Mathematics Research Seminar Series: Swap Volatility and Systemic Risk in Hong Kong Banking: A Machine Learning Approach

Mathematics Research Seminar Series
Swap Volatility and Systemic Risk in Hong Kong Banking: A Machine Learning Approach

by Paul McNelis,
Gabelli School of Business, Fordham University, New York

Date: Wednesday, 26 October 2022
Time: 4:00 pm
Venue: https://bit.ly/AdmuMathSeminar

In this talk, sources of contagion emanating from both within the Hong Kong banking sector as well as from external sources will be assessed. For robustness, two alternative measures of contagion will be used, one based on Forecast Error Variance Decomposition (FEVD) of daily realized volatility, and the other based on Delta Conditional Variance at Risk, with weekly share-market returns. Recent advances in the Machine Learning literature will be utilized, based on Elastic Net, Cross-Validation and Neural Network dimensionality-reduction methods. The results show that the US and Hong Kong Swaptions market volatilities are significant sources of systemic risk for Hong Kong banks. The Hong Kong Swaptions implied volatility not only responds to movements in the implied volatility of United States Swaptions, but also contains added information important for understanding ex-post realized volatility in the share prices of Hong Kong banks.