Misbehaving

The book opens Part I, “Beginnings: 1970–78”, with a classroom story. A hard midterm left Thaler’s students angry at an average of 72 out of 100, even though the curve meant their grades were no worse than usual. On the next exam he made the total 137 points. The average came out at 96, nobody’s grade changed, and the complaints stopped.

He uses the story to set up the book’s two characters. The models assume an idealised decision-maker he nicknames the Econ; the rest of us are Humans. His target is not people but the model: in his words, “the problem is with the model being used by economists, a model that replaces homo sapiens with a fictional creature called homo economicus”.

He names the gap. A detail that should not matter to an Econ but does matter to a Human — the total printed at the top of the exam, the price paid for food already on the plate, the date of an anniversary — is a “supposedly irrelevant factor”, or SIF. The rest of the book is a catalogue of SIFs and a history of the fight to get economists to take them seriously.

The chapter starts from a wine cellar. Richard Rosett, who chaired Thaler’s economics department, would not sell bottles that had risen in value, yet would not buy a bottle at the same market price either. The owned bottle and the unowned bottle were worth the same in the market and different amounts to him.

Thaler’s diagnosis is that people do not treat opportunity costs and cash costs alike. Passing up a chance to sell feels mild; handing over money feels like a loss. He named the result the endowment effect: things a person already holds are valued above the same things not yet held.

Once he started looking, Thaler kept a running list of such lapses on his office blackboard. Three items carry most of the book’s later work. Two friends with free tickets skip a game in a snowstorm, though they say they would have driven had they paid. A man who will not pay a teenager $10 to mow his lawn will not mow a neighbour’s for $20. A shopper will drive ten minutes to save $10 on a $45 clock radio but not on a $495 television.

Each item is a separate charge against the standard model — sunk costs, the gap between buying and selling prices, and savings judged as a share of the price rather than in dollars. The list is the seed of the rest of the book.

Reading Kahneman and Tversky’s work on prospect theory gave Thaler the frame his list lacked. The chapter reduces their value function to three features: people judge outcomes as changes from a reference point, not as levels of wealth; each extra dollar of gain or loss matters less than the one before; and losses weigh more than gains.

The third feature is the one he says left him “flabbergasted”, because it explained the wine cellar at once: “Roughly speaking, losses hurt about twice as much as gains make you feel good.” He calls loss aversion the most powerful tool the field has.

Thaler groups the standard objections to his work into a gauntlet he had to run at every seminar: people behave as if they were rational; the stakes in his examples are too small; people learn; and markets will correct the errors. The chapter answers each in turn. On stakes, he cites Caltech economists David Grether and Charlie Plott, who expected real money to make a known lab anomaly, preference reversals, go away; paying for real made them more frequent.

His reply on learning is the one most worth keeping. People get practice and quick feedback on small, frequent purchases, but very little on the large, rare ones — a house, a career, a retirement plan. On his account, if learning is what fixes mistakes, then decisions should get worse as the stakes rise, not better.

He gives the market objection a name, “the invisible handwave”, because nobody can say how a market turns a person into a rational agent. Most mistakes are not fatal, so nothing forces them out. His summary: “As cruel as the market may be, it cannot make you rational.”

The chapter opens on Vince, who paid $1,000 for a season at a tennis club, developed tennis elbow after two months, and played in pain for three months more rather than waste the fee. Economic theory says money already spent and not recoverable should not enter the decision. Thaler’s point is that for Humans it does.

He cites work showing the pull fades with time. Gym attendance rises in the month after a twice-yearly bill and drifts down until the next one. In Hal Arkes and Catherine Blumer’s experiment, theatre-goers who paid full price for season tickets went to more plays than those given discounts — but only for the first semester.

He also notes Barry Staw’s term for the organisational version, “escalation of commitment”: the more spent, the harder it becomes to stop.

At a regular low-stakes poker game, Thaler noticed that winners did not seem to treat their winnings as their own money. Gamblers call this playing with “the house’s money”. He and Eric Johnson then tested it. After being told they had just won $30, 70% of subjects took a coin flip to win or lose $9 — a bet most people refuse.

Losses work in two ways. After being told they had lost $30, only 40% took the same coin flip. But when offered a long shot that would let them win the $30 back, 60% took it. Being behind breeds risk-taking only when there is a chance to get back to even.

Thaler draws both lessons out of the casino. He links the house money effect to bubbles — “The house money effect—along with a tendency to extrapolate recent returns into the future—facilitates financial bubbles” — and the break-even effect to fund managers who take more risk late in a year they are trailing their benchmark, and to rogue traders who raised their bets as their losses mounted.

Part III, “Self-Control: 1975–88”, turns to the problem of wanting one thing and doing another. With Hersh Shefrin, Thaler modelled a person as two parts: a far-sighted planner and a series of short-lived doers who care only about now.

They borrowed the frame from the theory of firms, where an owner must control employees whose interests differ from the owner’s. The planner has the same two tools an owner has: incentives that leave the doer some choice, and rules that remove the choice. Thaler stresses the model is an “as if” device, not a claim that two people live in one head.

He is candid that the field later preferred a simpler formal model, David Laibson’s beta–delta approach, which values “now” above all later periods. He still finds the planner and the doer the more useful way to think about the problem.

Part IV, “Working with Danny: 1984–85”, covers a run of telephone surveys with Kahneman and Jack Knetsch on which business practices make customers angry. The best-known question: a hardware store raises the price of snow shovels from $15 to $20 the morning after a large storm. Eighty-two per cent called it unfair — though it is the textbook answer to a jump in demand.

The pattern they found ties fairness to the endowment effect. Buyers and workers treat the terms they are used to as theirs, so a worse deal reads as a loss. A cut in the nominal wage is judged unfair; a 5% raise during 12% inflation, which cuts real pay, was judged acceptable by 78%. Raising prices because costs rose is almost always accepted.

The practical rule Thaler draws for firms is to set the highest price as the regular one and call anything lower a discount, since “Removing a discount is not nearly as objectionable as adding a surcharge.”

To convince the experimental economists, Thaler, Kahneman and Knetsch ran markets in which half the subjects were given a coffee mug and could sell it to the other half. Owners asked about twice what non-owners would pay, and far fewer mugs changed hands than theory predicted.

The effect appeared within minutes of ownership — Kahneman’s “instant endowment effect”. Thaler pairs loss aversion here with inertia, which William Samuelson and Richard Zeckhauser called “status quo bias”: people stay where they are unless given a reason to move, and sometimes even then.

Thaler put a question to twenty-three executives at a print media firm, each the head of a separate publication. Would they take a project with an even chance of making $2 million or losing $1 million? Three said yes. The chief executive, watching, wanted all twenty-three: taken together they promised $11.5 million with under a 5% chance of losing money. One manager explained his no — a win meant at best a bonus of a few months’ pay, a loss might mean his job.

The chapter’s lesson is that judging bets one at a time makes a firm too timid, and the cure is to judge them as a pool. The fault sits in how the firm rewards risk, not in the managers’ nerve.

With Shlomo Benartzi, Thaler applied the same idea to stocks. Investors who check their holdings often see more losses, and loss aversion makes those losses loom large. In an experiment, subjects shown results eight times a year put 41% into stocks; those shown results once a year put in 70%. Their conclusion is that the equity premium “is so high because investors look at their portfolios too often.”

Part VI, “Finance: 1983–2003”, opens with Keynes’s comparison of professional investing to a newspaper contest where the prize goes to whoever picks the faces other entrants will pick. Thaler adds Keynes’s warning that “it is better for reputation to fail conventionally than to succeed unconventionally.”

He tests the idea with a game: guess a number from 0 to 100, aiming for two-thirds of the average guess. In theory everyone should say zero. When the Financial Times ran the game for its readers in 1997, the winning guess was 13. The lesson he takes is that the right answer depends on what you think others will do, not on the theory.

With Werner De Bondt, Thaler asked whether investors push prices too far after a run of good or bad news. They ranked New York Stock Exchange stocks by several years of past returns and compared the extreme winners with the extreme losers.

The losers won. In one test, over the five years after the portfolios were formed, the losers beat the market by about 30% while the winners trailed it by about 10%. The efficient-market view says past returns cannot predict future ones; this was a prediction drawn from psychology that came true.

The reply from defenders of efficient markets was that the losers must be riskier, and the extra return was pay for bearing that risk. Thaler explains why that argument can never be fully settled: any test of market efficiency is also a test of the risk model used.

By the standard measure, beta, the risk ran the other way. In the three-year tests, the winners averaged 1.37 and the losers 1.03. Adjusting for risk made the anomaly bigger.

Thaler splits the efficient market hypothesis in two: nobody can beat the market, and prices are right. This chapter is about the second claim, and about Robert Shiller’s 1981 paper on it.

Shiller’s test was that a forecast should not swing more than the thing it forecasts. Using data back to 1871, he found the present value of dividends moved little while stock prices moved a great deal. Thaler points to October 1987: prices fell more than 25% from a Thursday close to the next Monday’s with no news to explain it. For a fuller account of this line of work, see Irrational Exuberance.

A closed-end fund holds a portfolio of listed shares and trades on its own, so its price should match the value of what it holds. Thaler and two co-authors, one of them Andrei Shleifer, list four puzzles. New funds are sold with a broker’s fee of around 7% and then, within six months, usually trade more than 10% below their holdings. Discounts persist, they swing over time and between funds, and they close when a fund converts to the open-ended kind.

Their explanation is investor mood. Individuals own most of these funds, and the discounts moved with the gap between returns on small and large company stocks, which individuals also favour. Thaler quotes Benjamin Graham’s verdict on the funds: “an expensive monument erected to the inertia and stupidity of stockholders.”

In March 2000, 3Com sold a small slice of its Palm subsidiary to the public and promised to hand each 3Com shareholder 1.5 Palm shares later. Palm, priced at $38, closed its first day a little above $95. 3Com, which held the rest of Palm, closed at $82. The market was valuing the remaining 3Com business, which made a profit, at minus $61 a share — about minus $23 billion — and the gap lasted for months.

Thaler calls such cases the fruit flies of finance: rare, but clean enough to test, because here the right relationship between two prices is known for certain. He argued such cases were “the tip of the iceberg of market mispricing”; Eugene Fama’s view, he reports, was that “we had beheld the entire iceberg.” His wider point: if the market could get something this simple wrong, it seemed possible that the technology-heavy Nasdaq was overpriced too.

Part VII, “Welcome to Chicago: 1995–Present”, includes Thaler’s work with Cade Massey on the National Football League draft. Its emblem is the 1999 trade in which the New Orleans Saints gave up every pick they held that year, plus their first- and third-round picks for the next, to move from twelfth to fifth and take one running back, Ricky Williams.

They predicted early picks would be overvalued, for five reasons from the decision literature: overconfidence, forecasts that are too extreme, the winner’s curse, the belief that others want the same player, and the wish to win now. The data agreed. Measured as performance minus salary, the surplus a pick produces rises through the first round. Thaler says of it: “In all my years of studying market efficiency, this is the most blatant violation I have ever seen.”

The trade market for future picks implied teams were discounting next year at 136% a year. The advice was to trade down for more, later picks, and to lend picks this year for better ones next year.

Deal or No Deal let researchers test the poker-table findings with hundreds of thousands of euros at stake. Contestants were only moderately risk averse, and they reacted to the path they had taken, not just to the odds in front of them.

Frank, on the Dutch show, fell behind early, refused an offer of €75,000 — 74% of what his remaining cases were worth on average — and left with €10. Susanne, on the German show, was ahead and refused a sure €125,000 for an even chance of €100,000 or €150,000. She won the larger sum. The same two effects from the casino study appeared at a scale no laboratory could pay for.

Thaler judges finance the field where the “invisible handwave” has been tested hardest, because it holds cases, like 3Com and Palm, where part of a firm sold for more than the whole. Economists now study the “limits of arbitrage”, which, he says, “could just as easily be called the limits of handwaving.” He adds that would-be smart money can sometimes earn more by riding a bubble than by betting against it.

The poker-table chapter is the most direct tool on the shelf for a stock after a long run. A holder sitting on a large gain may be treating it as house money, and Thaler reports that fund managers trailing their benchmark take more risk in the last quarter of the year. Neither says where a price goes next; both say why a number that should give pause may not. The Palantir lens applies the first.

Reference prices and the fairness findings explain why a pricing change is fought in public: customers treat the old terms as theirs. That is the argument of the Salesforce lens, on a per-seat price under pressure. The inertia behind Save More Tomorrow is one candidate reading of the installed-base gap in the inference layer.

The finance chapters are the bridge to the rest of the shelf. Chapter 21 is Keynes’s beauty contest, set out in his own book on the General Theory page, and Chapter 24 is the start of Shiller’s case in Irrational Exuberance. The lens built from this book is at Richard Thaler.