Back to search

Boston University

Nonparametric approaches to economic models with interactions

Abstract

dc:description.abstract

The dissertation is built on the central theme of using nonparametric approaches to economic models with interactions. The first chapter studies nonparametric estimation of treatment effects in the presence of network interference. The second chapter develops a nonparametric test for homophily in networks. The third chapter lays out a theoretical foundation for studying how altruistic preferences are affected by market competition and education, and also involves nonparametric methods. The first chapter studies nonparametric identification and estimation of treatment effects in networks. When individual outcomes depend on peer outcomes, treating an individual in a network affects all connected individuals. This causes the absence of a control group and threatens the validity of causal inference. Existing methods either assume linear functional forms and exogenous networks or exclude the dependence on peer outcomes. By introducing a nonparametric peer-effect model, I prove that the treatment effect is identified by comparing individuals who share the same neighbors but have different treatment status, without relying on the above assumptions. Estimation is performed by combining kernel smoothing, which relaxes the identical-neighbor condition, and the method of sieves. I establish the consistency of the proposed estimator and its convergence rate, which suggests that more complex sieve spaces need to be complemented by tighter bandwidths to control the bias from kernel smoothing. I apply my method to an anti-violence campaign and find that the campaign's effect on individual attitude increases with the average neighbor attitude. Policymakers intending to reduce campus violence may target students whose friends are also committed to anti-violent acts, or spread anti-violent messages among students and their peers simultaneously, to boost the effectiveness. The second chapter constructs a nonparametric test for homophily in network formation models. Homophily refers to the tendency of agents to connect with others of similar characteristics. Existing methods rely on specific distance measures and other functional-form assumptions to model homophily. Such choices are arbitrary and lack support from economic theory. I exploit the `location invariance' property of common distance measures and construct a nonparametric test for homophily without knowing the distance measure. Applying my method to the formation of risk-sharing networks, I find that individuals are concerned about differences in household consumption but not risk attitude. The third chapter lays out a theoretical foundation for studying how the altruistic preferences of future physicians are affected by market competition and education, using a laboratory experiment. Altruism is important in health care markets where asymmetric information and unmonitored actions are critical. Extensive education is required for health care professionals, yet they also face competition in real markets. In the planned experiment, subjects at different stages of their medical studies are asked to choose health care qualities for hypothetical patients under different market competition: monopoly, simultaneous-move duopoly, and Stackelberg duopoly. We construct a theoretical model of strategic interaction that characterizes altruism as the marginal rate of substitution between patient benefit and own profit. This provides a mapping from altruism and belief to observed quality choices in the experiment. Altruism is estimated by inverting this mapping. We compare altruism distributions nonparametrically using a bootstrapped Kolmogorov-Smirnov test to account for additional uncertainty induced by belief estimation. We also perform regression to estimate the effect on the mean of altruism distributions, while accounting for induced dependence via belief estimation. The validity of the proposed procedures is confirmed through simulations.

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
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Pan, Haoran
Advisor dc:contributor.advisor
  • Kaido, Hiroaki

Subjects

dc:subject × 1

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*
Handle dc:identifier.uri
https://hdl.handle.net/2144/53616

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
related terms
citation

Pan, Haoran. Nonparametric approaches to economic models with interactions. doctoral thesis, Boston University, 2026. https://hdl.handle.net/2144/53616