Dynamic Volatility Management: From Conditional Volatility to Realized Volatility

Abstract

The volatility of concern in conventional volatility-managed strategies such as volatility-targeting strategy and mean-variance optimization is the expected conditional volatility. However for investors, it is the realized volatility that is important, because there is only one realization in the market. Simply managing the conditional volatility may not manage well the realized volatility of the actual portfolio performance. This article provides a multi-period strategy that directly manages the realized volatility over a long horizon. More specifically, the strategy maximizes the expected portfolio value subject to an upper cap on the realized volatility. Our out-of-sample backtesting results show that this novel strategy delivers higher risk-adjusted returns compared to the volatility-targeting strategy, and that it successfully caps the realized volatility below the targeted level. The results are consistent across twelve equity markets and five targeted volatility levels.

Keywords: volatility management, volatility target, realized volatility, multi-period portfolio management, least squares Monte Carlo

JEL Classification: G11, G17, C61, C63