Some recent developments on model uncertainty in risk measurement
主 题: Some recent developments on model uncertainty in risk measurement
报告人: Dr. Ruodu Wang (University of Waterloo)
时 间: 2016-04-14 14:00 - 15:00
地 点: 理科一号楼 1560
In this talk we discuss some recent issues on model uncertainty in risk measurement. The two regulatory risk measures, the Value-at-Risk (VaR) and the Expected Shortfall (ES) enjoy different properties regarding model uncertainty. In terms of the classic notion of robustness, VaR has an advantage over ES. We bring in some new perspectives into discussion. In particular, we show that ES is robust with respect to dependence modeling in risk aggregation, VaR has a larger dependence uncertainty spread, and VaR suffers from serious issues in the optimization of risks. The content in talk is complemented by a second part “quantile-based risk sharing”.