The Factor Interaction Effect
How many times have you read about Price/Earnings ratios and their impact on future returns? Whether you read academic articles or the popular press, discussions of individual factors are prevalent. People typically write about the Price/Earnings ratio, the Price/Book ratio, Price Momentum, Beta, Short Interest, and Price Targets. Our research has found that these investment factors are weak predictors of future returns when used by themselves. The deficiencies of these factors can be overcome with an understanding of behavioral investing. This edition of our blog will expand beyond individual factors and detail a critical but often underappreciated aspect of stock selection: the factor interaction effect.
The factor interaction effect refers to how individual investment factors, such as specific valuation, sentiment, quality, risk, or growth factors, can be combined to create a new factor that exhibits an improved level of efficacy in stock selection. When this effect is efficiently utilized, this newly created factor can generate materially more robust returns than any individual component factors that comprise it over time. While this interaction effect can be achieved with a range of individual factor combinations, we would note that this effect is most prominently observed when different styles of factors, such as valuation and sentiment factors, are combined.
We can highlight this interaction effect in historical market data by examining the return characteristics of three representative investment factors: Earnings to Price Yield, EPS Estimate Diffusion, and Return on Equity. Each of these factors was chosen because of their widespread use among investors and their ability to represent various styles of factors. In particular, the Earnings to Price Yield factor is definitively a valuation factor, EPS Estimate Diffusion is a sentiment factor, and Return on Equity could best be classified as a quality factor. More detailed descriptions of these factors can be found in the table below.
With each of these individual investment factors now defined, the charts below illustrate the magnitude and scope of the interaction effect generated among these three factors. The chart on the left exhibits the average twelve-month relative return of companies with a positive score for the factors above versus the broader equity universe over the test period. Importantly, the factor scores are sector neutralized, meaning that the positive scoring companies represented in the relative returns below are better than the average of the respective sector according to each factor. The leftmost chart also displays the same average twelve-month relative return of the combination factor, a new factor comprised of the individual factors that specifically represent companies with a positive score on all three individual factors. The rightmost chart below examines the consistency of the relative returns for each previously detailed group of companies by displaying the percentage of trailing twelve-month periods during the test window where companies with a positive score outperformed the broader market. The results in both charts reflect a decade-trailing test period from 03/2010 to 12/2020 that was run with a quarterly frequency and with the Russell 3000 used as the universe.
As can be seen above, each of the three individual factors has demonstrated a fair degree of efficacy during the specific period tested. Companies with positive scores for these factors have all generated returns above the average company in the equity universe and demonstrated this outperformance in 70% or more of the trailing twelve-month periods that comprise the decade-long backtest. However, the combination factor, which represents companies with a positive score on all three of these factors, notably generated even better return characteristics. This can be seen as the combination factor generated an average relative return of 2.4%, a level nearly 1.0% better than the best-performing individual factor. We would also note that the combination factor demonstrated strong consistency. This new factor generated outperformance in 82.5% of historical trailing-twelve-month periods, a material improvement relative to the individual factors.
The attractive return characteristics of the combination factor noted above result from the factor interaction effect. In particular, this combination factor can generate returns that are markedly superior to the factors that comprise it is because these individual factors enhance the return of the whole when joined together. The underlying driver of this factor synergy has deeply behavioral underpinnings related to the core process of security mispricing. Decades of research have indicated that market price and intrinsic value divergences do not arbitrarily appear in isolation but instead follow a predictable cycle of mispricing. This cycle of mispricing, which Hillcrest refers to as the Behavioral Cycle, is illustrated in the chart below. It occurs as the market price of a company oscillates around its intrinsic value in a systematic pattern over time.
The notable power of the factor interaction effect accordingly exists because combining various individual factors can enable investors to capture a broader portion of this repeated cycle of security mispricing. This is a critical distinction as solely utilizing a single factor, or a closely-related set of factors, results in investors only capturing an isolated segment of the mispricing cycle given the narrow factor focus. However, by joining a disparate set of individual factors, especially those with low correlations, investors can utilize the factor interaction effect and generate higher returns through a larger portion of the mispricing cycle.
In closing, we would note that the factor interaction effect is complex. The example utilized in this Behavioral Insight piece is intended to serve as a rudimentary demonstration to highlight the effect. Not every factor can be effectively combined to improve the results of a security selection model, and optimally capturing this key effect is a challenging and research-intensive process. With that said, there are indeed compelling opportunities to properly harness this interaction effect for investors with both the research capabilities and an understanding of its behavioral drivers.
Any forecasts, figures, opinions, or investment techniques and strategies explained are Hillcrest Asset Management, LLC's as of the date of publication. They are considered to be accurate at the time of writing, but no warranty of accuracy is given, and no liability in respect to error or omission is accepted. They are subject to change without reference or notification. The views contained herein are not to be taken as advice or a recommendation to buy or sell any investment. The material should not be relied upon as containing sufficient information to support any investment decision.
Data is provided by various sources and prepared by Hillcrest Asset Management, LLC and has not been verified or audited by an independent accountant. Test results are not indicative of future results and should not be relied upon. While the information provided above is not based on the performance of any individual security or group of securities, the methodology used to provide the information can be obtained by contacting Hillcrest Asset Management, LLC.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.