Predictably Irrational

by Dan Ariely

  In this final chapter, I will present an experiment that offers one more example of our predictable irrationality. Then I will further describe the general economic perspective on human behavior, contrast it with behavioral economics, and draw some conclusions. Let me begin with the experiment.

  TO GET TO the bottom of the sudsy barrel of questions that we thought of at the Carolina Brewery, Jonathan and I decided to plunge in—metaphorically, of course. We started by asking the manager of the Carolina Brewery to let us serve free samples of beer to the customers—as long as we paid for the beer ourselves. (Imagine how difficult it was, later, to convince the MIT accountants that a $1,400 bill for beer is a legitimate research expense.) The manager of the bar was happy to comply. After all, he would sell us the beer and his customers would receive a free sample, which would presumably increase their desire to return to the brewery.

  Handing us our aprons, he established his one and only condition: that we approach the people and get their orders for samples within one minute of the time they sat down. If we couldn’t make it in time, we would indicate this to the regular waiters and they would approach the table and take the orders. This was reasonable. The manager didn’t know how efficient we could be as waiters, and he didn’t want to delay the service by too much. We started working.

  I approached a group as soon as they sat down. They seemed to be undergraduate couples on a double date. Both guys were wearing what looked like their best slacks, and the girls had on enough makeup to make Elizabeth Taylor look unadorned in comparison. I greeted them, announced that the brewery was offering free beer samples, and then proceeded to describe the four beers:

  (1) Copperline Amber Ale: A medium-bodied red ale with a well-balanced hop and malt character and a traditional ale fruitiness.

  (2) Franklin Street Lager: A Bohemian pilsner-style golden lager brewed with a soft maltiness and a crisp hoppy finish.

  (3) India Pale Ale: A well-hopped robust ale originally brewed to withstand the long ocean journey from England around the Cape of Good Hope to India. It is dry-hopped with cascade hops for a fragrant floral finish.

  (4) Summer Wheat Ale: Bavarian-style ale, brewed with 50 percent wheat as a light, spritzy, refreshing summer drink. It is gently hopped and has a unique aroma reminiscent of banana and clove from an authentic German yeast strain.

  Which would you choose?

  □ Copperline Amber Ale

  □ Franklin Street Lager

  □ India Pale Ale

  □ Summer Wheat Ale

  After describing the beers, I nodded at one of the guys—the blond-haired guy—and asked for his selection; he chose the India Pale Ale. The girl with the more elaborate hairdo was next; she chose the Franklin Street Lager. Then I turned to the other girl. She opted for the Copperline Amber Ale. Her boyfriend, who was last, selected the Summer Wheat Ale. With their orders in hand, I rushed to the bar, where Bob—the tall, handsome bartender, a senior in computer science—stood smiling. Aware that we were in a hurry, he filled my order before any of the others. I then took the tray with the four two-ounce samples back to the double-daters’ table and placed their beers in front of them.

  Along with their samples, I handed each of them a short survey, printed on the brewery’s stationery. In this survey we asked the respondents how much they liked their beer and whether they had regretted choosing that particular brew. After I collected their surveys, I continued to observe the four people from a distance to see whether any of them took a sip of anyone else’s beer. As it turned out, none of them shared a sample.

  Jonathan and I repeated this procedure with 49 more tables. Then we continued, but for the next 50 tables we changed the procedure. This time, after we read the descriptions of the beers, we handed the participants a small menu with the names of the four beers and asked each of them to write down their preferred beer, rather than simply say it out loud. In so doing, we transformed ordering from a public event into a private one. This meant that each participant would not hear what the others—including, perhaps, someone they were trying hard to impress—ordered and so could not be influenced by it.

  What happened? We found that when people order out loud in sequence, they choose differently from when they order in private. When ordering sequentially (publicly), they order more types of beer per table—in essence opting for variety. A basic way to understand this is by thinking about the Summer Wheat Ale. This brew was not very attractive to most people. But when the other beers were “taken,” our participants felt that they had to choose something different—perhaps to show that they had a mind of their own and weren’t trying to copy the others—and so they chose a different beer, one that they may not have initially wanted, but one that conveyed their individuality.

  What about their enjoyment of the beer? It stands to reason that if people choose beer that nobody has chosen just to convey uniqueness, they will probably end up with a beer that they don’t really want or like. And indeed this was the case. Overall, those who made their choices out loud, in the standard way that food is ordered at restaurants, were not as happy with their selections as those who made their choices privately, without taking others’ opinions into consideration. There was, however, one very important exception: the first person to order beer in the group that made its decisions out loud was de facto in the same condition as the people who expressed their opinion privately, since he or she was unencumbered, in choosing, by other people’s choices. Accordingly, we found that the first person to order beer in the sequential group was the happiest of his or her group and just as happy as those who chose their beers in private.

  BY THE WAY, a funny thing happened when we ran the experiment in the Carolina Brewery: Dressed in my waiter’s outfit, I approached one of the tables and began to read the menu to the couple there. Suddenly, I realized that the man was Rich, a graduate student in computer science, someone with whom I had worked on a project related to computational vision three or four years earlier. Because the experiment had to be conducted in the same way each time, this was not a good time for me to chat with him, so I put on a poker face and launched into a matter-of-fact description of the beers. After I finished, I nodded to Rich and asked, “What can I get you?” Instead of giving me his order, he asked how I was doing.

  “Very well, thank you,” I said. “Which of the beers can I get you?”

  He and his companion both selected beers, and then Rich took another stab at conversation: “Dan, did you ever finish your PhD?”

  “Yes,” I said, “I finished about a year ago. Excuse me; I will be right back with your beers.” As I walked to the bar to fill their order, I realized that Rich must have thought that this was my profession and that a degree in social science would only get someone a job as a beer server. When I got back to the table with the samples, Rich and his companion—who was his wife—tasted the beers and answered the short questionnaire. Then Rich tried again. He told me that he had recently read one of my papers and liked it a lot. It was a good paper, and I liked it, too, but I think he was just trying to make me feel better about my job as a beer server.

  ANOTHER STUDY, CONDUCTED later at Duke with wine samples and MBA students, allowed us to measure some of the participants’ personality traits—something the manager of the Carolina Brewery had not been thrilled about. That opened the door for us to find out what might be contributing to this interesting phenomenon. What we found was a correlation between the tendency to order alcoholic beverages that were different from what other people at the table had chosen and a personality trait called “need for uniqueness.” In essence, individuals more concerned with portraying their own uniqueness were more likely to select an alcoholic beverage not yet ordered at their table in an effort to demonstrate that they were in fact one of a kind.

  What these results show is that people are sometimes willing to sacrifice the pleasure they get from a particular consumption experience in order to project a certain image to others. When people order food
and drinks, they seem to have two goals: to order what they will enjoy most and to portray themselves in a positive light in the eyes of their friends. The problem is that once they order, say, the food, they may be stuck with a dish they don’t like—a situation they often regret. In essence, people, particularly those with a high need for uniqueness, may sacrifice personal utility in order to gain reputational utility.

  Although these results were clear, we suspected that in other cultures—where the need for uniqueness is not considered a positive trait—people who ordered aloud in public would try to portray a sense of belonging to the group and express more conformity in their choices. In a study we conducted in Hong Kong, we found that this was indeed the case. In Hong Kong, individuals also selected food that they did not like as much when they selected it in public rather than in private, but these participants were more likely to select the same item as the people ordering before them—again making a regrettable mistake, though a different type of mistake, when ordering food.

  FROM WHAT I have told you so far about this experiment, you can see that a bit of simple life advice—a free lunch—comes out of this research. First, when you go to a restaurant, it’s a good idea to plan your order before the waiter approaches you, and stick to it. Being swayed by what other people choose might lead you to choose a worse alternative. If you’re afraid that you might be swayed anyway, a useful strategy is to announce your order to the table before the waiter comes. This way, you have staked a claim to your order, and it’s less likely that the other people around the table will think you are not unique, even if someone else orders the same dish before you get your chance. But of course the best option is to order first.

  Perhaps restaurant owners should ask their customers to write out orders privately (or quietly give their orders to the waiters), so that no customer will be influenced by the orders of his or her companions. We pay a lot of money for the pleasure of dining out. Getting people to order anonymously is most likely the cheapest and simplest way to increase the enjoyment derived from these experiences.

  But there’s a bigger lesson that I would like to draw from this experiment—and in fact from all that I have said in the preceding chapters. Standard economics assumes that we are rational—that we know all the pertinent information about our decisions, that we can calculate the value of the different options we face, and that we are cognitively unhindered in weighing the ramifications of each potential choice.

  The result is that we are presumed to be making logical and sensible decisions. And even if we make a wrong decision from time to time, the standard economics perspective suggests that we will quickly learn from our mistakes either on our own or with the help of “market forces.” On the basis of these assumptions, economists draw far-reaching conclusions about everything from shopping trends to law to public policy.

  But, as the results presented in this book (and others) show, we are all far less rational in our decision making than standard economic theory assumes. Our irrational behaviors are neither random nor senseless—they are systematic and predictable. We all make the same types of mistakes over and over, because of the basic wiring of our brains. So wouldn’t it make sense to modify standard economics and move away from naive psychology, which often fails the tests of reason, introspection, and—most important—empirical scrutiny?

  Wouldn’t economics make a lot more sense if it were based on how people actually behave, instead of how they should behave? As I said in the Introduction, that simple idea is the basis of behavioral economics, an emerging field focused on the (quite intuitive) idea that people do not always behave rationally and that they often make mistakes in their decisions.

  In many ways, the standard economic and Shakespearean views are more optimistic about human nature, since they assume that our capacity for reasoning is limitless. By the same token the behavioral economics view, which acknowledges human deficiencies, is more depressing, because it demonstrates the many ways in which we fall short of our ideals. Indeed, it can be rather depressing to realize that we all continually make irrational decisions in our personal, professional, and social lives. But there is a silver lining: the fact that we make mistakes also means that there are ways to improve our decisions—and therefore that there are opportunities for “free lunches.”

  ONE OF THE main differences between standard and behavioral economics involves this concept of “free lunches.” According to the assumptions of standard economics, all human decisions are rational and informed, motivated by an accurate concept of the worth of all goods and services and the amount of happiness (utility) all decisions are likely to produce. Under this set of assumptions, everyone in the marketplace is trying to maximize profit and striving to optimize his experiences. As a consequence, economic theory asserts that there are no free lunches—if there were any, someone would have already found them and extracted all their value.

  Behavioral economists, on the other hand, believe that people are susceptible to irrelevant influences from their immediate environment (which we call context effects), irrelevant emotions, shortsightedness, and other forms of irrationality (see any chapter in this book or any research paper in behavioral economics for more examples). What good news can accompany this realization? The good news is that these mistakes also provide opportunities for improvement. If we all make systematic mistakes in our decisions, then why not develop new strategies, tools, and methods to help us make better decisions and improve our overall well-being? That’s exactly the meaning of free lunches from the perspective of behavioral economics—the idea that there are tools, methods, and policies that can help all of us make better decisions and as a consequence achieve what we desire.

  For example, the question why Americans are not saving enough for retirement is meaningless from the perspective of standard economics. If we are all making good, informed decisions in every aspect of our lives, then we are also saving the exact amount that we want to save. We might not save much because we don’t care about the future, because we are looking forward to experiencing poverty at retirement, because we expect our kids to take care of us, or because we are hoping to win the lottery—there are many possible reasons. The main point is that from the standard economic perspective, we are saving exactly the right amount in accordance with our preferences.

  But from the perspective of behavioral economics, which does not assume that people are rational, the idea that we are not saving enough is perfectly reasonable. In fact, research in behavioral economics points to many possible reasons why people are not saving enough for retirement. People procrastinate. People have a hard time understanding the real cost of not saving as well as the benefits of saving. (By how much would your life be better in the future if you were to deposit an additional $1,000 in your retirement account every month for the next 20 years?) Being “house rich” helps people believe that they are indeed rich. It is easy to create consumption habits and hard to give them up. And there are many, many more reasons.

  The potential for free lunches from the perspective of behavioral economics lies in new methods, mechanisms, and other interventions that would help people achieve more of what they truly want. For example, the new and innovative credit card that I described in Chapter 7, on self-control, could help people exercise more self-control within the domain of spending. Another example of this approach is a mechanism called “save more tomorrow,” which Dick Thaler and Shlomo Benartzi proposed and tested a few years ago.

  Here’s how “save more tomorrow” works. When new employees join a company, in addition to the regular decisions they are asked to make about what percentage of their paycheck to invest in their company’s retirement plan, they are also asked what percentage of their future salary raises they would be willing to invest in the retirement plan. It is difficult to sacrifice consumption today for saving in the distant future, but it is psychologically easier to sacrifice consumption in the future, and even easier to give up a percentage of a salary increase that one
does not yet have.

  When the plan was implemented in Thaler and Benartzi’s test, the employees joined and agreed to have their contribution, as a percentage, increase with their future salary raises. What was the outcome? Over the next few years, as the employees received raises, the saving rates increased from about 3.5 percent to around 13.5 percent—a gain for the employees, their families, and the company, which by now had more satisfied and less worried employees.

  This is the basic idea of free lunches—providing benefits for all the parties involved. Note that these free lunches don’t have to be without cost (implementing the self-control credit card or “save more tomorrow” inevitably involves a cost). As long as these mechanisms provide more benefits than costs, we should consider them to be free lunches—mechanisms that provide net benefits to all parties.

  IF I WERE to distill one main lesson from the research described in this book, it is that we are pawns in a game whose forces we largely fail to comprehend. We usually think of ourselves as sitting in the driver’s seat, with ultimate control over the decisions we make and the direction our life takes; but, alas, this perception has more to do with our desires—with how we want to view ourselves—than with reality.

  Each of the chapters in this book describes a force (emotions, relativity, social norms, etc.) that influences our behavior. And while these influences exert a lot of power over our behavior, our natural tendency is to vastly underestimate or completely ignore this power. These influences have an effect on us not because we lack knowledge, lack practice, or are weak-minded. On the contrary, they repeatedly affect experts as well as novices in systematic and predictable ways. The resulting mistakes are simply how we go about our lives, how we “do business.” They are a part of us.

  Visual illusions are also illustrative here. Just as we can’t help being fooled by visual illusions, we fall for the “decision illusions” our minds show us. The point is that our visual and decision environments are filtered to us courtesy of our eyes, our ears, our senses of smell and touch, and the master of it all, our brain. By the time we comprehend and digest information, it is not necessarily a true reflection of reality. Instead, it is our representation of reality, and this is the input we base our decisions on. In essence we are limited to the tools nature has given us, and the natural way in which we make decisions is limited by the quality and accuracy of these tools.