Abstract

This dissertation studies topics in experimental economics and microeconomic theory. In the first chapter, I study the relationship matching problem through theoretical analysis and lab experiments. In particular, I consider a two-stage matching model where agents can only divide a fraction of match surplus when forming a match, with the remaining subject to bargain within the match. I characterize stable matching outcomes under the assumption of the standard Nash bargaining solution within the match. Two special cases are Bargaining Before Match (BBM), where agents divide the entire surplus before forming a match, and Bargaining After Match (BAM), where the entire surplus is divided after a match is formed. While matching experiments in BAM produce more stable matches than in BBM, our experiments show the opposite in the non-extreme cases. I explain the experiment outcomes with the quantal response equilibrium. In the second chapter, I investigate when and why observing actions of others can either facilitate or hinder decision-making in a sequential social learning environment where all information is public. I conduct a laboratory experiment in which participants receive the same total information but differ in whether they observe others’ actions or raw signals first. Once a herd forms, seeing actions before signals significantly reduces the likelihood of making an optimal decision, and providing the signals afterward does little to remedy this effect. Analysis of time spent on signals and players’ reliance on herding actions suggests that initially seeing actions steers attention away from subsequently revealed signals, thus amplifying errors. In the third chapter, we study voters’ endogenous acquisition of costly signals in a setting where pivotality and state magnitude exert opposing effects on incentives. We consider a continuous state space, where signals and the stakes of the decision to vary with the underlying state. When the state has a large magnitude, signals tend to align, reducing the chance of an individual being pivotal but increasing the stakes of the decision. In contrast, states near zero generate conflicting signals, increasing pivotality while lowering the stakes. We show that when marginal costs of nearly irrelevant information are zero and exhibit diminishing returns in precision, the probability that the electorate selects the correct alternative converges to a state-dependent value strictly above one-half. If the marginal cost of acquiring initial information is nearly flat, the probability of a correct decision converges to one.

Committee Chair

SangMok Lee

Committee Members

Brian Rogers; John Nachbar; Jonathan Weinstein; Mariagiovanna Baccara

Degree

Doctor of Philosophy (PhD)

Author's Department

Economics

Author's School

Graduate School of Arts and Sciences

Document Type

Dissertation

Date of Award

4-30-2025

Language

English (en)

Download

Available for download on Monday, May 12, 2031

Included in

Economics Commons

Share

COinS