Dimensional vs. Indexing
Re-posted from Swedroe: Value Premium Lives!
There has been much debate about the “death” of the value premium. Those making the case it has disappeared believe the publication of research on the premium has led to cash flows, which in turn have eliminated it.
They point to the last 10 calendar years of data, during which the value premium was slightly negative, at -0.8%. The value premium (HML, or high minus low) is the annual average return on high book-to-market ratio (value) stocks minus the annual average return on low book-to-market ratio (growth) stocks.
Before taking a deeper dive into the data, it’s important to note that all factor premiums, including market beta, have experienced long periods of negative returns. The following table, covering the period 1927 through 2017, shows the odds (expressed as a percentage) of a negative premium over a given time frame. Data in the table is from the Fama/French Data Library.
No matter the horizon, the value premium has been almost as persistent as the market beta premium. Even the market beta premium has been negative in 9% of 10-year periods and in 3% of 20-year periods.
Investors who know their financial history understand that what we might call “regime change,” with value underperforming for a fairly long time, is to be expected. In fact, even though the value premium has been quite large and persistent over the long term, it has been highly volatile. According to Ken French’s data, the annual standard deviation of the premium, at 12.9%, is 2.6 times the size of the 4.8% annual premium itself (for the period 1927 through 2017).
While 10 years may seem like an eternity to most investors (even institutional investors fire managers after a few years of underperformance), it can be nothing more than noise. Having to remain disciplined over such long periods is perhaps why Warren Buffett famously said that investing is simple, but not easy.
As Andrew Berkin and I point out in our book, “Your Complete Guide to Factor-Based Investing,” investors should have confidence in the value premium because it has been persistent across long periods of time and various economic regimes and has been pervasive around the globe and across asset classes. In addition to stocks, a value premium has existed in bonds, commodities and currencies (in other words, buying what is cheap has outperformed buying what is expensive).
Furthermore, the value premium has been robust to various definitions, not just the price-to-book (P/B) measure, or HML, popularized by Fama and French. Indeed, the premium is also found when using other value metrics, such as price-to-earnings (P/E), price-to-cash flow (P/CF), price-to sales (P/S), price-to-earnings before interest, taxes, depreciation and amortization (P/EBITDA), price-to-dividend (P/D) and price to almost anything—however it’s measured, cheap has outperformed expensive.
With that in mind, we can look at the size of the value premium for the 10-year period ending 2017 against some of those other metrics (all data is from Ken French’s website). While the value premium was negative by P/CF (-1.2%) and slightly negative by P/B, it was actually positive using P/E (1.2%) and P/D (0.6%)—so, on average, about zero.
Before moving on, it’s important to note many have questioned the sole use of P/B as the best value metric, especially in an age when many assets are now intellectual capital that often don’t appear on balance sheets (externally acquired intangibles do appear on balance sheets).
Reasons not to believe P/B—or for that matter, any single metric—is the superior measure of value include the fact that one metric could work better than others do for some industries, and various metrics have provided the highest premium in different economic regimes. That is why many practitioners and mutual fund families now use multiple metrics, which I believe can provide a diversification benefit.
Having covered that ground, let’s go to our trusty videotape and see how value performed outside the U.S. If value is “dead,” we should find confirming evidence in other markets.
The following table—again, using data from Ken French’s website—shows the premium for the 10 years from 2008 through 2017 in international developed markets.
|2008-2017||Average Annual Premium (%)|
No matter which metric I use, there was a value premium in developed international markets. This result actually surprised me, because research shows value stocks are more exposed to economic cycle risks—they perform best in periods of stronger economic growth because they become less risky. While the last 10 years saw the slowest economic recovery in the post-World War II era, economic growth has been even more anemic in the non-U.S. developed world.
Having shown the value premium in international markets has been fairly similar to historical levels, I can also examine live results.
Specifically, I’ll compare the returns of the passively managed, structured international value funds my firm uses with those of marketlike ETFs in three asset classes. Data is from Morningstar and covers the 10-year period ending May 4, 2018. (Full disclosure: My firm, Buckingham Strategic Wealth, recommends Dimensional funds in constructing client portfolios.)
|iShares MSCI EAFE ETF (EFA)||2.25%|
|DFA International Value Portfolio III (DFVIX)||2.20%|
|iShares MSCI EAFE Small-Cap ETF (SCZ)||6.05%|
|DFA International Small Value (DISVX)||5.89%|
|iShares MSCI Emerging Markets ETF (EEM)||1.29%|
|DFA Emerging Markets Value (DFEVX)||1.66%|
Note: Dimensional’s international strategies include Canada; EAFE indexes do not.
The average return of the three ETFs was 3.20%. The average return of the three Dimensional value funds was 3.25%. Thus, while there was a value premium, compound returns were no better (as the roughly 2-3% higher annual volatility of value stocks negatively affected their compound returns), but no worse, either.
One way to look at the results is that, over one of the worst 10-year periods for the performance of value stocks, value investors were not “penalized” for their strategic decisions. Another is to believe the value premium is dead.
My view is that there is a large body of evidence demonstrating logical, risk-based explanations for the value premium. While cash flows can shrink premiums, risk-based premiums cannot be arbitraged away. On the other hand, it’s possible to arbitrage away behavioral-based premiums, such as momentum (though limits to arbitrage often prevent this from occurring).
Among the risk-based explanations for the premium are that value stocks contain a distress (default) factor, have more irreversible capital, have higher volatility of earnings and dividends, are much riskier than growth stocks in bad economic times, have higher uncertainty of cash flow, and, as previously mentioned, are more sensitive to bad economic news.
Given these risk-based explanations, it is hard to believe the value premium can be arbitraged away. With that understanding, let’s consider whether cash flows have eliminated the premium.
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 those equities and sold growth stocks.
Yet the last 10 years have witnessed the reverse in terms of performance. In addition, if cash flows had eliminated the premium, the buying of “undervalued” value stocks and selling of “overvalued” growth stocks should have led to a reduction in the valuation spreads between value stocks and growth stocks.
I happen to have kept a table from a seminar Dimensional gave in 2000. It shows that, at the end of 1994, the P/B ratio of large growth stocks was 2.1 times as big as the P/B ratio of large value stocks.
Using Morningstar data, as of May 3, 2018, the iShares S&P 500 Growth ETF (IVW) had a P/B ratio of 4.7, and the iShares S&P 500 Value ETF (IVE) had a P/B ratio of just 2.0—the spread has actually widened from 2.1 to 2.4. Thus, value stocks are cheaper today, relative to growth stocks, than they were shortly after Fama and French published their famous research.
We can also look at the P/E metric. In 1994, according to the Dimensional table, the ratio of the P/E in large growth stocks relative to the P/E in large value stocks was 1.5. As of May 3, 2018, and again using Morningstar data, IVW had a P/E ratio of 20.0, and IVE had a P/E ratio 14.4. Thus, the ratio, at 1.4, was almost unchanged. There’s no evidence here that cash flows have eliminated the premium.
We see similar results when we look at small stocks. The Dimensional data shows that, at the end of 1994, the P/B of the CRSP 9-10 (microcaps) was 1.5 times as large as the P/B of small value stocks. Using Dimensional data, and the firm’s microcap fund (DFSCX) for microcaps and its small value fund (DFSVX) for small value stocks, as of April 30, 2018, the ratio of the funds’ respective P/B metrics was the same 1.5 (1.9÷1.3). When we look at P/E, again, the results are similar. At the end of 1994, the ratio of those funds’ respective P/E metrics was 1.2; it is now the same 1.2 (20.2÷16.3). Again, I see no evidence that cash flows have eliminated the premium.
Using data from Ken French’s website, we also see similar results when we look at the period starting in 2008. In 2008, the ratio of the P/B of U.S. growth stocks (4.8) to U.S. value stocks (1.1) was 4.4. By the end of 2017, the ratio had increased to 5.3 (6.3÷1.2), the opposite of what you would expect if cash flows had eliminated the premium.
Based on the logical, risk-based explanations for the value premium, and the lack of evidence pointing to shrinking valuation spreads, my conclusion is that the most recent 10 years of performance is likely just another of those occasionally occurring but fairly long periods in which the value premium is negative. If periods such as these did not occur, there would be no risk, and no risk premium. This is true of all risky investments, including stocks (and thus the market beta premium).
The bottom line is that, at least in my opinion, it’s hard to conclude the value premium is dead. Being able to stay the course during long periods of relative underperformance is what led Warren Buffett to assert that successful investing has a lot more to do with temperament (meaning discipline and patience) 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.