Startseite A Comment on the Gibbard–Satterthwaite Theorem
Artikel
Lizenziert
Nicht lizenziert Erfordert eine Authentifizierung

A Comment on the Gibbard–Satterthwaite Theorem

  • Daniel Sandroni und Alvaro Sandroni EMAIL logo
Veröffentlicht/Copyright: 13. Juni 2025

Abstract

This paper re-examines the Gibbard–Satterthwaite theorem when some anomalous choice functions are allowed. We show non-dictatorial social aggregators that limit manipulation and non-dictatorial social aggregators do not permit manipulation.

JEL Classification: D71

Corresponding author: Alvaro Sandroni, Department of Managerial Economics and Decision Sciences, Kellogg School of Management, Northwestern University, Evanston, IL 60208, USA, E-mail: 

Appendix: Proofs

Proof of Theorem 1:

Let X be an axiom that allows for an upper weakening of WARP. Let C 1 = ρ(≻1) be the choice function produced by 1’s context-dependent ranking ≻1. Let B C 1 A be the set of all alternatives x ∈ B C such that if the choice function C′ is such that C′(B) = C(B), BA, and C′(A) = x, then C X . Let Φ : Ψ X n X be a social aggregator such that

C s ( B ) = C 1 ( B ) , B A , C s ( A ) A 2 y , for all  y B C 1 ,  y C 1 ( A ) , C s ( A ) = C 1 ( A ) when  B C 1 = { C 1 ( A ) }

Given that C 1 ( A ) B C 1 (because C 1 X ), it follows that Φ is well-defined. By construction, C s X . If C 1 satisfies WARP, then for any x ∈ A, if C′ is such that C′(B) = C(B), BA, and C′(A) = x, then C X u . Thus, if C 1 satisfies WARP, then B C 1 = A . It follows that if C 1 = ρ(≻1) satisfies WARP and C 2 = ρ(≻2) is such that C 2(A) ≠ C 1(A), then C s (A) = C 2(A) ≠ C 1(A). Hence, Φ is not dictatorial. The context-dependent ranking of individuals i ∈ {3, …, n} do not influence the social choice. Thus, i ∈ {3, …, n} cannot manipulate Φ. Individual 2 has no influence on any issue BA, and, hence, 2 cannot manipulate Φ in any issue BA. The social choice is the same as 1’s choice in any issue BA and, hence, 1 cannot manipulate Φ in any issue BA. Individual 2 cannot manipulate Φ in any issue BA because 2 has no influence in issue A. Individual 1 cannot manipulate Φ in issue A because, given the choices made in issues other than A, 1 has no influence in the social choice in A. Individual 2 cannot manipulate Φ in any issue A because the social choice is the one that maximizes A 2 among a set determined by choices made in other issues. Finally, Φ has at least three outcomes because any alternative in A can be the social choice in A.□

Proof of Theorem 2:

Let X be an axiom that allows for lower weakening of WARP. Let B B ̄ be the set of all binary issues (i.e., a binary issue B contains two options). If ( B ̄ ) > n , then let B ̂ B ̄ be an arbitrary set such that ( B ̂ ) = n . If ( B ̄ ) < n , then let B ̂ = B ̄ . Let ν : B ̂ { 1 , , n } be an arbitrary function such that for any pair of issues B B ̄ and B B ̄ , ν(B) ≠ ν(B′) when BB′. Let ≻ A be an arbitrary preference order on A (without indifference), which is not necessarily related to the any individual ranking on A. Let f : P X n X be a general social aggregator such that

C s ( B ) = C ν ( B ) ( B ) , for any  B B ̂  ; C s ( B ) A y , for all  y B ,  y C s ( B ) ,  B B ̂ .

Given that in all C s violations of WARP, the subissue is binary, it follows that C s X l . Thus, C s X . Hence, f is well-defined. Given that ν(B) ≠ ν(B′) for two different issues B and B′, it follows that as many individuals as the cardinality of B ̂ have influence on f and, hence, f is non-dictatorial.

To see that f is non-dictatorial, consider an arbitrary individual i, and some other individual ı ̃ i such that ν ( B ı ̃ ) = ı ̃ for some issue B ı ̃ . Consider a profile ( P 1 , , P n ) P X n of preferences such that B ı ̃ ı ̃ and B ı ̃ i have a different maximum in B ı ̃ . Let C s  = f(P 1, …, P n ) and let C′ be such that C′(B) = C s (B) when B B ı ̃ and C ( B ı ̃ ) B ı ̃ i y , for all y B ı ̃ , y C ( B ı ̃ ) . Given that B ı ̃ B ̄ and in all non-binary issues C′ is equal to C s , and in all non-binary issues C s optimizes a single order ≻ A , it follows that in any C′ violation of WARP, the subissue is binary. Hence, C X l . Thus, C X . By definition, C ( B ı ̃ ) B ı ̃ i C s ( B ı ̃ ) . Moreover, C′ and C s produce the same choice on any issue other than B ı ̃ . By (3.1), C′ P i C s . Thus, f is not dictatorial.

Consider a profile ( P 1 , , P n ) P X n and an alternative P i P X for individual i. Let

C s = f ( P 1 , , P n )  and  C s = f ( P 1 , , P i 1 , P i , P i + 1 , , P n ) .

Let B i be the only issue such that ν(B i ) = i. If B i does not exist, then, by construction, C s = C s and, therefore, it cannot be that C s P i C s because P i is asymmetric. By construction, C s ( B ) = C s ( B ) for any issue BB i . By (3.1), C s P i C s if and only if C s ( B i ) B i C s ( B i ) . By definition of f, C s (B i ) = C i (B i ) and C s ( B i ) = C i ( B i ) . By definition of B i either C i ( B i ) = C i ( B i ) or C i ( B i ) B i C i ( B i ) . In the former case, C s = C s and, hence, it cannot be that C s P i C s because P i is asymmetric. In the latter case, C s P i C s and, therefore, it cannot be that C s P i C s because P i is asymmetric. Thus, f is non-manipulable. Finally, f has more than 3 outcomes because any alternative in A may the aggregate choice in A.□

Proof of Theorem 3:

Let X = X l . Let B B ̄ be the set of all binary issues (i.e., a binary issue B contains two options). Let f : P X n X be a general social aggregator such that

C s ( B ) = C 1 ( B ) , if  B B ̄ ; C s ( B ) = C 2 ( B ) , if  B B ̄ .

Consider a profile ( P 1 , , P n ) P X n and an alternative P 1 P X for individual 1. Let

C s = f ( P 1 , , P n )  and  C s = f ( P 1 , , P i 1 , P i , P i + 1 , , P n ) .

Assume that C s C s . Let B ̂ B ̄ be the set of all binary issues B such that C s ( B ) C s ( B ) . Note that if B is a non-binary issues, then C s ( B ) = C s ( B ) because aggregate choices do not depend upon individual 1’s preferences. Let’s say that B ̂ has k elements and let’s order these issues arbitrarily so that B ̂ = { B 1 , . , B k } . Let C 1 be the choice function such that C 1 ( B 1 ) = C s ( B 1 ) C s ( B 1 ) = C 1 ( B 1 ) and C 1(B) = C s (B) for every issue BB 1. Let ≻1 be the context-dependent ranking determined by P 1. By definition, C s ( B 1 ) = C 1 ( B 1 ) B 1 1 C 1 ( B 1 ) = C s ( B 1 ) . Given that C 1 and C s are equivalent outside B 1, it follows that C s P 1 C 1. Now let C 2 be the choice function such that C 2 ( B 2 ) = C s ( B 2 ) C 1 ( B 2 ) = C 1 ( B 2 ) and C 2(B) = C 1(B) for every issue BB 2. By definition, C 1 ( B 2 ) = C 1 ( B 2 ) B 2 1 C s ( B 2 ) = C 2 ( B 2 ) . Given that C 1 and C 2 are equivalent outside B 2, it follows that C 1 P 1 C 2. Given C e , e = 1, …, k − 1, let C e+1 be the choice function such that C e + 1 ( B e ) = C s ( B e ) C s ( B e ) = C 1 ( B e ) and C e+1(B) = C e (B) for every issue BB e . By the same argument as above, C e P 1 C e+1, e = 1, …, k − 1. By transitivity, C 1 P 1 C k . Thus, by transitivity, C s P 1 C k . By construction, C k = C s . Thus, C s P 1 C s . It follows that individual 1 cannot manipulate this general social aggregator. The proof that individual 2 cannot manipulate this general social aggregator is entirely analogous. The preferences of all other individuals (i.e. not 1 or 2) do not affect the aggregate choices and, hence they cannot manipulate this general social aggregator as well.

It is immediate that this general social aggregator is non-dictatorial, has more than 3 possible outcomes, two or more individuals have influence on f, and it satisfies unanimity. As mentioned, as long as C 2 X l , it follows that for any profile ( P 1 , , P n ) P X n , C s  = f(P 1, …, P n ) ∈  X l because C s (B) = C 2(B) for any non-binary issues. If C 2 X l , then there are no WARP violation in 2’s choices on two non-binary issues. Therefore, there are no WARP violation in the aggregate choices on two non-binary issues as well.□

References

Aizerman, M., and F. Aleskerov. 1986. “Voting Operators in the Space of Choice Functions.” Mathematical Social Sciences 11: 201–42. https://doi.org/10.1016/0165-4896(86)90027-2.Suche in Google Scholar

Aizerman, M., and F. Aleskerov. 1995. “Theory of Choice.” In Studies in Mathematical and Managerial Economics, Vol. 38, edited by H. Gleiser, and S. Martin. North-Holland Elsevier Science B.V.Suche in Google Scholar

Aleskerov, F. 2002. “Categories of Arrovian Voting Schemes.” In Chapter 2 in the Handbook of Social Choice and Welfare, Vol. 1, edited by K. Arrow, A. Sen, and K. Suzumura, 95–129. North-Holland Elsevier Science B.V.10.1016/S1574-0110(02)80006-6Suche in Google Scholar

Aleskerov, F. 2010. “Arrovian Aggregation Models.” In Studies B: Mathematical and Statistical Methods, Vol. 39, edited by W. Leinfellner, and G. Eberlein. Kluwer Academic Publishers.Suche in Google Scholar

Arrow, K. 1950. “A Difficulty in the Concept of Social Welfare.” Journal of Political Economy 58 (4): 328–46. https://doi.org/10.1086/256963.Suche in Google Scholar

Borah, A., R. Garg, and N. Singh. 2023. “Arrowian Social Equilibrium: Indecisiveness, Influence and Rational Social Choices under Majority Rule.” The B.E. Journal of Theoretical Economics 23 (1): 181–94. https://doi.org/10.1515/bejte-2021-0149.Suche in Google Scholar

Cherepanov, V., T. Feddersen, and A. Sandroni. 2013. “Rationalization.” Theoretical Economics 8 (3): 775–800.10.3982/TE970Suche in Google Scholar

Duggan, J., and T. Schwartz. 2000. “Strategic Manipulability without Resoluteness or Shared Beliefs: Gibbard–Satterthwaite Generalized.” Social Choice and Welfare 17: 85–93. https://doi.org/10.1007/pl00007177.Suche in Google Scholar

Eliaz, K., and R. Spiegler. 2011. “Consideration Sets and Competitive Marketing.” The Review of Economic Studies 78 (1): 235–62. https://doi.org/10.1093/restud/rdq016.Suche in Google Scholar

Eliaz, K., D. Ray, and R. Razin. 2006. “Choice Shifts in Groups: A Decision-Theoretic Basis.” The American Economic Review 96: 1321–32. https://doi.org/10.1257/000282806779468580.Suche in Google Scholar

Fudenberg, D., and D. Levine. 2006. “A Dual-Self Model of Impulse Control.” The American Economic Review 96: 1449–76. https://doi.org/10.1257/aer.96.5.1449.Suche in Google Scholar

Gibbard, A. 1973. “Manipulation of Voting Schemes: A General Result.” Econometrica 41 (4): 587–601. https://doi.org/10.2307/1914083.Suche in Google Scholar

Gibbard, A. 1978. “Straightforwardness of Game Forms with Lotteries as Outcomes.” Econometrica 46 (3): 595–614. https://doi.org/10.2307/1914235.Suche in Google Scholar

Heller, Y. 2012. “Justifiable Choice.” Games and Economic Behavior 76 (2): 375–90. https://doi.org/10.1016/j.geb.2012.07.001.Suche in Google Scholar

Houy, N. 2007. “Rationality and Order-Dependent Sequential Rationality.” Theory and Decision 62: 119–34. https://doi.org/10.1007/s11238-006-9016-0.Suche in Google Scholar

Houy, N., and K. Tadenuma. 2009. “Lexicographic Compositions of Multiple Criteria for Decision Making.” Journal of Economic Theory 144 (4): 1770–82. https://doi.org/10.1016/j.jet.2009.01.009.Suche in Google Scholar

Lehrer, E., and R. Teper. 2011. “Justifiable Preferences.” Journal of Economic Theory 146: 762–74. https://doi.org/10.1016/j.jet.2010.12.007.Suche in Google Scholar

Lleras, J., Y. Masatlioglu, D. Nakajima, and E. Ozbay. 2017. “When More Is Less: Limited Consideration.” Journal of Economic Theory 170: 70–85. https://doi.org/10.1016/j.jet.2017.04.004.Suche in Google Scholar

Manzini, P., and M. Mariotti. 2007. “Sequentially Rationalizable Choice.” The American Economic Review 97 (5): 1824–39. https://doi.org/10.1257/aer.97.5.1824.Suche in Google Scholar

Masatlioglu, Y., and E. Ok. 2005. “Rational Choice with Status Quo Bias.” Journal of Economic Theory 121: 1–29. https://doi.org/10.1016/j.jet.2004.03.007.Suche in Google Scholar

Masatlioglu, Y., D. Nakajima, and E. Ozbay. 2012. “Revealed Attention.” The American Economic Review 102 (5): 2138–205. https://doi.org/10.1257/aer.102.5.2183.Suche in Google Scholar

Sandroni, A., and A. Sandroni. 2021. “A Comment on Arrow’s Impossibility Theorem.” The B.E. Journal of Theoretical Economics 21 (1): 347–54. https://doi.org/10.1515/bejte-2019-0175.Suche in Google Scholar

Satterthwaite, M. 1975. “Strategy-Proofness and Arrow’s Conditions: Existence and Correspondence Theorems for Voting Procedures and Social Welfare Functions.” Journal of Economic Theory 10 (2): 187–217. https://doi.org/10.1016/0022-0531(75)90050-2.Suche in Google Scholar

Taylor, A. 2002. “The Manipulability of Voting Systems.” The American Mathematical Monthly 109 (4): 324. https://doi.org/10.1080/00029890.2002.11920895.Suche in Google Scholar

Tyson, C. 2013. “Behavioral Implications of Shortlisting Procedures.” Social Choice and Welfare 41 (4): 941–63. https://doi.org/10.1007/s00355-012-0704-0.Suche in Google Scholar

Received: 2024-08-30
Accepted: 2025-05-20
Published Online: 2025-06-13

© 2025 Walter de Gruyter GmbH, Berlin/Boston

Heruntergeladen am 23.11.2025 von https://www.degruyterbrill.com/document/doi/10.1515/bejte-2024-0095/html
Button zum nach oben scrollen