Essays on Philosophy, Politics & Economics

Integration & Common Research Projects

Gerald Gaus, Christi Favor, Julian Lamont, Gerald Gaus, Christi Favor, Julian Lamont

• Publisher: Stanford Economics and Finance
• Year: 2010
• ISBN:
  9780804774642
• Edition: 1
• Topic: Volkswirtschaftslehre
• Subtopic: Wirtschaftstheorie

This volume brings together distinguished philosophers with interdisciplinary expertise to show how the resources of philosophy can be employed in the tasks of evaluating economics and fostering policy debates. Contributors offer analyses of basic ideas in economics, such as the notion of efficiency, "economic man", incentives, self-interest, and utility maximization. They discuss key concepts in political theory such as desert, compensation, autonomy, equality, consent or fairness. The book then offers examples of how philosophical resources can be applied to specific, timely debates, such as discrimination, affirmative action, and ethical considerations in Social Security. These applications demonstrate how philosophy, politics, and economics can be fruitfully combined, while the more theoretical chapters clarify fundamental relationships across these related disciplines. Ultimately, the text guides students and scholars in expanding their perspectives as they approach the necessarily complex research questions of today and tomorrow.

PART I

THE FOUNDATIONS: RATIONALITY AND HUMAN VALUES

CHAPTER 1

UTILITY AND UTILITARIANISM

Edward F. McClennen

1. INTRODUCTION

The modern concept of utility makes its first appearance in the work of the mathematician D. Bernoulli, where it is employed to “save” a particular decision rule from an embarrassment.¹ The rule in question calls for choosing a gamble that maximizes expected monetary return. The embarrassment arises in connection with a particular gamble, in which the gambler flips a coin repeatedly until heads comes up, and the person who has purchased the gamble receives $2ⁿ, where n is the number of the trial on which heads first occurs. This gamble has infinite expected monetary return and, according to the rule in question, is a good purchase at any (finite) offering price—no matter how large. This is contrary to an intuitive understanding of what counts as a rational choice. Bernoulli’s suggestion was to salvage the expectation rule by taking expectation with respect to the utility (value) of the monetary return instead of monetary return itself and by assuming diminishing marginal utility for money.

The link established between a concept of utility, taken as the measure of the value of something, and the salvaging of a simple “sum-ranking” rule for evaluating alternatives (namely, taking the sum of the probabilistically discounted utilities), prefigures the close connection found today between the concept of utility as developed in the work of mathematically oriented economists and decision theorists and different “sum-ranking” approaches to the evaluation of both personal choices and public policies.² The modern theory of utility includes a construction, which ensures that the value of any gamble is simply its expected utility, and a natural extension of this construction supports a utilitarian sum-ranking approach to the evaluation of policies. By a careful definition of functions of individual utilities, sets of which define the components on which a social aggregation rule operates, we can hold on to a utilitarian-type sum-ranking rule and also defend it against many standard objections. We can defend a utilitarian-type evaluation procedure from the charge that it is insensitive to deprivation and to inequalities between persons.³

To give a historical perspective, utilitarianism came into prominence in the eighteenth and nineteenth centuries, especially in England. As aggressively formulated by Jeremy Bentham, the perception was that it was a needed antidote to a variety of philosophical traditions that allegedly served merely to rationalize different types of vested interests, by trapping them out in the guise of natural laws or eternal truths. Bentham’s attack on these vested interests, like the attack that Marxists would subsequently level, had the great disadvantage that it could be used against itself. Utilitarianism, as presented by Bentham, was subject to the objection that it, no less than the doctrines it was designed to replace, spoke to a set of interests (even if not yet vested): the interests of the greater number, the majority. Leaving the problem of ideology to one side, there is still a serious problem with Bentham’s argument. He argued that utilitarianism is the only serious alternative to either falling into the anarchy of permitting everyone to judge the issues of public policy from the perspective of personal interests and/or moral commitments or the despotism of supposing that the interests of some one person or group of persons should be regulative for all. This argument proceeds by way of a faulty disjunctive syllogism. If utilitarianism qualifies as an objective, impartial perspective, then there exist other positions as well that are deserving of a hearing.

Within the context of the intellectual development of the second third of the twentieth century, all of this became moot. In general terms, the theory of knowledge that came into vogue, as expressed in the program of the logical positivists for example, rejected any claim to objectivity on behalf of any normative principle. Normative principles could not be objectively valid, since they are neither certifiable by reference to logic or meaning alone (cannot be shown to be true in any possible world), nor are they capable of empirical confirmation (cannot be shown to be true in this world). Such principles are not the bearers of truth-values at all, but simply the objects of an emotional commitment on the part of individuals.⁴ This is a doctrine that, it will not escape the observant reader, provides an underpinning for an ideological critique much more sweeping than that envisioned by either the Benthamites or, subsequently, the Marxists.

Even if normative theories in general had not fallen into disrepute, utilitarianism still faced a serious problem concerning the intelligibility, or methodological propriety, of making the requisite interpersonal comparisons of utility. A principle that requires us to choose a policy maximizing the net sum of utilities as distributed to different individuals clearly presupposes that we can meaningfully compare these individual utilities. Within economic theory, however, the trend during this same period toward both an “ordinalist” interpretation of utility and skepticism with regard to interpersonal comparisons of utility served effectively to block any proposal to use the classical version of utilitarianism as a fundamental norm for social policy.⁵ The connection between the ordinalist interpretation of utility and the problem of interpersonal comparisons is more complicated than this brief remark might suggest. A person can embrace a “cardinal” conception of utility and still deny the meaningfulness of interpersonal comparisons; or, alternatively, can accept an ordinalist perspective and still insist on the possibility of “level” comparisons from one person to another.⁶

The last few decades have brought significant winds of change with regard to the methodological issues noted above. The theory of knowledge that is now in favor is not in principle as hostile to the suggestion that there might be an objective (or rational) approach to policy evaluation.⁷ More recently, the problem of interpersonal comparisons of utility has come to be seen as no longer so intractable.⁸ Not surprisingly, the last few decades have witnessed a considerable revival of interest in a utilitarian approach to policy evaluation, a revival that has been encouraged in no small part by the work of economists and philosophers.⁹

What we find in the more formal literature is a variety of ways to underpin utilitarianism as a fundamental perspective on social policy. Some of the arguments proceed axiomatically by way of establishing conditions on what is to count as an adequate social welfare function. Alternatively, we can try to make the perspective of individual choice pivotal and argue that utilitarianism is just the principle to which we should be rationally committed, given that we must make a choice between principles under conditions of substantial uncertainty as to our prospective position in society.¹⁰

These arguments (unlike the one put forward by Bernoulli with regard to gambling) do not simply beg the question in favor of a simple sum-ranking rule. The power of these contemporary constructions resides precisely in the consideration that they start at least some distance from such a presupposition. The point of these exercises is to provide a formal reconstruction of a simple sum-ranking principle, and one that is antecedently taken by many to constitute the most appropriate way to approach both individual decision-making and the evaluation of alternative public policies. But the sum-ranking rule comes in as a theorem, instead of an axiom.

Despite these interesting developments, there is still a substantial issue that arises in connection with utilitarianism. As my opening remarks suggested, it’s the issue of the sum-ranking feature of the utilitarian principle: the presupposition that one collection of individual utilities is as good as another if and only if it has at least as large a sum total of utility (or, alternatively, as large an average utility). Many have questioned this feature on the grounds that it yields an approach to public policy that is not sufficiently sensitive to issues both of deprivation in well-being for some, and inequality. The drift of much of the formal work done in recent decades has been to establish that we can, by one or another device, incorporate such concerns into a sum-ranking approach. I have my doubts as to whether the way in which such concerns are incorporated serves to resolve these issues. But to give this matter even a relatively cursory hearing would double the length of this paper and deflect attention from the equally pressing problem which is my focus.¹¹

The focus, then, will be on this question: Granting that there exist valid axiomatic constructions that can exhibit some version or other of the utilitarian principle as a theorem, what can we say by way of defense of the axioms upon which these constructions rest? This is a matter of no small moment. In normative theory, no less than in theories regarding natural phenomena, we are concerned with the soundness and not merely the validity of the arguments we employ. But if we consult the enormous body of work that has been published in the last fifty or so years since the initial publication of John von Neumann and Oskar Morgenstern’s Theory of Games and Economic Behavior in 1944, we find that little attention has been devoted to the evidentiary status of the axioms employed. In what is to follow, then, I want to review some of the leading axiomatic theories from this perspective.

2. UTILITARIANISM AS A THEOREM

Some of the more recent axiomatic approaches to utilitarianism are connected with Kenneth Arrow’s striking theorem concerning the impossibility of a rational principle for the ordering of social alternatives. Arrow’s theorem turns critically on a presupposition—buried in a very strong ordering assumption that he introduces, the independence of irrelevant alternatives axiom (IIA)—that the preferences of individuals cannot be numerically represented in a way that makes interpersonal comparisons meaningful. Abstracting from the extraneous feature that rules out interpersonal comparisons, IIA simply requires that the ordering of any two social alternatives is independent of whatever other alternatives happen to be available. It requires that the ordering of any two alternatives depends only on the features of those alternatives. As such, it forms a component of what has come to be known as the weak-ordering requirement.¹² As many constructions have demonstrated, with an appropriately reformulated version of IIA (and some others of Arrow’s axioms) and an explicit allowance for different kinds of interpersonal comparability, it is possible to escape the impossibility result. By such means we can defend, as a matter of fact, not only a utilitarian aggregation principle, but also, by a slight modification of the axioms, Rawls’s rival lexicographic maximin principle.¹³

Within this more relaxed framework, which permits interpersonal comparisons, the results that can be obtained all involve the appeal to a distinct axiom known as the strong independence axiom. Stated abstractly, as a requirement on the value of any combination of two items, strong independence requires that if A is ranked at least as good as B, then the combination of A and some third item, C, must be at least as good as the combination of B and C. Just where independence...