CMStatistics 2017: Start Registration
View Submission - CFE
A1287
Title: A factor-model approach to correlation stress testing Authors:  Natalie Packham - Berlin School of Economics and Law (Germany) [presenting]
Fabian Woebbeking - Goethe University Frankfurt (Germany)
Abstract: A factor-model is developed for parameterising correlation matrices of financial portfolios. The factor-model structure allows to understand various drivers of correlation amongst portfolio constituents. The approach can be used to translate economic scenarios into constraints and changes on the dependence structure allowing to measure the impact of specific scenarios on portfolio risk. In a detailed case study, we apply the factor-model approach to the credit derivatives trading strategy by JP Morgan Chase, dubbed the London Whale, that led to losses in the magnitude of 6.2 bln USD in 2012. By choosing the factors according to the stated investment goals of the strategy (e.g. hedging investment grade credit derivatives with high-yield credit derivatives), we derive insights into the specific correlation risks inherent in the trading strategy.