The Rise in Stock Valuations
and
Future Equity Returns
The great bull market of 1982 to 2000 witnessed the greatest increase in total real stock returns in US history. Despite the robust 120% rise in real per share earnings, real stock prices increased by more than six-fold over this period. Real returns on equities averaged 13.6% per year during the bull market, nearly double the 7% real return that has
characterized the last two centuries of US stock market returns. By year 2000 the valuation of share prices relative to every traditional measure of firm value – earnings, dividends, sales, book value, replacement cost – had reached historic highs.
Many investors rationalized the high valuations in the late 1990s as consistent with the US entering a new era of rapid technological change that would generate ever-rising corporate profits and stock prices. This enthusiasm for technology reached a point that by March 2000, six of out the top 20 most valuable firms in the US had price-to-earnings ratios in excess of 100, levels that no other large firm had ever reached before.1
Although the turn of the millennium witnessed a bubble in technology stocks, the subsequent crash in the price of these shares, the economic recession, and the terrorist attacks of September 2001 did not deflate investors’ enthusiasm for equities. Relative to history, stock prices at the beginning of 2002 remained very high compared to forecast earnings. The rich valuation of the equity market has caused justifiable concern among analysts as to whether future real stock returns can match their historical average of 7% per year.
It is the purpose of this article to address that question by examining the causes of the rise in the P-E ratio over the past decade and determining whether this rise is supported by the fundamental changes that have taken place in the economy, the equity markets, and investors’ attitudes. The relation between the P-E ratio and long-term equity returns is derived and it is determined whether historical stock returns have been too high to be justified on the basis of standard risk and return analysis. The structural changes that have taken place in the economy and equity market over the past several decades are shown to give reason for a higher valuation level for stock prices. Finally the consequences of this higher valuation level for future equity returns are considered.
Price to Earnings Ratio
The most basic and fundamental yardstick for valuing stocks is the price-earnings ratio. Figure 1 depicts the P-E ratio since 1870 on the S&P 500 Index (or, before 1957, the
1
These six were Cisco, Sun Microsystems, Oracle, Nortel, EMC, and AOL. See “Big Cap Stocks are s Sucker’s Bet,” Wall Street Journal, March 14, 2000, p. .
Standard and Poor Composite average or a like group of stocks). The P-E ratio is computed for three earnings measures: (1) the last twelve months of reported earnings, (2) a centered five-year moving average of reported earnings, and (3) a five-year centered moving average of operating earnings.
The P-E ratio based on a moving average of earnings eliminates many of the spikes caused by recessions that temporarily depress earnings.2 In the last decade, investors have often used operating earnings (which exclude special charges such as restructuring, write-offs, etc.) in place of reported earnings. In gap between the two earnings concepts has been especially wide in the 2001 recession. Because both of these measures are used to judge firm value, P-E ratios computed for smoothed operating and reported earnings are shown in Figure 1.3
Figure 1 clearly shows that over the past several years the P-E ratio on stocks has reached historically high levels, especially when computed using reported earnings. The average P-E ratio based on reported earnings over this 131-year period has been 14.45, but the ratio reached xx in March 2000 and xy by January 2002. The record level of the P-E ratio has led a number of economists and money managers, such as Shiller (2000), Grantham ( ) , and Asness ( ) to conclude that the market is significantly overvalued and that much lower, and perhaps even negative equity returns are likely for an extended period into the future.
The P-E ratio and future stock returns
The P-E ratio has a fairly good record of forecasting future equity returns in the long run, because the inverse of the P-E ratio, called the earnings yield is analogous to the
current yield on a bond. The current yield on a bond measures the ratio of the “coupon,” or cash received per year over the price paid. The earnings yield on a stock is the annual earnings per share divided by the price and measures the cash return that a stockholder would receive if all earnings were paid out as dividends.
But an important difference between current yield on a bond and the earnings yield on a stock is that stockholders have ownership in the underlying assets of the firm, assets that rise and fall in value with the general level of prices. These assets include physical and intellectual property, land and other tangible assets.
The earnings yield on stocks is therefore properly considered a real or inflation-adjusted return. This contrasts to the nominal return earned from standard fixed income assets, where all the coupons are fixed in money terms and do not vary with inflation.4
2
This technique is motivated by Shiller (2000) who ascribes it to Benjamin Graham. 3
Although this paper is not the proper place to debate the complex issues surrounding the proper measure of earnings, it is my feeling that the best measure of sustainable earnings used to value stocks is somewhere between reported and operating earnings.
4
Inflation-indexed bonds, which the US government first floated in January 1997 combines the protection against inflation with a guarantee of the government.
The long-run data bear out the contention that the average earnings yield on the stock market is a good long-run estimate of real stock returns. The 14.45 average P-E ratio noted above corresponds to an average earnings yield of 6.8%. This earnings yields almost exactly equals the 6.9% real return that stockholders have earned over the same period. Shiller and Campbell (1996) find that the earnings yield is also a good predictor of ten-year real stock returns, using a P-E ratio based on the average of the past ten years of earnings.
If the P-E ratio is a good long-term predictor of future stock returns, then the high P-E ratios of recent years portend poor future returns for stocks. P-E ratios of 20 to 25 that have been calculated from forecast earnings predict future real equity returns of only 4% to 5% per year. Although these real returns may be higher than those currently available on government bonds, these returns are significantly smaller than what investors have received in the past.
Common Explanations of higher stock valuations
The record high stock valuations have prompted equity investors to search for reasons to justify these high prices. The two most prominent explanations offered are: (1) low inflation and low interest rates and (2) faster economic growth. Unfortunately both of these reasons for higher stock prices are flawed on theoretical and empirical grounds.
It is true that bonds are the major asset class that competes with stocks in an investors’ portfolio, so one might expect that low interest rates would be favorable for stocks. But since in the long run low interest rates are caused by low inflation, the rate of growth of earnings, which depends in large part on the rate of inflation, will be lower also. Over long periods of time, changes in the inflation rate cause changes in earnings growth of the same magnitude and do not change the valuation of stocks.
This proposition is borne out by the historical data. The average P-E ratio for stocks in the nineteenth and early part of the twentieth century was the same as since World War II, although before the war inflation averaged zero and interest rates were quite low while since the war, average inflation and interest rates have been significantly higher.5 The fact that stocks are claims on real assets makes their value immune to changes in the overall rate of inflation that significantly impact interest rates.
Faster economic growth
The second reason often given for higher P-E ratios relies on the belief that accelerating productivity gains that will drive earnings growth to higher levels. Since
5 It turns out that in the US over the past 50 years, low inflation is good for stocks because of the way that
the US tax system penalizes capital gains caused by inflation. This explanation, which will be discussed below, has little resemblance to the reasons offered by most investors for why low interest rates are good for stocks.
stock prices are the present value of future dividends, it would seem natural to assume that economic growth would raise future earnings and dividends and be an important factor influencing stock prices.
But this is not necessarily so. The determinants of stock prices are earnings and dividends calculated on a per share basis. Although economic growth influences
aggregate earnings and dividends favorably, economic growth does not necessarily increase the growth of per share earnings or dividends.
The reason for this is that economic growth requires increased capital expenditures and this capital does not come freely. Implementing and upgrading technology requires substantial investment. These expenditures must be funded either by borrowing in the debt market or by floating new shares. The added interest costs and the dilution of profits that this funding requires place a burden on the firm’s bottom line profits.
Many investors believe that investment in productivity-enhancing technology can increase profit margins and spur earnings growth to permanently higher levels. But “cost-saving investments,” frequently touted as a source of increasing profit margins, only temporarily affect bottom-line earnings. As long as these investments are available to other firms, competition will force management to reduce product prices by the amount of the cost savings and extra profits will quickly be competed away. In fact, capital expenditures are often undertaken not necessarily to enhance profits, but rather to
preserve profits when other firms competitively adopt the same cost-saving measures. The data on economic growth and earnings back up these points. Real per share earnings growth from 1871 through 2001 has been a paltry 1.25% per year, considerably below the nearly 4% growth rate of real GDP.6 Because of the requirement to fund capital expenditures, earnings per share growth falls far behind aggregate economic growth over the long run. In fact, the 1.25% growth in real per share earnings can be explained solely by the investment of retained earnings and not by aggregate economic growth.
If neither lower inflation, interest rates, nor faster growth are satisfactory
justifications of higher stock valuations, then investors must look elsewhere. Fortunately, it turns out that there are valid reasons for the increase in these valuation ratios. These involve the recognition by the investing public of the superior historical returns to equity and the changes in economic risk, taxation, and liquidity in the market.
THE EQUITY PREMIUM
Before exploring these other factors, it is important to consider whether the extra returns that investors have earned on stocks above those on safe assets – called the equity premium – have been justified on the basis of economists’ knowledge of the risk and return on stocks. The historical real returns on long-term government bonds have
averaged 3.5% over the past 200 years, about one-half the 7% return enjoyed by stocks.
6 Se
Therefore the equity premium has averaged about 3½% per year over the past two centuries. Over the past 50 years, primarily because the real return on bonds has fallen since World War II, the equity premium has averaged over 5%.
In 1985, just a few years after the great bull market in stocks commenced, economists Rajnish Mehra and Ed Prescott published a paper entitled “The Equity Premium Puzzle.”7 In their work they showed that the standard models of risk and return that economists had developed over the past generation could not explain the high level of that premium. They claimed that economic models predicted that either the rate of return on stocks should be lower, or the rate of return on fixed income assets should be higher, or both. In fact, their studies justified an equity premium as low as one percent or less.8
Mehra and Prescott were not the first to note that stock returns appeared to be “too high” in relation to the return on other financial assets. In 1937, Alfred Cowles III, founder of the Cowles Commission for Economic Research, constructed capitalization-weighted stock indexes back to 1871 of all stocks traded on the New York Stock
Exchange. Cowles examined stock returns including reinvested dividends and concluded: During that period [1871–1926] there is considerable evidence to support the conclusion that stocks in general sold at about three-quarters of their true value as measured by the return to the investor.21
Later research also found that historical equity returns were very high compared to those obtained by safe assets. The works by Fisher and Lorie in 1964, Ibbotson and Sinquefield in 1975 and my own book, Stocks for the Long Run, strongly supported the case for long-term equity investment.
One explanation for the rise in the level of equity prices relative to economic fundamentals is the recognition by investors that stocks were in fact underpriced during most of their history so that historical P-E ratios were not the correct yardstick by which to measure equity valuations today. Stocks are indeed worthy of higher prices that investors have paid for them in the past.
This underpricing of equities is earlier years resulted from several factors. Clearly one was the ignorance of investors of the true long-term risk and return characteristics on stocks. To ordinary investors, stocks appeared much riskier than the long-term data suggested. Investors had no knowledge of the considerable evidence that stock returns display mean reversion, a phenomenon that makes the long-term risk of holding stocks
7
Mehra, Rajnish and Prescott, Edward C., “The Equity Premium: A Puzzle,” Journal of Monetary Economics, March 1985, 15, 145-62.
8
Mehra and Prescott used the Cowles Foundation data going back to 1972. In their research that did not even mention the mean reversion characteristics of stock that would have shrunk the equity premium even more.
not as high as predicted by examining their short-term volatility. The history of equity returns, which makes such a compelling case for owning stocks, was simply absent.
Drop in Transactions Costs
Another factor contributing to the underpricing of equities was the high cost of acquiring and maintaining a diversified portfolio of equities. Even if investors were aware of the superior returns to equities that were calculated from stock indices, it would have been impossible to attain them.
The increase in the liquidity of stocks and the concomitant drop in transactions costs is an important factor increasing share prices. Charles Jones of Columbia University has documented the precipitous fall in stock trading costs over the last century.9 These costs include both the fees paid to brokers and the “bid-asked spread” or the difference
between the buying and selling cost for stocks. His analysis shows that the average one-way cost to either buy or sell a stock have dropped from over 1% of value traded as late as 1975 (before the deregulation of brokerage fees) to under .18% today.
The fall in transactions costs suggests that the price of obtaining and maintaining a diversified portfolio of common stocks, which is necessary to replicate index returns, could have easily cost from 1% to 2% per year over much of the nineteenth and twentieth century. Because of these costs, investors in earlier years purchased fewer stocks and were less diversified, thereby bearing more risk than implied by the volatility of the stock averages. If investors attempted to diversify their holdings by owning dozens of different companies from different industries, their realized real returns would have been
substantially below the seven percent that can be calculated from stock indices.
The collapse of transactions costs over the past two decades means that stockholders can now obtain a completely diversified portfolio at an extremely low cost.10 It has been well established that liquid assets, i.e., assets that can be sold quickly in the public market and at little cost on short notice, command a premium over illiquid securities. Through most of the past two centuries, stocks were far less liquid than today and therefore sold at a significant discount to such safe and liquid assets as government bonds. As stocks become more liquid, their price relative to earnings and dividends should rise.11
LATER.
9
“A century of stock market liquidity and trading costs,” working paper by Charles M. Jones, June 2000.
10
The cost of some index mutual or exchange traded funds is less than 0.2% per year for even small investors.
11
See John B. Carlson and Eduard A. Pelz, “Investor Expectations and Fundamentals: Disappointment ahead?” Federal Reserve Bank of Cleveland, Economic Commentary, May 1, 2000.
This was not the first time that stocks have been viewed as historically underpriced. As early as 1937, Professor Chelcie Bosland of Brown University stated that one of the consequences of the spread of knowledge of superior stock returns in the 1920s would be a narrowing of the equity premium:x
Paradoxical though it may seem, there is considerable truth in the statement that widespread knowledge of the profitability of common stocks, gained from the studies that have been made, tends to diminish the likelihood that correspondingly large profits can be gained from stocks in the future. The competitive bidding for stocks which results from this knowledge causes prices at the time of purchase to be high, with the attendant smaller possibilities of gain in the principal and high yield. The discount process may do away with a large share of the gains from common stock investment and returns to stockholders and
investors in other securities may tend to become equalized.12,13 Increased Stability of Economy
The drop in transactions costs and the recognition of the superior historical returns on equity are not the only reasons why stock prices have increased. The increase in the overall stability of the economy has played an important role in boosting stock prices. Although policymakers have not been able to eliminate the business cycle (and probably never will), monetary authorities are able to avoid the sharp swings in real output and inflation that in the past made stocks far riskier investments.
One could counter that if the real economy is more stable, why has there not been a trend towards more stable stock returns? It is true that the volatility of equity returns has remained relatively constant over time despite the greater stability of the overall economy. One reason for this is that firms have taken advantage of the increased
economic stability to leverage their equity stake, capitalizing on the gap between the cost of borrowing and the return on equity.
But the principal channel by which greater economic stability causes higher equity prices is quite different. Greater economic stability means that labor income, which is about three-quarters of all national income, has become more stable. This implies that workers can be persuaded to put a larger share of their savings in riskier assets such as equities since they feel more confident with their employment income.14 The strong gains in share ownership over the past decade were aided immeasurably by the public’s justifiable confidence that the government could avoid periods of severe unemployment and would establish a safety net, such an unemployment compensation, for those that were laid off.
12 Chelcie C. Bosland, The Common Stock Theory of Investment, 1937, op. cit., p. 132.
13 14
See Heaton, John, and Deborah Lucas (2000), “Portfolio Choice in the Presence of Background Risk,” Economic Journal 110 (January), 1-26.
As a consequence, the average worker is willing to take investment risks that would be considered unthinkable in the nineteenth and early twentieth centuries. Safe assets such as government bonds should lose their appeal relative to more risky assets, such as equity, thereby raising the level of stock prices.
Tax induced changes in Equity Valuations Inflation and the Capital Gains Tax
In the United States, capital gains taxes are paid on the difference between the cost of an asset and the sale price, with no adjustment made for inflation. An investor must pay taxes on any gain in price, whether or not he has realized any real gain. This means that an asset that appreciates by less than the rate of inflation—meaning the investor has lost purchasing power—will nevertheless be taxed upon sale.
Even though there is overwhelming evidence that the before-tax real return on
equities has remained constant over long periods of time, inflation dramatically alters the after-tax real return to taxable stockholders. For even a moderate inflation of 3%, an investor with a 5-year average holding period suffers a 50 basis point reduction in real return compared the case when there is no inflation. If the inflation rises to 6% the loss of return is more than 94 basis points.15 Furthermore, the inflation-tax is more severe the shorter the holding period of the investor because turning over stocks more frequently gives the government more opportunities to collect the tax on nominal capital gains. Other Tax Factors and their Impact on After-Tax returns.
The reduction in inflation over the past decade has not been the only favorable tax development for equities. The current capital gains tax of 20% and long-term gains and 18% on gains held more than five years is the lowest in more than sixty years.
Furthermore, firms have taken advantage of the low capital gains tax by repurchasing shares to produce lightly-taxed capital gains instead of paying fully taxable dividends. Another source of the shift to capital gains and away from dividends is the proliferation of employee stock options. The value of these options are related to the price of the shares and this price can be boosted using earnings to repurchase shares instead of paying dividends. All these factors have caused the dividend payout ratio of corporations to fall to an all-time low of 32%, mostly in response to the tax law and employee stock options.
The changes in tax treatment of stock returns have been very important in boosting the after-tax return on equities and are described in detail in Chapter 4. The reductions in the capital gains tax, the dividend-payout ratio, and the inflation rate have added about 1½ percentage points to the after-tax return of upper middle-income investors compared to the average levels of these variables over the past 50 years.16
15
These reductions are discussed in greater detail in Chapter 4, Stocks for the Long Run, 3rd edition, 2002.
New Justified P-E ratios
What do all these favorable developments mean for the stock market? First, they mean that the average historical P-E ratio or 14, or 15 is no longer appropriate in today’s market. In Chapter 4, I calculated how much changes in taxes and inflation altered the after real tax rate of return of investors. If we assume that investors bid up or down stock prices in response to changing taxes and inflation to obtain the same after tax real return, we can calculate how shifts in these variables impact the P-E ratio.17 This is shown in Figure 7-3. If in addition we incorporate the fall in transactions costs, the justified P-E ratio is further increased.
Figure 7-3 shows that these factors – taxes, inflation, and transactions costs – can explain the major movements of the market’s P-E ratio over the past 50 years. One can see that the justified P-E ratio rises to the low 20s excluding changes in transactions costs and the high 20s including the fall in transactions costs.
It is true that the assumption that investors demand the same after-tax real return whenever taxes and inflation changes is quite strong and may overstate their impact on the P-E ratio. Nevertheless, Figure 7-3 does not include other factors likely to boost P-E ratios, such as the greater stability of the economy or the increased demand for stocks generated by the investing public’s recognition of the superior returns to equity. As a result, I feel that the future price-earnings ratio on the market should be higher than the historical average. An average P-E ratio for the market in the low 20s is fully warranted as long as inflation stays low and tax policy remains favorable.
FUTURE EQUITY RETURNS
In Chapter 6 we pointed out that in the long run the historical real return on stocks of nearly 7% closely matched the average earnings yield, which is one divided by the
average P-E ratio. We have also shown in this chapter that the favorable economic factors justify a P-E ratio in the low 20s, which corresponds to an earnings yield of between 4% and 5%.
Is it possible that forward-looking real equity returns can be as low as 4% to 5%? Most certainly, but there are a number of favorable factors that make this forecast of too pessimistic.
16
See Table 4-2. Ellen McGrattan and Edward Prescott, “Is the Stock Market Overvalued?” Federal Reserve Bank of Minneapolis working paper, November 2000 provide another tax-motivated argument for high stock prices. They claim that the high and growing fraction of stocks held in tax-exempt form boosts their after-tax yield and stock price.
17
Assume Rt(Tt, Pt) equals the before tax rate of return, dependent on tax parameters T and stock price level P at time t. If we assume P changes as the tax parameters T change so that the
after tax real rate of return rt(Tt, Pt) is a constant, r* (which is dependent on the long-run growth characteristics and time preferences of the economy), then the valuation level Pt is a function of Tt and r* (See Prescott and McGrattan, 2000).
The formula that relates the real returns on equity to the earnings yield depends on the market value of capital equaling its replacement cost. When the market value of equity exceeds the replacement cost of capital (Tobin’s Q is greater than one), such as exists 2001, the earnings yield provides an underestimate of future returns. 18 This occurs because higher equity prices allow for a smaller offering of new shares to fund capital expenditures. Furthermore, with the sharp drop in the cost of information and
communication technology, the overall cost of productivity-enhancing capital investment has declined markedly. This leaves firms with more funds to buy back shares or increase dividends.
One of the reasons for the collapse of technology prices is the over-investment in many technology areas, particularly networking and telecommunications. Although this over-investment has been extremely painful for the firms producing such technology, it is ultimately favorable to those firms using this technology. The excess supply of capital lowers funding costs and will increase future shareholder returns. This is not unlike the aftermath of the railroad mania that hit the UK in the 1840s and produced a tremendous overbuilding of rail lines. After the prices collapse, the cheaper and faster transportation brought many benefits to the Victorian economy.19
Furthermore the US is still viewed in the global economy as the center for
entrepreneurship and the incubator of new technology. Financial and intellectual capital are flowing into the United States from abroad as the openness and flexibility of the US economy are seen as a magnet for the type of firms that will spur growth in this new century. This flow is enhanced since the economic policies of many developing countries do not promote stability or entrepreneurship. There is a large and growing flow of funds from abroad seeking safe haven, a factor that boosts US equity prices.
Finally, there is still room in the short run for US corporate profits to grow without distorting the returns to capital and labor. We noted in Figure 7-1 that the share of corporate profit relative to national income was at or slightly above its historical average in year 2001. But given the shift towards the corporate form of ownership and away from partnerships and sole proprietorships, one might have expected the corporate profit share to increase. Yet it did not and, as Figure 7-1 illustrates, the total return to capital,
including corporate profits and proprietors’ income, has had a declining trend. As a result, total capital income does not appear especially high and this leaves corporate profits with some room to grow before impinging significantly on the share of labor income in our national accounts.
18
The general formula relating the earnings yield to the expected return on equity, r, is r = EY + g(1 – RC/MV), where EY is the earnings yield (inverse of the P-E ratio), and RC/MV is the ratio of the replacement cost of capital to the market value. As long as RC/MV < 1, then r > EY. See Thomas K. Philips, “Why Do Valuation Ratios Forecast Long-Run Equity Returns?” The Journal of Portfolio Management, Spring 1999, pp. 39- 44.
19
See Edward Chancellor, The Devil Take the Hindmost,” 1999, Farrar, Straus and Giroux, New York, Chapter 5.
These positive factors should boost the future real return on equities to about 6%, somewhat lower than the long run average but above the 4½% return calculated from a market P-E in the low 20s. This means that the equity premium is likely to be in the range of 2% to 3%, about one half the level that has prevailed over the past 70 years.
P-E Ratio of S&P 500 (1871-2001)
45.7
14.5
(mean)
27.6*
23.6**
0
5
10
15
20
25
30
35
40
45
50
1871 1881 1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001
PE Rati
os
PE with Reported Earnings
* PE with Smoothed Reported Earnings * * PE wih Smoothed Operating Earnings
D
a
ta as of D
ecember
