Defining the Value in Value Stocks

Value is in the eyes of the beholder. Every value stock investor has a different opinion of what value is, and there is no right answer. Everything works sometimes, and nothing works all the time.

There is also no consensus among index providers of what a value stock is or how portfolios of value stocks should be formed. Index strategies range from single value factors, such as low price-to-earnings or low price-to-book, to complex multi-factor formulas that incorporate many variables.

The data in this article shows that in the short-term, the returns using different value strategies can vary widely, but in the long-term, they tend to regress to a single mean. The exception is high dividend yielding stocks because they show weaker value characteristics than other factors.

I used Ken French’s Data Library to form the tables and figures in this article. The database includes return information on four value factors. The four factors are high earnings-to-price (E/P), high book-to-market (BtM), high cash flow-to-price (CF/P) and high dividend-to-price (D/P). Gene Fama and Ken French update the research data at least once a year. The latest update was July, 2011.

U.S. stocks in the database are sorted based on the four value factors. Table1 and Table 2 reflect the excess returns of portfolios formed on the high 30 percent of stocks from each of the four factors. It is worth noting that the database provides the return of concentrated portfolios. It would not be possible to form these portfolios in real time because of trading costs, liquidity constraints and high turnover. Accordingly, the excess returns shown below are probably about twice as high as what could be achieved in an actual portfolio.

Table 1 highlights the annualized 5-year and 10-year returns of U.S. stocks along with the excess return of the four concentrated value factors. The Total Market is the performance of the CRSP 1-10 index. The boxes with green shading were the factor with the highest excess return over the market each year, and the column with the red shading had the lowest excess return (sometimes negative).

Table 1: Five Year Annualized Market Return and Excess Factor Returns

Table 2: Returns by Decade (ending in ’09 of each period)

Returns by Decade

There was a return premium from all value factors in most periods, with the notable exception being in the second half of the 1990s. Table 2 shows that high E/P stocks outperformed the other factors during 3 out of 5 decades. High D/P stocks underperform the other factors in 4 out of 5 decades.

All four value strategies outperformed the market over the entire 60 year period. Figure 1 illustrates the cumulative return of the U.S. stock market and the cumulative returns from investing in the four value strategies. The notable lagging value strategy was high dividend yield, although the cumulative excess return was still quite high relative to the market.

Figure 1: Cumulative Returns from July 1951-July 2011

Figure 1 suggests that CF/P, E/P and BtM performed better than dividends and significantly better than the market over the 60-year period. It also shows that all three strategies resulted in near identical performance in the long-term. Consequently, it didn’t matter which of the three factors was used to achieve a value portfolio, assuming an investor was consistent.

Investment trends go in and out of favor, and factor preferences are no exception. Sometimes investors prefer high earnings-to-price (the inverse of low P/E) and other times they focus on cash flow or book value. I thought it would be interesting to map these changing preferences over time to see if there are patterns or trends.

Figure 2 illustrates the 12-month rolling excess returns of CF/P, E/P and BtM relative to the average of all three. When a factor line is above zero, it is outperforming the average of all three, and when it is below zero, that factor is underperforming the average.

Figure 2: Rolling 1-Year Relative Performance of Value Factors

To be clear, Figure 2 is not showing factor performance relative to the total market. It is showing each factor relative to the average of the three factors (the 0% line). A single factor may be outperforming the other two at the same time all three factors are underperforming the total market. This paper does not consider how each factor performed during bull and bear markets.

There is a lot of volatility in factor returns. High BtM stocks significantly outperformed both CF/P and E/P stocks during 2003 and into 2004, however this trend reversed sharply in late 2007 and 2008 when BtM underperformed. In 2011, BtM is again underperforming as investors have focused on stocks with high cash-flows-to-price.

Every value index fund has a different way to select value stocks. Some methods use one single factor while others use multiple factors. When you research which factors are used in an index, and the weight given to those factors, it helps explain the performance of the index.

Gene Fama and Ken French prefer to use BtM as their benchmark because book values are the most stable of the factors. This means a portfolio has less turnover. Since Fama and French are directors and consultants to Dimensional Fund Advisors (DFA), a mutual fund firm for advisors, that company uses a single BtM factor when selecting value stocks. It’s the only remaining fund company that does.

S&P was utilizing only BtM in the style indexes until December 2005 when they switched to a multi-factor model. In retrospect, that was pretty good timing. Book value fell out of favor with investors as the financial crisis took hold in late 2007. This was when bank stocks began their long and deep bear market. Bank stocks traditionally trade at low multiples of book value. Consequently, they commanded an overweight position in high BtM portfolios.

Table 3 is a comparison of calendar year returns between the DFA Large Cap Value fund (DFLVX), its peer group of large cap value managers, and the Morningstar Large Cap Value index (a multi-factor index). Since stocks for DFLVX are chosen using only high BtM stocks, when significant differences in BtM performance occur in Figure 2, it should be reflected in DFLVX performance.

Table 3: Comparing DFA Large Value to Peers and the Morningstar Large Value Index*

*Data through September 2011 from Morningstar Principia.

DFLVX outperformed its large cap value peer group and the Morningstar large value index by a wide margin during 2003 and 2010. These were the years when BtM meaningfully outperformed CF/P and E/P. The fund underperformed in 2007 and so far in 2011, the same years the high BtM stocks unperformed the other two factors.

A few observations and conclusions can be drawn from this casual analysis. They may help you decide which value factors to follow and which value funds to choose:

  1. BtM, E/P and CP/P essentially return the same gross performance over the long-term. Accordingly, any value index fund formed with one or more of these factors should perform close to another fund formed using similar factors.
  2. Attempting to time individual factors is futile because there is no autocorrelation in the data. It’s not possible to predict which factor will outperform the others in the short-term.
  3. In the past decade, value investors have enjoyed some of the highest excess returns in the history of the stock market. This superb performance isn’t likely to continue. It doesn’t mean value won’t outperform this decade, that can’t be known. It does mean invest for the long-term.

There are many ways to approach value investing and this article touched on a few. It’s helpful to know the types of value exposures in your index fund, and know that different factors take different return paths in the short-term.

All value fund investors should strive to keep investment costs low, including taxes, and expect long periods of underperformance. Value investing has rewarded those who have stayed the course.