Quantile random-coefficient regression with interactive fixed effects: heterogeneous group-level policy evaluation, joint with Jiti Gao, Tatsushi Oka and Yoon-Jae Whang. Econometric Reviews, 2024, 1-19.
Abstract: We study the estimation of heterogeneous effects of group-level policies, using quantile regression with interactive fixed effects. Our approach can identify distributional policy effects, particularly effects on inequality, under a type of difference-in-differences assumption. We provide asymptotic properties of our estimators and an inferential method. We apply the model to evaluate the effect of the minimum wage policy on earnings between 1967 and 1980 in the United States. Our results suggest that the minimum wage policy has a significant negative impact on the between-inequality but little effect on the within-inequality.
Multi-population modelling and forecasting life-table death counts, joint with Han Lin Shang and Steven Haberman. Insurance: Mathematics and Economics, Volume 106, 2022, 239-253.
Abstract: When modelling the age distribution of death counts for multiple populations, we should consider three features: (1) how to incorporate any possible correlation among multiple populations to improve point and interval forecast accuracy through multi-population joint modelling; (2) how to forecast age distribution of death counts so that the forecasts are non-negative and have a constrained integral; (3) how to construct a prediction interval that is well-calibrated in terms of coverage. Within the framework of compositional data analysis, we apply a log-ratio transform to transform a constrained space into an unconstrained space. We apply multivariate and multilevel functional time series methods to forecast period life-table death counts in the unconstrained space. Through the inverse log-ratio transformation, the forecast period life-table death counts are obtained. Using the age-specific period life-table death counts in England and Wales and Sweden obtained from the Human Mortality Database (2022), we investigate one-step-ahead to 30-step-ahead point and interval forecast accuracies of the proposed models and make our recommendations.
Change point detection for COVID-19 excess deaths in Belgium, joint with Han Lin Shang. Journal of Population Research, Volume 39, 2022, 557-565.
Abstract: Emerging at the end of 2019, COVID-19 has become a public health threat to people worldwide. Apart from deaths with a positive COVID-19 test, many others have died from causes indirectly related to COVID-19. Therefore, the COVID-19 confirmed deaths underestimate the influence of the pandemic on society; instead, the measure of ‘excess deaths’ is a more objective and comparable way to assess the scale of the epidemic and formulate lessons. One common practical issue in analysing the impact of COVID-19 is to determine the ‘pre-COVID-19′ period and the ‘post-COVID-19′ period. We apply a change point detection method to identify any change points using excess deaths in Belgium.
Single-Index Quantile Factor Model with Observed Characteristics, joint with Qingliang Fan
Abstract: We propose a characteristics-augmented quantile factor (QCF) model, where unknown factor loading functions are linked to a large set of observed individual-level (e.g., bond- or stock-specific) covariates via a single-index projection. The single-index specification offers a parsimonious, interpretable, and statistically efficient way to nonparametrically characterize the time-varying loadings, while avoiding the curse of dimensionality in flexible nonparametric models. Using a three-step sieve estimation procedure, the QCF model demonstrates high in-sample and out-of-sample accuracy in simulations. We establish asymptotic properties for estimators of the latent factor, loading functions, and index parameters. In an empirical study, we analyze the dynamic distributional structure of U.S. corporate bond returns from 2003 to 2020. Our method outperforms the benchmark quantile Fama-French five-factor model and quantile latent factor model, particularly in the tails. The model reveals state-dependent risk exposures driven by characteristics such as bond and equity volatility, coupon, and spread. Finally, we provide economic interpretations of the latent factors.
Abstract: This paper presents a panel quantile functional-coefficient regression framework with a latent factor structure. The model captures the time-varying quantile co-movement in large financial time series and improves Value-at-Risk (VaR) forecasts. Utilizing an iterative estimation approach, the model demonstrates high accuracy in both in-sample and out-of-sample contexts using extensive simulations. Asymptotic properties are established through an approximation of generalized sequences. The model is applied to U.S. financial institutions' weekly return data from 1993 to 2013. I show that the proposed model outperforms several established VaR forecasting methods due to (i) capturing the time-variation in the risk exposure and (ii) effectively eliminating serial correlation in residuals with the latent factors.