科学研究

Yi He is an Associate Professor in the Quantitative Economic Section at the University of Amsterdam. He earned his master’s degree from the University of Cambridge and his PhD from Tilburg University in 2016. Prior to returning to the Netherlands, he served as a tenured Assistant Professor in the Department of Econometrics and Business Statistics at Monash University in Australia. His research focuses on high-dimensional econometrics, random matrix theory, extreme value statistics, bootstrapping, and machine learning. His work has been featured in prestigious journals, including the Journal of the American Statistical Association, The Annals of Statistics, Journal of the Royal Statistical Society - Series B, Journal of Business & Economic Statistics, and Journal of Econometrics. Yi's breakthroughs in extreme value statistics have earned him a nomination for the 2025 Van Dantzig Award in Statistics and Operations Research in the Netherlands. His current research explores dense time series models with complex network interactions in high-dimensional econometrics.

Measuring systemic risk has been an important topic in recent research on risk management. Adrian and Brunnermeier (2016, AER) pioneered the CoVaR measure for systemic risk, which is defined as the conditional quantile of system loss given some state predictors and one stress event of an individual loss at a certain risk level. They proposed to estimate CoVaR by a two-quantile-regression approach, which is simple and easy to implement in practice but without stating or developing the necessary asymptotic theory.  In this paper, we develop tests to reveal that using the classical regularity condition in quantile regressions for the two-quantile-regression model may be unrealistic by revisiting the datasets in Adrian and Brunnermeier (2016). Under the proposed weaker conditions, we further develop the asymptotic theory of the CoVaR inference and introduce a random weighted bootstrap method to quantify the inference uncertainty. We examine the finite sample performance of the proposed tests in terms of size and power, as well as the coverage probabilities of confidence intervals for CoVaR. Finally, we demonstrate the CoVaR inference across financial institutions and time.