Dogs of the Dow With Bonds

Should you find yourself aswirl in a cloud of antimatter and flung, thus, back to the year 1968, I have an investment strategy that will turn $10,000 in a tax-free individual retirement account (if only there had been such things in 1968) into $5.7 million. The strategy is informally called (in a recent Barron’s, anyway) "the dogs of the Dow with bonds." It is based on Michael O’Higgins’ new book, Beating the Dow With Bonds: A High-Return, Low-Risk Strategy for Beating the Pros Even When Stocks Go South.

That same $10,000 invested in O’Higgins’ earlier dogs-of-the-Dow formulation without bonds — that is, simply investing each year in the five cheapest of the Dow’s 10 highest-dividend-yielding stocks — would have grown to but $1.4 million assuming no taxes.

And yet that is still a heck of a lot better than having just put $10,000 in the Dow itself and hung on. That theoretical exercise (theoretical because the 30 stocks that make up the Dow change a bit from time to time) would have produced a mere quarter million or so.

The new plan — with bonds — calls for you to be out of the market altogether when the yield from high-grade corporate bonds exceeds the implied "yield" from the S&P 500.

If the S&P is selling for 25 times earnings, then its implied yield (even though you only get a portion of it in actual cash, as dividends) is 4%. At 20 times earnings — 5%. At 10 times earnings — 10%. At 50 times earnings — 2%.

Lately … well, even since 1980, actually, this new system would have had you out of stocks and in bonds. That’s right: no stocks in your portfolio these last 18 years. And you would have done great, because what people forget about bonds is that in addition to paying interest, when the general level of interest rates falls, bond prices go up.

The added twist to O’Higgins’ new system has you choosing one of two polar-opposite kinds of bonds (in years like 1980-1998 when the system dictates bonds): one-year Treasuries when inflation is feared; long-term, zero-coupon Treasuries when it’s not. (To determine this, according to the O’Higgins system, you simply look at the price of gold. If its price is lower than a year ago, inflation is not in the cards. If gold is up from a year earlier, watch out.)

And by following this system, there in your 1968 blue jeans and radical sideburns, if you could have conned your parents out of $10,000 (and found a way to shelter the interest and capital gains from taxes), you’d have your $5.7 million today. You’d be — what was it we used to say in 1968? — made in the shade.

All with what the author suggests could be 15 or 20 minutes’ effort each year.

So now we know the second best way you could possibly have invested your money these last 30 years: this system. (The even better way, you will be sick by now of hearing, is to have put your money into the stock of Berkshire Hathaway and never, ever thought about it again.)

The question is: How about the next 30 years? Personally, because I rarely do find myself aswirl in a cloud of antimatter (or if I do, it resets my memory chip, so I don’t know it), this question is of more interest to me.

And for this I’m afraid the system may be nearly worthless — particularly for money you don’t have under the shelter of a tax-deferred account.

Yes, in many years the system might do fine, just as, looking back, the Dogs of the Dow did well. (Yet in the last couple of years, when the system became widely popular, the Dogs actually trailed the Dow — oh, no!)

But what if you had a year when the system dictated all your money should be out of the stock market and in long-term, zero-coupon Treasury bonds? Putting all your money in one thing is risky. And long-term zeroes are riskier still. They zoom when interest rates fall, but crash when interest rates rise.

So there you are, having sold all your stocks (and owing Uncle Sam and his local counterparts perhaps 30% of your profits) and having put the proceeds into zeroes, as per O’Higgins, because the price of gold is lower today than it was a year ago. And now, having spent your 15 minutes on this, you go on auto-pilot for a year. Fine. But what if, even though gold didn’t predict it, inflation fears should take hold mid-year? Down come your bonds, up goes gold … so come January you have to sell all the zeroes at a huge loss (only $3,000 of which serves to lower your taxable income, with excess losses carried over to future years) and buy one-year Treasuries (or else stocks, if they’ve finally gotten cheap enough).

So it was foolproof in hindsight (as any good system constructed with the benefit of hindsight is). But — going forward?

Basically, with this system you are betting you can time U.S. interest-rate movements by assuming that the price of gold will work as an inflation predictor in the future in the same way, and with the same timing, it has in the past.

And it may! In which case you’ll do very nicely.

Or it may not. In which case you won’t.

All that said, and taxes aside, the very broadest strokes of this idea are useful. We should be nervous about U.S. stocks at today’s levels. And there are times when a safe bet in one-year Treasuries will prove a lot more lucrative than something more exciting.

But beyond that, it doesn’t seem to me this system is the golden key after all.

Shucks!

(Note one other serious problem: With zero-coupon bonds owned outside a tax-sheltered account, you have to pay ordinary income tax each year on the imputed "interest" you earned — even though the actual amount of cash interest you got was: zero. So zeroes really make sense only within a tax-sheltered account.)