Irrational Exuberance

The book takes its title from Alan Greenspan’s speech of 5 December 1996, and Shiller opens with the reaction: markets from Tokyo to Frankfurt fell within hours on two words in an otherwise dull talk. He then sets out the boom the phrase came to stand for. The Dow was near 3,600 in early 1994, passed 10,000 in March 1999 and peaked at 11,722.98 on 14 January 2000 — roughly tripling in five years while GDP rose less than 40% and corporate profits less than 60%.

His yardstick is the real S&P Composite divided by ten years of average real earnings, a smoothing he credits to Graham and Dodd. By that measure the ratio reached 47.2 intraday on 24 March 2000, higher than ever before; the nearest rival was 32.6 in September 1929.

He names three earlier high readings — 25.2 in June 1901, the 1929 peak, and 24.1 in January 1966 — and follows each forward. After September 1929, real returns averaged −13.1% a year for five years and stayed below 1% a year over twenty. After January 1966, real prices were down 56% by December 1974 and did not regain that level until May 1992.

This chapter is new to the second edition. Finding no long series of home prices for any country, Shiller built one for the United States back to 1890 from existing indexes and new work by his assistants.

The series shows no steady upward drift. Real home prices were 66% higher in 2004 than in 1890 — about 0.4% a year — and nearly all of that came in two bursts: around the Second World War, and from 1998 on. Between those spells, prices were mostly flat or falling. He suggests people think otherwise because they remember the nominal price of a house bought decades earlier.

He measures the strain of the latest rise against income. In the eight most volatile states, the median home went from 4.9 years’ per capita income in 1985 to 7.7 years’ in 2002. Abroad, the boom began sharply around 1997 in Boston, Los Angeles, Paris and Sydney, with London a little earlier.

Shiller lists twelve events outside the market that helped start the boom, among them what he calls the capitalist explosion, the Internet, the baby boom, more business news, analysts’ optimistic forecasts, the switch to defined contribution pensions, mutual funds, falling inflation, cheap trading and the spread of gambling. He treats them as causes of a mood, not as reasons the prices were right.

The Internet section asks the sharpest question. New technology need not raise the value of existing companies, since they do not own it, and the Dow held no Internet stocks until 1999; upstarts such as eBay and Amazon took the gains. His point is about perception: “What matters for a stock market boom is not, however, the reality of the Internet revolution, which is hard to discern, but rather the public impressions that the revolution creates.”

Those impressions, he argues, depend on how easily examples come to mind, and people who spend hours online can summon them at once. A vivid personal experience of a technology is easily taken as proof of its economic weight.

The chapter’s thesis is in its closing section: “The media actively shape public attention and categories of thought, and they create the environment within which the speculative market events we see are played out.” Daily price moves give reporters a story every day, and the competition for attention rewards emotional colour and familiar characters.

Shiller names one habit “record overload”. Reporters who look at enough data find something near a record on most days, and records measured in Dow points rather than percentages come easily. The result is a public that cannot tell when something truly new is happening.

He also tests the common belief that big price moves follow big news. A 1989 study he cites listed the fifty largest post-war moves in the S&P and the explanations the press gave; most matched no unusual news. His own survey of investors in the week of the 1987 crash asked what news they had been reacting to. News often works instead as the trigger for an “attention cascade”, in which one story turns the public’s attention to a market that then moves on its own momentum.

Shiller grants the efficient markets theory its strong point: day-to-day price changes cannot be very predictable, or getting rich would be too easy. His quarrel is with the long run.

His 1981 paper in the American Economic Review asked whether stock prices moved more than later dividends could justify, and found that they did. As the book reports it, the paper concluded that no movement of U.S. stock prices beyond their trend had ever been justified by later dividends.

He gives cases of plain mispricing. When 3Com sold 5% of Palm in March 2000, the Palm shares implied a value for the rest of Palm greater than the whole of 3Com, which owned it. The cost of borrowing Palm shares to sell short reached 35% a year by July 2000, which kept arbitrageurs from closing the gap.

The chapter’s chart plots each January’s price-earnings ratio from 1881 to 1995 against the next ten years’ real return. Shiller’s reading: “as a rule and on average, years with low price-earnings ratios have been followed by high returns, and years with high price-earnings ratios have been followed by low or negative returns.” He notes the exceptions and the small number of independent ten-year spans in the data, and still treats the relation as significant.

A common defence of high prices in the late 1990s held that the public had learned stocks were less risky than it once thought, helped by Jeremy Siegel’s Stocks for the Long Run (first edition 1994). If so, lower expected returns were simply the rational price of that knowledge.

Shiller’s main target is James Glassman and Kevin Hassett, whose 1999 book Dow 36,000 argued that investors were still learning the lesson and would keep bidding prices up, with early 2005 as “a sensible target date”. He points out that the book’s own pitch — a lot of money to be made while others caught up — sat oddly with its claim that stocks were as safe as government bonds.

His wider point is that the sense of having just learned an important truth recurs near peaks, and that what is “learned” in a long rise is often just the rise.

The last chapter turns to policy. Shiller’s first advice is diversification, and much of the rest concerns pensions. The shift since 1982 from defined benefit to defined contribution plans moved market risk onto employees, and he finds the warnings plans give about concentration weak and of little effect.

He calls it “a serious mistake” to replace social security systems with defined contribution plans invested in the stock market. Reform, in his view, should improve how the systems share risk between groups, not tie retirement income to market returns.

The remaining sections ask monetary policy to lean gently against bubbles, ask opinion leaders to offer steadying views, and ask that the public be helped to hedge its risks.

Chapter 5 is the reason a coverage record matters at all. If the press shapes what the public attends to, then how much is written about a company, measured against its own past, is evidence in itself — and so is a sudden silence. That is what this site counts.

Chapter 6’s test of order applies directly to the AI trade: did the story of a new era come before the price rose, or after? The AI capex page tracks the spending; this book is the check on the explanations offered for the prices that rode on it.

The ten-year earnings measure in Chapter 1 descends from Security Analysis, and the behavioural mechanisms overlap with Thaler’s Misbehaving. The lens built on this book is at Robert Shiller.