Skulls, Financial Turbulence, and Risk Management

Based on a methodology introduced in 1927 to analyze human skulls and later applied to turbulence in financial markets, this study shows how to use a statistically derived measure of financial turbulence to measure and manage risk and to improve investment performance. View a webinar based on this article. We extended the research of those who have been analyzing the creation of optimal portfolios during times of financial turbulence. That research produced a mathematical measure of financial turbulence that captured the statistical unusualness of a set of asset returns given their historical pattern of behavior, including extreme price moves, decoupling of correlated assets, and convergence of uncorrelated assets. We showed that this measure of financial turbulence is nearly identical to the Mahalanobis distance, which was derived decades ago to analyze human skulls. Then, we provided evidence that this mathematical measure coincides with well-known episodes of financial turbulence, such as the stagflation of the late 1970s and early 1980s, the 1987 stock market crash, the Gulf War, Russia’s default on its sovereign debt, the technology bubble, 9/11, and the recent global financial crisis. We next discussed two intriguing features of financial turbulence. First, returns to risk are substantially lower during turbulent periods than during nonturbulent periods, and second, turbulence is persistent. It may arrive unexpectedly, but it does not immediately subside. It typically continues for weeks as market participants digest and react to its cause. Finally, we explored a variety of useful ways to apply this measure of financial turbulence. We showed how to stress-test portfolios by estimating value at risk from the covariances that prevailed during the turbulent subsample. We showed how to construct portfolios that are relatively resistant to turbulence by conditioning inputs to the portfolio construction on the performance of assets during periods of turbulence. In addition, we showed how to enhance the performance of certain risky strategies by using turbulence as a filter for scaling exposure to risk.