Back to search

Boston University

Identification and testability in models of decision-making under uncertainty

Abstract

dc:description.abstract

This thesis studies identification and testability for three models of choice under uncertainty. Identification is concerned with whether the parameters of a specified model are uniquely recoverable from observable behavior. Testability is concerned with how to test the consistency between the model’s predictions and choice data. All models considered deviate from classic models in one of two ways: either randomness in preference varies with decision problems, or preferences violate expected utility theory. Chapter 1 studies a model of rational inattention. An agent can acquire costly information about uncertain states of the world before choosing an action from a menu. The choice of information depends on the menu of actions and is assumed unobservable. Due to the unobservability of private information, the choice of action appears random from an outside analyst’s perspective. I show that, given only stochastic choice from menus of actions, an analyst can identify the agent’s risk attitude, prior belief, and information cost function. Chapter 2 studies an instance of the classic question of whether preference can be identified from the choice behavior it implies. Specifically, the question is whether the distribution of risk preferences is identifiable from random choice of lotteries. It is known that the answer is affirmative under random expected utility. I show that such uniqueness fails if risk preferences are not restricted a priori to conform to expected utility, for instance, if they are assumed to conform only to weighted utility. I discuss the reason for such non-uniqueness and argue that uniqueness may be restored if data includes the joint distributions of choice across a limited number of feasible sets. Chapter 3 studies how to test non-expected utility theory given that data contain only finitely many observations on choice of lottery. While consistency of data with expected utility has been thoroughly studied in the literature, I derive conditions on data that are both necessary and sufficient for consistency with maximization of a betweenness preference. The conditions employ novel geometric arguments and provide a more stringent test than what is used in the experimental literature, which is to check for direct violations of the key axiom.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Economics
Grantor
Boston University
Year dc:date.issued
2019

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Lin, Yi-Hsuan
Advisor dc:contributor.advisor
  • Epstein, Larry G.

Subjects

dc:subject × 1

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/2144/41523

Chain of custody

source
Harvested from
Boston University
Base URL
open.bu.edu/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Lin, Yi-Hsuan. Identification and testability in models of decision-making under uncertainty. doctoral thesis, Boston University, 2019. https://hdl.handle.net/2144/41523