Small Cap Banner

The Small Cap Tilt: Is It Still a Reliable Source of Higher Returns?

Small Cap Banner

There is no arguing the fact that 2014 has been a difficult year for small cap stocks around the world (with the exception of emerging markets) and particularly in the U.S., as shown in the table below:

2014 Returns of IFA Indexes
As of 12/31/2014
Region Large Cap Equities Small Cap Equitites Small Minus Large
United States 13.53% 4.44% -9.09%
International Developed -5.24% -6.30% -1.06%
Emerging Markets -1.71% 3.00% 4.71%

For the U.S., 2014 has been the worst year for small cap stocks (relative to large cap stocks) since 1998. Recently, we have seen a few articles claiming to have “busted the myth” of the size premium. However, as we specifically demonstrated in this article, the size premium meets the five qualifications of a risk premium shown below:

It is crucial to understand that when we say size is a risk factor, we mean that it is an explanatory factor of equity returns. It is an equally valid explainer of returns in years when large outperforms small as it is in years when small outperforms large. Furthermore, when we say that there is a size premium, we are referring to the fact that in the very long-term data, small cap stocks have delivered higher returns than large cap stocks. In this article, we will focus on the long-term data showing that the size premium is alive and well in most segments of the market. While it is true that the magnitude of the size premium has diminished in the years after it was first documented by Rolf Banz in 1981, we do note that this decrease is mostly explained by extreme small growth which stocks have delivered much lower returns than we would have expected from the Fama/French multifactor models, but DFA (and hence IFA) has excluded those stocks for the past few years. The table below illustrates the small growth anomaly in the post-Banz period.

Difference of Fama/French Factors
34 Years (1/1/1981 to 12/31/2014)
  Average Difference t-statistic
Small Value - Large Value 3.91 1.94
Small Neutral - Large Neutral 3.68 2.02
Small Growth - Large Growth -3.43 -1.50

Furthermore, once the small cap and large cap growth segments are adjusted to reflect the impact of tilting towards higher profitability stocks, the average difference becomes positive.

The bar charts below show the annual returns of the Fama/French research factors for 88 calendar years from 1927 to 2014.

To emphasize the point that small cap value stocks have a higher expected return than large cap growth stocks, we created the bar chart below.

Fama and French obtained a quantitative measurement of the size premium (SmB = Small Cap minus Big Cap) by sorting the US equity market into deciles based on size and assigning the five smallest deciles to the “small cap” classification. For the 88 calendar years from 1927 to 2014, the average value of SmB was 2.94% with a standard deviation of 12.64%. When we apply a statistical test (the t-test) to determine if chance alone could explain the magnitude of SmB, we find that the probability of that occurrence to be less than 5% (t >2). Furthermore, if we break up the 88-year period into the two pieces before and after the publication of the Fama/French paper1, we find that the average value of SmB was 3.24% in the 1927-1992 period and 2.05% in the 1993-2014 period. Contrary to the expectation that the small cap premium would disappear once it was publicly exposed due to numerous investors acting on it and thus driving up the prices of small cap stocks, the small cap premium actually persisted, although not as strongly.

A fair question is how long an investor must wait to be reasonably certain of capturing the size premium. The table below based on calendar year rolling periods shows that as the length of the holding period increases, so does the frequency with which small caps outperformed large caps.

Fama/French SmB (Small Minus Big) Research Factor
88 Years (1/1/1927 to 12/31/2014)
  Total Number
of Periods
Number of Periods Where
Small Caps Beat Large Caps
Percentage of Periods Where
Small Caps Beat Large Caps
1-year 88 49 56%
5-year 84 53 63%
10-year 79 54 68%
15-year 74 57 77%
20-year 69 59 86%

One very important aspect of the small cap premium is the standard deviation of about 13%. A standard deviation of this size means that it would not be uncommon for small cap stocks to lag large cap stocks by at least 10% in a single year (a one standard deviation event which has occurred in 10 of the last 88 years, including 2014). The two standard deviation event of small cap stocks lagging by 23% or more has only occurred twice (1929 and 1973) in the last 88 years.

One of the arguments made by the deniers of the size premium is that if we were to exclude a few months from the 1930s when the returns difference was very large, then the statistical significance of the premium would disappear. Our response is that if we begin the computation of the premium in 1935 (after the abnormally high months) and continue it to the present (12/31/2014), we get a t-stat above 2, implying statistical significance.  

Another argument is that the size premium is due to a reporting bias that does not properly account for delisted stocks. We disagree with this assertion, as we have been reassured by senior researchers at Dimensional Fund Advisors that the Fama/French Research Factor Returns use the CRSP delisting returns which are explained as follows:

“The Delisting Return is calculated by comparing the security's Amount After Delisting with its price on the last day of trading…If there is evidence that no distributions will ever be paid to shareholders, then the stock is considered worthless. The delisting return is set to -1 (i.e. a 100% loss).”

Another argument is that the size premium does not survive trading costs, but as the data below shows, the DFA U.S. Small Cap Portfolio (DFSTX) has captured 95% of the returns of the lowest five size deciles of the market.

Annualized Average Returns
22 Years and 9 Months (4/1/1992 to 12/31/2014)
DFA U.S. Small Cap Portfolio 11.06%
CRSP Deciles 6-10 Index 11.66%

Further ammunition against this argument is provided by the DFA U.S. Micro Cap Portfolio (DFSCX) which actually has obtained a higher return than the lowest two size deciles of the market.

Annualized Average Returns
33 Years (1/1/1982 to 12/31/2014)
DFA U.S. Micro Cap Portfolio 12.32%
CRSP Deciles 9-10 Index 11.75%

Another common argument against the size premium is that since small cap stocks are more volatile than large cap stocks, the premium is simply compensation for acceptance of the higher volatility. In other words, on a risk-adjusted basis, small cap stocks provide no advantage over large cap stocks. Our response to that argument is that even the skeptics of the size premium acknowledge that small cap stocks provide a diversification benefit to large caps, as shown in the chart below (please note that the 50/50 allocation shown below is for illustrative purposes only and should not be taken as an investment recommendation).

As we have stated in prior articles, the question an investor should ask is not "should I be in large caps or small caps?" but rather "what is a sensible blend of large and small companies that will allow me to capture a risk premium without being tempted to sell out of my portfolio when it performs differently from the Dow Jones Industrial Average or the S&P 500?" Despite a disappointing 2014, an exposure to small cap stocks is still a reliable source of higher returns.

1Fama, Eugene F. and Kenneth R. French. 1992. “The Cross-Section of Expected Stock Returns.” Journal of Finance (47):427-465.