, Research Paper
Rational Choice Theory is perhaps one of the best known methodological approaches to the explanation of individual action. In this essay I hope to outline Rational Choice Theory and discuss certain areas of weakness in this theory, where its explanatory powers arguably breakdown and expand on the theory’s formulation. Individuals are a unit of analysis at which to study society, but it should be remembered that society is not just made up of a large number of individuals, but contains groups and organisations and so any general sociological theory should be able to explain how such social structures come into being and how they are maintained. I shall attempt to evaluate rational choice theory, concluding that it should not ‘retreat to the banal claim that people have reasons for what they do’. Rational choice theory is based on the idea that individuals act for a particular purpose, that is to maximise their utility. In its simplest form rational choice theory states that given a number of options people do what they believe is likely to have the best overall outcome. This concern with outcome is same as Weber’s concept of ‘zweckrational’ action, where an individual performs the action which is most likely to further the individual’s pursuit of a particular goal. Rational choices involve three distinct processes. To illustrate this point I shall use the example of an individual looking for a job. Firstly the individual must collect an optimal amount of information before taking the decision. This would involve visiting the local job centre and looking in that vacancies section of the newspapers to ascertain the kind of employment that is currently available. It is not the case that the more information that an individual gathers, the better of they will be because there are costs associated with the gathering of information. Once the job seeker has received a number of offers of employment and is trying to decide which one to take, he cannot spend too long collecting information, since by the time he comes to his decision, it may be that the best is no longer available, the deadline for accepting the offer may have passed. Time is therefore an important factor in determining the level of information collection that is optimal. When driving a car which has lost its breaks, it is useless for the rational actor to spend time deliberating on his action as there is simply not the time to make a choice, based on the careful analysis of the outcome of each mode of action. Time will play a crucial role in choice situations, especially if the consequences of actions are spread out over time. One choice may yield immediate gains, the alternative a larger but delayed benefit. Therefore to choose between the two, the rational agent must have a justified method of comparing the benefits of alternatives that have different temporal patterns. Depending on the consistency of the time preferences, they can either be rational or irrational. Strotz showed that time preferences that are inconsistent are irrational, and that inconsistency is the result of non-exponential time preferences. Exponential time preference means that the present value of the future decays at a constant rate as the future develops; conversely non-exponential time preferences imply that some parts of the future lose their value more rapidly than others. The time element is particularly important in strategic situations, as it may alter what is considered to be rational. Consider Prisoners Dilemma, if there is a fixed number of moves, the rational behaviour will always be to defect, because on the last move it will be rational to do so, and this continues back to the beginning of the game. However, in ‘repeated’ prisoners dilemma, the individuals interact an indefinite number of times, over an indefinite period of time, it may become rational to attempt to co-operate, in order to build up trust and therefore gain the maximum benefit. The next stage in a rational decision involves the beliefs that are held by the individual, given the information that they have at their disposal. It is not rational for our job seeker to believe that he will earn £300 per week if the highest paid job on offer is only £150 per week. Similarly the belief that he stands a good chance of getting a job for which he is not qualified according to the information that he has collected is not rational. Finally the individual should choose the best means of obtaining his goal, given the beliefs that individual holds. So our job seekers’ goal might be to obtain sufficient money to pay off a debt. If he believes that he can get a job which will achieve this goal and this belief is rational based upon the information which he has collected, then it would be a rational choice of action to get a job, rather than say sell a valuable treasured possession to get the money. It is possible to identify 3 distinct elements in the choice situation. The first is the feasible set or the set of all courses of action available. The second is the set of rational beliefs about the causal structure of the situation, this determines what courses of action will lead to what outcomes. The third is the subjective ranking of the feasible set, derived from ranking the outcomes to which they are expected to lead. The rational action, therefore is to choose the highest ranking element in the feasible set. It is important to remember, that when considering human action, it is necessary to know what to believe with respect to the relevant factual matters, and so the theory of rational choice must be supplemented by a theory of rational belief. While the rationality of an action is ensured by its standing in the right kind of relation to the goals and beliefs of the agent, the rationality of beliefs depends on their having the right kind of relation to the evidence available to him. This is a normative theory which can then be employed in an explanatory function. The question of whether a given action was performed because it was rational, which involves proving that the action arose through the proper kind of connections between desires, beliefs and evidence, therefore arises. It unusual in everyday life to be faced with decisions for which perfect information is obtained. The choice situation must be separated into two main dimensions. The first is the distinction between perfect and imperfect information about the outcomes. For example, when a farmer chooses between two crop varieties, their expected yields depend on the state of the weather next year, which cannot be predicted with certainty. Most decisions are like this, although some come very close to the limiting case of full certainty. Of the incomplete information choice situations, there is further subdivision into those characterised by risk or uncertainty. The difference between the two is that statistical probabilities can be attached to risks, whereas this is impossible for uncertainties. Under conditions of risk it is rational to choose the option that maximises expected utility, where in this situation the expected utility of the action is defined as the weighted average of the utilities yielded under different situations, and the weights being the probabilities of the states. In this situation, the phenomenon of risk aversion occurs. An example of ‘risk aversion’ might be as follows. Consider a farmer who must choose between two grain varieties, a and b, and the size of the yield next year is dependent on the severity of the winter, either mild ( state 1) or harsh ( state 2 ). Assume also the income of the farmer in the various states to be given as follows: Crop a is chosen Crop b is chosen State 1 occurs 30 50 State 2 occurs 25 15 Source Elster (1986) If we stipulate that the two states have equal chance of occurring, then crop b clearly has the highest expected yield. However, this does not mean it has the highest expected utility and thus may not be the crop chosen by a rational agent. If the minimum required yield for survival is 20, then the agent would be foolish to choose the crop type that has a 50 % chance of starving to death. Thus rationality may dictate the choice of the option with the largest expected utility, but this is not necessarily the option with the largest expected income – even when utility is derived form income. Elster states that it is difficult to imagine the existence of choices under uncertainty, because we always have some information, however vague and diffuse, with which we can assess the probabilities of the various outcomes. In many choice situations the probabilities calculated are to some extent subjective, and thus once the probabilities have been given, the principle of maximising utility is the same as before. Under true uncertainty therefore, rationality can give us little normative guidance. The rational choice is still that choice which is perceived by the actor to minimise loss of utility, despite the lack of the role of information in his decision process. I shall expand this notion later on. The other main distinction that applies to the rational choice situation is the difference between parametric and strategic decisions. In the parametric situation the agent faces external constraints, constraints in some sense given or parametric. He must first estimate their nature as best he can and then decide on the rational course of action. A strategic decision on the other hand is characterised by the interdependence of decisions. Each agent has to anticipate what others are likely to do and this may include an estimate of what they anticipate he will do. This results in his decision becoming a part-determinant of the constraints that shape his decision. An example of a strategic decision is the kind made by agents in game theory, most notably ‘Prisoner’s Dilemma’. In order to provide a rational choice explanation of an action, it must satisfy three sets of requirements. Firstly, there are three sets of optimality conditions. The action is the best way for an agent to satisfy his desire, given his belief; the belief is the best he could form, given the evidence; and the amount of evidence itself is optimal, given his desire. Second there are a set of consistency conditions. Both the belief and the desire must be free of internal contradictions, thus the agent must not act on the desire that , in his own opinion, is less weighted than other desires which are the reasons for not performing the action. Finally there are a set of causal conditions. The action must not only be rationalised by the desire and the belief; it must also be caused by them, and moreover, caused in the ‘right way’. In rational choice explanations as actually practised by social scientists, the focus is almost totally on the optimality conditions, because the goal is normally not to explain individual action, but the behaviour of large numbers of people in similar external situations. “Rational choice theory can fail because it does not tell us what rationality requires or because people don’t behave as rationality requires them to do”(Elster 1986). It is argued by many sociological theorists, and as the title of this essay questions, rational choice theory fails to explain irrational action by actors, and in doing so poses serious doubts as to the validity of this theory. There are a number of points to be raised here. The first is that rational choice theory is a sociological theory used as a tool for explanatory analysis of social action and interaction. As with any model, the purpose of the theory is to simplify highly complex phenomena for easier understanding. It must also be as widespread in its range of application as possible and inevitably through this simplification, universal explanatory power is forfeited. Given that an action to be rational must be the final result of three optimal decisions. First it must be the best means of realising a person’s desire, given his beliefs. Next, these beliefs must themselves be optimal, given the evidence available to him. And finally, the person must collect an optimal amount of evidence, neither too much nor too little. That amount depends both upon his desires, on the importance he attaches to the decision, and on his beliefs about the costs and benefits of gathering more information. Source: Elster 1989 pp 31 The above diagram summarises the argument so far. Elster notes that the only independent element to which all others are subservient is ‘desires’. To quote David Hume, “Reason is, and ought only to be, the slave of the passions.” This does not imply that reason should yield to every whim of the passions, indeed, as the blocked arrow on the diagram indicates one’s beliefs should not be determined by desire, as in wishful thinking. On closer inspection it is evident that Hume’s remark implies the notion of rational desires is meaningless. Elster argues that reason cannot dictate to the passions what their objects should be. Although it would be irrational to say, ‘want one’s cake and eat it too’, Elster suggests that desires are not subject to rational assessment. One may dislike a desire, or think it immoral, yet it becomes tenuous to term it irrational. One of the main reasons for the failing of rational choice theory is through indeterminacy. Elster identifies two forms. There may be several actions that are equally and optimally good, such that a decision between two actions is arguably based not on rational choice. The often cited example is that of choosing a can of soup from a shelf in a supermarket. The choice between two identical cans might be classed as indifferent, yet the choice could equally be based upon something as trivial as the proximity of one can to the actor. Another case is that of incommensurability whereby, for given beliefs and desires, there is no optimal action when the actor is able to compare and rank all the options. The actor might therefore arrive at a situation where he is unable to say which of them he prefers, and unable to say that they are equally good. Elster terms this incommensurability. For two options, x and y, I have no preference for either. If I am indifferent between them, I should always prefer x with and extra pound over y. If I do not prefer x plus a pound over y, it shows that x and y are incommensurable. An inability to rank and compare all the options may also arise in certain bizarre cases. If I am told that, if I write down a positive number (strictly larger than zero), I will get a pound reward equal to 1 divided by that number. Clearly, for any number I write down there is a smaller number that would give me a greater reward. Other important decisions often involve incommensurable options. Choosing between Universities may be such an example. Lacking the experience of having experienced both and then being able to make an informed choice between the two, I know too little to make a rational decision. It then becomes necessary to choose the action according to some second variable, say proximity to London, or the politics of the local council. Indeterminacy of beliefs arises when the evidence is insufficient to justify a judgement about the likelihood of various outcomes of action. There are two main ways of this occurring, firstly through uncertainty, and secondly through strategic interaction. Elster uses the example of R and D investment for firms in a winner takes all scenario to illustrate the role of future uncertainty. Bayesian decision theorist would disagree will this factor of total uncertainty. They argue that since we always have some knowledge about the choice situation, it must be better to use some form of subjective probability estimates than none at all. Elster argues however, that although such probabilities may be formulated, they are no basis for rational action. A classic failure of rational choice theory is in instances of accident, or random human error. Pressing the wrong button, or dropping a plate. If this is a genuine mistake it cannot be termed ‘irrational’. Alternatively, a dieting actor may decide to accept the offer of a Mars Bar. This is irrational behaviour, and may be blamed upon weakness of will. Irrational behaviour may also stem from irrational beliefs, beliefs which are influenced by the passions they are supposed to serve. The Freudian ‘Pleasure Principle’, the mind’s tendency to seek immediate gratification, may act without the actors conscious will to lead to irrational action. Elster identifies two forms of irrational belief: ‘Hot’ and ‘Cold’ irrationality. Wishful thinking is an example of hot irrationality, which will make the choices of the actors appear irrational in the long run. Cold irrationality is mainly based on the errors of perceiving statistics, such as drawing conclusions from samples based on small sample sizes. People also have the habit of overestimating the importance of personal experiences and current events at the expense of impersonal sources and past events when forming beliefs. Furthermore at the information collection stage, there is a tendency for people to stop collecting when the beliefs they hold are supported, rather than collecting more information at the risk of having their beliefs being inaccurate. Many actions are performed out of habit, tradition, custom, or duty, and are not in this sense rational, and are instead determined by the notion of ’social norms’. Advocates of the importance of social norms, suggest that individual action is more frequently determined by the norms and values of society, imposing external constraints on the opportunity set of the individual actor. However, the theory that individuals act purely under the influence of norms, either external or internal, cannot replace rational choice theory, as there are many examples of traditional behaviour being discarded when new opportunities become available – not because they are better embodiments of the value inherent in customary behaviour, but because the individual finds that they serve his goals better and it is therefore rational to adopt them. Norms are nevertheless a useful concept for explaining behaviour that does not appear to follow the ‘right’ pathway through desires, beliefs and goals. It might also be possible to argue that adherence to social norms is rational in itself, due to the impositions of sanctions by society, after breaking the norm of action, thereby reducing net utility. Rational choice theory is an attempt to perform what appears to be an almost impossible task, that is to explain the behaviour of individual human beings. Given the size of this task and the uncertainty involved, combined with the relative simplicity of the theory, it is hardly surprising that it is not a perfect method of explanation. It is important to remember to function and aim of a theory or model, that is to simplify complex phenomena and inevitably some accuracy will be lost and over-generalisations made. It is in my opinion that rational choice theory provides a solid basis from which to analyse individual action, and it is unfair to suggests that ’since it is clear that individuals do not always act rationally in any strong sense, rational choice theory must either severely limit its scope of application or retreat to the banal claim that people have reasons for what they do’.