A fundamental structure of strategy-proof social choice correspondences with restricted preferences over alternatives

I prove that under each strategy-proof and unanimous social choice correspondence, there is at least one agent who is decisive. Because the result is established on a weak requirement on preferences over sets, the existence of a decisive agent is an underlying feature of most strategy-proof and unanimous social choice correspondences. Moreover, I consider a restriction on the space of preferences over alternatives. I prove that circular sets of preferences over alternatives are sufficient for the existence of a decisive agent.     

Coalitionally strategy-proof social choice correspondences and the Pareto rule

This paper studies coalitional strategy-proofness of social choice correspondences that map preference profiles into sets of alternatives. In particular, we focus on the Pareto rule, which associates the set of Pareto optimal alternatives with each preference profile, and examine whether or not there is a necessary connection between coalitional strategy-proofness and Pareto optimality. The definition of coalitional strategy-proofness is given on the basis of a max–min criterion. We show that the Pareto rule is coalitionally strategy-proof in this sense. Moreover, we prove that given an arbitrary social choice correspondence satisfying the coalitional strategy-proofness and nonimposition, all alternatives selected by the correspondence are Pareto optimal. These two results imply that the Pareto rule is the maximal correspondence in the class of coalitionally strategy-proof and nonimposed social choice correspondences.     

Choosing subsets: a size-independent probabilistic model and the quest for a social welfare ordering

“Subset voting” denotes a choice situation where one fixed set of choice alternatives (candidates, products) is offered to a group of decision makers, each of whom is requested to pick a subset containing any number of alternatives. In the context of subset voting we merge three choice paradigms, “approval voting“ from political science, the “weak utility model” from mathematical psychology, and “social welfare orderings” from social choice theory. We use a probabilistic choice model proposed by Falmagne and Regenwetter (1996) built upon the notion that each voter has a personal ranking of the alternatives and chooses a subset at the top of the ranking. Using an extension of Sen's (1966) theorem about value restriction, we provide necessary and sufficient conditions for this empirically testable choice model to yield a social welfare ordering. Furthermore, we develop a method to compute Borda scores and Condorcet winners from subset choice probabilities. The technique is illustrated on an election of the Mathematical Association of America (Brams, 1988). Received: 18 August 1995 / Accepted: 13 February 1997     

Generalized median social welfare functions

The article considers the construction of social welfare functions when the set of alternatives is the two-dimensional nonnegative orthant, as would be the case if there are two divisible public goods which can be consumed in any nonnegative quantities. With individual and social preferences required to be linear and strictly monotonic, but otherwise unrestricted, we characterize all of the social welfare functions which satisfy binary independence of irrelevant alternatives and anonymity and which satisfy binary independence of irrelevant alternatives, anonymity, and weak Pareto. These classes of social welfare functions are shown to be formally equivalent to the classes of social choice functions characterized by Moulin in his study of strategy-proof social choice with single-peaked preferences.We are grateful to Charles Blackorby, Hervé Moulin, and an anonymous referee for their comments.     

Single-Crossing, Strategic Voting and the Median Choice Rule

This paper studies the strategic foundation of the Representative Voter Theorem (Rothstein in: Pub Choice 72:193–212, 1991), also called the “second version” of the Median Voter Theorem. As a by-product, it also considers the existence of strategy-proof social choice functions over the domain of single-crossing preferences. The main result shows that single-crossing constitutes a domain restriction over the real line that allows not only majority voting equilibria, but also non-manipulable choice rules. In particular, this is true for the median rule, which is found to be group strategic-proof over the full set of alternatives and over every nonempty subset. In addition, the paper also examines the relation between single-crossing and order-restriction. And it uses this relation together with the strategy-proofness of the median rule to prove that the outcome predicted by the Representative Voter Theorem can be implemented in dominant strategies through a simple mechanism. This mechanism is a two-stage voting procedure in which, first, individuals select a representative among themselves, and then the winner chooses a policy to be implemented by the planner.     

Incremental cost sharing: Characterization by coalition strategy-proofness

Each one of n users consumes an idiosyncratic commodity produced in indivisible units. The n commodities are jointly produced by a central facility and total cost must be shared by the users.  A “sequential stand alone mechanism” shares costs incrementally according to a fixed ordering of the users: the first user always pays stand alone cost, the second pays the stand alone cost of the first two users minus that of the first and so on. If the second derivatives of costs are of a constant sign, such a method yields a unique strong equilibrium at every profile of convex preferences in the game where each user chooses his own demand. This equilibrium, in turn, defines a coalition strategy-proof social choice function.  Under decreasing marginal costs and submodular costs, the sequential stand alone mechanisms are almost characterized by these properties; the only exception is the binary demand case (each agent consumes zero or one unit) where a rich family of cost sharing methods (the Shapley value among them) yields a coalition strategy-proof equilibrium selection. Under increasing marginal costs and supermodular costs, coalition strategy-proofness characterizes a richer family of cost sharing methods: they give out one unit at a time while charging marginal costs, with the users taking turns according to a sequence fixed in advance. These methods contain serial cost sharing as a limit case. Received: 8 July 1997/Accepted: 22 January 1998     

Strategy-proof allocation of multiple public goods

This paper characterizes strategy-proof social choice functions (SCFs), the outcome of which are multiple public goods. Feasible alternatives belong to subsets of a product set . The SCFs are not necessarily “onto”, but the weaker requirement, that every element in each category of public goods A _k is attained at some preference profile, is imposed instead. Admissible preferences are arbitrary rankings of the goods in the various categories, while a separability restriction concerning preferences among the various categories is assumed. It is found that the range of the SCF is uniquely decomposed into a product set in general coarser than the original product set, and that the SCF must be dictatorial on each component B _l . If the range cannot be decomposed, a form of the Gibbard–Satterthwaite theorem with a restricted preference domain is obtained.     

Does majoritarian approval matter in selecting a social choice rule? An exploratory panel study

This study is an attempt to empirically detect the public opinion concerning majoritarian approval axiom. A social choice rule respects majoritarian approval iff it chooses only those alternatives which are regarded by a majority of “voters” to be among the “better half” of the candidates available. We focus on three social choice rules, the Majoritarian Compromise, Borda’s Rule and Condorcet’s Method, among which the Majoritarian Compromise is the only social choice rule always respecting majoritarian approval. We confronted each of our 288 subjects with four hypothetical preference profiles of a hypothetical electorate over some abstract set of four alternatives. At each hypothetical preference profile, two representing the preferences of five and two other of seven voters, the subject was asked to indicate, from an impartial viewpoint, which of the four alternatives should be chosen whose preference profile was presented, which if that is unavailable, then which if both of the above are unavailable, and finally which alternative should be avoided especially. In each of these profiles there is a Majoritarian Compromise-winner, a Borda-winner and a Condorcet-winner, and the Majoritarian Compromise-winner is always distinct from both the Borda-winner and the Condorcet-winner, while the Borda- and Condorcet-winners sometimes coincide. If the Borda- and Condorcet-winners coincide then there are two dummy candidates, otherwise only one, and dummies coincide with neither of the Majoritarian Compromise-, Borda- or Condorcet-winner. We presented our subjects with various types of hypothetical preference profiles, some where Borda respecting majoritarian approval, some where it failed to do so, then again for Condorcet, some profiles it respected majoritarian approval and some where it did not. The main thing we wanted to see was whether subjects’ support for Borda and Condorcet was higher when this social choice rule respected majoritarian approval than it did not. Our unambiguous overall empirical finding is that our subjects’ support for Borda and Condorcet was significantly stronger as they respect majoritarian approval.     

Generalized condorcet-winners for single peaked and single-plateau preferences

When preferences are single peaked the choice functions that are independent of irrelevant alternatives both in Nash's and in Arrow's sense are characterized. They take the Condorcet winner of the n individual peaks plus at most n-1 fixed ballots (phantom voters). These choice functions are also coalitionally strategy-proof.Next the domain of individual preferences is enlarged to allow for singleplateau preferences: again, Nash's IIA and Arrow's IIA uniquely characterize a class of generalized Condorcet winners choice functions. These are, again, coalitionally strategy-proof.     

Approval Voting on Dichotomous Preferences

The aim of this paper is to find normative foundations of Approval Voting when individuals have dichotomous preferences. We show that a social choice function is anonymous, neutral, strategy-proof and strictly monotone if and only if it is Approval Voting and interpret this result as an extension of May’s theorem (Econometrica 20:680–684, 1952). Then, we show that Approval Voting is the only strictly symmetric, neutral and efficient social choice function. This result is related to a characterization of Baigent and Xu (Math Soc Sci 21:21–29, 1991).     

Strategy-proofness of continuous aggregation maps

 We consider continuous aggregation maps (e.g., social welfare functions). By assuming that the voters have preferences over social outcomes, we regard the social decision procedure as a noncooperative game, with the aggregation map as a game form. The map is called strategy-proof if it is a Nash equilibrium that the voters report their most preferred outcomes. We give sufficient topological conditions on the space of outcomes so that only dictatorial maps are strategy-proof. Received: 28 February 1994/Accepted: April 22, 1996     

Choosing the level of a public good when agents have an outside option

The work examines strategy-proof social choice functions which select a level or the location of a public good when its consumption is not compulsory. We assume that agents have single-peaked preferences on the consumption of the public good as well as a reservation utility. Agents opt out of the good whenever they get a utility smaller than their reservation utility. We characterize strategy-proof and efficient social choice functions as well as the ones which are anonymous and group strategy-proof. Finally, we observe that for given preferences a Condorcet winner may not exist.This article constitutes the first chapter of my Ph.D. dissertation at Universitat Autònoma de Barcelona. I wish to thank especially Jordi Massó for supervising my research, Dolors Berga and Matthew Jackson. I also thank José Alcalde, Salvador Barberà, Carmen Beviá, Jeremy Heald, Hervé Moulin, Antonio Romero-Medina and an anonymous referee for their comments and suggestions. The European Commission is gratefully acknowledged for its financial support through the T.M.R. fellowship ERBFMICT960619.     

Strategy-proofness and markets

If a market is considered to be a social choice function, then the domain of admissible preferences is restricted and standard social choice theorems do not apply. A substantial body of analysis, however, strongly supports the notion that attractive strategy-proof social choice functions do not exist in market settings. Yet price theory, which implicitly assumes the strategy-proofness of markets, performs quite well in describing many real markets. This paper resolves this paradox in two steps. First, given that a market is not strategy-proof, it should be modeled as a Bayesian game of incomplete information. Second, a double auction market, which is perhaps the simplest operationalization of supply and demand as a Bayesian game, is approximately strategy-proof even when the number of traders on each side of the market is quite moderate. Received: 24 March 1999/Accepted: 25 October 1999     

A characterization of strategy-proof social choice functions for economies with pure public goods

We characterize strategy-proof social choice functions when individuals have strictly quasi-concave, continuous and satiated utility functions on convex subsets of IR ^m , representing preferences for the provision of m pure public goods. When specialized to the case m=1, these assumptions amount to requiring that preferences are single peaked, and for such a domain there exists a wide class of strategy-proof social choice functions. These were studied by Moulin (1980) under strong additional assumptions. Our first results characterize the complete class, after an appropriate extension of the single-peakedness condition. The new characterization retains the flavour of Moulin's elegant representation theorem. For the general m-dimensional case, previous results have shown that there is no efficient, strategy-proof, nondictatorial social choice function, even within the domain restrictions under consideration (Border and Jordan 1983; Zhou 1991). In fact, Zhou's powerful result indicates that nondictatorial strategy-proof s.c.f.'s will have a range of dimension one. This allows us to conclude with a complete characterization of all strategy-proof s.c.f.'s on IR ^m , because restrictions of preferences from our admissible class to one dimensional subsets satisfy the slightly generalized notion of single-peakedness that is used in our characterization for the case m=1. We feel that a complete knowledge of the class of strategy-proof mechanisms, in this as well as in other contexts, is an important step in the analysis of the trade-offs between strategy-proofness and other performance criteria, like efficiency.This paper was written while both authors were visiting GREMAQ, Université des Sciences Sociales de Toulouse. We are thankful for its hospitality and good research atmosphere. Barberà's work is supported by the Instituto de Estudios Fiscales and by research grant PB89-0294 from the Secretaría de Estado de Universidades e Investigación, Spain. Jackson acknowledges the support of NSF grant SES8921409. We thank Jacques Crémer, Beth Allen, John Weymark and two anonymous referees for helpful comments on earlier drafts.     

Manipulability measures of common social choice functions

All social choice functions are manipulable when more than two alternatives are available. I evaluate the manipulability of the Borda count, plurality rule, minimax set, and uncovered set. Four measures of manipulability are defined and computed stochastically for small numbers of agents and alternatives.  Social choice rules derived from the minimax and uncovered sets are found to be relatively immune to manipulation whether a sole manipulating agent has complete knowledge or absolutely no knowledge of the preferences of the others. The Borda rule is especially manipulable if the manipulating agent has complete knowledge of the others. Received: 5 January 1996/Accepted: 31 July 1998     

Nash Implementation with Lottery Mechanisms

Consider the problem of exact Nash Implementation of social choice correspondences. Define a lottery mechanism as a mechanism in which the planner can randomize on alternatives out of equilibrium while pure alternatives are always chosen in equilibrium. When preferences over alternatives are strict, we show that Maskin monotonicity (Maskin in Rev Econ stud 66: 23–38, 1999) is both necessary and sufficient for a social choice correspondence to be Nash implementable. We discuss how to relax the assumption of strict preferences. Next, we examine social choice correspondences with private components. Finally, we apply our method to the issue of voluntary implementation (Jackon and Palfrey in J Econ Theory 98: 1–25, 2001).I thank Toyo Sakai for his comments on a previous draft. I also thank two anonymous referees and an editor of this journal for helpful comments that improved this paper. A previous version circulated as “A note on Maskin monotonicity”. After the results presented here were obtained, I became aware of a new unpublished paper by Benoit and Ok (2004). The result of Theorem 2 and the discussion that follows is partially similar to their Theorem 1.     

On strategy-proofness and essentially single-valued cores: A converse result

In a general model of indivisible good allocation, S?nmez (1999) established that, whenever the core is nonempty for each preference profile, if an allocation rule is strategy-proof, individually rational and Pareto optimal, then the rule is a selection from the core correspondence, and the core correspondence must be essentially single-valued. This paper studies the converse claim of this result. I demonstrate that whenever the preference domain satisfies a certain condition of `richness', if the core correspondence is essentially single-valued, then any selection from the core correspondence is strategy-proof (even weakly coalition strategy-proof, in fact). In particular, on the domain of preferences in which each individual has strict preferences over his own assignments and there is no consumption externality, such an allocation rule is coalition strategy-proof. And on this domain, coalition strategy-proofness is equivalent to Maskin monotonicity, an important property in implementation theory. Received: 22 February 2000/Accepted: 22 January 2002 I am grateful to Ryo-ichi Nagahisa, Shinji Ohseto, Hiroshi Ono, Tomoichi Shinotsuka and Shigehiro Serizawa for valuable comments. And I am indebted to two anonymous referees for useful suggestions. Especially, I owe the present proof of Lemma 2 to one referee. I also thank Yukihiko Funaki, Atsushi Kajii, Mamoru Kaneko, Eiichi Miyagawa, Tatsuyoshi Saijo, Manimay Sengupta, Yves Sprumont, Yoshikatsu Tatamitani, Manabu Toda, Takashi Ui, Takehiko Yamato, Naoki Yoshihara and the participants of the seminars in Hokkaido University, Kansai University, ISER (Osaka University), Otaru University of Commerce and Tsukuba University. All errors are my own responsiblity.     

Strategy-proofness on Euclidean spaces

 In this paper we characterize strategy-proof voting schemes on Euclidean spaces. A voting scheme is strategy-proof whenever it is optimal for every agent to report his best alternative. Here the individual preferences underlying these best choices are separable and quadratic. It turns out that a voting scheme is strategy-proof if and only if (α) its range is a closed Cartesian subset of Euclidean space, (β) the outcomes are at a minimal distance to the outcome under a specific coordinatewise veto voting scheme, and (γ) it satisfies some monotonicity properties. Neither continuity nor decomposability is implied by strategy-proofness, but these are satisfied if we additionally impose Pareto-optimality or unanimity. Received: 18 October 1993/Accepted: 2 February 1996     

Domains, ranges and strategy-proofness: the case of single-dipped preferences

We characterize the set of all individual and group strategy-proof rules on the domain of all single-dipped preferences on a line. For rules defined on this domain, and on several of its subdomains, we explore the implications of these strategy-proofness requirements on the maximum size of the rules’ range. We show that when all single-dipped preferences are admissible, the range must contain two alternatives at most. But this bound changes as we consider different subclasses of single-dipped preferences: we provide examples of subdomains admitting strategy-proof rules with larger ranges. We establish exact bounds on the maximal size of strategy-proof functions on each of these domains, and prove that the relationship between the sizes of the subdomains and those of the ranges of strategy-proof functions on them need not be monotonic. Our results exhibit a sharp contrast between the structure of strategy-proof rules defined on subdomains of single-dipped preferences and those defined on subsets of single-peaked ones.     

Implementation in perfect equilibria

This paper considers implementation when the feasible outcomes are lotteries over a finite set of alternatives. The following weak condition is sufficient for implementation in trembling hand perfect equilibria (with three or more players): if all but one player agree on which alternative is the best, this alternative is (among those that are) chosen by the social choice rule, and if all but one player agree on which alternative is the worst, this alternative is not chosen. Many interesting social choice rules that are not Nash implementable satisfy this condition. On the other hand, there are social choice rules that are implementable in Nash equilibria but not in perfect equilibria.I am grateful to Luis Corchón, Eric Maskin, William Thomson, seminar participants at Harvard and Rochester, and two anonymous referees for helpful comments and suggestions. This research was supported by grants from the Royal Swedish Academy of Sciences and the Bank of Sweden Tercentenary Foundation.