In many countries around the world, it is common for the state to ask its citizens if they will volunteer to be organ donors. Now, organ donation is one of those issues that elicit strong feelings from many people. On the one hand, it’s an opportunity to turn one person’s loss into another person’s salvation. But on the other hand, it’s more than a little unsettling to be making plans for your organs that don’t involve you. It’s not surprising, therefore, that different people make different decisions, nor is it surprising that rates of organ donation vary considerably from country to country. It might surprise you to learn, however, how much cross-national variation there is. In a study conducted a few years ago, two psychologists, Eric Johnson and Dan Goldstein, found that rates at which citizens consented to donate their organs varied across different European countries, from as low as 4.25 percent to as high as 99.98 percent. What was even more striking about these differences is that they weren’t scattered all over the spectrum, but rather were clustered into two distinct groups—one group that had organ-donation rates in the single digits and teens, and one group that had rates in the high nineties—with almost nothing in between.1
What could explain such a huge difference? That’s the question I put to a classroom of bright Columbia undergraduates not long after the study was published. Actually, what I asked them to consider was two anonymous countries, A and B. In country A, roughly 12 percent of citizens agree to be organ donors, while in country B 99.9 percent do. So what did they think was different about these two countries that could account for the choices of their citizens? Being smart and creative students, they came up with lots of possibilities. Perhaps one country was secular while the other was highly religious. Perhaps one had more advanced medical care, and better success rates at organ transplants, than the other. Perhaps the rate of accidental death was higher in one than another, resulting in more available organs. Or perhaps one had a highly socialist culture, emphasizing the importance of community, while the other prized the rights of individuals.
All were good explanations. But then came the curveball. Country A was in fact Germany, and country B was … Austria. My poor students were stumped—what on earth could be so different about Germany and Austria? But they weren’t giving up yet. Maybe there was some difference in the legal or education systems that they didn’t know about? Or perhaps there had been some important event or media campaign in Austria that had galvanized support for organ donation. Was it something to do with World War II? Or maybe Austrians and Germans are more different than they seem. My students didn’t know what the reason for the difference was, but they were sure it was something big—you don’t see extreme differences like that by accident. Well, no—but you can get differences like that for reasons that you’d never expect. And for all their creativity, my students never pegged the real reason, which is actually absurdly simple: In Austria, the default choice is to be an organ donor, whereas in Germany the default is not to be. The difference in policies seems trivial—it’s just the difference between having to mail in a simple form and not having to—but it’s enough to push the donor rate from 12 percent to 99.9 percent. And what was true for Austria and Germany was true across all of Europe—all the countries with very high rates of organ donation had opt-out policies, while the countries with low rates were all opt-in.
Understanding the influence of default settings on the choices we make is important, because our beliefs about what people choose and why they choose it affect virtually all our explanations of social, economic, and political outcomes. Read the op-ed section of any newspaper, watch any pundit on TV, or listen to any late-night talk radio, and you will be bombarded with theories of why we choose this over that. And although we often decry these experts, the broader truth is that all of us—from politicians and bureaucrats, to newspaper columnists, to corporate executives and ordinary citizens—are equally willing to espouse our own theory of human choice. Indeed, virtually every argument of social consequence—whether about politics, economic policy, taxes, education, healthcare, free markets, global warming, energy policy, foreign policy, immigration policy, sexual behavior, the death penalty, abortion rights, or consumer demand—is either explicitly or implicitly an argument about why people make the choices they make. And, of course, how they can be encouraged, educated, legislated, or coerced into making different ones.
Given the ubiquity of choice in the world and its relevance to virtually every aspect of life—from everyday decisions to the grand events of history—it should come as little surprise that theories about how people make choices are also central to most of the social sciences. Commenting on an early paper by the Nobel laureate Gary Becker, the economist James Duesenberry famously quipped that “economics is all about choice, while sociology is about why people have no choices.”2 But the truth is that sociologists are every bit as interested in how people make choices as economists are—not to mention political scientists, anthropologists, psychologists, and legal, business, and management scholars. Nevertheless, Duesenberry had a point in that for much of the last century, social and behavioral scientists of different stripes have tended to view the matter of choice in strikingly different ways. More than anything, they have differed, sometimes acrimoniously, over the nature and importance of human rationality.
To many sociologists, the phrase “rational choice” evokes the image of a cold, calculating individual who cares only for himself and who relentlessly seeks to maximize his economic well-being. Nor is this reaction entirely unjustified. For many years, economists seeking to understand market behavior invoked something like this notion of rationality—sometimes referred to as “homo economicus”—in large part because it lends itself naturally to mathematical models that are simple enough to be written down and solved. And yet, as countless examples like the ultimatum game from the previous chapter show, real people care not only about their own welfare, economic or otherwise, but also the welfare of others for whom they will often make considerable sacrifices. We also care about upholding social norms and conventions, and frequently punish others who violate them—even when doing so is costly.3 And finally, we often care about intangible benefits, like our reputation, belonging to a group, and “doing the right thing,” sometimes as much as or even more than we care about wealth, comfort, and worldly possessions.
Critics of homo economicus have raised all these objections, and many more, over the years. In response, advocates of what is often called rational choice theory have expanded the scope of what is considered rational behavior dramatically to include not just self-interested economic behavior, but also more realistic social and political behavior as well.4 These days, in fact, rational choice theory is not so much a single theory at all as it is a family of theories that make often rather different assumptions depending on the application in question. Nevertheless, all such theories tend to include variations on two fundamental insights—first, that people have preferences for some outcomes over others; and second, that given these preferences they select among the means available to them as best they can to realize the outcomes that they prefer. To take a simple example, if my preference for ice cream exceeds my preference for the money I have in my pocket, and there is an available course of action that allows me to exchange my money for the ice cream, then that’s what I’ll choose to do. But if, for example, the weather is cold, or the ice cream is expensive, my preferred course of action may instead be to keep the money for a sunnier day. Similarly, if buying the ice cream requires a lengthy detour, my preference to get where I am going may also cause me to wait for another time. Regardless of what I end up choosing—the money, the ice cream, the walk followed by the ice cream, or some other alternative—I am always doing what is “best” for me, given the preferences I have at the time I make the decision.
What is so appealing about this way of thinking is its implication that all human behavior can be understood in terms of individuals’ attempts to satisfy their preferences. I watch TV shows because I enjoy the experience enough to devote the time to them rather than doing something else. I vote because I care about participating in politics, and when I vote, I choose the candidate I think will best serve my interests. I apply to the colleges that I think I can get into, and of those I get accepted to, I attend the one that offers the best combination of status, financial aid, and student life. When I get there, I study what is most interesting to me, and when I graduate, I take the best job I can get. I make friends with people I like, and keep those friends whose company I continue to enjoy. I get married when the benefits of stability and security outweigh the excitement of dating. We have children when the benefits of a family (the joy of having children who we can love unconditionally, as well as having someone to care for us in our old age) outweigh the costs of increased responsibility, diminished freedom, and extra mouths to feed.5
In Freakonomics, Steven Levitt and Stephen Dubner illustrate the explanatory power of rational choice theory in a series of stories about initially puzzling behavior that, upon closer examination, turns out to be perfectly rational. You might think, for example, that because your real estate agent works on commission, she will try to get you the highest price possible for your house. But as it turns out, real estate agents keep their own houses on the market longer, and sell them for higher prices, than the houses of their clients. Why? Because when it’s your house they’re selling, they make only a small percentage of the difference of the higher price, whereas when it’s their house, they get the whole difference. The latter is enough money to hold out for, but the former isn’t. Once you understand the incentives that real estate agents face, in other words, their true preferences, and hence their actions, become instantly clear.
Likewise, it might at first surprise you to learn that parents at an Israeli day school, when fined for picking up their children late, actually arrived late more often than they did before any fine was imposed. But once you understand that the fine assuaged the pangs of guilt they were feeling at inconveniencing the school staff—essentially, they felt they were paying for the right to be late—it makes perfect sense. So does the initially surprising observation that most gang members live with their mothers. Once you do the math, it turns out that gang members don’t make nearly as much money as you would think; thus it makes perfect economic sense for them to live at home. Similarly, one can explain the troubling behavior of a number of high-school teachers who, in response to the new accountability standards introduced by the Bush Administration’s 2002 No Child Left Behind legislation, actually altered the test responses of their students. Even though cheating could cost them their jobs, the risk of getting caught seemed small enough that the cost of being stuck with a low-performing class outweighed the potential for being punished for cheating.6
Regardless of the person and the context, in other words—sex, politics, religion, families, crime, cheating, trading, and even editing Wikipedia entries—the point that Levitt and Dubner keep returning to is that if we want to understand why people do what they do, we must understand the incentives that they face, and hence their preference for one outcome versus another. When someone does something that seems strange or puzzling to us, rather than writing them off as crazy or irrational, we should instead seek to analyze their situation in hopes of finding a rational incentive. It is precisely this sort of exercise, in fact, that we went through in the last chapter with the ultimatum game experiments. Once we discover that the Au and Gnau tradition of gift exchange effectively transforms what to us looks like free money into something that to them resembles an unwelcome future obligation, what was previously puzzling behavior suddenly seems as rational as our own. It is just rational according to a different set of premises than we were familiar with before. The central claim of Freakonomics is that we can almost always perform this exercise, no matter how weird or wonderful is the behavior in question.
As intriguing and occasionally controversial as Levitt and Dubner’s explanations are, in principle they are no different from the vast majority of social scientific explanations. However much sociologists and economists might argue about the details, that is, until they have succeeded in accounting for a given behavior in terms of some combination of motivations, incentives, perceptions, and opportunities—until they have, in a word, rationalized the behavior—they do not feel that they have really understood it.7 And it is not only social scientists who feel this way. When we try to understand why an ordinary Iraqi citizen would wake up one morning and decide to turn himself into a living bomb, we are implicitly rationalizing his behavior. When we attempt to explain the origins of the recent financial crisis, we are effectively searching for rational financial incentives that led bankers to create and market high-risk assets. And when we blame soaring medical costs on malpractice legislation or procedure-based payments, we are instinctively invoking a model of rational action to understand why doctors do what they do. When we think about how we think, in other words, we reflexively adopt a framework of rational behavior.8
The implicit assumption that people are rational until proven otherwise is a hopeful, even enlightened, one that in general ought to be encouraged. Nevertheless, the exercise of rationalizing behavior glosses over an important difference between what we mean when we talk about “understanding” human behavior, as opposed to the behavior of electrons, proteins, or planets. When trying to understand the behavior of electrons, for example, the physicist does not start by imagining himself in the circumstances of the electrons in question. He may have intuitions concerning theories about electrons, which in turn help him to understand their behavior. But at no point would he expect to understand what it is actually like to be an electron—indeed, the very notion of such intuition is laughable. Rationalizing human behavior, however, is precisely an exercise in simulating, in our mind’s eye, what it would be like to be the person whose behavior we are trying to understand. Only when we can imagine this simulated version of ourselves responding in the manner of the individual in question do we really feel that we have understood the behavior in question.
So effortlessly can we perform this exercise of “understanding by simulation” that it rarely occurs to us to wonder how reliable it is. And yet, as the earlier example of the organ donors illustrates, our mental simulations have a tendency to ignore certain types of factors that turn out to be important. The reason is that when we think about how we think, we instinctively emphasize consciously accessible costs and benefits such as those associated with motivations, preferences, and beliefs—the kinds of factors that predominate in social scientists’ models of rationality. Defaults, by contrast, are a part of the environment in which the decision maker operates, and so affect behavior in a way that is largely invisible to the conscious mind, and therefore largely absent from our commonsense explanations of behavior.9 And defaults are just the proverbial tip of the iceberg. For several decades, psychologists and, more recently, behavioral economists have been examining human decision-making, often in controlled laboratory settings. Their findings not only undermine even the most basic assumptions of rationality but also require a whole new way of thinking about human behavior.10
In countless experiments, for example, psychologists have shown that an individual’s choices and behavior can be influenced by “priming” them with particular words, sounds, or other stimuli. Subjects in experiments who read words like “old” and “frail” walk more slowly down the corridor when they leave the lab. Consumers in wine stores are more likely to buy German wine when German music is playing in the background, and French wine when French music is playing. Survey respondents asked about energy drinks are more likely to name Gatorade when they are given a green pen in order to fill out the survey. And shoppers looking to buy a couch online are more likely to opt for an expensive, comfortable-looking couch when the background of the website is of fluffy white clouds, and more likely to buy the harder, cheaper option when the background consists of dollar coins.11
Our responses can also be skewed by the presence of irrelevant numerical information. In one experiment, for example, participants in a wine auction were asked to write down the last two digits of their social security numbers before bidding. Although these numbers were essentially random and certainly had nothing to do with the value a buyer should place on the wine, researchers nevertheless found that the higher the numbers, the more people were willing to bid. This effect, which psychologists call anchoring, affects all sorts of estimates that we make, from estimating the number of countries in the African Union to how much money we consider to be a fair tip or donation. Whenever you receive a solicitation from a charity with a “suggested” donation amount, in fact, or a bill with precomputed tip percentages, you should suspect that your anchoring bias is being exploited—because by suggesting amounts on the high side, the requestor is anchoring your initial estimate of what is fair. Even if you subsequently adjust your estimate downward—because, say, a 25 percent tip seems like too much—you will probably end up giving more than you would have without the initial suggestion.12
Individual preferences can also be influenced dramatically simply by changing the way a situation is presented. Emphasizing one’s potential to lose money on a bet, for example, makes people more risk averse while emphasizing one’s potential to win has the opposite effect, even when the bet itself is identical. Even more puzzling, an individual’s preferences between two items can be effectively reversed by introducing a third alternative. Say, for example, that option A is a high-quality, expensive camera while B is both much lower quality and also much cheaper. In isolation, this could be a difficult comparison to make. But if, as shown in the figure below, I introduce a third option, C1, that is clearly more expensive than A and around the same quality, the choice between A and C1 becomes unambiguous. In these situations people tend to pick A, which seems perfectly reasonable until you consider what happens if I introduce instead of C1 a third option, C2, that is about as expensive as B yet significantly lower quality. Now the choice between B and C2 is clear, and so people tend to pick B. Depending on which third option is introduced, in other words, the preference of the decision maker can effectively be reversed between A and B, even though nothing about either has changed. What’s even stranger is that the third option—the one that causes the switch in preferences—is never itself chosen.13
Illustration of preference reversal
Continuing this litany of irrationality, psychologists have found that human judgments are often affected by the ease with which different kinds of information can be accessed or recalled. People generally overestimate the likelihood of dying in a terrorist attack on a plane relative to dying on a plane from any cause, even though the former is strictly less likely than the latter, simply because terrorist attacks are such vivid events. Paradoxically, people rate themselves as less assertive when they are asked to recall instances where they have acted assertively—not because the information contradicts their beliefs, but rather because of the effort required to recall it. They also systematically remember their own past behavior and beliefs to be more similar to their current behavior and beliefs than they really were. And they are more likely to believe a written statement if the font is easy to read, or if they have read it before—even if the last time they read it, it was explicitly labeled as false.14
Finally, people digest new information in ways that tend to reinforce what they already think. In part, we do this by noticing information that confirms our existing beliefs more readily than information that does not. And in part, we do it by subjecting disconfirming information to greater scrutiny and skepticism than confirming information. Together, these two closely related tendencies—known as confirmation bias and motivated reasoning respectively—greatly impede our ability to resolve disputes, from petty disagreements over domestic duties to long-running political conflicts like those in Northern Ireland or Israel-Palestine, in which the different parties look at the same set of “facts” and come away with completely different impressions of reality. Even in science, confirmation bias and motivated reasoning play pernicious roles. Scientists, that is, are supposed to follow the evidence, even if it contradicts their own preexisting beliefs; and yet, more often then they should, they question the evidence instead. The result, as the physicist Max Planck famously acknowledged, is often that “A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die.”15
Taken together, the evidence from psychological experiments makes clear that there are a great many potentially relevant factors that affect our behavior in very real and tangible ways but that operate largely outside of our conscious awareness. Unfortunately, psychologists have identified so many of these effects—priming, framing, anchoring, availability, motivated reasoning, loss aversion, and so on—that it’s hard to see how they all fit together. By design, experiments emphasize one potentially relevant factor at a time in order to isolate its effects. In real life, however, many such factors may be present to varying extents in any given situation; thus it’s critical to understand how they interact with one another. It may be true, in other words, that holding a green pen makes you think of Gatorade, or that listening to German music predisposes you to German wine, or that thinking of your social security number affects how much you will bid for something. But what will you buy, and how much will you pay for it, when you are exposed to many, possibly conflicting, subconscious influences at once?
It simply isn’t clear. Nor is the profusion of unconscious psychological biases the only problem. To return to the ice cream example from before, although it may be true that I like ice cream as a general rule, how much I like it at a particular point in time might vary considerably, depending on the time of day, the weather, how hungry I am, and how good the ice cream is that I expect to get. My decision, moreover, doesn’t depend just on how much I like ice cream, or even just the relation between how much I like it versus how much it costs. It also depends on whether or not I know the location of the nearest ice cream shop, whether or not I have been there before, how much of a rush I’m in, who I’m with and what they want, whether or not I have to go to the bank to get money, where the nearest bank is, whether or not I just saw someone else eating an ice cream, or just heard a song that reminded me of a pleasurable time when I happened to be eating an ice cream, and so on. Even in the simplest situations, the list of factors that might turn out to be relevant can get very long very quickly. And with so many factors to worry about, even very similar situations may differ in subtle ways that turn out to be important. When trying to understand—or better yet predict—individual decisions, how are we to know which of these many factors are the ones to pay attention to, and which can be safely ignored?
The ability to know what is relevant to a given situation is of course the hallmark of commonsense knowledge that I discussed in the previous chapter. And in practice, it rarely occurs to us that the ease with which we make decisions disguises any sort of complexity. As the philosopher Daniel Dennett points out, when he gets up in the middle of the night to make himself a midnight snack, all he needs to know is that there is bread, ham, mayonnaise, and beer in the fridge, and the rest of the plan pretty much works itself out. Of course he also knows that “mayonnaise doesn’t dissolve knives on contact, that a slice of bread is smaller than Mount Everest, that opening the refrigerator doesn’t cause a nuclear holocaust in the kitchen” and probably trillions of other irrelevant facts and logical relations. But somehow he is able to ignore all these things, without even being aware of what it is that he’s ignoring, and focus on the few things that matter.16
But as Dennett argues, there is a big difference between knowing what is relevant in practice and being able to explain how it is that we know it. To begin with, it seems clear that what is relevant about a situation is just those features that it shares with other comparable situations—for example, we know that how much something costs is relevant to a purchase decision because cost is something that generally matters whenever people buy something. But how do we know which situations are comparable to the one we’re in? Well, that also seems clear: Comparable situations are those that share the same features. All “purchase” decisions are comparable in the sense that they involve a decision maker contemplating a number of options, such as cost, quality, availability, and so on. But now we encounter the problem. Determining which features are relevant about a situation requires us to associate it with some set of comparable situations. Yet determining which situations are comparable depends on knowing which features are relevant.
This inherent circularity poses what philosophers and cognitive scientists call the frame problem, and they have been beating their heads against it for decades. The frame problem was first noticed in the field of artificial intelligence, when researchers started trying to program computers and robots to solve relatively simple everyday tasks like, say, cleaning a messy room. At first they assumed that it couldn’t be that hard to write down everything that was relevant to a situation like this. After all, people manage to clean their rooms every day without even really thinking about it. How hard could it be to teach a robot? Very hard indeed, as it turned out. As I discussed in the last chapter, even the relatively straightforward activity of navigating the subway system requires a surprising amount of knowledge about the world—not just about subway doors and platforms but also about maintaining personal distance, avoiding eye contact, and getting out of the way of pushy New Yorkers. Very quickly AI researchers realized that virtually every everyday task is difficult for essentially the same reason—that the list of potentially relevant facts and rules is staggeringly long. Nor does it help that most of this list can be safely ignored most of the time—because it’s generally impossible to know in advance which things can be ignored and which cannot. So in practice, the researchers found that they had to wildly overprogram their creations in order to perform even the most trivial tasks.17
The intractability of the frame problem effectively sank the original vision of AI, which was to replicate human intelligence more or less as we experience it ourselves. And yet there was a silver lining to this defeat. Because AI researchers had to program every fact, rule, and learning process into their creations from scratch, and because their creations failed to behave as expected in obvious and often catastrophic ways—like driving off a cliff or trying to walk through a wall—the frame problem was impossible to ignore. Rather than trying to crack the problem, therefore, AI researchers took a different approach entirely—one that emphasized statistical models of data rather than thought processes. This approach, which nowadays is called machine learning, was far less intuitive than the original cognitive approach, but it has proved to be much more productive, leading to all kinds of impressive breakthroughs, from the almost magical ability of search engines to complete queries as you type them to building autonomous robot cars, and even a computer that can play Jeopardy!18
The frame problem, however, isn’t just a problem for artificial intelligence—it’s a problem for human intelligence as well. As the psychologist Daniel Gilbert describes in Stumbling on Happiness, when we imagine ourselves, or someone else, confronting a particular situation, our brains do not generate a long list of questions about all the possible details that might be relevant. Rather, just as an industrious assistant might use stock footage to flesh out a drab PowerPoint presentation, our “mental simulation” of the event or the individual in question simply plumbs our extensive database of memories, images, experiences, cultural norms, and imagined outcomes, and seamlessly inserts whatever details are necessary in order to complete the picture. Survey respondents leaving restaurants, for example, readily described the outfits of the waiters inside, even in cases where the waitstaff had been entirely female. Students asked about the color of a classroom blackboard recalled it as being green—the normal color—even though the board in question was blue. In general, people systematically overestimate both the pain they will experience as a consequence of anticipated losses and the joy they will garner from anticipated gains. And when matched online with prospective dates, subjects report greater levels of liking for their matches when they are given less information about them. In all of these cases, a careful person ought to respond that he can’t answer the question accurately without being given more information. But because the “filling in” process happens instantaneously and effortlessly, we are typically unaware that it is even taking place; thus it doesn’t occur to us that anything is missing.19
The frame problem should warn us that when we do this, we are bound to make mistakes. And we do, all the time. But unlike the creations of the AI researchers, humans do not surprise us in ways that force us to rewrite our whole mental model of how we think. Rather, just as Paul Lazarsfeld’s imagined reader of the American Soldier found every result and its opposite is equally obvious, once we know the outcome we can almost always identify previously overlooked aspects of the situation that then seem relevant. Perhaps we expected to be happy after winning the lottery, and instead find ourselves depressed—obviously a bad prediction. But by the time we realize our mistake, we also have new information, say about all the relatives who suddenly appeared wanting financial support. It will then seem to us that if we had only had that information earlier, we would have anticipated our future state of happiness correctly, and maybe never bought the lottery ticket. Rather than questioning our ability to make predictions about our future happiness, therefore, we simply conclude that we missed something important—a mistake we surely won’t make again. And yet we do make the mistake again. In fact, no matter how many times we fail to predict someone’s behavior correctly, we can always explain away our mistakes in terms of things that we didn’t know at the time. In this way, we manage to sweep the frame problem under the carpet—always convincing ourselves that this time we are going to get it right, without ever learning what it is that we are doing wrong.
Nowhere is this pattern more evident, and more difficult to expunge, than in the relationship between financial rewards and incentives. It seems obvious, for example, that employee performance can be improved through the application of financial incentives, and in recent decades performance-based pay schemes have proliferated in the workplace, most notably in terms of executive compensation tied to stock price.20 Of course, it’s also obvious that workers care about more than just money—factors like intrinsic enjoyment, recognition, and a feeling of advancement in one’s career might all affect performance as well. All else equal, however, it seems obvious that one can improve performance with the proper application of financial rewards. And yet, the actual relationship between pay and performance turns out to be surprisingly complicated, as a number of studies have shown over the years.
Recently, for example, my Yahoo! colleague Winter Mason and I conducted a series of Web-based experiments in which subjects were paid at different rates to perform a variety of simple repetitive tasks, like placing a series of photographs of moving traffic into the correct temporal sequence, or uncovering words hidden in a rectangular grid of letters. All our participants were recruited from a website called Amazon’s Mechanical Turk, which Amazon launched in 2005 as a way to identify duplicate listings among its own inventory. Nowadays, Mechanical Turk is used by hundreds of businesses looking to “crowd-source” a wide range of tasks, from labeling objects in an image to characterizing the sentiment of a newspaper article or deciding which of two explanations is clearer. However, it is also an extremely effective way to recruit subjects for psychology experiments—much as psychologists have done over the years by posting flyers around college campuses—except that because workers (or “turkers”) are usually paid on the order of a few cents per task, it can be done for a fraction of the usual cost.21
In total, our experiments involved hundreds of participants who completed tens of thousands of tasks. In some cases they were paid as little as one cent per task—for example, sorting a single set of images or finding a single word—while in other cases they were paid five or even ten cents to do the same thing. A factor of ten is a pretty big difference in pay—by comparison, the average hourly rate of a computer engineer in the United States is only six times the federal minimum wage—so you’d expect it to have a pretty big effect on how people behave. And indeed it did. The more we paid people, the more tasks they completed before leaving the experiment. We also found that for any given pay rate, workers who were assigned “easy” tasks—like sorting sets of two images—completed more tasks than workers assigned medium or hard tasks (three and four images per set respectively). All of this, in other words, is consistent with common sense. But then the kicker: in spite of these differences, we found that the quality of their work—meaning the accuracy with which they sorted images—did not change with pay level at all, even though they were paid only for the tasks they completed correctly.22
What could explain this result? It’s not completely clear; however, after the subjects had finished their work we asked them some questions, including how much they thought they ought to have been paid for what they had just done. Interestingly, their responses depended less on the difficulty of the task than on how much they had been paid to do it. On average, subjects who were paid one cent per task thought they should have been paid five cents; subjects who were paid five cents thought they should have been paid eight cents; and subjects who were paid ten cents thought they should have been paid thirteen cents. In other words, no matter what they were actually paid—and remember that some of them were getting paid ten times as much as others—everyone thought they had been underpaid. What this finding suggested to us is that even for very simple tasks, the extra motivation to perform that we intuitively expect workers to experience with increased financial incentives is largely undermined by their increased sense of entitlement.
It’s hard to test this effect outside of a laboratory setting, because workers in most real environments have expectations about what they should be paid that are hard to manipulate. But consider, for example, that in the United States women get paid on average only 90 percent as much as men who do exactly the same jobs, or that European CEOs get paid considerably less than their US counterparts.23 In either case, could you really argue that the lower-paid group works less hard or does a worse job than the higher-paid group? Or imagine if next year, your boss unexpectedly doubled your annual pay—how much harder would you actually work? Or imagine a parallel universe in which bankers got paid half of what they get in ours. No doubt some of them might have chosen to go into other professions, but for those who remained in banking, would they really work less hard or do a worse job? The outcome of our experiment suggests that they would not. But if so, then you have to wonder how much influence employers can have on worker performance simply by changing financial incentives.
A number of studies, in fact, have found that financial incentives can actually undermine performance. When a task is multifaceted or hard to measure, for example, workers tend to focus only on those aspects of their jobs that are actively measured, thereby overlooking other important aspects of the job—like teachers emphasizing the material that will be covered in standardized tests at the expense of overall learning. Financial rewards can also generate a “choking” effect, when the psychological pressure of the reward cancels out the increased desire to perform. Finally, in environments where individual contributions are hard to separate from those of the team, financial rewards can encourage workers to ride on the coattails of the efforts of others, or to avoid taking risks, thereby hampering innovation. The upshot of all these confusing and often contradictory findings is that although virtually everyone agrees that people respond to financial incentives in some manner, it’s unclear how to use them in practice to elicit the desired result. Some management scholars have even concluded after decades of studies that financial incentives are largely irrelevant to performance.24
No matter how many times this lesson is pointed out, however, managers, economists, and politicians continue to act as if they can direct human behavior via the application of incentives. As Levitt and Dubner write, “The typical economist believes that the world has not yet invented a problem that he cannot fix if given a free hand to design the proper incentive scheme.… An incentive is a bullet, a lever, a key: an often tiny object with astonishing power to change a situation.”25 Well, maybe, but that doesn’t mean that the incentives we create will bring about the changes we intended. Indeed, one of Levitt and Dubner’s own vignettes—of the high-school teachers cheating on the tests—was an attempt by policy makers to improve teaching through explicit performance-based incentives. That it backfired, producing outright cheating and all manner of lesser gaming, like “teaching to the test,” and focusing exclusively on marginal students for whom a small improvement could generate an additional passing grade, should give one pause about the feasibility of designing incentive schemes to elicit desired behavior.26
And yet common sense does not pause. Rather, once we realize that some particular incentive scheme did not work, we conclude simply that it got the incentives wrong. Once they know the answer, in other words, policy makers can always persuade themselves that all they need to do is to design the correct incentive scheme—overlooking, of course, that this was precisely what they thought they were doing previously as well. Nor are policy makers uniquely susceptible to this particular oversight—we all are. For example, a recent news story about the perennial problem of politicians failing to take long-term fiscal responsibility seriously concluded blithely, “Like bankers, politicians respond to incentives.” The article’s solution? “To align the interests of the country with those of the politicians who are guiding it.” It all sounds so simple. But as the article itself concedes, the history of previous attempts to “fix” politics has been disappointing.27
Like rational choice theory, in other words, common sense insists that people have reasons for what they do—and this may be true. But it doesn’t necessarily allow us to predict in advance either what they will do or what their reasons will be for doing it.28 Once they do it, of course, the reasons will appear obvious, and we will conclude that had we only known about some particular factor that turned out to be important, we could have predicted the outcome. After the fact, it will always seem as if the right incentive system could have produced the desired result. But this appearance of after-the-fact predictability is deeply deceptive, for two reasons. First, the frame problem tells us that we can never know everything that could be relevant to a situation. And second, a huge psychological literature tells us that much of what could be relevant lies beyond the reach of our conscious minds. This is not to say that humans are completely unpredictable, either. As I’ll argue later (in Chapter 8), human behavior displays all sorts of regularities that can be predicted, often to useful ends. Nor is it to say that we shouldn’t try to identify the incentives that individuals respond to when making decisions. If nothing else, our inclination to rationalize the behavior of others probably helps us to get along with one another—a worthy goal in itself—and it may also help us to learn from our mistakes. What it does say, however, is that our impressive ability to make sense of behavior that we have observed does not imply a corresponding ability to predict it, or even that the predictions we can make reliably are best arrived at on the basis of intuition and experience alone. It is this difference between making sense of behavior and predicting it that is responsible for many of the failures of commonsense reasoning. And if this difference poses difficulties for dealing with individual behavior, the problem gets only more pronounced when dealing with the behavior of groups.