November 25, 2002
Double Counting Retained Earnings--Glassman and Hassett

James Glassman demonstrates that he doesn't know the difference between payouts-to-shareholders on the one hand and retained earnings on the other. And so he double counts retained earnings in order to get to Dow 36000.

The Stock Market By James K. Glassman and Clive Crook: ...A good way to express values is by price-to-earnings ratios. Currently, the average stock sells at a P/E of about 25, meaning that it costs \$25 to buy \$1 worth of profits. Or, as a reciprocal, the stock's return in the first year is \$1/\$25 or 4 percent.

That doesn't seem like much, especially compared with a 30 year Treasury bond, paying 5.9 percent. If stocks and bonds have equal risk, shouldn't the P/E for stocks be 1/.059, or about 17?

No. There's a big difference between a stock and a bond. A stock increases its profits year after year, while a bond simply pays the same return.

Hassett and I assumed that earnings would rise annually at the same rate as nominal GDP (in fact, they have been rising far faster lately), or 4.9 percent, according to the Congressional Budget Office. Then we applied a simple finance formula: value = 1/(r-g), where r is the interest rate on 30 year bonds (5.9 percent right now) and g is that 4.9 percent growth rate. The result: 1/.01, or a P/E of 100.

Or, put another way, a stock can return a mere 1 percent when you buy it and still equal the cash flow from a 5.9 percent bond if you hold both to perpetuity.

a name="24094">From: James K. Glassman
To: Clive Crook
Friday, May 8, 1998, at 12:30 AM PT

Dear Clive,

No, I'm not pulling your leg about a Dow Jones industrial average of 36,000. To explain the reasoning, I need a little more space than the average Slate "Dialogue" message, so bear with me. You write that "over runs of years the market tends to revert to the mean." In other words, if the market has been flying high, it is bound to fall in order to get back to the averages.
Mean-reversion sounds perfectly logical. How many years in a row can anything go up 30 percent? But consider this: What if the big run-up in the stock market reflects a changing perception of market fundamentals? Then, there is no need for the market to reverse itself.
My colleague Kevin Hassett and I believe that, in fact, such a change is happening. It is a change from an irrational view of the market to a rational one.
What's so irrational? Well, investors have always demanded higher returns from stocks than the actual riskiness of those stocks requires. For example, between 1926 and 1997, Ibbotson Associates found that the average annual return for U.S. large-company stocks was 11 percent. The return for long-term government bonds was 5.2 percent. So the "risk premium" for stocks is nearly 6 percentage points a year.
That premium is way out of whack, since stocks, in truth, are no riskier than bonds over the long run. In fact, according to research by Jeremy Siegel, they are less risky than both bonds and Treasury bills over periods of at least 20 years.
So why the premium? Economists have tried to figure this puzzle out for years, but to no avail. Some, like Robert Shiller and John Campbell, have explained it with complicated but unconvincing models; others chalk it up to irrationality on the part of investors--a risk aversion that does not meet the facts.
Now, imagine a world in which the risk premium disappears--as it should. How would stocks be valued? Well, obviously, their prices would rise since their returns would fall. But by how much?
A good way to express values is by price-to-earnings ratios. Currently, the average stock sells at a P/E of about 25, meaning that it costs \$25 to buy \$1 worth of profits. Or, as a reciprocal, the stock's return in the first year is \$1/\$25 or 4 percent.
That doesn't seem like much, especially compared with a 30 year Treasury bond, paying 5.9 percent. If stocks and bonds have equal risk, shouldn't the P/E for stocks be 1/.059, or about 17?
No. There's a big difference between a stock and a bond. A stock increases its profits year after year, while a bond simply pays the same return.
Hassett and I assumed that earnings would rise annually at the same rate as nominal GDP (in fact, they have been rising far faster lately), or 4.9 percent, according to the Congressional Budget Office. Then we applied a simple finance formula: value = 1/(r-g), where r is the interest rate on 30 year bonds (5.9 percent right now) and g is that 4.9 percent growth rate. The result: 1/.01, or a P/E of 100.
Or, put another way, a stock can return a mere 1 percent when you buy it and still equal the cash flow from a 5.9 percent bond if you hold both to perpetuity.
All this, of course, is a back of the envelope calculation, but P/Es of at least 50 seem to be in the ballpark. Even at 50, P/Es would be twice as high as previous records. How can this possibly be so?
Let's take a look at dividends, which are one way that companies use their profits (another way is to reinvest them in their own businesses or to buy back their own stock, in both cases boosting the stock price).
Say that in 1988 you bought \$100 worth of General Electric stock. It paid a dividend amounting to \$3.50 a year, for a return of 3.5 percent. But that's just in the first year. By 1998, the dividend had risen to \$12 a year. So the return on your original investment is now 12 percent annually--and rising. By 2011, if GE continues to increase its dividend at the same rate it has since 1993, you will be earning an annual return of 50 percent. Don't laugh. Philip Morris currently pays a dividend that is 140 percent of the price of its stock in 1977!
Our P/E ratio of 50 to 100 (or more) does depend on some controversial assumptions. Interest rates could go through the roof, earnings growth could fall below our assumptions--or we could be wrong about the dissipation of risk. Investors could continue to demand high returns from stocks--thus holding down stock prices.
If we are right, however, our analysis justifies a Dow of about 36,000--not in five or 10 years, but right now.
There is a precedent. In the early 1980s, Michael Milken began preaching that the junk-bond market had priced too much risk into interest rates. Typical noninvestment-quality corporate bonds were yielding 9 percentage points more than Treasuries. But Milken's research showed that the historic default rates on such bonds justified spreads far lower--and that investors were far too risk-averse. Over time, investors changed their minds; they were educated into rationality. Today, junk bond spreads have dropped from 9 percentage points down to about 3 percentage points.
Thanks to this reassessment of risk, returns on junk bonds are lower--just as we expect stock returns to be ultimately. If you had bought junk bonds 15 years ago, you would have scored a huge capital gain as the price of your original bonds rose. This, too, is the pattern we expect with stocks.
Why are investors just now becoming less risk-averse? Two reasons: First, they're learning. Second, something has changed fundamentally in the economy.
I'll try to convince you on both these counts, Clive, in the next message.

Sincerely,
Jim

From: Clive Crook
To: James K. Glassman
Thursday, May 14, 1998, at 12:30 AM PT

Dear Jim,

All right, you've convinced me. You aren't kidding about 36,000. But I'm still puzzled.
I don't have much trouble with the first part of your argument, which concerns the "risk premium" on equities. You say equities have been safer than bonds over the long term. It's true this is odd, because for decades the stock market has priced equities to deliver a higher return than bonds, as if compensating investors for a risk that doesn't exist. At last, you say, the market is waking up to this. That's why equity prices are rising. They should keep rising, you reckon, until they reach a point where the return on equities and bonds is the same. So far, I understand you.
It's when you move on to the second stage, and work out what this equalizing price will be, that you lose me. Only when the Dow is 36,000 and the profit-to-equity ratio is 100, you reckon, will sensible investors no longer prefer stocks to bonds. This calculation must be wrong, but let's come back to that in a minute. First, a word on the risk premium.
I agree it's an anomaly. A key question is whether the historic premium was a straightforward error or a reflection of some intelligible but unexplained investor preference. If it's the first, as you say, then the mistake might be corrected all of a sudden, and the implications for the stock market would be big (but not, I emphasize, 36,000 big). If it's the second, then it's economists--not investors--who need to think harder. Perhaps it's both, but I would definitely not dismiss the second kind of explanation. (Suppose, as the economist Richard Thaler has argued, that investors are more upset by their losses than they are pleased by their gains. That would make them wary of the stock market, where the risk of short-term losses is great, leading them to demand a premium despite the market's long-term safety. Sounds plausible.)
Anyway, the premium has had a long history (hardly to be compared with the junk-bond episode), and investors probably noticed that the risk of long-term loss was small even before you came along to point it out. I'd be cautious about saying the premium has had its day.
But what I find really odd about your argument is this: Anybody who believes the risk premium has gone for good should find today's dizzy market about fairly valued, not vastly undervalued, as you do. At a P/E of 25, the market is yielding 4 percent. This, of course, is a real yield, because earnings rise with inflation. The corresponding yield on government bonds is about 6 percent, less, say, 2 percent for inflation, which is roughly 4 percent. In other words, there is no risk premium in this market. I expect one to reappear in due course, so I see the market as dangerous and expensive in the short term. I might be wrong about that. But if I thought the right premium was zero, as you argue, I would only think the market was now about where it should be--not that it's undervalued by a factor of four.
How then do you get to this zany valuation? Through a cash-flow calculation which seems to assume that companies can simultaneously pay their earnings out as dividends and reinvest them in the business in order to grow. Real (as opposed to nominal) growth in earnings doesn't come from nowhere, as your calculation assumes. It comes from investment, and if that investment is financed out of retained earnings, the growth of dividends is correspondingly reduced. As I expect you noticed, Jeremy Siegel from Wharton (whose book on equities over the long term seemed to set you off on this track) explained this in a letter to the Wall Street Journal last month. "[Your] analysis contains a serious flaw which vastly overstates the value of stocks," he said. "[The] greatest danger for the market today stems from unrealistic expectations of what stocks can earn. In no way can the high stock returns of the past five years or even the past 15 years persist." I agree.

Sincerely,
Clive
Posted by DeLong at November 25, 2002 07:25 AM | Trackback

Email this entry
Email a link to this entry to:

Message (optional):