Files in this item

FilesDescriptionFormat

application/pdf

application/pdfFUNG-DISSERTATION-2016.pdf (962kB)
(no description provided)PDF

Description

Title:Three essays in applied market design
Author(s):Fung, Juan Francisco
Director of Research:Williams, Steven R
Doctoral Committee Chair(s):Williams, Steven R
Doctoral Committee Member(s):Bernhardt, Mark D; McMillen, Daniel P; Perry, Martin K
Department / Program:Economics
Discipline:Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Market design, matching
Abstract:The market design approach to economics recognizes that markets do not arise naturally but are rather an amalgamation of various rules and norms. From this perspective, an economist can reverse engineer the rules that are consequential to a functioning market and evaluate the effects of those rules on market outcomes, with an eye toward potentially re-engineering certain components to achieve some objective. In this thesis, I present three market settings inspired by real-world applications, viewed broadly from the lens of market design. The first chapter, joint work with Chia-Ling Hsu, explores the consequences of various market details on equilibrium outcomes. Specifically, we consider a situation in which a matching problem between two sets of agents is solved by a platform serving as an intermediary. For instance, an artist who wants to find donors and a backer who wants to find artists to support can find each other on a crowdfunding platform like Kickstarter. Existing models of platform markets restrict agent heterogeneity and so the matching problem is secondary. However, it is possible that different artists target different types of backers and even likely that backers differ in their preferences for artists. In this chapter, we introduce agent heterogeneity by proposing a matching model of platform markets. In such markets, stability eliminates the possibility of an individual or group of agents switching in equilibrium, thus ensuring successful coordination. The model allows exploration into the properties of equilibrium with heterogeneous agents, offering a new approach to studying platform markets. In the second chapter, I empirically quantify the value of public school choice. Traditionally, public school assignment is determined by a family’s residence in the district. An alternative policy is to allow families to apply to any school in the district. Such school choice programs provide families with more options, but it is unclear how much families value these options over ii having a guaranteed school. In this chapter, I exploit a natural experiment in Champaign-Urbana, IL: in 1998, Champaign school district adopted school choice while the neighboring district of Urbana did not. Using variation in housing prices in each district, before and after the policy change, I estimate the marginal willingness to pay for school choice relative to residence-based assignment. I find that, on average, households are willing to pay between 5- 7% more for school choice relative to residence-based assignment. The results are robust to regularization and alternative model specifications. The third chapter, joint work with Blake Riley, is motivated by decentralized matching: the process by which agents find matches on their own. We show that, without revealing information to a centralized matchmaker and without coordination, agents can find stable matches on their own. Existing work on uncoordinated matching, based on the random better reply dynamics of Roth and Vande Vate (1990), shows that agents do find stable matches but that in the worst case it could take exponentially long. We introduce a new process that, in various numerical experiments, appears to converge in polynomial time. The key to our proposal process is mitigating a major bottleneck in uncoordinated matching: the possibility that an agent is single for a very long time before finding a match. In the worst case, our process converges in O(n^3 ) time in moderate sized balanced markets with n agents on each side. We also consider unblanaced markets, in which there are more agents on one side of the market. While convergence to stability is not guaranteed in polynomial time, we show numerically that typical outcomes of our proposal process are more egalitarian than stable outcomes. This chapter thus sheds some light on the value of centralizing a matching market, as opposed to allowing the market to clear on its own. The common thread in all three chapters is that, while markets should not be taken as given, it is important to evaluate the relative importance of particular design elements. The first chapter characterizes equilibrium outcomes under various designs; the second considers the relative value of two particular designs; and the third questions the value of designing at all. In the spirit of market design, each application is driven by actual markets and a variety of methodologies.
Issue Date:2016-07-11
Type:Thesis
URI:http://hdl.handle.net/2142/92785
Rights Information:Copyright 2016 Juan F. Fung
Date Available in IDEALS:2016-11-10
Date Deposited:2016-08


This item appears in the following Collection(s)

Item Statistics