Charles F. Manski, Eytan Sheshinski
arXiv 17 Oct 2024 · Econometrics
arXiv:2410.13658 · PDF · DOI · OpenAlex
We study optimal policy when a paternalistic utilitarian planner has the power to design a discrete choice set for a heterogeneous population with bounded rationality. We show that the policy that most effectively constrains or influences choices depends in a particular multiplicative way on the preferences of the population and on the choice probabilities conditional on preferences that measure the suboptimality of behavior. We first consider the planning problem in abstraction. We then study two settings in which the planner may mandate an action or decentralize decision making. In one setting, we suppose that individuals measure utility with additive random error and maximize mismeasured rather than actual utility. Then optimal planning requires knowledge of the distribution of measurement errors. In the second setting, we consider binary treatment choice under uncertainty when the planner can mandate a treatment conditional on publicly observed personal covariates or can enable individuals to choose their own treatments conditional on private information. We focus on situations where bounded rationality takes the form of deviations between subjective personal beliefs and objective probabilities of uncertain outcomes. To illustrate, we consider clinical decision making in medicine. In toto, our analysis is cautionary. It characterizes the subtle nature of optimal policy, whose determination requires the planner to possess extensive knowledge that is rarely available. We conclude that studies of policy choice by a paternalistic utilitarian planner should view not only the population but also the planner to be boundedly rational.