This paper considers economies involving one public good, one private good, and convex technology and proposes an informationally decentralized dynamic nontatonnement procedure that converges in general from the initial endowments to an allocation in the core. The procedure may be seen as enunciating a plausible method of cooperation among the agents for achieving an optimal provision of a public good and an equitable sharing of its cost. The viewpoint of noncooperative game theory is also considered and it is shown that there exists a trade-off between the requirements of local incentive compatibility and equitable cost sharing.

[This paper develops a framework of analysis for studying the informational properties of a certain class of "parametric" resource allocation processes. It is shown that the Taylor process (related to certain ideas for planning in the so-called socialist economies as put forward by Taylor [126] is informationally efficient in the sense that any informationally decentralized resource allocation process which has similar (static) properties (Pareto optimality) must use a message space which is dimensionally at least as large as that of the Taylor process. We also show that in general greater informational decentralization can be achieved through parametric than through "nonparametric" processes.]

[This paper describes a decentralized planning process which is a formalization of an iterative method proposed by Taylor [12] for the construction of an optimal plan in a socialist state. Further, it shows that this process is computationally and informationally more efficient than an alternative process proposed by Malinvaud [10] also as a formalization of Taylor's method.]