Read Predictably Irrational Online
Authors: Dr. Dan Ariely
Predictably Irrational (28 page)
The novel aspect of this experiment had to do with the third group. Before these participants began, each was asked to sign the following statement on the answer sheet: “I understand that this study falls under the MIT honor system.” After signing this statement, they continued with the task. When the time had elapsed they pocketed their answer sheets, walked to the front of the room, told the experimenter how many problems they had correctly solved, and were paid accordingly.
What were the results? In the control condition, in which cheating was not possible, participants solved on average three problems (out of 20). In the second condition, in which the participants could pocket their answers, they claimed to have solved on average 5.5 problems. What was remarkable was the third situationâin which the participants pocketed their answer sheets, but had also signed the honor code statement. In this case they claimed to have solved, on average, three problemsâexactly the same number as the control group. This outcome was similar to the results we achieved with the Ten Commandmentsâwhen a moral reminder eliminated cheating altogether. The effect of signing a statement about an honor code is particularly amazing when we take into account that MIT doesn't even have an honor code.
So we learned that people cheat when they have a chance to do so, but they don't cheat as much as they could. Moreover, once they begin thinking about honestyâwhether by recalling the Ten Commandments or by signing a simple statementâthey stop cheating completely. In other words, when we are removed from any benchmarks of ethical thought, we tend to stray into dishonesty. But if we are reminded of morality at the moment we are tempted, then we are much more likely to be honest.
At present, several state bars and professional organizations are scrambling to shore up their professional ethics. Some are increasing courses in college and graduate schools, and others are requiring brush-up ethics classes. In the legal profession, Judge Dennis M. Sweeney of the Howard County (Maryland) circuit published his own book,
Guidelines for Lawyer Courtroom Conduct
, in which he noted, “Most rules, like these, are simply what our mothers would say a polite and well raised man or woman should do. Since, given their other important responsibilities, our mothers (and yours) cannot be in every courtroom in the State, I offer these rules.”
Will such general measures work? Let's remember that lawyers do take an oath when they are admitted to the bar, as doctors take an oath when they enter their profession. But occasional swearing of oaths and occasional statements of adherence to rules are not enough. From our experiments, it is clear that oaths and rules must be recalled at, or just before, the moment of temptation. Also, what is more, time is working against us as we try to curb this problem. I said in Chapter 4 that when social norms collide with market norms, the social norms go away and the market norms stay. Even if the analogy is not exact, honesty offers a related lesson: once professional ethics (the social norms) have declined, getting them back won't be easy.
T
HIS DOESN'T MEAN
that we shouldn't try. Why is honesty so important? For one thing, let's not forget that the United States holds a position of economic power in the world today partly because it is (or at least is perceived to be) one of the world's most honest nations, in terms of its standards of corporate governance.
In 2002, the United States ranked twentieth in the world in terms of integrity, according to one survey (Denmark, Finland, and New Zealand were first; Haiti, Iraq, Myanmar, and Somalia were last, at number 163). On this basis, I would suspect that people doing business with the United States generally feel they can get a fair deal. But the fact of the matter is that the United States ranked fourteenth in 2000, before the wave of corporate scandals made the business pages in American newspapers look like a police blotter.
29
We are going down the slippery slope, in other words, not up it, and this can have tremendous long-term costs.
Adam Smith reminded us that honesty really is the best policy, especially in business. To get a glimpse at the other side of that realizationâat the downside, in a society without trustâyou can take a look at several countries. In China, the word of one person in one region rarely carries to another region. Latin America is full of family-run cartels that hand out loans to relatives (and then fail to cut off credit when the debtor begins to default). Iran is another example of a nation stricken by distrust. An Iranian student at MIT told me that business there lacks a platform of trust. Because of this, no one pays in advance, no one offers credit, and no one is willing to take risks. People must hire within their families, where some level of trust still exists. Would you like to live in such a world? Be careful, because without honesty we might get there faster than you'd imagine.
What can we do to keep our country honest? We can read the Bible, the Koran, or whatever reflects our values, perhaps. We can revive professional standards. We can sign our names to promises that we will act with integrity. Another path is to first recognize that when we get into situations where our personal financial benefit stands in opposition to our moral standards, we are able to “bend” reality, see the world in terms compatible with our selfish interest, and become dishonest. What is the answer, then? If we recognize this weakness, we can try to avoid such situations from the outset. We can prohibit physicians from ordering tests that would benefit them financially; we can prohibit accountants and auditors from functioning as consultants to the same companies; we can bar members of Congress from setting their own salaries, and so on.
But this is not the end of the issue of dishonesty. In the next chapter, I will offer some other suggestions about dishonesty, and some other insights into how we struggle with it.
Reflections on Dishonesty and Toilet Paper
About a year after
Predictably Irrational
was first published, I received an interesting e-mail from a woman who used elements from our honor code experiment to design her own test, which effectively solved an important problem involving stolen toilet paper.
Dear Professor Ariely,
I found your experiments on cheating particularly interesting (e.g., the nonexistent “MIT honor code” and “The 10 Commandments”) and actually modeled an experiment of my own on yours.
I live in a house near the UC Berkeley campus. I share the house with many housemates; none of us knew each other before moving in. Moreover, a bunch of new students moved into the house for the summer months last year. This living arrangement led to a number of problemsânamely, the stealing of toilet paper.
Every Sunday, a housekeeper restocks each toilet with three fresh rolls of toilet paper. Last summer, however, I noticed that by Monday evening, all the toilet paper would be gone. This is not unlike a “tragedy of the commons”
*
situation, and it has become clear that certain housemates were hoarding the toilet paper for their private use. I tried to think of ways to stop the hoarding. I did not want to confront anyone, because my goal was to keep the TP from disappearing, not to alienate people.
With your experiment in mind, I put a note in the upstairs bathroom asking housemates not to take toilet paper outside the restroom for personal use.
A couple of hours later, one fresh roll of toilet paper magically reappeared in the bathroom where I had placed the note. This had never happened before. The experiment worked! I checked the downstairs restroom, where I had not left a noteâno rolls to be found. Two days later, another fresh roll of toilet paper reappeared in the upstairs bathroom.
I was quite pleased that my note helped put an end to the hoarding/stealing.
Thank you for the inspiration! Behavioral economics is certainly applicable to daily life!
Regards,
Rhonda
I must confess that when I thought about the implications of the honor-reminder experiments for day-to-day life, issues around toilet paper didn't come to mind. But I am happy that Rhonda made the connection and applied it to her own situation. I am also quite pleased that she placed the note in one bathroom but not the otherânice experimental control!
Appendix: Chapter 13
The Ten Commandments
*
Catholic - I, the Lord, am your God. You shall not have other gods besides me.
Jewish - I am the Lord your God who has taken you out of the land of Egypt.
Protestant - You shall have no other gods but me.
Catholic - You shall not take the name of the Lord, your God, in vain.
Jewish - You shall have no other gods but me.
Protestant - You shall not make unto you any graven images.
Catholic - Remember to keep holy the Lord's Day.
Jewish - You shall not take the name of the Lord your God in vain.
Protestant - You shall not take the name of the Lord your God in vain.
Catholic - Honor your father and your mother.
Jewish - You shall remember the Sabbath and keep it Holy.
Protestant - You shall remember the Sabbath and keep it Holy.
Catholic - You shall not kill.
Jewish - Honor your mother and father.
Protestant - Honor your mother and father.
Catholic - You shall not commit adultery.
Jewish - You shall not murder.
Protestant - You shall not murder.
Catholic - You shall not steal.
Jewish - You shall not commit adultery.
Protestant - You shall not commit adultery.
Catholic - You shall not bear false witness.
Jewish - You shall not steal.
Protestant - You shall not steal.
Catholic - You shall not covet your neighbor's wife.
Jewish - You shall not bear false witness.
Protestant - You shall not bear false witness.
Catholic - You shall not covet your neighbor's goods.
Jewish - You shall not covet anything that belongs to your neighbor.
Protestant - You shall not covet anything that belongs to your neighbor.
M
any of the dormitories at MIT have common areas, where sit a variety of refrigerators that can be used by the students in the nearby rooms. One morning at about eleven, when most of the students were in class, I slipped into the dorms and, floor by floor, went hunting for all the shared refrigerators that I could find.
When I detected a communal fridge, I inched toward it. Glancing cautiously around, I opened the door, slipped in a six-pack of Coke, and walked briskly away. At a safe distance, I paused and jotted down the time and the location of the fridge where I had left my Cokes.
Over the next few days I returned to check on my Coke cans. I kept a diary detailing how many of them remained in the fridge. As you might expect, the half-life of Coke in a college dorm isn't very long. All of them had vanished within 72 hours. But I didn't always leave Cokes behind. In some of the fridges, I left a plate containing six one-dollar bills. Would the money disappear faster than the Cokes?
Before I answer that question, let me ask you one. Suppose your spouse calls you at work. Your daughter needs a red pencil for school the next day. “Could you bring one home?” How comfortable would you be taking a red pencil from work for your daughter? Very uncomfortable? Somewhat uncomfortable? Completely comfortable?
Let me ask you another question. Suppose there are no red pencils at work, but you can buy one downstairs for a dime. And the petty cash box in your office has been left open, and no one is around. Would you take 10 cents from the petty cash box to buy the red pencil? Suppose you didn't have any change and needed the 10 cents. Would you feel comfortable taking it? Would that be OK?
I don't know about you, but while I'd find taking a red pencil from work relatively easy, I'd have a very hard time taking the cash. (Luckily for me, I haven't had to face this issue, since my daughter is not in school yet.)
As it turns out, the students at MIT also felt differently about taking cash. As I mentioned, the cans of Coke quickly disappeared; within 72 hours every one of them was gone. But what a different story with the money! The plates of dollar bills remained untouched for 72 hours, until I removed them from the refrigerators.
So what's going on here?
When we look at the world around us, much of the dishonesty we see involves cheating that is one step removed from cash. Companies cheat with their accounting practices; executives cheat by using backdated stock options; lobbyists cheat by underwriting parties for politicians; drug companies cheat by sending doctors and their wives off on posh vacations. To be sure, these people don't cheat with cold cash (except occasionally). And that's my point: cheating is a lot easier when it's a step removed from money.
Do you think that the architects of Enron's collapseâKenneth Lay, Jeffrey Skilling, and Andrew Fastowâwould have stolen money from the purses of old women? Certainly, they took millions of dollars in pension monies from a lot of old women. But do you think they would have hit a woman with a blackjack and pulled the cash from her fingers? You may disagree, but my inclination is to say no.
So what permits us to cheat when cheating involves nonmonetary objects, and what restrains us when we are dealing with money? How does that irrational impulse work?
B
ECAUSE WE ARE
so adept at rationalizing our petty dishonesty, it's often hard to get a clear picture of how nonmonetary objects influence our cheating. In taking a pencil, for example, we might reason that office supplies are part of our overall compensation, or that lifting a pencil or two is what everyone does. We might say that taking a can of Coke from a communal refrigerator from time to time is all right, because, after all, we've all had cans of Coke taken from us. Maybe Lay, Skilling, and Fastow thought that cooking the books at Enron was OK, since it was a temporary measure that could be corrected when business improved. Who knows?
To get at the true nature of dishonesty, then, we needed to develop a clever experiment, one in which the object in question would allow few excuses. Nina, On, and I thought about it. Suppose we used symbolic currency, such as tokens. They were not cash, but neither were they objects with a history, like a Coke or a pencil. Would it give us insight into the cheating process? We weren't sure, but it seemed reasonable; and so, a few years ago, we gave it a try.
This is what happened. As the students at one of the MIT cafeterias finished their lunches, we interrupted them to ask whether they would like to participate in a five-minute experiment. All they had to do, we explained, was solve 20 simple math problems (finding two numbers that added up to 10). And for this they would get 50 cents per correct answer.
The experiment began similarly in each case, but ended in one of three different ways. When the participants in the first group finished their tests, they took their worksheets up to the experimenter, who tallied their correct answers and paid them 50 cents for each. The participants in the second group were told to tear up their worksheets, stuff the scraps into their pockets or backpacks, and simply tell the experimenter their score in exchange for payment. So far this experiment was similar to the tests of honesty described in the previous chapter.
But the participants in the last group had something significantly different in their instructions. We told them, as we had told the previous group, to tear up the worksheets and simply tell the experimenter how many questions they had answered correctly. But this time, the experimenter wouldn't be giving them cash. Rather, she would give them a token for each question they claimed to have solved. The students would then walk 12 feet across the room to another experimenter, who would exchange each token for 50 cents.
Do you see what we were doing? Would the insertion of a token into the transactionâa piece of valueless, nonmonetary currencyâaffect the students' honesty? Would the token make the students less honest in tallying their answers than the students who received cash immediately? If so, by how much?
Even we were surprised by the results: The participants in the first group (who had no way to cheat) solved an average of 3.5 questions correctly (they were our control group).
The participants in the second group, who tore up their worksheets, claimed to have correctly solved an average of 6.2 questions. Since we can assume that these students did not become smarter merely by tearing up their worksheets, we can attribute the 2.7 additional questions they claimed to have solved to cheating.
But in terms of brazen dishonesty, the participants in the third group took the cake. They were no smarter than the previous two groups, but they claimed to have solved an average of 9.4 problemsâ5.9 more than the control group and 3.2 more than the group that merely ripped up the worksheets.
This means that when given a chance to cheat under ordinary circumstances, the students cheated, on average, by 2.7 questions. But when they were given the same chance to cheat with nonmonetary currency, their cheating increased to 5.9âmore than doubling in magnitude. What a difference there is in cheating for money versus cheating for something that is a step away from cash!
If that surprises you, consider this. Of the 2,000 participants in our studies of honesty (described in the previous chapter), only four ever claimed to have solved all the problems. In other words, the rate of “total cheating” was four in 2,000.
*
But in the experiment in which we inserted nonmonetary currency (the token), 24 of the study's 450 participants cheated “all the way.” How many of these 24 extreme cheaters were in the condition with money versus the condition with tokens? They were all in the token condition (24 of 150 students cheated “all the way” in this condition; this is equivalent to about 320 per 2,000 participants). This means that not only did the tokens “release” people from some of their moral constraints, but for quite a few of them, the extent of the release was so complete that they cheated as much as was possible.
This level of cheating is clearly bad, but it could have been worse. Let's not forget that the tokens in our experiments were transformed into cash within a matter of seconds. What would the rate of dishonesty have been if the transfer from a nonmonetary token to cash took a few days, weeks, or months (as, for instance, in a stock option)? Would even more people cheat, and to a larger extent?
W
E HAVE LEARNED
that given a chance, people cheat. But what's really odd is that most of us don't see this coming. When we asked students in another experiment to predict if people would cheat more for tokens than for cash, the students said no, the amount of cheating would be the same. After all, they explained, the tokens represented real moneyâand the tokens were exchanged within seconds for actual cash. And so, they predicted, our participants would treat the tokens as real cash.
But how wrong they were! They didn't see how fast we can rationalize our dishonesty when it is one step away from cash. Of course, their blindness is ours as well. Perhaps it's why so much cheating goes on. Perhaps it's why Jeff Skilling, Bernie Ebbers, and the entire roster of executives who have been prosecuted in recent years let themselves, and their companies, slide down the slope.
All of us are vulnerable to this weakness, of course. Think about all the insurance fraud that goes on. It is estimated that when consumers report losses on their homes and cars, they creatively stretch their claims by about 10 percent. (Of course, as soon as you report an exaggerated loss, the insurance company raises its rates, so the situation becomes tit for tat). Again it is not the case that there are many claims that are completely flagrant, but instead many people who have lost, say, a 27-inch television set report the loss of a 32-inch set; those who have lost a 32-inch set report the loss of a 36-inch set, and so on. These same people would be unlikely to steal money directly from the insurance companies (as tempting as that might sometimes be), but reporting what they no longer haveâand increasing its size and value by just a little bitâmakes the moral burden easier to bear.
There are other interesting practices. Have you ever heard the term “wardrobing”? Wardrobing is buying an item of clothing, wearing it for a while, and then returning it in such a state that the store has to accept it but can no longer resell it. By engaging in wardrobing, consumers are not directly stealing money from the company; instead, it is a dance of buying and returning, with many unclear transactions involved. But there is at least one clear consequenceâthe clothing industry estimates that its annual losses from wardrobing are about $16 billion (about the same amount as the estimated annual loss from home burglaries and automobile theft combined).
And how about expense reports? When people are on business trips, they are expected to know what the rules are, but expense reports too are one step, and sometimes even a few steps, removed from cash. In one study, Nina and I found that not all expenses are alike in terms of people's ability to justify them as business expenses. For example, buying a mug for five dollars for an attractive stranger was clearly out of bounds, but buying the same stranger an eight-dollar drink in a bar was very easy to justify. The difference was not the cost of the item, or the fear of getting caught, but people's ability to justify the item to themselves as a legitimate use of their expense account.
A few more investigations into expense accounts turned up similar rationalizations. In one study, we found that when people give receipts to their administrative assistants to submit, they are then one additional step removed from the dishonest act, and hence more likely to slip in questionable receipts. In another study, we found that businesspeople who live in New York are more likely to consider a gift for their kid as a business expense if they purchased it at the San Francisco airport (or someplace else far from home) than if they had purchased it at the New York airport, or on their way home from the airport. None of this makes logical sense, but when the medium of exchange is nonmonetary, our ability to rationalize increases by leaps and bounds.
I
HAD MY
own experience with dishonesty a few years ago. Someone broke into my Skype account (very cool online telephone software) and charged my PayPal account (an online payment system) a few hundred dollars for the service.
I don't think the person who did this was a hardened criminal. From a criminal's perspective, breaking into my account would most likely be a waste of time and talent because if this person was sufficiently smart to hack into Skype, he could probably have hacked into Amazon, Dell, or maybe even a credit card account, and gotten much more value for his time. Rather, I imagine that this person was a smart kid who had managed to hack into my account and who took advantage of this “free” communication by calling anyone who would talk to him until I managed to regain control of my account. He may have even seen this as a techie challengeâor maybe he is a student to whom I once gave a bad grade and who decided to tweak my nose for it.
Would this kid have taken cash from my wallet, even if he knew for sure that no one would ever catch him? Maybe, but I imagine that the answer is no. Instead, I suspect that there were some aspects of Skype and of how my account was set up that “helped” this person engage in this activity and not feel morally reprehensible: First, he stole calling time, not money. Next, he did not gain anything tangible from the transaction. Third, he stole from Skype rather than directly from me. Fourth, he might have imagined that at the end of the day Skype, not I, would cover the cost. Fifth, the cost of the calls was charged automatically to me via PayPal. So here we had another step in the processâand another level of fuzziness in terms of who would eventually pay for the calls. (Just in case you are wondering, I have since canceled this direct link to PayPal.)
Was this person stealing from me? Sure, but there were so many things that made the theft fuzzy that I really don't think he thought of himself as a dishonest guy. No cash was taken, right? And was anyone really hurt? This kind of thinking is worrisome. If my problem with Skype was indeed due to the nonmonetary nature of the transactions on Skype, this would mean that there is much more at risk here, including a wide range of online services, and perhaps even credit and debit cards. All these electronic transactions, with no physical exchange of money from hand to hand, might make it easier for people to be dishonestâwithout ever questioning or fully acknowledging the immorality of their actions.
