Sub-core solutions of the problem of strong implementation (Q1100984)

The problem of strong implementation is to determine which social choice correspondences (SCC) can be obtained as the strong equilibrium correspondence of a game form. We introduce the notion of the nucleus of an effectivity function. Under certain conditions, it yields the smallest implementable SCC having that effectivity function. We contrast it with the core, which yields the largest implementable SCC (as shown by \textit{H. Moulin} and \textit{B. Peleg} [J. Math. Econ. 10, 115-145 (1982; Zbl 0481.90004)] and argue that the smaller solution should be preferred when available. Another result is that the known necessary conditions for implementability are not sufficient, except in the case of (at most) 3 alternatives.