Behavioural Economics

Capitolo 1: How We Decide

Opportunity cost

Activities that you gave up in order to pursue this current activity. Economists call this “imputed value”; the idea that you could and should assign a value to activities. When you decide to undertake a particular activity, you have implicitly decided not to undertake another activity.

Rationality in decision making

Economists start by assuming a certain amount of rationality on the part of humans. When faced with choices, humans engage in a clearheaded calculation of the benefits and costs; we possess foresight and are able to undertake complex calculations, and finally, that we are interested in maximizing what economists refer to as utility. Furthermore, we do not have the luxury of choosing anything we want. Our choices are constrained, in the sense that they depend not only on what we want, but also on what we can afford.

Psychologists, on the other hand, typically start by assuming humans are “boundedly rational”. “Bounded rationality” is that we have limits on our cognitive abilities; we face constraints of time, computing ability and foresight; we often fall prey to biases and errors of judgement. Some problems are hard, some others we experience only infrequently. This implies that we have more experience and expertise with some problems than with others. Buying groceries does not pose much of a challenge; we do it all the time. But buying a computer requires more research; buying a car even more so. Buying a house is a decision we make infrequently. This means that the possibility of making mistakes are higher, the less facility we have with the task at hand.

How do economists and psychologists approach decision making?

This is the economics approach to understanding human behaviour: Start with a “model”, essentially a theory, based on some assumptions about behaviour. See what predictions this model makes about behaviour in different circumstances and then see if actual behaviour corresponds to this “model”. If not, then possibly tinker with the model to make it more realistic. All models are wrong, but some models are useful. Starting with a model and making adjustments along the way is a good way to approach many scientific questions.

Psychologists, on the other hand, typically start by focusing on the “target” behaviour. Instead of trying to write down an assumption-based model of such behaviour, psychologists prefer to collect data on such behaviour. Once a particular phenomenon has been documented many times, psychologists then call it a “theory”, which is really a different name for a model, because once a theory is formulated, it can then be used to make predictions about future behaviour under similar circumstances. The psychology approach then, is descriptive.

Consider the example of “ego depletion theory” proposed by Roy Baumeister. Baumeister had two groups of participants placed in two separate rooms around a table. In one room, the table had a bowl of freshly cut cauliflowers. The table in the other room had a bowl of freshly baked cookies, whose aroma filled the room. Participants were forbidden from consuming any of the contents of the bowls in either room. Afterwards, both sets of participants were given challenging mathematical problems to solve. It turned out that the group that had been in the room with the cookies gave up much more quickly on trying to solve the problems, as opposed to the people in the room with the cauliflowers.

Many tasks in day-to-day life require us to exert self-control. But there is a limit to how long and how much self-control we can exert. The people who had to refrain from eating the cookies had to exercise much greater self-control than those asked to stay away from the cauliflowers. As a consequence, the people in the room with the cookies had used up most of their quota of self-control, and hence had a harder time concentrating on the mathematical problems that followed. The people facing cauliflowers had a larger reserve of self-control left over and hence persisted at the problem-solving task much longer.

Baumeister went on to call this “Ego depletion theory”; the idea that over the course of time, our ability to exert self-control diminishes. This is why most of us find it much easier to resist dessert (or chocolates) earlier in the day when our stock of self-control is high, but find resistance more difficult at night, when our stock of self-control is running low.

The descriptive psychological approach: going from a regularly observed phenomenon to building a theory about a potential explanation of those observations. One drawback to this descriptive approach, however, is that there are instances when a particular phenomenon may be explained by more than one theory. In such cases, it becomes difficult to know whether something is caused by one reason or another. This also, at times, makes it difficult to predict future behaviour and has occasionally created problems with the replicability of experimental results. There is nothing inherently superior about one approach or the other. They are different ways to figure out the foibles of human behaviour and each has its advantages and disadvantages. This is partly why behavioural economists have come to rely on elements of both approaches and have done so with great benefit.

Infusing psychology into economics: The guessing game

Let us play the “guessing game”. I hand out small slips of paper and ask each member of the group to choose a number between 0 and 100. The winner is the person who comes as close as possible to one half of the average number chosen in the group. So, what number did you choose? If you are like Sheldon Cooper, then you probably chose zero! In fact, this is what dispassionate reasoning in line with traditional economic thinking will suggest. Why? Suppose everyone chooses 100. Then the average for the group is 100. One half of the average is 50. In that case, you should choose 50 to come as close as possible to one half of the average. But suppose everyone thought that! In that case, everyone chooses 50, so the average is 50 and one half of the average is 25. So, should you choose 25? No, because if everyone figured this out and chose 25, the average would be 25 and you should choose 12.5 to come as close as possible to one half of the average. If you continue this process (or “iterative” thinking to throw in some economic jargon) then the only number that survives is zero. Everyone should choose that. Again, using jargon, this is the equilibrium outcome.

Over time, everyone should converge to zero; over time, those who choose other things should learn the folly of their ways. Of course, that does not quite help if you only get one shot at playing this game. But I am sure it will not come as a surprise to you to learn that this is not what happens. In fact, you will seldom “win” if you chose zero but homing in on zero is not going to do the trick. You need to actively anticipate what others in the group are doing. A lot of students who chose zero. I do not believe it is the case that these people are necessarily smarter than others. They may be, but it is also possible that, in looking around the room, they are forming judgements about their peers: what they know, how many levels of thinking they will apply. Economists sometimes call these models of “cognitive hierarchy” or “Level-k thinking”.

There may be some naive types who stop at the first step, say, by choosing 50 in the one-half target game. Let us call this “Level Zero” thinking: everyone will choose 100, so the average is 100 and, therefore, one half of the average is 50. Then there is a second type that applies Level One thinking, which takes them to 25; then there is a Level 2 type that goes to 12.5. Actually, it is a little bit more complicated than that. Excluding the naive Level Zero types, each type might actually be considering how many players of which type there are in the group and conditioning their responses based on that.

Random group of 100 people and asked each of them the same question. The challenge for the team is to pick the most popular choice among those external respondents. So, for instance, if the most frequent choice among those other 100 people is “wallet”, then the team will earn the most points if they also pick “wallet”; the team gets fewer points if they pick the second most popular choice, even less if they choose the third most popular choice and so on.

In many of these instances then, if you applied nothing but equilibrium reasoning, you would most likely not “win”. Your better option would be to look around, get a sense of the room, think of how many levels of reasoning your peers might apply. In short, pure reasoning is not enough. Of course, you need to engage in some amount of reasoning, since stopping at the first step is also not the recipe for success. But this needs to be complemented by some amount of emotional input and psychological intuition about those you are interacting with; in short, reason leavened with passion: this is the very essence of behavioural economics.

Types of decision making

We can think of three types of decisions we are frequently called upon to make. First, decisions that involve one person and probably does not impact others. For example, buying a lottery ticket. Second, decisions that involve small groups or one-on-one bargaining situations. For example, bargaining for a good price. Here, what we do often depends on what others are doing and our beliefs about them. Many such decisions involve strategic thinking, in the sense that they often require us to anticipate how the other party may respond to choices we make. At times, we need the tools of game theory to think through these situations. But, equally, many such decisions are influenced by notions of social norms; what is fair; what is appropriate; what is expected. Third, there are decisions that we make in markets with many buyers and sellers. Often, an individual will have limited power over the prices that prevail in these markets, given that (s)he is one of many in the market. So, at times, these decisions may be similar to individual choice problems – how much do you want to buy at a given price in the supermarket – but not always.

Ana’s problem of choice

Ana’s dad gave her $100 for her birthday and told her that she can choose how to spend this money. Suppose that Ana can spend it on two goods. In a feat of imagination, let us call them Goods X and Y. Certainly, Ana can save some money for later, but let us assume that she does not need to worry about saving. So, for now, Ana intends to spend $100 that she has at her disposal.

Ana cannot buy everything she wants; her choices are constrained

So how should Ana spend her money? Well, it depends, first of all, on what X and Y cost. This will dictate what she can afford. Suppose X costs $20 per unit (P = $20) while Y costs $10 per unit (P = $10). How many of each can she afford? What if she spent her entire $100 on X? Since X costs $20 each, she can buy 5 of these. How about Y? She can buy 10 of these since they cost $10 each. This allows us to draw Ana’s budget constraint. She can buy either 5 units of X or 10 units of Y, or any combination of the two that add up to $100. At one extreme, Ana buys 5 of X and zero of Y; at the other extreme, she buys 10 of Y and zero of X. If we join these two extreme points (or the two intercepts) with a straight line, then this is Ana’s budget constraint. She can choose either of the two extremes, or, more likely, another intermediate point on the line such that the total is $100. For instance, she can buy 4 units of X, costing $80, and 2 units of Y, costing $20, for a total of $100. Or 3 of X costing $60 and 4 of Y costing $40, etc. The line is negatively sloped because, in order to buy more of one good, Ana needs to give up some of the other. We will refer to a possible pair of choices of X and Y as a “consumption bundle”.

Figure 1.7 shows that the absolute slope of the budget constraint is given by the ratio of the two prices (P_X/P_Y). This is because the slope of a line is given by the rise over the run. In this case, the rise is (Income/P_Y) while the run is (Income/P_X). Simple algebraic manipulation shows that this is the same as (P_X/P_Y). Of course, this is the absolute slope (or the magnitude of the slope) since the actual slope is negative. Given the supposed values of P_X and P_Y, this value is equal to ($20/$10) which is 2. This simply implies that you can buy 2 units of Y for each unit of X; that is, if you buy one more of X, then you will need to buy two fewer of Y. Figure 1.8 shows the nature of Ana’s choices. She can afford any consumption bundle that lies on the budget constraint but nothing that lies beyond (to the north and east) of that line. Of course, she can consume inside the shaded triangle, but, as I argued above, she wants to spend the entire $100 and, therefore, will operate on the line itself, rather than under the line.

What happens if Ana’s mum (who is more fiscally responsible and conscious about not spoiling the children) adds $20? In that case, Ana will have $120. If Ana has $120, then, at the two extremes, she can either buy 6 units of X ($120/$20) and zero of Y, or she can buy 12 units of Y ($120/$10) and zero of X. In this case, her budget constraint shifts out and to the right, parallel to its former self. If Ana has only $80, then she can buy 4 of X ($80/$20) and zero of Y, or she can buy 8 of Y ($80/$10) and zero of X. If that happens, then her budget constraint shifts inward to the left, but once again in a parallel manner. Figure 1.9 shows what happens in each case.

Let us go back to the case where Ana has $100. What happens if the price of either X or Y goes up or down? Suppose the price of X changes. Since the price of Y has not changed, the intercept of the budget constraint on the Y-axis remains unchanged. Ana can still buy 10 of Y and zero of X if she wishes. But now she can buy more or less of X depending on whether X has become less or more expensive. If X has become cheaper than $20 per unit and Ana spends her entire $100 on X, then she can buy more than 5 units. In this case, the budget line swivels outward around the Y intercept. However, if X has become more expensive than $20 per unit and Ana spends her entire $100 on X, then she can buy fewer than 5 units. In this case, the budget line swivels inward around the Y intercept.

On the other hand, if the price of X does not change but Y becomes more or less expensive, then the intercept of the budget constraint on the X-axis remains unchanged. Ana can still buy 5 of X and zero of Y if she wishes. But now, she can buy more or less of Y depending on whether Y has become less or more expensive. If Y has become cheaper than $10 per unit and Ana spends her entire $100 on Y, then she can buy more than 10 units. In this case, the budget line swivels outward around the X intercept, whereas if Y has become more expensive than $10 per unit and Ana spends her entire $100 on Y, then she can buy fewer than 10 units. In this case, the budget line swivels inward around the X intercept. Figure 1.11 shows the respective situations.

Ana knows what she can afford, but how does she choose among these?

How does Ana choose a particular combination of X and Y? Economists assume that consumption of goods yields utility, which is economese for happiness or satisfaction. Consumption bundles that offer more utility are preferred to those that provide less utility. Utility is measured on an “ordinal” scale. An ordinal measure is one where we assign a higher number to things that are more preferred, but the numbers are not meaningful per se. (On the other hand, a “cardinal” measure is one where the numbers have a particular meaning, such as when measuring height, weight, age or distance.) Let me digress briefly to highlight the distinction between “total utility” and “marginal utility”. Clearly, the value we ascribe to goods (or bundles) and, therefore, the price we are willing to pay, depends on the utility we get from them.

The word value, it is to be observed, has two different meanings, and sometimes expresses the utility of some particular object, and sometimes the power of purchasing other goods which the possession of that object conveys. The one may be called “value in use;” the other, “value in exchange.” The things which have the greatest value in use have frequently little or no value in exchange; on the contrary, those which have the greatest value in exchange have frequently little or no value in use. Nothing is more useful than water: but it will purchase scarcely anything; scarcely anything can be had in exchange for it. A diamond, on the contrary, has scarcely any use-value. Water may be more essential, but the higher price of diamonds is dictated by its scarcity. The fact that an additional unit of water is worth much less than an extra carat of diamond is not because of total utility calculation but because an additional unit of diamond is worth a lot more than the additional unit of water. In other words, the marginal utility, the extra utility we get from consuming one more unit of the same good, is higher for diamonds than for water.

Ana’s preferences to be “transitive”. If given two choices, E and G, Ana says that she prefers G over E, and given two choices E and F, she says that she prefers E to F, then if asked to choose between G and F, Ana should prefer G to F. This is a basic assumption regarding consistency of choices. This assumption is often violated in real life with consequences for many things.

An indifference curve shows a combination of consumption bundles that give Ana the same utility. So, just like the budget constraint, the indifference curve must also have a negative slope: if you want more X (Y) then you have to give up some Y (X). Indifference curves as being parallel to one another. This is because indifference curves for the same person cannot intersect. Or maybe they can, but we do not want them to. Why? Take a look at Figure 1.14. In that figure, bundles E and A lie on the same indifference curve; so, they must both yield the same utility to Ana. This means Ana would be indifferent between E and A. But bundles E and B both lie on the same indifference curve, so they must also yield the same utility to Ana. Hence, Ana should be indifferent between E and B. But then, using the transitivity assumption, given a choice between A and B, Ana must be...