When I wrote the first edition of Stocks for the Long Run in 1993, the Dow Jones Industrial Average was 3500 and the prices of stocks relative to earnings were close to their historical average. My research demonstrated that stocks in the long run not only offered investors higher returns, but, when uncertain inflation was taken into account, entailed lower risk than fixed income assets. The evidence was overwhelming that stocks should be the cornerstone of every investor=s long-term portfolio.

But circumstances are quite different today. The great bull market that began in August of 1982 is one of the longest and strongest in stock market history even counting the recent decline. Over the past sixteen years the real return on stocks has been almost 14% per year, nearly double the average 7 percent real return that has characterized stock returns over the past two centuries.

The rise in stock prices since 1982 has substantially outpaced the rise in corporate earnings. Although per share earnings on the S&P 500 Stock Index has risen 200% since 1982, stock prices have risen 800%, four times as much. As of September 1998, the S&P 500 Stock Index is selling for nearly 23 times projected 1998 earnings, far above the 15 to 20 multiple that has historically characterized periods of low inflation. Throughout history, stocks have only sold at peak multiples in economic downturns, when earnings are abnormally depressed. Stocks have never been so highly valued seven years into an economic cycle, the current age of the US expansion. Given these circumstances, one can justifiably ask, AAre stocks still right for the long run?@ That question is the subject of today=s lecture.
 

Long-term Stock and Bond Returns

There are two consequences implicit in the current level of stock prices. Either future stock returns will fall below the historical norm, or future earnings growth is going to exceed its historical levels so that stockholders can continue to enjoy the high stock returns despite today=s high prices. Before we examine which of these two consequences is most likely to occur, it is important to review the historical risks and returns to equity investing.

Since the beginning of the nineteenth century, the average compound rate of return on a fully diversified portfolio of common stocks, dividends plus capital gains, has been 7% per year after inflation. The after-inflation return on long-term government bonds has been exactly one-half of the stock return, 32 % per year. The difference between these returns is called the Aequity premium@ and has averaged 32 % over the past two centuries.

Real stock returns have remained remarkably constant over time, averaging between 62 % and 72 % annually over the three major periods (1802 to 1870, 1871 to 1925, and 1926 to the present) that span the research in my book, Stocks for the Long Run. In contrast to stocks, the real return on fixed income assets has fallen markedly. Since 1926, the real return on bonds has averaged only 2.2% per year, and only 1.1% per year since World War II. The low return on fixed income assets since 1926, a period made into a popular benchmark by the historical yearbooks of Roger Ibbotson and Rex Sinquefield, boosted the equity premium over the past 71 years to about 6% per year.

But it is very likely that the 6% equity premium since 1926 overstates the forward-looking difference between stock and bond returns even if stock returns do not fall. Real bond returns computed over the past half century are biased downward by the unanticipated double-digit inflation of the 1970s. This inflation caused severe losses to the holders of fixed income securities who did not incorporate a sufficient inflation premium in their nominal yields. Today expected real returns on government bonds can be found by observing the 3.7 % yields now offered by the Treasury=s new inflation-indexed bonds. These indexed bond yields are extremely close to the real returns I found on nominal government bonds over the past two centuries and far higher than those calculated since 1926.
 

Long-term Risks

Although the 32 % historical equity premium may seem reasonable given the apparent volatility of stock returns, this premium appears unjustifiably large when the long-term risks of stocks and bonds are taken into account. Figure 1 displays the
sandard deviation of average after-inflation returns on stocks, bonds and bills since 1802. It shows that the behavior of stock and bond volatility changes dramatically when the investment horizon is lengthened.

The standard deviation of average annual stock returns declines rapidly as the holding period increases, far more rapidly than calculated on the basis of the short-term volatility of stocks. For thirty-year holding periods, the historical risk in equities is only about one-half what would be predicted on the basis of the standard deviation of one-year returns if stock returns followed a random walk. Although the random walk theory of asset price movements was long believed to be the best description of security prices, the 200-year data do not support the random walk hypothesis. The data indicate that stock returns show strong evidence of mean reversion B the statistical tendency for returns to tend towards a long-run average despite substantial short-term volatility.

The real returns on fixed income investments, on the other hand, show no such mean reversion. In fact, the data confirm the opposite: the real returns on bonds display mean aversion, the tendency for returns to wander away from, instead of being drawn to a mean in the long-run. This behavior is primarily due to the impact of inflation, which is an autocorrelated process that has a greater short-term than long-term predictability.

The mean reversion of stock returns and mean aversion of bond returns cause the real risk of stocks to fall below that of bonds or bills for holding periods of 15 to 20 years or longer. This implies that the risk premium of stocks relative to nominal bonds should disappear C and might even turn negative C for long-term investors. The fact that the return on equity far exceeds that on bonds implies that short-term risks dominate the equity market.
 

The Earnings Yield and the Equity Return to Investors

What does the current above-average price-to-earnings ratio tell us about future stock returns? Over the long run, the earnings yield, or inverse of the price-to earnings ratio, has been an excellent indicator of long-term real returns to investors. Since 1871 the average price-to-earnings ratio of the stock market has been 14, which translates into an earning yield of about 7%, equal to the average real return I have found investors have earned on equities since 1802.

The equality between the earnings yield on stocks and the real returns to stockholders has a strong theoretical basis. Returns to shareholders arise from real assets, such as capital, land, labor, and technology whose combination produces cash flows which, in contrast to the fixed returns from bonds, rise over time with the overall price level. The return to shareholders thus represents a real return, much like the cash flows from the treasury inflation-indexed bonds.
 

The Equity Premium

If future earnings growth does not accelerate, the current price-earnings ratio of the S&P 500 Stock Index implies an earnings yield B and hence an expected future real return of less than 5% for stocks. This is not far above the current real return from nominal bonds and implies a far lower equity premium than has been found in past data.

But what is the right premium that stocks should yield over bonds? Ever since the publication of AThe Equity Premium Puzzle@ by Rajnish Mehra and Ed Prescott in 1985, economists have attempted to explain the historical magnitude of the equity premium. Mehra and Prescott did not find sufficient correlation between stock prices and changes in aggregate consumption to warrant a premium anywhere near the 6% excess return to stocks that has prevailed since 1926, or even the 32 % premium that has characterized the past two centuries. In fact, in the context of standard economic models calibrated on historical data, unless investors are extremely averse to risk, it is difficult to explain an equity premium higher than one percentage point. And this difficulty is made worse by the long-term mean reversion of stock returns and mean aversion of bond returns that should reduce or eliminate the premium altogether.

There have been numerous attempts to reconcile the high historical magnitude of the equity risk premium with those derived from economic models. Explanations range from making the average investor far more averse to risk than we are led to believe, or by introducing other, possibly psychological factors that induce investors to shun stocks and cling to safe short-run assets (which might not be all that safe after inflation in the long run).

All this means it is quite possible that future stock returns will be lower because they should be lower. Equities have offered investors premium returns because they have been underpriced and over-discounted by risk-averse investors. As investors respond to the extensive literature extolling stocks that has come from both academia and Wall Street, the price of equities will be bid up to levels that will lower future stock returns to a level that is more consistent with the historical evidence. In other words, once everyone learns about how good stocks have been in the past, their price will be pushed to levels that yield lower future returns.

Another factor justifying higher stock prices stems from the drastic reduction in transaction costs in recent years. Until recently, it was impossible for the average investor to obtain the average Amarket return@ C taken as the capitalization-weighted return on all stocks C without incurring significant brokerage and dealer costs. When these costs are taken into account, the realized returns from a fully diversified portfolio of common stocks could be one to two percentage points lower than those calculated from historical data. In contrast, index funds today offer even the small investors the capitalization-weighted market return with costs of 0.20% per year and less.

The availability of low-cost indexing lowers risks and increases returns to investors, and may be a factor sending stock prices to higher valuations. As a result, investors in the future may realize returns lower than those calculated from market indices, although not necessarily lower when costs are accounted for. It is very likely that the recognition of the superior returns and low long-term risk of stocks, as well as the lower costs of investing in a diversified stock portfolio, drive part of the current increase in equity prices.

Higher Future Earnings Growth

Most investors, of course, hope that the current high valuation of stock prices is due to increased future earnings growth and not a signal for lower future returns. The extraordinary rise in earnings over the past fifteen years has raised the hopes of many investors that a Anew era@ of faster earnings growth is at hand, justifying higher price-to-earnings ratios.

There are two sources from which higher earnings growth could become a reality. One is an increase in technological progress that will spur economic growth and raise earnings. The second is the opening of the global market economy that introduced literally billions of consumers to goods and services that were formerly unavailable.

Earnings Growth and Technological Progress

Many stock investors base their current optimism on the revolution in information and communication technology. There have indeed been many extraordinary advances in communication through the centuries: the telegraph, telephone, radio and television to name a few. But the current environment is unique. Not only is the cost of communicating around the world dropping precipitously, the cost of storing information is falling even faster than the amount of information is increasing.

One can picture a world in the not too distant future where all books, manuscripts, and current data, as well as markets for financial and real assets and goods and services can be instantly accessed anywhere in the world by anyone with an inexpensive portable computer. This revolution will give individuals unprecedented free time to pursue the leisure activities of their choice, and may usher in a new post-Industrial Revolution. The revolution in information technology has already aided the restructuring of US corporations and lowered costs by reducing the need for middle management.

But the relation between economic growth and stock prices is quite tenuous. In the short-run, when an economy recovers from a recession, strong economic growth means higher earnings. But an increase in long-term economic growth is different. Although the aggregate earnings of firms rise because of an expanding economy, per share earnings, which are critical for valuation, need not. This is because firms must undertake capital investment C by either borrowing externally or retaining current earnings C to achieve aggregate earnings growth. This requires that firms float more shares or issue more debt to finance economic growth. Since 1871 all of the increase in real per share earnings has come from reinvesting retained earnings C earnings not paid as dividends that were either reinvested in the firm or used to buy back shares. No separate contribution from real growth is evident in the long-term data.

The data on earnings growth are corroborated by the economic data on output and capital. Historically the fruits of increased growth have accrued to workers in the form of higher wages, while the returns to capital have remained approximately constant through time. Increased output has required increased capital, and this is supported by the fact that the capital to output ratio has remained relatively constant over time. In sum, it does not follow that technological change, even if it increases aggregate economic growth, will increase share prices.
 

The transformation of big cap stocks

Despite my doubt that technological change in an of itself can give rise to faster earnings growth, the nature of the leading US corporations has been transformed significantly over the past 30 years. The largest stocks ranked by market value are vastly different than those that have occupied those positions in the past. Table 1 shows the 20 stocks with the largest market value in the S&P 500 Stock Index in the summer of 1998 and those in 1964. I chose 1964 for comparison since that year was also marked with low inflation, low interest rates, and high price-earnings ratios.

In 1998, fifteen of the top twenty stocks are Agrowth@ stocks, possessing above average price-to-earnings and price-to-book ratios. The average growth rate of per share earnings of these stocks has exceeded 15% per year over the past five years, and several: Microsoft, Intel, Cisco, have exceeded 20% per year.

In sharp contrast, fifteen of the top 20 firms in 1964 were considered Avalue@ or cyclical stocks, possessing relatively low P-E and book ratios, while eight of the top twenty were oil companies. Only five were considered growth stocks. The average earnings growth of the 1964 group was only 8% per year, and no firm=s growth exceeded 20%. It should be noted that these firms in 1964 were very large and very profitable companies. In fact the market valuation of the top 20 in 1964 was nearly one-half of the total value of the S&P 500 Stock Index, while today it is less than 30%. But in 1964 the top companies were growing no faster than the overall economy C in part because they were a large part of the economy.

The fact that today=s top 20 are growing twice as fast as the top 20 in the past is significant. Today=s leaders are significantly smaller large relative to the total economy than those in 1964. The total employment of today=s top 20 is only 12% of the employment of all firms in the S&P 500 Stock Index, and Wal-Mart employs nearly one quarter of this total. Today=s firms can, if circumstances remain favorable, continue to grow significantly faster than the economy without bumping into economy-wide constraints on growth rates.
 

Globalization of the world economy

The emergence of the global marketplace is one of the defining events of this century. Just ten years ago more than one-half of the world=s population was either too poor to buy Western brand name goods or prohibited from doing so by governments that dictated consumer choice. The developing nations currently consume only about one-quarter of the world=s output, yet they contain well over one-half the world=s population.

Prospects for global economic growth have dimmed considerably over the past year. A review of the sources of the current crises is important to understanding its resolution. Over the past decade as countries freed their economies of import and capital restrictions, an unprecedented, and in many cases excessive level of investment capital flowed to the emerging markets.

There was no way that all this capital could be invested productively, especially over such a short period of time. With local financial markets unable to signal the efficient allocation of resources, capital was committed to industries under the assumption that double-digit growth would continue forever. Oversupply was inevitable.

When capital fled and the local currencies crumbled, dollar-denominated debt overwhelmed other asset values. The current turmoil has raised the prospects of protectionism, widespread currency controls, and a contraction of international trade that threatens world economic growth.

Despite their gloomy short-term trends, the long-run prospects for most of these developing countries are not bleak. Their workforces are educated and well motivated. Their currencies are suffering from crises are not unlike the banking and liquidity panics that the industrialized world suffered in the nineteenth century. Then investors and depositors sought to convert to gold. Today they seek the modern symbol of stable value C the US dollar.

Notwithstanding the rise in trade and capital restrictions, the world will not retreat to the closed societies of the past. The communications revolution means that images of a more affluent lifestyle C through movies, videos, television, and now the internet C are or will soon be available to almost all the world=s inhabitants. Their desires to improve their standard of living cannot be suppressed, and Western technology and brand names will be avidly pursued as the developing nations recover from their current crises. The recovery of the global economy, although it might take a good part of the next decade, can still be a significant long-term source of profit for US firms.

Implications for Today=s Valuations

There are reasons to be cautiously optimistic that future earnings growth may accelerate somewhat from its historical norm. But history should make us cautious of proclaiming that AThis time is different.@ At market peaks optimists have frequently argued that stock prices were justified on the basis of Anew era@ conditions that negated old yardsticks for measuring stock values. More often than not, the traditional yardsticks reassert themselves and send stock prices downward.

But even if earnings growth does not accelerate and all the optimistic claims of the new era advocates prove false, stocks at current elevated levels still emerge as the long-run asset of choice. As noted earlier, if stocks remain at a high price-to-earnings ratio of 20, this still implies a 5% after-inflation return, and a 6% to 7% nominal return if inflation remains at its current unusually low levels.

Even if the promise of global markets boosting profits fades and the communications and technology revolution does not enhance earnings, stocks returns should still exceed bond returns in the long run. Since the long bond yields close to 5% with no inflation protection, stocks rate to beat bonds by at least one to two percentage points per year. Although this advantage is far smaller than found in the historical data, and probably below what most investors hope to obtain, it is not an unreasonable margin given the long-term historical risks and returns of a diversified stock portfolio.

Barber, Brad, and Odean, Terrance, AThe Common Stock Investment Performance of Individual Investors,@ working paper, 1998.

Mehra, Ranish, and Prescott, Edward C., AThe Equity Premium: A Puzzle,@ Journal of Monetary Economics 15 (March 1985): 145-61.

Siegel, Jeremy, Stocks for the Long Run, McGraw Hill, New York, 1998, 301 pages.

Siegel, Jeremy, and Thaler, Richard, AAnomalies: The Equity Premium Puzzle@ (with Richard Thaler), Journal of Economic Perspectives, 11 (1), Winter 1997, pp. 191-200.

Stocks, Bonds, Bills, and Inflation, Ibbotson Associates, Chicago, Illinois, 1998