All Too Human

1.

To most observers of the ups and downs of today’s stock market, it defies common sense when eminent economists assert that the stock market works according to logical principles. But most economists believe just that. According to generally accepted economic theory, stocks have a true or intrinsic value.

This value is based on several factors, the first of which is a company’s dividends. In the long run, a shareholder who holds on to a company’s stock can extract a certain cash value from it only when the management returns profits in the form of dividends to its owners. Today, of course, investors buy many stocks, such as Microsoft and Cisco Systems, that do not pay dividends. These companies are growing so rapidly that they continuously reinvest their profits in new products, research, and expanding productive capacity, and investors generally believe their stock prices will rise as they generate more profits. But eventually, even these companies, or so it is presumed, will begin to pay out part of their earnings in the form of dividends as their businesses mature. If Microsoft, for example, retained all its earnings even as its core businesses grew more slowly or stagnated, investors might sell the stock. If they were to hold on to the stock, they might demand some of those earnings in the form of dividends so that they might invest elsewhere (even though they would have to pay taxes at ordinary rates on dividends rather than lower capital gains rates on a rising stock price).

General Electric, for example, is a widely admired and fast-growing company, which no longer mostly makes electric turbines but also owns, among other businesses, a large credit company as well as NBC. But it is also a mature company which prudently pays about half of its profits to investors in the form of dividends—fifty-five cents a share, a little more than 1 percent of its recent stock price. For the economist, the value of GE’s stock depends on future dividends. When analysts devise mathematical models to determine the value of GE’s stock, they usually assume that dividends will grow at about the same rate as profits. Because future dividends are so closely related to profits, if the outlook for profits falls, investors should pay less for stocks.

Money also has a time value, however, and this second factor significantly affects the current valuation of a stock. You will not pay a dollar today for a dollar in dividends twenty years from now because you can earn interest income on today’s dollar. Therefore, when interest rates rise, stock prices will usually fall because dividends to be received in the future will be worth relatively less. (Rising interest rates may also reduce corporate profits, because the cost of borrowing rises.) When interest rates rise, investors will usually pay less for stocks because they can earn more on their money from interest in bonds. A stock price can be seen theoretically as based on the value of estimated dividends discounted for the level of interest rates. When that discount is made, the result is called the “present value” of the dividends, and this determines the true value of the stock. (This method of assessing stock is called the “dividend discount model.”)

A third factor in estimating intrinsic value is the uncertainty of future estimates of dividends. The prospects for some companies, such as those with new businesses, are more difficult to forecast than for others. Similarly, some companies’ dividends or earnings will be affected more by the unforeseeable fluctuations of the economy, the dollar, or interest rates than other companies’. The more uncertain the outlook for a company’s profits and dividends, the lower will be its intrinsic value.

The uncertainty of future profits and dividends, it may seem, leaves a particularly wide margin for error in the evaluation of stock prices. But even so, many and perhaps most economists believe that stock prices usually reflect a reasonable estimate of a company’s future performance, and that stock prices deviate from this intrinsic value only temporarily. It is not necessary, moreover, that most investors be right about stock prices. A relative handful of well-informed investors will sell or buy stocks if their prices are irrationally driven too high or too low. In fact, so “efficiently” do these investors gather and evaluate new information that it is extremely difficult for one well-informed investor to have an advantage over others.

Rarely, however, have the movements of the stock market tested the validity of this thesis as they do today. It is hard to open the financial pages of the newspapers and avoid a discussion about whether there is a speculative “bubble” in stock prices that is about to burst. Compared to earnings and dividends, stock prices on average have in fact never been higher, and rarely have they risen so far so fast. A broad measure of stocks, the Standard & Poor’s 500 index, has tripled since the 1990s after tripling in the preceding dozen years. The average “price- earning multiple” of the S&P 500—the prices of all the stocks in the index divided by their earnings per share—recently stood at thirty compared to a former high of only twenty. The Nasdaq index of stocks, which is mostly dominated by high-technology companies such as Intel and Sun Microsystems, has risen by nearly six times since the mid-1990s, and it almost doubled in 1999 alone. Many individual stocks are trading at price-earnings multiples of more than one hundred, levels that are almost unknown for any but start-up companies. Relative newcomers such as America Online and Yahoo are now valued at well more than General Motors or American Express.

Even more disturbing to the proponents of “market efficiency” is the recent volatility of stocks. To take only one example, last March the Nasdaq had climbed to above 5,000, a record high. Within a few weeks it fell to roughly 3,300. It is surely difficult to make a case that the intrinsic value of the companies that make up the index declined by nearly 35 percent on average in only a few weeks. (In the October 1987 market crash, stocks fell by more than 20 percent in a single day.) The outlook for future earnings did not suddenly change by more than one third in March and April. Interest rates did indeed rise early this year, but not nearly so much as to drive down the value of the Nasdaq stocks in a matter of weeks by some 35 percent.

In fact, there has probably not been a more volatile period for stocks in post-World War II history. The Nasdaq index has fluctuated by more than 5 percent in a single day several times this year. Major stocks, such as Cisco Systems and Oracle, have fallen by 5 to 10 percent in morning trading only to regain all the losses in the afternoon. Securities such as stock options, which give investors the right to buy or sell stocks within a restricted period of time, have never fluctuated as much in price before.

In view of such constant changes, the claim that the stock market reflects logical behavior increasingly seems to defy common sense, whatever economists may say. Stock prices clearly reflect the hopes, frustrations, greed, and fear of the millions of investors who now own stocks. Hundreds of millions of shares trade hands each day; millions of individual decisions are made. In recent years, good economic news has dominated bad, and few investors have experienced extensive losses. To many, today’s stock market euphoria can be seen as an expression of one of the human species’ most frequent conditions, overconfidence. Why should stock prices, so subject to human foible and fantasies of wealth, generally reflect the true value of companies?

In Irrational Exuberance Robert J. Shiller, a Yale economist and long-time researcher into the vagaries of financial markets, contends that common sense may be right after all. He even implies that stock prices may rarely reflect companies’ true values. Although Shiller’s book is informative and well argued, what is particularly important about it is that a respected economist has written it.

Shiller makes three central points. If history is at all an appropriate guide, the rise in profits in the 1990s, though rapid, cannot justify the rise in stock prices. “No price action quite like this has ever happened before in US stock market history,” he writes. Stock prices rose far more rapidly than corporate profits in the 1990s, leaving price-earnings multiples, as noted, at record levels. Profits rose even more rapidly in the 1920s and stock prices did not rise as much. Profits grew about as rapidly in the 1990s as they did in three other periods: following the depression of the 1890s; following the Great Depression of the 1930s; and after World War II; but stock prices rose more slowly in those years than they did in the 1990s.

A second important point is that stock prices time and again significantly overestimate or underestimate the actual “present value” of future dividends—which, as we have seen, is considered the basis of the true value of a stock. Since we know the level of dividends going back to the late 1800s, Shiller computes the present value of actual future dividends (discounted for the time value of money) for more than a hundred years. In fact stock prices gyrated widely compared to the steady rise in the present value of dividends. In the 1920s, for example, the S&P composite index increased by more than 400 percent but the present value of dividends increased by only about 16 percent. In the 1930s, the S&P composite index fell by more than 80 percent but the present value of dividends fell by only 3 percent. Shiller thus clearly shows that stock prices regularly overreact to economic expansions and contractions as well as to the proliferation of new technologies.

Shiller’s third important point is that there is a largely psychological herd effect in the stock market. Higher stock prices encourage people to invest and therefore to bid prices still higher. Another round of higher prices encourages still more stock buying, and so on. Investors may become more confident; they may become accustomed to the higher level of prices; and increasingly they are investing their winnings so they may take bigger risks. No matter the underlying value of the companies, Shiller believes the herd effect has driven today’s stock prices to unjustifiable levels.

To make clear just how bold Shiller’s contentions are, one has to understand how strongly the academic case for market efficiency has been made. Most economists allow for the possibility that stock prices will temporarily overshoot the mark. But for the most part, they argue that these deviations in the stock market will not persist because, as noted, there are enough well-informed investors to push prices back to sensible levels.

There is considerable evidence to support such claims. If stocks were generally mispriced, professional investors who intensively analyze particular stocks should be able to outperform the stock market averages consistently. But they do not. Burton G. Malkiel, a Princeton economist, has long been one of the most respected proponents of the efficient markets theory. In an updated version of his book A Random Walk Down Wall Street, he finds that most mutual funds did not outperform the major stock market averages in most of the past thirty years. Nor was there any consistency in their performance. Funds that did well in the 1970s generally did not do well in the 1980s or 1990s. Malkiel and many other economists contend that investors get and process information so quickly and accurately that there is little room for mutual funds to outperform one another over a period of, say, five or ten years. Stock prices thus move up and down as new information becomes available about a company, its industry, or the economy—information which cannot be routinely predicted. On the contrary, Malkiel writes, this new information seems to surface randomly. The lesson for most individual investors is to place money in a variety of diversified mutual funds or in index funds that reproduce the broad market averages, such as the S&P 500. Most professional investment managers apparently can neither anticipate changes in market direction consistently nor systematically find undervalued stocks that will soon rise in price.

Still, we have all heard of investors who did well over a long period of time. For example, Warren Buffett, who runs Berkshire Hathaway, and hedge fund managers such as George Soros and Julian Robertson outperformed stock market averages by a wide margin almost every year for a generation. Theorists of efficient markets acknowledge that some exceptional investors may outperform the pack. But during the last two to three years, even these legendary investment professionals have seen their portfolios significantly outpaced by the S&P 500. Buffett is holding fast to his strategy of buying proven companies with strong track records that are selling for uncharacteristically low prices, which generally means that he avoids new high-technology companies. After failing to match their earlier returns, Soros and Robertson have liquidated all or part of their hedge funds.

Academic researchers have also discovered anomalies in the stock market that suggest markets are not consistently efficient, and an increasing number believe markets are influenced by psychological behavior that does not reflect an investor’s rational economic interests. In stock market history, for example, small stocks usually rose faster in the month of January than large stocks. Researchers have also begun to question whether the group of well-informed investors who will theoretically correct distortions in prices is itself biased. Such investors might, on average, be too risk-averse to truly correct distortions, for example. After all, stocks can be mispriced when you buy them, but there is always the possibility that they will be mispriced even further when you sell them.¹

Many economists, however, have become purists about the logic of markets to a degree that neither theory nor statistical evidence justifies. Malkiel acknowledges that “the market does get carried away with fads and fashions.” But he nevertheless claims that the stock market crash of October 1987 might not have been a herdlike overreaction but rather a sensible response to rising interest rates and growing uncertainty during the previous two months. Why, then, did so much of the downturn occur in one day? If they were rational, investors should have more gradually unwound an overpriced market.

Some academic analysts claim, however, that investors rationally go along with the pack for a long time because stock price movements tend to persist. Evidence shows, for example, that if stock prices rise over a sustained period of time, they will usually continue to rise for a long time—a momentum effect. One reason may be that a rising stock price suggests that new positive information about a company is being disseminated, if only slowly. But when a piece of information to the contrary suddenly emerges, such as an increase in interest rates or a disappointing new product, it is, Malkiel writes, only natural to sell immediately. Stock prices can fall precipitously because they have been bid up for so long. This, he contends, does not necessarily mean investors were investing irrationally up to this point.²

Such an argument has been widely used by a number of economists who now claim that bubbles themselves—periods of unusually rapid run-ups in stock prices—can be rational. Peter M. Garber has written a brief analysis of the market bubbles, Famous First Bubbles, including the tulip mania in seventeenth-century Holland, and claims that the dramatic run-up in prices had a rational basis because investors were generally aware of the risks. Apparently, Garber believes, many of them understood how severe a price they might have to pay when the market turned.

Edward Chancellor has pointed out the flaw in these analyses in Devil Take the Hindmost, his insightful if excessively sweeping analysis of financial bubbles around the world. At some point, if investors are truly rational, the risk of reversal and sharp losses should eventually outweigh the momentum effect in investors’ calculations. Speculation, if it is rational, should unwind slowly as the risks become clearer. Yet time and again, investors have been left with enormous losses, and such market reversals have had serious effects on national economies. The implications of market efficiency have been taken considerably beyond what the evidence, and common sense, support.

2.

Still, the argument that stock prices can reflect rational assessments of the prospects for thousands of businesses has historical support. Periods of similar speculation in the past, such as in the early 1900s, the 1920s, and the 1960s, reflected general economic transformations—from mass production at the turn of the century to cars and electricity in the 1920s to television, jets, and fast food in the 1960s—that were ultimately beneficial for the economy as a whole, even if investors generally became overoptimistic and stock prices eventually fell sharply. Moreover, a much greater demand for stocks has been created by the changes in the market in recent years, including the proliferation of thousands of mutual funds and the participation of many more individual investors, especially as baby boomers save for retirement.

One cogent explanation of the rapid run-up in stock prices is provided by Jeremy Siegel in his influential book Stocks for the Long Run, first published in 1994 and recently reprinted in a revised and expanded edition. Siegel diligently calculated long-term returns to investors in the stock market going back to the early 1800s, including the reinvestment of dividends. He then compared them to returns earned by investing in both long-term government bonds and short-term government securities (with interest payments reinvested). The stock market returned roughly 7 percent throughout the two centuries while the returns on government bonds fell from 3 to 4 percent a year to 1 to 2 percent.

In general, according to Siegel, investors wanted stocks to yield an average return of 5 percent greater than the return on bonds. This 5 percent—the so-called equity premium—was the margin they felt would compensate for the greater risk of stocks. Stocks are riskier than bonds essentially because companies are required to meet their interest obligations on debt, and other expenses, such as wages, before they can pay dividends. In general, stock prices fluctuate much more than bonds do. Just when a stockholder wants to sell, prices might fall. Thus, investors demand a premium to buy them.

But Siegel’s findings confounded many experts. It turned out that stocks returned more than bonds during most periods and did not appear so risky after all. Consider someone who invests for twenty years. In eight out of every ten twenty-year time periods since 1802 (from 1802-1822 to 1977- 1997), an investment in stocks would have returned more than an investment in bonds. In all thirty-year periods since the 1870s, stocks outperformed bonds. Moreover, the average rate of return for stocks was usually considerably higher. For shorter periods, however, bonds were often the better investment. Over five-year periods, an investment in bonds outperformed stocks one third of the time. Moreover, in some of those five-year periods, an investment in stocks fell significantly.

The historical lesson was clear and surprising. Investors who were willing to hold a broad range of stocks for relatively long periods—generally ten years or more—did not merely do better in stocks most of the time, but their average return was significantly greater—on average, about 5 percent a year better since the 1920s. For many and perhaps most long-term investors, it made sense to own stocks. (Stocks performed well on average, even for shorter holding periods; but if you feel you can’t wait longer than five or ten years, the odds increase that bonds will be a better investment.)

Academic studies have thus supported an idea that has been gaining currency ever since stock prices recovered from the 1987 crash. If you hold stocks for the long run, they are less risky than was once thought. In addition, the spread of mutual funds and especially index funds have made it possible for individual investors to buy a widely diversified portfolio of stocks. The transactions fees for these funds have also been minimal.

Such a shift in investor expectations cannot entirely account for the rise in stock prices. Stock prices rose so far by 1999 and early 2000 that many Wall Street models based on future dividends already assumed that the equity premium had fallen below 5 percent. In other words, investors were already willing to pay more for stocks because they assumed the risk was lower if they were held longer. In claiming that stocks are not overpriced at their current levels, even after the fall in the Nasdaq this spring, these models also assume that profits will continue to rise rapidly. In fact, to justify the current level of stock prices, Wall Street dividend discount models generally assume that profits will grow by 10 or 11 percent a year for at least another five years, which is much faster than the economy’s expected rate of growth of perhaps 5 or 6 percent before adjusting for inflation. In the past five years, profits easily grew at this rate; but were they to continue to do so, that would mean that they would become a record share of the nation’s gross domestic product.

Stock prices might continue to rise if investors were to assume that stocks are no riskier than bonds and therefore require no equity premium at all to invest. This is the thesis presented in Dow 36,000 by James Glassman and Kevin Hassett. The authors, an economics writer and an economist, base their analysis on Jeremy Siegel’s findings, but make at least one extraordinary assumption (in addition to assuming rapid growth of profits). They claim that long-term investors—who are willing to take the most risk and therefore pay the highest prices—will essentially set the price of stocks. But few if any other economists believe that the equity premium—the return expected from stocks in order to compensate for their being riskier than bonds—will completely disappear; they observe that investors with shorter time horizons dominate the market, and for such investors the risks of equity investment should remain important. Even for twenty-year investors, history suggests that there is a significant chance that bonds will outperform stocks. And we can by no means be sure that history will repeat itself precisely, anyway.

To the contrary, a growing number of normally circumspect economists believe that it is highly unlikely that stock prices can keep rising rapidly, because dividend yields—the dividends per share divided by the stock price—are already so low and investors will continue to demand some equity premium for their riskier investment. In a research paper for the Social Security Administration, Peter Diamond, an MIT economist, writes for example that under any normal set of expectations for returns on stocks, prices either have to fall from current levels or stop rising for many years to come. To earn the historical return of 7 percent a year on stocks, starting from current high levels, would require that the value of stocks rise so far in the next ten to twenty years that they would be an extraordinarily high proportion of the GDP. Those who propose to place part of the Social Security trust fund in the stock market, and proponents—like George W. Bush—of Social Security privatization who would allow individuals to do the same, typically ignore prospects for a stagnant or falling stock market. In Social Security: The Phony Crisis, economists Dean Baker and Mark Weisbrot show how a falling stock market could place pressure on both future Social Security payments and privatization schemes because earnings from the trust fund could actually fall.

The remaining argument for stock prices rising above current high levels is the one that everyone, it seems, is talking about—a new economy based on information technology will continue to increase the rate of growth of both profits and GDP substantially. Thus, profits will not have to become an outsize proportion of GDP, nor will stock prices in general, because GDP itself will grow faster than anticipated. The extraordinary prices being paid for many high-technology stocks certainly cannot be explained simply because investors are taking longer-term views.

It is customary for market analysts who believe in this new vision to claim that the old methods of valuation, such as price-earnings multiples and dividend levels, must be discarded. That was precisely the view of Alberto Vilar, one of the most successful investors in new technology, when I talked to him in late May. Vilar was an early investor in companies such as Microsoft and Cisco Systems, and his fund, Amerindo Technology, has, remarkably, returned on average 38 percent a year for the past twenty years, even after this spring’s decline. “The standard approach—the Graham & Dodd textbook—does not apply to emerging technology,” he says. “Take your Graham/Dodd, your MBA, your CFA [Chartered Financial Analyst] and put them on the shelf.”³

Vilar believes the US is “in the first inning” of an industrial revolution that can take these stocks much higher. The prime beneficiaries will be two general sectors. First, sales of supplies, parts, machinery, and services between businesses will increasingly be ordered and processed on the Internet. Right now, such business-to-business commerce on the Internet amounts only to some $100 billion of the trillions of dollars of sales made each year. Vilar thinks it will become a multitrillion-dollar business in five to seven years. Thirty-fold growth would represent an average annual rate of growth of 60 percent or more a year. Second, the global telecommunications network will have to accommodate the Internet as it becomes even faster and more sophisticated. This will require the deployment of vast new technologies.

Rates of growth of 60 percent a year in major sectors of the economy are tantalizing. But even if we are in the midst of a full-fledged industrial revolution, there is much we do not know about how the future will evolve. Which companies will ultimately dominate such a market? Even if business-to-business Internet commerce reaches a few trillion dollars, what will the profit margins for corporate participants be? Profit margins for today’s retailers and telephone companies, for example, are low.

Such uncertainty leads to investment strategies that many people may not realize are unconventional. Instead of investing in proven companies, of which there are few in emerging technologies, Vilar invests in a portfolio of the best companies he can find that might gain, and maintain, a significant share of the new and largely unknown markets. Although Vilar might disagree, this kind of investing seems akin to what movie companies do. Movie executives admit they cannot pick hits beforehand but they invest in a diverse portfolio of movies. They count on a few becoming hits, which will compensate for most losing money or, at best, breaking even. “I have a list of twelve or fourteen stocks,” Vilar says. “Half of them will go up by ten times, but I don’t know which ones.”

Such attitudes help to explain why so many unproven stocks are trading at such high prices, and why even those that have already grown large are being bid up to prices almost never seen before. But the question is whether there is leeway for error. Consider Cisco Systems, almost everyone’s favorite stock. It makes the so-called routers of electronic transmissions that allow the Internet to function. Strategists at Morgan Stanley Dean Witter figure that at its recent price of about $55, Cisco traded at more than 120 times past earnings per share but about 80 times the next twelve months’ earnings. Wall Street analysts expect Cisco’s earnings to increase by 30 percent a year for five more years. Perhaps this is not unreasonable, if forecasts of future growth such as Vilar’s come true, but it is a high rate of earnings improvement considering how rapidly Cisco has already grown.

If investors demand that Cisco stock increase by a relatively modest 15 percent a year, and if earnings do grow by 30 percent a year, its price-earnings multiple five years from now would have to be 30. This is well above the long-term average of good stocks but certainly not implausible. On such grounds, Cisco’s high stock price is not outrageous. But if Cisco’s earnings grow at a strong but less spectacular rate of 20 percent a year, the price- earnings multiple would have to reach almost 70 five years from now—a very high level for a maturing company.

Some Internet stocks are selling at considerably higher levels, for example eBay, the leading auctioneer on the Internet, which is younger and faster-growing than Cisco and one of Wall Street’s current favorites. Analysts expect its earnings to grow by 50 percent a year, and Morgan Stanley figures it has recently been selling at about 180 times the next twelve months’ earnings. Such stocks may be future winners, but they are still the subject of highly optimistic projections.

For Shiller, patterns of overconfidence in the past have so regularly been repeated that there is little reason to think they are not recurring once again. Investors have justified stock market speculation on the basis of an allegedly new economy time and again in the past, and Shiller documents these fashions. In the 1920s, it was a “new world,” for example. In the 1950s, US News and World Report called it a “new era.” In part, these analyses of a revolutionarily new economy were correct. For example, in 1920, only about one in five families owned a car; by 1929, just before the stock market crash, three in five did. In 1948, 3 percent of Americans had a television set; by 1955, 76 percent did. Stock prices rose in anticipation of the gains of these new ages, but they consistently overshot the mark, and in the 1990s, stock prices rose even faster.

What is especially disturbing today is how sound arguments for long-term stock appreciation, such as Jeremy Siegel’s, are repeatedly overstated. Shiller believes the business press has had a large part in this exaggeration. To take only one recent example, I have heard a well-known analyst on cable-TV’s CNBC tell investors that buying stocks is safe as long as you are willing to hold them for five years. But Siegel’s work, as has been seen, clearly shows that five years is by no means long enough to be safe. One out of three times, the investor will have been better off buying government bonds, and stock market losses can be enormous over five years. In fact, there were several fifteen-year periods in the twentieth century where investors would have done better owning bonds than stocks—the latest was between the late 1960s and the early 1980s.

Several other factors may be feeding overconfidence among investors. The Federal Reserve has adopted a clearly asymmetrical policy, which is influencing investors’ expectations. When markets fall sharply, the Fed injects money into the economy and thereby lowers interest rates, as it did in October 1987 and again in 1997 and 1998 during the Asian financial crises and the demise of Long-Term Capital Management. Thus investors have come to expect the Fed to save the day and limit their potential losses in bad times. But when stock prices rise to unusually high levels, the Fed does not explicitly try to dampen the speculation. In an understated but clear and controversial book about his years on Wall Street, On Money and Markets, Henry Kaufman, probably the most influential Wall Street economist of the 1970s and 1980s, says that the Fed’s asymmetrical policies have encouraged investors to take undue risks. Contrary to Fed chairman Alan Greenspan, he believes that the central bank should be concerned about the rising stock market. Greenspan maintains that the Fed will establish more restrictive policies only to dampen the growth of the economy when it is inflationary, not to deflate stock market speculation. Partly because the Fed did not take action in other ways earlier, such as raising margin requirements—the amount of cash needed to buy stocks—Kaufman thinks a stock market crash is likely.

A second source of overconfidence is the widespread belief that policy-makers now know how to manage the economy. A severe recession is thought to be unlikely. But if, as a result of such confidence, investors bid up stock prices to unprecedented levels, compared to earnings and dividends, even a modest downturn in the economy can have severe repercussions in the financial markets. It would, moreover, be historically naive to believe that there will never again be a severe recession.

The wild card, of course, is the new economy. Wall Street analysts and even some economists who should know better, including Burton Malkiel, repeatedly claim an industrial revolution is underway that is the equal of any in history.⁴ So far, this is hard to prove. The scope of past industrial revolutions was remarkable. Changes in our way of life, I believe, were much greater in the 1920s than in the past ten years, for example. Not only were cars widely available for the first time but so were all kinds of electrical products, from washing machines and refrigerators to radios, record players, not to mention that other great communications network device, the telephone. Movies became popular, cities at last had sanitation, agricultural productivity soared on the basis of new chemical fertilizers, seeds, and insecticides, people went to high school in vast numbers, and antibiotics were about to become available.

Information technology is certainly changing our world but we do not yet know its limits. It may yet hold surprises “on the upside,” as Wall Street puts it. But, in my view, any reasonable reading of history strongly suggests that stock prices are so high that the potential rewards will increasingly fail to justify the risks.

We rarely seek to call war or love rational. They are part of the drama of human life. But as a wise book editor has often told me, so is the stock market. Financial assets are different from most other products because they are mostly a function of the future, and even on Wall Street the future remains beyond confident prediction. Robert Shiller has provided a calm and reasonable antidote to today’s euphoria. He believes the good times are over, but he admits that he has been saying this since the mid-1990s. At the very least, it is encouraging to hear a first-rate economist take on his own colleagues. His careful arguments, well-grounded in history, suggest to me that he is closer to the truth than most of them.