Recency bias—the tendency to give too much weight to recent experience and ignore long-term historical evidence—underlies many of the mistakes commonly made by investors. It’s particularly dangerous because it causes investors to buy after periods of strong performance (when valuations are high and expected returns low) and sell after periods of poor performance (when valuations are low and expected returns high).
Recency can lead even investors with well-developed plans to abandon them. And it can also lead to other mistakes, such as overconfidence and the penchant to treat unlikely outcomes as impossible. It’s one of the reasons studies have found that investors tend to underperform the very mutual funds in which they invest.
Recency & Value
A great example of the recency problem involves the performance of value stocks. For the nine years from 2007 through 2015, the value premium (the annual average difference in returns between value stocks and growth stocks) was -3.4%. Cumulative underperformance for the period was 28%.
This type of underperformance often leads to selling. Unfortunately for investors who sold, 2016 turned out to be a year in which the value premium turned strongly positive (20.7%), although it turned negative again in the first 10 months of 2017 (-11.2%).
Investors who know their financial history understand that this type of what we might call “regime change” is to be expected. In fact, even though the value premium has been quite large and persistent over the long term, it’s been highly volatile. The annual standard deviation of the premium, at 12.7%, is 2.7 times the size of the 4.7% annual premium itself (for the period July 1926 through October 2017).
As further evidence, the value premium has been negative in 37% of years since 1926. Even over five- and 10-year periods, it has been negative 22% and 14% of years, respectively. Thus, periods of underperformance like the one we have seen recently should not come as any surprise.
In fact, they should be expected (the only thing we don’t know is when they will pop up), because periods of underperformance occur in every risky asset class and factor.
The following table, taken from my latest book, “Your Complete Guide to Factor-Based Investing,” which I co-authored with Andrew Berkin, provides evidence demonstrating this point.
Each of the factors in the table has been shown to have an economically large, persistent and pervasive premium. In addition, there are intuitive risk- or behavioral-based explanations for why we should believe the premiums are likely to persist in the future, and they are robust to various definitions as well as implementable (meaning they survive trading and other costs).
But each of them, including market beta, also experience long periods of underperformance. That is a good reason to diversify risk across factors rather than concentrating it in any single factor, including market beta.
However, a long period of underperformance should not cause investors to abandon a well-developed plan. Nor should it cause them to question the existence of the value premium any more than it should have caused them to question the market beta premium when it was negative for 3% of the 20-year periods from 1927 through 2016, as the table shows.
As I point out in my book, “Think, Act, and Invest Like Warren Buffett,” one of the great ironies today is that while investors idolize Buffett, many not only ignore his advice but tend to do exactly the opposite of what he recommends (like never trying to time the market).
For example, we know Buffett did not abandon his belief in the value premium following the 10-year period ending in 1999, when it posted an annual average return of just 0.5% and produced a cumulative return of -5.2%. And I suspect he has not abandoned his faith in it given its recent relatively poor performance. While this surely is the case, many investors still question the continued existence of the value premium.
Why Value Still Matters
There are various reasons you should continue to expect an ex-ante value premium. The first is that risk cannot be arbitraged away. And the research provides us with many simple and intuitive risk-based explanations for the persistence of the value premium, as I have written about before.
Second, if, as many people believe, the publication of findings on the value premium has led to cash flows that have caused it to disappear, we should have seen massive outperformance in value stocks as investors purchased value stocks and sold growth stocks. Yet the last 10 years have witnessed the reverse in terms of performance.
In addition, as David Blitz demonstrated in his February 2017 paper, “Are Exchange-Traded Funds Harvesting Factor Premiums?”, while some exchange-traded funds are specifically designed for harvesting factor premiums, such as the size, value, momentum and low-volatility premiums, other ETFs implicitly go against these factors.
Specifically, Blitz found that “from a factor investing perspective, smart-beta ETFs tend to provide the right factor exposures, while conventional ETFs tend to be on the other side of the trade with the wrong factor exposures. In other words, these two groups of investors are essentially betting against each other.”
The bottom line is that, despite what many investors believe, there has not been a massive net inflow into value stocks relative to growth stocks.
Third, academic research has found that valuation metrics, such as the earnings yield (E/P) or the CAPE 10 earnings yield, and valuation spreads have predictive value in terms of future returns. In other words, the higher the earnings yield, the higher the expected return, and the larger the spread in valuations between growth and value stocks, the larger the future value premium is likely to be—and it holds across asset classes, not just for stocks.
For example, the 2007 study “Does Predicting the Value Premium Earn Abnormal Returns?” by Jim Davis of DFA found that, despite book-to-market ratio spreads containing information regarding future returns, style-timing rules did not generate high average returns because the signals are “too noisy” (they don’t provide enough information to offer a profitable timing signal).
The October 2017 study “Value Timing: Risk and Return Across Asset Classes,” authored by Fahiz Baba Yara, Martijn Boons and Andrea Tamoni, offers further support that valuation spreads provide information.
The authors found that “returns to value strategies in individual equities, commodities, currencies, global government bonds and stock indexes are predictable by the value spread …. In all asset classes, a standard deviation increase in the value spread predicts an increase in expected value return in the same order of magnitude (or more) as the unconditional value premium.”
Is Value Still Value?
Given that valuation spreads have been shown to have predictive value, we can examine current valuation spreads to see if they have shrunk in a way that would be expected to eliminate the value premium.
As a simple test, using data from Ken French’s website, I took a look at the spread in book-to-market (BtM) values of the top and bottom three deciles in 2008 and compared them to where they were at the end of October 2017, the latest data available. This should tell us whether cash flows over the last 10 years have altered the very nature of the value premium.
The methodology used to compute BtM ratios is this: Take the book value of the stocks as of December 31 of the prior year and then the market value of the stock as of June 30 of the following year. For 2008, the BtM ratios were 0.21 (deciles 1-3) and 0.92 (deciles 8-10), for a value spread of 0.71. For 2017, the BtM ratios were 0.16 and 0.83, respectively, for a value spread of 0.67—virtually unchanged from 10 years earlier.
Note that the value premium from 1926 through 2007 was 5.6%. Thus, at least in terms of valuation spreads, there doesn’t seem to be any evidence to support the idea that the value premium has disappeared.
To help you avoid mistakes involving recency, keep the following example handy. The stock risk premium has been large, almost 8% a year. However, it’s also highly volatile, with an annual standard deviation of about 19% (about 2.5 times the size of the premium itself).
The premium is large because there’s a large amount of risk involved in equity investing, and investors demand it to take that risk. Because it’s a risk premium, the chance that investors may experience very long periods of underperformance must exist. In fact, the premium may never be realized. If such a risk did not exist, there really wouldn’t be any risk (and thus no premium).
As proof, consider that from 1969 through 2008, U.S. large-cap growth stocks returned 7.8% and underperformed long-term (20-year) Treasury bonds, which returned 9.0%. That’s a 40-year period in which investors took all the risks of stocks and underperformed long-term U.S. Treasuries. Should that have convinced investors the strategy of believing in a stock risk premium was wrong? Of course not.
The logic is still the same. Stocks are riskier and must have higher expected returns. It’s just that the risk involved showed up for this very long period. Those who abandoned their plans and sold stocks because they confused strategy and outcome may have missed out on the bull market that followed—the greatest since the 1930s.
What’s important to understand is that the premiums for the market overall, small stocks and value stocks have been earned only by those investors disciplined enough to stay the course through the periods when the asset classes (and factors) they have invested in underperform.
As we have seen, the periods can be quite long, long enough to test even the most disciplined of investors. That is, perhaps, why Warren Buffett has stated that his favorite holding period is forever. He has also said that successful investing has far more to do with temperament than intellect.
Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.