Autocorrelation Effects on CTA and Equity
Risk Measurement

Abstract

In recent years considerable research has examined the impact of historical return patterns on the measured risk of both traditional and alternative investments. At the heart of this analysis is the industry’s failure to take the strategy autocorrelation of a security or asset into account when measuring the asset’s underlying risk or in estimating the current risk environment (Lo [2002]). As the authors show in this article, the industry’s failure to take autocorrelation into account has produced estimates of equity volatility that are much lower than they should be, and estimates of CTA volatility that are much higher. Given the results reported in this article, an equity volatility of 16% really should be closer to 20%, and a CTA volatility of 10% should be closer to 6% or 7%.

TOPICS: Commodities, futures and forward contracts, statistical methods, risk management