It's relatively long, so I've attached it as a text file.
Hank Leland Research Analyst Service Employees International Union
Barbara Laurence wrote:
> Doug, am I crazy or is Glassman and Hasset's calculations based on earnings
> only, not capital gains from stock sales? If you assume that everyone will
> stay in the market no matter what, and that more people will enter as more
> join the labor force, then it would seem that they're right? or, again, am
> I crazy?
> Jim O'Connor
-------------- next part --------------
==========================================================================
Slate - E-Mail to the Editors - March 25, 1999
==========================================================================
? Address your e-mail to the editors to letters at slate.com. You must include your address and daytime phone number (for confirmation only).
(Posted Thursday, March 25, 1999)
Dowism
? ? ? ?In "Crapshoot," Bruce Gottlieb, like Clive Crook before him, thinks he has found a flaw in our argument that stocks are undervalued. He has not. In a Slate "Dialogue" last year, Crook claimed that we were contending that the value of a company is the present value of its stream of future earnings and that such a contention was false. Gottlieb's piece, in an easy-to-follow way, demonstrates Crook's argument that, if you value a firm by discounting its earnings, and all earnings are not paid out in dividends, then you are double-counting.
? ? ? ?But the problem with what Crook and Gottlieb are saying is that we never based our theory on earnings but instead, as we wrote in our Wall Street Journal piece on March 17, 1999, on the "money a stock will put in your pockets through the profits generated by the company that issued it." Neither in this year's piece nor in our March 30, 1998, Journal piece did we claim that all those profits--or earnings--would go into your pockets. What we said explicitly was that the proper figure to use to measure the cash generated by a company was somewhere between the lower bound of the dividend that a firm pays and the upper bound of its official after-tax earnings. Indeed, in our first piece we explicitly highlighted the fact that earnings are a problematic measure of cash flow because of the potential that a firm might grow simply through retentions.
? ? ? ?But Crook and Gottlieb are intelligent, well-intentioned journalists. What led them astray? They both cite an argument against our first piece raised by Jeremy Siegel of the Wharton School. He claimed that it is a "mistake" to assume that dividends per share will post real growth, as we did in our calculations, because dividends, in theory, should only just keep up with inflation.
? ? ? ?Crook and Gottlieb appear to have been convinced that real earnings growth only reflects retentions and that real dividend growth must be zero. While Siegel's point fits the simplest introductory economic models, it is contradicted by the facts: Historical data on dividends reveal significant real growth of dividends per share. For example, in the latest edition of Siegel's excellent book Stocks for the Long Run, Table 5-1 on Page 79 shows that the growth of real dividends per share has been 2.1 percent on average since 1946 and positive since the 19th century. (In our calculations, we assume 2.1 percent real growth in dividends.)
? ? ? ?What Siegel, Crook, and Gottlieb say can't happen--dividends growing faster than inflation--has been going on for almost 130 years. Where has all this growth come from? We treat that question at length in our book, relating the observation to theories at the frontier of the branch of economics known as "Industrial Organization," but here is a hint: The simplest textbook model of the perfectly competitive firm doesn't do a great job of describing the companies that have driven the market higher and higher.
? ? ? ?Discounting dividends is uncontroversial, and the fact that they grow is clear. It is interesting that Gottlieb says that, on the strength of dividends alone, the Dow should be about 14,000 using our theory. That is a good start. Even sticking with dividends alone, the number is pushed considerably higher than that if one accounts for repurchases and the tax advantage associated with them. So much for the crazy stock market bubble. Dow 14,000 is clearly a lower-bound estimate. Today, many enormous firms don't pay dividends at all. These firms have value because, ultimately, they will deliver cash to their shareholders.
? ? ? ?Crook and Gottlieb think that all valuation techniques are out the window for firms that don't pay dividends, but they are misinformed. There is a very large peer-reviewed academic literature on this topic. The basic idea is simple: You can base a value measure on earnings instead of dividends if you can identify those things that you are double-counting and make sure that you only count them once. For example, earnings themselves are a reasonable measure of the dividend that you use to construct the value of the firm if all earnings are paid out each year. Such an example is no pipe dream. Real Estate Investment Trusts, for example, pay out 95 percent of their earnings, and many of them post earnings growth well above inflation year after year. In 1999 alone, REITs, on average, are expected to increase their earnings by 10 percent, with inflation at about 2 percent. When firms retain lots of earnings it gets more complicated, with the growth of earnings increasing as more and more cash is retained in the firm. For those who can't wait until the fall release date of our book Dow 36,000 and want to start thinking through these issues, check out the Spring 1995 issue of the journal Contemporary Accounting Research, which provides a number of valuable review articles of the some of the relevant academic work.
? ? ? ?It is impossible to address every conceivable objection in a short article in the Wall Street Journal or Slate, which is one reason we are writing our book. We look forward to picking up this debate in the fall when we lay all our facts and arguments on the table. Until that time, we have a little homework assignment for Gottlieb and anyone else at Slate who would like to try. Microsoft's earnings have grown at an average rate of about 25 percent annually over the past 10 years. Microsoft pays no dividends. Were all Microsoft's earnings consumed in running in place, as Gottlieb's model suggests? Should Microsoft be worth nothing since it doesn't pay an actual dividend today? Was the price increase over that period justified? What would a fair price for Microsoft be if the risk premium were zero? We look forward to their answers.
--James K. Glassman and Kevin Hassett Washington
Bruce Gottlieb replies:
Glassman and Hassett write in to say I have misrepresented their argument. They do not, they say, value a stock by looking at future earnings. They even agree that this would be a big mistake. They claim to look instead at future cash flows to stockholders. The two authors might want to reread their original WSJ article, which says: "Assume that after-tax earnings are a reasonable estimate of the cash flow from a stock." In other words, there is no difference between my description of their argument and their own.
? ? ? ?Meanwhile, they have misrepresented my argument. My article does not say or imply that "real earnings growth only reflects retentions and that dividend growth must be zero" or that "all valuation techniques are out the window for firms that don't pay dividends." It simply asserts that, in calculating a firm's potential value, you can't assume that earnings are simultaneously retained and paid out--and that the Glassman-Hassett argument depends on precisely these conflicting assumptions. The argument is fallacious whether all the earnings are paid out, all are retained, or anything in between.
? ? ? ?Glassman and Hassett now claim that they will sort out the components of retained earnings to avoid double-counting in their forthcoming book. I look forward to it. But I boldly predict it won't work unless they have an entirely new thesis, since double-counting of corporate earnings is the core of their current one. The thesis is simply wrong and cannot be refined into sense. What's more, double-counting of all corporate earnings is how they get the figure in their title--Dow 36,000--so that will have to go if they even start down Refinement Road.
? ? ? ?It's obvious that our discussion must rest until this ambitious book is published. For right now, it is worth noting that this letter is the first place where Glassman and Hassett have explicitly admitted that equating earnings and dividends is a mathematical sin. They refer to it as "double-counting," which indeed it is.
?