Strategy Space Reduction in Maskin's Theorem: Sufficient Conditions for Nash Implementation (Q3798437)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 4064708
Language Label Description Also known as
default for all languages
No label defined
    English
    Strategy Space Reduction in Maskin's Theorem: Sufficient Conditions for Nash Implementation
    scientific article; zbMATH DE number 4064708

      Statements

      Strategy Space Reduction in Maskin's Theorem: Sufficient Conditions for Nash Implementation (English)
      0 references
      0 references
      0 references
      1988
      0 references
      Maskin's theorem
      0 references
      strategy space reduction
      0 references
      size of the message space
      0 references
      informational requirement
      0 references
      social choice correspondence
      0 references
      dispersion of information
      0 references
      acceptable outcomes
      0 references
      implementability
      0 references
      lower contour sets
      0 references
      Nash equilibrium
      0 references
      monotonicity
      0 references
      no veto power
      0 references
      A social choice correspondence (SCC) F: E'''A assigns a socially acceptable set of alternatives \((F(e)\subset A)\) for a given environment \(e\in E\). Can the norms set by F be achieved when knowledge of e is dispersed among the n members of society \((I\equiv {\mathbb{N}}\) (mod n))? F is implemented via a game form \(g: \times_{i\in I}S_ i\to A:\) an alternative \(g(s)\in A\) is chosen on the basis of the members' reports \((s\in S\equiv \times_{i\in I}S_ i\); also \(s_{-i}\in S_{-i}\equiv \times_{j\in I\setminus \{i\}}S_ i)\). The dispersion of information allows a member to manipulate her report to induce an outcome which that member prefers to the acceptable outcomes. \textit{E. Maskin} [``Nash equilibrium and welfare optimality'', unpublished (1977)] gives a sufficient condition for the implementability of F. Let \(E\equiv \times_{i\in I}E_ i\) to simplify the exposition. Member i's preferences are given by the lower contour sets \[ L(a,e_ i) = \{b\in A| \text{ a is no worse than b when i's characteristic is }e_ i\}. \] Let \(g(S_ i,\bar s_{-i}\}\equiv \{g(s)|\) \(s_{-i}=\bar s_{- i}\}\), the alternatives i can induce by varying his message when the others' are fixed. A vector \(\bar s\in S\) is a Nash equilibrium for e \((\bar s\in N_ g(e))\) iff \(g(S_ i,\bar s_{-i})\subseteq L(g(\bar s),e_ i)\) for all i. The game form g implements F if and only if \(g(N_ g(e))=F(e)\) for all e. Two conditions which are imposed on the SCC F are monotonicity: \[ (\forall e\in E)(\forall e'\in E)(\forall a\in F(e))([(\forall i)L(a,e_ i)\subseteq L(a,e_ i')]\quad \Rightarrow [a\in F(e')]), \] and no veto power: \[ (\forall a\in A)(\forall e\in E)([\#\{i\in I| \quad L(a,e_ i)=A\}\geq n-1]\Rightarrow [a\in F(e)]). \] Theorem. If \({\#}I\geq 3\), \({\#}E>0\), \({\#}A>1\), and the SCC F satisfies monotonicity and no veto power, then F is implementable. Maskin's proof is constructive. The message space is \(S_ i\equiv \{(e,a)\in E\times A|\) \(a\in F(e)\}\) for each member i. This does not economize on information: \(S_ i\) is large, and its size increases with n. Saijo's contribution is to reduce i's message space to \(S_ i\equiv E_ i\times E_{i+1}\times A\times I\). Each member thus reports, not the entire environment, but only some ``local'' information: her own characteristic, her ``neighbor's'' characteristic, and the appropriate alternative. (The component of \(s_ i\) from I is used only out of equilibrium. It helps determine which of several disagreeing agents to ``punish'' for the disagreement.)
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references