Research 

Grants and Awards

National Natural Science Foundation of China under Grant No.72203114 (国家自然科学基金青年项目), 2023.1-2025.12 

Fundamental Research Funds for the Central Universities under Grant No. 63212145, 2021.1-2023.12 

Dean’s Graduate Research Excellence Award of Monash Business School, 2016.2 - 2019.2.

Publications

Binary Response Models for Heterogeneous Panel Data with Interactive Fixed Effects 

Abstract: In this paper, we investigate binary response models for heterogeneous panel data with interactive fixed effects by allowing both the cross-sectional dimension and the temporal dimension to diverge. From a practical point of view, the proposed framework can be applied to predict the probability of corporate failure, conduct credit rating analysis, etc. Theoretically and methodologically, we build a link between a maximum likelihood estimation and a least squares approach, provide a simple information criterion to detect the number of factors, and establish the corresponding asymptotic theory. In addition, we conduct intensive simulations to examine the theoretical findings. In an empirical study, we focus on the sign prediction of stock returns and then use the results of sign forecast to conduct portfolio analysis. 

Nonparametric Time-Varying Panel Data Models with Heterogeneity

Abstract: Since Bai (2009), considerable extensions have been made to panel data models with interactive fixed effects (IFEs). However, little work has been conducted to understand the associated iterative algorithm, which,  to the best of our knowledge,  is the most commonly adopted approach in this line of research. In this paper, we refine the algorithm of panel data models with IFEs using the nuclear-norm penalization method and duple least squares (DLS) iterations. Meanwhile, we allow the regression coefficients to be individual-specific and evolve over time. Accordingly, asymptotic properties are established to demonstrate the theoretical validity of the proposed approach. Furthermore, we show that the proposed methodology exhibits good finite sample performance using simulation and real data examples.

Tail Connectedness: Measuring the Network Connectedness of Equity Markets During Crises

Abstract: This paper studies the global volatility connectedness network among 16 stock markets under different market conditions. We construct measures of tail connectedness following Ando et al. (2022) by introducing quantile regression into the classic Diebold-Yilmaz network model. We demonstrate the advantages of using tail connectedness for measuring extreme systemic risk and examine the dynamic evolution of volatility connectedness from 2005 to 2021 at different quantiles. Our empirical results suggest that when the market is calm, the strength of volatility connectedness is determined by the closeness of economic and trade ties. Although Euro-American countries tend to act as net risk providers during crises, Asian markets have become increasingly influential in the past two decades. We also find that the spillover of extreme risks occurs layer by layer, with either the U.S. or China sitting at the centre of the spillover network and transmitting risks to the regional centres and peripheral markets.

Estimation and Inference for a Semiparametric Time-Varying Panel Data Model

Abstract: This paper introduces a new semiparametric panel data model that accounts for time-varying coefficients and aligns with recent advancements in factor models featuring nonparametric loading functions. We propose a profile marginal integration (PMI) method to jointly estimate the unknown quantities in a series of easily implementable steps. The asymptotic properties of these estimators are established. Additionally, we provide a hypothesis test to assess the validity of parametric model specifications in applied settings. Simulation studies and an empirical application on U.S. mutual fund returns are conducted to evaluate the finite sample performance of the proposed method. The empirical results suggest that traditional parametric methods, which ignore time variation, may lead to invalid inference.