In day care centers, parents sometimes arrive late to pick their children up. Economists have studied this problem, and proposed that day care centers fine late parents (since it costs extra money to take care of children while they’re waiting for their parents to arrive). Strangely, however, when day care centers adopted such a policy, late arrivals went up, not down. How?
Like the Introduction, this chapter begins with a puzzle in need of a solution, creating a sense of suspense (and, as the authors suggested in the Introduction, making economics more exciting!). The authors will then use the theory of incentives to “solve” the puzzle.
In order to understand the day care problem, the authors say, we’ll need to think carefully about incentives. Economics is largely the study of how incentives drive human behavior. In simplest terms, an incentive is “a means of urging people to do more of a good thing and less of a bad thing”. There are two kinds of incentives: positive and negative (“carrots” and “sticks”). Some incentives are biological; for instance, we instinctively pull our hands away from a hot flame (a negative incentive). But most incentives have to be created artificially; this means that incentives are always changing. For example, by fining a big company for polluting the environment, the government could incentivize the company to decrease its pollutions (another negative incentive).
The first distinction that the authors make is a distinction between positive and negative incentives—in other words, “carrots” and “sticks” (after the proverbial horse chasing a carrot and running away from the stick). Such a distinction intuitively makes sense—we all understand the negative incentive that makes us pull our hands away from a hot flame. Put another way, the concept of incentives suggests that every behavior must have a cause: in all walks of life, people do things because they’re either trying to gain a positive incentive or avoid a negative incentive.
Another way to classify incentives is to label them as economic, social, or moral incentives. A government plan to fine smokers would be an economic incentive to reduce smoking. Now the authors apply the three forms of incentives to crime. Why, we might ask, isn’t there more crime than there already is? At some point, everyone has an opportunity to steal, cheat, or otherwise break the law. The reason more people don’t commit crimes is partly economic—people are frightened of going to jail and losing their jobs and incomes. The incentive is also moral—people think crime is wrong. There’s also a strong social incentive: people don’t want to be caught committing a crime and humiliated in front of their friends or peers.
The second distinction that the authors make is a distinction between economic, moral, and social incentives. As the crime example would suggest, however, it’s often difficult to disentangle the incentives that motivate an action. People refrain from committing crimes for a variety of economic, social, and moral reasons. A further implication of the crime example is that the three forms of incentive can be equally influential under different circumstances—in other words, the authors aren’t suggesting that humans always put economics above morality.
In terms of incentives, the problem with the day care center’s system of fining adults was that the fine the day care center proposed, three dollars, wasn’t big enough. If the fine had been one hundred dollars, it probably would have convinced some late parents to arrive on time. But there’s another interesting problem with the fine: by fining late parents, the day care center replaced a moral incentive with an economic incentive. In other words, parents who would ordinarily feel the moral guilt of being late to pick their children up could rationalize their lateness by paying a small fine to the day care center, thus freeing themselves from their guilt for a small monetary fee.
This is a particularly subtle example of how incentives can conflict with one another. The daycare fine inspired parents (who had previously conceived of their tardiness in moral terms) to conceive of their tardiness in strictly economic terms—a change that, counterintuitively, resulted in more tardiness. The parents who left their children late could seemingly think of their behavior in moral or economic terms, but less frequently in terms of both.
Another example of the clash between moral and economic incentives came in the 1970s. Doctors discovered that when people are paid for donating blood, less blood is donated overall. The problem with the blood donation incentive program was that it paid a small amount of money (less than fifty dollars) for an action that most people take for moral reasons. In the process, the blood donor center reduced the moral benefit of donating blood, resulting in fewer donations.
In this case, blood donors started out by acting for moral reasons, but eventually acted for economic reasons. Blood donors seemingly found it difficult to conceive of their donations as both an economic and a moral behavior; the fifty-dollar bonus tarnished the blood donation process with self-interest.
The practice of cheating is a good way to understand incentives. Almost everyone has cheated at some point in life: children cheat on tests, and CEOs cheat on their taxes. In 1987, the Internal Revenue Service required taxpayers who listed a dependent child to provide a social security number for each child. Seven million supposed “child dependents” disappeared from tax forms, suggesting that millions of people had cheated on their taxes, falsely claiming they had children.
The passage studies cheating in economic terms, rather than moral terms. While it may be true that cheating is morally wrong, the morality of cheating is largely irrelevant to the authors’ analysis. Their primary purpose is to describe data, not to offer recommendations for how people should behave. Thus, one could certainly say that the 7 million people who falsified their tax returns were morally wrong, but in economic terms they were just responding to strong economic incentives.
In the Chicago Public School system, the biggest cheaters of all might be teachers, not students. Federally mandated tests measure students’ success for each school year; if students don’t succeed on their tests, the students’ teachers may be punished—passed over for raises and promotions. Therefore, the introduction of federal student testing creates a new economic incentive for teacher cheating.
The authors don’t spend a lot of time discussing specific teachers who cheated; instead, they focus on what the group of Chicago schoolteachers did. In part, this is because the authors are taking an impartial, economic view of cheating, not a moral view. While a teacher who falsifies test scores might not be a very good teacher, he or she is simply responding to an economic incentive.
The Chicago Public School system released some of the data for its students’ test scores. This allows economists to study how pervasive cheating on federal tests might be. One common way for a teacher to cheat on student tests would be for the teacher to add correct answers to the end of a student’s test (i.e., the part of the test where incorrect answers are most common). Statistical analyses of Chicago classes’ scores on federal tests indicate that an unexpectedly high number of students in certain classes “choose” the same correct answers for the final ten questions on federal tests—the questions that should be the hardest. It is highly unlikely that students would choose the same correct answers to hard questions, but not the same correct answers to easy questions (or the same wrong answers to easy questions). Based on this principle, researchers estimate evidence of cheating in about five percent of all classrooms in Chicago. The changing economic incentives of cheating drove more teachers to cheat.
The Chicago Public School system study is interesting for a number of reasons. First, the study uses statistics and relative probability to identify teachers who were likely to have cheated. In other words, it is difficult for the study to prove to a certainty that certain teachers cheated; the results of the study can only suggest the likelihood of cheating on certain tests.
There’s been some controversy over the prevalence of cheating in Chicago schools. In 2002, the CEO of the Chicago Public School system, Arne Duncan, decided to reduce teacher cheating, reasoning that doing so would help the underprivileged students of Chicago, who needed to learn. Duncan identified 120 classrooms, some of which had been identified as having teachers who may have cheated. This time, teachers weren’t allowed to be in the room with their students when they were tested, or handle their students’ tests. When the results came in, students did considerably worse on their tests than they’d done originally: without teachers to help them cheat, the students didn’t succeed. Duncan publicized news of the cheating study, hoping that the news would act as a warning to teachers next year. Sure enough, cheating fell 30 percent the next year.
Duncan was able to 1) identify that cheating was, indeed, occurring in Chicago classrooms, and 2) use his influence, and the influence of his study, to reduce cheating the next year. Duncan’s actions foreshadow the ideas of the second chapter: Duncan used information and publicity to intimidate or shame Chicago teachers into changing their behavior. The next year, Chicago teachers had the same positive economic incentives for cheating, but they also had to consider the negative economic incentives of being caught.
Sports and cheating “go hand in hand.” Athletes have a huge economic, social, and even moral incentive to win. In Japan, sumo wrestling is a highly prestigious sport: sumo wrestlers are big celebrities, with the most famous earning millions of dollars. There is a complicated system for ranking sumo wrestlers, and that ranking system largely determines the wrestler’s success. If a wrestler wins more than half of his matches (i.e., 8 out of 15) at one of the prestigious sumo tournaments, then his ranking rises; if not, it goes down. For this reason, a wrestler’s eighth match is especially important in determining his rank. In terms of incentives, a wrestler with a 7-7 record has much more to gain from a victory than does an opponent with an 8-6 record. So it’s possible that in tournaments, wrestlers with 8-6 records will allow opponents with 7-7 records to win.
For the next example the authors discuss, the economic incentives are plain: successful sumo wrestlers make a lot of money, and they enjoy a tremendous amount of social prestige in Japan. Therefore, it would seem that the social and economic incentives for cheating in sumo wrestling outweigh the negative moral incentives of doing so. Sumo wrestling is a particularly good example of the power of economic studies, since, on the surface, it seems almost impossible to measure whether sumo wrestlers cheat: sumo wrestling is such an unpredictable sport that it would be difficult to separate legitimate matches from rigged matches.
But how can we measure cheating in sumo wrestling? First, the authors focus on bouts between 7-7 wrestlers and 8-6 wrestlers. One reason to do so is that it’s the simplest way to isolate the wrestlers’ incentives. A wrestler with a 14-0 record will have his own conflicting reasons for taking a bribe and intentionally losing his 15th match (on one hand, he wants the first-place prize money; on the other hand, he might get a bribe for losing). An 8-6 wrestler in the same position, however, would not have these strong confounding motives for turning down a bribe.
In order to analyze sumo wrestling, the authors begin by isolating some variables. A 7-7 wrestler will have a very strong positive incentive for winning a match, while an 8-6 wrestler will have a smaller incentive. Thus an 8-6 wrestler has a strong incentive for accepting a bribe and very little incentive for turning it down; he probably won’t win the tournament either way.
Based on all past data, a 7-7 wrestler should beat an 8-6 wrestler about 48 percent of the time. In actuality, however, 7-7- wrestlers defeat 8-6 wrestlers about 80 percent of the time. There is such a large difference between the real and expected outcomes that it stands to reason that many 8-6 wrestlers take bribes or otherwise plan to lose their matches. Another good way to estimate bribery in sumo matches is to look at what happens the next time the same 7-7 and 8-6 wrestler compete; i.e., when neither wrestler is in a “bubble” match. Statistics show that 7-7 wrestlers defeat 8-6 wrestlers for a second time only 40 percent of the time. This is probably because the two wrestlers make a deal: the 7-7 wrestler wins the first time, and the 8-6 wrestler wins the second time. A final way to measure cheating in sumo is to look at 7-7 vs. 8-6 matches shortly after there have been major allegations of cheating. In this situation, 7-7 wrestlers win their matches against 8-6 wrestlers about 50 percent of the time. Thus it is highly likely, based on the data, that sumo wrestling is a corrupt sport in which wrestlers regularly take bribes to throw matches.
The authors’ analysis persuasively suggests the prevalence of bribery in the sumo wrestling world. On the surface, one would expect an 8-6 wrestler to beat a 7-7 wrestler slightly more than half of the time; in reality, the 7-7 wrestler wins considerably more often. Notice that the authors do not (and, in fact, cannot) pinpoint which wrestlers do and don’t take bribes. It would be very difficult to analyze specific sumo matches and decide which ones are legitimate and which ones are rigged. By studying the sport of sumo wrestling as a whole, however, economists can estimate that a significant number of wrestlers cheat, without having a very good idea of which wrestlers they are.
Another illuminating example of corruption lies with a man named Paul Feldman. Feldman was a government researcher during the 1960s, but among his colleagues, he was famous for being “the guy who brings in the bagels.” Feldman always made a point of bringing bagels to work. Years later, Feldman decided to quit his job and “bring bagels” full-time. Feldman would travel to hundreds of companies and bring fresh bagels. Instead of charging the companies upfront for the bagels, Feldman used an honor-system collection format—he would come back to companies in the afternoon to see if anybody had eaten a bagel and left some money. Amazingly, Feldman made a healthy living bringing bagels to workers.
So far, one could argue, the examples from this chapter have painted a pretty cynical picture of human nature. While the authors don’t focus their attention on the moral implications of the data, it would seem that a significant number of people are willing to break the law or bend the rules in order to protect their own interests. For the final case study in this chapter, however, we’ll see that humans can also be surprisingly honest and trustworthy.
As a “bagel guy,” Feldman would personally go to different companies. Sometimes, he would find that people hadn’t obeyed the honors system, and had eaten bagels without paying for them. Sometimes, Feldman would leave collection boxes at his various companies, and come back to collect the boxes later. Although company employees would occasionally eat bagels without paying for them, very few people would steal the collection boxes themselves.
Feldman’s career suggests some interesting things about human behavior. People will occasionally “cheat” by eating bagels without paying, and yet they will almost never steal entire boxes of money, despite the fact that their economic incentive for stealing boxes is much greater than the economic incentive for stealing one bagel.
Feldman’s example tells us a lot about what’s usually called “white-collar crime.” At large companies (like the ones to which Feldman delivered bagels), there are certain people who embezzle company money—that is, people who cheat and steal. White-collar crime is relatively rarely prosecuted, and often unsolved (whereas murders and burglaries are solved and prosecuted in the majority of cases). So perhaps “bagel theft” could be used as a measure of white-collar crime in a business setting.
One of the premises of Freakonomics is that we can understand a lot about society and humanity by studying seemingly trivial things like sumo wrestlers, bagel thieves, etc. While stealing a couple of bagels might seem unimportant, it’s a useful benchmark of the overall amount of crime and cheating in a group.
In recent years, there have two major trends in the bagel payment rate for Feldman’s company. First, bagel payment rates slowly declined after 1992. Second, payment rates increased noticeably after September 11, 2001, perhaps reflecting a “patriotic surge” in the businesses Feldman served. Feldman has also observed that smaller businesses tend to be more honest than large ones. Furthermore, the bagel data suggests that one’s personal mood correlates with one’s likelihood to commit a crime. Pleasant weather often correlates with a higher payment rate. The Christmas holiday correlates with a lower payment rate, while other holidays, like the 4th of July, correlate with a higher pay rate.
The same rule is seemingly true of cheating teachers in Chicago and bagel thieves in Washington, D.C.: the actions of the minority who break the rules are not as random as they seem. One might think that the prevalence of bagel theft is basically unpredictable, but in fact, bagel theft is subject to a wide array of environmental influences, including holidays and the weather. Even if individual human behavior is unpredictable, economics can analyze the behavior of a group.
In short, Feldman’s bagel data reflects “the intersection of morality and economics.” The vast majority of people Feldman serves do not steal bagels—a conclusion that perhaps reflects the writings of Adam Smith, the famous 18th century economist. In his book, The Theory of Moral Sentiment, Smith posits that humans are innately honest; by default, they care about helping other people and making others happy. Of course, many thinkers and economists take exactly the opposite point of view. The ancient Greek philosopher Plato repeated a fable about the “ring of Gyges,” in which a humble shepherd discovers a ring of invisibility. Without any consequences for his actions, the shepherd wore his ring and used it to kill, rape, etc. Plato and Smith exemplify two competing views of human nature. Feldman’s data suggests that, the vast majority of the time, human beings will be honest, perhaps suggesting that Smith was right about human nature.
Although the authors have been focusing on crime and cheating in various sectors of life, they end the chapter by looking at the “big picture.” While it’s surprising that five percent of teachers cheated in Chicago schools, perhaps the more surprising fact is that 95 percent of teachers did not—the moral and social incentives of obeying the rules kept them honest. While the authors don’t attempt a philosophical analysis of good and evil, they do suggest that humans have an innate sense of good that leads them to obey the rules, even when they have no practical reasons for doing so. Sometimes, goodness (and the moral incentive that accompanies it) is its own reward.