Financial researchers have uncovered many relationships between investment factors and security returns. For investors, an important question is whether these relationships will continue after the research on them has been published. Said another way, should we expect a premium to continue outside of the sample period if everyone knows about it?
As my co-author Andrew Berkin and I explain in our new book, “Your Complete Guide to Factor-Based Investing,” unless such a relationship is persistent across long periods of time and economic regimes, pervasive around the globe, robust to various definitions, investable (survives transaction costs) and has logical, risk-based and/or behavioral-based explanations, we should be skeptical about its persistence going forward.
This is especially true for technical trading rules that rely solely on historical prices and, thus, don’t have risk-based explanations (which cannot be arbitraged away).
If anomalies are the result of behavioral errors—or even investor preferences—and the publication of research about them draws the attention of sophisticated investors, it is possible that post-publication arbitrage would cause premiums to disappear. Investors seeking to capture an identified premium could quickly move prices in a way that reduces the return spread between assets with high and low exposure to the factor.
However, limits to arbitrage (such as aversion to shorting and its high cost) can prevent arbitrageurs from correcting pricing mistakes. And the research shows that this tends to be the case when mispricings exist in less-liquid stocks where trading costs are high.
Publish & Anomalies Perish?
Two recent studies—the 2015 paper “When Anomalies Are Publicized Broadly, Do Institutions Trade Accordingly?” by Paul Calluzzo, Fabio Moneta and Selim Topaloglu; and the paper “Does Academic Research Destroy Stock Return Predictability” by R. David McLean and Jeffrey Pontiff, published in the January 2016 issue of the Journal of Finance—provide us with insights into the question of return predictability.
The authors of both studies found that, post-publication, institutional investors and hedge funds—and to a lesser degree, active mutual funds—trade on the published information, exploiting the pricing mistakes of retail investors.
Thus, institutional trading and anomaly publication are integral to the arbitrage process, which helps bring prices to a more efficient level. Both studies also found that, on average, anomalies shrink by about one-third. Where anomalies occur in more liquid, easily traded stocks, anomalies can even disappear.
Bollinger Bands & Popularity
Jiali Fang, Ben Jacobsen and Yafeng Qin contribute to the literature addressing the effect of publication on anomaly returns through their study “Popularity versus Profitability: Evidence from Bollinger Bands,” which appeared in the Summer 2017 issue of The Journal of Portfolio Management.
Among numerous technical indicators, methods that involve Bollinger Bands are some of the most widely used. In 1983, John Bollinger introduced Bollinger Bands on the Financial News Network (which eventually became CNBC), where he was chief market analyst. Bollinger Bands generally include three parameters, with the following default settings:
- A middle band = 20-day moving averages of underlying prices
- An upper band = the middle band plus two standard deviations of the underlying prices
- A lower band = the middle band minus two standard deviations of the underlying prices
According to Fang, Jacobsen and Qin, who cite a 2001 book that he authored, Bollinger used 20 days to “capture reasonable intermediate-term price fluctuations and, in statistical terms, the ±2 standard deviations should contain about 95% of the price variations. This means that the price falling outside the bands signals a potential market change.”
They continue: “The basic application of Bollinger Bands, namely, the volatility breakout method, generates a buy (sell) signal when the underlying price closes outside the upper (lower) band.”
Since 1983, Bollinger Bands have gained popularity among investors. In 2001, Bollinger published his influential work on this indicator in a book, “Bollinger on Bollinger Bands.”
Today, his work and the trading strategies based on Bollinger Bands continue to be popular. The question Fang, Jacobsen and Qin sought to answer was: Should they still be popular? Or, has publication consigned them to the graveyard of once-successful strategies?
Bollinger Bands seem like a logical candidate for efficient markets to arbitrage away.
The authors, for instance, note: “Bollinger Bands use only information derived from historical prices to predict future returns. This means that investors have easy access to the data. In addition, the strategy itself is relatively easy to implement, since it does not involve sophisticated financial modeling or parameter estimation.”
Fang, Jacobsen and Qin found their evidence is consistent with the hypothesis that potentially profitable trading strategies indeed quickly self-destruct with increasing popularity. Bollinger Band-based trading strategies seem to have worked well before the mid-1980s, and returns were generally positive for nearly 60 years. However, from 1993 through 2013, they found that, for strategies based on the S&P 500 and Bollinger Bands, “returns are mostly negative … and such losses even have worsened over time.”
In addition to the U.S., the authors examined profitability on a number of major international stock markets, including Australia, France, Germany, Hong Kong, Italy, Korea, Japan, New Zealand, Singapore, Spain, Switzerland and the U.K. For each market, they used the longest sample available that starts between 1885 (DJIA) and 1971 (Madrid SE General Index), and all samples end in 2014.
Their results were consistent with those they found in the U.S. They write: “In most international markets, the forecastability of Bollinger Bands disappeared after … 2001.” They also found that their results held up to robustness tests—such as using a different number of days for the moving average in the bands.
Fang, Jacobsen and Qin concluded: “Our results suggest that no matter how profitable a trading result has been in the past, future performance may be strongly affected by how well known and popular the trading strategy becomes.”
Economists Dwight Lee and James Verbrugge once offered investors this insight into market efficiency: “The efficient market theory is practically alone among theories in that it becomes more powerful when people discover serious inconsistencies between it and the real world. If a clear efficient market anomaly is discovered, the behavior (or lack of behavior) that gives rise to it will tend to be eliminated by competition among investors for higher returns.”
In his wonderful new book, “Adaptive Markets,” economics professor Andrew Lo added this important insight: While the market may never be completely efficient, it certainly does have a tendency to become more efficient over time as anomalies that had previously generated significant profits stop making money.
The Bollinger Bands method had received growing attention since its introduction in 1983 in the U.S. and, in particular, since publication of the book “Bollinger on Bollinger Bands” in 2001.
Associated with this growing popularity, Fang, Jacobsen and Qin found gradual downward profitability of using Bollinger Bands in international stock markets. Since 2002, the method has largely lost its predictive ability in major stock markets.
The authors concluded: “Our results indicate the impact of investor overuse on the profitability of a useful trading strategy and warn of the danger of investing in many so-called return predictability anomalies.”
The bottom line is that the profitability of the Bollinger Bands method never met the five criteria you should require before investing in a strategy. While it may have been profitable and persistent, there wasn’t any risk-based explanation, or any limits to arbitrage to prevent other investor from eliminating profit opportunities.
This commentary originally appeared November 1 on ETF.com
By clicking on any of the links above, you acknowledge that they are solely for your convenience, and do not necessarily imply any affiliations, sponsorships, endorsements or representations whatsoever by us regarding third-party Web sites. We are not responsible for the content, availability or privacy policies of these sites, and shall not be responsible or liable for any information, opinions, advice, products or services available on or through them.
The opinions expressed by featured authors are their own and may not accurately reflect those of the BAM ALLIANCE. This article is for general information only and is not intended to serve as specific financial, accounting or tax advice.
© 2017, The BAM ALLIANCE