The Dynamics of Volatility and Correlation During Periods of Crisis : Implications for Active Asset Management

During the 2007-2008 financial crisis, idiosyncratic and market volatilities across the world increased to level never seen before. The financial econometrics literature focused on the spectacular increase in aggregate idiosyncratic volatility. Bekaert, Hodrick, and Zhang (2012) showed the high correlation across countries, documenting how most of the time variation in idiosyncratic volatility can be attributed, among others, to variation in U.S. market volatility and a business cycle sensitive risk indicator. However, what worried most asset managers and commentators was not so much the increase in volatility per se but the dramatic increase in correlation between and within stock markets. A phenomenon interpreted by the financial press as a sort of structural break in stock market dynamics. Sandoval and De Paula (2011) studied the phenomenon of the increase in correlation during period of crisis. They found that markets tend to behave as one in time of crisis. I will show that it is possible to calibrate a simple dynamic CAPM model over the last 20 years that fits very well the observed dynamic of stock markets' volatility and correlation. I use the model to investigate if the 2007-8 environment should have been detrimental to bottom-up managers and favourable for top-down manager, as far as “alpha” creation is concerned. There is in fact a sort of consensus view about the fact that the increase in correlation observed during the 2007-2008 should have impaired the capabilities of bottom-up managers to produce alpha. As there is good and bad cholesterol, there is good and bad volatility for bottom-up alpha generators