INFORMATION-PROCESSING IN DYNAMIC DECISION-MODELS - AN INSURANCE DEMAND EXAMPLE

This paper deals with a multiperiod model of insurance demand in which losses are not assumed to be stochastically independent over time. The expected utility of end-of-period wealth is to be maximized; we assume constant relative risk aversion. Additional information during the planning process is part of the optimal decision. Structural properties of optimal insurance strategies (e.g., myopicity) are derived. For the special case of a Bayesian learning process we give sensitivity rules. © 1991.