Swedroe: Are Dividends A Value Strategy?

It's worth re-examining whether high-dividend strategies really are good value strategies.

As I’ve previously discussed, the Federal Reserve’s zero-interest-rate policy has “pushed” many investors—especially those who use a cash-flow approach as opposed to a total-return approach—to look to stocks and equity funds that have a high dividend yield; that is, that have a low price-to-dividend ratio.

Adding to their attraction is that a high-dividend strategy has outperformed the market over the long term. And this has been true around the globe.

However, it’s long been known that a high-dividend strategy is basically a value strategy, one that correlates highly with value strategies such as buying stocks with low prices-to-book value, earnings, sales or cash flow. The question then is, Is a high-dividend strategy the best, or even a good, value strategy?

One problem with dividend-focused strategies is that because today only about one-third of stocks pay dividends, such a strategy is less diversified than strategies that screen for the other value factors.

In addition, they can lead to concentration in sectors such as utilities, creating other risks, such as term risk. And for taxable investors, dividends are less tax efficient than capital gains because you are taxed on the full amount of the dividend, instead of just the portion that is profit.

Despite these issues, the attraction of current income has led to a heightened interest in these strategies. With that in mind, Gregg Fisher decided to take a close look at whether the focus on high dividends actually added or subtracted from returns. His study, “Dividend Investing: A Value Tilt in Disguise?” appeared in the April 2013 edition of the Journal of Financial Planning.

Fisher’s study covered the 33-year period from August 1979 through July 2012. His objective was to determine which factors best explained stock returns. Using Barra’s performance-attribution methodology, various risk factors such as value, growth, momentum and company size were separated out to determine the contribution of each to the portfolio’s returns.

For the high-dividend strategy, he used the highest-yielding 10 percent of the Russell 3000 Index. The Russell 3000 was used to represent the market portfolio. The following is a summary of his findings:

The high-dividend-yield portfolio’s annualized return was 1.27 percentage points greater than that of the total market portfolio (12.42 percent versus 11.15 percent).

The high-dividend strategy had, as should be expected, a high loading on yield: 1.60. Other loadings of note were leverage (0.42), momentum (-0.22), growth (-0.41), value (0.53), volatility (-0.37) and earnings yield (-0.47).

By tilting the portfolio to high-dividend-paying stocks, investors were—perhaps unknowingly—tilting their portfolios to value stocks.

The factors with the largest contributions to the excess return of the high-dividend strategy were: earnings yield (2.28 percent), volatility (1.22 percent), value (0.41 percent) and growth (0.40 percent).

The dividend yield factor actually detracted from performance, contributing -1.02 percent to the excess return. Other negative contributors were leverage (-0.32 percent) and momentum (-0.21 percent). Note that value strategies typically are negatively correlated with momentum (thus having negative exposure to the momentum premium leads to lower returns when the premium is positive). The bottom line is that it was the value factor, not the yield factor, that was responsible for the excess performance over the period studied. And finally, the dividend-yield factor tilt brought with it a high exposure to the earnings yield factor, a commonly used method for identifying value stocks and that is a strong contributor to positive returns.

Fisher then tested the strategy using a loading of 0.5 on dividends. He found basically the same results: The high dividend strategy outperformed the market by 2.89 percent.

However, the dividend yield provided basically no contribution (was actually a small negative contribution of -0.05 percent). The most positive contribution came from the earnings yield factor (2.28 percent). A small negative exposure to the size factor contributed 0.37 percent. The value factor was the other important contributor (0.28 percent).

Another interesting finding was that when Fisher analyzed the return streams on a monthly basis, he found that the dividend yield factor had a negative contribution to return in 231 months and a positive contribution in 170 months.

In addition, he found the dividend yield factor’s contribution to return to be negative more often than it was positive—and the average downside (0.40 percent) was slightly larger than the average upside (0.35 percent).

Fisher drew this conclusion: “If it is long-term outperformance over the broader market that investors are seeking, findings from this study suggest a more direct approach would be to employ a portfolio tilt toward value and high-earnings-yield stocks.”

I would add this note of caution for those currently using a high-dividend strategy. The popularity of such strategies has driven valuations to high levels—levels that cause such strategies to lose exposure to the value factor that has produced large excess returns.

The table below compares the value metrics of price-to-earnings (P/E) and price-to-book [P/B] of high-dividend ETFs, as opposed to ETFs that focus on companies that have experienced growth of dividends) with that of the SPDR Russell 3000 ETF (THRK | B-100) and the Guggenheim S&P 500 Pure Value ETF (RPV | A-64).

Data: Morningstar as of May 23, 2014

What’s clear from the above data is that while, historically, high-dividend strategies were once value strategies—which produced above-market returns, the popularity of the strategy has driven valuations to be very marketlike. The dividend strategies no longer even look anything like a large value strategy. And it’s valuations that are the best predictor of future returns. Forewarned is forearmed.

When trades get overcrowded, the endings can be ugly. One reason for this is that, as author William Bernstein has pointed out, when a strategy becomes popular, not only will it have low expected returns due to the crowding, but the investors are now “weak hands” that tend to panic at the first sign of trouble. That leads to the worst returns occurring at the worst times, when the correlations of all risky assets move toward 1.

Larry Swedroe is a director of Research for the BAM Alliance, a community of more than 130 independent registered investment advisors throughout the country.