism that also implements the original decision rule. The proposed supermodular mechanism can be chosen in a way that minimizes the size of the equilibrium set, and we provide two sets of sufficient conditions to this effect. This is followed by conditions for supermodular implementation in unique equilibrium.     

STRATEGY‐PROOF ALLOCATION OF INDIVISIBLE GOODS AMONG COUPLES

We consider the problem of allocating indivisible goods among couples. Agents in a couple share the indivisible good assigned to them. The main result is that an allocation rule is strategy‐proof, neutral and non‐bossy if and only if it is serially dictatorial. An allocation rule is serially dictatorial if there is a priority order of couples and a function that identifies who chooses in each couple, such that for all preference profiles, a good assigned to couple i is the best element according to the preference of the identified agent in couple i among the remaining goods when the couples with higher priorities have made their choice.     

Anonymous monotonic social welfare functions

This paper presents two results about preference domain conditions that deepen our understanding of anonymous and monotonic Arrovian social welfare functions (ASWFs). We characterize the class of anonymous and monotonic ASWFs on domains without Condorcet triples. This extends and generalizes an earlier characterization (as Generalized Majority Rules) by Moulin (Axioms of Cooperative Decision Making, Cambridge University Press, New York, 1988) for single-peaked domains. We also describe a domain where anonymous and monotonic ASWFs exist only when there are an odd number of agents. This is a counter-example to a claim by Muller (Int. Econ. Rev. 23 (1982) 609), who asserted that the existence of 3-person anonymous and monotonic ASWFs guaranteed the existence of n-person anonymous and monotonic ASWFs for any n>3. Both results build upon the integer programming approach to the study of ASWFs introduced in Sethuraman et al. (Math. Oper. Res. 28 (2003) 309).     

Voting by committees under constraints

We consider social choice problems where a society must choose a subset from a set of objects. Specifically, we characterize the families of strategy-proof voting procedures when not all possible subsets of objects are feasible, and voters’ preferences are separable or additively representable.