CMStatistics 2021: Start Registration
View Submission - CFE
A1635
Title: Measuring model risk for market risk models Authors:  Emese Lazar - University of Reading (United Kingdom) [presenting]
Radu Tunaru - University of Sussex (United Kingdom)
Ning Zhang - Shanghai University (China)
Abstract: A scoring function-based model risk estimation methodology is proposed for measuring the joint model risk of Value-at-Risk and Expected Shortfall. We apply the proposed model risk measure across various market risk models and in our simulation study we find that our technique captures a large proportion of true model risk. We show that model risk is not always subadditive and when model risk is present, the ranking of market risk models is affected by the scoring function.