Misbehaving: The Making of Behavioral Economics

by Richard H. Thaler

  Rather than doing the rational self-interested thing of agreeing to take any positive offer (the smallest offer allowed in our version was 50 cents), when the Responders were playing with what they considered to be their own money (rather than “house money”) they became even more concerned with being treated fairly. In the experiments Kahneman, Knetsch, and I had run years before, the average minimum request by Responders was $1.94. In these new experiments, that average jumped to $3.21 for a group of MIT MBA students, $3.73 for Chicago MBA students, and $3.35 for Chicago law students. And in all three groups, many of the Responders demanded their full $5 back. Making the experiment more “real” had made the Responders less consistent with self-interested income maximization! As we had hoped, the audience greeted this result with some consternation.

  This experiment was germane to the behavioral analysis of the Coase theorem. Willingness to walk away from an “unfair” offer is another reason why the predictions of the Coase theorem often fail. I had discovered this firsthand many years earlier in Rochester. Our home there had a willow tree in the backyard that shed its leaves late in the fall and continued to do so even after snow arrived. This made raking the leaves especially arduous. The tree was located very close to the border of my neighbor’s land, and the neighbor hated that tree. He asked me to have the tree removed.

  I was ambivalent about the tree. It was nice to look at and provided some shade, factors that roughly offset the cleanup issue. Still, in the interest of neighborhood harmony I inquired about the cost of removing the tree and found out the price was $1,000, roughly one month’s income for me at the time. I was not willing to pay that much money to get rid of the tree. But I knew the Coase theorem. In fact, I was teaching a course in which it played a central role. So I went to talk to my neighbor and told him that, while the tree did not bother me, if he felt strongly about it, I would let him arrange to remove it at his own expense. He thought this was the most outrageous suggestion he ever heard, slammed the door in my face, and never broached the subject again.

  When people are given what they consider to be unfair offers, they can get angry enough to punish the other party, even at some cost to themselves. That is the basic lesson of the Ultimatum Game. As the willow tree story illustrates, the same can occur in situations in which the Coase theorem is often applied. After a lawsuit, both sides are typically upset with each other, and this is particularly true for the person who loses the case. For the Coase theorem to work, that losing party has to be willing to make an offer to the other side if he puts a greater value on the property right he just lost. But if people are angry, the last thing they want to do is talk to the other side. Law professor Ward Farnsworth documented this reluctance by interviewing attorneys from over twenty civil cases in which injunctive relief was sought and either granted or denied after full litigation before a judge. In not a single case did the parties even attempt to negotiate after the court had issued its order.

  In addition to the Coase theorem, the other part of the paper that got people’s blood boiling was something we left for the very end of it—the topic of paternalism. The core principle underlying the Chicago School’s libertarian beliefs is consumer sovereignty: the notion that people make good choices, and certainly better choices than anyone else could make for them. By raising the specters of bounded rationality and bounded self-control, we were undercutting this principle. If people make mistakes, then it becomes conceivable, at least in principle, that someone could help them make a better choice.

  We knew this was treacherous, inflammatory territory for the Chicago law and economics crowd, so we approached the topic in the mildest possible manner using a term Cass had coined: “anti-antipaternalism.” The double negative implied that we were not ready to put forward a positive argument for paternalism. Instead, we noted that the knee-jerk claim that it is impossible to help anyone make a better decision is clearly undercut by the research. The short, two-page section on this topic was followed by a longer section on “behavioral bureaucrats.” It was, for Cass and me, the first of many times that we went out of our way to say that if the government bureaucrat is the person trying to help, it must be recognized that the bureaucrat is also a Human, subject to biases. Frustratingly, no matter how many times we repeat this refrain, we continue to be accused of ignoring it.

  After the workshop, we retreated to the faculty club. Christine had a glass of wine, I had a double scotch, and Cass had three Diet Cokes—his strongest and favorite elixir. We had not converted any of the key participants, but we had survived. Better still, we had confirmed that that our paper was going to cause a stir.

  Postscript: It is not possible to say what impact our paper had. We do know that it has been frequently cited, but cannot determine whether we successfully nudged anyone to take up the cause of behavioral law and economics. What I can say is that today there is a lot of behavioral law and economics research being done, enough to fill an 800-page Oxford Handbook of Behavioral Economics and the Law, edited by Eyal Zamir and Doron Teichman. One of the prominent contributors to this field, UCLA law professor Russell Korobkin, is ready to declare victory: “The battle to separate the economic analysis of legal rules and institutions from the straitjacket of strict rational choice assumptions has been won.” Ever fearful of overconfidence, I am not ready to declare “mission accomplished,” but for sure we can safely declare “mission launched.”

  ________________

  * France has now switched from marketing to photography. In my highly biased opinion, her images are worth a look. See for yourself at francleclerc.com.

  † Later we would call this a nudge.

  ‡ Another important proviso of the Coase theorem, along with no transaction costs, is that the stakes be “small” relative to the wealth of the disputing parties. I will ignore it for the purposes of this discussion.

  § The latest findings by Stanford Law professor John Donoghue and his colleagues suggest that if anything, the passage of so called “right to carry” laws increases crime rates (Aneja, Donohue III, and Zhang, 2014).

  ¶ The experiment makes money because the students had provided the stakes and many offers were rejected, meaning that both players got nothing. We always figured out a way of returning this money to the students, often by playing the beauty contest game discussed earlier and giving the leftover money to the winner.

  28

  The Offices

  Normally the Booth School of Business at the University of Chicago is a research hotbed. You can almost feel those scientific frontiers creak as they are pushed ever outward. Except, that is, for a few months in the spring of 2002. During that period, research, at least among the tenured faculty members of the school, took a pause. Offices needed to be picked.

  The task at hand was seemingly simple. After years in charming but cramped and rustic quarters on the university’s main quadrangle, the business school was building a new home two blocks away. Designed by the world-famous architect Rafael Viñoly, it was to be a stunning, modern edifice with a spectacular atrium. The site was across the street from the famous Robie House, the first home built by Frank Lloyd Wright, and Viñoly had paid subtle homage to Wright in designing the corner of the building that faces Wright’s iconic house. The palatial building was full of light, and virtually everyone was looking forward to the move. All that was left to do was to decide who would get which office. What could go wrong?

  There are many possible ways to assign offices, but the deans settled on an unusual process. There would be an office draft. Faculty would receive a time slot to pick, and then choose any open office, with full knowledge of all the selections made up to that point. This all seems fine, but there remained the important question of how the order would be determined. Seniority seems like one obvious choice, but there is a famous saying around Chicago that you are only as good as your last paper. Strike seniority as a possibility. A lottery was also not seriously considered; office locations were too important to leave entirely
to chance.

  The deans decided that the selection order would be based on “merit,” and the judge of that merit would be Deputy Dean for Faculty John Huizinga. He already had the duty of negotiating with new faculty members over the terms of their contracts, as well as that of dealing with any current faculty members who were unhappy with their teaching assignments, pay, colleagues, students, research budget, or anything else. In spite of several years on the job, John was greatly admired by the faculty, who considered him an honest, if at times blunt, straight shooter.*

  The other deans had the sense to make it clear that this job was going to be handled solely by John, to whom all complaints should be taken. After considerable deliberation, he announced how the picking order (and pecking order) would be determined. First, there would be a certain number of categories (bins, they were called, a term from statistics). John would decide how many bins there would be, and which faculty members would be assigned to each bin, but the order within each bin would be determined by random drawing. The number of bins was not announced, and has still not been revealed. As we will see, this created some ambiguity about the process.

  On the day of the draft, faculty members would have fifteen minutes each to select their office. They would do so with the aid of one of the architects working on the project. The building was a steel cage at this point, so it would not be possible to go see the offices, but architectural drawings and a scale model of the building were made available. Two other rules of interest: offices could not be traded and, after one senior faculty member inquired, the deans emphatically ruled out the possibility of buying an earlier draft pick from a colleague. This ruling, and the fact that the school decided not to simply auction off the draft picks, reveals that even at the University of Chicago Booth School of Business—where many favor an open market in babies and organs—some objects are simply too sacred to sell in the marketplace: faculty offices.

  It appeared most of the faculty had expected a process vaguely like this, and nearly all the senior faculty were content in the knowledge that they would be chosen to make one of the early picks. A few weeks of calm ensued.

  In time, all the faculty members received an email announcing that the draft would occur in a few weeks, and that our time to pick was, say, from 10:15 to 10:30 a.m. on a Wednesday. The email gave no hint about the pecking order. We were in the dark . . . for about thirty minutes. Anil Kashyap, a hyperenergetic senior faculty member in the finance and economics groups, took it upon himself to make the draft order known to everyone. An email went out asking people to reply with their time slot. Within hours, the basic outline of the draft order became clear.

  Seniority had not been ignored altogether. All the tenured full professors would choose before the (untenured in our system) associate professors, who would pick before the assistant professors, who would pick before the adjuncts, and so forth. The order of the picks within the groups of untenured faculty members seemed clearly random, and at that point the junior faculty went back to work, trying to get tenure and have a chance to do work in one of those senior faculty offices someday. Meanwhile, all hell broke loose among the senior faculty members.

  John has never revealed to me (or anyone else, as far as I know) exactly how the draft order for the senior faculty was determined. What follows is my best guess.† I believe that there were three full-professor bins. The first bin (bin A) had about a dozen people who were considered stars and/or were the obvious senior figures in their respective groups. There was at least one faculty member from each faculty group, such as accounting, economics, and so forth, but there were several people from finance, which is by far the largest department. So far, so good. No one would have complained if Gene Fama had been given the first choice. He was the most distinguished faculty member in the draft pool.

  Bin B contained most of the rest of the tenured faculty, and bin C consisted of faculty members who were no longer doing active research. In a classy move, John had slotted himself as the last tenured faculty member to choose. I believe that John selected people to be in the first bin with several purposes in mind. One was to reward those who had made significant contributions to the school. Another was to scatter the star faculty members around the building; the most attractive offices were those in corners and as such they were far apart, since the five-story building takes up an entire city block with faculty offices spread over the top three floors.

  The most distressed people were the ones in bin B who thought they deserved to have been in bin A, and then got unlucky in the lottery within their bin. There were several people in this category, but the angriest of them all was “Archie.”‡ Someone else from his group, “Clyde,” had been included in bin A, and he had lucked into the second pick. Meanwhile, Archie was picking near the end of the second group, after two of his much younger colleagues.

  To call Archie furious at this turn of events is a serious understatement. He was hopping mad, or jumping mad if there is such a thing. He was corybantic, if that means anything to you. As far as Archie was concerned, the entire draft had been rigged and the considerable evidence to the contrary would not sway him. The first pick had gone to Doug Diamond, one of the most respected and likable members of the faculty but not a household name outside of academia. Fama was third. I remember thinking at the time that the only person who was truly happy with their spot in the draft was Doug. But no one was as unhappy as Archie.

  About a day after the draft order was pieced together, Anil Kashyap got back to work and decided that it was essential to test how this draft would play out. Someone with a high pick might be interested to see the “neighborhoods” that could develop out of the later picks. We conducted a “mock” draft via email. A spreadsheet was passed around by email from Doug, to Clyde, to Gene, and so forth, on which everyone would indicate their choice of office.

  Someone circulated floor plans but the faculty demanded more information, specifically the size of each office and whether the office had a thermostat. There were thermostats in about one in three offices and, at least in theory, the occupant of the office could control the temperature with the thermostat. I suggested to John that they install “placebo” thermostats in the rest of the offices to make everyone happy, and based on my experience with the thermostat in the office I chose, the placebos would have been equally effective in controlling the temperature. The mock draft took days to complete, leading to loud complaints of “Where the hell is X, doesn’t he read his email?” Everyone was captivated by the exercise, so we ran it again to see if things would change. This was important!

  Finally, the day of the draft arrived, and we began making picks at 8:30 in the morning. The only early hiccup was when someone picked an office that someone below him had claimed in a mock draft, producing a “That was my office, you bastard!” It seems that the endowment effect can occur even for an office that was selected in what had been clearly labeled a practice exercise. Then, something strange happened. Picking at 1:15, the finance professor Luigi Zingales had his eye on a fifth-floor office near where his corporate finance colleagues were congregating. Luigi is suspicious by nature—he attributes it to his Italian upbringing—and he questioned the estimated square footage of the office he had selected.

  The architect tried to put him off, but Luigi persisted. She hauled out the real floor plans only to discover that he was right. The office he had selected was 20 square feet smaller than indicated. (The offices are all large, mostly between 180 and 230 square feet.) Luigi quickly switched his pick to a larger one nearby, and went back to his office to share his discovery. Naturally, he had not mentioned his suspicion to anyone before making his choice, lest he lose his competitive advantage. Word travelled quickly. People who had picked earlier were descending on the office that was being used for the selection process, demanding that their office be remeasured. Other mistakes in the office size estimates were found and people wanted to switch. Mayhem! John, who was out of town at a conference, was finally reached, and
sometime around 3 p.m. the draft was suspended for remeasurement.

  It took a few days for the new measurements to be announced, and this time the unhappy people included some with early picks. A few of their offices had “shrunk,” and they wanted to switch to offices that others, lower in the draft, had taken. John now weighed in via email. The draft would start over the following week. People were free to switch their picks; however, they could not choose any office that someone had already taken, even if that person were drafting later. More uproar. Around this time, John wandered into the faculty lounge during lunchtime wearing a pair of plastic Groucho Marx glasses, as if he were there incognito. It brought the house down, but the ranks of the pissed off did not laugh quite as loudly.

  Postmortem

  A year or so later, we moved into the new building and for the most part all was well. In hindsight, the most remarkable thing about the entire fiasco is that, except for the nine corner offices, the rest of the offices are pretty much the same. They are all nice, much nicer than what we had in the old building. Sure, some are a bit bigger than others, and some have slightly nicer views, but many of the differences that are now apparent were not fully appreciated at the time of the draft. For example, the offices on the fifth floor were grabbed early, perhaps on the basis of a flawed “higher is better” heuristic, but there is no view advantage to the fifth floor versus the fourth, and it has the disadvantage of being served by only one of the three elevator banks in the building, and the busiest one at that. The offices along the north face of the building have the nicest views, including the Chicago skyline, but were not among the first offices picked.