Risk 2.0: EVT Predicted Downside Risk of 2007

755209_windsockAs in 1929, the current financial crisis involved irrational exuberance and leverage. Banks were badly exposed this time as well, although not through direct equity investments. They were also very highly levered, even in comparison to 1929.

In retrospect, one of the consequences was obvious. There was a bubble in financial industry stocks worldwide and in bank stocks in particular. Many large equity portfolios were exposed to severe losses as this bubble burst.

This week we turn to the question of what our Extreme Value Theory based Expected Shortfall technology would have told an investor about the downside risk in his share holdings in two large Canadian banks, CIBC and RBC, from January 2005 through July 2010.

We’ll use expected shortfall as a common risk measure to compare the two. Our EVT technology is used to estimate the 1-day 99% Value at Risk (VaR) and the 1-day 99% Expected Shortfall (ES). All of our estimates are based on end of day returns from the previous year of trading–so the results are available daily at market close and can be acted upon during the following market day.

As a test of the reliability of the 99% EVT VaR estimates, we should see 1 VaR violation per 100 days out of sample. There were 1395 days out of sample and each bank had 15 VaR violations in that period–so there is a very good fit between our predictions and the observations. Click here for the full article. Click to review Part 1 or Part 2.