Social choice with transfers
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.
Ruth: We're looking at the paper: "Social choice with transfers".
Warren: Here's the gist of the paper. Social choice with transfers considers the problem of social choice when agents have quasi-linear utility in a num´eraire good and transfers between agents are possible.
Ruth: The focus is on identifying a fair allocation assuming the preferences of the agents are known and not on implementation.
Warren: Problem Formulation An n-player social-choice problem, or bargaining set, is defined as a set B ⊂ R n that is nonempty, closed, convex, comprehensive, and bounded above.
Ruth: The feasible set of allocations is then defined as the set of x ∈ R n such that xN ≤ maxy∈B yN.
Title and authors: Ruth: So, we're looking at this paper called "Social choice with transfers." It tackles those tricky situations where agents can actually trade things with each other, you know?
Warren: Yeah. The authors are Gregorio Curello and Sam Jindani. What that title really points to is the idea of social choice when you have these transfers between agents possible, which includes things like public good provision or settling debts.
Iris: It sets up the core problem: given a set of alternatives, how should we choose one and what transfers should happen? The authors are focusing on finding a fair allocation assuming everyone's preferences are known.
Ruth: Exactly. They're not worried about how to actually implement it in the real world, just figuring out what the fair outcome is based on the preferences. This paper is setting up a framework for this kind of problem.
Warren: It suggests that there’s a specific way to solve this, and it claims that this solution has some very desirable properties that make it particularly useful.
The paper's summary: Ruth: Let's look at what they actually propose. Basically, the paper argues that the Shapley value of the stand-alone game is the only solution to satisfy a set of axioms. This includes things like efficiency, anonymity, and the dummy property.
Warren: So they are saying that if you want a solution that meets those specific criteria—efficiency, anonymity, and so on—the Shapley value is it. It's not just *a* solution; it’s the *only* one under those rules.
Iris: That sounds pretty strong for a result. The authors are proving that this specific value satisfies efficiency, anonymity, the dummy property, continuity, additivity, and monotonicity all at once.
Ruth: That's a lot of desirable properties bundled up in one solution. It means we don't have to worry about just getting one nice property; we get a whole set of good behaviors with this one formula.
Warren: The example they use is building a power plant for two cities, where city one derives eight in monetary units from location A and two from location B, and city two gets two from A and four from B.
Iris: In that example, the stand-alone claim for city one is eight, but the grand coalition's claim is ten because it's eight plus two. The Shapley value calculation results in a transfer of one monetary unit from city one to city two, leading to an outcome where both cities get seven and three utility respectively.
The paper's improvements: Ruth: Beyond just stating that the Shapley value works, the authors extend this model. They cover applications involving surpluses or deficits by adding an extra entry for those things at the alternative level.
Warren: So they're not just looking at pure gains or losses; they're looking at scenarios where there might be something left over or a deficit that needs to be compensated for.
Iris: The solution on this expanded set is given by a formula that’s analogous to what we see in Theorem two of the original paper, which allows for principled compensation for forgone benefits.
Ruth: It lets us think about things like building shared infrastructure or settling debts where you need to account for what each party actually attaches to owning the property outright.
Warren: That framing is important because it moves beyond just simple valuation and looks at the actual attachment of each player to the outcome.
Conclusion: Ruth: So, to wrap up, the main finding of this paper is that a solution on a bargaining set satisfies efficiency, anonymity, continuity, additivity, and monotonicity if and only if it's the Shapley value of the stand-alone game.
Warren: That really establishes the Shapley value as the unique fair allocation for these types of social choice problems when you have those desirable properties in place. It’s a very specific tool for this class of problem.
Iris: And they also show that it's uniquely determined on B zero, which is the set where no transfers happen, even if there are multiple players involved. That's a concrete result for the two-player case.
Ruth: It’s interesting how this solution emerges even though coalitions aren't explicitly part of the model or the axioms themselves. The paper also shows that L dagger is contained within S sub s for every set S in N, which is a technical detail they found important.
Warren: It’s a solid piece of work because it gives us a definitive way to find fairness in these transfer problems, even when things get complicated with surpluses and deficits.
Iris: So, the paper "Social choice with transfers" gives us a unique fair allocation based on the Shapley value under very specific conditions. We'll be looking at how this applies to real-world resource management next.
University of Mannheim · National University of Singapore
econ.TH
Submitted: 2025-02-01
Updated: 2026-11-11
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 98/100
The gist: Social choice with transfers considers the problem of social choice when agents have quasi-linear utility in a num´eraire good and transfers between agents are possible.
Key concepts
- Bargaining Set (B)
- This is a set of possible outcomes for an n-player social choice problem. It must be non-empty, closed, convex, comprehensive, and bounded above. The feasible allocations are those that respect the constraints defined by this set.
- Shapley Value
- This value represents the fair allocation derived from a stand-alone game. It is shown to be the unique solution that satisfies key desirable properties like efficiency and anonymity in social choice problems with transfers.
- Axioms (Efficiency, Anonymity, Dummy Property)
- These are desirable properties that a fair allocation should possess. The paper proves that the Shapley value of the stand-alone game is the only solution to satisfy these specific axioms simultaneously.
- Surplus or Deficit Extension
- The model can be extended to handle situations where there are surpluses or deficits in alternatives. This allows for principled compensation for forgone benefits, such as settling debts or compensating parties for not choosing a certain option.
Terminology
Summary
Social choice with transfers considers the problem of social choice when agents have quasi-linear utility in a num´eraire good and transfers between agents are possible. p. 0 The focus is on identifying a fair allocation assuming the preferences of the agents are known and not on implementation. p. 1 p. 1
Problem Formulation
An n-player social-choice problem, or bargaining set, is defined as a set B ⊂ R n that is nonempty, closed, convex, comprehensive, and bounded above. The feasible set of allocations is then defined as the set of x ∈ R n such that xN ≤ maxy∈B yN. p. 3 A solution to this problem is a map ψ: B → R n such that ψN (B) ≤ maxx∈B xN for every B ∈ B. p. 3 The Shapley value of the stand-alone game is given by the formula in equation (1). p. 0
Axiomatic Foundation
The paper establishes that the Shapley value of the stand-alone game is the only solution to a set of desirable properties, including efficiency, anonymity, and monotonicity. These axioms include Axiom 1 (Efficiency), Axiom 2 (Anonymity), and Axiom 3 (Dummy property). p. 8 The authors prove that the Shapley value satisfies these axioms in section 3. p. 81.
Uniqueness Proof
The uniqueness of the solution is demonstrated by showing that any solution satisfying axioms 1 to 6 must be the Shapley value of the stand-alone game. This involves proving that solutions satisfying axioms 1 to 5 can be represented by a specific form involving a finite Borel measure µ on S+ and a Borel-measurable function h: S+ → S0. p. 9 The subsequent steps utilize the Lebesgue differentiation theorem to show that h(v) must be zero for every v ∈ L†. p. 9 This leads to the conclusion that there exist unique weights (¯wk)n−1k=1 such that ψ is uniquely determined on B0. p. 2
Extensions and Applications
The model is extended to cover applications involving surpluses or deficits, where the bargaining set C includes an extra entry for the surplus or deficit at the alternative. A solution on C is given by a formula analogous to Theorem 2 in section 3. p. 2 This framework allows for principled compensation for forgone benefits, as illustrated in examples such as building shared infrastructure and settling debts. p. 4
Welfarist Alternatives
The literature also presents two welfarist solutions that do not satisfy the dummy property, namely 'equal allocation of non-separable costs' (EANS) and 'equal sharing from the ideal level' (ESIL). These alternative solutions are discussed by Moulin in section 4. p. 9 The Shapley value solution is superior because it satisfies the dummy property, which is not satisfied by EANS or ESIL. p. 8
Conclusion
The main result is that a solution on B satisfies efficiency, anonymity, the dummy property, continuity, additivity, and monotonicity if and only if it is the Shapley value of the stand-alone game. This establishes the Shapley value as the unique fair allocation for social choice problems with transfers under these desirable properties. p. 0 p. 0 The model also covers surplus and deficit applications, providing a general approach to compensation. p. 2 p. 2 The paper concludes by showing that the Shapley value is uniquely determined on B0 for n=2. p. 9 p. 9 It is especially interesting that this solution emerges despite the fact that coalitions play no explicit role either in the model or in any of the axioms. p. 8 p. 8 The paper also shows that L† ⊆ Ss for every S ∈ N. p. 9 p. 9 The paper is 2502. p. 0 p. 000308. This paper is 2502. p. 0
--- Page 1 ---
Social choice with transfers considers the problem of social choice when agents have quasi-linear utility in a num´eraire good and transfers between agents are possible. The focus is on identifying a fair allocation assuming the preferences of the agents are known and not on implementation.
The main result is that a solution on B satisfies efficiency, anonymity, the dummy property, continuity, additivity, and monotonicity if and only if it is the Shapley value of the stand-alone game. This establishes the Shapley value as the unique fair allocation for social choice problems with transfers under these desirable properties. The model also covers surplus and deficit applications, providing a general approach to compensation. The paper concludes by showing that the Shapley value is uniquely determined on B0 for n=2. It is especially interesting that this solution emerges despite the fact that coalitions play no explicit role either in the model or in any of the axioms. The paper also shows that L† ⊆ Ss for every S ∈ N. The paper is 2502.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
1
No-transfer social choice models can be solved uniquely by calculating a specific value, stating that the Shapley value of the stand-alone game is the only solution to satisfy certain desirable properties.
This allows AI systems modeling resource allocation problems without transfers to find a single, fair outcome based on marginal claims.
2
AI systems can determine fair allocations for complex resource management problems by using the Shapley value of a stand-alone game defined by coalition claims in a bargaining set. This applies to public-good provision, common resource management, and debt settlement where agents have quasi-linear utility and transfers are possible.
3
The improved AI system can ensure that its allocation solutions satisfy desirable properties such as efficiency, anonymity, the dummy property, continuity, additivity, and monotonicity simultaneously. This is achieved by basing the solution on the stand-alone claims corresponding to the problem
rather than relying on arbitrary coalitional games where many solutions satisfy these axioms.
4
The system can handle problems involving surpluses or deficits in alternatives by extending the model to include a surplus or deficit entry, and solving for a solution that satisfies the strengthened axiom of compromise. This allows for compensation based not only on market value but also on what each party attaches to owning the property outright.
5
The AI system can be designed to solve problems where costs are borne by agents rather than central funds, and location choices involve trade-offs between different alternatives with associated costs. This is demonstrated in applications like building shared infrastructure where the final net benefit depends on the chosen alternative.