What is it about?
Financial institutions have to use Credit Default Swap data to estimate the default risks for their trading counterparties; however, for most of their counterparties, no such data are available. Thus, the institutions have to proxy them. Existing methods lead to model-arbitrage and lead to significant estimation errors. Our Machine Learning technique-based approach achieve superior accuracies and account for counterparty-specific risks based on comparisons between ours and existing methods that are being used by financial insitutions.
Featured Image
Why is it important?
Financial institutions need a sound CDS-proxy method, which we prove our approach is. In absence of such sound method, the lessons learned from the 2007-09 financial crisis is wasted. The financial institutions will continue to calculate the risks of counterparty default incorrectly as they did during the financial crisis for companies such as Lehman Brothers.
Perspectives
Financial crisis showed that financial institution failed to properly calculate the counterparty credit risk and systemic risk; for the former, banks have made progresses in using so-called Valuation Adjustments (or XVA) to rectify such problem. However, to that end, a sound CDS proxy method; our study showed that methodologically, empirically our Machine Learning-based method is more accurate, more sensible.
Zhongmin Luo
Read the Original
This page is a summary of: CDS Rate Construction Methods by Machine Learning Techniques, SSRN Electronic Journal, January 2017, Elsevier,
DOI: 10.2139/ssrn.2967184.
You can read the full text:
Contributors
The following have contributed to this page
