Nevertheless, this model of economic behavior based on a population consisting only of Econs has flourished, raising economics to that pinnacle of influence on which it now rests. Critiques over the years have been brushed aside with a gauntlet of poor excuses and implausible alternative explanations of embarrassing empirical evidence. But one by one these critiques have been answered by a series of studies that have progressively raised the stakes. It is easy to dismiss a story about the grading of an exam. It is harder to dismiss studies that document poor choices in large-stakes domains such as saving for retirement, choosing a mortgage, or investing in the stock market. And it is impossible to dismiss the series of booms, bubbles, and crashes we have observed in financial markets beginning on October 19, 1987, a day when stock prices fell more than 20% all around the world in the absence of any substantive bad news. This was followed by a bubble and crash in technology stocks that quickly turned into a bubble in housing prices, which in turn, when popped, caused a global financial crisis.
It is time to stop making excuses. We need an enriched approach to doing economic research, one that acknowledges the existence and relevance of Humans. The good news is that we do not need to throw away everything we know about how economies and markets work. Theories based on the assumption that everyone is an Econ should not be discarded. They remain useful as starting points for more realistic models. And in some special circumstances, such as when the problems people have to solve are easy or when the actors in the economy have the relevant highly specialized skills, then models of Econs may provide a good approximation of what happens in the real world. But as we will see, those situations are the exception rather than the rule.
Moreover, much of what economists do is to collect and analyze data about how markets work, work that is largely done with great care and statistical expertise, and importantly, most of this research does not depend on the assumption that people optimize. Two research tools that have emerged over the past twenty-five years have greatly expanded economists’ repertoire for learning about the world. The first is the use of randomized control trial experiments, long used in other scientific fields such as medicine. The typical study investigates what happens when some people receive some “treatment” of interest. The second approach is to use either naturally occurring experiments (such as when some people are enrolled in a program and others are not) or clever econometrics techniques that manage to detect the impact of treatments even though no one deliberately designed the situation for that purpose. These new tools have spawned studies on a wide variety of important questions for society. The treatments studied have included getting more education, being taught in a smaller class or by a better teacher, being given management consulting services, being given help to find a job, being sentenced to jail, moving to a lower-poverty neighborhood, receiving health insurance from Medicaid, and so forth. These studies show that one can learn a lot about the world without imposing optimizing models, and in some cases provide credible evidence against which to test such models and see if they match actual human responses.
For much of economic theory, the assumption that all the agents are optimizing is not a critical one, even if the people under study are not experts. For example, the prediction that farmers use more fertilizer if the price falls is safe enough, even if many farmers are slow to change their practices in response to market conditions. The prediction is safe because it is imprecise: all that is predicted is the direction of the effect. This is equivalent to a prediction that when apples fall off the tree, they fall down rather than up. The prediction is right as far as it goes, but it is not exactly the law of gravity.
Economists get in trouble when they make a highly specific prediction that depends explicitly on everyone being economically sophisticated. Let’s go back to the farming example. Say scientists learn that farmers would be better off using more or less fertilizer than has been the tradition. If everyone can be assumed to get things right as long as they have the proper information, then there is no appropriate policy prescription other than making this information freely available. Publish the findings, make them readily available to farmers, and let the magic of markets take care of the rest.
Unless all farmers are Econs, this is bad advice. Perhaps multinational food companies will be quick to adopt the latest research findings, but what about the behavior of peasant farmers in India or Africa?
Similarly, if you believe that everyone will save just the right amount for retirement, as any Econ would do, and you conclude from this analysis that there is no reason to try to help people save (say, by creating pension plans), then you are passing up the chance to make a lot of people better off. And, if you believe that financial bubbles are theoretically impossible, and you are a central banker, then you can make serious mistakes—as Alan Greenspan, to his credit, has admitted happened to him.
We don’t have to stop inventing abstract models that describe the behavior of imaginary Econs. We do, however, have to stop assuming that those models are accurate descriptions of behavior, and stop basing policy decisions on such flawed analyses. And we have to start paying attention to those supposedly irrelevant factors, what I will call SIFs for short.
It is difficult to change people’s minds about what they eat for breakfast, let alone problems that they have worked on all their lives. For years, many economists strongly resisted the call to base their models on more accurate characterizations of human behavior. But thanks to an influx of creative young economists who have been willing take some risks and break with the traditional ways of doing economics, the dream of an enriched version of economic theory is being realized. The field has become known as “behavioral economics.” It is not a different discipline: it is still economics, but it is economics done with strong injections of good psychology and other social sciences.
The primary reason for adding Humans to economic theories is to improve the accuracy of the predictions made with those theories. But there is another benefit that comes with including real people in the mix. Behavioral economics is more interesting and more fun than regular economics. It is the un-dismal science.
Behavioral economics is now a growing branch of economics, and its practitioners can be found in most of the best universities around the world. And recently, behavioral economists and behavioral scientists more generally are becoming a small part of the policy-making establishment. In 2010 the government of the United Kingdom formed a Behavioural Insights Team, and now other countries around the world are joining the movement to create special teams with the mandate to incorporate the findings of other social sciences into the formulation of public policy. Businesses are catching on as well, realizing that a deeper understanding of human behavior is every bit as important to running a successful business as is an understanding of financial statements and operations management. After all, Humans run companies, and their employees and customers are also Humans.
This book is the story of how this happened, at least as I have seen it. Although I did not do all the research—as you know, I am too lazy for that—I was around at the beginning and have been part of the movement that created this field. Following Amos’s dictum, there will be many stories to come, but my main goals are tell the tale of how it all happened, and to explain some of the things we learned along the way. Not surprisingly, there have been numerous squabbles with traditionalists who defended the usual way of doing economics. Those squabbles were not always fun at the time, but like a bad travel experience, they make for good stories after the fact, and the necessity of fighting those battles has made the field stronger.
Like any story, this one does not follow a straight-line progression with one idea leading naturally to another. Many ideas were percolating at different times and at different speeds. As a result, the organizational structure of the book is both chronological and topical. Here is a brief preview. We start at the beginning, back when I was a graduate student and was collecting a list of examples of odd behaviors that did n
ot seem to fit the models I was learning in class. The first section of the book is devoted to those early years in the wilderness, and describes some of the challenges that were thrown down by the many who questioned the value of this enterprise. We then turn to a series of topics that occupied most of my attention for the first fifteen years of my research career: mental accounting, self-control, fairness, and finance. My objective is to explain what my colleagues and I learned along the way, so that you can use those insights yourself to improve your understanding of your fellow Humans. But there may also be useful lessons about how to try to change the way people think about things, especially when they have a lot invested in maintaining the status quo. Later, we turn to more recent research endeavors, from the behavior of New York City taxi drivers, to the drafting of players into the National Football League, to the behavior of participants on high-stakes game shows. At the end we arrive in London, at Number 10 Downing Street, where a new set of exciting challenges and opportunities is emerging.
My only advice for reading the book is stop reading when it is no longer fun. To do otherwise, well, that would be just misbehaving.
________________
* One economist who did warn us about the alarming rate of increase in housing prices was my fellow behavioral economist Robert Shiller.
2
The Endowment Effect
I began to have deviant thoughts about economic theory while I was a graduate student in the economics department at the University of Rochester, located in upstate New York. Although I had misgivings about some of the material presented in my classes, I was never quite sure whether the problem was in the theory or in my flawed understanding of the subject matter. I was hardly a star student. In that New York Times Magazine article by Roger Lowenstein that I mentioned in the preface, my thesis advisor, Sherwin Rosen, gave the following as an assessment of my career as a graduate student: “We did not expect much of him.”
My thesis was on a provocative-sounding topic, “The Value of a Life,” but the approach was completely standard. Conceptually, the proper way to think about this question was captured by economist Thomas Schelling in his wonderful essay “The Life You Save May Be Your Own.” Many times over the years my interests would intersect with Schelling’s, an early supporter and contributor to what we now call behavioral economics. Here is a famous passage from his essay:
Let a six-year-old girl with brown hair need thousands of dollars for an operation that will prolong her life until Christmas, and the post office will be swamped with nickels and dimes to save her. But let it be reported that without sales tax the hospital facilities of Massachusetts will deteriorate and cause a barely perceptible increase in preventable deaths—not many will drop a tear or reach for their checkbooks.
Schelling writes the way he speaks: with a wry smile and an impish twinkle in his eye. He wants to make you a bit uncomfortable.* Here, the story of the sick girl is a vivid way of capturing the major contribution of the article. The hospitals stand in for the concept Schelling calls a “statistical life,” as opposed to the girl, who represents an “identified life.” We occasionally run into examples of identified lives at risk in the real world, such as the thrilling rescue of trapped miners. As Schelling notes, we rarely allow any identified life to be extinguished solely for the lack of money. But of course thousands of “unidentified” people die every day for lack of simple things like mosquito nets, vaccines, or clean water.
Unlike the sick girl, the typical domestic public policy decision is abstract. It lacks emotional impact. Suppose we are building a new highway, and safety engineers tell us that making the median divider three feet wider will cost $42 million and prevent 1.4 fatal accidents per year for thirty years. Should we do it? Of course, we do not know the identity of those victims. They are “merely” statistical lives. But to decide how wide to make that median strip we need a value to assign to those lives prolonged, or, more vividly, “saved” by the expenditure. And in a world of Econs, society would not pay more to save one identified life than twenty statistical lives.
As Schelling noted, the right question asks how much the users of that highway (and perhaps their friends and family members) would be willing to pay to make each trip they take a tiny bit safer. Schelling had specified the correct question, but no one had yet come up with a way to answer it. To crack the problem you needed some situation in which people make choices that involve a trade-off between money and risk of death. From there you can infer their willingness to pay for safety. But where to observe such choices?
Economist Richard Zeckhauser, a student of Schelling’s, noted that Russian roulette offers a way to think about the problem. Here is an adaptation of his example. Suppose Aidan is required to play one game of machine-gun Russian roulette using a gun with many chambers, say 1,000, of which four have been picked at random to have bullets. Aidan has to pull the trigger once. (Mercifully, the gun is set on single shot.) How much would Aidan be willing to pay to remove one bullet?† Although Zeckhauser’s Russian roulette formulation poses the problem in an elegant way, it does not help us come up with any numbers. Running experiments in which subjects point loaded guns at their heads is not a practical method for obtaining data.
While pondering these issues I had an idea. Suppose I could get data on the death rates of various occupations, including dangerous ones like mining, logging, and skyscraper window-washing, and safer ones like farming, shopkeeping, and low-rise window-washing. In a world of Econs, the riskier jobs would have to pay more, otherwise no one would do them. In fact, the extra wages paid for a risky job would have to compensate the workers for taking on the risks involved (as well as any other attributes of the job). So if I could also get data on the wages for each occupation, I could estimate the number implied by Schelling’s analysis, without asking anyone to play Russian roulette. I searched but could not find any source of occupational mortality rates.
My father, Alan, came to the rescue. Alan was an actuary, one of those mathematical types who figure how to manage risks for insurance companies. I asked him if he might be able to lay his hands on data on occupational mortality. I soon received a thin, red, hardbound copy of a book published by the Society of Actuaries that listed the very data I needed. By matching occupational mortality rates to readily available data on wages by occupation, I could estimate how much people had to be paid to be willing to accept a higher risk of dying on the job.
Getting the idea and the data were a good start, but doing the statistical exercise correctly was key. I needed to find an advisor in the economics department whom I could interest in supervising my thesis. The obvious choice was the up-and-coming labor economist mentioned earlier, Sherwin Rosen. We had not worked together before, but my thesis topic was related to some theoretical work he was doing, so he agreed to become my advisor.
We went on to coauthor a paper based on my thesis entitled, naturally, “The Value of Saving a Life.” Updated versions of the number we estimated back then are still used in government cost-benefit analyses. The current estimate is roughly $7 million per life saved.
While at work on my thesis, I thought it might be interesting to ask people some hypothetical questions as another way to elicit their preferences regarding trade-offs between money and the risk of dying. To write these questions, I first had to decide which of two ways to ask the question: either in terms of “willingness to pay” or “willingness to accept.” The first asks how much you would pay to reduce your probability of dying next year by some amount, say by one chance in a thousand. The second asks how much cash you would demand to increase the risk of dying by the same amount. To put these numbers in some context, a fifty-year-old resident of the United States faces a roughly 4-in-1,000 risk of dying each year.
Here is a typical question I posed in a classroom setting. Students answered both versions of the question.
A. Suppose by attending this lecture you have exposed yourself to a rare fatal disease. If you contract the disease
you will die a quick and painless death sometime next week. The chance you will get the disease is 1 in 1,000. We have a single dose of an antidote for this disease that we will sell to the highest bidder. If you take this antidote the risk of dying from the disease goes to zero. What is the most you would be willing to pay for this antidote? (If you are short on cash we will lend you the money to pay for the antidote at a zero rate of interest with thirty years to pay it back.)
B. Researchers at the university hospital are doing some research on that same rare disease. They need volunteers who would be willing to simply walk into a room for five minutes and expose themselves to the same 1 in 1,000 risk of getting the disease and dying a quick and painless death in the next week. No antidote will be available. What is the least amount of money you would demand to participate in this research study?
Economic theory has a strong prediction about how people should answer the two different versions of these questions. The answers should be nearly equal. For a fifty-year-old answering the questions, the trade-off between money and risk of death should not be very different when moving from a risk of 5 in 1,000 (.005) to .004 (as in the first version of the question) than in moving from a risk of .004 to .005 (as in the second version). Answers varied widely among respondents, but one clear pattern emerged: the answers to the two questions were not even close to being the same. Typical answers ran along these lines: I would not pay more than $2,000 in version A but would not accept less than $500,000 in version B. In fact, in version B many respondents claimed that they would not participate in the study at any price.
Misbehaving: The Making of Behavioral Economics Page 2
