Version classiqueVersion mobile

Economic Fables

Ariel Rubinstein

1. Rational, Irrational

Texte intégral

My father’s rationality

1I consider myself a very rational man, explaining all randomness with statistical tools and refusing to recognize the existence of the hand of fate and supernatural forces. I was born on Friday the 13th and scorn lucky numbers. I like Yeshayahu Leibowitz’s rationalistic definition of belief in God as accepting the yoke of Torah and commandments. This definition freed me from feeling ashamed of my lack of belief in a God who showers me with mercy and is furious about my sins. I believe in making a calculated analysis of decision-making situations, and always consider it appropriate to ask myself what I can do, what my objective is and what action would be best in my pursuit of this objective. I am often unsuccessful in doing this, but I still continue to try. While I regard rationality as an ideal way of life, it plays an even more prominent role in my consciousness as a polar opposite to behavior that I do not respect: I feel revulsion toward astrologers, readers of coffee grounds, and all sorts of experts adorned with academic degrees, who manipulate their audiences emotionally and whom I suspect of being driven by self-interest.

2I associate the rational approach with the path of study and learning. The opposite, irrational world is identified with urges and emotions. Rational versus irrational, mind versus emotion, study versus prayer – for me, all these terms describe the contrast between my father and my mother.

3My mother was proud to be a scion of a family from the Slonimer Hasidic community, a small branch of Hasidism founded in the mid-nineteenth century that was nearly wiped out in the Holocaust. My father, on the other hand, came from a devout family of Mitnagdim (opponents of Hasidism). When my father wanted to berate my mother, he would accuse her of coming from a Hasidic family. Nonetheless, on the eve of Passover my father would send me to the son of the rebbe (the Hasidic rabbi) to get three shmura matzos (special unleavened bread eaten during Passover), and the rebbe’s son would tell me that Eli, my mother’s brother, had studied in heder (a traditional Jewish elementary school) with his father – none other than the great rebbe himself. The thought that my mother’s brother had played as a child with the rebbe was for me like an encounter with royalty.

4I hate situations of emotional frenzy. The only ecstatic experience in which I participated was when my father would take me to the Simhat Torah celebrations at the Slonimer Shtibel, a synagogue that appeared to have been copied from a faded photograph of a lost Jewish town and rebuilt on the edge of the field between the Beit Shmuel neighborhood of Jerusalem and the border. I was squeezed into a dense ring of men in their holiday kapotes (traditional Hasidic long black coats). The smell was at first like that of the bath house, until the scent of soap gave way to the stink of male sweat. The elders of the community would carry the smaller, lighter Torah scrolls, while younger honorees would be called upon to carry the heavy ones. The Torah bearers would be surrounded by young yeshiva students, who would cling to the inner circle. I, as an appendage to the outer circle, would bump into tables and benches stacked haphazardly at the sides of the room, grasping my father’s hand on one side and the warm and moist hand of a stranger on the other side. I would try to move with the rhythm of the dance and mumble the words of the refrain that was repeated endlessly, despite the fact that I was unable to understand most of them. I still remember the words: ”My God, my God, why hast thou forsaken me?”

5But Simhat Torah was just once a year. During the rest of the days of the year, the Mitnagdim won.

6My adolescent rebellion erupted when I was about fourteen: I wanted to become fervently religious and asked my father to buy me a tallit katan (Jewish undergarment with fringes). On an afternoon stroll, at the entrance to Mea She’arim, my father told me that he had no objection to the change, but that before I don the external trappings I should learn Talmud and study halacha (Jewish law). He even offered to find a suitable teacher for me. My father responded in a rational way by demanding that I ”hear and understand” before ”doing”. He kept his cool and stamped out the rebellion without a battle.

7My mother believed in every person’s freedom of choice, even a child’s, and her wish for me was that I would fulfill everything I felt good about, as long as it was not dangerous. My father, just like the rational man who stars in economics textbooks, always had a clear objective: ensuring survival – physical and economic. Therefore, my father supported every military action the government conducted, vetoed every request from my mother to take a loan in order to buy living room furniture, and scolded me about unnecessary expenses. That was why my father wanted me to be an accountant or economist.

The rational man

8The rational man’s pattern of thought constitutes the foundation for most economic models. At the beginning of each course I teach on microeconomics or game theory, I find myself conducting an induction ceremony into the world of economic theory in which I expose the students to the portrait of the rational man in economics. In order to endow this ceremony with the appropriate air of dignity, I tell them that I should launch the presentation with a presidential fanfare. And then I begin to recite:

The rational man has preferences regarding the consequences that are likely to result from choosing various alternatives. When he is required to choose, he:

  1. asks himself what alternatives he has;
  2. clarifies to himself the consequence that would follow upon choosing each of the possible alternatives;
  3. chooses the alternative that leads to the best consequence (as expressed in 2) in accordance with his preferences, from among all of the reasonable alternatives in the situation (as expressed in 1).

9This is quite abstract and I will try to explain. In economics, models are nothing more than stories about interactions between units called decision makers. A decision maker in the model encounters decision problems to which he must respond. We have in mind situations like a buyer choosing a car from a catalogue. A fellow choosing a life insurance policy from among the plans offered in the insurance market. Parents deciding how many children they will have (a non-negative integer) by choosing at some point not to have any more. If I had to draw a decision maker, I would portray him as a person with two eyes, two ears, a nose and a mouth. He sits before a keyboard with many keys. His hand reaches out and must press one – and only one – of the keys. Pressing one of the keys leads to a particular consequence.

10A decision problem is described in economic models as a set of alternatives available to the decision maker, who is required to choose exactly one of the alternatives in the set. In presenting the decision problem as a set of alternatives, it is assumed that the way in which the alternatives are presented to the decision maker does not affect his decision. This means, for example, that a default option – if one exists – does not influence the decision: the decision of someone who is deliberating whether to sign his consent to donate organs after his death is not influenced by whether the appropriate box on the form has a check mark that he can remove, or whether the box is empty and he can add a check mark to indicate his consent to donate organs. This also means that the decision maker’s choice is the same regardless of whether the set of alternatives is described by a characteristic or explicitly. For example, the problem ”Choose one of the four largest countries in the world” is supposed to be identical to the problem ”Choose between the U.S., China, Canada and Russia.” The order of presenting the alternatives is also not supposed to have an impact on the decision maker’s choice: the choice between ”yes” and ”no” is not different from the choice between ”no” and ”yes.”

11All this is concealed in the innocent definition of the decision problem as a set of alternatives from which to choose. Of course, we are aware that factors such as a default option, the way a set of alternatives is described, and the order in which the alternatives enter the consciousness of the decision maker are indeed likely to have an influence on what most people choose in real life. However, we presume that the rational man does not take such non-essential factors into consideration and, therefore, these factors are excluded from the description of the decision problem. Thus, as if by the way, we insert elements of rationality into the description of the decision maker in economics.

12In regard to the identity of the decision maker – we view a decision maker as an independent decision making unit whose choices are not influenced by other decision makers. We generally refer to a single individual, but sometimes a decision maker is a group of individuals such as a family, committee or commercial enterprise. On the other hand, there are cases when an individual, let’s call him Moses, is split into two decision makers, Moses 1 is Moses after being slapped by his brother Aaron, and Moses 2 is the Moses who has calmed down the next morning. Moses 1 and Moses 2 have the same ID number and genetic makeup, but the considerations of Moses 1 and Moses 2 can be completely different, and neither Moses controls the actions of the other.

13When we introduce a decision maker into an economic fable, we equip him with operating instructions that guide him in responding to all of the decision problems that are likely to arise during the tale. Thus, for example, the description of a young person who applies to a group of universities should include details about which university he would choose if accepted by more than one, i.e., his order of priority. If he applies to three universities – Jerusalem, Tel Aviv and Haifa – the descriptions should include what he would do if accepted by all three, if only accepted by Jerusalem and Tel Aviv, if only accepted by Tel Aviv and Haifa, or only by Jerusalem and Haifa. Another example: many economic models include a type of decision maker we call a consumer. We imagine the consumer receiving a monetary budget that he can use to purchase products that are of value to him. Each product has a price – the number of units of money he must allocate to acquire a unit of the product. The consumer must decide how to allocate the budget between the various products. A complete description of the consumer should explain what he would consume (that is, which combination of products) at every budget level and at every possible price structure.

14It is possible to think of a decision maker as a machine that receives data about a set of alternatives he must choose from, and then creates output in the form of one of these alternatives. Psychologists and brain science researchers are interested in the structure of the machine that processes the data and reaches a decision. Traditionally, the economist is not interested, at least not as an economist, in the technical details of the machine’s operation. He is interested only in the connection between the input and output, between the decision problem and the action chosen.

15Here we reach the core of the assumption of rationality in economic models. A rational decision maker has a clear ranking of consequences in mind. We call this ranking preferences. When a rational decision maker faces a decision problem, he chooses the alternative that produces the best result in accordance with his preferences.

16Let’s take for example a decision maker who is entitled to a monetary reimbursement. The amount of the reimbursement is likely to depend on the date on which the decision maker collects what is owed to him. The result might take the form of ”receives X dollars T days from now.” Let’s assume, for example, that the decision maker feels that postponing collection of the money is equivalent to a loss of $3 a day. Such a decision maker will prefer, for example, to receive $100 in another ten days (which is equivalent to receiving $70 immediately, from his perspective) rather than to receive $200 in another 45 days (which is equivalent to receiving $65 immediately, from his perspective). In general, he assesses the result ”receives X dollars T days from now” by using the formula X-3T. This rule unequivocally defines the decision maker’s preferences. The decision maker’s objective function is expressed by X-3T. The choice of the best alternative means choosing the one for which the objective function gives the maximum value.

17According to the assumption of rationality in economics, the decision maker is guided by his preferences. But the assumption does not impose a limitation on the reasonableness of preferences. The preferences can be even in direct contrast with what common sense might define as the decision maker’s interests. According to this definition, a company director who chooses a strategy of production and marketing that minimizes the company’s profits is rational. An employer who evaluates an employee according to the numerical value of the letters in his name (known in Hebrew as gematria) is rational. Someone who is concerned only with the welfare of someone else is rational, even if this concern is detrimental to his own existential interests. A person who does what he does because ”they told me to do so” or because ”it was the first thing that came to mind” or ”it is not the best but it is nearly the best,” or someone who tries very hard to do what is impossible, all are irrational according to the accepted meaning in economics.

18It is not always clear whether a person is acting according to the paradigm of the rational man. When my daughter, Michal, was one year old, I wanted to check whether she displayed consistency in her choices. I knew that she distinguished between colors. We sprawled on the floor of the room. I placed three colored blocks in front of her – green, red and blue – and asked her to choose one of them. She took the green one. I rearranged the blocks in a different order and she chose the red one. Even after a dozen repetitions, no consistency was evident in her choices. I was really happy and so proud of her. Here was my daughter violating the most basic assumption of rationality – that there is some consistency in choices made. But then I realized that my excitement was premature. My daughter always chose the cube on the left. In other words, her choices were indeed consistent if seen from the perspective of the position of the cube rather than its color.

As if

19According to the traditional approach, economists are not interested in the question of whether the decision maker’s choice was preceded by a stage in which he actually conducted an explicit act of maximization, that is, if he chose the alternative that maximizes an objective function. The only thing that is important to the economist is that it is possible to describe the behavior of the decision maker as if he had conducted maximization. Let’s look, for example, at the worker who wakes up after sleeping for S hours. Let’s assume that the worker cannot control the number of hours of sleep but that he can decide how many of the remaining (24-S) hours in the day will be devoted to work and how many to leisure activity. Assume that in exchange for his labor, the worker will receive wages of W per hour. Let’s assume that the worker always allocates half of his waking hours to work. That is, he works (24-S)/2 hours and devotes the same number of hours to leisure. The worker does not conduct any maximization, but acts according to a rule that is ingrained in him, perhaps without even being aware of its existence.

20Ostensibly, it seems that the worker’s behavior does not fit the model of the rational man in economics. Nonetheless, we will see that the worker’s behavior can be described as if he has a particular objective function and always chooses the solution that maximizes it. The result for this worker is the combination of the number of hours of leisure (L) and the sum of money (M) he will earn that day. Let’s assume that the worker’s preferences among the group of outcomes are determined according to an odd criterion: the product of the leisure hours and the sum of money he earns. If he chooses L hours of leisure, he will work (24-S-L) hours and receive (24-S-L) × W units of money. Consequently, he chooses the number of leisure hours that will maximize the function L × (24-S-L) × W. In algebra class in high school, we learned that this function has a single maximum point at L = (24-S)/2. We found, therefore, that an employee who allots his time so that the function L x M receives the highest value possible will always (whatever S and W are) devote half of his waking hours to work. Maximizing this strange function is like describing the employee’s behavior. I cannot imagine a reasonable person who consciously maximizes the product of leisure hours and the sum of wages. But for the economist, in order to qualify the decision maker as being rational, it is enough that he can describe his behavior as if he maximized some preferences.

21I remember the moment as a student when I realized that the models in economic theory do not assume that the decision maker consciously tries to maximize his preferences, but only assume that the behavior of a decision maker can be described as if he had maximized some objective function. The words ”as if” were magic for me then. Suddenly, economics appeared more abstract and sophisticated than I had thought previously. Years went by until I understood that there is also a potential for deception in this sophistication. On the one hand, decision makers’ preferences in economics are presented only as a means to describe their behavior. In the previous paragraph we used the function of the product of wages and leisure hours only to describe the behavior of the worker who devotes half of his waking hours to work. Most economists, however, apply economic models to policy questions which require a criterion for the welfare of the individuals to be specified. Economists often identify an individual’s level of happiness with the preferences that explain his behavior. By this approach, if the decision maker chooses alternative A when B is also possible, this means that indeed he prefers A to B. It is far from obvious that the preferences used to describe the decision maker’s behavior correspond to his degree of happiness. Even if the decision maker’s behavior can be described as the result of maximizing some objective function, the objective might not relate to promoting his happiness. For example, it is possible that he consistently works to achieve a ”wrong” goal. And here is an absurd case: a decision maker has a clear concept of the essence of happiness, but he acts determinedly and consistently to actually diminish his happiness. A person like this is seen by economists as rational in the sense that he maximizes a clear goal – to worsen his situation as much as possible. But to identify his happiness with the preferences used to describe his behavior is like saying that the person is happier the less happy he is. This does not sound good.

22In other words, ”as if” is not only a magical phrase, as I felt back in the days when I was thrilled to discover the hidden treasures of economic theory. The underlying economic theory is not only that the decision maker can be described as someone who maximizes some objective, but also that he maximizes a function that expresses his happiness. It ultimately became clear to me that the phrase ”as if” is a way to avoid taking responsibility for the strong assumptions upon which economic models are founded.

23My entire professional life revolves around the definition of the rational man – someone who aspires to advance a well-defined objective or whose behavior can be expressed as the result of a process in which he asks himself what is desirable and what is possible, and chooses the best possible alternative. At the end of the induction ceremony that I conduct for my students I add some reservations about the status of the rational man in the real world. I emphasize that by using this definition we are not claiming that every person makes decisions in a way that is consistent with the definition of the rational man. I also note that the accepted definition in economics demands that the rational decision maker maximizes some function – but not necessarily his happiness. Yet it seems to me that these reservations fade away in the face of the magic of the clear formulation and joy of the formal description of the decision maker as someone who maximizes an explicit function. I believe that the image of the rational man that remains in the student’s mind is that of someone who maximizes some function, and this function also quantifies his happiness.

Dependence on presentation

24The professional literature is full of experiments that clash with the assumption of rationality. Psychologists, philosophers and economists conduct these experiments. Prominent among them are Amos Tversky and Daniel Kahneman and their followers. Those who worship economic thinking get angry. Those who do not like the ”square” mindset of the rational man applaud.

25Here is one well-known example (on the book’s website, you can experiment with the situations discussed in this example, and the rest of the chapter:​product.php/​148). Physicians at the university hospital at Stanford were randomly divided into two groups. The physicians in one group were asked the following question (Question 1):

An epidemic threatens the lives of 600 people. Two alternative and mutually exclusive plans of action are proposed:

Under Plan A, 400 people will die.

Under Plan B, there is a 1/3 chance that no one will die and a 2/3 chance that 600 people will die.

You must choose one plan of action.
Which would you choose?

26The physicians in the second group were asked the following (Question 2):

An epidemic threatens the lives of 600 people. Two alternative and mutually exclusive plans of action are proposed:

Under Plan C, 200 people will be saved.

Under Plan D, there is a 1/3 chance that 600 people will be saved and a 2/3 chance that no one will be saved.

You must choose one plan of action.
Which would you choose?

27The need to choose between plans A and B poses a dilemma: Plan A is bad, because it will not prevent the certain death of 400 human beings. In Plan B, there is hope (a 1/3 chance) that the plan will succeed and no one will die; but there is also a significant risk that the number of dead will reach 600.

28The choice between plans C and D also poses a dilemma: in Plan D, there is hope of saving all the patients, along with a significant risk that all 600 will die. Plan C is not especially encouraging, but it offers the certainty of saving a substantial number of patients.

29Some readers will be surprised to discover that Plan A and Plan C are identical: the death of 400 people is equivalent in this story to saving 200 people. Plans B and D are also identical: 0 mortalities means 600 survivors, and 600 mortalities is equivalent to 0 survivors. The problem of choosing between A and B is the same as the problem of choosing between C and D. Therefore, a rational decision maker in the first group who chooses A would also choose C if responding to Question 2; if he chooses B in Question 1, he would choose D in Question 2. The fact that Ais identical to C and that B is identical to D is obvious once it is pointed out. But it turns out that there are many intelligent people who fail to note they are identical, even when they are asked to respond to the two questions one after the other.

30In the original experiment, 78 % of the physicians who were asked to choose between A and B chose Plan B, the alternative that involves risk. 72 % of those who had to choose between C and D chose the sure plan, C. The physicians were divided into the two groups randomly. There is no reason to think that the hand of fate divided a large group of individuals into two groups in such a way that the characteristics of the physicians who responded to the first question differed significantly from the characteristics of those who responded to the second question. Thus, it is possible to conclude that about 78 % of the physicians at Stanford would choose B if asked Question 1 and 28 % would choose D if asked Question 2. The data do not rule out the possibility that 22 % of the physicians would choose A and C, and that 28 % of them would choose B and D. But, the data indicate that at least half of the physicians would choose B if asked to choose between A and B, and would choose C if asked to choose between C and D.

31We have noted a lack of consistency among at least half of the physicians who participated in the experiment. In decision problems under conditions of uncertainty that entail losses, we exercise different considerations than in decision problems in conditions of uncertainty that involve gains. When we look at the decision problem from the perspective of loss, we are drawn to a plan that offers real hope of preventing the tragedy, even if it entails a risk of enlarging its dimensions if it occurs. When we examine the decision problem from the perspective of salvation, as a choice between two life-saving medical actions, we prefer to go for the sure thing rather than take a gamble that would enable us to save a larger number of patients but entails a significant risk that no one would be saved.

32Kahneman and Tversky conducted this experiment using a routine experimental technique. As noted, the respondents were randomly divided into two groups, with one group assigned to respond to Question 1 and the other group assigned to respond to Question 2. This technique provides a solution for the concern that if the same person is asked to respond to both questions, one after the other, his response to the second question will be influenced by the fact that he had just answered the first question. It turns out that even when the same people are asked both questions one after the other, there is significant inconsistency between the answers to the two questions. More than 5,000 students of game theory were asked to respond to two questions posted on the web that are similar in format to the questions the reader can find on this book’s Internet site. The students were asked the two questions in the order presented here, with a number of completely different questions inserted between the two questions. 72 % of the students chose alternative B in the first question and far fewer, 49 %, chose D in the second question. The result regarding the choice between A and B is very similar to those of the original experiment whereas in regard to the second question, the percentage of respondents who chose D rose significantly, from 28 % in the original experiment to 48 % in the web-based experiment. Apparently, when the two questions were presented consecutively, some of the students did indeed become aware of the connection between the two questions and made sure to remain consistent with the choice they made on the first question. Nevertheless, more than a quarter of the students made the combination of choices B and C.

Begin and Rabin

331977 was one of the most memorable years of my life. I was in my first year as a doctoral student, full of enthusiasm about the world of formal models I had just discovered. In one of the classes, the lecturer referred to a simple model I had formulated and I was in seventh heaven. I had stepped out into the big world and was living away from my parents’ home for the first time. We were a few serious youngsters who had founded the ”Movement for a Different Zionism.” We envisioned it as a harbinger of the formation of a large political group that would step into the breach against Gush Emunim, the Israeli settlers’ movement. We were fearful of the messianism and extreme nationalism of certain national religious groups, and were disgusted with the settlement policy the Begin government was committed to and began to implement as soon as it took office. The excitement of political activity blended with the romantic intoxication of a Jerusalem summer and meetings that lasted into the small hours of the night in the quaint Nahlaot neighborhood.

34I had already encountered Begin’s rhetorical style when I was a child. My father took me to a soccer game only once, but many times to election rallies. At Menorah Square in Jerusalem and at the entrance to the Mea She’arim neighborhood, I heard Begin speak vehemently against the ruling Mapai semi-socialist party. My father would make fun of Begin, but still admired him enough to take me to shake his hand at a barmitzvah celebration where Begin was among the guests. When I was a child, I thought Begin’s rhetoric made him look as if he were playing the fool or clowning. Fifteen years later, in 1977, I was amazed to watch him enthrall the masses. I felt helpless and frustrated by the reactions of many of my friends, who extolled Begin for his rhetorical prowess and in the same breath criticized the rhetorical poverty of our own forces. I, who believed in the power of level-headed argument, did not regard Begin as a role model.

35Begin often explained his decisions in terms of carrying out duties and honoring rights: ”We must all make an effort to… We have to… But we are also obliged…” He would start by saying ”We must make sure that…” and ask ”What should we have done?” In a meeting with President Carter on 19 July 1977, Begin reached new heights of rhetoric:

Mr. President, in your country there are many cities with biblical names. You have eleven places with the name Hebron; five with the name Shiloh and seven with the name Bethlehem. Can you imagine a governor in one of these states prohibiting Jews from living in these cities? The Israeli government also cannot prohibit Jews from living in Hebron, Bethlehem or Beit El. It is our duty to…

36Begin’s arguments were generally based on ”our rights” and ”our duty.” One could think that there is room for discussion and disagreement regarding rights and duties. Did our forefathers command us to settle in Beit El in 1977? Why are we bound by the wishes of our forefathers? Are there other obligatory commands that contradict this ”duty”? However, in Begin’s rhetorical realm, there was no room to examine the limits of the possible and to identify the desirable. The preferred status of an action derived from its being considered part of our rights and our duties and not from its being the best action in light of the limitations of the possible, according to our worldview.

37Rabin, in contrast to Begin, had a measured tone and a matter of fact, down to earth style. I remember the satisfaction I felt one morning in the late 1970s when I heard Rabin interviewed on the radio. Speaking just like the rational man from the economic definition, he outlined the possible and the desirable, and after careful consideration, drew conclusions. During his victorious election campaign in 1992, Rabin made frequent use of the concept of priorities. In his book Service Notebook, he wrote: ”I have no doubt that the dangers of peace are a thousand times preferable to the gloomy certainty of war.” He asked himself what possibilities were available, even after he made a decision:

I shall always remember the moment just after deciding to mount an action: the hush, the sound of the door closing; and then the silence in which I remain alone… In that moment of great tension just before the finger pulls the trigger, just before the fuse begins to burn; in the terrible quiet of that moment, there is still time to wonder, alone: Is it really imperative to act? Is there no other choice? No other way? (from Rabin’s speech at the award ceremony of the Nobel Peace Prize in 1994).

38The distinction between Begin and Rabin corresponded to the difference in style between the slogans of the right and those of the left in Israel at that time. The leading slogans of the right claimed rights (”The Land of Israel, all of it is ours”), made statements that were ostensibly factual (”The Golan is an inseparable part of the State of Israel”), and expressed prayers (”Messiah now”), while the slogans of the left quoted preferences (”Peace is preferable to the greater Land of Israel”) and a demand (”Peace now”). It seems to me that the differences in style between the slogans of the left and right became less acute in later years, as all of Israeli politics was taken over by the culture of media advisors and public relations.

39In the final analysis, did Rabin’s preferences really have an advantage over Begin’s rights? In my view, the answer is not unequivocal. Begin entangled Israel for generations with the settlement enterprise and embarked on the unnecessary Lebanon War, but also signed the historic peace treaty with Egypt. Rabin ordered the army to break the limbs of the demonstrators in the first intifada and was responsible as IDF chief of staff in the Six Day War for Israel’s inhumane treatment of the remnants of the Egyptian Army in the Sinai, but his signing of the Oslo Accords ensured his place in history (and led to his death).

40As years went by, I realized that I think more like Begin than Rabin in regard to the occupation and the occupied territories. My unconditional opposition to ruling over another people did not derive from my formulation of the objectives that the State of Israel is supposed to achieve or from asking myself which possible policy would generate the best result in terms of these objectives. I simply feel an absolute duty not to be on the side of the occupier and oppressor, even if the occupation is economically beneficial and brings peace closer. Nonetheless, I do not have a shred of sympathy for Begin. Even his signing of the peace treaty with Egypt and the fact that he was subject to periodic bouts of depression did not soften my anger over his demagogic antics. Like the times when I was a child and wanted to use a book of logic to prepare myself for asserting irrefutable arguments against evil, I still find myself looking for ways to understand rhetoric and long to defeat demagoguery.

Mental arithmetic

41The following example is Kahneman and Tversky’s. Students were divided into two groups. Students in one group were asked:

Imagine that you have decided to see a show that costs $20. You purchased a ticket, but upon your arrival you discover that you have lost it. Will you purchase another ticket for the show?

42The question for the students in the second group was:

Imagine that you have decided to see a show that costs $20. Upon your arrival, you discover that you have lost a $20 bill. Will you purchase a ticket for the show?

43Rational principles, it seems, dictate that the respondents should answer the two questions in the same way. That is, regardless of whether the decision maker lost a ticket or a $20 bill, he will have $40 less than he had before if he purchases a ticket and sees the show; and if he does not purchase another ticket he will find himself with $20 less than what he originally had and misses the show. Thus, whether the decision maker lost a ticket or a $20 bill, he faces identical problems of choice. Therefore, we can expect that a rational man would make the same decision in the two cases. But in the original study, only 46 % of the participants said they would purchase a ticket after they lost a previous ticket, while 88 % answered that they would purchase a ticket after losing a $20 bill. On my website, among 1,500 students who were randomly assigned to respond to one of the two questions, the results were less pronounced but similar to the original results: 64 % and 80 %, respectively. We can summarize: decision makers are more inclined to purchase a ticket after losing a $20 bill than after losing a previously purchased $20 ticket.

44What is happening here? Decision makers assess the price of purchasing a ticket as the total expense involved in seeing the show. After losing a ticket, a decision maker tends to assess the price of an additional ticket at $40 rather than at its nominal value. After losing a $20 bill, he regards the loss as an event that is not directly connected to purchasing a ticket to the show, and he calculates the cost of the new ticket at only $20. This pattern of thinking is called mental accounting. The difference in the cost of a ticket, as it is calculated mentally, explains the readiness of some of the participants in the experiment to purchase a ticket in certain circumstances and not in others – those for whom the show is worth sacrificing $20 and but for whom a payment of $40 is already too much will purchase a ticket after losing a $20 bill, but not after losing a ticket.

45Kahneman and Tversky also report that when the two questions were presented one after the other, the percentage of participants with inconsistent responses declined. The inclination of nearly all of the participants is not to purchase another ticket after losing the first ticket. But some feel embarrassed after becoming aware that they would have purchased a ticket after losing a $20 bill. A person’s embarrassment in light of a certain choice reinforces the view that the choice is not rational.

My father died

46After the army of physicians and interns and nurses finished treating my father, the physicians permitted me to enter the scrubbed and polished room, which showed no traces of the last battle. The room was lit by a pale fluorescent light typical of hospital evenings. My father lay unconscious on a wide bed, attached to tubes. The instruments rose and fell at the pace of an ebbing life. A physician told me that it was important that I hold his hand. I tried for a moment, but did not discern any reaction and let go. I sat near my father, I looked out from the hospital window on Mount Scopus at Jerusalem on Earth and thought about the nurse in the white robe with two open buttons who came every half hour to replace the cocktail of medications being pointlessly infused into my father’s veins. Around midnight, I left and went home. I was awoken the next morning by a phone call informing me that the situation had deteriorated. When I entered the room, I saw only a white screen that hid the bed on which apparently my father’s body was lying. The physician did not say a word. He assumed that I understood. He asked me if I wanted to approach the bed. Without hesitation I said no. My father was no longer there and as a rational man I found no reason to see the body of a dead person. And then the physician gently asked me whether I objected to having an autopsy performed.

47I first encountered the concept of autopsies in posters by the Association for Defending the Dignity of the Dead, which often covered the walls of northern Jerusalem. The posters called on the public to fast and participate in demonstrations against autopsies performed by the ”butchers” from Hadassah hospital. Sometimes the advertisements featured black and white photographs of a stomach or brain after the physicians had sawed, cut, emptied and sewn it. My father, despite his rejection of religion, still observed Jewish tradition to some extent. I do not know whether he feared God or felt nostalgic toward a lost world. I also do not know what he thought about autopsies, because we never dared to speak about death.

48When the physician asked for my consent, I reminded myself that I was a rational man and that the operation would be performed on my father’s corpse and not on my father himself. The lump of flesh laying there was at most a container in which my ex-father had resided. It was no longer my father whose body was to be dissected, and therefore why should I object to the advancement of science? If the dead were not operated on, the physicians would be less proficient and medical knowledge would be the poorer. If I did not consent to the autopsy, the responsibility for the death of future patients would also be on my conscience.

49It was the dawn of the Friday before the Purim Festival. Mount Scopus was always the object of my dreams. I still sing to myself the ever popular hymn to Jerusalem ”From on Mount Scopus” in times of stress. My sister was born on Mount Scopus during the pre-state era. My father worked there as a clerk during the British Mandate. We were cut off from Mount Scopus before I was born. Walking hand-in-hand with my mother on Ezra Street in the early evening, I would stare at the trees and gray buildings on Mount Scopus, which beckoned like an enchanted garden whose gate – the Mandelbaum Gate – was locked. I would sometimes watch Maman, the police officer, as he peered through his binoculars from the balcony on the third floor of the house across the street, anxiously tracking the twice-a-month convoy ascending the hill, until announcing that the convoy had passed. A feeling of relief would then sweep over the street. I did my first year of university studies on Mount Scopus. Later, my two children would be born in the same building in which my father died. Now, on that Friday before Purim, I thought about the limits of rationality.

50The physician explained to me that by law I could not sign the forms for an autopsy on the spot, and needed to wait six hours after being notified of the death before signing the authorization. I was impressed by the logic and humaneness in the law. But it was impossible to tarry. The autopsy needed to be completed before the start of the Jewish Sabbath on Friday evening, and Friday was a short day. I controlled myself and did not expose my emotions. I just asked the physician, with the fear that characteristically accompanies questions posed to physicians, whether it was probable that someone’s life would be saved as a result of the post mortem. Without hesitation he answered in the affirmative. He explained that an autopsy would help physicians and students understand my father’s heart illness and that this might help in treating similar patients in the future. The physician had a pleasant countenance and wore a yarmulke (skull cap). At that moment, the yarmulke was for me a guarantee of his honesty.

51Six hours later, I hurried back to the hospital to sign the required form. Before I signed, I stipulated that the post mortem should only involve the parts of his body affected by the disease. The physician expressed satisfaction with my response and promised that my request would be honored.

52On Sunday afternoon, a few of my father’s acquaintances and some friends gathered at the funeral home in the Sanhedria neighborhood. The burial society in Jerusalem had a bad reputation at the time, but I was impressed by the respect and sensitivity the thick-bearded undertakers showed to the mourners. As is customary, I was asked to accompany the undertakers to identify the corpse, which had been cleansed, wrapped in a shroud and prepared for burial. One of the undertakers asked me in a despondent tone whether I had consented to an autopsy. I nodded hesitantly, like someone who was caught red-handed. I stuttered that I had authorized operating only on the parts of the body related to my father’s illness. The undertakers glanced at me with skepticism. Slowly, they unwrapped the bandages from the skull. It was the first time in my life that I looked upon the face of a dead person. My father, with his large face and without his false teeth, had already looked shrunken when he was in the hospital. I saw now that the skull had been cut open and sewn, as if they had emptied it. Then the undertaker covered the head again with the shroud.

53Apparently, I had allowed myself to be deceived. Later, I also failed to fulfill my father’s last wish – that I recite kaddish, the mourners’ prayer, during the shiva, the week-long period of mourning. I went to the synagogue once for mincha, the afternoon prayer service, but I felt out of place there and did not go back again. My father was dead, and a rational man has no reason to honor the wishes of dead people. Nonetheless, why did it bother me that my father’s brain had been cut? And why will I not forgive the physician who did not keep his word? And why did I not fulfill my father’s last wish? And why does this bother me at all?

Looking for reasons

54Tversky and Simonson asked the individuals in one group to choose between two alternatives: $6 in cash or a high-quality pen. Some 36 % of them chose the pen. The individuals in the second group were asked to choose between three alternatives: $6 in cash, the same high-quality pen or a simple pen that was obviously inferior to the high-quality one. The percentage of those who chose the high-quality pen rose to 46 %.

55What is the explanation for this finding? Some of the individuals presumably have no interest in a pen and they would choose $6 when responding to either of the two questions. Some of the individuals may actually need a pen and in any case would choose the high-quality pen. But there are some individuals for whom the choice between the high-quality pen and the cash is not clear-cut. The existence of an additional alternative, a simple pen, provides ”a good reason” to resolve this indecision and to choose the high-quality pen: the high-quality pen is clearly preferable to the simple pen, while the $6 is not obviously preferable to the simple pen.

56On this book’s website I cannot give out pens (and do not wish to do so). Therefore, two similar, but hypothetical, questions are posed on the site. Half of the readers are asked to respond to the following question:

Assume that you are planning a trip and are interested in buying a digital camera. You find a store that sells three models produced by the same manufacturer. The models and their prices are quite similar to one another. The only differences between them is the score they received in a professional magazine and the number of pixels.

ModelAreceived a score of 9.1 and has 6 megapixels.
Model B received a score of 8.3 and has 8 megapixels.
Model Creceived ascore of8.1 and has 7megapixels.

Which model would you choose?

57The other readers are asked to respond to an identical question – except for the fact that they must choose only between models A and B.

58More than 1,500 students were asked to respond to one of the two questions prior to the publication of the book and the following results were received: among those who had to choose between models A, B and C, 65 % chose B and only 1 % chose C. These results show the alertness of the participants in the experiment, even without offering real prizes. Among those who had to choose between A and B only, the percentage choosing B declined to 53 %.

59In this experiment, we compared the choices of decision makers who chose between A and B with those who chose between A, B and C. It turns out that the percentage of individuals who chose B from A, B and C was significantly higher than the percentage who chose B from only A and B. The decision makers who would choose B if choosing from the larger selection and A if asked to choose from the smaller selection cannot be described as rational: if a rational decision maker has preferences and B is the preferred alternative from among A, B and C, then B should also be the preferred alternative between A and B.

60Here is another, similar example, this time from Shafir, Simonson and Tversky: individuals in one group had to choose, hypothetically of course, between A, a camera priced at $170, and B, a more sophisticated camera priced at $240. A low percentage of individuals in this first group chose B. The individuals in another group were given a third possibility, C, a much more sophisticated camera, priced at $470. And here, with the addition of the very expensive camera, the percentage of those choosing B rose (to some extent). A reasonable explanation for the change is that adding the expensive camera C made B into the middle option between A and C. When we arrange the alternatives in our minds linearly, the central location of an alternative is a reason to choose it.

61Was it necessary to conduct experiments to prove the tendency of people to choose the middle option? This is well known not only to those engaged in marketing but also to my son, Yuval. When he was small, he told me that when his favorite fast food chain added an extra-large cup to the drinks menu, in addition to the large and medium cups, it was conducting a marketing ploy aimed at encouraging people to choose the large cup. This logic should have also led to the conclusion that in an election battle between a candidate from the right and a candidate from the left, it would be beneficial for the former if an extreme right-winger joins the race. Political experience shows that this is not always true: adding an extreme right-wing candidate often arouses aversion toward the moderate right-winger.

62No one argues that the irrational considerations attributed here to decision makers are always decisive considerations. All of the examples presented here are only intended to illustrate the existence of considerations that sometimes influence decision makers and are contrary to the assumption of rationality. We have a lot to learn before understanding when these considerations arise in decision making and what determines their weight vis-à-vis other considerations.

What would my mother have said?

63We use expressions like a dead person rolling over in his grave, my grandmother would have been happy to see me now, or, I observe the Sabbath because my forefathers observed the Sabbath. In our dreams, we see our deceased loved ones summoning us to do things that we might not have done if they had not appeared to us. For those who believe that the deceased are situated somewhere and are watching over us until we are gathered up to join them, it is easier to justify such behavior as rational. But as someone who believes that the souls or spirits of the dead are not hovering anywhere, I do not have much respect for those who use the kingdom of the dead to try to influence what happens on the planet of the living.

64On one Rosh Hashanah (Jewish New Year) morning in faraway Princeton, I was informed that my sister had been rushed to a hospital in Jerusalem. At the end of the holiday, en route to Israel, I prepared myself for the worst. At the entrance to the hospital ward, a physician met me and explained her condition in detail: a ”manufacturing defect” of nature had been discovered in one wall of an artery in the brain; nature had been negligent and had not provided a warranty. The wall of the artery was too thin and, with age, could not withstand the pressure. It had inflated like a balloon and exploded; its contents had leaked and perhaps caused damage to adjacent brain cells. There was a probability of 90 %, the physician said, that the initial leak would lead to a second and fatal leak. The solution: a dangerous operation – though its description sounded like a routine plumbing operation. And here the physician confronted me with the most difficult decision of my life: the operation could be performed in Jerusalem, where the surgeons had limited experience with this procedure, or we could fly my sister to a medical center in the U.S., with the most experienced physicians in this type of surgery. On the one hand, undergoing the operation in the U.S. would diminish the risk of irreversible damage that could occur during surgery. On the other hand, it would involve a logistical project the like of which I had never had to contend with in the past. I was concerned about detaching her from the supportive environment. I also felt a twinge of fear about the possibility that in a foreign environment she might ultimately be operated on by a surgeon who did not have extensive experience. It was impossible to receive an unequivocal recommendation from the physicians. This situation arose soon after the community of neurosurgeons in Israel had been harshly attacked by a public figure who had traveled overseas for a brain operation and complained about the ostensibly inferior professional capabilities of neurosurgeons in Israel. The physicians were on the defensive and showed a lack of confidence in their recommendations. Friends I visited at 2am offered me conflicting advice.

65Only rarely had I faced a real dilemma that was so similar to the decision problems under uncertainty that we discuss in our models. The alternatives were clear: an operation in Israel, or an operation in the U.S., or no operation. The last option was clearly inferior to the other two. The results were a matter of chance and the dice were in the hands of God. I had time to consult, learn and weigh the arguments. It was solely my decision to make. I was an adult and had even written several papers on the theory of decision making. If decision making theory has any practical value, it should have become evident then.

66I lay on my bed and felt a tightness in my chest, signs of the tension and distress I was experiencing. I could not decide. I needed a convincing reason. And then I found it. ”What would my mother have said if she were alive?” I asked myself, despite the fact that my mother had died several months earlier and even before her death had lost the ability to make real decisions. ”My sister belongs to our mother. Our mother is responsible for my sister’s fate and she will decide.” Everything became clear and the dilemma was resolved: it was clear that my mother would have chosen for my sister to be operated on in the place where she was born and in the city in which she lived. The decision was made.

They do not understand the causal connection

67Let’s return to the picture of the decision maker who sits at the keyboard and knows that each keystroke leads to a given result. Let’s say that we know and that he knows that pressing the A key leads to receiving $1 and pressing the B key leads to receiving $2. It is clear to us that the decision maker is interested in receiving as much money as possible: he says so, and every time that we offer the decision maker two envelopes with $1 and $2 he chooses the larger sum. Rationality, as we understand it in life and also as it is perceived in economics, requires that the decision maker understands the connection between pressing the A or the B key and the consequences ($1 or $2, respectively) and that he will press the B key. We can state this in a more abstract way: when there is a causal connection between an action and a result, the assumption of rationality includes the requirement that the decision maker is aware of the causal connections and will choose the action that will generate the preferred result, without making an error.

68But human beings err, and do so systematically. Here is another example by Tversky and Kahneman:

Imagine a die with six sides, four of them green (G) and two of them red (R).

You must choose one of three strings of the letters G and R. The die will be rolled 20 times and you will receive $25 if the series of results includes the string you chose.

The three strings are:

  1. RGRRR

Which string would you choose?

69Choosing any of the three strings will lead to a result that in our professional terminology is called a lottery. If the string appears in the series of rolls of the die, the decision maker will receive a prize; if the string does not appear, he will not receive anything. It is reasonable to assume that the decision maker will prefer one lottery over another if it offers him a higher probability of winning a prize.

70We note that every series that includes the second string (GRGRRR) also includes the first string (RGRRR). Every series of rolls of the die that awards a prize to someone who chooses the second string will also award a prize to those who choose the first string. Moreover, the string RRGRRR is the first string with the addition of another R at the beginning; thus, there are series that award a prize to those who choose the first string but not to those who choose the second string. Therefore, without calculating probabilities, one can discern that the first alternative offers a higher probability of winning a prize than the second alternative. Nonetheless, the overwhelming majority (65 %) of the 125 students who participated in the original study chose the second string (GRGRRR). A similar result was found among 2,000 students who responded to the question (without receiving any real reward) as presented on the book’s website: 59 % chose the second string. And, incidentally, only 4 % chose the third string (GRRRRR), which is easily seen to be less likely than the second string.

71As noted, the rational decision maker must be aware of the causal connections between his choice and the result that derives from it, and he is supposed to choose an action that generates the result that he views as preferable. This assumption is sometimes violated due to errors, as in the experiment discussed here. It is clear that most of the respondents misunderstood the connection between the choice of the string and the probability of winning the prize. They did not understand that choosing the string GRGRRR means choosing an option with a lower probability of winning than the string RGRRR. The reason, apparently, is related to the fact that probability is often confused with representativeness. Given that the probability of the result G on any roll of the die is 2/3, the string RGRRR, in which only 1/5 of the letters are G, is perceived as less typical than the string GRGRRR, in which 1/3 of the letters are G. But, the probability of finding the string RGRRR in the series of rolls of the die is always greater than the probability of finding the string GRGRRR.

72Not only errors hinder the understanding of the connection between a choice and its result. Sometimes the problem stems from the fact that we feel that we are able to influence the connection between the action and result, despite knowing full well that we have no such influence.

73Imagine that you are a student. The teacher, who generally means what he says, announces that a memorization test will be held the next day on a text you have never read, and you would like to achieve a high score. You have to decide whether or not to prepare for the test. If you prepare, you will have to give up your favorite leisure activity, but you will ensure your success in the test. If you do not prepare, you will be able to spend your time as you wish, but you will fail the test. You clearly understand the connection between the action and the consequence. As a rational person, you will weigh the fun you are missing while preparing for the test and the satisfaction of success, on the one hand, against the pleasure derived from your leisure activity and the unpleasantness of failing the test, on the other. But let’s assume that you noticed that in the past a strange coincidence has occurred: when you repeated a secret word, the test did not take place. If you choose to use your magical power and do not prepare for the test, you will fit the definition of being irrational despite the fact that you are doing your best within the framework of your beliefs and are trying to affect the connection between an action and its result. It will immediately become clear that you would not be the only one I would label as irrational…


74In high school, I was a good student, but only in the subjects that interested me: mathematics, Jewish law and citizenship. I did not do homework in the rest of the subjects. I would come to class before everyone, lie in wait for the first students and copy from them the homework in all the subjects I detested – literature, history, chemistry and biology – in return for the homework I prepared in math. The school conducted many surprise exams. The teacher would enter the classroom and pronounce the words that even now unnerve me a bit: ”Take out paper for a test.” I usually did not know the answers to the questions in the test and I became expert in copying them from my neighbor behind the broad back of my friend Yuval, who sat in front of me. During the moments of dread, when the teacher entered the classroom, I discovered a surprising correlation. Always, yes always, if I muttered the words ”There is going to be an exam, there is going to be an exam” when the teacher was about to enter the classroom, there wouldn’t be one. All of the exams took place on days when I failed to mutter my abracadabra.

75This is how I encountered the power I have over the events in the world. When I say to myself: ”This bad thing will happen,” it actually does not occur. Just yesterday I lost my university ID card. The loss ignited an exaggerated burst of emotions in me. The ID card can be reissued. In the worst case, I would need to inquire about the procedures for replacing a lost ID card, to report the loss, to call and check the office hours, to stand in line until the new ID is issued… that’s all. Really not so terrible. Nonetheless, I was terribly agitated. I looked in my wallet again and again; I searched everywhere and did not find it. I returned home to check whether the ID had perhaps fallen from my pocket when I was undressing or dressing. Nothing. It was not on the floor, not on the bed, and not under the bed. There was no ID card anywhere. I decided to check just one more time in case it was hiding in my wallet after all. But before I began the final probe, I muttered to myself: ”The ID card is lost, the ID card is lost.” I again rummaged through the wallet and… there it was, in one of the pockets of the wallet, tucked behind a credit card.

76I did not pray to God. I did not speak with creatures from outer space. All I did was note the connection between what I say and what happens, a dependence that has proven itself time and again as an act of magic. As a rational person, I am struck by the urge to look for an explanation for this strange coincidence. Of course, I suspect that the success of my spell is related to the fact that I chose to use it in cases when it was likely that the thing I feared would not really occur in any case. The rational person within me protests: why should I shatter this wonder? After all, I had found an effective tool for protecting myself in stressful situations. And perhaps I had simply reached the limits of my own rationality.


77The heroes of most economic models are rational decision makers. The traditional economic view, the one that dominates in the textbooks at least, asserts that the assumption of decision makers’ rationality enables us to describe human behavior quite accurately. The flood of experimental results in the field of economics, primarily from cognitive psychology, contradicts this view.

78Results that are inconsistent with the model of the rational person elicit skeptical reactions (sometimes justifiably). The attitude of economists toward results that conflict with the paradigm of the rational man reminds me of the natural reaction to the optical illusion generated when subconscious mechanisms create a picture that does not conform to our perception of the world. In both cases, we smile, become irritated, suspect that someone is pulling a fast one on us, are happy to discover that even ”perfect” beings fall into the snares that nature sets for us all, and we look for explanations.

79Some criticize the experiments pertaining to decision making because the participants are offered insignificant incentives. The important decisions of life, the critics say, involve interests that are much more significant than a few dollars and therefore we cannot infer what people will do when faced with major economic decisions from the way they behave when offered the chance to earn negligible sums. I do not agree with this criticism. First, not only fateful decisions are important. In life, people make many small decisions and the cumulative economic impact of these decisions is significant. And second, most of the major economic decisions are made by people for whom such decisions are an everyday matter. The small decisions we make in life are no less important to us than the major decisions of the rich and powerful are to them. I see no reason to think that when senior executives decide on matters involving millions they exercise different considerations than those that common folks use when making decisions about a few dollars.

80In most of the experiments we mentioned in the previous sections, the participants were only asked to imagine a decision problem and did not receive any compensation (not even symbolic) related to their choices. In my view, material compensation for participants in experiments is completely unnecessary. People are very good at imagining hypothetical situations. The fact that the results of the experiments in which no material incentives were provided are very similar to those received when participants were given material incentives indicates that offering incentives usually makes no real contribution to the identification of patterns of thinking used in decision making.

81It is customary to evaluate research results via a statistical test. At best, this entails an attempt to assess the chance that we are drawing an erroneous conclusion from the data. For example, let’s return to the question of the epidemic that was presented to physicians at Stanford. In the original experiment, of 152 physicians who responded to Question 1, 109 (78 %) chose option A. Of 155 physicians who responded to the same but differently worded question, Question 2, only 34 (28 %) chose C. We deduced from these results that the differences in wording between the two questions influenced the responses. We only assigned the title rational to those who chose A in the first question and C in the second, or B in the first question and D in the second. We concluded that a large group of individuals acted irrationally. But perhaps this conclusion is wrong and the differences in percentages between those who chose A in the first question and C in the second question can be attributed to pure coincidence? Perhaps all of the physicians at Stanford are rational, but some prefer the sure plan (A or C) while others prefer the more risky plan (B or D)? Perhaps, by sheer coincidence, the physicians were divided into the two groups in such a way that it just happened that more of those who preferred A (and, therefore, also C) were in the group asked Question 1, while more of those who prefer red B (and, therefore, also D) were in the group asked Question 2?

82A statistical test is designed to examine the extent of unreasonableness of the assumption that the results we received were only coincidental. We assume that all 307 participants are rational and would respond in the same way to the two questions: 143 of them prefer A to B (and, therefore, also C to D), while the other 164 participants prefer B to A (and, therefore, also D to C). A statistical test is based on assumptions about the random factors involved. The test we employ here (Fisher’s exact probability test) assumes that the 307 participants were randomly assigned to two groups, one with 152 participants who responded to Question 1 and the second with 155 participants who responded to Question 2. The test assumes that all of the possible distributions of the 307 participants into two groups of 152 and 155 have the same probability. The test calculates the probability that the randomness created a distribution that is so biased that at least 109 of those who preferred A (and C) happened to be in the first group. Statistical programs greatly facilitate such calculations and it turns out that the probability of such an event is very (very!) close to zero. It is so close to zero that the possibility that the results are consistent with the hypothesis of rationality is unreasonable, and we reject the hypothesis. Incidentally, it was not necessary to have such dramatic results in order to reject thehypothesis. According to the conventional criteria, the test results would be considered significant even if the percentage of those choosing C dropped from 78 % to just 72 % (and not to 28 % as occurred).

83I have not addressed statistical tests at all in this chapter. The findings are so clear that it seems to me that such an addition would constitute nothing more than paying lip service to the professional conservatism expected in such reports. Incidentally, Kahneman and Tversky, for example, also did not bother to conduct (or at least to report) a statistical test that examined the results they obtained in the epidemic problem.

84In general, the mechanical use of the concept of statistical significance is dangerous. The logic in its use rests upon important assumptions that are usually ignored or taken for granted but which should be examined. Researchers and newspaper readers love to use indices and rarely ask themselves what stands behind them. For example, the conventional tests of significance in economics completely ignore factors such as errors in measurement, documentation, analysis and reporting – factors that could have a major impact on the validity of the results. And, of course, the researchers have interests and biases and these, consciously or subconsciously, are liable to influence the reported results. It seems to me that the uncertainty regarding the credibility of the researchers is much more significant that the uncertainty taken into account in the conventional statistical tests. Therefore, I would be more impressed by two experiments conducted by two different researchers with small samples and results that are not considered significant, than an experiment conducted by one researcher with a sample twice as large and a result that is considered significant.

Is there any need for experiments at all?

85Experiments are born in the feverish and fruitful mind of a researcher who is not detached from reality and knows how to surmise the considerations that pass through the minds of human beings faced with decision making problems. Whoever conceived the problem of the lost theater ticket certainly knew that people feel compelled not to purchase another ticket after losing the original one because the loss makes the purchase seem more expensive. The researcher understood that some of the participants find this consideration to be decisive when deliberating whether to purchase another ticket, while some take comfort in thinking that it really does not cost $40, but just $20. The product of the experiment is qualitative, supporting the assumption that more people would include the loss in their calculation of cost after losing a previous ticket than after losing a $20 bill.

86Quantitative results have very limited significance in any case because the sample of participants does not represent more than a group of psychology, economics, or MBA students in a particular university. The qualitative results are usually insights that common sense had already suggested prior to the experiment. So why is there a need for experiments at all? Why is it not possible to suffice with self-reporting by researchers, as is customary in philosophy? For years I thought that experiments in economics were nothing but a waste of research funds. I still do not find great relevance in quantitative results, and I believe that the most reliable test of the reasonableness of an idea is that of common sense. But my appreciation of experiments grew after I realized that composing an array of questions that succeeds in illustrating a certain thought process is a work of art, and after I became convinced that common sense sometimes does deceive even the most experienced of us.

87In recent years, I have myself been guilty of conducting experiments. I felt the captivating excitement when results started to arrive and a hypothesis became a proven fact, or the pervasive disappointment when it became apparent that ”something went wrong here.” I have never bet at horse races, but I imagine that the feelings of a researcher as the results of an experiment come in are similar to a gambler’s feelings when the horses are racing round the track. And incidentally, the risk taken by researchers is no less than that of professional gamblers. At stake are the researcher’s honor, satisfaction, professional advancement and the monetary compensation awarded to him.

88My personal experience with experiments has led me to doubt the validity of the results of experiments in economics. Economists are not cheats, but like everyone else they make mistakes, for the most part unwittingly, and have a tendency to further their own interests. The economist wants his results to confirm his hypothesis. He is sure that he is right and the experiment is merely designed to confirm what he knows. He regards caution and meticulousness as unnecessary obstacles on the path of advancing human knowledge. I myself felt an urge to refrain from extending an experiment after the results I received in the initial stage were favorable, and I felt compelled to check the results seven times when they did not support a hypothesis I ”knew” to be correct. The fear of being embarrassed if the conclusions are refuted by other researchers is almost non-existent in economics because we have no tradition of checking data and repeating experiments.

89It is strange that according to the economic view of the world, people are economic agents who respond to incentives, primarily material ones. The economist describes them as aspiring to attain an objective that rates money, and perhaps status, highly, and gives a low rating to moral values. In this world view, all human beings are economic agents and their actions should be regarded from the perspective of their motives. All of them, except a group of angels who look at the world from above: the economists.

Rationality on the defensive

90There are those who defend the assumption of rationality, arguing that behavior that appears to be irrational at first glance is indeed rational if only we define the decision problem correctly. Take for example the case in which we must choose one alternative from among a very large set of possibilities. What we normally do is we examine a relatively small number of options until we find an alternative that satisfies us. At that moment, we know that there might be better options than the one we are about to choose. Ostensibly, we are not acting rationally. We are not choosing the best alternative from among the set of choices. But it turns out that this pattern of behavior can be explained within the framework of the rational decision maker. In order to do this, one must describe the decision maker not only as someone who decides which alternative to choose, but rather as someone who also must decide when to bring the decision process to a close, taking into account the physical or mental demands the decision process entails.

91There are those who justify the assumption of rationality by evolutionary considerations: if people in the world had used patterns of behavior that clash with rationality, a manipulator would have emerged to exploit their irrationality for his own benefit and to their detriment – until they realized this and changed their behavior or until they became extinct.

92For example, consider a decision maker who is ready to pay a dollar to exchange object A for object B, and is ready to pay a dollar to exchange B for object C, and is ready to pay a dollar to exchange C for A. Let’s also assume that he prefers more money to less money in his pocket. The decision maker is not rational: we cannot attribute to him an order of priorities that explains his behavior. If someone like this initially holds A and $1 million, he would be susceptible to a manipulator offering him a series of exchanges for a dollar. At first the manipulator would offer to exchange A for B for one dollar. After this offer was accepted, he would offer to trade B for C for one dollar, and then C for A for one dollar… and so on. Before long, our decision maker would wake up and change his behavior – and if not, he would find that his bank account had been depleted.

93This is a fascinating argument for the assumption that individuals whose behavior is inconsistent with the assumption of rationality cannot survive for long. But it is not so obvious that such individuals would become extinct. The manipulators waste energy in hunting the irrational creatures, and may be expected to cease their activities before they destroy all their source of sustenance (or they may become extinct themselves). The survival of those who eradicate the irrational individuals requires the survival of their irrational victims. This is similar to the argument that in nature the fact that one creature is stronger than another and depends on it for food does not mean that the weaker creature will become extinct, but actually explains the mutual existence of both.

94And there are those who approach the criticism of the assumption of rationality as if they were a labor union, arguing that the criticism is fundamentally destructive, that it does not offer alternative working frameworks and should therefore be ignored.

95In recent years, economic theory has actually responded positively to criticism and we have witnessed the development of fields of research called ”bounded rationality” and ”behavioral economics” – fields that lay an infrastructure for building economic models in which the rational person is replaced by decision makers with other characteristics.

And do I want to be rational?

96I refuse to answer this question rationalistically. That would require me to grapple with the meaning of life, to define the goals of my life, to clarify which paths of life are open to me, to face the certainty of death and to embrace maximization – which I have no idea how to describe, let alone solve. I prefer not to go down that route.

97I can only say that I take pleasure in watching the rational man defeated. I do not like his perfection. The more I imagine him as flesh and blood, the more I realize that he is unbearable, even inhuman. The fact that this person does not actually exist – and that it is possible, using simple tricks, to make fun of anyone who considers himself rational – really makes me happy. I like the fact that I myself would make ”problematic” choices in most of the problems I presented in this chapter. When explaining the concept of the rational man and discussing these issues in my economics courses, I make a point of saying that even after discerning the inconsistency in my behavior, I will continue to break all the rules and promise to violate the economic principles of rationality.

98”Why?” I ask myself. First, I recoil from the aura of dogmatism emanating from the assumption of rationality in economics. I have a feeling that economics preaches that there is something that should be called correct behavior. That makes me feel as if I am denied the right to be myself and am forced into a mold designed to train me to behave as the economic models assume I behave. I refuse to obey.

99Second, the assumption of rationality is supposed to make me predictable. I do not want someone to anticipate my moves and I do not want to be able to predict the activity of other people. I never understood why the world would be better if there were someone who knows what we will do before we do it. I am prepared to make significant sacrifices if only to do the opposite of what the perfect prophet predicts when he appears one day. And even if he tarries – I will wait for him. I so much want to defeat him.

100Nonetheless, when all is said and done, I see myself as a very rational person, explaining every coincidence with statistical tools, refusing to recognize the existence of the hand of fate and supernatural forces.


Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.


Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search