Abstract
We introduce a new operator for general rationing problems in which, besides conflicting claims, individual baselines play an important role in the rationing process. The operator builds onto ideas of composition, which are not only frequent in rationing, but also in related problems such as bargaining, choice, and queuing. We characterize the operator and show how it preserves some standard axioms in the literature on rationing. We also relate it to recent contributions in such literature.
Similar content being viewed by others
Notes
The reader is referred to Thomson (2003) for a survey.
The notion of operators for the domain of rationing rules was first introduced by Thomson and Yeh (2008).
These properties are reminiscent of the so-called “path independence” axiom for choice functions (e.g., Plott 1973). They also have a relative in the theory of axiomatic bargaining: the so-called “step-by-step negotiations” axiom introduced by Kalai (1977), which is the basis for the characterization of the egalitarian solution in such context. The same principle has also been frequently used in other related contexts like taxation, queuing, or resource allocation (e.g., Moulin 2000; Moulin and Stong 2002; Moreno-Ternero and Roemer 2012).
For each \(N\in\mathcal{N}\), each M⊆N, and each \(z\in\mathbb{R}^{n} \), let z M ≡(z i ) i∈M . For each i∈N, let z −i ≡z N∖{i}.
For the case E≥T(b,c), both operators agree with the proposal made by Pulido et al. (2002) for bankruptcy problems with objective entitlements, which are a specific instance of our bankruptcy problems with baselines.
Such a question was initially posed by Thomson and Yeh (2008) with respect to the operators they study.
Note that the last two properties are dual, whereas the first one is self-dual.
Although the axiom of consistency is often described as an operational principle, solidarity underpinnings for it have also been argued (e.g., Thomson 2012).
Note that the first two properties in this block are dual, whereas the last two are self-dual.
For ease of exposition, we skip the straightforward definitions of the general versions of each axiom introduced above.
The terminology is borrowed from Hokari and Thomson (2008).
Note that t N∖{i}(b,c′)≡t N∖{i}(b,c), t i (b,c′)≥t i (b,c) and thus, T′(b,c)≥T(b,c).
References
Aumann, R. J., & Maschler, M. (1985). Game theoretic analysis of a bankruptcy problem from the Talmud. Journal of Economic Theory, 36, 195–213.
Chun, Y. (1999). Equivalence of axioms for bankruptcy problems. International Journal of Game Theory, 28, 511–520.
Hokari, T., & Thomson, W. (2008). On properties of division rules lifted by bilateral consistency. Journal of Mathematical Economics, 44, 1057–1071.
Hougaard, J. L., Moreno-Ternero, J. D., & Østerdal, L. P. (2012). A unifying framework for the problem of adjudicating conflicting claims. Journal of Mathematical Economics, 48, 107–114.
Hougaard, J. L., Moreno-Ternero, J. D., & Østerdal, L. P. (2013). Rationing in the presence of baselines. Social Choice and Welfare, 40, 1047–1066.
Ju, B.-G., Miyagawa, E., & Sakai, T. (2007). Non-manipulable division rules in claim problems and generalizations. Journal of Economic Theory, 132, 1–26.
Kalai, E. (1977). Proportional solutions to bargaining situations: interpersonal utility comparisons. Econometrica, 45, 1623–1630.
Kaminski, M. (2004). Hydraulic rationing. Mathematical Social Sciences, 40, 131–155.
Kaminski, M. (2006). Parametric rationing methods. Games and Economic Behavior, 54, 115–133.
Moreno-Ternero, J., & Roemer, J. (2006). Impartiality, priority, and solidarity in the theory of justice. Econometrica, 74, 1419–1427.
Moreno-Ternero, J., & Roemer, J. (2012). A common ground for resource and welfare egalitarianism. Games and Economic Behavior, 75, 832–841.
Moreno-Ternero, J. D., & Villar, A. (2006). New characterizations of a classical bankruptcy rule. Review of Economic Design, 10, 73–84.
Moulin, H. (2000). Priority rules and other asymmetric rationing methods. Econometrica, 68, 643–684.
Moulin, H., & Stong, R. (2002). Fair queuing and other probabilistic allocation methods. Mathematics of Operations Research, 27, 1–30.
O’Neill, B. (1982). A problem of rights arbitration from the Talmud. Mathematical Social Sciences, 2, 345–371.
Plott, C. (1973). Path independence, rationality, and social choice. Econometrica, 41, 1075–1091.
Pulido, M., Sanchez-Soriano, J., & Llorca, N. (2002). Game theory techniques for university management: an extended bankruptcy model. Annals of Operations Research, 109, 129–142.
Pulido, M., Borm, P., Hendrichx, R., Llorca, N., & Sanchez-Soriano, J. (2008). Compromise solutions for bankruptcy situations with references. Annals of Operations Research, 158, 133–141.
Roemer, J. (1986). Equality of resources implies equality of welfare. The Quarterly Journal of Economics, 101, 751–784.
Thomson, W. (1983). The fair division of a fixed supply among a growing population. Mathematics of Operations Research, 8, 319–326.
Thomson, W. (2003). Axiomatic and game-theoretic analysis of bankruptcy and taxation problems: a survey. Mathematical Social Sciences, 45, 249–297.
Thomson, W. (2007). Consistent allocation rules. Book manuscript, University of Rochester.
Thomson, W. (2012). On the axiomatics of resource allocation: interpreting the consistency principle. Economics and Philosophy, 28, 385–421.
Thomson, W., & Yeh, C.-H. (2008). Operators for the adjudication of conflicting claims. Journal of Economic Theory, 143, 177–198.
Young, P. (1987). On dividing an amount according to individual claims or liabilities. Mathematics of Operations Research, 12, 398–414.
Young, P. (1988). Distributive justice in taxation. Journal of Economic Theory, 44, 321–335.
Young, P. (1994). Equity. Princeton: Princeton University Press.
Author information
Authors and Affiliations
Corresponding author
Additional information
We thank Jorge Alcalde-Unzu, William Thomson and three anonymous referees for helpful comments and suggestions. Financial support from the Spanish Ministry of Science and Innovation (ECO2011-22919) as well as from the Andalusian Department of Economy, Innovation and Science (SEJ-4154, SEJ-5980) via the “FEDER operational program for Andalusia 2007–2013”, and the Danish Strategic Research Council is gratefully acknowledged.
Appendix: Proof of the results
Appendix: Proof of the results
Proof of Theorem 1
It is straightforward to see that, for any (standard) rule R, R c satisfies the four axioms. Thus, we focus on the converse implication. Let S be a rule satisfying the four axioms and let (N,b,c,E) be a problem with baselines. If E≥T(b,c), then, by baseline invariance, and baseline truncation, S(N,b,c,E)=S(N,0,c−t(b,c),E−T(b,c)). If E≤T(b,c), by baseline truncation and truncation of excessive claims, S(N,b,c,E)=t(b,c)+S(N,t(b,c),t(b,c),E). Moreover, by polar baseline equivalence, S(N,t(b,c),t(b,c),E)=S(N,0,t(b,c),E).
Let \(R:\mathcal{D}\to\bigcup_{N\in\mathcal{N}}\mathbb{R}^{n}\) be such that, for any \((N,c,E )\in\mathcal{D}\),
In other words, R assigns to each problem the allocation that S yields for the corresponding problem with baselines in which baselines are null. Then,
which implies that S≡R c. □
Proof of Theorem 2
Note first that, by Theorem 1 in Hougaard et al. (2013), the three properties are preserved by the baselines-first extension operator. As this operator coincides with the composition extension operator for the case in which truncated baselines are collectively feasible, we only need to focus on the opposite case to prove the statements of the theorem.
Let R be a rule satisfying resource monotonicity. Let \((N,c,E )\in\mathcal{D}\) and each E′>E, with E′≤∑c i . Let \(b\in\mathbb{R}^{n}\) be a baselines profile and let i∈N be a given agent. If E<E′≤T(b,c), then \(R^{c}_{i}(N,b,c,E)=R_{i}(N,t(b,c),E)\) and \(R^{c}_{i}(N,b,c,E^{\prime })=R_{i}(N,t(b,c),E^{\prime })\). Now, as R satisfies resource monotonicity, the desired inequality follows.
Let R be a rule satisfying consistency. Let \((N,c,E)\in\mathcal{D}\) and \(b\in\mathbb{R}^{n}_{+}\). Let x=R c(N,b,c,E). The aim is to show that, for any M⊂N,
Fix M⊂N and let E′=∑ j∈M x j and T′(b,c)=∑ j∈M t j (b,c). Then, it is straightforward to show that E≤T(b,c) if and only if E′≤T′(b,c). If E≤T(b,c), then x i =R i (N,t(b,c),E) for each i∈N, and thus E′=∑ i∈M R i (N,t(b,c),E). Therefore, \(R^{c}_{i}(M,b_{M},c_{M},E^{\prime })=R_{i}(M,t_{M}(b,c),E^{\prime})\) for each i∈M. Now, as R is consistent, it follows that R i (N,t(b,c),E)=R i (M,t M (b,c),E′), for each i∈M, as desired.
To conclude, the statement on resource-population uniformity follows from the fact that such axiom is equivalent to the combination of resource monotonicity and consistency (e.g., Hougaard et al. 2013). □
Proof of Theorem 3
Again, by Theorem 3 in Hougaard et al. (2013), we only need to focus on the case in which truncated baselines are collectively unfeasible to prove the statements of the theorem.
Let R be a rule satisfying claims monotonicity and linked claims-resource monotonicity. Our aim is to show that R c satisfies the general versions of the two properties.
Claims monotonicity Let \((N,c,E )\in\mathcal{D}\) and i∈N, such that \(c_{i}\leq c_{i}^{\prime }\). Let \(b\in\mathbb{R}^{n}_{+}\) be a baselines profile, T(b,c)=∑ j∈N t j (b,c), and T′(b,c)=∑ j∈N t j (b,c′).Footnote 12
If E≤T(b,c), then, \(R^{c}_{i}(N,b,c,E)=R_{i}(N,t(b,c),E)\) and
As R satisfies claims monotonicity, the desired inequality follows.
If T(b,c)<E<T′(b,c), then, \(R^{c}_{i}(N,b,c,E)= t_{i} (b,c)+ R_{i}(N,c -t(b,c), E- T(b,c)) \), and \(R^{c}_{i}(N,b,c^{\prime },E)=R_{i}(N,t(b,c^{\prime }),E)\). Now, this case implies that t i (b,c)=c i (otherwise, t i (b,c)=b i and hence T(b,c)=T′(b,c)). Thus, by boundedness, \(R^{c}_{i}(N,b,c,E)= c_{i}\). Now, by resource monotonicity and claims monotonicity of R, R i (N,t(b,c′),E)≥R i (N,t(b,c),T(b,c))=t i (b,c)=c i , from where the desired inequality follows.
Linked claims-resource monotonicity Let \((N,c,E )\in \mathcal{D}\) and i∈N. Let \(b\in\mathbb{R}^{n}_{+}\) be a baselines profile, ε>0 and c′=(c i +ε,c N∖{i}). Let T′(b,c)=T(b,c)+t i (b,c′)−t i (b,c). Then, t i (b,c′)≤t i (b,c)+ε and T(b,c)≤T′(b,c)≤T(b,c)+ε.
If E≤T′(b,c)−ε, then \(R^{c}_{i}(N,b,c,E)=R_{i}(N,t(b,c),E)\) and \(R^{c}_{i}(N,b,c^{\prime },E+\varepsilon)=R_{i}(N,t(b,c^{\prime }),E+\varepsilon)=R_{i}(N,(t_{i}(b,c^{\prime }),t_{-i}(b,c)),E+\varepsilon)\). By claims monotonicity of R, R i (N,t(b,c′),E+ε)≤R i (N,(t i (b,c)+ε,t −i (b,c)),E+ε). By linked claims-resource monotonicity of R, R i (N,(t i (b,c)+ε,t −i (b,c)),E+ε)≤R i (N,t(b,c),E)+ε, from where the desired inequality follows.
If T′(b,c)−ε<E<T(b,c), then \(R^{c}_{i}(N,b,(c_{i}+\varepsilon ,c_{N\setminus\{i\}}),E+\varepsilon)=t_{i}(b,c^{\prime })+R_{i}(N,(c_{i}+\varepsilon-t_{i}(b,c^{\prime }) ,(c-t(b,c))_{-i}),E+\varepsilon-T^{\prime }(b,c))\), and \(R^{c}_{i}(N,b,c,E)=t_{i}(b,c^{\prime })-R^{d}_{i}(N,t(b,c),T(b,c) -E)\). Thus, the desired inequality becomes
Now, by balance and boundedness, the right hand side of (3) is bounded above by T(b,c)−E+E+ε−T′(b,c), which is precisely the left hand side of (3).
As for the second statement of the theorem, let R be a rule satisfying resource monotonicity, population monotonicity and linked resource-population monotonicity. By Theorem 2, R c satisfies the general property of resource monotonicity. Our aim is to show that R c also satisfies the general properties of population monotonicity and linked resource-population monotonicity.
Population monotonicity Let \((N,c,E )\in\mathcal{D}\) and \((N^{\prime},c^{\prime},E^{\prime} )\in\mathcal{D}\) be such that N⊆N′, \(c_{N}^{\prime}=c\) and E=E′. Let \(b\in \mathbb{R}^{n}_{+}\) and \(b^{\prime}\in\mathbb{R}^{n^{\prime}}_{+}\) be two baselines profiles such that \(b_{N}^{\prime}=b\). Note that t j (b′,c′)=t j (b,c) for each j∈N. Finally, let T(b,c)=∑ j∈N t j (b,c) and T′(b,c)=∑ j∈N′ t j (b′,c′). If E≤T(b,c), \(R^{c}_{i}(N^{\prime},b^{\prime},c^{\prime},E)=R_{i}(N^{\prime},t(b^{\prime},c^{\prime}),E)\) and \(R^{c}_{i}(N,b,c,E)=R_{i}(N,t(b,c),E)\). As R satisfies population monotonicity, the desired inequality follows.
Linked resource-population monotonicity Let \((N,c,E )\in \mathcal{D}\) and \((N^{\prime},c^{\prime},E )\in\mathcal{D}\) be such that N⊆N′ and \(c_{N}^{\prime}=c\). Let \(b\in\mathbb{R}^{n}_{+}\) and \(b^{\prime}\in\mathbb{R}^{n^{\prime}}_{+}\) be two baselines profiles such that \(b_{N}^{\prime}=b\). Note that t j (b′,c′)=t j (b,c) for each j∈N. Finally, let T(b,c)=∑ j∈N t j (b,c), and T′(b,c)=∑ j∈N′ t j (b′,c′). If \(E\le T(b,c)-\sum_{N^{\prime}\setminus N} (c^{\prime}_{j}-t_{j}(b^{\prime},c^{\prime}) )\), then E≤T(b,c) and E′≤T′(b,c) and, therefore, \(R^{c}_{i}(N^{\prime},b^{\prime},c^{\prime},E^{\prime})=R_{i}(N^{\prime},t(b^{\prime},c^{\prime}),E^{\prime})\) and \(R^{c}_{i}(N,b,c,E)=R_{i}(N,t(b,c),E)\). By resource monotonicity and population monotonicity of R, the desired inequality follows. □
Proof of Theorem 4
In order to prove the first statement, let R be a rule satisfying order preservation and let (N,b,c,E) be an extended problem for which baselines are ordered like claims. Let i,j∈N be such that c i ≤c j . As b i ≤b j it follows that t i (b,c)≤t j (b,c). Now, if E≤T(b,c), \(R^{c}_{i}(N,b,c,E)=R_{i}(N, t(b,c),E)\) and \(R^{c}_{j}(N,b,c,E)=R_{j}(N,t(b,c),E)\). As R is order preserving, and t i (b,c)≤t j (b,c), it follows that \(R^{c}_{i}(N,b,c,E)\le R^{c}_{j}(N,b,c,E)\), as desired. If, on the other hand, E≥T(b,c), the result follows from the proof in Hougaard et al. (2013).
As for the second statement, let R be a rule satisfying order preservation and let (N,b,c,E) be an extended problem for which baselines, and claim-baseline differences, are ordered like claims. Let i,j∈N be such that c i ≤c j . It then follows that t i (b,c)≤t j (b,c) and that c i −t i (b,c)≤c j −t j (b,c).
Now, if E≤T(b,c), \(R^{c}_{i}(N,b,c,E)=R_{i}(N, t(b,c),E)\) and \(R^{c}_{j}(N,b,c,E)= R_{j}(N, t(b,c),E)\).
As R is order preserving, and t i (b,c)≤t j (b,c), it follows that \(R^{c}_{i}(N,b,c,E)\le R^{c}_{j}(N,b,c,E)\).
As R d is order preserving, and c i −t i (b,c)≤c j −t j (b,c), it follows that
i.e., \(c_{i}-R^{c}_{i}(N,b,c,E)\le c_{j}-R^{c}_{j}(N,b,c,E)\), as desired.
If, on the other hand, E≥T(b,c), \(R^{c}_{i}(N,b,c,E)=t_{i}(b,c)+R_{i}(N, c-t(b,c),E-T(b,c))\) and \(R^{c}_{j}(N,b,c,E)=t_{j}(b,c)+R_{j}(N,c-t(b,c),E-T(b,c))\). As t i (b,c)≤t j (b,c), c i −t i (b,c)≤c j −t j (b,c), and R is order preserving, it follows that \(R^{c}_{i}(N,b, c,E)\le R^{c}_{j}(N,b,c,E)\) and \(c_{i}-R^{c}_{i}(N,b,c,E)\le c_{j}-R^{c}_{j}(N,b,c,E)\), as desired. □
Rights and permissions
About this article
Cite this article
Hougaard, J.L., Moreno-Ternero, J.D. & Østerdal, L.P. Rationing with baselines: the composition extension operator. Ann Oper Res 211, 179–191 (2013). https://doi.org/10.1007/s10479-013-1471-8
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10479-013-1471-8