Robin writes:
I must say, I'm not the biggest fan of factor-based investing. It's not just that I prefer the simplicity of market-cap-weighted indexing. I'm also doubtful as to whether, for most people, the alpha generated by factor investing will continue to justify the additional cost and volatility involved in trying to capture it.I'm certainly not against it altogether. If your investment time horizon is long enough, and you're sufficiently patient and disciplined, there is a case for trying to beat the market with at least a portion of your portfolio, by tilting it towards reliable risk factors. The problem is, as Joachim Klement explained the other day, reliable factors are few and far between. A recent study called Replicating Anomalies looked at nearly 450 factors factor — yes, there really that many — and found that the vast majority were "insignificant". For LARRY SWEDROE, it's another reminder that markets are really very efficient. But what else can we learn from this latest study?
As professor John Cochrane observed, the literature on investment factors now fills a veritable “zoo of factors”, with hundreds that are anomalies for the capital asset pricing model (CAPM) and other asset pricing models.
Kewei Hou, Chen Xue and Lu Zhang contribute to the literature with their study Replicating Anomalies, published in the May 2020 issue of The Review of Financial Studies. The authors conducted a gigantic replication of the bulk of the published anomalies literature in finance and accounting by compiling a largest-to-date data library with 447 anomaly variables.
The list includes 57, 68, 38, 79, 103 and 102 variables from the momentum, value-versus-growth, investment, profitability, intangibles and trading frictions categories, respectively. To control for micro-caps (stocks that are smaller than the 20th percentile of market equity for New York Stock Exchange, or NYSE, stocks), they formed testing deciles with NYSE breakpoints and value-weighted returns.
Their data sample was from January 1967 to December 2014. Financial firms and firms with negative book equity were excluded.
How significant are they?
The following is a summary of their findings:
— Out of 447 anomalies, 286 (64%) were insignificant at the 5% confidence level.
— Raising the bar to the 1% confidence level in order to control for what is called “p-hacking,” which increases the risk of false positives, raised the number of insignificant anomalies further, to 380 (85%).
— The biggest casualty was the liquidity literature. In the trading frictions category that contained mostly liquidity variables, 95 of 102 variables (93%) were insignificant.
— Prominent variables that did not survive replication included short-term reversal, idiosyncratic volatility, the distress anomaly (for example, z- and o-scores), accruals, and operating profits-to-book equity, dividend yield and payout yield.
— Even for significant anomalies, their magnitudes were often much lower (though often still economically large) than originally reported. Among them were price momentum; cash flow-to-price (a value metric); earnings momentum formed on standardised unexpected earnings, abnormal returns around earnings announcements and revisions in analysts’ earnings forecasts; and asset growth.
Disappearing anomalies
Hou, Xue and Zhang answered the question, “Why does our replication differ so much from original studies? The key word is micro-caps.” While micro-caps represent only 3% of the total market capitalisation of the NYSE-Amex-NASDAQ universe, they account for 60% of the number of stocks and more than 60% of the stocks in extreme deciles.
Micro-caps have the highest equal-weighted returns and also the largest cross-sectional standard deviations in returns. The authors noted: “With equal-weights, micro-caps earn on average 1.32% per month relative to 1.03% for big stocks. In contrast, the value-weighted market return of 0.93% is close to 0.92% for big stocks.”
Many studies overweight micro-caps with equal-weighted returns. Hou, Xue and Zhang explained: “Due to high costs in trading these stocks, anomalies in micro-caps are more apparent than real. More important, with only 3% of the total market equity, the economic importance of micro-caps is small.”
They added that micro-caps represented 2.5% of the total market equity in 1967 (increasing to 4.6% with the addition of Nasdaq), reached its maximum of 6.2% in 1984, and showed a downward trend afterward. At the end of 2014, micro-caps represented only 1.4% of the total market. The bottom line is that the percentage of small public firms has declined, with one likely explanation being the increased costs of being public.
The authors went on to note that their q-factor model (beta, size, investment and profitability) explained the bulk of the remaining anomalies though still left 46 with alphas significant at the 1% confidence level (t-stat = 3). Among them were abnormal returns around earnings announcements, operating and discretionary accruals, cash-based operating profits-to-assets, and R&D-to-market.
Markets "more efficient than previously reported"
Summarising their findings, the authors of Replicating Anomalies stated: state: “In totality, our evidence suggests that capital markets are more efficient than previously reported.”
In an interview with Alpha Architect’s Wes Gray, Lu Zhang was asked what strategy worked best. His answer, in practitioner terminology, boiled down to buying baskets of small firms that are cheap (i.e., value) and have high returns on equity (i.e., profitability/quality).
Given these findings, how can investors navigate their way through the factor zoo and minimise, if not eliminate, the risks of data mining?
Navigating your way through the factor zoo
In our book Your Complete Guide to Factor-Based Investing Andrew Berkin and I established the following criteria for a factor to be considered for investment. It must have provided a premium that has demonstrated:
Persistence
— It holds across long periods of time and different economic regimes.
Pervasiveness
— It holds across countries, regions, sectors and even asset classes.
Robustness
— It holds for various definitions (for example, there is a value premium whether it is measured by price-to-book, earnings, cash flow or sales).
Investibility
— It holds up not just on paper but also after considering actual implementation issues, such as trading costs.
Intuitiveness
— There are logical risk-based or behavioural-based explanations for its premium and why it should continue to exist.
The factors that met all these criteria were beta, size, value, momentum and profitability/quality. We also noted that the low-beta factor met the criteria as well because low-beta stock portfolios consistently have had more attractive Sharpe ratios than high-beta stock portfolios and remain attractive after transaction costs.
However, as is the case with all factors, the low-beta premium is time-varying and can become crowded. The popularity of the strategy, fuelled both by the bear market of 2008 and the strategy’s acceptance as a legitimate investment option, raised valuations.
In Your Complete Guide to Factor-Based Investing, we showed that when low-beta strategies are in a value regime (which they have been about 60% of the time), they produce above-market returns with below-market risks. However, when they are in a growth regime (as is currently the case), while they still produce below-market risks, they also produce below-market returns.
Summary
Replicating Anomalies is a must-read for those interested in the academic research on investment strategies. It might even make you question your deepest-held investment beliefs. For example, if you believe that searching for high dividend-yielding stocks is a sound approach for generating higher returns, the paper was unable to replicate the results!
Finally, while the authors noted that much of the premiums are found in the smallest stocks, this is not a new finding — it’s a well-known one. It’s also why many anomalies can persist, because the limits to arbitrage that prevent sophisticated investors from correcting mispricings (such as trading costs and the costs of borrowing securities in order to short them) are greatest in the smallest stocks.
Investors should keep these facts in mind as they choose the funds they will use to build their portfolios, making sure the manager has demonstrated the ability to keep trading costs low through patient trading strategies, and keep assets under management from growing too much so the fund can keep its average market cap at lower levels (and thus can keep exposure to the smallest stocks where the premiums have been the largest).
Investors should also seek out value strategies that incorporate exposure to the profitability/quality factor as well as momentum.