Nonpaternalistic intergenerational altruism (Q1092782)

The paper develops a concept of equilibrium behaviour and establishes its existence in a model of nonpaternalistic intergenerational altruism. Each generation derives utility from its own consumption and the utilities of its successors. Equilibrium capital stocks are time-monotone and therefore converge to a steady state. This paper conducts a direct study of the behavioural foundations of the nonpaternalistic model. The first line of business is the definition of an equilibrium outcome, and a demonstration of its existence in a general setting. This is one of the main tasks of the paper. The notion of an equilibrium outcome is adapted in an obvious way from the concept of subgame perfect Nash equilibria in extensive-form games. The author states and proves an existence theorem for a class of models that exhibits an extremely general form of nonpaternalistic altruism. In particular, each generation's utility may depend on those of any number of future generations. The existence result asserts the following, in the context of a stationary model. There exist an indirect utility function and a savings policy, both depending on current endowment, such that each generation finds it optimal to adopt that savings policy provided its descendants use the same policy and exhibit the given indirect utility. In addition, the indirect utility function generated by the generation's maximization problem is also the same as that ``announced'' by its descendants. Furthermore, the author describes a class of cases where equilibrium programs are inefficient in the sense that more consumption can be feasibly provided at every date. The equilibrium path can therefore display efficiency failures of a drastic sort. Such failures, however, are only possible if altruism extends to the utilities of descendants other than one's immediate successor.