Yesterday, the Bank of International Settlements released consultative document sets out a revised market risk framework that proposes a number of specific measures to improve trading book capital requirements. (BIS Fundamental Review of the Trading Book) Amongst the points to consult on is whether to replace the "poster child of failure" in the 2008 financial crisis - the value-at-risk (VaR) measure, with expected shortfall (aka CVaR), a risk measure that better reflects "black swan" events. Is this really news?
One could argue not in certain respects. Academic research (from at least a decade back) has showed CVaR to hold advantages compared to VAR from a mathematical standpoint. Also, many financial institutions have already incorporated expected shortfall as part of their internal dashboard of risk measures. So what is the issue?
- For one, good or bad, because expected shortfall also measures highly improbable risks, it opens up the door to a greater degree of ambiguity related to modeling assumptions.
- Secondly, the metric is also sensitive to rare "tail risk" events that are many practitioners consider low probability. Some consider that these events should be determined by a firm’s view of the world and not be imposed through regulatory mechanisms for every firm.
- With expected shortfall, this is likely to increase regulatory capital required (in addition to what is currently already on the table to be met by banks)
In a time where there is a drought of liquidity and capital in the banking industry, most bankers will be hoping (and arguably lobbying hard) that the proposals in this consultation should be carefully weighted out rather than hastily imposed. Let the debates begin!
