This is a summary of “Risk Management and Optimal Combination of Equity Market Factors,” by Roger Clarke, Harindra de Silva, CFA, and Steven Thorley, CFA, published in the Third Quarter 2020 issue of the Financial Analysts Journal.
Combining factors in a multi-factor portfolio using forecast risk management can add substantially to investment returns. Backtesting showed such a strategy run over 54 years would have made annualized returns of 10.79%, vs. 7.77% for a similar non-risk-managed portfolio.
What Is the Investment Issue?
What is the best way to combine multiple factors into a single integrated portfolio? Can the ongoing risk management of a multi-factor portfolio enhance long-term returns and improve risk-adjusted investment performance? Factor investing involves identifying groups of securities where the financial performance is driven by characteristics associated with each group. Five commonly used factors are value, momentum, profitability, small size, and low beta. Combining these in an optimal way into one portfolio is a challenge faced by all multi-factor investors.
How Did the Authors Tackle the Issue?
The authors analyzed three different approaches to constructing five-factor portfolio strategies using data from a 54-year period from 1966 through 2019. That dataset, which includes the largest 1,000 US stocks throughout the period, was used to backtest each of the five common strategies (value, momentum, profitability, small size, and low beta).
The first approach involved designing a portfolio with an equally weighted mix of so-called primary factors: The securities selected for each factor grouping had exposures to other factors. For instance, the small-cap factor securities might also have exposure to the momentum factor or the value factor or any of the others. As a result, primary factors likely wouldn’t perform independently of each other, with “the other factors in some cases dominating the intended factor exposure,” the authors state.
The second approach looked at an equally weighted portfolio of pure factor investments. In this portfolio, there were zero exposures to secondary factors. The securities in one factor, such as value, would not have exposure to the profitability factor, for instance.
The third approach tested how a portfolio of the five pure factors would perform if the investments were risk managed—that is, if the factor exposures were raised or lowered each month in response to recent changes in factor volatility. Specifically, the idea was to increase exposure to a factor when its volatility was forecast to be lower and vice versa. The authors note that the riskiness of factors change in persistent ways and can, therefore, be more predictable than forecast factor returns.
What Are the Findings?
The results of the backtested performance analyses showed that all three multi-factor approaches delivered performance superior to that of an investment in the market.
Furthermore, the authors found that adding risk management had a beneficial effect on performance that was of the same order as adding an extra factor to the portfolio! The actively managed pure portfolio performed best, with mean annual returns of 10.79% and a Sharpe ratio of 0.765. The equal mix of pure factors and primary factors resulted in mean returns of 7.77% and 7.39% and Sharpe ratios of 0.537 and 0.516, respectively. These results all compare favorably with average returns of 6.29% and a Sharpe ratio of 0.417 for the market benchmark.
Although the backtesting itself didn’t include transaction costs or portfolio turnover, an additional analysis gave some estimates of the likely costs assuming the turnover of the portfolio was 10% a year. For the dynamic portfolios, the authors infer these transactions costs would have reduced gross returns for the risk-managed portfolio by around 30 bps per year, versus a 10 bp reduction for the static portfolios. Even with these costs added, the risk-managed portfolio returned more alpha than the other strategies.
What Are the Implications for Investors and Investment Managers?
The analyses show that a multi-factor portfolio can substantially outperform the broader market. That base performance improvement, whether from using primary or pure factors, can be further enhanced by combining factors using ongoing risk management whereby the exposures to the various factors are increased or decreased in response to changes in their forecast risk.
Furthermore, the authors develop a closed-form formula to implement this multi-factor strategy. The authors explain, “Because closed-form formulas are used to create the optimal multi-factor portfolios, our empirical research can easily be extended to other country markets or the global equity market.”