A1666
Title: A copula-based data augmentation strategy for the sensitivity analysis of extreme operational losses
Authors: Amir Khorrami Chokami - University of Cagliari (Italy) [presenting]
Giovanni Rabitti - Heriot-Watt University (United Kingdom)
Abstract: The aim is to assess the importance of macroeconomic and financial variables for UniCredit Bank's operational losses. To achieve this, the Shapley effects is considered a variance-based measure of importance. However, the small number of observations of extreme losses makes the estimation of the Shapley effects challenging. To address this issue, augmenting the sample of extreme observations is proposed, using vine copulas and calculating the Shapley effects on the augmented sample. The effectiveness of this procedure is supported by a numerical simulation. Findings obtained with the methodology applied to the UniCredit Bank data show its usefulness for the risk management of operational losses.
