Recursive utility and optimal growth with bounded or unbounded returns.

In this paper we propose a unifying approach to the study of recursive economic problems. Postulating an aggregator function as the fundamental expression of tastes, we explore conditions under which a utility function can be constructed. We also modify the usual dynamic programming arguments to include this class of models. We show that Bellman's equation still holds, so many results known for the additively separable case can be generalized for this general description of preferences. Our approach is general, allowing for both bounded and unbounded (above/below) returns. Many recursive economic models, including the standard examples studied in the literature, are particular cases of our setting.recursive utility, dynamic programming, Bellman equation, unbounded returns