In my original article on a simple dividend strategy, I chose to equally weight the 10 sectors in my model portfolio to make it easy to implement and because it offered a higher yield at a lower beta than a market-cap weighting. In this article, I will highlight the results from S&P’s Dividend Aristocrats and a comprehensive study of dividend paying stocks performed by researchers in the U.K. in 2009. From its title, I was expecting it to provide supporting data for dividend stock strategies, which it did, but there was so much more! I strongly recommend that everyone take a look at the research, as I can’t do it justice in 1,500 words. I am not going to include the full tables from the U.K. research, but I will note the appropriate exhibit as results are discussed. The research also examined several filters that identified additional traits that led to higher returns. I will present those in a separate article, so that this one can stay focused on weighting strategy.
U.K. Study Parameters
The U.K. study examined data from the London Share Price Database (LSPD) from 1975-2006, as the U.K. market has a high-proportion of dividend payers, allowing for well-diversified portfolios. A firm that paid out a greater cash dividend in the most recent 12 months, compared to the prior 12 months, was classified as having dividend growth; assessment of growth occurred at the end of the calendar year. Investment trusts and similar vehicles were excluded. Portfolios were equally weighted and formed at the beginning of the year. If a firm was delisted (acquired or bankruptcy), any residual value was invested in the remaining constituents; rebalancing due to dividend cuts only occurred at the end of the year. The average portfolio size for firms with 5 years of dividend growth was 297, 10 years – 127, and 20 years – 32. As the sizes were over 30, they imply that the benefits of diversification were captured.
click to enlarge images
Market-Capitalization Weighting vs. Equal Weighting
The table above summarizes selected results from Exhibit #1 and Exhibit #2 of the U.K. study, focusing on compound annual rate of growth (CAGR), standard deviation of annualized returns (volatility), and the lowest annual return for the specified time period. Portfolios were created based on minimum years of consistent dividend growth (5-year increments), and then weighted equally (left side results, most in green) and by market capitalization (right side results). Two broad-based market indices were used as benchmarks; they were also weighted using both systems for comparison purposes.
Key highlights are noted below and strongly support the case for equally weighted portfolios.
- The CAGR for market-cap weighted dividend portfolios generally performed slightly worse than the market-cap weighted benchmarks, and their standard deviations were comparable to those of the indices (similar volatility). The study found little advantage to a market-cap weighted consistent dividend portfolio.
- The CAGR for equally weighted consistent dividend portfolios outperformed the benchmarks in all three time period ranges, regardless of the minimum years of dividend growth requirement. This does not mean these portfolios outperformed each year; see the full Exhibit #1 data table for annual results. Over longer time periods though, these portfolios yielded superior returns.
- The equally weighted portfolios also had the lowest average standard deviations (volatility) for each of the three time periods. Therefore, the equally weighted consistent dividend portfolios provided higher absolute and higher risk-adjusted returns compared to the overall market benchmarks.
- The equally weighted portfolios had smaller losses than the market-cap weighted portfolios, as measured by finding the largest decline for each time period range. Most notably in 2002, when the overall markets declined ~22% and a market-cap weighted dividend portfolio (15-20 yrs of growth) declined ~19%, the equally weighted portfolios (15-20 yrs) declined just 2%.
S&P Data Follows U.K. Findings
In its 2008 paper on Dividend Aristocrats, S&P includes the above comparison between an equally weighted portfolio of the current year Dividend Aristocrats, the S&P 500, and an equally weighted S&P 500 index. For each time period, the Dividend Aristocrats outperformed the S&P 500 (market-cap weighted) and the equally weighted S&P 500 portfolio, and did so with lower standard deviation.
For shorter time frames, the market-cap S&P 500 outperformed the equally weighted S&P 500, which differs from the U.K. study results, though it didn’t examine time ranges less than 10 years. For 10- and 15-year ranges, the equally weighted S&P 500 beat the market-cap S&P 500, which is consistent with the U.K. results. While its standard deviation was higher, the equally weighted S&P 500 still delivered better risk-adjusted returns (Sharpe ratio) for the 10 and 15-year periods.
Yield-Weighted Portfolio Results
The table above, excerpted from Exhibit #9 of the U.K. study, calculated the average compound return and standard deviations for a yield-weighted portfolio of stocks with 10 or more years of consistent dividend growth from 1986-2006. It also calculated the results for each quartile based on yield. The overall result (15.43% average return) is comparable to the equally weighted dividend portfolio results, though the volatility is a little higher. Also, it beat the market-cap and equally-weighted index returns, so both equally-weighted and yield-weighted dividend portfolios were found to be “far superior” to the market-cap weighted approach.
The U.K. study concluded that “consistent dividend payers have outperformed the wider market on an equally weighted basis for 1986-2006, particularly when the minimum requirement is set at 10 years of continuous growth.” (ap Gwilym, et. al., 2009, p. 123). This group of stocks also had lower variance of returns, and had smaller drawdowns over shorter time periods.
However, ALL of these benefits disappeared when market-cap weighted portfolios were used instead of equally weighted ones. This may have been due to a loss of diversification, as the universe of consistent dividend payers consisted mainly of large-cap firms; the majority of funds in the market-cap weighting portfolio resided with a few firms. For example, on average, the largest 10 firms in the 10-year portfolio made up 58.4% of the total portfolio market value (ap Gwilym, et. al., 2009, p. 117).
In addition, equally weighted consistent dividend portfolios invest more in smaller-cap dividend paying firms, which typically have higher growth opportunities than larger-cap firms. These portfolios would also invest more in sectors such as utilities or telecom, which make up a smaller portion of the overall market, but provide more stable returns at lower volatility. In my simple dividend portfolio, I went a step further and equally-weighted the sectors, which raised the representation for utilities and telecom even more.
I found it interesting that market-cap weighted dividend portfolio results were essentially equal to overall market results. The study noted that in the U.K., 95% of firms paid dividends in 1979 and 75% in 1999, so this may have limited the differences, as most benchmark firms paid dividends. A recent screen on FinViz.com found that only 3056 of 5953 (51.3%) U.S. firms pay a dividend, so the results could be different in the U.S.
Also, I would have expected lower volatility for the market-cap dividend portfolio compared to the market-cap indices, since the overall market includes more small firms and non-dividend payers that should have higher betas, but this was not the case. Whether the U.K. results fully translate to the U.S. markets is up for debate, but the S&P data provides some support for equally weighted dividend portfolios, as this is the system used by the Dividend Aristocrats index.
Overall, I believe this information is worthy of serious consideration. It validated my initial simple dividend portfolio rationale (use Dividend Aristocrats and equally weight), and is causing me to rethink my core long-term portfolio strategy. The U.K. evidence suggests that the conventional wisdom of classifying dividend stocks for income purposes, as opposed to growth, may be incorrect. Stocks with 10+ consistent years of dividend growth have yielded higher total returns than the overall market with lower volatility. Perhaps we can have our cake and eat it too, if we watch our weightings!
My last planned segment for this series will be on some additional filters investigated by the U.K. study that identified sub-groups of consistent dividend growth stocks that outperformed. I may try to create a second model portfolio based on this research using the Dividend Champions/Challengers lists as the universe. Not quite as simple as using the Aristocrats, but it would be interesting to see how it compares to my SDS-30 portfolio returns over time.
- I am long KMB, MO, COP, AFL, CB, INTC, and T. I may initiate positions in some of the stocks mentioned in my Simple Dividend Strategy model portfolio (SDS-30) in the next 72 hours.
- I’m not affiliated with any of the research authors or S&P.
- I haven’t found a similar study on U.S. markets, but if anyone finds one, I would be very interested in the seeing the results.
- Owain ap Gwilym , Andrew D Clare , James Seaton , Stephen H Thomas. (Winter 2009). Consistent Dividend Growth Investment Strategies. The Journal of Wealth Management, Vol. 12, No. 3: pp. 113-124. Working Paper version (here).
- Standard & Poor’s (December 2008). S&P 500 Dividend Aristocrats. The McGraw-Hill Companies (here).