Instead, they suggested, new factors--such as structural changes in the economy, new rules of economics, and the value of intangible assets and brands--justified the lofty stock prices. Today those valuations seem ludicrous, though the fundamental question remains: Has the market changed what it factors into share values?
Using a simple model based on changes in earnings, inflation, and interest rates, we found that the traditional factors alone explain most of the medium- and long-term movement in S&P 500 stocks over the past 40 years. We uncovered scant evidence that the market had changed the things it consistently factored into stock prices.
This reality doesn?t mean that the market never deviates from fundamental values: A strong case can be made that the performance of Internet and high-tech stocks during the second half of the 1990s added up to a "bubble." But such deviations tend to be short-lived. The economy and the market are closely connected, making the market?s long-term performance quite predictable--at least at an aggregate level. Given that connection, we can be confident that real long-term returns from stocks won?t exceed about 7 percent a year.
The record, clarified
What happened to the bull market? When we examined the performance of the S&P 500 from 1980 through May 2001, we discovered that a majority of the growth during those years resulted from the natural and expected growth of the market; only a small portion could be assigned to the amazing run-up in the Internet and high-tech sectors, which, some investors came to believe, had rewritten classical theories of how markets behave. Yet the plunge in the prices of a few "megacapitalization" stocks did play a major role in driving the markets down.
Between January 1, 1980, and December 31, 1999, the S&P 500 rose to 1,469, from 108, representing a compound annual growth rate of almost 14 percent (excluding dividends). In the 17 months that followed, from January 1, 2000, to May 31, 2001, the S&P 500 fell to 1,256.
We identified three factors responsible for almost all of the change in the index. The first two--growth in earnings and changes in interest rates and inflation--are precisely the factors that would traditionally have been expected to drive share prices. The third is the temporary and somewhat irrational emergence of megacap stocks. Together, these three factors account for over 80 percent of the run-up in stocks from 1980 to 1999. The retreat in the value of megacap stocks accounts for 50 percent of the decline in the market from January 2000 through May 2001.
Consider earnings growth. From 1980 to 1999, forecast earnings per share for the S&P 500 rose to $56, from $15. If the forward price-to-earnings ratio of the 1980 S&P 500 had remained constant over that time, earnings growth alone would have boosted the index by 302 points. This annual growth in earnings of 6.9 percent (3.2 percent in real terms) isn?t exceptional, since the nominal U.S. gross domestic product grew by 6.6 percent over the same period. As a result, corporate profits remained a relatively constant share of overall GDP.
Simultaneously, interest rates in the United States were falling dramatically, along with inflation. Long-term U.S. government bond yields peaked at nearly 15 percent in 1981 and then fell, more or less steadily, to 5.7 percent by 1999. Falling interest rates reduced the cost of capital for corporations, thereby enabling them to earn a larger premium as well as generating higher values for stocks.
To quantify the impact on valuations of falling interest rates and expected inflation, we built a simple model shaped by five variables: current earnings, inflation, interest rates, long-term earnings growth, and returns on equity. The model demonstrated that falling interest rates and inflation accounted for an increase in the S&P?s forward P/E ratio of nearly 8 points, corresponding to a 440-point increase in the S&P index. Combined, the increase in earnings and the decline in both inflation and interest rates accounted for 742 points (or 55 percent) of the increase in the S&P 500.
The megacap boost and bust
Much of the remaining increase can be explained by the uneven distribution of value within the index. In the relatively brief period from 1997 to 1999, a handful of companies--such as Cisco Systems, EMC and General Electric--attained huge market capitalizations as well as very high P/E ratios.
By 1999, the P/E of the 30 largest companies was double that of the other 470. Such a divergence was new: In 1980 and 1990 the average P/E ratio of the largest 30 stocks in the index (measured by market capitalization) was close to that of the other 470 companies and of the index as a whole. In fact, the outsized gains of the largest stocks from 1997 to 1999 had no precedent in the previous 40 years. The forward P/E ratio increase resulting from the emergence of this gap accounted for an additional 376 points of the increase in the S&P 500 from 1980 to 1999.
Taken together, earnings growth, inflation, interest rates and the megacap phenomenon explain more than 80 percent of the 1,361-point increase in the S&P 500 from 1980 to 1999. The rest of the change reflects a combination of other factors, such as the impact of the bubble on the index as a whole, the simplicity of our model, and the imprecision with which variables such as earnings are measured. Whatever the source of this residual, it largely vanishes from 1999 to 2001.
The same factors explain the drop in the S&P 500 since the end of 1999. While an increase of more than $1 in analysts? forecast earnings per share produced an increase of 34 points in the index level, the weight of other factors has produced a net decline. The impact on the index of a small increase in long-term interest rates and expected inflation was minor. However, the closing of the gap between megacap stocks and the rest of the index caused the index as a whole to lose 106 points.
Our conclusions about market behavior--conclusions that we derive from our analysis, which covered the period from 1962 to the present--apply to intervals as short as 3 to 5 years and as long as 40 years. Whatever the duration of the period, the primary factors driving the aggregate market are earnings, inflation, and interest rates, just as economic theory suggests. (Notwithstanding economic theory, the market isn?t driven by returns on capital, since in aggregate they are remarkably stable.) It is reassuring that the market actually works the way theory predicts it will.
Even so, the market sometimes does deviate from fundamental values; in view of the behavior of Internet and high-tech stocks over the past several years, it would be hard to argue otherwise. A strong case can be made that Internet and high-tech stocks did go through an upward deviation, or bubble--that is, a period when factors that would have validated high P/Es were absent.
Academic researchers are finding more and more instances of such deviations, though we cannot yet predict when they will begin or end--or even know with certainty when we are in the middle of one. Fortunately, in the United States these deviations have tended to be concentrated in a small number of stocks. By contrast, the behavior of the typical, or median, company is remarkably true to theory.
Predicting the future, broadly
If the past can be explained relatively easily, forecasting the future shouldn?t be extremely difficult--at least within broad bands. Although our analysis says nothing about short-term fluctuations, it can explain longer-term movements of the market.
In the light of past performance, the most you might expect from investing in stocks is a return of about 7 percent a year in real terms. Why? In the aggregate, future returns from stocks will be driven by earnings growth, changes in P/E ratios, and dividends. We have observed that U.S. corporate profits have remained a relatively constant 5.5 percent of US GDP over the past 50 years. Assuming that aggregate earnings increase at the same rate, history suggests that real corporate earnings will grow, along with GDP, at a rate of 3 percent to 4 percent a year.
If long-term interest rates don?t drop further, aggregate P/E ratios are about as high as they can be. Currently, expected inflation and interest rates are quite low; in fact, long-term rates haven?t been so low since the late 1960s, when, not coincidentally, P/E ratios were about the same as they are today. Assuming, optimistically, that P/E ratios remain constant and that earnings grow by 3 percent to 4 percent a year in real terms, stock prices alone should also increase by 3 percent to 4 percent a year.
The current dividend yield of roughly 1.5 percent and annual repurchases of 1.5 percent of outstanding shares generate an additional 3 percent of the return on stocks, for a total expected real return of 6 percent to 7 percent. Once again, this finding is in line with history. Jeremy Siegel, a professor of finance at the University of Pennsylvania?s Wharton School, has shown that the long-term real return on stocks during the past 200 years has averaged about 6.7 percent a year.
Stocks could exceed a real return of about 7 percent only if the GDP were to grow significantly faster than it has in the past or if the real cost of capital for companies were to decline. McKinsey research has found that their real cost of capital has been stable over the past 40 years. Real GDP growth has averaged 3.5 percent a year over the past 70 years or so and has been nearly 3.3 percent for the past 20. If economic growth slows significantly or if inflation and interest rates rise, returns from stocks could trend lower.
For more insight, go to the McKinsey Quarterly Web site.
Copyright © 1992-2001 McKinsey & Company, Inc.