assignment

BIASES AND BLUNDERS 23

Although rules of thumb can be very helpful, their use ca¡r also lead to systematic biases. This insight, first developed decades ago by two Israeli

+), has changed the way about thinking Their original work or rules of thumb-an- choring, availability, d the biases th"t

"re asõ-

.i"t.ã *iiliêã.tt.Th.ir research program has come to be known as the "heu¡istics and biases" approach to the study of human judgment. More recentl¡ psychologists have come to understand that theÀe heuristics and biases emerge from the interplay between the Automatic System and the Reflective System. Let's see how.

Anchoring

Suppose we are asked to guess the population of Milwaukee, a city about two hours north of Chicago, where we live. Neither of us knows much about Milwaukee, but we think that it is the biggest city in Wisconsin. How shor:ld we go about guessingf Well, one thing we could do is start with something we do know, which is the population of Chicago, roughly tluee million. So we might think, Milwaukee is a major city, but clearly not as big as Chicago, so, hmmm, maþe it is one-third the size, say one million. Now consider someone from Green Bay, Wisconsin, who is asked the same question. She also doesn't know the answer, but she does know that Green Bay has about one hundred thousand people and knows that Milwaukee is larger, so guesses, say, three times larger-three hundred thousand.

This process is called "anchoring and adjustment." You start with some anchor, the number you know, and adjust in the direction you think is ap- propriate. So far, so good. The bias occurs because the adjustments are_qp-

ically insufficient. Experiments repeatedly show that, in problems similar to our example, people from Chicago are likely to make a high guess (based on thei¡ high anchor) while tåose from Green Bay guess low (based on their low anchor). As it happens, Milwaukee has about 58o,ooo people.a : '"

Even obviously irrelevant anchors creep into the decision-making pro- .ìì cess. Try this one yourself. Take the last three digits ofyour phone number\ .. ' and add two hundred. Write the number down. Now, when do you think ' Attila the l{un sacke d Europel Was it before or after that yearf What is your best guessf (We will give you one hint: It was after the birth of Jesus. ) Even

22 HUMANSANDECONS

sibly be right would give that as an answer, but research by Shane Freder-

ick (zoo5) (who calls this series of questions the cognitive reflection test)

finds that these are the most popular arLswers even among bright college

students.

The correct answers are 5 cents, 5 minutes, and 47 days, but you knew

that, or at least your Reflective System did if you bothered to consult it. Econs never make an important decision without checking with their Re-

flective Systems (if they have time). But Humans sometimes go with the

answer the lizard inside is giving without pausing to think. Ifyou are a tele -

vision fan, think of Mr. Spock of Stør Trek.fame as someone whose Reflec-

tive System is always in control. (Captain IGrk: "You'd make a splendid computer, Mr. Spock." Mr. Spock: "That is very kind of you, Captain!")

In contrast, Flomer Simpson seems to have forgotten where he put his Re-

flective System. (In a commentary on gun control, llomer once replied to a gun store clerk who informed him of a mandatory five-.day waiting pe-

riod before br'tloog a weapon, "Five daysl But I'm mad now!") One of our major goals in this book is to see how the world might be

made easier, or safer, for the Flomers among us (and the Flomer lurking

somewhere in each of us). If people can rely on their Automatic Systems without getting into terrible trouble, their lives should be easier, better, and longer.

Rules of Thumb

Most of us are busy, our lives are complicated, and we can't spend

all our time thinking and analyzing everlthing. When we have to make judgments, such as guessing Angelina )olie's age or tlte distance between

Cleveland and Philadelphia, we use simple rules of thumb to help us. We

use rules of thumb because most of the time they are quick and useful.

In fact, tlrere is a great collection edited by Tom Parker tttJed Rules of Tbørnb. Parkerwrote the book by asking füends to send him good rules of thumb. For example, "One ostrich egg will serve 24 people for brunch." t "Ten people will raise the temperature of an average size room by one de- gree per hour." And one to which we will return: "No more than z5 per- cent of the guests at a university din¡rer party can come from the econom-

ics department without spoiling the conversation."

24 HUMANSAND ECONS

ifyou do not know much about E'ropean history, you do know enough to know that wheneverAttila did whatever he did, the date has nothing to do witì"your phone number. Still, when we conduct this experiment with our students, we get answers t}rat are more tfian tfuee hund¡ed years later from students who start with high anchors rather than low ones. (The right an- swer is 4rr.)

Anchors can even influence how you think your life is going. In one ex- periment, college students were asked two questions: (a) How happy are

, i you¡ (b) How often are you datingl when the two quesrions were asked in , ffr order the correlation between tìe two questions was quite low (.rr).

l- But when the question order was reversed, so that the dating question was asked first, tåe correlation jumped to .62. Apparently, when prompted by the dating question, rhe srudenrs use what might be called t}'e ..dating heuristic" to answer t}re question about how happy they are. ..Gee, I can't remember when I last had a date! I must be miserable." Similar results ca¡ be obtained from married couples if the dating question is replaced by a lovemaking question.s

In the language of this book, anchors serve as nudges. we can influence the figure you will choose i"

" ffiffiFriffiã-Ëy-.rr.r-so-subtly sug- t gesting a starting point for your thought process. when charities ask you

for a donation, they typically offer vou a range of options such as $roo, \t $25o, $r,ooo, $5,ooo, or "other." If the charity's fund-raisers have an idea

r, ofwhat they are doing, these values are not picked at rand.om, because the options influence rhe amount of money people decide to donate . people will give more if the options are $roo, $25o, $r,ooo, and $5,ooo, than if the options are $5o, $75, $roo, and $r5o.

.\ In many domains, the evidence shows that, within reason, the more you )i' ask for, the more you tend to get. Lawyers who sue cigarette companies oÊ

. È ten win astronomical amounts) in part because they have successfirlly in-\' duced juries to anchor on multimillion-dollar figures. clever negotiators often get amazing deals for their clients by producing an opening offer i that makes treir adversary thrilled to pay half that very high amount.

Atøiløbility

How much should you worry about hurricanes, nuclear power, terrorism, mad cow disease, alligator attacl$, or avian flur And how much

BIASES AND BLUNDERS 25

care should you take in avoiding risks associated with eachl What, exactly, should you do to prevent the kinds of dangers that you face in ordinary läef :+ jrp +_.,?

In answering questions of this kind, most people use what is called the availability heuristic. They assess the likelihood of risks by asking how readily examples come to mind. If people can easily think of relevant ex- amples, they are far more likely to be frightened and concerned than ifthey cannot. A risk that is familiar, like *rat associated with te¡rorism in the af- termath of 9 /r,will be se en as more serious than a risk that is less familiar, like that associated with sunbathing or hotter summers. Homicides a¡e more available tha¡ suicides, and so people tend to believe, wrongly, that more people die from homicide.

Accessibility and salience are closely related to availability, and t}rey are important as well. Ifyou have personally experienced a serious earthquake , you're more likely to believe that an earthquake is likely t}ran if you read about it in a weekly magazine. Thus vivid and easily imagined causes of death (for example, tornadoes) often receive inflated estimates of proba- biJity, and less-vivid causes (for example, asthma attacks) receive low esti- mates) even if they occur with a falr greater frequency (here a factor of twenty). So, too, recent events have a greater impact on our behavior, and on our fears, than earlier ones. In all these highly available examples, the Automatic System is keenly aware of the risk (perhaps too keenly), without having to resort to any tables of boring statlstlcs.

The availability heuristic helps to explain much risk-related behavior, in- cluding bot}r public and private decisions to take precautions. Whether people buy insurance for natural disasters is gready affected by recent ex- periences.ó In tÏe aftermath of an earthquake, purchases of new earth- quake insurance policies rise sharply-but purchases decline steadily from that point, as yi$"d_4gg¡ori-es qççede. Iffloods have not occurred in the im- mediate past, people who live on floodplains are far less likely to purchase insurance. And people who know someone who has experienced a flood are more likely to buy flood insurance for t-hemselves, regardless of the flood risk they actually face.

Biased assessments of risk can perversely influence how we prepare for artd respond to crises, business choices, artd the political process. When In- ternet stocks have done very well, people might well buy Internet stocks,

' t. /n'\ ' Í:/ J''c

4^ BlasEs ÀNÞ BLUNDER' 37

tdп L^Ì' attention, so consider the following information campaigns: (a) Ifyou use energy conservation methods, you will save $3jo per year; (b) If you do not use energy colserv4þn ne3¡hods, you will lose $3So per year. Itt;rns 4 out that information campaign (b), framed in terms of losses, is far more effective than information campaign (a). If the government wants to en- courage energy conservation, option (b) is a stronger nudge.

works because tend to be somewhat mindless, passive

questions would pro- duce a different answer. One reason they don?t do this is that they wouldn't know what to make of the contradiction. This implies that frames are powerful nudges, and must be selected with caution.

So Whatl

Our goal in this chapter has been to offer a brief glimpse at human fallibility. The picture that emerges is one of busy people trying to cope in a complex world in which they cannot afford to think deeply about every choice they have to make. People adopt sensible rules ofthumb tlrat some - times lead tfrem astray. Because they are busy and have limited attention, they accept questions as posed rather than trying to determine whether their answers would vary under alternative formulations. The bottom line, from our point of view, is that people are, shall we say, nudge-able . Their choices, even in life's most important decisions, are influenced in ways that would not be anticipated in a standard economic framework. Here is one final example to illustrate.

One of the most scenic urban thoroughfares in the world is Chicago's Lake Shore Drive, which hugs the Lake Michigan coastline that is the city's eastern boundary. The drive offers stunningviews of Chicago's mag-

nificent sþline . There is one stretch of this road that puts drivers through a series of S curves. These curves are dangerous. Many drivers fail to take heed of the reduced speed limit (25 mph) and wipe out. Recend¡ the city has employed a newway of encouraging drivers to slow down.

At the beginning of the dangerous curve, drivers encounter a sign painted on the road warning of the lower speed limit, and then a series of white stripes painted onto the road. The stripes do not provide much ifany

Framing

Suppose that you are suffering from serious heart disease and that your doctor proposes a grueling operation. You're unde¡standably curious about the odds. The doctor says, "Of one hund¡ed patients who have this operation, ninety are alive after five y€ars." What will you dol Ifwe fill in the facts in a certain way, the doctor's statement \Mill be pretty comforting, and you'll probably have the operation.

But suppose the doctor frames his answer in a somewhat dìfferent way. Suppose that he says, "Of one hundred patients who have this operation, ten are dead after five years." Ifyou're like most people , the doctor's state - ment will sound pretty alarming, and you might not have the operation. The Automatic System thinks: "A significant number of people are dead, and I might be one of t-hem!" In numerous experiments, people reactvery differently to the information that "ninety of one hundred are alive" than to the information that "ten of one hundred are dead'-even though the

,i .' content ofthe two statements is exactly tÏe same . Even experts are subject .{ì

.. to framing effects. \4hen doctors are told that "ninety of one hundred are ,:: alive," they are more likely to recommend the operation than if told that

"ten ofone hundred are dead."l4 Framing matters in many domains. \44hen credit cards started to become

popular forms of payment in the r97os, some retail merchants wanted to charge different prices to their cash and credit card customers. (Credit card companies typically charge reteilers r percent of each sale.) To prevent this, credit card companies adopted rules that forbade their retailers from charging different prices to cash and credit customers. FIowever, when a bill was introduced in Congress to outlaw such rules, the credit card lobby turned its attention to language. Its preference was that if a company charged different prices to cash and credit customers, the credit price should be considered the "normal" (default) price and the cash price a dis- count-rather than tåe alternative of making the cash price the usual pricé and charging a surcharge to credit card customers.

The credit card companies had a good intuitive r¡nderstanding ofwhat psychologists would come to call "framing." The idea is that choiçes de- pend, in part, on the way in which problems are stated. The point matters a great deal for public poliry. Energy conservation is now receiving a lot of

36 HUMANSAND ECONS

RESISTINGTEMPÎATION 5T

might see him take the money he has won and Put it i¡to one pocket and

put rhe money he brought with him to gamble that evening (yet another

mental account) intq3.{iË-erentJocket. Gamblers even have a term for

this. The money thãt has recently been won is called "house money" be-

cause in gambling parlance the casino is referred to as the house . Betting

some of t}¡e money that you have just won is referred to as "gambling with

the house's money," as if it were, somehow, different from some otler

kind of money. Experimental elidesce reveals that people are more willing

to gamble with money that they consider house money.4 .'

This same mentality affects people who never gamble. v/hen invest-

ments pay off, people are willing to take big chances with tleir "win- nings." For example, mental accounting contributed to the large increase

in stock prices in the r99os, as many people took on more and more risk

with the justification that they were playing only with their gains from the

past few years. Similarly, people are far more likely to splurge impulsively

on a big luxury purchase when they receive an unexpected windfall than

with savings that they have accumulated over time, even if those savings

are firlly available to be sPent'

Mental accounting matters precisely because the accounts are treâted as

nonfungible . True, rhe mason jars used by Dustin Hoffman (and his par-

ents' generation) have largely disappeared. But many households continue

to designate accounts for various uses: children's education, vacations, re-

tirement, a¡d so forth. In many cases these are literally different accounts'

as opposed to entries in a ledger. The sanctity of these accounts can lead to

seemingly bizarre behavior, such as simultaneously borrowing and lending

at very different rates. David Gross and Nick Souleles (zooz) found that

the typical household in their sample had more than $5,ooo in liquid assets

(typically in savings accounts earning less than 5 percent a year) and nearly

$3,ooo in credit card balances, carrying a typical interest rate of t8 percent

or more. Using the money from the savings account to pay off the credit

card debt amounts to what economists call an arbitrage oPPortunity-

buyrng low and selling high-but the vast majority of households fail to take advantage.

]ust as with Christmas clubs) though, this behavior might not be es stu-

pid as it looks. Many of these households have borrowed uP to the limits

that their credit cards set. They may realize that if they paid off the credit

50 HUMANS AND ECONS

\ ),,.,

coønting.Mental accounting is the system (sometimes implicit) that house-

holds use to evaluate , regulate, and process their home budget' Almost all

.¡ ' of us use mental accounts) even ifwe 're not awafe that wè're doing so. The concept is beautifully illustrated by an exchange between

the actors

Gene Hackman and Dustin Hoffman in one of those extra features offered

onDvDs.IlackmanandHoffmanwerefüendsbackintheirstarvingartist

other"utilities,"andsoforth'Hackmanaskedwh¡ifHoffmanhadso much money in jars, he could possibly need a loan' whereupon

Hoffman

Iloffrnan's kitchen counter.

At the household level, violations of fungibility are everyr,vhere. One of

t}remostcreativeexamplesofmentalaccountingwasinventedbyafinance

bler who is lucþ enough to win some money early in the evening' You

*You might think that this d'eprives the United Way of money' but not so' The pro-

fessorhastomakesurehisintendedgiftislargeenoughtocoverallhismishaps.

INgREASING ORGAN DONATIONS IAI

system) only rz percent of the citizens gave their consent) whereas in Aus-

ûia, nearly everyone (99 percent) did'

-+ :tF. . +¿è

Some ComPlexities

So far, presumed consent looks awfrrlly good, but we must stress

per million people, compared with a bit more than twenty donors per mil-

lion in the United States. But the U.S. donation rate is higher than in many

presumed consent countries because of the superiority of the American

medical system in quickly matching consenting donors with recipients, de-

livering the organs, and performing successfirl transplants' The default

consent rule, therefore, is not the only thing that matters. Still, carefirl sta-

tistical analyses by the economists Abadie and Gay (zooa) find that' hold-

ing everlthing else constant, switching from explicit consent to presumed

consent increases the donation rate in a country by roughly ró percent'

fohnson and Goldstein obtain a slightly smaller but similar effect. what-

ever the precise figure, it is clear that the switch would save thousands of

lives every year.

Determining the exact effect of changing the default rule is difficult be-

cause countries vary widely in how they implement the law. France is tech-

nically a presumed consenr counrry, but physicians routinely ask the family

members of a donor for their permission, and they usually follow the fam-

ily's wishes. This policy blurs the distinction between presumed consent

and explicit consent.

Still, the default rule does matter. In the united states, if there is no ex-

plicit donor card for survivors to see) families reject requests for donations

about half the time . The rejection rate is much lower i¡ countries with pre -

sumed consent rules, even though there is typically no record of the

donor's wishes. In Spain the rate is about 20 percent, and in France it is

about 3o percent.4 As one rePort put it: "The next of kin can be ap-

proached quite differently when the decedent's silence is presumed to in-

dicate a decision to donate rather than when it is presumed to i¡dicate a

IAO HEALTH

pose low costs, and if possible no costs, on those who go tlteir own way' Al-

though presumed consent is, in a sense, the opposite of explicit consent)

there is a key similariry: under both regimes, those who don't hold the de-

fault preference will have to register in order to opt out'

L.t's ,,rppore, for the sake of argument, that both explicit consent and

presumed consent could be implemented with "one-click" technology.

Specifically, imagine that the state could successfully contact every citizen

(andtheparentsofminors)byemail,askingt}lemtoregister.Inaworldof Econs'thetwopolicieswouldproduceidenticaloutcomes.Becausetlle costs of registering are trivial' everyone would click the preferred choice'

But even in a one-click world, the default will matter if the population is

made up ofl{umans. Of course that is how the population is composed, and thanks to

an im-

portant experiment conduced by Eric Johnson and Dan Goldstein (zoo3),

we know something about how much the choice of the default matters in

this domai¡. Using an online survey' the resea¡chers asked people' in dif-

ferent ways, whether they would be willing to be donors. In the explicit

consent condition, participants were told that they had just moved to a

new State where the default was not to be an organ donor, and they were

given the option of confuming or changing that status' In the presumed

consent version, the wording was identical but the default was to be a

donor. In the third, neutral, condition' there was no mention of a de-

fault-they just had to choose. under all three conditions, the response

was entered literally with one click.

As you will now expect' the default mattered-a lot' When parricipants

had to opt in to being an organ donor, only +z percent did so' But when

they had to opt out, 82 percent agreed to be donors. Surprisingl¡ almost

as many peo ple (ZgPercent) agreed to be donors in the neutral condition'

Although nearly all states in the united states use a version of explicit

consent) many countries in Europe have adopted presumed conserft laws

(though the cost of opting out varies, and always involves more t\an a

click). lohnson and Goldstein have analyzed the effects of such laws by

comparing countries with presumed consent to those with explicit con-

sent. The effect on consent rates is enormous. To get a sense of the power

of the default rule, consider the difference in consent rates between two

similar countries, Austria and Germany' In Germany, which uses an opt-in

2O6 FREEDOM

locals would get preferenrial treatment. You woulcl have only srightly bet- ter odds of getting inro one than of winning the powerball jackpot. (you think property in cambridge and Palo Alto is expensive nowr M/hat if liv- ing there guaranteed your child a se ar at Ha¡vard or stanfordf ) clever par- ents who do not happen to live in cambridge, but who have been dream- ing of sending their child to Harvard since the diaper days, would realize the futility oflisting it fust. The Boston sysrem attempts to march as many fust choices as possible, so if every honest parent in America risted Harvard fust, only Cambridge residents could sleep well at night.

rnstead of taking their chances orì a long shot, parents outside cam- bridge would be better served to select as their fust choice a slightly less popular school such as Dartmouth or cornell, say, where rhere are also fewer students nearby getting preferential treatment. In the Boston sys- tem, parents who rank a school second or third lose out to everyone who ranks it fust-making it risþ ro use a füst choice on a highly sought-after school if a child has a low priority, and a complete waste to list such a school as a second choice. Information about school demand is usually available online, giving parents an incentive to tweak rankings based on ac- ceptance rates and where their child has priority.

when the Boston system was first developed, almost no one intuited this strategy. (only a handful of people even knew how the algorithm worked!) But over time, some parents figured out ways to gain an edge. Not surprisingl¡ affluent, educated parents with large social networks (they volunteer at school with other affluent, educated parents) learned the tricks fust. They performed better than less afiluent, less educated par- ents, who routinely listed an overdemanded school as a second choice, the worst mistake they could make. who knows how many of thei¡ child¡en lost out on access to fust-rate educations because of it?

The Boston system is still in place around the countr¡ though not in Boston. In zoo3 a group of economists le d by Al Roth at Harvard pointed out these problems to initially skeptical Boston school administratort.'Af- ter letting the economists poke around in the internal data, the adminis- trators became convinced of their system's flaws.ó

rn response, they adopted the economists' new strategy-proof choice mechanism, based on one used to match hospitals and medical residents.

tMPROVtNG SCHOOL CHOICES 207

The mechanism does not penalize parents who a¡e unsophisticated about

the choice process, allowing them to spend time visiting schools and see-

ing teachers, rathel{a¡l.gtimating t}re level of competition to get into each school. In return, administrators do not have to guess about parents'

true preferences so that the policy can be adjusted properþ based on fu-

ture feedback.

Nudging High Schoolers Toward College

Good choice architecture doesn't need to originate with a wonk-

ish professor and a powerful computer algorithm. It can be the brainchild of a local school official or two. In San Marcos, Texas, the school superin-

tendent and an administrator at nearby Austin Community College were

Iooking for a way to get more of San Marcos's largely Latino student pop-

ulation into college. They hit on a nudge so simple a¡d effective it spread through the state faster than a YouTube clip. (Well, maybe not that fast.)

The nudge was this: in order to graduate from San Marcos High, a student

would have to complete an application to nearbyAustin Community Col-

lege. Because all it takes to get admitted to the community college is a high

school degree and a record ofhaving taken a standardizedtest, completing

the application properly was tantamount to accePtance.

In San Marcos, schools run on a tight budget, and two-thirds of high

schoolers never experience higher education. The superintendent had no

outside funding to implement the idea, so she asked her teachers and the

community college for help. Students were pulled from Engl-ish classes to

meet with the college's staff counselors. In a smart piece of mapping, the counselors didn't try to sell the students on the high-mindedness of edu-

cation. Instead, they hooked them with the universal symbol of teenage

freedom: the automobile. They talked about how much more money col-

lege graduates earned compared with high school graduates, explaining it as the difference between a Mercedes and a KIA. Next, community college

administrators took a standardized admissions exam to the high school and tested the students free of charge. The administrators also gave stu-

dents financial aid information a¡rd had ta-r consultants offer weekend ses-

sions for parents.

2OA FREEDOM

In the end, the nudge produced big results. From zoo4 to 2ooj the per- centage of San Marcos High students who went to Texas colleges rose rr percentage points, ro 4j percent. Now more than forty-five Texas high schools have similar programs) and schools in Florida and california have created programs modeled after sa¡ Marcos's. In Maine a state legislator is proposing a law requiring high school seniors to submit at least one col- Iege application before they graduate.

We have covered a lot ofterritoryin a short space . Milton Friedmanwas right: at least in the abstract, school choice is an excellent idea, because it increases freedom a¡d offers real promise for improving education. Of course) reforms should be assessed empirically, not in tÏe abstract. For- tunately, existing evidence suggests that school choice has considerable promrse .

The major problem, and our principal concern here , is that what is true for investments and prescription drugs is true for education as well: it is not enough to make lots of choices available and then hope parents choose wisely. School sysrems need to pur parents in a position to think through their choices, and to exercise their freedom rather than to rely on the de- fault option. Both parenrs and children need the right incentives. FDR s *right to a good education" is not part of the Constitution, but it has be- come a cultural commitment,.and a few simple steps could enable many more child¡en to enjoy that right.

-+

ja.Þ +¿

l4 SHOULD PATIENTS BE FORCED

TO BUY LOTTERY TIqKETS?

Every election cycle, presidential contenders unveil plans to make

health care coverage available to the tens ofmillions ofAmericans who lack health insurance. The candidates decry our government's failure, thus far,

to implement an effective plan.

Whatever happens in the long run, such plans are hard to design for a simple reason: health care is really expensive . It is expensive in part because Americans want access to all the best services: doctors, hospitals, prescrip-

tion drugs, medical devices, and nursing homes, to nâme a few. Of course , we can try to keep health care affordable on our own, by

maintaining healthy lifestyles, and by buying only the health care products

and services that we need. We can save money by visiting the doctor no

more often than necessary, and ifwe purchase insurance) we can choose a plan that covers only catastrophic illnesses instead ofcoverage witlt low de- ductibles, which is much more expensive. But there is something that every health care customer in America is forced to buy, whether she wants

it or not: the right to sue the doctor for negligence. Our principal claim here is that patients and doctors should be free to

make their own agreements about that right. Ifpatients want to waive the

right to sue, they should be allowed to do exactly that. This increase in freedom is likely to help doctors and patients alike, and to make avaluable,

even if modest, contribution to the health care problem.

It may seem strange to t¡jnk that we "purchase" the right to sue. Of course) that right is not an itemized portion of the insurance bill-but it is

209