Another Look at Dogs of the Dow Theory: Still Not a Compelling Standalone Strategy

|
Includes: DIA, F, GE, KO, MO, MRK, PFE, SPY, T, VZ, XOM
by: Bennington Investment Ideas

Feedback from my previous article on the Dogs of the Dow (DoD) theory prompted me to review another group of stocks. If you have not read the previous article or are unfamiliar with DoD, DoD postulates that one can buy the highest dividend yielding stocks in the Dow Jones Industrial Average and outperform the Dow over the following year. The theory is based on the notion that the highest yielding dividend stocks will be bid up as investors seek the higher dividends until the yields drop to a more reasonable range. In some ways, this is a reversion to the mean concept with the mean being a dividend yield.

My previous article analyzed a select number of securities, including General Electric Co. (NYSE: GE), Exxon Mobil Corporation (NYSE: XOM), Ford Motor Company (NYSE: F) , The Coca-Cola Company (NYSE: KO), and SPDR S&P 500 Trust ETF (NYSE: SPY). This analysis used historical data from Yahoo Finance for both prices and dividends. For each dividend payment, I looked at the associated yield at the ex-dividend date and the one year return. I then looked at the correlation between the yield and subsequent 1-year return. By using dividend and split adjusted prices from Yahoo Finance, the return calculation would account for the intermediate dividends as well. High positive correlations between dividend yields and returns would strengthen DoD. Unfortunately, this was not the case for the securities that I initially analyzed. Correlations ranged from around 0% for F to 20% for GE and KO to 30% for XOM. SPY offered the highest correlation at around 34%

The feedback suggested revisiting the analysis with stocks offering higher dividend yields. Verizon Communications (NYSE: VZ) and AT&T (NYSE: T) are often in the DoD selection. I also included the more obvious choice of SPDR Dow Jones Industrial Average ETF (NYSE: DIA). The following table lists the entire group of securities considered for this analysis:

Securities Analyzed


Ticker Name Current Yield Historical Average Above Average
MRK Merck & Company, Inc. 4.30% 2.68% 1.62%
MO Altria Group 5.60% 3.98% 1.62%
PFE Pfizer, Inc. 4.00% 2.73% 1.27%
T AT&T Inc. 5.70% 4.62% 1.08%
VZ Verizon Communications Inc. 5.50% 4.87% 0.63%
DIA SPDR Dow Jones Industrial Average ETF 2.41% 2.09% 0.32%
KO Coca-Cola Company (The) 2.90% 2.67% 0.23%
GE General Electric Company 3.30% 3.08% 0.22%
XLU SPDR Sector Select Utilities 3.94% 3.91% 0.03%
SPY SPDR S&P 500 Trust ETF 1.73% 1.79% -0.06%
XOM Exxon Mobil Corporation 2.40% 4.76% -2.36%
F Ford Motor Company 0.00% 4.11% -4.11%
Click to enlarge

Source: Yahoo Finance. Historical Averages are across the entire history but exclude one time large dividends. Average is taken as a straight average. MO is not currently a Dow Component; although, it was previously. XLU and DIA are also obviously not Dow Components. As noted in the previous article, neither is F.

This second analysis used Trailing Twelve Month dividends with large one time dividends excluded (in the case of MO). This list also includes several stocks with higher dividend yields as well as the SPDR Sector Select Utility ETF (NYSEARCA:XLU). The first graph shows the relationship for DIA followed by its historical dividend yields.

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit.

The correlation of returns to dividend yield is about 50% for the entire 13 year history. If recent history is excluded this climbs to 81%.

Source: Created from Yahoo Finance data

At first glance one might say that the relationship of DIA’s returns and dividend yields has a reasonable correlation and that this might support DoD. However, one needs to also consider the time frames and ask whether or not there is sufficient sample size. DIA had the shortest available history of approximately 13 years. Furthermore a 50% correlation is not very strong. The following graphs show the relationships between dividend yields and returns for VZ, T, MO, XLU, MRK, and PFE.

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit. Correlation is 58%

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit. Correlation is 41%

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit. Correlation is 50%

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit. Correlation is 47%

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit. Correlation is 16%

Source: Created from data from Yahoo Finance. Red line is the least squares linear fit. Correlation is about -4%

Clearly having high dividend yields does not mean there will be a stronger relationship in all cases since PFE and MRK both showed very low correlations between returns and dividend yields. In fact, PFE had a slightly negative correlation which is counter to the premise of DoD. However, the higher dividend yield stocks showed a generalization that their returns were correlated with dividend yields. The following table summarizes the correlations across the securities from both articles:

Correlation Between Yields and Returns


Ticker Correlation (all history) Correlation (up to around 12/2006)
VZ 57.5% 60.3%
DIA 49.5% 81.0%
MO 48.9%* 50.2%
XLU 46.6% 40.2%
T 41.1% 44.5%
SPY 34.0% 45.5%
XOM 30.2% 23.9%
KO 23.5% 23.1%
GE 21.3% 26.3%
MRK 16.4% 20.9%
F 4.3% 4.3%
PFE -3.5% 15.0%
Click to enlarge

Source: Author calculations. Data from Yahoo Finance. *MO all history is up to year end 2007. A very large dividend was made in March 2008 for $51 related to the PM spin off. Hence subsequent analysis is not really relevant to the previous time due to a very substantial change in the business.

Conclusion

It appears that this strategy might be appropriate for communication stocks, VZ and T. Among individual stocks, VZ had the overall highest correlation between yield and returns. However, the long term challenge would be that the fundamentals of long distance and land line phones are declining in favor of wireless communications. Since the VZ dividend yield is somewhat above its average, it might be a reasonable long play. The graph below shows the historical dividend yields:

Source: Created from data from Yahoo Finance. It should be noted that Yahoo Finance's dividend entry for June 24, 1985 is incorrectly listed as $0.02125. The correct dividend is $0.2125 as noted from www.verizon.com. The above chart reflects the correct dividend.

This strategy may also be appropriate for T, MO, and XLU. XLU is right at the historical average which would indicate a possible long. T is also a little above. MO might be more questionable since I excluded the recent history.

I would not apply this strategy to PFE and MRK even though they are above their historical average dividend yield. Both stocks show almost no correlation between returns and yields. PFE's graph is below:

Click to enlarge

Source: Created from data from Yahoo Finance.

However, this is not to say that PFE and MRK are not good possible long investments especially for those seeking a solid dividend yield. Additional fundamental analysis and research might prove them to be good opportunities. Overall, I am skeptical of this strategy. On standalone basis it will probably result in disappointment; however, as supporting evidence for an investment decision it might be useful since for some stocks, their clearly is some correlation. I suspect there are other stocks, especially in the utility space. Conceptually, this strategy will probably be more applicable for stocks in mature markets that provide good dividend yields.

Disclosure: I am long SPY.

Disclaimer: This article is for informational and educational purposes only and shall not be construed to constitute investment advice. Nothing contained herein shall constitute a solicitation, recommendation or endorsement to buy or sell any security.