A Comparison of Value Strategies
November 27, 2012
Dividend-focused strategies have won the allegiance of many prominent investors, including Rob Arnott, Bill Gross and Jeremy Siegel. Others claim value-based strategies offer superior risk-adjusted returns. Both sides can claim a partial victory in this debate, but I will show that, when understood properly, dividend strategies offer a crucial edge – one that many investors will find attractive.
There are a variety of reasons why investors may concentrate on dividends when choosing equities, including:
- A preference for income, rather than having to sell assets
- A belief that dividend-paying stocks will outperform non-dividend stocks
- A belief that stocks that pay dividends are less risky that stocks that do not
- A belief that dividends have historically comprised a substantial component of total return
Implicit in all of the above is the idea that dividends are a more consistent form of returns than price appreciation. Research backs up that conclusion. Dividend-paying companies have been shown to be less likely to play accounting games with their earnings (a practice sometimes referred to, generously, as “earnings management”) and have been shown to have higher-quality earnings.
Stock Investors buy a fraction of the entire future stream of income generated by a company. If earnings are more predictable and consistent, total returns should be steadier as well.
But the key differentiator is another special feature of dividend stocks – reduced estimation risk, the risk associated with the difficulty of estimating future returns. In a 2011 article, I described research that demonstrates how dividend stocks lower the estimation risk for the total return of a portfolio. Today, I’ll clarify how this particular benefit of dividend investing differentiates it from other value strategies.
Let’s explore some of arguments posed by critics and proponents of dividend-focused investing, in order to understand why neither side of the argument is capturing the full story. Then we’ll see why estimation risk is the missing piece of this discussion.
Is a portfolio selected on dividend yield sub-optimal?
A recent article by Michael Nairne of Toronto-based Tacita Capital compared the performance of portfolios selected on the basis of high-dividend yield to other value-oriented strategies. Using data from 1952 through 2011, courtesy of Ken French’s website, Nairne found that, while high-dividend stocks outperformed the S&P 500 by 1.8% per year over the period studied, other value strategies performed considerably better. Portfolios chosen on the basis of high earnings-to-price ratio (E/P), in particular, returned 4.9% per year more than the S&P 500 and 2.7% more than the high-dividend-yield portfolio. Nairne’s results are consistent with a 2010 analysis by Buckingham Asset Management (BAM) that used the same data set (though with data through 2009, rather than 2011). I confirmed Nairne’s results as well.
Nairne and BAM concluded that portfolios built on the basis of high-dividend yields are a sub-optimal value strategy and that others (especially a low P/E strategy) are superior.
I see a couple of important caveats for anyone who plans on extrapolating these results forward, however.
First, neither Nairne nor BAM noted that the high-E/P portfolio has a beta (with respect to the S&P 500) of 1.10, while the high-dividend-yield portfolio’s beta is only 0.90. This is a substantial difference that goes a long way towards explaining the apparent outperformance of the high E/P strategy. The beta-corrected outperformance of the high-dividend portfolio is 3.2% per year, versus 3.8% for the high-E/P portfolio. Thus, in practical terms, the apparent advantage for the high-E/P strategy is just 0.6%.
So the relative under-performance of dividend-based portfolios compared to high-E/P portfolios and other value strategies is not fully explained by the portfolios’ beta. There is a missing element if we are to fully explain away the underperformance of dividend-based stocks relative to other value strategies.
As we’ll see, that missing element is reduced estimation risk.
Would you like to send this article to a friend?Remember, if you have a question or comment, send it to .