The separability axiom and equal-sharing methods (Q1083005): Difference between revisions

Let \((N,A)\) be a social choice problem, where \(N\) is the set of voters and \(A\) a set of alternatives, both finite. Each voter \(i\) has a utility \(u_i(a)+t_i\), where \(a\) is an alternative and \(t_i\) is a side payment. If \(a\) is chosen, it generates a cost \(c(a)\) which must be paid. A social choice function \(S\) is a map from the characteristics \((u_i;c)\) to the set of voters' utility possibilities. \(S\) satisfies the separability axiom if the social choice is not affected by the removal of voter \(i\), when \(i\) retains his former utility level. The paper studies social choice functions \(S\) which are Pareto optimal, anonymous, independent of utility and cost function zeroes, and satisfy the separability axiom. This class of \(S\) includes the Vickrey-Clarke-Groves mechanism, utilitarian mechanisms without side-payments, and mechanisms which share the surplus equally.