To print article, click
Print on your browser's File menu.
Go back
Go back
Posted 6/28/2001
Related Resources
Keep track of Victor Niederhoffer and Laurel Kenner's picks on their Recommendations page.
Also follow their Dow Reversals portfolios.
Details
 |
|
The Speculator
5 Dow greyhounds ready to run
We adjust our measurements for the uncertainties of an imperfect world, overlook the junkyard dogs and pick what we think will be the Dow's racers in the second half of 2001: Philip Morris, JP Morgan, Honeywell, Boeing and Merck.
By Victor Niederhoffer and Laurel Kenner
More than a century following its creation, the Dow Jones Industrial Average remains at the pinnacle of respect in the realm of finance and investments.
-- David Elias, "The Dow 40,000 Portfolio"
 Money Plus. Easy online tools and free Bill Pay, too. |
Like all things of power and potency, the Dow Jones Industrial Average is imbued with considerable mythology. One legend is that buying a group of temporarily suffering Dow stocks can generate superior returns.
Michael O'Higgins, in his 1990 book "Beating the Dow," was the chief builder of this myth. "Multinationals without exception, the Dow companies in their different ways are positioned to benefit from the mega-trends of the 1990s: globalization of markets, cleaning up the environment, repairing the infrastructure, depletion of energy sources, revival of manufacturing, increased literacy rate, aging of the population, expansion of the free world," O'Higgins wrote. "With companies of Dow stock caliber, there is more opportunity than risk. Bad news is usually good news, because it makes strong stocks cheap."
The myth's chief proselytizers, until last December, were David and Tom Gardner, of Motley Fool fame. Their version called for the purchase of four high-yielding, low-priced Dow stocks at the end of each year.
Unfortunately, myths are invented to provide humans consolation, not superiority; following blindly is not unduly profitable. A number of things were wrong with the O'Higgins/Fool approach from an analytical standpoint, and we'll enumerate the problems later. The worst drawback was that the system didn't work very well. In fact, the Gardners publicly renounced the "Foolish Four" system last December, after their researchers studied 50 years of data and concluded that the system offered, on average, only a 1.74% annual edge over the Dow itself (a difference that could be eaten away by taxes and trading commissions).
We applaud the Fools' willingness to examine their premises -- but we think they missed an opportunity to develop a better system.
Our own study has led us to conclude that the recent massacre of growth stocks, combined with the veneration accorded Dow stocks, affords the chance to employ a systematic approach with alluring profit potential.
Why study the 30 Dow stocks at all, when there are more than 7,000 stocks to choose from? The answer lies in the fact that to the public, the Dow is the market. Dow stocks command the sort of reverence that ancient Greeks had for Zeus and Americans reserve for Alan Greenspan. The Dow 30 are seen as icons, leaders in leading industries chosen by The Wall Street Journal's wise men to provide a continuously updated portrait of the U.S. economy's character and strength. Hence the transformation from Charles Dow's original dozen in 1896, when America was shifting from an agrarian to an industrial economy, to today's average, dominated by consumer stocks and advanced technology.
Worth a fresh look
We think it's worth taking a fresh look at the Dow, particularly because past studies made some basic errors. They tended to consider the numbers and key ratios of its stocks one at a time. They assumed -- incorrectly, as we'll explain below -- that the effects of such variables are constant between years and within years. They included no measure of uncertainty. Furthermore, the studies lacked forward-looking variables such as earnings estimates to enliven the results with real-world phenomena.
We attempted to rectify these errors in our own efforts. To do so, we calculated the impact of price/earnings ratio, dividend yield, 12-month return, Value Line timeliness ranking and price level at Dec. 31 and June 30, for the five years ending Dec. 31, 2000. (We chose Value Line's ranking system because it includes a consistent forward estimate of earnings.)
We generated our material from Standard & Poor's stock guides and Value Line's database, so as to be sure that the data would actually have been available to investors as of the end of the year. We performed the analysis as of midyear and year-end to take account of the changing cycles within the year.
First we looked to see if any variables exerted consistent predictive influence. Regrettably, we found nothing significant for predicting yearly performance. Worse yet, the relations that did exist were weak and unstable from year to year. For example, the price-to-earnings (or E/P) ratio correlated close to 0 with yearly returns in all years except 2000, when it correlated to 0.5. If you had to put all your money on one horse, the dividend yield was the best choice -- the higher, the better.
Stats of the Dow: What works?
Correlation of selected ratios at prior year-end with annual returns.
| Correlation variable | 2000 | 1999 | 1998 | 1997 | 1996 |
| Earnings/price ratio | 0.51 | 0.06 | -0.08 | -0.04 | 0.09 |
| Dividend yield | 0.57 | -0.08 | 0.10 | 0.35 | -0.07 |
| Year-end price | -0.30 | 0.31 | 0.62 | 0.28 | 0.15 |
In short, a can of shifting and malevolent worms.
(The correlation represents the percentage reduction in error of prediction that can be created by using one variable to predict the other, compared with using just one variable alone. In numerical terms, it represents the degree of linear association between two variables, on a scale of -1 to 1, with 1 representing perfect direct association and -1 representing perfect negative.)
We then looked at the first half and second half to see if there were any consistent predictive relations within the years. Unfortunately, we found absolutely nothing of predictive value between any of the year-end variables and performance in the first six months.
However, when we turned to the predictive relations for the second half, we found some great tells.
The most important variable for the second half was the Value Line ranking (the more favorable, the better), followed by the previous 12 months' return and dividend yield (in both cases, the higher, the better).
The remaining variables did not add anything significant to the accuracy of the prediction, in either half of the year.
We will send a spreadsheet to readers who would like to see the complete workout. If anyone has difficulty opening the file, or any of the spreadsheets we offered in past columns, we'll send a fax.
Two of the three variables that are helpful in the second half relate to momentum: the percentage move in price the previous year and the Value Line ranking, which is based mainly on the trend in earnings (the more favorable the comparisons, the better). Apparently during the last half of the year dividends become more important to investors.
Something similar often takes place on the racetrack. In the first half, the tendencies are to back and fill, to gain space. In the second half, the real strength and class show.
Profit potential
We alluded earlier in this column to profit potential, and we will not disappoint readers by omitting this part of our research. Unfortunately, we don't have a simple formula. (Fortunately, we suspect that after the past year in the market, nobody believes that making money is simple.)
After considering all the interrelations between the variables, we computed a formula that over the past five years has provided the best predictive fit between the key ratios of the Dow and the subsequent returns for the last six months of the year. The formula's efficacy is highly significant from a statistical point of view, and it seems to have practical value.
Without further ado:
Expected Return = 5.3-(3.57 x Value Line Rank) + (0.12 x 2000 Return) + (433 x Dividend Yield)
(The dividend yield and Value Line rankings are both as of year-end 2000.)
To calculate our formula, we used regression analysis, a statistical technique first employed by the 19th century scientist Francis Galton, to describe relations between dependent variables (those on the left side of the equals sign) and independent variables (those on the right side).
Take Philip Morris (MO, news, msgs). At the end of 2000, Value Line's ranking on the stock was 2. Last year's return was 91.3%. The dividend yield at year-end was 4.8%.
Thus, the predicted return for Philip Morris would be:
5.3-(3.57x2) + (0.12x91.3) + (433x0.048)
= 5.3-7.14 + 10.96 + 20.78 = 29.9%
This equation explains some 9% of the total variation in the returns of the 30 stocks over the last half of the year. (For the statisticians out there -- everyone else cover your ears -- the R-squared is 9%. The square of the correlation coefficient, R, is the proportionate reduction of the total variation in the predicted variable that comes from introducing the explanatory variables.)
Nine percent doesn't sound like much, but statistically it's very significant; to obtain anything much higher, you'd need a clairvoyant. And from a practical standpoint, when you're dealing with companies that have a combined market value of more than $3 trillion, being able to explain some 9% of the variation in returns of these stocks in any six-month period seems worthwhile.
We'll call our formula "Greyhounds of the Dow," to set it apart from the Dogs strategy and highlight its emphasis on earnings momentum rather than low price. The following table lists the five stocks we expect, based on our formula, to be the Dow's top performers in the second half.
Dow greyhounds
| Company | Value Line rank | Dividend yield | 2000 return | Predicted return |
| Philip Morris (MO, news, msgs) | 2 | 4.8% | 91% | 29.9% |
| JP Morgan (JPM, news, msgs) | None | 2.5% | -41% | 11.1% |
| Honeywell (HON, news, msgs) | None | 1.6% | -18% | 10.1% |
| Boeing (BA, news, msgs) | 2 | 1.0% | 59% | 9.6% |
| Merck (MRK, news, msgs) | 2 | 1.5% | 39% | 9.4% |
• 5 high-priced stocks ready to shoot skyward, 6/21/01
• 20 stock speculations in time for summer, 6/14/01
• Slicing and dicing to find Nasdaq’s best stocks, 6/8/01
more...
One of the reasons that the Foolish Four system and others like it have fallen by the wayside is that it is not possible to make money over the long run by following fixed rules. Nobody has ever put it better than Robert Bacon in his 1956 book, "Secrets of Professional Turf Betting:"
"The principle of ever-changing trends works to force quick and drastic changes of results sequences when the public happens to get wise to a winning idea."
Bacon was writing for racetrack bettors, but the principle applies with equal force to the stock market. (We are proud to have brought Bacon's insights to the attention of the masses. Bacon's book has become a cult classic, and we are constantly receiving requests as to how to locate it. Please feel free to contact our researcher, daveciocca@yahoo.com, if you would like us to fax some pertinent pages.)
One final word: Readers should bear in mind that we are vulnerable as anyone else to Bacon's principle of ever-changing cycles -- and that goes for our formulas, no matter how painstakingly calculated.
At the time of publication, neither Vic Niederhoffer nor Laurel Kenner owned any stocks mentioned in this column.
MSN Money's editorial goal is to provide a forum for personal finance and investment ideas. Our articles, columns, message board posts and other features should not be construed as investment advice, nor does their appearance imply an endorsement by Microsoft of any specific security or trading strategy. An investor's best course of action must be based on individual circumstances.