The Economic Way of Thinking
Economics is best considered not as an area of study, but rather as an approach to understanding the social world. Economists do spend a lot of time studying markets, but the economic method can be applied to topics as diverse and seemingly non-economic as marriage (Becker 1981), religion (Iannaccone 1998), terrorism (Phillips 2011) and, as we show in this book, politics and public administration. At the core of the economic approach is the assumption of instrumental rationality : that individuals have preferences and act in order to satisfy them.
On its strongest interpretation, instrumental rationality can be taken to mean that individuals know whatās best for them and will ruthlessly pursue their own interests at all costs. This notion gives rise to Homo Economicus, the hyper-rational and sociopathic character who has earned the scorn of critics from outside the economics profession, as well as heterodox economists. More modestly, instrumental rationality forms the basis of rational choice theory as an organizing principle around which theoretical and empirical investigation can be structured. This requires only that preferences satisfy the basic conditions required for coherence. Beyond that, any instrumental preference (i.e. preferences over states of the world) can be straightforwardly modelled using rational choice theory.
By making rationality assumptions about agential behaviour, we enable scientific prediction and hence explanation of that behaviour. These rationality conditions are assumptions that are designed for scientific prediction; considered in this way, they have no normative status beyond that, though we will see later that evaluative standards have been built on top of these assumptions. The idea of rational choice or revealed preference theory is to construct preference orderings or utility functions for agents by examining their behaviour in one context and using those orderings to predict their behaviour in another context. Scientific prediction is not the same as forecasting or pragmatic prediction, though it can contribute to such forecasting (Dowding and Miller 2019). However, many forecasting models make no use of revealed preference theory.
Rational choice uses consistency conditions in order to interpret actions. To see how this works in rational choice theory, we need to introduce a few technical definitions. We define the following terms and symbols. We assume there is a finite set X composed of elements a, b, c and so on. These elements can be considered as observable outcomes such as āvoting for the Liberal candidateā or ātaking a bribeā, and we will refer to these as āalternativesā in an opportunity set. We then define a set of āpreference relations ā in terms of three categories:
Weak preference . An individual i weakly prefers x to y when she considers x at least as good as y. We represent this relationship by the symbol ā½. So individual i is represented as weakly preferring x to y by x ā½ iy.
Indifference . An individual i is indifferent between x and y when she weakly prefers x to y and weakly prefers y to x: x ā½ iy and y ā½ ix. This is represented by the symbol ~; thus x ~ iy.
Strict preference. An individual strictly prefers x to y when she weakly prefers x to y and does not weakly prefer y to x: x ā½ iy and Ā¬ (y ā½ ix) (where Ā¬ means ānotā). This relationship is represented as x ā» iy.
The relation ā» is asymmetric, which means that if x ā½iy then necessarily Ā¬ (y ā½ ix). But ~ is symmetric, because x ~ iy implies x ~ iy. ā½ is thus composed of an asymmetric and a symmetric part.
If we suppose that an agent is choosing between three alternativesāsay a voter between three parties or a politician between three policiesāx, y and z, the axioms of rational choice are:
- Reflexivity . For an alternative x, x = x;
- Completeness . For any two alternatives x and y, either x ā½ y or y ā½ x or both;
- Transitivity . If x ā½ y and y ā½ z then x ā½ z;
- Continuity . For any alternative x, we define a set A(x) as the āat least as good asā set of alternatives and B(x) as the āno better than xā set. Sets A and B are āclosedā, meaning they include everything in their boundaries. Then, if x ā» y and z is an alternative close enough to x, z ā» y.
From all this, we can define agentsā preferences over all alternatives when reflexivity, completeness and transitivity hold. When they hold, we can view an agentās behaviour in several choice situations to construct their preferences, which we can then apply to a different choice setting and predict their behaviour. These assumptions give us the basis of revealed preference theory or rational choice. The three assumptions, often called ārationality assumptions ā, are better thought of as predictive assumptions. When they hold, we can scientifically predict and therefore explain behaviour. That is, our interpretation of behaviour in the first setting enables our interpretation in other settings. We consider criticisms of these assumptions in Chapter 3.
Continuity is a condition that shifts us from revealed preference to cardinal utility functions. This enables us not just to calculate the order in which people rank alternatives, but to give a scale showing the size of the differentials. Once we can scale alternatives, we can start the process of welfare calculations by some form of social utility function. We say no more here about cardinal utility functions, as all of the simple models in this book utilize only preferences.
We should also note that preference and utility thus considered are completely empty concepts. They do not represent happiness or satisfaction or other psychological experiences; they are just mathematical formulation to enable scientific prediction. Our interpretation of behaviour might well fill in some of these psychological experiences for biological agents, but the conditions work equally well, if not better, for agents without such psychology, such as firms or political parties.
Simply using ordering is often thought to bring advantages, since we do not have to attempt to make interpersonal comparisons of utility, which are problematic (Binmore 2009; Dowding 2009). This is one important difference between markets and government. Markets operate with a pricing tool which automatically provides cardinal comparisons. When a person purchases private goods, she pays only up to the amount that the good is worth to her. Now, that price cannot be simply equated with that individualās utility and certainly price cannot measure interpersonal utility. The amount of money one has available to spend in any given week does not equate to the amount of utility one can gain in any given weekānot even the amount of utility one gains from market goods. The fact that people have very different levels of income and wealth means that money cannot be used for interpersonal comparisons. Nevertheless, the market operates on cardinal grounds. When we aggregate utility in a public setting, we usually do so by voting, and that only m...