Social Sharing
Extended Viewer
In the Market - April 2008
Apr 15, 2008
Volatility Clustering
Many investors have been caught off guard by the recent high volatility in equity markets. However, looking at past equity price movements we can see that this volatility increase is not out of the ordinary. By some measures the current period has not been as extreme as the previous high volatility period from 1998-2003. In that span there were 19 extreme days which were greater than the biggest moves we’ve seen in the current environment. The chart below also illustrates volatility clustering – where low volatility breeds low volatility and high volatility breeds high volatility. Prior to this current rise in volatility we saw 4 years of lower and lower volatility and high returns which helped to encourage a level of complacency among investors. This complacency mixed with a large systemic shock has contributed to today’s high volatility period which – if history is our guide – may be with us for a while.
Volatility clustering illustrates the challenges risk managers face in identifying periods to use for predictive stress testing and VaR analysis. Consequently risk managers are now linking together multiple discrete days and periods to model new correlation periods. This technique which is as much art as science is improving risk managers’ ability to model volatility and correlation related risks.
Download