DOW 36,000 Redux

Michael Pollak mpollak at
Fri Aug 20 04:20:18 PDT 1999

On Thu, 19 Aug 1999, Enrique Diaz-Alvarez wrote:

> The point is that, even assuming this is true, the DOW is seriously
> overvalued, and they can only sustain their projections by lying about the
> numbers.

I love the way they use an entirely different formula from the first time around and somehow finagle it to end up with exactly the same weird estimate of 36,000 that just happens to be in the title of the forthcoming book that they've been flogging since June. So I guess that means they frantically re-wrote the math path parts in the last two two months and left the rest of the text as is. It reminds me of nothing so much as faking lab results in high school science class: find out the answer and then manipulate the data until they produce it.

This new approach makes it look like they've "learned" from their opponents. In the exchange Glassman had with Clive Crook from The Economist Magazine in Slate (below), Crook pointed out that one could get huge differences from tweaking assumptions, which were matters of opinion, but that Glassman got his super-huge difference from a math mistake, which wasn't. So now I guess what Glassman and Hassett took away from that is that should depend on tweaked assumptions all the way down. At the end of his exchange with Crook, Glassman seems backed into admitting a Dow of "only" 18,000. But now he's managed to tweak it back to the magic number with ever more absurd assumptions.

I guess the one number he can count is book sales, and the only number that's invariant is in the title. He couldn't change that without looking like a fool to the book-buying public (as opposed to math and finance experts). But now all this "controversy" will probably just generate more sales. I'm beginning to think he's not dumb so much as a flat out charlatan liar with an idea to make money -- for him.


The Stock Market From: Clive Crook To: James K. Glassman Posted: Monday, June 22, 1998, at 4:30 p.m. PT

Dear Jim,

I haven't much to add to my previous letter. If you had cast your original argument in terms of dividends rather than earnings, we would have had less to quarrel about. Even when you do cast it that way (in other words, when you put the right variable into your formula), you still conclude that the Dow should be higher--but now only 50 percent to 100 percent higher, not four times higher. That's progress, I suppose.

A gap of 50 percent-100 percent between the actual and (you say) warranted levels of the Dow is still big, sure enough. Nonetheless, a difference of this order is explicable in terms of quite small changes in underlying assumptions. This is a consequence of adding up payouts to infinity--and a reminder to handle valuation formulas with care. Changes of mere tenths of a percentage point in your forecasts for inflation, dividend growth, and so on would close the gap between your "lower bound" dividend-based estimate of where the Dow should be and where the market actually is. (You use a real bond yield of 3.1 percent, for instance, in your calculation. But, as you say in your last letter, the real return on T bills is currently 3.7 percent. Using the second figure rather than the first would be enough to bring your estimate of where the Dow should be down to 9,000.)

So, once we are discussing lower bound dividend-based estimates, we are in the realm of difference of opinion and forecasting error. Our earlier discussion on the risk premium becomes relevant again. On all this we can agree to differ. But the enormous valuation anomaly you said you had spotted, the idea that attracted all the attention in the first place--the remarkable claim that the Dow should right now be at 36,000, on a price-to-earnings ratio of 100, and all that--comes not from tweaking the assumptions but from your conceptual breakthrough of an "upper bound" formula based on earnings. As I explained last time, that is not in fact a legitimate formula but a piece of nonsense, a plain mathematical absurdity.

What you say about Microsoft shows you still don't see this, or won't admit it anyway, but I can't explain it more clearly than I already have. I advise readers who are still perplexed to look at Pages 59-62 of the current edition of Principles of Corporate Finance, by Brealey and Myers, a standard text. It explains how you use a discounted cash-flow formula (the simple formula, Jim, of your second letter) to value a stock. It explains why you must use dividends, not earnings, in such a calculation, and why it is a fallacy to suppose (as you do) that using dividends instead of earnings is equivalent to "ignoring" capital gains.

But what the hell, if the Dow hits 36,000 by the end of the year, lunch is on me.

Sincerely, Clive


© 1999 Microsoft Corporation. All rights reserved

__________________________________________________________________________ Michael Pollak................New York City..............mpollak at

More information about the lbo-talk mailing list