서브메뉴
검색
Essays on Causal Inference and Network Econometrics
Essays on Causal Inference and Network Econometrics
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202103557
- ISBN
- 9798288865015
- DDC
- 302
- 저자명
- Gao, Mengsi.
- 서명/저자
- Essays on Causal Inference and Network Econometrics
- 발행사항
- [Sl] : University of California, Berkeley, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 190 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-01, Section: A.
- 주기사항
- Advisor: Graham, Bryan S.;Ding, Peng.
- 학위논문주기
- Thesis (Ph.D.)--University of California, Berkeley, 2025.
- 초록/해제
- 요약This dissertation includes three chapters, each developing methods for estimating causal effects in network experiments, a setting characterized by interference among units. In the first chapter, I investigate the identification and inference of treatment effects in randomized controlled trials with social interactions. Two key network features introduce endogeneity: (1) latent variables influencing both network formation and outcomes, and (2) treatment-induced changes to the network structure that mediate treatment effects. I first define parameters in a post-treatment network framework, distinguishing direct effects from indirect effects mediated by network changes, and provide a causal interpretation of coefficients in a linear outcome model. To address endogeneity, I propose a shift-share instrument variable strategy and establish consistency and asymptotic normality of the IV estimator in relatively sparse networks. For denser networks, I introduce a denoised SSIV estimator based on eigendecomposition to restore consistency. Finally, I revisit Prina (2015) as an empirical illustration, showing that treatment can influence outcomes both directly and through network structure changes. In the second chapter, coauthored with Peng Ding, we study the spillover effects using regression-based estimators under the design-based framework. Network experiments are powerful tools for studying spillover effects, which avoids endogeneity by randomly assigning treatments to units over networks. However, it is non-trivial to analyze network experiments properly without imposing strong modeling assumptions. We show that regression-based point estimators and standard errors can have strong theoretical guarantees if the regression functions and robust standard errors are carefully specified to accommodate the interference patterns under network experiments. We first recall a well-known result that the Hajek estimator is numerically identical to the coefficient from the weighted-least-squares fit based on the inverse probability of the exposure mapping. Moreover, we demonstrate that the regression-based approach offers three notable advantages: its ease of implementation, the ability to derive standard errors through the same regression fit, and the capacity to integrate covariates into the analysis to improve efficiency. Recognizing that the regression-based network-robust covariance estimator can be anti-conservative under nonconstant effects, we propose an adjusted covariance estimator to improve the empirical coverage rates.In the third chapter, I investigate the Hausman test for network endogeneity by revisiting the linear-in-means framework in Chapter 1, which restricts the attention to models without endogenous peer effects. First, I establish the consistency and asymptotic normality of the OLS estimator when there is no endogeneity and show that its estimator for the average peer effect may converge at a rate slower than √n when variation in the peer-fraction regressor vanishes. I then develop a Hausman test for network exogeneity by contrasting OLS and IV estimators. Simulation results demonstrate that the test maintains correct size and delivers strong power.
- 기타저자
- University of California, Berkeley Economics
- 기본자료저록
- Dissertations Abstracts International. 87-01A.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017357764
■00520260202103557
■006m o d
■007cr#unu||||||||
■020 ▼a9798288865015
■035 ▼a(MiAaPQ)AAI32042102
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a302
■1001 ▼aGao, Mengsi.
■24510▼aEssays on Causal Inference and Network Econometrics
■260 ▼a[Sl]▼bUniversity of California, Berkeley▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a190 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-01, Section: A.
■500 ▼aAdvisor: Graham, Bryan S.;Ding, Peng.
■5021 ▼aThesis (Ph.D.)--University of California, Berkeley, 2025.
■520 ▼aThis dissertation includes three chapters, each developing methods for estimating causal effects in network experiments, a setting characterized by interference among units. In the first chapter, I investigate the identification and inference of treatment effects in randomized controlled trials with social interactions. Two key network features introduce endogeneity: (1) latent variables influencing both network formation and outcomes, and (2) treatment-induced changes to the network structure that mediate treatment effects. I first define parameters in a post-treatment network framework, distinguishing direct effects from indirect effects mediated by network changes, and provide a causal interpretation of coefficients in a linear outcome model. To address endogeneity, I propose a shift-share instrument variable strategy and establish consistency and asymptotic normality of the IV estimator in relatively sparse networks. For denser networks, I introduce a denoised SSIV estimator based on eigendecomposition to restore consistency. Finally, I revisit Prina (2015) as an empirical illustration, showing that treatment can influence outcomes both directly and through network structure changes. In the second chapter, coauthored with Peng Ding, we study the spillover effects using regression-based estimators under the design-based framework. Network experiments are powerful tools for studying spillover effects, which avoids endogeneity by randomly assigning treatments to units over networks. However, it is non-trivial to analyze network experiments properly without imposing strong modeling assumptions. We show that regression-based point estimators and standard errors can have strong theoretical guarantees if the regression functions and robust standard errors are carefully specified to accommodate the interference patterns under network experiments. We first recall a well-known result that the Hajek estimator is numerically identical to the coefficient from the weighted-least-squares fit based on the inverse probability of the exposure mapping. Moreover, we demonstrate that the regression-based approach offers three notable advantages: its ease of implementation, the ability to derive standard errors through the same regression fit, and the capacity to integrate covariates into the analysis to improve efficiency. Recognizing that the regression-based network-robust covariance estimator can be anti-conservative under nonconstant effects, we propose an adjusted covariance estimator to improve the empirical coverage rates.In the third chapter, I investigate the Hausman test for network endogeneity by revisiting the linear-in-means framework in Chapter 1, which restricts the attention to models without endogenous peer effects. First, I establish the consistency and asymptotic normality of the OLS estimator when there is no endogeneity and show that its estimator for the average peer effect may converge at a rate slower than √n when variation in the peer-fraction regressor vanishes. I then develop a Hausman test for network exogeneity by contrasting OLS and IV estimators. Simulation results demonstrate that the test maintains correct size and delivers strong power.
■590 ▼aSchool code: 0028.
■653 ▼aSocial interactions
■653 ▼aNetwork structure
■653 ▼aAsymptotic normality
■690 ▼a0501
■690 ▼a0511
■71020▼aUniversity of California, Berkeley▼bEconomics.
■7730 ▼tDissertations Abstracts International▼g87-01A.
■790 ▼a0028
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17357764▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
