Research
My research interests are Econometrics, Statistics, Financial Modeling, Computational Finance, Machine Learning, and Deep Learning. Following are my current research projects.
Abstract: This paper introduces a censored-outcome simultaneous-equation model with social interactions. The construction of the microeconomics foundation for this model is from the equilibrium in a large-network-based game with incomplete information, in which each agent conducts multiple actions and interacts with other agents through a network and a linear quadratic utility function. The sufficient condition of the unique Bayesian Nash Equilibrium (BNE) existence is characterized. We also discuss the identification of the econometric model. We propose a two-stage method to estimate the model in which we apply the nested pseudo-likelihood (NPL) to estimate the reduced parameters and then derive the structural form parameters by Amemiya Generalized Least Square estimator (AGLS). Monte Carlo simulation shows that the estimation performs well in finite samples. The estimation also shows the feasibility of the computation when the network size is large.
A Spatial Autoregressive Model with Endogenous Network Structure and Limited Dependent Variable (Work in progress)
Abstract: Spatial Autoregressive model (SAR) has been widely discussed in recent decades. This paper focuses on the SAR model with limited dependent variables (Probit and Tobit) and rational expectations when the network structure is endogenous.
Simultaneous Tobit Model with Social Interactions (Work in progress)
Abstract: This paper proposes a simultaneous Tobit model with social interactions. Derive the Bayesian Nash Equilibrium (BNE) and find the sufficient condition for the unique fixed point existence. We develop the identification and a nested pseudo-likelihood (NPL) estimation of the econometric model.