There have only been a handful of truly monumental “discoveries” in the history of investment theory. The three-factor model, developed in 1992 by Nobel Laureate Eugene Fama (University of Chicago) and Professor Kenneth French (Tuck School of Business, Dartmouth College), is on that short list.
Prior to the three-factor model, accepted wisdom was that any given stock’s expected return was determined primarily by just one factor, the stock’s “beta,” or its systematic risk as compared to the overall market. The three-factor model hypothesized that two other factors – a stock’s size and its book-to-price ratio – also meaningfully contribute to its expected return[1].
To say that those two factors contribute to expected return is an understatement. Since 1926, small company[2] (aka “small cap”) US stocks have outperformed large cap stocks by 1.7% per year, while high book-to-price[3] (aka “value”) stocks have outperformed low book-to-price (“growth”) stocks by 3.0% per year. These differences in return are known as the small cap and value “premiums.”
Annualized total return from July 1926 – September 2019[4]
| Small cap stocks | Large cap stocks | Small cap “premium” |
| 11.7% | 10.0% | 1.7% |
| Value stocks | Growth stocks | Value “premium” |
| 12.7% | 9.7% | 3.0% |
These premiums are robust across geographies and time periods
Growth of $1 from July 1926 – September 2019[5]
| Small cap stocks | Large cap stocks |
| $29,256 | $7,083 |
| Value stocks | Growth stocks |
| $69,409 | $5,552 |
Investors disagree as to why these premiums exist: do they represent a market inefficiency, or are they simply reflective of investor risk preference? This distinction is important. If it’s an inefficiency (i.e., the stock’s price does not incorporate all publicly-available information), then we should expect the anomaly to disappear as it becomes known. If instead it reflects investors’ collective risk preferences (i.e., all else equal, investors prefer to own large cap stocks to small and prefer growth stocks to value), then the premiums are simply the compensation that investors demand in order to be willing to own riskier (or otherwise less desirable) stocks, in which case we would expect the premiums to persist. (Note that the difference in risk also explains why stocks would be expected to outperform bonds, a phenomenon known as the “equity risk premium.”)
As with most good debates, the truth almost certainly lies somewhere in between. We at Geometric are believers in the efficiency of the markets, so we side more with the risk preference argument. For what it’s worth, so do Professors Fama and French.
Because we think it’s a good bet that the premiums will persist going forward, we choose to “tilt” our clients’ portfolios – and our own, of course – towards small cap and value stocks. This means that while we still “own the market” by holding essentially every public stock in the world (via mutual funds or ETFs), we modestly overweight[6] small cap and value stocks and underweight large cap and growth stocks.
As an oversimplified example, Microsoft – a large cap growth company if there ever were one – represents roughly 3.6% of the total US stock market, but only 2.4% of our clients’ US equity portfolios. The remaining 1.2% is used to overweight the universe of US small cap and value stocks.
Historically, a portfolio with this degree of tilting would have returned 11.0% per year, as compared to 9.8% for the total US stock market[7]. The difference between the two (1.2%) represents the portion of the small cap and value premiums that we hope to capture for our clients.
Dimensional Fund Advisors was founded in part to allow advisors to tilt portfolios – and thereby capture these premiums – as efficiently as possible. It’s a big part of the reason we prefer Dimensional to traditional index funds.
But while investment theory is clean, the real world is messy.
Just as stocks can underperform bonds for extended periods (i.e., the equity risk premium can be negative in any given 1-, 5-, or even 10-yr period), the small cap and value premiums are often negative in spurts as well.
This short-term underperformance is a feature and not a bug. If stocks never underperformed bonds, then investors wouldn’t require the added compensation to own them and the premium wouldn’t exist in the first place. The same is true for small cap and value stocks: if the premiums were never negative, there would be no premiums.
Indeed, the small cap and value premiums have been negative over the past ten years (ending in 2018), particularly the most recent two:
Annualized total return for last ten years, 2009 – 2018[8]
| Small cap stocks | Large cap stocks | Small cap “premium” |
| 13.1% | 13.5% | (0.4%) |
| Value stocks | Growth stocks | Value “premium” |
| 11.0% | 15.1% | (4.1%) |
Annualized total return for last two years, 2017 – 2018[9]
| Small cap stocks | Large cap stocks | Small cap “premium” |
| 1.3% | 8.3% | (7.0%) |
| Value stocks | Growth stocks | Value “premium” |
| 0.4% | 13.9% | (13.5%) |
Just as it would be a mistake to abandon stocks during periods when they underperform bonds, it’s important to stay committed to small cap and value stocks even during bad stretches, as the periods that follow are often when premiums are at their greatest. Analyzing all of the ten-year periods since 1927 in which the small cap premium was negative, the average premium over the following ten-year period was a healthy 4.6%. For the value premium, it was an incredible 8.3%![10]
So what would make us abandon this strategy? As evidence-based investors, we follow the data. The evidence in peer-reviewed academic literature strongly favors tilting portfolios towards small cap and value stocks. If the supporting evidence ever changes, we would be forced to reevaluate.
Few investing concepts are both as potentially powerful and as academically-supported as tilting towards small cap and value. But like so much else in investing, the benefits from tilting don’t come free – there will always be periods that test investors’ resolve. Only those with the discipline to stay the course get to earn the potential long-term premiums.
————————————————————————————————————
[1] Note that Fama and French have more recently enhanced their model to include two additional factors – “Profitability” and “Investment.” Those factors are also incorporated into Dimensional Fund Advisors’ mutual fund and ETF construction but are beyond the scope of this article.
[2] Fama and French define a small company as one whose market capitalization is in the bottom 10% of the total market.
[3] Fama and French define a value company as one whose book-to-market ratio is in the top 30% of the total market.
[4] Annualized total return data sourced from Dimensional Fund Advisors’ Returns Web program.
[5] Ibid.
[6] Geometric’s preferred US equity mutual fund/ETF (DFA US Core Equity 2) has historic “loading factors” of 0.21 on the small cap coefficient and 0.17 on the value coefficient. The coefficients (expressed as percentages) approximate how much more exposure to each factor (small cap and value) Core 2 has relative to a market-capitalization-weighted total US stock market fund. While there is no “correct” level of tilting, we believe these coefficients balance our desire to capture the potential long-term premiums without creating a portfolio that would deviate too far from the overall market.
[7] Annualized total return data sourced from Dimensional Fund Advisors’ Returns Web program for the CRSP Total Market Index and Dimensional’s Adjusted Market 2 Index, which is a historical index that has the same tilts as DFA US Core Equity 2. Data were pulled for the longest overlapping time frame – June 1, 1927 through September 30, 2019.
[8] Annualized total return data sourced from Dimensional Fund Advisors’ Returns Web program.
[9] Ibid.
[10] Using monthly return data from June 1927 to December 2018, annualized premiums are calculated over the subsequent 10-year periods following a negative 10-year small cap or value premium. The average is calculated across all such 10-year periods for each premium.