Egalitarian division under Leontief preferences (Q2434963)

In the paper under review the authors study the problem of fairly dividing a divisible goods among \(n\) agents with generalized Leontief preferences, and characterize the class of generalized egalitarian rules. Results on agent-specific and endowment-specific egalitarian rules are also obtained. Section 1 (Introduction) states that \textit{M. Ghodsi} et al. [``Dominant resource fairness: fair allocation of heterogeneous resources in datacenters'' Technical Report, No. UCB/EECS-2010-55, EECS Department, University of California, Berkeley (2010)], \textit{A. Nicolò} [Rev. Econ. Des. 8, No. 4, 373--382 (2004; Zbl 1083.91061)] and \textit{E. Pazner} and \textit{D. Schmeidler} [``Egalitarian equivalent allocations: a new concept of economic equity'', Q. J. Econ. 92, 671--687 (1978)] inspired this work. Leontief preferences and the corresponding non-wasteful rules are of practical interest, e.g., for such as multiple resource sharing problems in cloud computing systems. The paper extends existing results in a more general setup. First, the authors define generalized egalitarian rules assuming that there is a continuous monotonic benchmark preference in the commodity space owned by the society. Under Leontief preferences, a class of non-wasteful rules generalizes the egalitarian equivalence rule proposed by Pazner and Schmeidler [loc. cit.] and satisfies efficiency group strategy-proofness and almost all the fairness axioms. Egalitarian rules set essentially a standard for society to measure difference ordinal preferences of the agents so that they are treated equally by this standard. \textit{W. Thomson} [Math. Oper. Res. 8, 319--326 (1983; Zbl 0524.90102)] proposed a concept of equity to capture the notion of equal opportunities. It turns out that a general egalitarian rule always picks the Pareto-optimal equal opportunity allocation relative to a corresponding family of nested choice sets. One of the authors' main result in Theorem 1, shows that a generalized egalitarian rule satisfies efficiency, group strategy-proofness, anonymity, resource monotonicity, population monotonicity, envy-freeness and consistence; and conversely, given an efficient resource monotonic and consistent rule with those properties, it must be a generalized egalitarian rule. Another main result of the paper in Theorem 2, shows that the axiomatized characterization holds for generalized Leontief preferences. In real life, generalized Leontief preferences are relevant when the agents are production units and the goods are inputs. For example, a group of people are dividing some cotton, silk and lace to make clothes. They would like to use these materials in different proportions according to their own tastes. More material of one kind may be useless, which captures the essence of a Leontief preference. The paper is crucially relying on non-wasteful rules. The authors provide an example to illustrate that the restriction to non-wasteful rules is important for the obtained results. Theorems 3 and 4 extend generalized egalitarian rules in two directions: a rule assigns to each agent a personal welfare index and equalizes their utilities according to these agent-specific welfare indices, and the second is the voluntary participation of the agents in the social reallocation. The welfare indices are always an equal treatment allocation, in this case, to give to every agent a minimum bundle that provides him/her the same level of welfare as his private endowment. The authors prove that endowment-specific egalitarian rules are efficient, group strategy-proof, anonymous, consistent and individual rational. Both agent-specific and endowment-specific rules lack of axiomatization as in the previous Theorem 2. Section 2 sets up the model, notations and definitions as well as the rules. The authors define a generalized egalitarian rule and characterize these rules in Theorem 1 in Section 3, examples are provided. Section 4 devotes to generalized Leontief preferences, and Theorem 2 gives necessary and sufficient conditions to characterize the class of generalized egalitarian rules under the generalized Leontief preferences. The proof of Theorem 2 takes the whole Section 5. Section 6 gives the tightness of the characterization in Theorem 2. Section 7 studies the agent-specific egalitarian rules, and Section 8 studies the endowment-specific egalitarian rules and private property, to possess those properties. The paper concludes in Section 9. The appendix section provides the proofs of the results in Section 3.