Shortcuts to Factor Investing: Multifactor Portfolios and Benchmarking
As alternative investing models like active management struggle, the factor investing approach provides a refreshing alternative, picking out underlying drivers of investment returns such as inflation, interest rates, or GDP growth.
A multifactor approach combines these identified drivers into an optimal portfolio to better satisfy investor objectives using macroeconomic, fundamental, and statistical factor models.
The first article in this series described the intellectual history of factor investing. The second took a deeper dive into different factor investing approaches for stocks and wider applications to other asset classes including fixed income. Here, we look at combining factors in multifactor portfolios and consider issues of performance measurement in factor investing.
More than 50 years after its introduction, the most powerful concept in forming investing portfolios remains traditional mean variance optimization (MVO) based on quadratic concepts in modern portfolio theory (MPT) — which follows the popular intuition of not putting all your eggs in one basket.
MPT has weathered any number of attempts to overturn it, such as the seemingly insightful but somewhat impractical prospect theory and behavioral finance, an academic field dependent on elegant but questionable models and hard-to-replicate experiments. In contrast, multifactor investing retains the robust and familiar framework of MPT but with different inputs.
With their underlying common exposures, certain underpinning asset classes can be readily remixed to improve portfolio diversification, according to Eugene L. Podkaminer, CFA, in “Risk Factors as Building Blocks for Portfolio Diversification: The Chemistry of Asset Allocation.” Citing common and overlapping risk exposures such as currency, volatility, and inflation, Podkaminer shows how correlations among asset classes can be high, detracting from diversification.
But implementing factor investing remains a challenge. No single opportunity set, such as the market portfolio, encompasses all the factors. Weighting factors can also be complicated, with many researchers and practitioners using the shortcut of an equal weighting technique. An individual stock can overlap several different factors leading to complex interaction effects. Such factors as momentum or GDP growth may not be easily investable. Correlations, just like past and future factor risk premia, can be fast-moving targets and necessitate frequent re-balancing. Use of the long and short exposures necessary to implement some factors may be disallowed for certain investors.
One survey summarized in CFA Digest finds issues and challenges in key stages of the decision-making and implementation process of factor investing. Yet, the potential for practical applications is substantial enough to suggest that interest in the area is likely to endure. Some potential benefits include:
- Better evaluation of a portfolio across the various macroeconomic scenarios using the factor lens.
- De-risking liability-driven investing (LDI) portfolios on both the asset and liability sides.
- Use of factors within and across manager structures to identify cross-correlations.
A new monograph by a team of researchers at PIMCO, Factor Investing and Asset Allocation: A Business Cycle Perspective, suggests that real-life portfolio allocation problems can be better solved by adopting a broad rather than narrow range of models and working with simpler models.
Performance Measurement and Factor Investing
New investing techniques demand rigorous criteria to measure and attribute their success or failure. In his timely article, Andrew Lo asks the existential question, “What Is an Index?” Lo finds that dynamic indices, such as a volatility-controlled index, not tied to market-cap weights have many advantages in terms of meeting investor objectives, but they also require greater sophistication to evaluate and avoid unintended consequences.
Factor analysis can be used to assess whether a manager is achieving alpha or using a factor-based strategy to capture returns from beta says Deborah Kidd, CFA, in “Factor Investing: When Alpha Becomes Beta.” The benchmarking performance of factor investing seems to offer many challenges since direct comparisons with readily available indices are not always useful.
Many end investors revert to oversimplified comparisons of individual factor performance with T-bill rates or long-run (market-cap weighted and long-only) equity or bond returns. These are unsatisfactory pairings, especially since equity and bonds have shown such high recent correlation. Peer group comparisons across factors are also difficult, especially once leverage, fees, and costs — which are not equal across factors and strategies — are taken into consideration. Nevertheless, it would be odd if the same skepticism about active management that drove investor flows into factor investing was suspended for underperforming factors.
Ultimately, investors should be forecasting as much (ex ante) as they are looking backward (ex post) at factor performance in their portfolios. To do that effectively, a better understanding of the selected units of risk and the assignment of more thoughtful benchmarks at the outset should help investors obtain more return at lower risk.
A full topical collection of recommended links for all the related material on factor investing covered in this series can be found below along with a Glossary.
Factor Investing: Applies to a wide range of risk-based approaches that sit between active and passive investment management but possess attributes of both. Uses both long and short techniques.
Smart Beta: A marketing label describing simple, rules-based, and transparent approaches to building portfolios that deliver fairly static exposures (relative to capitalization-weighted benchmarks) to characteristics historically associated with excess risk-adjusted returns. Often long only.
Alternative Beta: A subset of “smart beta,” alternative beta is distinguished from smart beta by its use of short as well as long investing.
Fundamental Indexation: Another subset of “smart beta” with a focus on using accounting, economic, and weighting data to develop new indices.
Topical Reading Collection
1. What Is Factor Investing?
“I have got bad news as a starter,” Antti Ilmanen told the audience at the 2016 CFA Institute European Investment Conference. “It is not only a low interest rate world, it is also a low expected return world on any long-only investment.” Ilmanen, a principal and researcher at hedge fund AQR, said low expected returns are going to anchor bad news for all of us for the rest of our working lifetimes. And maybe beyond.
Smart beta products are a disruptive financial innovation with the potential to significantly affect the business of traditional active management. Ronald N. Kahn and Michael Lemmon provide an important component of active management via simple, transparent, rules-based portfolios delivered at lower fees.
The quant manager has the same set of tools that any active manager has: Quants simply apply them using the ever-increasing power of computers, Gina Marie N. Moore, CFA, observes. These tools allow the manager to pursue reward and deal with risk, costs, fees, and buying themselves the time necessary to distinguish investment skill from luck.
Combining long-only constrained factor subportfolios is generally not a mean variance efficient way to capture expected factor returns, Roger G. Clarke, Harindra de Silva, CFA, and Steven R. Thorley, CFA, observe. For example, a combination of four fully invested factor subportfolios — low beta, small size, value, and momentum — captures less than half (e.g., 40%) of the potential improvement over the market portfolio’s Sharpe ratio. In contrast, a long-only portfolio of individual securities, using the same risk model and return forecasts, captures most (e.g., 80%) of the potential improvement.
2. Industry Scope and Challenges to Factor Investing
Since the beginning of the 20th century, institutional investors have gained prominence in the UK and US financial markets not only because of changes in economic access but also because of changes in the way governments protect investors, according to Janette Rutterford and Leslie Hannah.
Financial research has uncovered many new factors (small cap, value, momentum, low beta) that explain stock returns. In fact, many of these factors have already been commercialized into financial products. The authors examine whether these historical insights and return patterns remain after the academic research discovering them is published.
Because of the potential for data mining and multiple testing, it is common practice to haircut reported Sharpe ratios by 50% when evaluating backtests of trading strategies. Campbell R. Harvey and Yan Liu propose an approach that calculates a haircut to the Sharpe ratio to account for data mining and multiple testing.
Exchange-traded funds (ETFs) have been growing in popularity with recent developments in factor-tilted strategies. Some investors have observed that these portfolios derive most of their outperformance from exposure to only two factors — value and small size — and the portfolios outperform even if randomly put together or turned upside down (monkey portfolios), according to Noël Amenc, Felix Goltz, and Ashish Lodh.
Investors are wary of the robustness of the outperformance of smart beta strategies. Noël Amenc, Felix Goltz, Sivagaminathan Sivasubramanian, and Ashish Lodh address this concern by providing measures of relative and absolute robustness. They examine the causes of a lack of robustness and propose remedies for these problems. Their conclusions focus on the dangers of data mining and a lack of transparency.
The multifactor investing framework has become very popular in the indexing community. Both academic and practitioner researchers have documented hundreds of equity factors. But which of these factors are likely to profit investors once implemented? Noah Beck, Jason Hsu, Vitali Kalesnik, and Helge Kostka find that many of the documented factors lack robustness.
3. Equities Factor Investing
Eugene Fama and Kenneth French introduce a five-factor asset pricing model that outperforms the well-known Fama–French three-factor asset pricing model in explaining stock returns. Surprisingly, when the two additional factors of profitability and investment are added to the original three-factor model, the value factor becomes superfluous. Although the five-factor model is not without its challenges, it is useful in describing the cross-sectional variance of the factors’ expected return.
By adding profitability and investment factors to their earlier three-factor model, Eugene Fama and Kenneth French explain the market β, net share issues, and volatility anomalies. The accruals and momentum anomalies cannot be explained by the five-factor model.
Managed volatility and covered call writing are two of the few systematic investment strategies that have been shown to perform well across a variety of empirical studies and in practice. So far, they have been studied mostly as separate strategies. It turns out that when combined, these two strategies create a powerful toolset for portfolio enhancements, according to Anna Dreyer, CFA, Robert L. Harlow, CFA, Stefan Hubrich, CFA, and Sébastien Page, CFA.
The idea that seemingly cheap securities, according to measures of fundamental and intrinsic value, outperform seemingly expensive securities has been scrutinized by academics for more than 30 years, yet the value strategy is still widely misunderstood. Recent research that updated the extensively cited Fama–French three-factor model introduced two new factors that claim to make the value factor redundant. Clifford Asness, Andrea Frazzini, Ronen Israel, and Tobias Moskowitz identify a number of facts and fictions about value investing that need clarification.
The term “value investing” is increasingly being adopted by quantitative investment strategies that use ratios of common fundamental metrics (book value, earnings) to market price. A hallmark of such strategies is that they do not involve a comprehensive effort to determine the intrinsic value of the underlying securities. We argue that these strategies should not be confused with value strategies that use a comprehensive approach in determining the intrinsic value of the underlying securities.
In factor investing, assets are viewed as bundles of underlying risk factors, according to Andrew Ang. Investors should hold factors whose losses they can endure more easily than the typical investor can. Ideally, the benchmark for factor investing is dynamically based on investor-specific circumstances rather than on market capitalization.
4. Bond Factor Investing
A risk factor–based approach can be used for managing fixed-income portfolios. Ramu Thiagarajan, Douglas J. Peebles, Sonam Leki Dorji, Jiho Han, and Chris Wilson show that a limited set of factors — rate, growth, and volatility — explain the return on fixed-income portfolios. Investors can use this approach in managing and analyzing their portfolios and in incorporating their macro views into their asset allocation decisions.
To investigate the fundamental indexing methodology, Lidia Bolla, CFA, applies it to global government bond markets and examines its exposure to several newly introduced risk factors. She finds that the fundamental indexing approach outperforms a market-value-weighted index. However, her results show statistically significant and economically relevant exposures of fundamentally weighted indexes to the risk factors term and duration risk, default risk, convexity risk, liquidity risk, and carry trade risk. The increased risk exposure explains the outperformance of the fundamental indexing methodology in government bond markets.
Fixed-income attribution explains the sources of a manager’s active return, Deborah Kidd, CFA, observes. A complex process, attribution can be challenging to implement and often plagued by large, unexplained residual returns. Understanding the assumptions underlying a manager’s attribution model and their relation to the investment process, along with a qualitative assessment, can help determine how well the attribution reflects the manager’s decision-making skills and provide a clearer picture of performance.
5. Other Asset Classes
Patrick Houweling and Jeroen van Zundert, CFA, offer empirical evidence that size, low-risk, value, and momentum factor portfolios generate economically meaningful and statistically significant alphas in the corporate bond market. Because the correlations between the single-factor portfolios are low, a combined multi-factor portfolio benefits from diversification among the factors: It has a lower tracking error and a higher information ratio than the individual factors.
The number of alternatively weighted equity indices, also called “alternative beta indices,” has risen dramatically since their introduction into the index scene in the mid-2000s. Deborah Kidd, CFA, provides an overview of popular alternatively weighted index schemes and gives investors a framework with which to understand and gauge the suitability of an alternative index for their desired risk exposures.
Having created a monthly dataset of US security prices between 1801 and 1926, Christopher C. Geczy and Mikhail Samonov conduct out-of-sample tests of price-return momentum strategies that have been implemented in the post-1925 datasets. The additional time-series data strengthen the evidence that price momentum is dynamically exposed to market risk, conditional on the sign and duration of the trailing market state.
Xi Li, Rodney N. Sullivan, CFA, and Luis Garcia-Feijóo, CFA, CIPM, explore whether the well-publicized anomalous returns associated with low-volatility stocks can be attributed to market mispricing or to compensation for higher systematic factor risk. The results of their study, covering a 46-year period, indicate that the relatively high returns of low-volatility portfolios cannot be viewed solely as compensation for systematic factor risk.
Research showing that the lowest-risk stocks tend to outperform the highest-risk stocks over time has led to rapid growth in so-called low-risk equity investing in recent years. Luis Garcia-Feijóo, CFA, CIPM, Lawrence Kochard, CFA, Rodney N. Sullivan, CFA, and Peng Wang, CFA, examine the performance of both the low-risk strategy previously considered in the literature and a beta-neutral low-risk strategy that is more relevant in practice.
6. Multi-Factor Approaches
After completing this chapter of Quantitative Investment Analysis, Third Edition, readers will be able to describe arbitrage pricing theory (APT), including its underlying assumptions and its relation to multifactor models; define arbitrage opportunity and determine whether an arbitrage opportunity exists; calculate the expected return on an asset given an asset’s factor sensitivities and the factor risk premiums; and more.
Asset classes can be broken down into factors that explain risk, return, and correlation characteristics better than traditional approaches.
This monograph, by Vasant Naik, Mukundan Devarajan, Andrew Nowobilski, Sébastien Page, CFA, and Niels Pedersen, draws heavily on the vast body of knowledge that has been built by financial economists over the last 50 years. Its goal is to show how to solve real‐life portfolio allocation problems. The authors have found that using a broad range of models works best and prefer simple over complex models.
Diversification in portfolios is desirable, but the models we use to achieve this objective may be misleading because they are divorced from macroeconomics. Individual and institutional investors should pay close attention to improvements to the traditional approach to asset allocation, including consideration of forward-looking macro views.
Despite the shortcomings of traditional asset allocation policies, most investment portfolios are still constructed based on direct asset class exposure. In addition, it may not be feasible for investors to implement policy-level decisions using a factor-based allocation framework. Daniel Ung, CFA, and Xiaowei Kang, CFA, discuss three approaches to risk factor–based portfolio construction and offer their reflections on the practical aspects of implementation.
7. Performance Measurement of Factor Investing
The recent advances in computational and financial technology and resultant financial innovation have created the possibility of a new perspective on indexes, indexation, and the distinction between active and passive investing, writes Andrew Lo.
Many sources of alpha have become easy to identify and widely replicated over time, Deborah Kidd, CFA, notes. Such systematic return drivers, or factors, occupy the space between traditional beta and alpha. They represent investment strategies that require skill beyond passive investing but not the complexity necessary for alpha generation
In attempting to profit from the anomaly that the observed returns for high-beta stocks inadequately compensate for their higher exposure to market risk, practitioners have increasingly “bet against beta” — selling short high-beta assets and buying low-beta assets. Scott Cederburg, CFA, and Michael S. O’Doherty challenge the existence of any such anomaly.
If you liked this post, don’t forget to subscribe to the Enterprising Investor.
All posts are the opinion of the author. As such, they should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute or the author’s employer.
Image credit: ©Getty Images/erhui1979