At the invitation of School of Mathematical Sciences, Research Associate Shiqi Ye from Chinese Academy of Sciences, gave a lecture entitled “ Double Pooling for Dynamic Tail Estimation” at Academic Activity Room 526 of Xingjian Building on the morning of April 29, 2026. The seminar was chaired by Dr. Hongfang Sun. Faculty members and students from the Statistical and Financial Mathematics Research Group attended the lecture.Shiqi Ye is a Assistant Research Fellow at the Center for Forecasting Science, Academy of Mathematics and Systems Science, Chinese Academy of Sciences. His research focuses on high-dimensional complex system modeling, macroeconomic risk management, and big data methodsand applications.At the seminar, Shiqi Ye first introduced the background of the study. Effectively estimating and managing tail risk is an important topic in macroeconomics and finance. Tail risk generally refers to potential losses arising from extreme market fluctuations and low-probability events. Commonly used measures such as Value-at-Risk (VaR) and Expected Shortfall (ES) are often difficult to capture in panel data due to substantial cross-country heterogeneity and the rarity of extreme events; traditional approaches struggle to simultaneously accommodate the dynamics at multiple quantile levels and across multiple cross-sectional units. To address this issue, Ye presented the newly proposed Double Pooling Dynamic Tail Estimation (DPDTE) framework. The core breakthrough of the DPDTE framework lies in constructing a unified loss function that pools information in two directions simultaneously: across cross-sectional units (countries or sectors) and across quantile levels. This “double pooling” design considerably improves the efficiency of tail risk estimation, allowing multiple-levels VaRs and the extreme ES to be recovered jointly, while yielding more stable estimates. For the dynamic specification, the framework integrates a score-driven model with extreme value theory, enabling tail dynamics to respond promptly to newextreme observations while maintaining reasonable constraints on the shape of the tail. The presenter demonstrated the consistency and asymptotic normality of the estimators, and simulation experiments confirmed favorable finite-sample performance. In the application, DPDTE was employed to estimate Growth-at-Risk (GaR) and extreme ES for 14 European economies from 1985 to 2025, uncovering the long-run evolution of macroeconomic tail risk and its cross-countrydifferences. Ye emphasized that the value of the DPDTE framework lies not only in its theoretical innovation but also in its natural adaptability to “ragged-edge” data structures and sparse tail information. After the presentation, the attending faculty members engaged in a thorough discussion with Research Associate Shiqi Ye on the details of the DPDTE framework. The presentation sparkedlively debate among the participating scholars on frontier issues such as tail risk modeling and cross-economy risk spillovers, and opened up new directions for macro-tail risk research.