Strategic social choice. Stable representations of constitutions (Q606269)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Strategic social choice. Stable representations of constitutions |
scientific article |
Statements
Strategic social choice. Stable representations of constitutions (English)
0 references
17 November 2010
0 references
In the 1950's Arrow defined a constitution as a social welfare function assigning a social ranking of the alternatives to every profile of individual linear orderings. In this book the authors give a another definition that models actual real-life constitutions much better. A society is defined as a list \({\mathcal S} = (N, A, \rho, \alpha, \gamma)\), where \(N\) is the finite set of members of \({\mathcal S}\), \(A\) is the set of \textit{social states} or \textit{alternatives}, \(\rho\) is a finite set of \textit{rights}, \(\alpha: P(N) \rightarrow P(\rho)\) with \(\alpha(\emptyset) = \emptyset\) is the current assignment of rights to groups of members of \({\mathcal S}\), \(\gamma: P(N) \times P(\rho) \rightarrow P(P_{0}(A))\), where \(P_{0}(A)\) is the set of all non-empty subsets of \(A\), such that \(\gamma(\emptyset, \theta) = \emptyset\) and \(A \in \gamma(S, \theta)\) for all \(\theta \subseteq \rho\) and \(S \in P_{0}(N)\), determines the sets of attainable social states for groups of members of \(S\) as a function of their rights. Intuitively, rights are means to reach certain social states. If \({\mathcal S} = (N, A, \rho, \alpha, \gamma)\) is a society, the triple \((\rho, \alpha, \gamma)\) is called a \textit{constitution}. Of course, one may combine the functions \(\alpha\) and \(\gamma\) into a function \(E: P(N) \rightarrow P(P_{0}(A))\), defined by \(E(S) = E(S; \alpha, \gamma) = \cup_{T \subseteq S} \;\gamma(T, \alpha(T))\). So, according to \(E\), a coalition \(S\) is entitled to all sets of social states to which some sub-coalition of \(S\) is legally entitled. \(E\) satisfies (i) \(E(\emptyset) = \emptyset\) and (ii) \(A \in E(S)\) for every \(S \neq \emptyset\). In order to introduce the concept of an effectivity function, we define a \textit{structure} on \(A\) as a set \({\mathcal T}\) \(\subseteq P_{0}(A)\) such that \(A \in {\mathcal T}\) and \(B_1 \cap B_2 \in {\mathcal T}\) for all \(B_1, B_2 \in {\mathcal T}\) with \(B_1 \cap B_2 \neq \emptyset\). Of course, \({\mathcal T}\) = \(P_{0}(A)\) is a structure on \(A\). Given a structured space \((A, {\mathcal T})\), an \textit{effectivity function} is a function \(E: P(N) \rightarrow P({\mathcal T})\) that satisfies (i) \(E(\emptyset) = \emptyset\), (ii) \(A \in E(S)\) for every \(S \neq \emptyset\) and (iii) \(E(N) = {\mathcal T}\). \(E\) is \textit{monotonic with respect to coalitions} iff \(S\subseteq T \rightarrow E(S) \subseteq E(T)\) for all \(S, T \in P(N)\), and \(E\) is \textit{monotonic with respect to outcomes} iff for all \(B, B' \in {\mathcal T}\) and \(S \subseteq N\), \(B \in E(S) \;\& \;B \subseteq B' \rightarrow B' \in E(S)\). \(E\) is called \textit{monotonic} if it satisfies both properties. \(E\) is \textit{superadditive} iff for all \(S_1, S_2 \in P_{0}(N)\) with \(S_1 \cap S_2 = \emptyset\) and for all \(B_1 \in E(S_1)\) and \(B_2 \in E(S_2)\) we have \(B_1 \cap B_2 \in E(S_1 \cup S_2)\). Let \({\mathcal S}\) be a society and suppose the constitution is described by the function \(E: P(N) \rightarrow P({\mathcal T})\). Then \(E\) describes for any coalition the sets of social states to which this coalition is legally entitled, but it does not tell us how the members of society can actually exercise their rights. Therefore, we assume that every society member has at its disposal a set of ``legal'' strategies. A \textit{game form} is a list \(\Gamma = (N; \Sigma^1, \ldots, \Sigma^n; g; A)\) where \(N\) is the set of members of the society or \textit{players}, \(\Sigma^i\) is the non-empty set of \textit{strategies} of \(i \in N\), \(g: \Sigma = \Sigma^1 \times \ldots \times \Sigma^n \rightarrow A\) is the \textit{outcome function} and \(A\) is the set of social states or outcomes. Given a game form \(\Gamma = (N; \Sigma^1, \ldots, \Sigma^n; g; A)\), a structure \({\mathcal T}\) on \(A\), a non-empty coalition \(S \subseteq N\) and a set \(B \in {\mathcal T}\), \(S\) is called \textit{effective} for \(B\) iff there exists \(\sigma_{0}^{S} \in \Sigma^{S} = \Pi_{i \in S} \Sigma^i\) such that \(g(\sigma_{0}^{S}, \sigma^{N \setminus S}) \in B\) for all \(\sigma^{N \setminus S} \in \Sigma^{N \setminus S}\). If \(g\) is surjective, then \(E^{\Gamma}: P(N) \rightarrow P({\mathcal T})\) defined by \(E^{\Gamma}(\emptyset) = \emptyset\) and \(E^{\Gamma}(S) = \{B \in {\mathcal T} \;| \;S \text{ is effective for } B\}\) is the \textit{effectivity function} for \((A, \mathcal{T})\) \textit{associated with} \(\Gamma\). Surjectivity of \(g\) implies that \(E^{\Gamma}(N) = {\mathcal T}\), so \(E^{\Gamma}\) is indeed an effectivity function. Let \({\mathcal T}\) be a structure on \(A\) and let \(E: P(N) \rightarrow P({\mathcal T})\) be an effectivity function. Then a game form \(\Gamma\) is called a \textit{representation} of \(E\) iff \(E^{\Gamma} = E\). The most basic theorem in this monograph is the following one (Theorem 2.4.7): \(E\) has a representation \(\Gamma\) iff \(E\) is monotonic and superadditive. The monograph studies representations of effectivity functions by game forms that satisfy, at least, the minimal stability requirement of having a Nash equilibrium for any profile of individual preferences. Part I studies strategic stability properties of those game forms, Pareto optimality of equilibrium outcomes of those game forms and continuity properties of game forms. A social choice function assigns an alternative to any profile of individual preferences and is, thus, a game form where the strategies of the players are their individual preferences. In part II the focus is on social choice functions that admit a strong Nash equilibrium resulting in the sincere outcome. The book is well written and the definitions are illustrated by nice examples. It is not an easy book, but the material is presented in a clear way with motivating discussions and examples. The book is based on work that has appeared over the last thirty years, but it also contains new results and new or improved proofs of existing results. It is a fine contribution, bringing together a lot of material in a single book in a coherent way.
0 references
constitution
0 references
effectivity function
0 references
game form
0 references
Nash consistent
0 references