Quantile Partial Correlation for Growth Vulnerabilities
报告人： Hongqi Chen（Hunan University）
时间：2026-05-21 15:10-17:00
地点：Room 217, Guanghua Building 2
Abstract:
We extend the theory of Quantile Partial Correlation by Ma et al. (2017) to high-dimensional time-series settings by establishing selection consistency of Quantile Partial Correlation Regression (QPCR) under dependent, heteroskedastic, and potentially unbounded processes. We introduce an EBIC-type tuning criterion that recovers the true quantile model with high probability. Applying QPCR to predict monthly industrial production growth using a large number of macrofinancial predictors, we find that labour-market and financial conditions are the primary drivers of downside growth risk, while there is little predictive content for upside risk. Decomposing downside risk into its individual components, we construct sector-specific indices that predict it, while controlling for information from other sectors, thereby isolating the downside risks emanating from each sector.
About the Speaker:
陈弘琦现任湖南大学金融与统计学院助理教授，博士毕业于伊利诺伊大学厄巴纳香槟分校。研究方向主要包括计量理论、应用计量经济学与统计学，重点关注高维方法与分位数回归等问题。相关成果发表于Journal of Statistical Planning and Inference。
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
