Job Market Paper:

Identification and Estimation in Semi-parametric Social Interaction Models

This paper investigates the identification and estimation of endogenous social interactions using a flexible semiparametric model to control for confounding factors. The rationale for considering nonparametric controls is that, if the groups or networks are not randomly assigned, or if the contextual effects are heterogeneous, identifying the endogenous social interaction effect is difficult without adequate controls. Based on the semiparametric model, the identification is attained by using the instrumental variable (IV) approach after partialling out the nonparametric controls. To estimate the endogenous social interaction effect, this paper proposes a semiparametric two-step GMM estimator with the optimal weight matrix clustered at the group or network level. For the semiparametric estimators that use the first step series method, this paper provides primitive regularity conditions for consistency and asymptotic normality. To detect severe nonlinearities and higher-order interactions, more flexible machine learning methods are also applied in the first step nonparametric estimation. The asymptotic properties of the semiparametric machine learning estimators are proved given high level conditions. Monte Carlo simulations are conducted to investigate the finite sample performance of semiparametric estimator using different first-step estimators, such as series or machine learning methods. The results suggest that no estimation method dominates across all the Data Generating Processes (DGPs) considered. However, applying the post-LASSO estimator in the first step is stable and performs relatively good across the settings considered. For this reason, the post-LASSO estimator is recommended for use in empirical studies.

Publications:

"Significance test in nonstationary logit panel model with serially correlated dependent variable", with Chia-Shang J. Chu and Lina Zhang. Economics Letters, Volume 159, 2017, PP. 37-41.

"Significance test in nonstationary multinomial logit model", with Chia-Shang J. Chu and Lina Zhang. Economics Letters, Volume 143, 2016, PP. 94-98.