COVID volatility underscores need for next generation in risk management
New academic research by Plato Investment Management’s Head of Long Short Strategies, Dr David Allen, has highlighted the need for investors to move beyond the traditional “bell-shaped” normal distribution assumption that underpins much of industry practice.
Dr Allen’s research paper, titled A comparison of non-Guassian VaR estimation and portfolio construction techniques, has recently been published in the prestigious Journal of Empirical Finance.
“It is widely accepted by academics that asset returns do not follow the well behaved bell-shaped normal distribution of economic textbooks, and that extreme returns occur much more frequently than one would expect,” said Dr Allen.
“For example, if stock returns really were ‘normally’ distributed as practitioners tend to assume, the 12% fall in the S&P 500 that occurred on the 16th March as COVID-19 fears gripped the world would not have occurred even if the stock market had been open every day since the Big Bang. The 20% drop in the S&P 500 that occurred on Black Monday, October 19th, 1987, would not have occurred even if the history of the universe was repeated one billion times.
“Using the normal distribution in a non-normal world is to court disaster. Nevertheless, the normal distribution forms the bedrock of modern financial practice, primarily because it is mathematically tractable and easy to use.
“Unfortunately, almost all of the widely available approaches for building portfolios and measuring risk are based on how asset prices should behave rather than on how they do behave in the real world.”
While completing his PhD at Cambridge University, Dr Allen developed a new model of asset returns that accounts for “fat tails”, volatility clustering (where extreme returns beget extreme returns), and correlations increasing during times of stress.
The results, published in Journal of Empirical Finance, show that investors can significantly increase their risk adjusted returns by moving beyond the traditional “normal” assumptions of modern finance.
“In 2007-9 the CDO (Collateriased Debt Obligations) pricing model was based on the normal distribution and subsequently described as the formula that felled Wall Street due to the mispricing of risk. The recent market turbulence has again underscored the need for approaches to risk management that reflect how asset prices really do behave, rather than how we would like them to.”
