Are there strategic asset allocation methodologies that reliably beat equal weight? In the February 2012 version of their paper entitled “Portfolio Optimization Using Forward-Looking Information”, Alexander Kempf, Olaf Korn and Sven Sassning investigate the performance of a minimum variance portfolio based on returns implied by equity options rather than historical returns. They argue that, since option prices reflect the expectations of market participants, the former approach is inherently forward-looking. The methodology involves calculating option-implied volatilities and option-implied correlations. Using daily prices for the Dow Jones Industrial Average (DJIA) stocks and associated option-implied return statistics during 1998 through 2009 for out-of-sample testing, and DJIA stock prices for 1993 through 1997 for historical data tests, they find that:
- A minimum variance portfolio constructed solely from option-implied return statistics, rebalanced monthly, consistently outperforms both a comparable portfolio based on historical returns and a simple equally weighted portfolio.
- Outperformance holds with and without shorting restrictions, for quarterly rebalancing, and with and without 0.3% one-way trading frictions.
- Outperformance is particularly strong during crises, when information asymmetry between options and stock markets is likely high.
- Combining implied and historical data does not work well, sometimes underperforming purely historical approaches.
In summary, evidence indicates that investors can better capture stock diversification benefits by relying solely on option-implied return expectations rather than historical returns.
Cautions regarding findings include:
- Calculating option-implied return statistics would involve burdensome and costly data collection/processing for many investors.
- There may not be sufficient options data to test or use the methodology for robust diversification across asset classes.
