## Abstract

Agents endowed with powers compete for a divisible prize by forming coalitions. When a coalition wins, all non-members are eliminated. The winning coalition then divides the prize among its members according to a given sharing rule. We investigate the case where the sharing rule satisfies a property we call *consistent ranking*. Consistent ranking ensures that agents’ rankings of competing coalitions coincide. Sharing rules such as equal and proportional sharing satisfy this property. We also examine a larger class of sharing rules that satisfy a property we call *generalized consistent ranking* where agents can rank coalitions even though the sharing rule does not satisfy consistent ranking. For instance, a convex combination of equal and proportional sharing, which we call *combination sharing*, violates consistent ranking but satisfies generalized consistent ranking under certain conditions. For these different sharing rules, we characterize rules on choosing coalitions (called *transition correspondence*) that satisfy two main axioms: *self-enforcement*, which requires that no further deviation happens after a coalition has formed, and *rationality*, which requires that agents pick the coalition that gives them their highest payoff. We find that a transition correspondence that satisfies self-enforcement and rationality always exists for sharing rules that satisfy generalized consistent ranking (and hence, consistent ranking). In the case of combination sharing, one way to satisfy generalized consistent ranking is to restrict the configuration of powers in society to satisfy *size-power monotonicity*, where larger coalitions have higher powers.

This is a preview of subscription content, access via your institution.

## Notes

A real-world example of this setting is the purge in the Russian Politburo documented in Acemoglu et al. (2008).

Elsewhere in this paper, we interchange the term ultimate ruling coalition with the term

*limit coalition*or*final coalition*.Equal sharing is widely used not only on theoretical grounds but also for practical purposes such as inheritance bequest (Erixson and Ohlsson 2014).

The Sen share is a convex combination of an equal share of the surplus among productive agents and a proportional share according to an agent’s labor to the total labor supplied in the economy. (Fabella 1988, 2000) shows that under increasing returns to scale, the Sen share of the production surplus can support Pareto optimal production.

The consistency critique of the traditional concept of the core stems from the requirement that the subcoalitions that object to the formation of the grand coalition must themselves be tested against further objections (see Ray and Vohra 1997 for a discussion of this consistency critique).

Using a laboratory experiment, Jandoc and Juarez (2019) show that agents do not display farsighted behavior when playing a simplified version of the game informed by the model of this paper.

The precise definition of self-enforcement is stated in Axiom 1.

A transition correspondence is a mapping that defines which coalitions form over time. The precise definition is given in Sect. 2.

For convenience, the power vector \(\pi \in {\mathbb {R}}^{N}_{+}\) can be normalized such that \(\sum _{i \in N} \pi _{i} =1\).

Juarez et al. (2022) considers a more general version where power can be any arbitrary monotonic function.

Our results can be easily adapted to require winning coalitions to have relative power larger than 50%, that is, \(\alpha \in [0.5,1)\).

Note that these three sharing rules are cross-monotonic.

A correspondence is continuous if for any sequence of power vectors \(\pi ^{1}, \pi ^{2}, \dots \rightarrow \pi ^{*}\) where \(S\in \phi (N, \pi ^{i}) \ \forall i\) and

*S*is winning in \(\pi ^{*}\), then \(S\in \phi (N, \pi ^{*})\).We do not restrict which coalition will be selected from \(\phi (S^{t-1}, \pi ^{t-1})\). This allows our results to be more robust since the evolution of the game includes any potential path of coalitions such that \(S^{t}\in \phi (S^{t-1}, \pi ^{t-1})\) for all

*t*.As long as cross-monotonicity is satisfied.

Note that \(\phi ^{**}\) is well-defined because \({\bar{G}}\) is feasible. That is, if the game \((N,\pi )\) is SPM, then clearly, for any \(S\subset N\), the game \((S,\pi _S)\) is also SPM.

## References

Acemoglu D, Egorov G, Sonin K (2008) Coalition formation in non-democracies. Rev Econom Stud 75(4):987–1009

Acemoglu D, Egorov G, Sonin K (2009) Do juntas lead to personal rule? Am Econom Rev Papers Proc 99(2):298–303

Arnold T, Schwalbe U (2002) Dynamic coalition formation and the core. J Econom Behav Organiz 49(3):363–380

Banerjee S, Konishi H, Sönmez T (2001) Core in a simple coalition formation game. Social Choice Welfare 18(1):135–153

Bernheim D, Peleg B, Whinston M (1987) Coalition-proof nash equilibria I concepts. J Econom Theory 42(1):1–12

Bloch F (1996) Sequential formation of coalitions in games with externalities and fixed payoff division. Games Econom Behav 14(1):90–123

Bloch F, Dutta B (2011). Formation of Networks and Coalitions. In Handbook of Social Economics, volume 1, pages 729–779. Elsevier,

Bogomolnaia A, Jackson M (2002) The stability of hedonic coalition structures. Games Econom Behav 38(2):201–230

Carraro C (1999) Environmental conflict, bargaining and cooperation. In: Van den Bergh J (ed) The handbook of environmental and resource economics. E. Elgar, Cheltenham

Casajus A, Huettner F (2014) Weakly monotonic solutions for cooperative games. J Econom Theory 154:162–172

Cechlarova K, Romero-Medina A (2001) Stability in coalition formation games. Int J Game Theory 29(4):487–494

Chae S, Heidhues P (2004) Nash bargaining solution with coalitions and the joint-bargaining paradox: a group bargaining solution. Math Soc Sci 48(1):37–53

Chae S, Moulin H (2010) Bargaining among groups: an axiomatic viewpoint. Int J Game Theory 39(1):71–88

Chatterjee K, Dutta B, Ray D, Sengupta K (1993) A noncooperative theory of coalitional bargaining. Rev Econom Stud 60(2):463–477

Chwe M (1994) Farsighted coalitional stability. J Econom Theory 63(2):299–325

Diamantoudi E, Xue L (2003) Farsighted stability in hedonic games. Social Choice Welfare 21(1):39–61

Dimitrov D, Haake C (2008) A note on the paradox of smaller coalitions. Social Choice Welfare 30(4):571–579

Ehlers L (2002) Coalitional strategy-proof house allocation. J Econom Theory 105(2):298–317

Erixson O, Ohlsson H (2014). Estate Division: Equal Sharing as Choice, Social Norm, and Legal Requirement. IFN Working Paper No. 1006,

Fabella R (1988) Natural team sharing and team productivity. Economics Lett 27(2):105–110

Fabella R (2000) Generalized sharing, membership size and pareto efficiency in teams. Theory Decis 48(1):47–60

Farrell J, Scotchmer S (1988) Partnerships. Quart J Econom 103(2):279–297

Greenberg J (1990) The Theory of Social Situations: an Alternative Game-Theoretic Approach. Cambridge University Press, MA

Han L, Juarez R, Vargas M (2023) Robust equilibria in tournaments. Games Econom Behav (Forthcoming)

Harsanyi J (1977) Rational behavior and bargaining equilibrium in games and social situations. Cambridge University Press, MA

Iñarra E, Kuipers J, Olaizola N (2002) Absorbing Sets in Coalitional Systems. Universidad del País Vasco, Facultad de Ciencias Económicas

Jandoc K, Juarez R (2017) Self-enforcing coalitions with power accumulation. Int J Game Theory 46(2):327–355

Jandoc K, Juarez R (2019) An experimental study of self-enforcing coalitions. Games 10(3):1–31

Jandoc K, Juarez R, Roumasset J (2015) Towards an economics of irrigation networks. In: Burnett K, Howitt K, Roumasset J, Wada CA (eds) Handbook of water economics and institutions. Routledge, New York

Jordan JS (2006) Pillage and Property. J Econom Theory 131(1):26–44

Juarez R (2013) Group strategyproof cost sharing: the role of indifferences. Games Econom Behav 82:218–239

Moulin H (1988) Axioms of Cooperative Decision Making. Cambridge University Press, Cambridge MA

Papai S (2004) Unique stability in simple coalition formation games. Games Econom Behav 48(2):337–354

Piccione M, Razin R (2009) Coalition Formation under Power Relations. Theoretical Economics 4(1):1–15

Pycia M (2012) Stability and Preference Alignment in Matching and Coalition Formation. Econometrica 80(1):323–362

Ray D (1989) Credible Coalitions and the Core. Int J Game Theory 18(2):185–187

Ray D, Vohra R (1997) Equilibrium binding agreements. J Econom Theory 73(1):30–78

Ray D, Vohra R (2013) Coalition formation. In: Young P, Zamir S (eds) Handbook of game theory, vol 4. North-Holland, Amsterdam

Ray D, Vohra R (2015) The farsighted stable set. Econometrica 83(3):977–1011

Roth A (1984) The evolution of the labor market for medical interns and residents: a case study in game theory. J Polit Economy 92(6):991–1016

Sen A (1966) Labour allocation in a cooperative enterprise. Rev Econom Stud 33(4):361–371

Shenoy P (1979) On coalition formation: a game-theoretical approach. Int J Game Theory 8(3):133–164

Xue L (1998) Coalitional stability under perfect foresight. Econom Theory 11(3):603–627

## Author information

### Authors and Affiliations

### Corresponding author

## Additional information

### Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

We appreciate the financial support from the AFOSR Young Investigator Program and the Philippine Center for Economic Development (PCED). Any errors or omissions are the authors’ own. We are grateful for the helpful comments from Daron Acemoglu, Anna Bogomolnaia, Vikram Manjunath, Hervé Moulin, Elena Iñarra, and seminar participants at the Econometric Society Conference in Mexico, Conference on Economic Design in Lund, Public Economic Theory Conference in Lisbon, Social Choice and Welfare Conference in Seoul, and the International Conference on Game Theory in Stony Brook .

## A. Appendix: Proofs

### A. Appendix: Proofs

** Proof of Lemma **
1

### Proof

Consider the sets

and

We will prove by induction on the size of *u* that \(A^u=B^u.\)

This is clearly true if \(u=1,\) because any singleton coalition is self-enforcing.

For the induction hypothesis, assume that \(A^{u-1}=B^{u-1}.\)

Consider \(S\in A^u.\) Then \(S\in \phi (S, \pi )\). Therefore, since \(\phi \) is minimalistic, there is no \(Q\subset S\) such that \(Q\in W_{(S, \pi )}\) and \(Q\in \phi (Q, \pi _{Q}).\)

Therefore, since \(A^{u-1}=B^{u-1}, \) there is no \(Q\subset S\) such that \(Q\in W_{(S, \pi )}\) and \(Q\in {\tilde{\phi }} (Q, \pi _{Q}).\)

Hence, \(S\in {\tilde{\phi }}(S, \pi )\) and \(S\in B^u.\) Thus \(A^u\subseteq B^u\).

We can similarly prove that \(B^u\subseteq A^u.\) \(\square \)

** Proof of Proposition** 1

### Proof

**Step 1.** \(\phi ^*\) is SE and minimalistic.

Proof. To show SE, take any \(X \in \phi ^*(S, \pi )\). There are two cases: either \(X=S\) or \(X \in Q(S,\pi )\). If \(X=S\), then \(X \in \phi ^*(S, \pi )=\phi ^*(X, \pi _{X})\). If \(X \in Q(S,\pi )\), then \(X \in \phi ^*(X, \pi _{X})\) by definition of the set \(Q(S,\pi )\).

On the other hand, \(\phi ^*\) is minimalistic because \(\xi \) is cross-monotonic. That is, at the coalition formation game \((S, \pi )\), the set *S* is chosen only if \(Q(S,\pi )=\emptyset \).

**Step 2.** \(\phi ^*\) satisfies RAT.

Proof. Take \(T\in \phi ^*(S, \pi )\) and consider a coalition *Z* such that \(Z\in W_{(S, \pi )}\) such that \(Z\in \phi ^*(Z, \pi _{Z})\).

(\(\Rightarrow \)) First assume that \(Z\not \in \phi ^*(S, \pi ).\) Since \(T\in \phi ^*(S, \pi )\) we have that

Notice that *Z* is winning and self-enforcing within *S*, therefore \(Z\in Q(S, \pi )\cup \{S\}\). Moreover, since \(Z\not \in \phi ^*(S, \pi )\), we have that \(Z \not \in \displaystyle {\text {*}}{arg\,max}_{M \in Q(S, \pi )\cup \{S\}} R^\xi (M,\pi _M)\). Hence, \(R^\xi (T,\pi _T)>R^\xi (Z,\pi _Z)\).

(\(\Leftarrow \)) Now, assume that \(R^\xi (T,\pi _T)>R^\xi (Z,\pi _Z)\). Then, \(Z \not \in \displaystyle {\text {*}}{arg\,max}_{M \in Q(S, \pi )\cup \{S\}} R^\xi (M,\pi _M)\). Hence, \(Z \not \in \phi ^*(S, \pi )\)

**Step 3.** Consider any cross-monotonic sharing rule and transition correspondences \(\phi \) and \({\tilde{\phi }}\) that are self-enforcing and minimalistic. Then, the sets of coalitions that are self-enforcing coincide. That is,

Proof. The proof of this result is a straightforward consequence of Lemma 1.

**Step 4.** There exists a unique transition correspondence that meets SE and RAT.

Proof. Consider a transition correspondence \(\phi \) that is SE and RAT. Then, \(\phi \) is minimalistic because the sharing rule is cross-monotonic. We will show that \(\phi =\phi ^*.\)

Since \(\phi \) and \(\phi ^*\) are SE and RAT, by step 3, we have,

Suppose \(X\in \phi (X, \pi )\). Then, by equation 2, \(X\in \phi ^*(X, \pi )\). By RAT, for any \(S\subset X\), \(S\ne X\), then \(S\not \in \phi (X,\pi )\). Hence, \(\phi (X, \pi )=\phi ^*(X, \pi ).\)

On the other hand, suppose \(S\in \phi (X, \pi )\), where \(S\not =X\). Then, by RAT, \(\xi _i(S,\pi _S)\ge \xi _i(T,\pi _T)\) for \(i\in S\cap T\) for any coalition *T* such that \(T\in \phi (T, \pi _T)\) and \(T\in W_{(X,\pi )}\). Therefore, by consistent ranking, \(R^\xi (S,\pi _S)\ge R^\xi (T,\pi _T)\) for any coalition *T* such that \(T\in \phi (T, \pi _T)\) and \(T\in W_{(X,\pi )}\). Hence, \(S\in \phi ^*(X, \pi )\) and \(\phi (X, \pi )\subset \phi ^*(X, \pi )\). We could similarly show that \(\phi ^*(X, \pi )\subset \phi (X, \pi ).\) Therefore, \(\phi (X, \pi )=\phi ^*(X, \pi ).\)

**Step 5.** \(\phi ^{*}\) is superior to any transition correspondence that is SE and minimalistic.

Proof. We prove this step by contradiction. Suppose \(\phi ^{*}\) is not superior to the SE and minimalistic transition correspondence \({\hat{\phi }}\). Then, there exists a game \((N, \pi )\) such that \(S, T \subset N\) where \(S\in \phi ^{*}(N,\pi )\) and \(T\in {\hat{\phi }}(N,\pi )\) such that \(\xi _i(T,\pi _T)\ge \xi _{i}(S,\pi _S)\) for some \(i\in T\cap S\).

By step 3, since *T* is self-enforcing for \({{\hat{\phi }}}\), it follows that it is also self-enforcing for \(\phi ^*.\) Therefore, \(T\in Q(N, \pi )\).

Since \(T\not \in \phi ^{*}(N,\pi )\), it follows that \(T \not \in \displaystyle {\text {*}}{arg\,max}_{M \in Q(N, \pi )\cup \{N\}} R^\xi (M,\pi _M)\). Since \(S \in \phi ^{*}(N,\pi )\), it follows that \(S \in \displaystyle {\text {*}}{arg\,max}_{M \in Q(N, \pi )\cup \{N\}} R^\xi (M,\pi _M)\). Since *S* and *T* are winning within *N* (by the definition of a transition correspondence), we have \(T\cap S\ne \emptyset \). Therefore, \(R^\xi (S,\pi _{S})>R^\xi (T,\pi _T)\), which implies that \(\xi _{i}(S, \pi _S)>\xi _{i}(T, \pi _T)\) for \(i\in T\cap S\). This is a contradiction.

\(\square \)

** Proof of Proposition** 2

### Proof

We prove this result in three steps.

**Step 1**. Suppose \(\phi \) and \({\tilde{\phi }}\) are transition correspondences that satisfy SE and RAT under the sharing rule \(\xi \). Then, \(\phi ={\tilde{\phi }}\).

Proof. Suppose there is a game \((N,\pi )\) such that \(\phi (N,\pi )\not ={\tilde{\phi }}(N,\pi )\). Without loss of generality, let \(T\in {\tilde{\phi }}(N,\pi )\) and \(T\not \in \phi (N,\pi )\). Since RAT implies MIN, by Lemma 1, the set of self-enforcing coalitions generated under \(\phi \) and \({\tilde{\phi }}\) coincide. Since *T* was not chosen in \(\phi \), by rationality, there exists \(S\in SEC(N,\pi )\) such that \(\xi _i(S,\pi _S)> \xi _i(T,\pi _T)\) for some \(i\in T\cap S\). By rationality of \({\tilde{\phi }}\), \(\xi _i(T,\pi _T)\ge \xi _i(S,\pi _S)\) for all \(i\in T\cap S\). This is a contradiction.

**Step 2**. Show that \(UD^{\xi }(N,\pi )\not =\emptyset \).

Since \(\phi \) is a transition correspondence, \(\phi (N,\pi )\not =\emptyset \). Let \(T\in \phi (N,\pi )\). Since \(\phi \) satisfies SE, \(T\in SEC(N,\pi )\). By RAT, for all \(S\in SEC(N,\pi )\), \(\xi _i(T,\pi _T)\ge \xi _i(S,\pi _S)\) for all \(i\in S\cap T\). Therefore, \(T\in UD^{\xi }(N,\pi )\).

**Step 3**. We show that \(\phi (N,\pi )= UD^{\xi }(N,\pi )\) satisfies SE and RAT.

If \(T\in {\phi }(N,\pi )= UD^{\xi }(N,\pi )\), by Definition 6, \(T\in SEC(N,\pi )\). Thus, \(\phi \) satisfies SE.

Let coalition \(T\in \phi (N,\pi )=UD^{\xi }(N,\pi )\). By definition of \(UD^{\xi }(N,\pi )\), we have that \(\xi _i(T,\pi _T)\ge \xi _i(S,\pi _S)\) for all \(S\in SEC(N,\pi )\) and for all \(i\in S\cap T\). Thus, \(\phi \) satisfies RAT.

\(\square \)

** Proof of Proposition** 3

### Proof

To prove part (i), let the game be \((N,\pi )\in {\bar{G}}\) and assume combination sharing and fix a \(\lambda \).

For any game \((S,\pi _{S})\subset (N,\pi )\), agent *i*’s share in a coalition *S* is:

Since \((N,\pi )\in {\bar{G}}\), if there exists another game \((T,\pi _{T})\subset (N,\pi )\) such that \(|S|<|T|\), it must also be true that \(\pi (S)<\pi (T)\). Therefore,

for all agents in \(i\in S\cap T\).

Hence, we can construct a ranking function

where coalitions with smaller size (and therefore lesser power by size-power monotonicity) will yield a higher value of \(R^{CS^{\lambda }}(S,\pi )\). Hence, the sharing rule satisfies consistent ranking, and by Proposition 1, the transition correspondence \(\phi ^{**}\) is the unique transition correspondence satisfying SE and RAT over the domain of SPM games.

To prove part (ii), assume a transition correspondence \(\phi \) that satisfies SE and RAT. Suppose we have two arbitrary coalitions *S* and *T*, such that \((S,\pi _{S})\in {\bar{G}}\) and \((T,\pi _{T})\in {\bar{G}}\). With combination sharing, agents in the intersection of *S* and *T* will have a higher share in *S* whenever

As \(\lambda \rightarrow 1\), \(\xi _{i}(S,\pi _{S})\rightarrow \frac{1}{|S|}\) and \(\xi _{i}(T,\pi _{T})\rightarrow \frac{1}{|T|}\). Hence to maintain the inequality, it must be true that \(|S|<|T|\). As \(\lambda \rightarrow 0\), \(\xi _{i}(S,\pi _{S})\rightarrow \frac{\pi _{i}}{\pi (S)}\) and \(\xi _{i}(T,\pi _{T})\rightarrow \frac{\pi _{i}}{\pi (T)}\). Hence to maintain the inequality, it must be true that \(\pi (S)<\pi (T)\). Therefore, for a ranking function \(R^{CS^{\lambda }}\) to exist for all values of \(\lambda \) and for all games in \({\bar{G}}\), we should have that for any two coalitions *S* and *T*, such that \((S,\pi _{S})\in {\bar{G}}\) and \((T,\pi _{T})\in {\bar{G}}\), if \(|S|<|T|\) then \(\pi (S)<\pi (T)\).

The proof that \(\phi =\phi ^{**}\) is similar to Step 4 in the proof of Proposition 1, restricted to the domain of SPM games.

\(\square \)

## Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

## About this article

### Cite this article

Jandoc, K., Juarez, R. Generalized consistent ranking and the formation of self-enforcing coalitions.
*Rev Econ Design* (2023). https://doi.org/10.1007/s10058-023-00335-9

Received:

Accepted:

Published:

DOI: https://doi.org/10.1007/s10058-023-00335-9