Research

Abstract:  This paper investigates peer effects in discrete choice models with incomplete data on social links.  Following Graham (2017), I set up an undirected dyadic link formation model where links are based on homophily (similarities in characteristics) and individual fixed effects.  Homophily effects are identified through available configurations among tetrads (groups of four agents).  I then identify the fixed effects as unknown parameters through available configurations among triads (groups of three agents).  I propose an estimation strategy based on the estimated link formation probabilities to study peer effects on individual decision-making, and establish its large sample properties.  Simulations illustrate that the finite sample performance of the estimator is close to that obtained when the true network were known. Using data on household microfinance participation in rural India from Banerjee et al. (2013) where network data are available, I detect positive peer effects even with partial network data.

Presentations: 2025 World Congress of the Econometric Society (scheduled), Chinese Economists Society (CES) North America Annual Conference 2025, American Economic Association Annual Meeting 2025, Global GLO-JOPE Conference 2024 (Job Market Session), The North East Universities Development Consortium (NEUDC) 2024 Conference, Asian Meeting of the Econometric Society 2024,  North American Summer Meeting of the Econometric Society 2024 [Certificate], TEXPOP Conference 2024 by the Population Research Center of the University of Texas at Austin, Texas Camp Econometrics XXVII 2024, Southern Economic Association 93rd Annual Meeting 2023, Causal Data Science Meeting 2023, 18th Annual Economics Graduate Student Conference of Washington University in St. Louis,  33rd Annual Midwest Econometrics Group Conference,  SMU Brown Bag Seminar, 2nd Annual SMU Research Computing Day 2023 (Poster)

Award:  Dean's Dissertation Fellowship 2024, SMU

Abstract: We study the identification and estimation of social interactions in large endogenous networks. Our analysis focuses on binary-action games of incomplete information in which an agent's expected payoff relies on her characteristics, peers' average characteristics, the average of her beliefs about peers' actions, and some preference shocks. Endogeneity in networks results from unobserved characteristics that affect both link formation and individual decision-making. To identify the utility functions, we express the unobserved characteristics as some unknown function of observed variables and address the endogeneity issue through a control function approach. The identification strategy holds even if multiple equilibria exist. We employ the strategy to develop a semiparametric estimator and assess finite-sample performance through Monte Carlo simulations. 

Presentations: Texas Camp Econometrics XXVIII, Southern Economic Association 94th Annual Meeting 2024, 34th Annual Midwest Econometrics Group Conference 2024

Abstract: This paper documents the negative influence of overly ambitious parental aspirations on children's academic outcomes, utilizing data from the Education Longitudinal Study 2002. After controlling for family socioeconomic status, I find that parents with better education are less likely to form overly ambitious aspirations, which represents a new channel to account for the positive effects of parental education on children's academic performances. To address the possibility that parental aspirations are endogenous, I construct instruments based on children's social network. For robustness, I use an alternative approach based on bounding the effect of parental aspirations under weaker assumptions. The results are consistent across both approaches.

Presentations: TEXPOP Conference 2025 by the Population Research Center of the University of Texas at Austin (scheduled), Stata Texas Empirical Microeconomics Conference 2024, Poster, Dallas Fed 

Abstract:  Estimating treatment effects is one of the most fundamental problems in econometric studies.  However, estimation is complicated when spillovers across individuals are present.  In this research, we focus on the spillovers in the treatment assignments and study the resulting Bayesian Nash equilibrium of a binary participation game, also known as peer-influenced propensity scores. In addition, we allow for the case when the network is unobservable. Under network sparsity, the estimation problem is formulated as a penalized maximum likelihood estimation, where the missing binary links are treated as high-dimensional nuisance parameters. Simulation studies demonstrate the performance of our treatment effects estimators in finite samples: our approach has good performance for small networks.

Presentations: American Economic Association Annual Meeting 2024 (Poster),  1st CIREQ Interdisciplinary PhD Student Conference on Big Data and Artificial Intelligence at McGill University,  Midwest Economics Association 87th Annual Meetings 2023, Texas Camp Econometrics XXVI,  3rd Annual SMU Economics PhD Alumni Conference, Southern Economic Association 92nd Annual Meeting 2022, SMU Research & Innovation Week Graduate Poster Session 2022, SMU Microeconomics Workshop 2021

Awards:  Cobb Fellowship (Best Third-Year Paper Prize) [link]; Dean's Award, SMU Research & Innovation Week Graduate Poster Session 2022 [link]

Abstract:  Social network data collection often emphasizes missing data on networks rather than characteristics, leading to situations where we have abundant social network data, but missing covariates can also cause estimation bias. This paper addresses the question by formulating it as a high-dimensional estimation problem. We use machine learning techniques to predict the conditional distribution of missing covariates and recover the peer effects in a partially linear model.