Thursday, April 15, 2010

Follow up: The difference a daily risk model makes

From my last post, a commenter posed the question:

"The benefit for steplike intra month risk increase is nice, but what about a single day peak - getting nervous ... Is it measuring or forecasting risk numbers?"
This is a very valid concern and worth addressing. I took a look at the days prior to and after a jump in stock specific risk for a number of companies. I generated these charts to show what happened to the levels of risk after the change.
The bars in yellow is the first day of the "spike" in risk. As you can see, risk levels maintain the new higher level of risk, although we must acknowledge there is some amount of correction in the days following (take BBBY for example). I went through my entire testing universe and could not find a "single day peak" where there is an increase in risk followed by an similar magnitude decrease in risk.

Thank you for your question. Please leave a comment if you have further questions about this or any other blog entry.

To receive new posts by e-mail, subscribe to this blog.

1 comment:

  1. Looks like a case of volatility clustering due to regime switching ?