Testing risk proxies for financial collateral haircuts: adequacy of capturing tail risk

What is it about?

The purpose of this study is to empirically test several statistical methods for analysing the adequacy of the proxy assignment in capturing tail risk. Currently, there are no academic studies or practical white papers addressing the need for a validation guidance on the index and benchmark assignment in banks’ internal models. The lack of guidance is especially evident for the under-researched financial collateral haircuts models. Upon the empirical review of the available statistical tools, this paper suggests using the Two-Sample Kolmogorov–Smirnov test in order to validate the adequacy of capturing tail risk by the assigned proxy and the Harrell’s C test to capture the discriminatory power of the proxy-based collateral haircuts models. The paper also suggests a tool that compares the reactions of risk proxies to tail events in order to verify possible underestimation of risk in times of significant stress.

Featured Image