Games and Decisions 2

Dividing Dynamic Decisions into Two Stages

speaker: Goreti Faria (LSE, London)

abstract: In orthodox models dynamic choice amounts to the static choice of an ex ante optimal strategy, using ex ante expected utility computed as in the static case. Such models might give a reasonable approximation to reality when consequences are limited to the immediate future, but they will fail to account for intuitively allowed rationalisations of behaviour when that is not the case. Therefore, alternatives to the orthodox fashion of modelling choices need to be explored. However, it is difficult for a model to account for all possible rationalisations of behaviour, and making use of the type of problem as a first step to deciding which model to use could be advantageous. Thus, in this paper I suggest the division of a dynamic decision into two stages: in the first the decision maker classifies the decision into a type according to her motivation, and in the second she chooses the model that best solves the type of problem in question. I focus on the case of a retirement decision and I apply this general strategy to it. I show how, if this decision is of the type where the agent only gets utility if she gets the best outcome, then maximising the probability of obtaining the best outcome will be preferable to applying the unqualified solution of maximising expected utility.

timetable:
Tue 8 Jul, 16:45 - 17:15, Sala Azzurra
<< Go back