3
Selfishness, Rationality, and Utility

In the previous chapter, I already used a word that needs careful definition: selfishness. We say a person is selfish when she lacks kindness toward and responsibility for others.1

Let me also take the opportunity to introduce another word: rationality. We say a person is rational if she takes actions that are consistent with her goals. Otherwise, she is irrational.

There are three main kinds of irrationality. One is to forget or otherwise neglect the goals one has. It is typically irrational to get heavily drunk the night before an important exam.2

A second kind of irrationality is to draw wrong conclusions from the information at hand. In complicated situations, we are all prone to some irrationality, because our cognitive capacity is limited. A perfect human chess player will never exist. However, in this book, we shall only be concerned about systematic kinds of erroneous conclusions. For example, many people are unable to understand how little weight should be attached to new information when we already have a lot of existing information.

The third kind of irrationality is to hold incorrect beliefs for other reasons. Sometimes, we misread ourselves. Many people, especially males, are prone to overestimate their own abilities. Others, often females, underestimate theirs. Sometimes, we misread the external world. For example, during the Coronapandemic some people believed the virus did not actually exist. I will classify such beliefs as irrational even if the people holding them have been fed with misinformation.3

According to our definition, a selfish person can be rational or irrational. Likewise, an unselfish person can be rational or irrational.4

For centuries, social scientists have thought hard about how to describe rational decision-making in a simple way. For most purposes, the answer is expected utility theory. Expected utility theory describes the decision-maker's preferences through a utility function and the beliefs through a probability function. The decision-maker chooses the action that gives the highest expected utility.5

Do not be confused. Nobody in their right mind thinks that people carry around utility functions that they consciously maximize. Rather, the idea is that if the decision-maker can rank the alternatives she faces in a consistent manner, then the ranking can be described by certain utility and probability functions. And, if she consistently chooses according to that same ranking, the decision-maker behaves as if she maximizes expected utility. In this sense, we can even describe the behavior of animals, plants, and other organisms as being rational.

Utility and Desire

For a selfish person i who only cares about their own material consumption, we can describe the preferences by any increasing function Ui(ci), where Ui denotes person i's utility and ci denotes person i's consumption. That may sound abstract, but it just means that if consumption is higher, then utility is higher too.

If the person is not entirely selfish, but also cares about others' consumption, then we must include their consumption in the person's utility function also; we may then write it Ui(c), where c = (c1, ., cn) is the consumption of the n people that i cares about.6 The function then needs to describe how i weighs others' consumption relative to her own consumption. For example, suppose the person's own consumption goes up by five percent and a friend's consumption goes down by ten percent, does the person's utility go up or down?7

If a person makes decisions under uncertainty, the utility function needs to rank consumption lotteries as well as consumption levels. In this book, I simplify whenever I can.8 Therefore, in the presence of uncertainty I (almost always) assume people maximize the expectation of whatever their objective is. In the terminology of decision theory, we assume people are risk neutral.

Exercise 3.1

1. Write down a simple utility function for a selfish person.
2. Write down a simple utility function for an altruistic person.
3. Write down a simple utility function for a person who loves equality.

Utility and Duty

Economists often assume people care only about consequences. Sociologists object that this assumption is typically wrong; people are also directly concerned about the actions they take. One reason is that society imposes its understandings and values. Teachers and religious leaders actively promote some values and criticize others. Within the family, parents similarly reinforce values they want their kids to hold.9 Last, but not least, governments impose laws and regulations.

When parents, teachers, or government officials forbid or advise against some action, people may feel bad about taking that action regardless of any sanctions. To represent the desire to obey duties, we consider utility functions of the general form

where s = (s1, ., sn) are the actions that relevant people take. We call these actions s (for strategy) rather than a (for action) for reasons that will become clear.

The function v denotes social values. It does not have an index i, since social values are determined at the group level. The parameter ?i indicates the extent to which i cares about social values, i.e., the person's loyalty or dutifulness.10 In general, a person's loyalty to group values might be affected by the perceived legitimacy of the imposed values. For example, when a country is occupied or ruled by a tyrant, many people will feel less loyalty to the government's decrees than they do to those of a democratically elected government. However, to keep notation simple we shall mostly treat ?i as a fixed number here, in which case loyalty is entirely a characteristic of the individual rather than a function of the situation.

The function v may take several forms, but I shall usually imagine that it only takes non-positive values.11 The idea is that social values are constraining; they get us to do things we otherwise might not want to do. Hence, when possible, we might want to avoid situations where such social pressure applies.12

When society has clear rules about right and wrong, it is reasonable to assume that the function v(s) jumps down, as s goes from being permitted to being forbidden. For example, in a society where people are supposed to tell the exact truth and not to steal, a moral individual feels guilty about a selfish lie or theft, even if the transgression has little material consequence.

There is a famous saying that economics is about individuals' choices, sociology about how individuals don't have any choices to make.13 Equation 3.1 allows both of these extreme views. When ? goes to infinity, the individual does what society asks. They follow a logic of appropriateness. In a society without social pressure, the individual focuses entirely on attaining her own goals.14 They follow a logic of consequences.

Note that under the "standard" economic assumption that people ignore social pressure, we have ui(s) = Ui(c(s)). This formulation still allows altruism and fairness, for example. Only if person i is selfish, do the own goals coincide with the own consumption, or formally ui(s) = Ui(ci(s)). Unless the society is completely without social pressure or populated only by people immune to it, both U(c(s)) and v(s) matter for understanding behavior, to different degrees.

It is easy to think that in individualistic societies, social pressures are more limited than in collectivistic societies and that people will therefore be inherently more selfish in individualistic societies. However, this is not correct. Rather, collectivistic societies use punishments and rewards to enforce conformism, whereas individualistic societies encourage people to respect others and to be kind. Thus, trust in the kindness of others is greater in individualistic societies than in collectivistic societies.15 Liberalism is not a license to be selfish, neither in theory nor in practice.16 As an illustration of this point, let me share with you a favorite segment from Torbjørn Egner's liberal children's book When the Robbers Came to Cardamom Town from 1955. The Constable is singing about the Law of Cardamom, which is short and simple (my translation from Norwegian):

You should never bother others
You should be forever kind
But beyond that there is nothing more to mind

That is, only when justice and charity have received their dues are you free to do what you want.17

Utility and Evolution

Where do preferences come from ultimately? Evolutionary theory posits that behavior will be geared to maximize the reproduction of the organism or, more precisely, the organism's genes. Simple organisms do not...