June 17, 2002
How Far Out of Whack the Stock Market Really Is: The Answer? Very
Few people recognize how far out of whack the stock market still is today. Even though stock prices--especially the prices of high-tech stocks--have fallen substantially relative to their early-2000 peak values, there is still a large disconnect between current stock-market values and traditional valuation ratios relative to measures like earnings and dividends.
Typically, over the last fifty years, if you took a diversified and representative basket of stocks, you would have found that it sold for about 30 times its annual dividends and for about 18 times its annual earnings. It might go as low as 6 times earnings or as high as 23, but whenever it reached one of those boundaries you could bet that valuation ratios were about to move back to normal.
Today, however, stocks are selling for more than 30 times earnings, and more than 60 times dividends. You can leave the dividend numbers aside--there are, after all, better and cheaper ways of getting cash out to shareholders than dividends, and the mystery is not why dividends today are less than 50% of earnings, but why in the post-WWII past they were more than 60% of earnings (and still higher percentages of earnings back in the pre-WWII period). But why the extraordinarily high--unprecedented--earnings multiple?
You can take any of three positions:
- Investors today are more tolerant of risk than in the past--have a better assessment of just how much they should be willing to pay for stocks than investors in the past who were irrationally scared of equities.
- Investors today understand that the new economy is going to boost growth, and so are rationally willing to pay higher multiples than in the past.
- The stock market is coming down far and fast, soon.
My view? A 60% chance of the third--that the stock market is coming down far and fast, soon--alternative; a 30% chance of the first--that investors are more risk tolerant--alternative; and a 10% chance of the second--faster growth from the new economy--alternative. (The problem with the second is not that the new economy is fake, but that faster growth does not necessarily mean higher profits if it is accompanied by better information and search capabilities, and that faster growth is likely to be accompanied by higher interest rates as well--which put downward pressure on stock valuations.)
Posted by DeLong at June 17, 2002 02:58 PM
My friend and I constantly discuss what exactly these high PE ratios are telling us. I do think there is another possible outcome, that instead of a crash the market may just not grow at all throughout the next decade as earnings may catch up with current stock prices.
There are some other important issues that this brings up:
1)If there has been a change in risk tolerance, what does that mean to the risk premium? Does this mean that average real returns oo stocks in the future be lower than 7%? (I believe it was Mehra who has a paper saying that the risk premium was an artifact of high capital gains taxes, and with the current low rate the premium may be gone)
2)What effect does the life-cycle savings pattern of baby-boomers hold for future movment in the stock market? I posit that a good deal of the run up in the market in the 90s was due to baby-boomers reaching their prime earning years and starting to save greater amounts for retirement. Along with the explosion of 401(k)s, more money than before flowed into the market, leading to the great increase in stock holding in the population. This gives a reason why the stock market won't crash since much of this money will stay in the market until baby-boomers approach retirement. Even if no new money is added, there will be a desire to avoid current paper loses instead of selling (especially given the low level of interest rates on bonds). And forgive me if I don't trust broad fund managers to move portfolios well. So what happens as boomers follow the typical advice and move from stocks to bonds in the next ten years?
As to dividends-Part of the reason I think ther has been a move away from dividends to stock buy backs is that capital gains are now taxed at a lower rate than income for most holders. Is this move related to the current accounting problems, since it is easier to fake profits in statements than to come up with funds to pay dividends?
There's a fourth real possibility (not necessarily inconsistent with the other three) to be considered, as per Prof. Baruch Lev of the Stern Graduate School of Business, NYU, http://pages.stern.nyu.edu/~blev/
This is that the accounting rules used to compute earnings are breaking down and producing unreliable numbers as business value derives ever more from intellectual assets and human capital, rather than from the tangible and financial assets that the accounting rules were developed to handle.
As one example, acquisitions of intellectual assets and human capital generally must be expensed, while those of tangible assets are capitalized, which produces quite a difference on the financial statements.
When IBM bought Lotus it paid $2 billion that was allocated to the technology it was acquiring. But under GAAP rules it had to write off the $2 billion as a charge against earnings with no resulting asset appearing on the books -- whereas if it had paid the same amount for land or buildings there would have been no reduction of earnings and $2 billion more of assets on its books.
The first result in such a case is that the price/earnings and price/asset ratios rise. There is no real reason for the stock price to go down since fair value is acquired in the deal, and earnings and assets are only artificially reduced.
But later there'll be an artificial swing the other way and earnings will be inflated, because when the acquired technology starts producing a return there'll be no depreciation charge against it -- since there's no asset carried on the books to depreciate.
For IBM itself this case was no huge deal because it is so big and well understood. But for a lot of smaller businesses in the wide range of industries that now get almost 100% of their value from intellectual and human capital, real whipsaws can occur in financial statements that investors don't understand. Even inside managers may be misled by the distorted numbers into misallocating resources and capital.
The overall result for these is that distorted financial statements no longer show clear trend lines to investors. When in an optimistic mood, investors can be enthralled by exaggerated up numbers, and when in a pessimistic mood scared off by exaggerated down numbers. It's easy to see how this can contribute to overshooting in both directions. Small, innovative, tech-dependent firms are affected with the most volatility and that's just where the bubble-and-bust was most extreme. (If reported earnings don't mean much, I'll give money to anybody ... oops, now nobody).
Lev says that the basic relation between movement in reported earnings and stock price has been weakening since the 1970s, when intellectual assets started surging in value on the balance sheet. (Today even old-line industrials get like Dow get 70% of their value from non-tangible assets).
Anyone who roots around on his web site can find a lot of interesting stuff on this, both academic papers and press articles.
>My view? A 60% chance of the third--that the stock market
>is coming down far and fast, soon...
Why so pessimistic given the strong productivity numbers you report? Assuming they keep up.
"New economy" hyperbole apart, I'd think strong productivity growth would have to translate into either higher profits or more consumer spending (which is good for business cash flow, at least) or, if they don't spend, more saving with strengthening of balance sheets and easing pressure on interest rates. Or a mix. *Something* good would have to happen.
If the markets see at least something good happening, I'd give them a reasonable chance of holding ground, treading water, for a couple or three years or so until earnigns catch up. Not accept higher multiples permanently, but avoiding a crash-and-burn.
But then I invested in Kaypro over Microsoft, so if your weren't fibbing about your working ouija board I won't argue.
If equities are overvalued today, then equities should have outperformed Treasury bonds by more than the long historic average during the 1990s. This did not happen. During the 1980s and the 1990s, the total return on the S&P 500 has been around 3%/year above a constant-maturity 10-year Treasury (where all proceeds reinvested). This is broadly in line with the 1941-80 experience for the S&P. Today's high P/E (i.e. low earnings yield) is explained by (a) low bond yields, and (b) temporarily low S&P earnings due to recession. The gap between an H-P filtered earnings yield and T-bond yields today is in line with the historic average for this yield gap.
> Today's high P/E (i.e. low earnings yield) is explained by
> (a) low bond yields, and (b) temporarily low S&P earnings due to recession.
The P/E of 30 is on trailing (depressed) earnings, while on forecast future earnings it is 18. Plugging that in to the "Fed Model", which puts the P/E up against the return on 10-year T-notes, the market looks about fairly valued now. So you have a fair point.
Yes, forecast earnings are crapulous, the Fed doesn't have a ouija board, and Mr. Greenspan invests in T-bills. But this gives an alternate measure of the order of magnitude by which the P/E may be out of line, indicating it need not take until "the next decade", as a prior poster said, for earnings to catch up to prices.
Also, the P/E should be somewhat higher than the historic average because, as I mentioned above, the expensing of intellectual and human capital which represents a growing part of business value on a macro level systematically inflates the P/E; and the historical average includes periods of war and inflation and the 70s crunch.
It seems the main problem of the moment is past overinvestment in tech. Fortunately, unlike the Japanese overinvestment in real estate which will take 40 years to write down, tech depreciates in two or three.
Until then I wouldn't expect big growth in the economy or rising stock prices. But as long as the rest of the economy is basically OK (especially productivity) and absent new bad news, I don't see any reason for the market to plunge from here. So, by default, it seems most likely to tread water for a while, IMHO, FWIW.
Do you have a way to either post or link the H-P filtered data that you referenced? (I could do it myslef but am lazy). I guess what I am really interested in seeing is the deviations from trend the last decade produced.
I have nothing to add myself, except an old article from Martin Wolf of Financial Times because it made a very important point regarding bond yield and equities. And I think I read from him as well saying higher growth potential economy wide will lead to higher real interest rate, not higher valuation multiples.
A POOR DEFENCE FOR SHARE PRICES
Many depend on the relationships between bond yields and earnings yields to value equities. But they are wrong to do so, says Martin Wolf - Oct 24 2001 00:00:00
I know that I do not know what is going to happen to equity prices. All any economist can contribute is a little logic and a little history. From these one can derive two things: an idea of probabilities and an ability to distinguish the arguable from the nonsensical.
From 1997 to early 2000, the US stock market reached valuation levels that were, on the basis of empirically and theoretically sound indicators, higher than at any other time in the 20th century. The chances that the market would fall were higher than the chances that it would rise. Nevertheless, the market continued to rise until it ceased to do so.
As I argued last week (October 17), even at present valuations returns on US stocks will probably be far below their historic average of 6-7 per cent a year, in real terms. But there is a huge stockbroking industry dedicated to convincing its clients that this cannot be true. Their job is to find measures that show equities are cheap when they are expensive.
At present, the indicator of choice for market bulls is the relationship between the yield on bonds and the earnings yield on equities. Conveniently, in the US the ratio has fallen to 1.3 from a peak of 2.1 in January 2000. Unfortunately, as I stated last week, the ratio is "worthless", because it divides a nominal return by a real one.
What I thought would be an uncontentious statement proved quite the opposite. People have told me that there has to be a relationship between interest rates and earnings on equities. So how, they ask, can the yield/earnings ratio not be telling us something? The answer is that what the ratio says, if anything, depends on why it is changing.
Earnings on equity are a claim on a share of profits. Profits are derived from real activity, more precisely on a company's ability to use real things (its assets and its inputs) to make real things (its outputs). But the yield on a conventional government bond is a promise to pay a given stream of money until maturity. That yield will depend on the real interest rate and inflation. The real interest rate will depend, in turn, on time preference and the prospective rate of economic growth. The inflation component will depend on expected inflation and inflation risk. Inflation is a sustained rise in the general price level, not a change in relative prices.
To simplify, consider an irredeemable bond issued with a coupon of $7 and a face value of $100, when the expected real rate of interest is 3 per cent and expected inflation is 4 per cent. Suppose expected inflation falls to 2 per cent. If one ignores rounding errors, the price of the bond rises to $140 and the yield falls to 5 per cent. The owner of the bond has made a $40 capital gain - but the yield for any new purchaser gives the same real return as before.
Now what does the fall in inflation do to the earnings yield and price of an equity? To simplify, the answer is absolutely nothing. To see this, assume for simplicity that the real return demanded from equities is the same as the real interest rate - the "equity risk premium" is zero. Assume also that the pay-out ratio is 100 per cent. The price of an equity expected to give earnings this year of $3 is $100. If inflation were to be 4 per cent, next year's equity price would be $104. If inflation were 2 per cent, next year's price would be $102. If one purchased a share and sold it next year, the money return would be $7 and $5, respectively, with $3 from the earnings and $4 and $2, respectively, from the rise in the price. The nominal and real returns would be the same as on the bond, as required.
If inflation falls, the ratio of the bond yield to the earnings yield will fall but there will be no effect on the price of equities now. The change in the yield ratio will also say nothing about the advisability of buying shares. All it does is indicate an alteration in the expected path of future nominal earnings and equity prices.
That is the theory. What about reality? The era of disinflation, from the early 1980s to the late 1990s, was a period when bond yields and earnings yields both fell. In contrast, the era of rising inflation between 1948 and 1968, saw a rise in bond yields and a fall in earnings yields. In Japan, in the 1990s, there has been no relationship: bond yields have collapsed with no visible effect on the earnings yields on equities.
This is a broken indicator. So why is there ever a relationship? To answer this, one needs to consider the two elements: inflation and real interest rates.
In the 1950s and 1960s, a combination of buoyant profitability with the desire for hedges against inflation gave the inverse relationship between bond yields, which rose, and earnings yields, which fell.
In the 1970s inflation exploded, partly because of the adverse shift in the terms of trade, after the oil price shock, and partly because of struggles over the distribution of income between wages and profits. Profits were squeezed, even more so than published accounts suggested, because of the fiction of historic cost accounts. So bond yields soared, together with equity yields.
In the 1980s and 1990s inflation fell, as oil prices declined and labour markets were deregulated. Profits recovered and the quality of reported profits improved. As bond yields fell, so did the earning yield.
Thus the link between inflation, bond yields and the price of equities is complex. It is neither mechanical nor easily predicted.
Sometimes there are changes not in inflation but in real interest rates. If expected future growth rates rise, so should real rates of interest. But if the equity risk premium were unchanged, rising real rates of interest would lower the price of equities. The familiar argument that superior growth prospects necessitate higher prices for existing equity is wrong. The opposite is more likely.
There are at least four conclusions.
First, a change in the expected rate of inflation will alter the yield on bonds but should have no effect on today's price of shares.
Second, if that is not true, it is because inflation is having real effects - on wealth or the level and distribution of income - or is altering the accuracy of reported earnings.
Third, where changes in the yield on bonds reflect changes in real rates of interest, the price of equities should change. In general, higher prospective rates of growth are as likely to lower the equilibrium price of equities as to raise them.
Last but not least, there is no consistent exploitable relationship between the yield on bonds and the price of equities.
Interest rate movements are telling us something: but what they are saying is complex. People who market shares on nothing more than the ratio of bond yields to earnings are quacks.