Tao Chen

"Time-Varying Jackknife Model Averaging for Quantile Regression"

with Zhongwen Liang

We propose a time-varying jackknife model averaging (TV-JMA) estimator for quantile regression that allows model averaging weights to adapt smoothly over time via local kernel estimation. A centering correction to the cross-validation criterion eliminates the bias from the non-zero conditional mean of the quantile regression error, enabling MSE-based weight selection. We establish convergence rates, asymptotic normality with diverging model dimensions, and asymptotic optimality of the selected weights under a scale-bias orthogonality condition that is automatically satisfied for median regression, pure location models, and heteroskedastic models with appropriately specified candidates. Monte Carlo simulations show that TV-JMA dominates constant-weight and constant-coefficient alternatives, with the gains increasing at tail quantiles and larger sample sizes. An empirical application to conditional quantile forecasting of excess stock returns and GDP growth-at-risk demonstrates the practical gains.

"Time-Varying Partially Linear Quantile Regression Under α-Mixing Conditions"

Solo-authored

This paper develops estimation and inference for partially linear quantile regression with time-varying coefficients under α-mixing dependence. The model accommodates covariates with smoothly evolving effects alongside covariates with time-invariant effects, embedding the framework within Dahlhaus's local stationarity paradigm. I propose a three-stage estimation procedure based on local linear quantile regression: initial local estimation treating all coefficients as time-varying, averaging to obtain preliminary constant coefficient estimates, and re-estimation using partial residuals. I establish a Bahadur representation for the local estimator under mixing, prove √T-consistency and oracle efficiency for the constant coefficient estimator, and derive uniform convergence rates for the time-varying coefficient functions. For inference, I develop a supremum-type test for coefficient constancy with a block bootstrap that resamples (Y, X, Z) jointly to preserve the mixing dependence structure. Monte Carlo simulations demonstrate excellent finite-sample size control across all quantile levels and error structures, with good power against various alternatives. An empirical application to U.S. equity returns reveals significant time variation in market beta at lower quantiles, with risk exposure amplified during economic downturns.

"Quantile Regression Approach to Study Risk Heterogeneity in PMI as a Business Cycle Indicator"

with Karan Bhasin and Kajal Lahiri, Under review, Journal of Applied Econometrics

In its monthly newsletter announcing the latest Purchasing Managers Index (PMI), the Institute of Supply Management (ISM) provides a breakeven value to suggest if the latest PMI value is indicative of continued growth or a recession. For instance, after being around 43.2 over last decade, this value has been revised to 47.5 since January 2026. Using U.S. real time macro data and partially linear time-varying quantile regression, we estimate PMI thresholds for predicting growth in Industrial Production (IP) and GDP growth. The median PMI thresholds for IP and GDP were found to be 49.5, and 43.5 respectively that varied very significantly across macroeconomic regimes since mid-80's. By estimating thresholds across conditional distributions of output growth using one-sided kernels, our cross-quantile variation quantifies the fundamental uncertainty in the PMI-output relationship and provides a principled framework for interpreting economic indicators. We provide fitted conditional quantile tables that translate any PMI reading into a probabilistic growth forecast, enabling practitioners to select thresholds appropriate for their risk tolerance, given all available information.

"Conformalized Time-Varying Quantile Regression: Distribution-Free Prediction Intervals for Locally Stationary Processes"

"Jackknife Model Averaging for Partially Linear Time-Varying Coefficient Quantile Regression"

"“Time-Varying Quantile Regression of Real GDP Growth on the Leading Economic Index and Its Components” (with Kajal Lahiri)."

"Community Nursing on Subjective Well-Being of the Elderly: Evidence from CLASS Data"

with Fulei Jin

BMC Nursing, 2025, 24(1), 757.