As for whether rationality alone is “sufficient”—meaning that by itself, it alone can deliver important predictions—Arrow argued convincingly that rationality alone does not get you very much. To derive useful results, theorists have to add auxiliary assumptions, such as assuming that everyone has the same utility function, meaning the same tastes. This assumption is not only demonstrably false, but it immediately leads to all kinds of predictions that are inconsistent with the facts. We are not Econs and we are certainly not identical Econs.
Arrow also noted an inconsistency inherent in the behavior of an economic theorist who toils for months to derive the optimal solution to some complex economic problem, and then blithely assumes that the agents in his model behave as if they are capable of solving the same problem. “We have the curious situation that scientific analysis imputes scientific behavior to its subjects.” At the end of the talk, Arrow declared his allegiance: “Obviously I am accepting the insight of Herbert Simon on the importance of recognizing that rationality is bounded.”
But my role in this conference was not just listening to academics I admired; I was given the intimidating task of acting as the discussant for a set of three papers authored respectively by Herbert Simon, Danny Kahneman with Amos Tversky, and Hillel Einhorn with Robin Hogarth (the conference organizer). In this situation, I largely agreed with what the authors had said, so I was not sure what to do. Discussants are expected to critique and elaborate. For me to just say, “Yeah, what he said,” would not serve me well. The papers that I thought had real conceptual problems were slated for sessions yet to come. I also had to keep in mind that I was at the “kids’ table”; there were two Nobel laureates on the program (Arrow and Simon), several others in the audience, and half a dozen more that were to win prizes later. How could I make my points to such a big-league crowd without seeming presumptuous?
I ended up deciding that my best strategy was to employ some humor. This can be risky, but I have found that if people are laughing, they tend to be more forgiving. I based my discussion on an obscure essay by George Stigler, one of the wittiest economists of his generation, and as a Chicago faculty member, he was sitting in the rationalists’ cheering section of the audience. Stigler’s essay was called the “The Conference Handbook,” and it, in turn, was based on an ancient joke:
A new prisoner arrives at a jail where everyone else has been locked up for a long time. He notices that occasionally someone shouts out a number, and everyone else laughs. He asks his cellmate what is going on and is told that they have been in jail so long together that they have all heard all the jokes that anyone knows, so to save time they have numbered the jokes. After hearing a few more numbers followed by howls of laughter, he decides to try it himself and shouts out “Thirty-nine!” No one laughs. He asks his cellmate why no one laughed and was told, “Well, some people just can’t tell a joke.”
Stigler’s essay proposed to apply the joke numbering system at conferences and departmental seminars where the same tiresome comments are repeated again and again. Stigler offered several introductory remarks, indicated by letters, followed by thirty-two specific comments that he suggested could be referenced by number. I quoted his introductory comment F, figuring we might hear a version of it soon: “It is good to have a non-specialist looking at our problem. There is always a chance of a fresh viewpoint, although usually, as in this case, the advantages of the division of labor are reaffirmed.”
In this spirit, I offered what I called the “Psychology and Economics Conference Handbook.” My idea was to list the tiresome comments I had been hearing anytime I gave a talk, those described in chapter 6 on the Gauntlet, along with suggested retorts. I figured that announcing them in advance might preempt some of the participants from hauling them out later. You can guess by now some of the comments: 1. If the stakes are high enough, people will get it right. 2. In the real world, people will learn and avoid these mistakes. 3. In aggregate, the errors will cancel . . . And so forth. For each one, I explained why the comment was not as devastating as the person delivering it might have thought.
I then concluded:
I will end my remarks with the following two false statements.
1. Rational models are useless.
2. All behavior is rational.
I have offered these false statements because both sides in the debate that will be taking place at this conference and at similar conferences in the future have a tendency to misstate the other side’s views. If everyone would agree that these statements are false, then no one would have to waste any time repudiating them.
People seemed to like the discussion. I even got a thumbs-up from Stigler as I was leaving the podium. The rest of the first day of the conference was reasonably calm.
The morning of the second day began with the announcement that Franco Modigliani had won the Nobel Prize in economics, in part for work that he had done jointly with Merton Miller, one of the primary speakers scheduled for the second day. Modigliani was then at MIT, but he had earlier been a colleague of Herb Simon’s at Carnegie Mellon, and at Simon’s urging the conference sent Modigliani a congratulatory telegram. That morning, Miller could not be blamed if he was thinking that this good news for his mentor and collaborator was bad news for him. Modigliani won the prize alone, and Miller might have felt that he had missed his chance. It turned out that he would win a Nobel Prize five years later, but he had no way of knowing that at the time. Nor did he know that morning, in this pre-Internet era, that the prize had been awarded primarily for Modigliani’s work on saving and consumption—the life-cycle hypothesis—rather than for his work with Miller on corporate finance.
In the morning festivities surrounding the news, Miller spoke briefly about Modigliani’s research. The press had asked him to summarize the work he had done with Modigliani, and, with his usual sharp wit, he said they had shown that if you take a ten-dollar bill from one pocket and put it into a different pocket, your wealth does not change. This line got a big laugh, to which Miller replied: “Don’t laugh. We proved it rigorously!”
The joke was meant to refer to their so-called “irrelevance theorem,” which proved that, under certain assumptions, it would not matter whether a firm chose to pay a dividend or instead use that money to repurchase their own shares or reduce their debts. The idea is that investors should care neither where money is stashed nor how it is paid out. But the joke actually applied equally well to the life-cycle hypothesis, since in that theory the only determinant of a household’s consumption is its wealth, not the manner in which that wealth is held, say in cash, retirement savings, or home equity. Both theories take as a working hypothesis that money is fungible. We have already seen that in the case of the life-cycle hypothesis, this assumption is wrong. It turns out, all jokes aside, the assumption was equally questionable in corporate finance, which was the topic of Miller’s talk that afternoon.
Miller’s paper had been provoked by a behavioral finance paper by Hersh Shefrin, my self-control collaborator, and Meir Statman, a colleague of Shefrin’s at Santa Clara University. In particular, they were offering a behavioral explanation for an embarrassing fact. One of the key assumptions in the Miller–Modigliani irrelevance theorem was the absence of taxes. Paying dividends would no longer be irrelevant if dividends were taxed differently than the other ways firms return money to their shareholders. And given the tax code in the United States at that time, firms should not have been paying dividends. The embarrassing fact was that most large firms did pay dividends.
The way taxes come into play is that income, including dividend income, was then taxed at rates as high as 50% or more, whereas capital gains were taxed at a rate of 25%. Furthermore, this latter tax was only paid when the capital gain was realized, that is, when the stock was sold. The effect of these tax rules was that shareholders would much rather get capital gains than dividends, at least if the shareholders were Econs. Importantly, a firm could easily transform a dividend into a capital gain by us
ing the funds that would go to paying dividends to repurchase shares in the firm. Instead of receiving a dividend, shareholders would see the price of their shares go up, and would save money on their tax bill. So the puzzle was: why did firms punish their tax-paying shareholders by paying dividends? (Those who pay no taxes, such as endowments or those saving in a tax-free account, would be indifferent between the two policies.)
Shefrin and Statman’s answer relied on a combination of self-control and mental accounting. The notion was that some shareholders—retirees, for instance—like the idea of getting inflows that are mentally categorized as “income” so that they don’t feel bad spending that money to live on. In a rational world, this makes no sense. A retired Econ could buy shares in companies that do not pay dividends, sell off a portion of his stock holdings periodically, and live off of those proceeds while paying less in taxes. But there is a long-standing notion that it is prudent to spend the income and leave the principal alone, and this idea was particularly prevalent in the generation of retirees around in 1985, all of whom had lived through the Great Depression.*
It is fair to say that Merton Miller was not a fan of the Shefrin and Statman paper. In his talk, he did not disguise this disdain, saying that the behavioral approach might have applied to his own Aunt Minnie and a few others like her, but that that was as far as it went.
The written version of Miller’s paper was less strident than his presentation, but was nevertheless quite odd. Most of the paper was devoted to a lucid tutorial on the very puzzle that Shefrin and Statman were trying to explain, rather than a critique of their hypothesis. In fact, I know of no clearer explanation for why, in a land of Econs, firms would not pay dividends under the tax regime then in place. Miller agreed that firms should not pay dividends, but most did so. He also agreed that the model that best described how firms decided how much to pay out in dividends was the one proposed by the financial economist John Lintner, a model Miller labeled “behavioral.” In Lintner’s model, firms only increase dividends when they are confident that earnings have gone up enough such that dividends will not have to be cut in the future. (Had the model been written later, Lintner might have used loss aversion to help explain why firms are so reluctant to cut dividends.) Lintner had arrived at this model after using the unfashionable strategy of interviewing the chief financial officers of many large companies. About this model Miller said: “I assume it to be a behavioral model, not only from its form, but because no one has yet been able to derive it as the solution to a maximization problem, despite thirty years of trying!”
So let’s summarize Miller’s paper. Theory tells us that firms should not pay dividends and yet they do. And a behavioral model admittedly best describes the pattern by which they pay them. This sounds like a paper written by someone who has come to praise behavioral finance, not bury it. But Miller was neither ready to praise nor to concede. He wrote: “The purpose of this paper has been to show that the rationality-based market equilibrium models in finance in general and of dividends in particular are alive and well—or at least in no worse shape than other comparable models in economics at their level of aggregation.” So, the strongest statement Miller could muster was to say that the standard rational model of financial markets—the efficient market hypothesis, to which we will turn in the next section, on finance—was not quite dead.
Not only did Miller concede that the best model of how firms pay dividends is behavioral, but he was also happy to grant the same about how individual investors behave. He said: “Behind each holding may be a story of family business, family quarrels, legacies received, divorce settlements, and a host of other considerations almost totally irrelevant to our theories of portfolio selection. That we abstract from all these stories in building our models is not because the stories are uninteresting, but because they may be too interesting and thereby distract us from the pervasive market forces that should be our principal concern.” Take a moment to absorb that: we should ignore the reasons why people do things, not because they are uninteresting, but because they are too interesting. I, for one, had trouble keeping track of which side of the case Miller was arguing.
Miller’s talk came in the afternoon session of the last day, chaired by Eugene Fama, another Chicago faculty member and a strong defender of the rational point of view. The other speaker during that session was Allan Kleidon, who like Miller was not so much presenting new research of his own, but rather attacking a paper by Robert Shiller that we will discuss in detail in chapter 24. Shiller was given the role of discussant, along with two efficient market defenders, Richard Roll and Steve Ross. Shefrin and Statman could only heckle from the audience. Clearly, during this part of the program the deck was stacked. Chalk it up to home field advantage.
Shiller was thrust into the unusual role of discussing a paper that critiqued his own work without having the chance to present his original research in any detail. Yet his remarks were, as usual for him, calm and well reasoned. He noted that both Miller and Kleidon had referred to Thomas Kuhn’s model of scientific revolutions, in which paradigms change only once a significant number of empirical anomalies are accepted as valid violations of the received wisdom. The papers by Kleidon and Miller amounted to a declaration that the revolution was, thankfully, not yet upon us. Here is the beginning of Shiller’s reply: “Maybe something as dramatic as a scientific revolution is in store for us. That does not mean, however, that the revolution would lead to ‘the abandonment of assumptions of rational expectations in favor of mass psychology.’” Instead, he explained: “I tend to view the study of behavioral extensions of these efficient market models as leading in a sense to the enhancement of the efficient market models. I could teach the efficient market models to my students with much more relish if I could describe them as extreme special cases before moving to the more realistic models.” Well said and still true.
As usual after such meetings, or after debates between political candidates, both sides were confident that they had won. The debate between behavioral finance researchers and defenders of the efficient market hypothesis was just beginning, and has been continuing for the last thirty years, but in some ways it all began that afternoon in Chicago. We will see where that debate has taken us in the next section of the book.
________________
* For a long time foundations and endowments operated in the same way, which was to leave the principal alone and spend the “income,” tending to push them to hold bonds and stocks that paid large dividends. Gradually this practice was recognized as silly, and these organizations adopted a more sensible rule, such as to spend a given percentage (say 5%) of a three-year moving average of the value of the endowment, allowing them to choose investments based on their long-term potential rather than their cash payouts. This change in policy allowed endowments to invest in new asset classes such as venture capital funds, which often do not pay any returns for many years.
18
Anomalies
An important aspect of Thomas Kuhn’s model of scientific revolutions, which came up at the end of the Chicago conference, is that paradigms change only once experts believe there are a large number of anomalies that are not explained by the current paradigm. A few scattered unexplained facts are not enough to upend the conventional wisdom. That conference was not the first time that the links between Kuhn’s ideas and what I was trying to do had crossed my mind. It was a topic I had thought about, but only on the sly. As someone who had until recently still been in the “promising” stage of his career, it would be viewed as brash, unseemly, and self-destructive to talk about my own work as something that could be part of a “revolution.” My goal was much more modest: just get a few more papers published and begin to establish the case that adding some psychology to economics was an activity worth pursuing. But I had certainly read Kuhn’s path-breaking book The Structure of Scientific Revolutions, and had secretly spent idle moments wondering whether anything like a paradigm shift could ever be possible in econo
mics.
A paradigm shift is one of the rare cataclysmic events in science when people make a substantial break with the way the field has been progressing and pursue a new direction. The Copernican revolution, which placed the sun at the center of the solar system, is perhaps the most famous example. It replaced Ptolemaic thinking, in which all the objects in our solar system revolved around the Earth. Given that the planets do not revolve around the Earth, it now seems odd to think that anyone could have made a geocentric model work at all. But for centuries astronomers using the geocentric system had in fact managed to do a pretty good job of explaining the movements of the planets, albeit with numerous somewhat ad hoc modifications of the basic model that were called epicycles: mini-circles around a main circular path along which the planets were thought to be rotating around the Earth.
At the Chicago conference, the speakers who were defending the status quo usually mentioned the idea of a paradigm shift with evident horror, with the gist of their remarks being that there was no reason to think we were standing on the precipice of a revolution. Of course, that they kept invoking it suggested there was at least some reason for concern among traditionalists. Their defense was usually to pick apart any given result and explain why it was not as critical as it seemed. If necessary, defenders of the traditional paradigm could always find some economics version of an epicycle with which to rationalize an otherwise embarrassing fact. And each single anomaly could be dismissed as a one-off puzzle, for which a satisfactory explanation was sure to exist if one looked hard enough. To create a real paradigm shift, I felt that we would require a whole series of anomalies, each calling for its own ad hoc explanation. At exactly the right time and place in my life, an opportunity to compile and document such a list of anomalies fell into my lap, and I had the good sense to seize the chance.
Misbehaving: The Making of Behavioral Economics Page 18
