I am sure that many readers are familiar with the "Dogs of the Dow" strategy. The strategy has been in use since 1972 and has been a proven success, with an average yield of over 14% per year, compared to an average yield by the Dow overall of over 11%. The results this year aren't quite so good, with the Dogs up by 5.7%, while the Dow's 30 stocks as a whole are up 11%. The Dow Jones Industrial Average was up 7.3%. The results for the Dogs are based on the "traditional" strategy of investing $10,000, divided equally among each of the ten Dogs. (The figures do not include dividends.) The table below shows the results for the 2012 Dogs:
The success of the "Dogs" strategy is well established. The strategy was "invented" in 1972, and over 40 years it has, on average, outperformed the Dow by about 300 BPS. Even more impressive is the strategy based on the five lowest-priced Dogs (alternatively called the "Pups of the Dow," the "Small Dogs of the Dow" or the "Flying Five"): they have averaged better than 20% yield, or nearly 900 BPS over the Dow. This year, the Pups managed a yield of 6.7% - somewhat less than the Dow.
I do have a question, however, that has been nagging at me ever since I took a close look at the Dogs of the Dow and the companies that make up the Dow 30. Given that the Dow Jones is very selective about the companies included in its group of industrial companies, how would those companies hold up to an analysis by a set of criteria intended to identify well-run companies? This would, of course, take a lot of time, so focusing on the Dogs of the Dow seemed like a reasonable starting point. I walked away from this initial encounter a little bit wiser about the "Dogs" strategy.
The Pedigree Approach
Those readers who have been following some of my articles and blogs may remember the set of criteria I have been using to identify what I (rather pretentiously) consider to be a "minimally acceptable investment." Such a company would exhibit good management, earnings adequate to pay its claimed dividends, the ability to cover its short-term debts, the ability to turn a profit, and a trend upwards in share value. I have worked on the criteria a little - sharpening its claws, so to speak. The set of criteria stated for use in a screener look like this:
- Return on Assets > 0 (or > 1, if you prefer)
- Return on Equity > 0 (ditto)
- Return on Investment (or, Invested Capital) > 0 (ditto)
- Quick Ratio > 1
- This is a change from the original set of criteria, where I used the Current Ratio. Quick Ratio is a more demanding condition, in that it is based on only the most liquid assets a company has available - excluding inventory, which cannot always readily be turned into cash. The quick ratio is sometimes referred to as the "acid test" ratio.
- Operating Profit Margin > 15
- This is a change from my use of net profit margin, as it measures a company's profitability independent of interest and taxes (EBIT), giving a better indication of the company's ability to manage its operating costs.
- Instead of accepting "just any" positive level of profit, in order to separate the wheat from the chaff, so to speak, I have arbitrarily set the minimum level at 15%.
- Dividend Cover > 1 or Payout Ratio < 100 (%)
- For those unfamiliar with the term, "dividend cover" is essentially the inverse of "payout ratio." While the payout ratio is determined by dividing a company's dividend payment (a year's payment) by earnings per share, dividend cover (a ratio used more in Britain) is determined by dividing EPS by the dividend. The latter approach makes more sense to me, as it measures the ability of a company to cover its dividend by the earnings it makes per share. A dividend cover of 1 or better means the EPS is adequate to pay the dividend. With payout ratio, the ratio must be lower than 100 (less than 100%), which happens when the EPS is greater than the dividend. Looking for a lower ratio just seems counterintuitive to me.
- I have started using Payout Ratio < 100, in deference to screeners' not usually having a filter for "Dividend Cover."
- Performance (trailing 12 months) > 0
- Performance (most recent quarter) > 0
- This last criterion I consider optional, for reasons that will be explained below.
Out of curiosity, I applied the criteria above along with a filter for the Dow, to see how many of the Dogs I'd catch. Of course, this is a bit biased, in that I did this in early November, and the Dogs are determined in January. The 10-month interval may have had some effect on how the companies selected performed, so I do not consider this an iron-clad test. The list below consists of the 2012 Dogs of the Dow. The companies with an asterisk next to them are companies that meet my criteria - all, that is, but for the Performance (MRQ). With that in place, only two Dogs survive (those have two asterisks).
- AT&T (T)
- Verizon (VZ)
- Merck (MRK)*
- General Electric (GE)
- Pfizer (PFE)**
- DuPont (DD)
- Johnson & Johnson (JNJ)**
- Intel (INTC)*
- Procter & Gamble (PG)
- Mondelez (MDLZ) (Mondelez replaces Kraft (KRFT) on the 2012 Dogs list.)
It's interesting that three of the four companies satisfying my criteria above are in the healthcare sector. The one company that isn't (Intel) pulled up lame and was down $3.23 for the year (-15%). What is really interesting is that the four companies selected by the criteria accounted for a yield of 4.10% on their own, which seems fairly respectable. The following table is extracted from the earlier table:
(click to enlarge)
Results like that led me to wonder if a set of "dogs" selected only by my criteria (rather than with an additional requirement about dividend yield) would pass the Dogs up. Using my criteria (this was done on November 16, by the way), I drew up a list of Dow Jones companies that met the criteria, again foregoing the performance for most recent quarter - that only added Cisco (CSCO) to Pfizer and J&J. I was quite surprised that fewer than nine companies in the Dow 30 could meet my criteria, so I removed the remaining performance criterion, which gave me the nine companies below:
- Cisco Systems
- Chevron Corp. (CVX)
- Intel Corporation
- Johnson & Johnson
- McDonald's Corp. (MCD)
- 3M Company (MMM)
- Merck & Co.
- Microsoft Corp. (MSFT)
- Pfizer Inc.
I tried weakening the operating margin, but that didn't help. This meant that whatever was keeping the other 21 companies out was something a bit more fundamental - either the quick ratio or the payout ratio were likely to blame, and I consider both of those to be essential. So nine it is. I decided to call them the "Pedigrees." These aren't just any old dogs. Assuming a start at the beginning of the year, the table below shows how they would have performed in 2012:
The results are not bad, but the Pedigrees did 225 BPS worse than the Dogs. I was curious to find out why the shortfall, as I had expected the Pedigrees would perform at least as well as the Dogs, and (really) had hoped they would outperform the Dogs by a fair margin. There were, however, two things that were interesting about the comparison between the Dogs and my Pedigrees: I had nine holdings, while the Dogs have ten - no fair playing 10 against 9 (bad Dogs); and only one of the Dogs had managed to outperform all of my Pedigrees - Pfizer was the best-performing Pedigree with a 15.9% growth, while General Electric had a growth of 17.2%.
It was clear to me that anything that had a lower growth rate than Pfizer would not improve the overall performance of the Pedigrees, but a company that performed better would. So I plugged GE into my "kennel" to see what would happen:
The result was interesting. I had brought the Pedigrees up to within 88 BPS of the Dogs - an improvement of 137 BPS - by adding the "runt" of the litter of the Dogs (GE was the lowest-priced Dog, after all). Considering a shortfall of anything less than 100 BPS to be at least competitive, I wondered if I was on to something here. Next test: how many companies in the Dow 30 had outperformed GE? As it turned out, only six:
Plugging each of these into the kennel in the "tenth cage," I achieved results in keeping with the expectations I had formed based on each company's 2012 yield: Travelers (TRV) moved the overall yield up to 5.25%; American Express (AXP) charged ahead a bit further to 5.3%; JPMorgan (JPM) really chased the Dogs, and brought the Pedigrees up to a 6.34% yield; Disney (DIS) artfully edged the yield even further to 6.39%; Home Depot (HD) built an even larger lead with 7.83%. The best results came from putting Bank of America (BAC) into the tenth slot, where it really cashed in by giving the Pedigrees a 14% yield overall.
Contemplating these results led to some insights into the nature of the Dogs and why they are able to outperform the Dow (on average).
The Dogs of the Dow are the ten companies (from the Dow 30) that have the highest dividend yield on December 31 (after close of business). What struck me as curious was that - when reporting on the performance of the Dogs - the actual dividend payments are not considered part of the yield at the end of the next year. Why, then, the focus on dividend yield in selecting the Dogs in the first place?
Then it occurred to me. The companies that comprise the Dow 30 are considered to be among the leading companies in the country, and one would expect that their valuation be stable or prone to growth. But even the best companies have their down days (or quarters, or years), and when the share price of a company goes down, its dividend yield goes up. Also, if a company is doing well, one result of its success can be an increase in dividend payments, which also results in a higher yield.
It would seem logical, then, that in either case investors would find a solid company with a higher than usual dividend yield to be an attractive buy - if the company is enjoying good times, one gets growth potential in the stock, plus a higher dividend to boot; on the other hand, if the company's stock is depressed for some reason, it is more than likely to jump back up to a more "appropriate" valuation. In either case, companies with high dividend yields would, in general, be companies that one could look to for increased growth potential.
The above considerations can only be counted on to a degree. Not every company whose value is depressed is going to go back up - some may be experiencing serious malfunctions resulting in the devaluation, and the temporary high dividend yield may constitute a false read as an indicator of future growth potential. There is also no way to conclusively determine when a depressed share price is going to shoot back up; it may be months - or even a year or so - before a company's valuation begins the long trek upwards (for example, consider Alcoa Aluminum (AA), whose shares dropped from the $40's in 2008 and have since languished around $10).
By the same token, not every company that is currently riding the gravy train is going to stay on track indefinitely. Pfizer's loss of exclusivity on Lipitor is an example of the sort of thing that can arise out of the normal course of business and result in diminished value - at least, in the perception of investors. Changes in management can change a company's momentum (Apple's (AAPL) - to use an example not in the Dow - loss of Steve Jobs and the ascension of CEO Tim Cook, plus other management adjustments, preceded the company's precipitous drop in share price).
This is no doubt why the Dogs of the Dow consists of the top ten dividend-yielding companies in the Dow. Note that both DuPont and Intel dropped in value over the course of 2012. Note that Procter & Gamble realized a yield of only 1.8% for 2012. While it is true that the "top dog" (AT&T) provided a yield of 11.8%, that kind of growth likely did not happen just because AT&T offered the highest yield of the Dow 30 (although its dividend yield may have played some role) - after all, GE outperformed AT&T even though it was (as I characterized it) the "runt" of the litter. By spreading your bet over the top-ten-dividend-yielding companies you take into account that some may be losers, and some may eke out only a small gain - but as a whole, the ten companies are likely to give a fairly nice return.
This may also be why the Dogs strategy states that one invest an equal amount on each of the Dogs. If one were to buy just one share (or any specific number of shares) of each stock, we come up with significantly different results. If one had played the Dogs this way, one would have realized a yield of 4.98% - certainly not bad, but also 73 BPS lower than the standard Dog strategy. (The Pedigrees would have fared worse - 2.74%, or 209 BPS less than the standard-strategy approach.) No matter how much you spend on the strategy - $10,000 or $100,000 or even $1,000 - as long as the same amount is invested in each company, you maximize the outcome of the strategy in terms of yield.
As a sort of segue to the Pups of the Dow, the standard strategy seems to place a little more emphasis on the lower-priced stocks. The highest priced of last year's Dogs was Procter & Gamble, at $66.71. A $1,000 investment would have bought almost 15 shares. That same $1,000 bought almost 56 shares of GE, priced at $17.91. In terms of percentage, the same percentage shift in either P&G or GE will pay out the same in terms of the holdings as a whole. Consider the following "test" using Pedigrees:
As you can see, the same yield is going to have the same overall effect on the value of an investment. Whether you are talking about a $65 share, a $24 share or a $106 share, if the same amount is invested in each, the same yield will press through to the resultant value of the investment. (All of this, by the way, presumes you can buy fractional shares.)
Now, consider this test:
In the test above, rather than adjusting each company by the percentage change in share prices, we adjust the share prices by the actual dollar difference between the beginning and ending prices. The results are interesting, and indicated to me how to resolve the issue of what to add to the Pedigrees.
Note that when we add $4.52 (the difference between J&J's starting and ending prices) to Intel, we get a significant return (basically, an 18.6% yield), although neither J&J's nor Chevron's results are anywhere near that substantial. Then note what happens when we subtract $3.63 (the difference between Intel's beginning and ending prices) - although Intel suffers a 15% meltdown, neither J&J nor Chevron see quite the dramatic loss.
As I see it, the Dogs of the Dow strategy incorporates three ideas (at least): first, as mentioned at the beginning of this section, the Dogs - by focusing on the highest-dividend-yielding companies - is biased towards companies that are either devalued or are experiencing growth; second, by including ten companies it helps to minimize the impact of any negative occurrences in a few stocks on the value of the portfolio as a whole; third, the equal initial investments give extra weight to the lowest-priced companies (arguably the companies more likely to increase in value), while at the same time providing a buffer in the event of adverse movement among the higher-valued companies, of which there will be fewer shares (and, arguably, greater chance of decrease in value).
As I acknowledge, the third point is arguable on both accounts, but it is the third point that I believe underlies the success of the Pups of the Dow. Being the lowest-priced Dogs, the Pups are going to realize more movement, simply due to the fact that more shares of each company are involved. Further, being the low dogs on the totem pole, they are - albeit arguably - the stocks most likely to appreciate in value. The fact that the Pups have tended to outperform both the Dow 30 - as well as the Dogs - as a whole, would seem to lend some credence to the claim. Whether this would be true of the market as a whole, I cannot say. My observations here are limited to the rather restrictive and rarefied environment of the Dow Jones Industrials. There is, after all, an intangible value to being a member of the Dow.
As a final point to this section, let me state upfront that I do not claim to have "invented" anything here, or to have uncovered some arcane secret. I am simply giving an account of the process I've gone through in trying to decide what to do with the idea of a "Pedigree of the Dow" strategy, using the Dogs as a standard against which to compare my work.
Back to the Kennels
I have the following options:
- Run the Pedigrees as is - there is nothing wrong with a strategy that doesn't pay out more than the Dogs of the Dow. The primary concern is to put forward a strategy that is not likely to tank on the investor, and the quality of the Pedigrees is such that they seem likely to do that much.
- Run the Pedigrees with a "Runt" as a play to boost the yield of the portfolio. This strategy takes two forms:
- One play would use the Dog with the lowest share price on January 1;
- An alternate play would use the member of the Dow 30 with the lowest share price.
Either of the "Runt" strategies seems viable. If one were prescient, one could always pick that company that was going to have the best overall performance for 2013, but that is not possible (could anyone have determined of one of the six best-performing companies for 2012 that it just had to be one of the six best?). Could anyone have determined beforehand that BoA was going to more than double in value? This is not to say that any of the six companies leading the Dow this past year shouldn't have finished the year with positive performance - just that I don't think one could have determined that any of them would have been one of the six best-performing companies of the Dow for 2012.
On the other hand, there is reason to believe that the lowest-priced company would move upwards - the reasoning in this regard has already been discussed. The question is, should the lowest-priced Dog be the choice, or the lowest-priced Dow 30 company be selected? I'm going to opt for the lowest-price Dog, for two reasons: first, the high dividend yield that got the company into the Dogs is at least some form of endorsement. Given that it has a high dividend, it is likely that the "Runt" of the litter is there because it is devalued, and since it has still maintained a significant dividend, it is a reasonable bet that it will gain some value back in the coming year. The track record for the Runt of the litter is fairly good - out of the past 16 years, the Runt has increased in value 10 times, while losing in only 6 years (a 62.5% performance).
The second reason is purely pecuniary: the Runt of the litter is going to offer a higher dividend yield than a lower-priced company with an inferior yield.
Before committing myself to either strategy (that is, to run with just the Pedigrees, or to add the Runt of the Dogs to my kennel) I thought I would run one more test. From November 16 through the end of the year I would run the Pedigrees against the Dogs. If the Pedigrees won, the Runt would not be necessary. Furthermore, if the Pedigrees lost with the Runt, I would run the Pedigrees alone. I would adopt the Runt only if the Runt's addition to the Pedigrees made enough of a difference to give the Pedigrees the win.
Because my first experiment ran the Pedigrees (determined on November 16) against the Dogs (determined on January 1), I decided to pick a new set of Dogs - those ten companies which, on November 16, had the highest dividend yields. The two "teams" would have their November 16 share prices as the starting point. There were some changes.
The result was a fairly decisive (130 BPS) victory by the Dogs. It was time to bring in the ringer - in this case, it would be Hewlett-Packard (HPQ), which replaced GE as the Runt.
Result: a 56 BPS win for the Pedigrees. I shall adopt the Runt of the litter. (It sounds so humane - how can I go wrong?)
Stating the "Pedigree Strategy" more precisely, then, the strategy will involve the application of my eight criteria to the Dow 30. If needed, the most-recent-quarter performance will be dropped. If further needed (and it was), both performance criteria would be dropped. If this leaves the Pedigrees with fewer than 10 companies, the lowest-priced members of the Dogs of the Dow for that year (the Runts) will be added to the kennel until the kennel has 10 members. Runts will be picked from the bottom up (lowest-priced first).
The 2013 Dogs of the Dow (and their starting prices) are:
The 2013 Pedigrees of the Dow (and their starting prices) are:
The only new addition to the kennel is the Pedigree IBM (IBM). Intel made it onto the Pedigree list once the performance criteria were eliminated. I did have some reservations about McDonald's Corporation, since Mickey D's had a Quick Ratio of 0.99 instead of a straight 1, but given the fact that that would require the addition of a Runt - and that the Runt would be Hewlett-Packard (which is, at present, embroiled in controversy over its acquisition of Autonomy) - led me to think that it would be overly persnickety of me to penalize McDonald's for one-one-hundredth of a point. This gives me a purely Pedigree pack. (That's gratitude for you. HP gave me the win over the November Dogs, and now I'm all too happy to say good-bye to it.)
I will update the progress of the comparison weekly on my Instablog (sometime during the weekend, using the Friday closing prices). I invite you to keep track of the proceedings there.
Disclosure: I am long PFE.
Additional disclosure: Data has been acquired from DogsoftheDow.com, the Motley Fool, YCharts and FINVIZ.