Rational Choice Explanation Research Paper

The most common criticism of rational choice explanation is that it is false to the facts of actual motivations of people. This is a by-inspection claim. Because very much of actual behavior surely is motivated by interests, the dispute must actually be whether rational choice explanation is likely to help understand certain activities but not others. This is a question that cannot be answered in the abstract. Sometimes the criticism of rational choice trades on the fact that we commonly resolve many apparent choices by acting from habit or routine. As A. N. Whitehead quipped, civilization advances by the reduction of more and more choices to habit. A similar criticism of utilitarianism is that, if we spend as much time in calculating costs and benefits of our choices as would seem to be necessary to get our choices right, we will have no time left for life. J. S. Mill (1969) responded, rightly, that a sailor need not work out the positions of the stars anew each trip to sea but can simply rely on an astrolabe.

To illustrate the nature and power of rational choice explanation, as well as its plausible limits, I will address the aggregate-level phenomena of collective or group action, majority rule, and democracy. I will then address game theory and the value theory of rational choice. Finally, I will address the application of rational choice explanation to norms and social order.

1. Fallacies Of Composition

A common criticism of rational choice explanation is that it cannot handle aggregate-level phenomena because it is fundamentally grounded in individual level interests, choices, and actions. On the contrary, the chief successes of such theory have been to relate aggregate-level phenomena to individual-level actions. Several of the most important findings have been the—typically negative—demonstration that various aggregate-level claims are fallacies of composition from individual-level actions, choices, or values. It is a fallacy of composition to suppose that a collection of items has itself the characteristics of those items. This can be true, but it must be shown and not merely assumed. Consider three of these demonstrations.

The logic of collective action suggests that, although individuals may share a general interest in supplying themselves some collective benefit, they individually may have no interest in contributing effort or resources to the project. My contribution will buy a small part of our good, of which an even smaller part will go to me, so that my cost of contribution will be larger than my benefit from it. All members of a group may therefore be individually worse off because they failed to take some action (Olson 1965, Hardin 1982). This is a problem that is pervasive in societies, ranging from very small societies with limited resources to very large and complex industrial societies.

The Arrow theorem (Arrow 1963, Barry and Hardin 1982) concludes that a collective utility or welfare function cannot be aggregated from individual utility functions if both are ordinal (see discussion of ordinal utility theory below). In essence, the theorem says that there cannot be a collective preference that has the simple features of an individual preference function. Among the implications of this theorem is that every standard choice procedure is potentially incoherent. Majority rule, for example, can produce cyclic preferences such that candidate A would defeat B by a majority, B would defeat C, and C would defeat A. Hence, simple majority rule is potentially indeterminate. It often works well enough—or seems to work—in practice, however, because there might not be cyclic preferences or because there are devices that block such preferences from being effective and even from being revealed. For example, plurality elections or the use of run-offs between the top two candidates in a multicandidate election can hide any cyclicities because there cannot be cycles over only two candidates.

The economic theory of democracy explains why candidates might not be substantially differentiated in certain common contexts and why voters often have little interest in voting. It also makes sense of the otherwise surprising level of ignorance of typical voters. In this latter case, however, there may be no fallacy of composition, because individual-level ignorance often translates into collective ignorance. Consider the problem of voting, as analyzed by Anthony Downs (1957). The chance that my vote would make a difference in an election decided by a large electorate is very nearly nil. In a 1974 statewide contest for senator in the very small state of New Hampshire, with somewhat more voters than Iceland, the vote was so close that errors in counting made it practicably undecidable who had won. No-one’s single vote mattered. Suppose that the trouble to me of voting is such that I would sooner spend several dollars to vote by telephone to avoid going to the polls. The value to me of electing one candidate rather than another would have to be in the many thousands of dollars to offset my costs of voting, because the benefit would have to be discounted by the very low likelihood that my vote would make a difference. Yet many people do vote. Much behavior seems not to fit the rational choice theory and must be explained in other ways.

If it would not be worth my very slight effort to vote, Joseph Schumpeter (1950) concludes, then it would be even less worth the much more substantial effort it would take for me to be well informed about the candidates and issues at stake. Evidence from many national contexts suggests, in fact, that citizens are abysmally ignorant when they vote. Here rational choice explanations seem to fit well. Editorial campaigns against voter ignorance are often misframed as being about the irrationality of citizens, who are acting all too rationally.

2. Game Theory

Rational choice explanations are commonly based in game theory, neoclassical economics, or commonsense arguments such as those of Smith and Hume. Game theory is perhaps the framework that best characterizes the nature of most of the problems in rational choice theory, although many of the most important insights in such theory predate game theory, which originated roughly during World War II (Von Neumann and Morgenstern 1953). Although it is called a theory, game theory is little more than a descriptive framework on which many theories about particular or quite general phenomena are founded. That framework takes many forms, the most transparent of which is the matrix form in which each of the players or participants faces a set of possible strategy choices. Outcomes of a play of a game are jointly determined by the strategy choices of all the players.

This characterization of our individual choices—of strategies, not of outcomes—makes game theory virtually antideontological. Judging action types as right or wrong makes no sense in interactive contexts, which are the principal contexts for most of moral theory. An action theory that takes what are essentially individual choices of outcomes as central is almost irrelevant to choice in social contexts in which, typically, important outcomes necessarily are not under the control of individual deciders (Hardin 1988, pp. 68–70). An ethic of rules makes nonsense of choices of strategies, which are chosen because of their effect on outcomes, as these are influenced by the strategy choices of others. Game theory can readily be fitted to consequentialism and especially utilitarianism. Utilitarianism essentially generalizes the principle of rational choice to choosing the welfare of all instead of only the welfare of the individual chooser.

To fit utilitarianism to game theory is to revamp its central concern with outcomes as valued by their benefits and costs. In interactive social choice, we typically cannot say that a particular action is right or wrong in producing a good overall outcome. We must, rather, judge strategy choices and we must judge them contingently. A strategy choice is right only to the extent that, in interaction with the choices of others, it produces a good overall outcome. If the others with whom I am interacting are sometimes utilitarian, sometimes self-seeking, and sometimes rule-guided in their choices, I may have difficulty determining my best strategy choice for producing a utilitarian outcome. This is not different from the difficulty that self-interested players have in determining the best strategy choices for themselves. Game theory is indeterminate in many contexts, and therefore both pragmatic and utilitarian choices in a real-world setting are apt to be indeterminate as well. Merely seeing the nature of interactive choice in descriptive game theory changes our understanding of the possibilities of rational choice theory and of utilitarianism.

3. Value Theory: Interests, Consumptions, And Welfare

Rational choice accounts are commonly grounded in interests. Sometimes, interests are the whole story of someone’s motivations in a particular context. But typically, my interest in having more resources, such as money, is itself grounded in my desire to consume various things. Resources, capabilities, and opportunities matter to us only because they enable us to consume or experience various things, and consuming or experiencing these things is what constitutes our welfare. Resources are worthless otherwise. Focusing only on interests is therefore a shorthand for what the resources can enable us to do. Moreover, we could relate our consumptions to the overall welfare that they bring us. It would be a mistake to suppose interests translate smoothly into welfare or consumptions. Indeed, consumptions typically trade off against each other and against interests. If I invest much of my resources in a holiday, I may not have enough left for other things, such as a new refrigerator. Rational choice explanations involve all three of these categories—interests, consumptions, and welfare.

It would also be a mistake to suppose that my welfare is wholly self-interested. One of the things I might wish to consume or experience might be your happiness. One may say this is against my interests because it means using resources that I could use for my own happiness in some other, perhaps more direct, way. But this is a confusion in at least two ways.

First, the only point in having resources is to be able to consume in order to contribute to one’s welfare, as is evident to everyone but a Scrooge for the resource of money. True, there is a conflict between using my resources for one consumption and saving them for another, but this is properly seen not as a conflict between consumptions and interests but as a conflict between different consumptions, or between different paths to welfare.

Second, there is no reason to suppose that, say, enjoying an opera, a good meal, or a lovely journey is a more direct or more successful route to enhancing my welfare with some bit of my resources than providing for my son’s or lover’s, or friend’s happiness with that same bit of resources. More generally, my welfare is enhanced best by a mixture of consumptions and experiences, including contributions to the welfare of others, especially certain others. With those certain others, my welfare will depend on our sharing intentions to do various things for and with each other.

If we could suppose that economic interests as measured in resources were the primary issue for people, we could fairly easily explain actions as though people were all alike except for their circumstances. Rational choice explanations typically reduce the focus to interests, rather than other demands, such as demands for moral positions (ranging from deontological to utilitarian). In part they do so for methodological reasons. Analysis is more productive if we keep it clean and focused, and simplification is required for virtually all social analyses. But in part, the focus on interests is often basically right, especially for aggregate or collective issues. Other concerns matter to collective groups and societies, but they matter the way they do in life more generally. They often do not conflict with interests although, of course, occasionally they do and they may even trump interests in many contexts. For example, voters may vote against their personal interests in support of welfare programs and policies favoring ill-represented ethnic groups. But those who voted against the British Raj in India, French control of Algeria, or the US involvement in Vietnam may generally have voted with their interests as well as with moral concern. Still, many social and individual phenomena cannot readily be explained as the product of interest-seeking by individuals.

4. Value Theory: Cardinal, Ordinal, Ramsey–Neumann Cardinal

Apart from the structure or content of the value theory discussed above, there has been extensive debate over the form of such theory. Jeremy Bentham supposed that pleasures and pains during a day of someone’s life could be defined in such a way that we could add the value of pleasures and subtract the value of pains for an overall measure of benefit to the person for that day. Today, we would speak of such a welfare or utility measure as cardinal in the sense that each experience one has could be assigned a utility number that could be added straightforwardly to other utilities.

Bentham’s terminology was confused in that he spoke of the utility one gets from consuming some object and also of the utility inherent in the object. The latter notion underlies the labor theory of value that corrupted much of economic thinking throughout the nineteenth century and often still corrupts it today, in part under the doleful influence—on this matter—of Karl Marx. Bentham contributed to two opposing streams of value theory: that of the labor theory of value and that of the subjective theory that holds value to be strictly what a person gets out of an experience or consumption. The value to me of consuming some- thing such as, say, beluga caviar might be far below the value to me of consuming other things that take equal expenditures of my resources. The value to you, even with the same overall level of resources, might be radically higher. Hence, you might consume a lot of beluga caviar and I might consume none after an initial taste. The subjective theory obviously has the disadvantage that we cannot always generalize from one person to others, although often we can generalize enough not to need to do in-depth studies of the tastes and preferences of everyone in order to explain overall behavior.

Bentham supposed further that the utility for one person could be added to that for another. This supposition would allow us to add up the utilities to all under one set of circumstances and to compare the total to that under another set of circumstances. Supposing we could do the measures, we could then say that, for example, one form of government is superior to another.

The so-called ordinal revolution in demand theory in economics overthrew all of the cardinal assumptions and set in their place a utility theory that is entirely subjective and that cannot be given interpersonal comparisons. The remarkable fact for economics is that such a limited theory clears up many prior problems and makes coherent sense of demand and supply relationships and, hence, prices in market exchange. Unfortunately, this revolution came only in the 1930s when moral theorists had ceased also to be economic theorists, as Smith, Hume, Mill, and Henry Sidgwick had been. It has therefore not had as much influence in general social theory as in economics, although it clearly should be brought to bear on the former as much as on the latter. A clear implication of interpersonally noncomparable value theory is that many traditional questions of interaction between people cannot be readily settled, so that moral and social theory must often be open on many issues.

In what might count as a minor sideline, Frank Ramsey (1931) and John von Neumann (Von Neumann and Morgenstern 1953) independently proposed a way to cardinalize preferences that are merely ordinal. Suppose you face a list of mutually exclusive outcomes, and suppose you can rank order any two outcomes. Assign a value of 0 to the least preferred, Z, and a value of 100 to the most preferred, A, outcomes in your list. For some outcome, K, between these two in your ranking, you can rank it against any lottery over A and Z. For example, suppose you would be indifferent to getting K and getting the combination of a p percent chance of getting A and a 100–p percent chance of getting Z (0 < p < 100). On the scale of 0 to 100, K has a cardinal value to you of p.

The Ramsey–Von Neumann cardinalization of utility (RN) clearly applies to personal decisions over outcomes. But we cannot use it to construct a Benthamite cardinal utility that is additive across persons, because there is no sense in which your 0 and my 0 are the same or your scale and mine, because each of them is arbitrary. If I assign a value of p and you assign a value of q to outcome K, we cannot meaningfully say that K has a joint value of p q. Indeed, we can perform a linear transformation on your utility scale without affecting the results that follow from it. For example, we could divide the value of each outcome by 10 and add 37 to it, so that your scale would then run from 37 to 47, and this transformed scale would get all the results that your original scale from 0 to 100 would get. In addition, we can assign utility values to outcomes that fall outside the original ranking from A to Z, so that some outcomes might have negative utility and others might have extremely large positive values. This is suggested by the fact that the original normalization was performed over the arbitrary set of outcomes from A to Z. Hence, there is no meaning to any claim that your RN utility can be added to mine to yield an overall joint utility. The RN theory is therefore primarily of interest in individual decision making under risk, and is seldom of interest for more general social theory. It is irrelevant to Benthamite hopes for a cardinally additive utilitarianism.

5. Norms

There is a growing literature on the rational choice explication of norms. Edna Ullmann-Margalit (1977) gave rational reconstructions of norms for dealing with strategic contexts that are modeled by prisoner’s dilemma, coordination, and partiality or unequal coordination games. In this categorization the descriptive apparatus of game theory is used to establish and characterize types of norms according to the strategic problems they resolve. The first of these norms essentially handle collective action problems because, without the force of the norms, many individuals would free ride, and any effort to provide a collective good would fail. The second handle such coordinations as the driving convention. That convention surely requires no norm to keep drivers in compliance with it, because the threat of grievous harm from violating the convention is sufficient. But many conventions are less regulated by such clear incentives and some of these might be more forceful if backed by norms. Norms in Ullmann-Margalit’s third category handle unequal coordinations such as the division of labor by gender roles or social order in any hierarchical society in which obedience to norms might keep even those at the bottom of the hierarchy loyal and orderly. Again, these conventions are grounded in coordination incentives and they might be fully self-enforcing, but norms that moralize the inequalities might make those in the less well off positions less sullen in acquiescing in the arrangements.

More generally we may say that there are two broad categories of norms: norms of exclusion and universalistic norms (Hardin 1995, Chaps. 4 and 5). Norms of exclusion are specific to particular groups, which can enforce the norms by sanctioning violators within the group; the sanctions can go so far as to shun and even exclude members. The sanctioning is grounded in the interests of individual sanctioners, so that the norms are commonly interest based, although group members might claim they are inherently moral or right. A norm of exclusion can work to help enforce cooperativeness in a group or community, or to enforce a coordination that might serve the interests of some but not all. For example, a minority culture might enforce upon its members the use of a minority language or a particular dress code. Using that language or following that dress code might reduce the possibilities for some members of the community to gain employment outside the community. The community might protect its hegemony over deviant members by shunning those who violate the language or dress code.

Universalistic norms can be dyadic, as in promising or truth telling, or large-number in the sense of applying to actions toward the whole society, as in not littering. Dyadic norms are commonly enforced through the sanctions inherent in the iterated prisoner’s dilemma. If you lie to me or break your promise to me, I can simply avoid further interaction with you, to your likely cost. If you break a norm that affects everyone, such as a norm against littering or in favor of voting, no particular individual is likely to have an interest in sanctioning you. Hence, such norms are likely to be relatively weak in their force and in our adherence to them. Oddly, therefore, it is the often odious norms of exclusion, such as racist and ethnic norms, that have strongest force, while universalistic norms are relatively weak unless they are played out essentially in dyadic interactions. Those who insist that, morally, norms should commonly trump interests, should be bothered that it is interests that trump any plausible moral content of many norms, and that norms not backed by interests are likely to be weak. Large-number universalistic norms may be backed primarily by moral or psychologically inculcated commitments. If so, then it is of interest whether they are, in fact, very effective in motivating people.

6. Social Order

There are three grand schools of social theory. One, associated with the French sociologist Emile Durkheim and many others, holds that society is held together by shared values. The second, associated with Karl Marx and Plato’s Thrasymachus, holds that conflict is pervasive and that social order must therefore be maintained through the imposition of power to coerce when necessary. The third, associated with Adam Smith, holds that society is organized by and for exchange, which involves a combination of coordination and conflict. Each of these schools has many strong contemporary advocates. A fourth plausible view is that much of social life is simply organized by coordination (Hardin 1999, Chap. 1). We need not share particular values in order to share interests in avoiding conflicts with each other. The ideal type of this last form of social order is coordination on traffic rules, especially on the rule to drive right (or to drive left). We can be hostile to each other in many ways and still find it sensible to avoid needless conflict over many things, so that we would willingly be coordinated. A shared value theory is also a coordination theory, but its advocates ground it in value commitments and are generally hostile to the inclusion of interests among those value commitments. The conflict and exchange schools are commonly grounded in interests, although the socially destructive conflict that sparked much of early liberal thought was over commitment to divergent religious views.

The pre-eminent coordination theorists among classical political philosophers were Thomas Hobbes and David Hume. They were both protogame theorists in their grasp of the fundamental nature of various social interactions. While Hobbes is commonly thought to be a major conflict theorist, his actual resolution of the problem of social order is to arrange for coordination on an all-powerful sovereign, whose task is essentially to coordinate us on order so that we may construct our own separate lives and engage with one another in productive pursuits (Hobbes 1968). Hume’s (1978, Book 3) articulation of a remarkable array of coordination problems qualifies him to be seen as the greatest coordination theorist, although his complex insights were largely neglected for two centuries, in part because he lacked the technical vocabulary of game theory and did not himself bother to work out the general principles of it. He did, however, work out a general moral and political theory that was substantially grounded in strategic understanding (Hardin 1988, Chap. 2).

In complex modern societies, coordination may be the central, dominant mode of social order. It does not require consensus on values—pluralists can coordinate on an order for the society in which they seek their diverse values. It does perhaps fit especially well with modern individualist liberalism that allows individuals to live by their own lights. Indeed, coordination is a simple version of Mill’s principle that we should be free to pursue our own ends without interference so long as our doing so does not interfere with others (Mill 1977). When there is such interference, of course, other principles may be required to regulate our interactions.

7. Concluding Remarks

Rational choice theory is a diverse set of approaches to the study of society that are based in assumptions of individual rationality. Indeterminacies in such theory often mirror indeterminacies in social relations and individual understandings of these. Indeed, many rational choice explanations have demonstrated the indeterminacy of social choice. Among the most important contributions of such theorizing is the descriptive clarity it often adds to our understanding of issues that have been poorly characterized before. Rational choice explanation is inherently about states of affairs, although it can include in a state of affairs concerns for anything that motivates people. It is therefore easy to see how it fits with utilitarianism.

Rational choice value theories, when introduced into the study of law, have led to the burgeoning movement of law and economics, which has produced the most systematic analysis of the broad sweep of law and legal rules that legal scholarship and practice have hitherto seen. The quick march of that movement— roughly beginning with the Coase theorem (Coase 1988)—through virtually all areas of law has provoked counterattacks, but none of these is itself grounded in a value theory systematic enough to permit more than piecemeal objections. The Coase theorem says, with audacious generality and simplicity, that if there are no transaction costs to impede economic cooperation between owners of various resources, then ownership will not affect production, although it presumably will affect how the profits from the production are divided. That theorem, along with earlier work of Ronald Coase and many others, has led to the massive study of transaction costs.

We have a rich understanding of the problems of rational choice value theories just because those theories have been a major focus in the development and articulation of the entire discipline of economics. The theories are a product of centuries of determined and often brilliant debate with dozens of major contributions. Alternative value theories have had far too few advocates and critics to yield much understanding at all. Indeed, value claims commonly seem to be ad hoc and not systematically generalizable. Such considerations are no argument for the superiority or rightness of any value theory. But they may be a tonic for those who wonder how the only well-articulated class of value theories is so easy to criticize: It has enough content to be subject to extensive and varied criticism.

Bibliography:

1. Arrow K J 1963 Social Choice and Individual Values, 2nd edn. Wiley, New York
2. Barry B, Hardin R (eds.) 1982 Rational Man and Irrational Society? Sage, Beverly Hills, CA
3. Coase R H 1988 The problem of social cost. In: Coase R H The Firm, the Market and the Law. University of Chicago Press, Chicago, pp. 95–156
4. Downs A 1957 An Economic Theory of Democracy. Harper, New York
5. Hardin R 1982 Collective Action. Johns Hopkins University Press, Baltimore, MD
6. Hardin R 1988 Morality within the Limits of Reason. University of Chicago Press, Chicago
7. Hardin R 1995 One for All: The Logic of Group Conflict. Princeton University Press, Princeton, NJ
8. Hardin R 1999 Liberalism, Constitutionalism, and Democracy. Oxford University Press, Oxford, UK
9. Hobbes T 1968 Leviathan. Penguin, London
10. Hume D 1978 A Treatise of Human Nature, 2nd edn. Oxford University Press, Oxford, UK
11. Mill J S 1977 On Liberty. In: Robson J M (ed.) Collected Works of John Stuart Mill. University of Toronto Press, Toronto, Vol. 18, pp. 213–310
12. Mill J S 1969 Utilitarianism. In: Robson J M (ed.) Collected Works of John Stuart Mill. University of Toronto Press, Toronto, Vol. 10, pp. 203–59
13. Olson M Jr 1965 The Logic of Collective Action. Harvard University Press, Cambridge, MA
14. Ramsey F P 1931 Truth and probability. In: Braithwaite R B (ed.) The Foundations of Mathematics and Other Logical Essays by F. P. Ramsey. Harcourt Brace, New York
15. Schumpeter J A 1950 Capitalism, Socialism and Democracy, 3rd edn. Harper, New York
16. Ullmann-Margalit E 1977 The Emergence of Norms. Clarendon, Oxford, UK
17. Von Neumann J, Morgenstern O 1953 Theory of Games and Economic Behavior, 3rd edn. Princeton University Press, Princeton, NJ