See instructions below.....

CAN UTILITARIANISM BE DISTRIBUTIVE? MAXIMIZATION AND DISTRIBUTION AS CRITERIA

IN MANAGERIAL DECISIONS

Robert Audi

Abstract: Utilitarianism is commonly defined in very different ways, sometimes in a single text. There is wide agreement that it mandates maximizing some kind of good, but many formulations also require a pattern of distribution. The most common of these take utilitarianism to characterize right acts as those that achieve "the greatest good for the greatest number." This paper shows important ambiguities in this formulation and contrasts it (on any plausible interpretation of it) with the kinds of utilitarian views actually defended by major proponents of utilitarianism. The aim is not to defend any of these views but to formulate them in a way that facilitates using them—or, more likely, some revised version suggested by the paper—in guiding decisions in business. The analysis provided here should also facilitate appraisal of utilitarianism, contribute to clarity in discussions of business ethics, and suggest a range of ethical standards that merit consideration for certain kinds of decision. If the results of the analysis are correct, a distributive reading of utilitarianism is at best misleading as a representation of its central thrust; it should not be described as the view that ethics calls for achieving "the greatest good forthe greatest number"; and, understood as its major proponents take it, utilitarianism differs more from Kantian ethics than distributive readings imply and is more difficult to defend than it appears to be when viewed as intrinsically distributive.

Utilitarianism is a major position in ethics and is treated at least implicitly invirtually every course in business ethics. If it is not the ethical position fa- vored by most (non-relativistic) social scientists; it is certainly widely held among them. The same may apply to the management field in particular, and some of our examples will indicate how utilitarian considerations can bear on managerial deci- sions. Utilitarianism also figures in management theory, for instance as providing a rationale—though not the only one—for a stakeholder view of the corporation.' A very common characterization of utilitarianism represents it as the view that ethically right acts are those that produce the greatest good (or greatest happiness) for the greatest number. Call this the distributive formulation.^ My chief aim is to examine this formulation—both as interesting in itself and as an interpretation of utilitarianism with important implications for decisions in business and elsewhere.

© 2007. Business Ethics Quarterly, Volume 17, Issue 4. ISSN 1052-150X. pp. 593-611

594 BUSINESS E T H I C S QUARTERLY

The formulation and its relation to what might be called standard formulations have not been examined as they will be here. Indeed, they are rarely examined in depth at all, if we may judge by the representative sources cited below. The result of this examination should be clarifying for many who are interested in utilitarianism and will also suggest some ethical principles meriting further scrutiny and some empirical hypotheses having significant interest for both management theory and business decisions.

/. Utilitarianism as Formulated by its Founders

It will help in achieving clarity concerning the various formulations of utilitarian- ism if we begin with classical sources. As Bentham introduces utilitarianism,

[b]y the principle of utility is meant that principle which approves or disap- proves of every action whatsoever, according to the tendency which it appears to have to augment or diminish the happiness of the party whose interest is in question.-*

He understood interest aggregatively: "The interest of the community" is "the sum of the interests of the several members who compose it."" Despite this aggregative definition of what is roughly the target of ethical conduct, Bentham is probably the main source of the distributive formulation. He says, "Of legislation the proper end may, it is hoped, be stated as being .. . the creation and preservation of the greatest happiness of the greatest number."^ He does not, however, present this as equivalent to his central principle, the principle of utility, nor resolve ambiguities in the former that will be brought out in this paper. He is, however, explicit regarding his main principle: in deciding what act is best, we "sum up all the values of all the pleasures on the one side, and those of all the pains on the other [and] take an account of the number of persons whose interests appear to be concerned.. . . Take the balance; which, if on the side oi pleasure, will give the general good tendency of the act" (Bentham 1961: 175). Here the number of persons matters because everyone's pleasure and pain is part of the relevant aggregate. There is, however, no requirement to maximize the number affected. On his view, one act is ethically preferable to another by producing a greater proportion of pleasure to pain, even if the second act produces almost as favorable a proportion and does some good for a larger number of people.

Bentham's standard is essentially like maximization formulations famihar in contemporary economics and social theory: the standard takes an act to be right if and only if it maximizes aggregate utility for the population in question, those "concerned," in Bentham's terminology. For him, the population in question is appar- ently those persons who would be affected by the alternatives under consideration, and for our purposes this understanding may be assumed.®

This same point holds for Mill (who is similarly vague regarding the composition of the population in question). Consider his formulation of utilitarianism, widely

CAN UTILITARIANISM B E DISTRIBUTIVE? 595

considered even more authoritative than Bentham's (despite greater vagueness regarding the population relevant for determining utility):

The creed which accepts as the foundation of morals "utility" or the "greatest happiness principle" holds that actions are right in proportion as they tend to promote happiness, wrong as they tend to produce the reverse of happiness. By happiness is intended pleasure, and the absence of pain; by unhappiness, pain, and the privation of pleasure.'

A leading twentieth-century utilitarian, G. E. Moore, is more explicit: "the assertion 'I am morally bound to perform this action' is identical with the assertion 'This action will produce the greatest possible amount of good in the Universe.'"* More recent defenders of utilitarianism also take aggregate good to be their fundamental criterion for rightness, whether they take it to apply directly to individual acts or, as in the case of rule-utilitarianism, only to rules.'

One way to see why utilitarianism must be understood as centered on a maxi- mization principle is to note that it is a "consequentialism": it judges the moral status of an act by the overall value of its consequences, as opposed to its intrinsic nature, such as being a lie or, more abstractly, being virtuous or befitting the dignity of persons. In the very same paragraph in which Mill introduced the principle of utility, he spoke of what might be considered a prior principle (valuational though not moral): of the (hedonistic) "theory of life on which this [utilitarian] theory of morality is founded—namely, that pleasure, and freedom from pain, are the only things desirable as ends.'"" To be sure, Moore was more pluralistic than Mill; but the quoted passage indicates that he is clearly a maximizing theorist.

Maximization is also called for by desire-satisfaction views of utility." The intuitive idea underlying all of these consequentialist views is roughly that what we ought to do is produce as much good (or utility) as we can. How that bears on distribution will be considered in detail shortly, but in preparation for that analysis it should be useful first to connect utilitarianism with a decision strategy familiar in business.

//. Utilitarianism and Cost-Benefit Analysis

Structurally, utilitarianism is isomorphic with cost-benefit analysis—a point that may partly account for why it appeals to some and is odious to others. To see the isomorphism, recall that cost-benefit analysis is aggregative; decision is not constrained, for instance, by who benefits except insofar as targeting some people or categories of persons produces a better overall result. Imagine monetary profit in place of welfare, conceived as the net balance of pleasure to pain (or the net of whatever quantifiable value a utilitarian posits). For some people approaching ethi- cal questions, this structural analogy is a merit of utilitarianism. For others, it will confirm a suspicion of insufficient sensitivity to ethical variables.'^ My purpose is not to assess utilitarianism (or cost-benefit analysis). But I am stressing two points. The first is the analogy just noted. The second is the disanalogy between hedonic

596 BUSINESS ETHICS QUARTERLY

and monetary value. Hedonic value seems far more difficult to measure. Seeing why will bring out some problems for ethical research that are important for ap- praising utilitarianism.

One problem for utilitarianism is clarification of its criteria of the good. What, for instance, are pleasure and pain? Is the former partly sensory, and is the latter entailed by, say, shame or embarrassment? Is suffering necessary for pain, and how are they related, empirically and ethically? Even if this definition problem is soluble, measurement problems remain. The difficulty is partly empirical. What kinds of behavior, from self-reports to approach or avoidance behavior, indicate pleasure and pain? Suppose we adequately answer this difficult question. There is a further question of how to measure degree. Self-reports are not always true, even if they should be reliable in the statistical sense. And is calling a toothache twice as painful as a pin prick even fully clear in meaning? Supposing it is, what of treating one person's toothache as twice as bad as another's? Duration of experiences is quantitatively measurable, but perhaps intensity is not, even in a particular person. Utilitarianism, however, requires interpersonal comparisons of pleasure and pain as a basis of ethical decision.'^

If pleasure and pain were not important for every major ethical position, these problems would weigh against utilitarianism more than they do. Pleasure and pain play positive and negative roles (respectively) in all major ethical views. But whereas for other major ethical views, such as those of Kant (t948), virtue ethics (Aristotle 2002; Swanton 2003), and intuitionist pluralism (Ross 1930; Audi 2004), promotion of well-being hedonically understood is (largely under the heading of beneficence) just one among other ethical demands, for utilitarianism it is central. Because of this centrality, prediction problems are more serious for utilitarianism. The pivotal problem is to predict the effects of options on pleasure and pain. In principle, there is no time limit on the chain of effects of our deeds. Utilitarians must either ignore the potentially infinite stream of consequences or estimate them from data at the time of decision. Eacts about the present may or may not be a good guide to the indefinite future.

The problems just noted do not disappear if utilitarianism is understood in terms of some interpretation of the distributive formula. But in part because that kind of understanding raises questions about distribution, the formulations of utilitarianism that go with it raise additional problems. Section III will indicate what seem the most important distributive formulations. These are ethically significant in their own right, but understanding them in contrast with the broad interpretation of utilitarianism just outlined is important for appraising utilitarian ethics in general.

///. Three Distributive Standards

Readers who know contemporary literature in ethics, including the business ethics literature, may have noticed that both the kinds of aggregative formulations of utilitarianism cited from Bentham, Mill, and Moore and the "greatest good for

CAN UTILITARIANISM B E DISTRIBUTIVE? 597

the greatest number" fonnulation—the distributive formulation—commonly occur in the same context with no contrast drawn between them.''' I do not assume that the writers in question would not contrast them if they pursued the question of their relation; but they are rarely contrasted, and they have not been compared and contrasted in the way I propose to do here with a view to widening the alternatives for ethical inquiry.'̂ A hypothesis that in part underlies this paper is that serious defenders of utilitarianism do not draw the contrast because they would not consider the distributive formulation remotely adequate as a definition of their view, whereas many who want to represent utilitarianism in vivid simple terms and as a plausible ethical view find that formulation attractive. There is little doubt that the fonnulation is vivid and memorable (in my experience, students who have heard of utilitarianism so described tend to have a difficult time shifting to any substantially different one). Moreover, it appears to reflect an element of justice not discernible in the classical fonnulations. But its appearance of simplicity conceals its ambiguity, and (as we shall see) the element of justice it seems to capture is something utilitarians must earn by argument and cannot presuppose by definitional fiat.

To show some contrasts between the aggregative definition and the distributive formulation, let me suggest three possible readings of the latter. Each is a plausible reading given certain ethical purposes, and each does justice to the two occurrences of "greatest." This duality requires identifying two variables to be maximized. To illustrate, I will consider, as the relevant population affected by one's options, a limited group of people who, as in corporate decision-making, might be one's ac- tual concern in a decision. The generalizability of the results to larger populations will be evident.

1. Inclusiveness

Suppose I am a benefits manager wanting to extend medical help to as many of the company's employees as possible. I might assume that the selected policy cannot help them all, if only because some will ignore their illnesses or never engage the health-care system and a few will (during the years they are covered) need medical help only with ineligible problems. But I intend the policy to help as many as possible, and my top priority is inclusiveness. Still, once I find a policy that helps a maximal number of the employees, I may have a secondary preference for policies in the order of the aggregate medical good they do. My ethical decision principle, then, is twofold. An act (such as choosing a medical plan) is right if and only if, as compared with its alternatives, first, it positively affects (i.e., distributes a net good to) at least as many as any altemative and secondarily—after this num- ber is reached—it maximizes aggregate happiness, i.e., produces at least as much aggregate happiness as any altemative that positively affects as many people. It reaches the "greatest number" of those concemed, and it does the "greatest good" for them, i.e., at least as much good as any altemative can do without ehminating anyone from receiving the benefits of my action.

598 BUSINESS E T H I C S QUARTERLY

Suppose, then (to make tlie example more definite), that act A provides ten units of (medical) benefit for all employees, including those who are part-time, whereas act B produces greater benefit for a smaller group, say just full-time employees, but a greater aggregate benefit. (Even if there are fewer full-time employees, their greater benefit multiplied by their number might yield a higher total.) The inclusive- ness interpretation of the distributive formulation requires ranking act A over act B: number of people positively affected takes priority over aggregate good produced. The comparison might look like this:

Option A: 1,000 employees receive benefits whose aggregate total is 10,000;

Option B: 750 employees receive benefits whose aggregate total is 11,500,

where the aggregate totals may represent an estimate of "medical good" based on an assessment of such factors as the coverage provided by the different policies for length of hospitalization, for surgical costs, for sick days required, and for dealing with subjective descriptions of discomfort. (Assigning the probabilities and values here is a very difiicult matter, but that point applies to assessment of medical benefits from the perspective of any plausible theory.)

Granted, someone who shares the utilitarian concern with maximizing happiness might have a qualified inclusivist principle that allows benefiting fewer persons given (for instance) a huge gain in aggregate good and only a small decrease in the number of people who would receive a benefit if an alternative act is performed. Such a formula might be quantified, say allowing for a narrower distribution provided the result is a gain of at least 20 percent overall or an average gain of 10 percent. The matter could also be left to competent judges, as Mill leaves questions of the quality of pleasure."" In either case one might pay special attention to the negative value of pain. The following are possible standards that might be used in business or social policy settings.

la. Preference for Pain Reduction, Other Things Equal One could weight reducing pain over enhancing pleasure whenever other things

are equal. Consider an application of this idea. A pharmaceutical company's research might indicate that equal profit is likely from producing a drug that builds muscle in healthy people or from producing one that relieves arthritic pain. We might hope that an ethical person (especially but not necessarily a utilitarian) would make the choice in terms of the best way of doing good. The specified preference for pain reduction would be significant, though minimal; it would play just a tie-breaker role where the number affected is the same for two alternatives, but pain reduc- tion is greater in one case. Treating arthritis, then, would be the preferable way to maximize the good.

lb. Prioritizing Pain Another possibility is giving pain reduction more than a tie-breaker role in

competition with producing pleasure. One might, for instance, allow the former to compensate for an action's benefitting many fewer people than some alternative.

CAN UTILITARIANISM B E DISTRIBUTIVE? 599

The pharmaceutical example can be easily altered to illustrate this. Simply let the quantity of pleasure enhancement for unafflicted persons be greater than the quantity of pain reduction for the arthritics.'^

lc. Prioritizing the Saving of Lives Utilitarians may plausibly take the value or disvalue of human life to be often

incalculable, in part because lives are hedonically so different and in part because their length—and consequences for the well-being of others—are so difficult to predict. Still, the value of human lives may surely be estimated by utilitarians (and others), and this might affect any uses that a utihtarian might make of the distribution formula. One idea would be to give children top priority in determining an optimal distribution. They might thus be given preference over adults. Various formulas are possible for determining tradeoffs in saving children's lives—say, by donating a miracle drug—as compared with those of adults positively affected by a decision. Medical suitability is still another criterion: one could give far more of a drug to those more treatable or even (perhaps at the same time) forgo helping all who are terminally afflicted with (say) AIDS if this would cost lives owing to leaving too little medicine for non-terminal cases.

Id. Quantifying with Respect to Stakeholder Groups Any potentially useful business appHcation of a qualified inclusivist standard

might apply it to stakeholder groups. Here one could use any of the variants just noted. We need not, however, regard a stakeholder group as like an individual: since it is composed of individuals, we could average or aggregate the benefits of a busi- ness decision for each group and then proceed according to the formula we have selected. All the variants noted can come into play. Here corporate policy-making might resemble legislative policy-making.

These are other possibilities well worth exploring—both normatively, in terms of the difference in value between reducing pain and enhancing pleasure, and empiri- cally, in terms of whether people tend to prefer the former benefit over the latter. My purpose, however, is simply to bring out the differing plausible interpretations of the distributive formulation. Appraising any one interpretation and exploring plausible modifications of it is a task for another place. Since most of the variants noted here for the inclusivist reading will apply to the other readings, we can presuppose their application to those and proceed more briefly.

2. Qualified Aggregation

As important as inclusiveness is in ethical decisions, it is not the only important ethical concern, and particularly if, with classical utilitarians, we consider pleasure and pain the only basic positive and negative values, we may approach the benefits question differently. We may above all want to produce as much total medical benefit as possible; our top priority is then maximizing aggregate good. Still, our second priority might be—whether for moral or prudential reasons or both—to institute a

600 B U S I N E S S E T H I C S QUARTERLY

policy that helps as many employees as we can without sacrifice in total medical benefit. Thus, suppose one policy does far more for cancer patients than another and yields much greater net happiness, but the other policy does something positive for more members of the group, perhaps by reimbursement for a wider range of medicines. The second policy, however, yields a much lower net gain in happiness. Since I foresee less aggregate good in choosing the more inclusive policy, I reject it. I would choose a more inclusive policy if and only if doing so costs nothing in aggregate good. Call this the aggregation interpretation of the distributive formu- lation. On this interpretation. Option B above is superior to Option A. Suppose, however, that the choice were like this:

Option C: 1,000 employees receive benefits whose aggregate total is 10,000;

Option D: 1,100 employees receive benefits whose aggregate total is 10,000;

Option E: 1,200 employees receive benefits whose aggregate total is 10,000.

In this case, D is preferable to C, but still wrong, since only E reaches the greatest number one can in the situation requiring a policy decision.

It should be clear that both the inclusiveness and the aggregation readings of the distributive formula represent natural interpretations of it. For the former, the ethically right action above all serves the greatest number of those concerned and, secondarily, maximizes the good it produces for that maximal group; for the latter, the ethically right action above all produces the greatest good for those concerned and, secondarily, distributes it to the largest number reachable without diminution in the aggregate good distributed. Thus, if our commitment is to the aggregation reading and we prefer the first policy (Option D) over the second (C) on grounds of aggregative hedonic superiority, we would also prefer a third to the first if, like (E), it does as much good but distributes some good to at least one additional person.

An important area of contemporary business decisions is the question of contribu- tions to deal with the AIDS epidemic. A plausible principle devised to deal with this kind of catastrophe calls for firms to devote at a minimum "the largest sum of (1) their most recent year's investment in social initiatives, (2) their five year average of investment in social initiatives, (3) their industry's average investment in social initiatives, or (4) the average investment in social initiatives by firms in their home nation."'* As stated, this is an aggregation principle with no specific indication of how the donation is to be distributed among beneficiaries. Distribution might be accomplished in any of the ways we have identified as interpretations of the dis- tributive utilitarian formula and, as we shall soon see, many other ways.

3. An Egalitarian Interpretation

The distributive formula admits of still another reading. Suppose I am a manager in a start-up company with all new people in my division, very much on a par in terms of their contributions during the first year, all recent recipients of an MBA, and all of roughly comparable experience. Particularly if I believe they will discover what percentage raise others receive, I may judge that the best thing (at least initially) is

C A N UTILITARIANISM B E D I S T R I B U T I V E ? 601

to treat them equally. I then might seek the largest percentage raise such that I can distribute this same determinate percentage to all (or as many as possible, suppos- ing I have some people who will, given their special contracts, have to receive a different percentage and so are not in the population I consider). My top priority is to produce, for the relevant population, the greatest determinate good that, within my resources, can go to "the greatest number" of them"(for simplicity, I am taking monetary good to be the only relevant good and using a percentage rather than a monetary amount). Thus, even if the CEO would give me more to distribute if I made the distinctions in monetary desert that the CEO thinks I should, I will forgo the greater quantity for distribution in my group in order to maintain my egalitarian policy. (I would naturally seek more for distribution if I could give it equally, for then I could give a greater good to the greatest number—here, the same number, since money, unlike many medical benefits, can be given to everyone in the group who is eligible.) One choice might look like this:

Option E: Twenty receive 5 percent raises, with an aggregate total of $500,000;

Option G: Twenty receive varying raises (say 2-10 percent) with an aggregate total of $600,000.

The egalitarian reading requires choosing E over G. Granted, one could qualify the principle in question and weigh equality in such a way that, say, inequalities of two percentage points or lower can be offset by an aggregate gain for the group of 20 percent. But the point is that there is a reading on which equal treatment is a non-negotiable standard having priority over aggregate good.

Is the reading that requires distributing a determinate good to the greatest number egalitarian rather than utilitarian? It is certainly not utilitarian on the aggregative views of Bentham, Mill, and many others, but it does give aggregate good enough weight to require a kind of maximization of the good. The imagined (maverick) utilitarian certainly seeks to do as much good as possible; such a theorist is simply unwilling to maximize at the expense of equal treatment. An underlying view here might be that we should try to produce as much good as possible in the right way. If this idea is combined with the egalitarian reading, to ask which act would produce the greatest good for the greatest number is to ask which act that would positively affect the greatest number of people one can reach would produce the most good for all equally.

Despite appearances, none of the views considered above implies that equal- ity is an intrinsic good—nor does any genuinely utilitarian view. Admittedly, the most Hkely ethical basis for the suggested egalitarian principle may be a moral commitment, perhaps based on a sense of justice, to equal treatment of persons as intrinsically desirable, supplemented by a secondary commitment to maximizing their good. Nonetheless, the principle might still be used by someone who is a utili- tarian in Mill's sense but, quite apart from any antecedent commitment to justice, is convinced that equal treatment is crucial for maximizing aggregate good. One might think, for example, that only equality preserves morale.

602 BUSINESS E T H I C S QUARTERLY

There may be other viable interpretations of the distributive formulation, but these three seem most plausible and each comports with ethical ideals many find compelling. None of the formulations, however, is equivalent to utilitarianism as Bentham or Mill introduced it. They all greatly differ from it. It does not even entail any of them. Recall that Mill's formulation mentions tendency, which is usually in- terpreted probabilistically. Consider the common assumption that on this "classical" utilitarianism, an act is right if and only if it has at least as much expected utility as any alternative, where the expected utility of an act is calculated by multiplying the probabilities times the (positive or negative) values of its possible outcomes and adding these products." A classical utilitarian may properly make a choice that would be wrong on any of the three interpretations of the distributive formulation. Recall the benefits manager choosing among insurance policies. Suppose two would bring equal aggregate good, but the second would benefit a smaller number and do so unequally, whereas the first would benefit a larger number of people—the largest that can be affected by one's action in the situation—and would do so equally. The choice might look like this:

Option H: 1,000 employees receive equal benefits whose aggregate total is 10,000;

Option I: 950 employees receive unequal benefits whose aggregate total is 10,000.

The classical, aggregative utilitarian may decide arbitrarily between these. The three distribution principles do not permit that. There are three points here. First, Option I benefits fewer people, hence does not reach the "greatest number" and is unacceptable under the inclusiveness formulation. Second, Option H produces equivalent aggregate good and so ranks equally in achieving the "greatest good," but also reaches a greater number, so is preferable by the aggregation formula- tion—which is secondarily inclusivist in giving priority to the "greatest number." Third, Option I is inegalitarian, since it treats those it affects unequally, and is thus ruled out by the egalitarian formulation.^"

IV. Some Incentives for Adopting a Distributive Version of Utilitarianism

In providing three main interpretations of the distributive formula, I framed examples that suggest how the resulting principles (or qualified versions of them) might be plausible for certain kinds of ethical decisions. I do not think that it is only the rhetorical elegance of "the greatest good for the greatest number" that makes the formula so natural for so many. Ironically, perhaps, whatever ethical incentive there is apparently does not come from "classical" utilitarianism (say, Bentham's or even Mill's versions of it).^' Identifying right action with aggregate utiUty-maximizing action does not strictly imply anything about how good or bad things (benefits and burdens, for instance) should be distributed among people affected.^^

It seems likely, to be sure, that if one of two acts that are equivalent in aggregate goodness yields a wider distribution of it, a utilitarian will tend to prefer it. But the

CAN UTILITARIANISM B E DISTRIBUTIVE? 603

theory does not require this. From a utilitarian point of view, a wide distribution may be sought because it is more positively received, hence not resented (or less resented). But here it may have a better aggregate score than alternatives which—apart from the resentment—would score higher in utility. A wide distribution is not in itself preferable for utilitarians.

Part of the incentive for taking utilitarianism to entail the aggregation inter- pretation of the distributive formula (or something close to it) may be empirical assumptions such as these, taken as built into calculation of the total consequences of a distribution: (1) people tend to resent unequal distributions (at least) of certain common kinds; (2) resentment is either painful or tends to cause pain; (3) people tend to accept equal distributions; (4) this acceptance, if not pleasurable, at least tends to reduce pain; and (5) wide distribution is desirable given a hedonic analogue of the declining marginal utility of money. These are ethically important assump- tions, especially for policy makers, even apart from utilitarianism. Recognition that such assumptions may be important for supporting utilitarianism (or close relatives of it) is further incentive to subject them to more intensive empirical investigation than they have received. All of them are pertinent to appraising the influential idea that "good ethics is good business."

It should be noted, however, that even if (l)-(5) are empirically sustainable, they do not directly support any of the three distributive formulations. This is in part because those formulations are framed in superlatives—we must, for instance, reach the "greatest" number our acts can positively affect. In real ethical deci- sions, whether in corporate policy matters or elsewhere, there are tradeoffs. Thus, the proportion of employees who will benefit from a policy is relevant given the goal of reaching the greatest number; but given the potentially conflicting goal of maximizing aggregate good, we might be ethically warranted in choosing a policy that reaches only the vast majority but does a great deal more aggregate good than one that reaches all employees.

As to (5), insofar as any non-hedonic good—such as monetary compensation—is used to measure contribution to happiness, the declining marginal utility assumption also indicates why the inclusivist formula is inadequate. It would likely be foolish to try to reach the greatest possible number if one's resources would produce just a few pennies for each. Tradeoffs again come in. Is there, however, such a thing as the declining marginal utility of pleasure? Bentham and Mill would deny this: for them pleasure—as enjoyment, not merely pleasant sensation—is the good. But a preference utilitarian, for whom a kind of basic desire satisfaction is (at least functionally) the good, might accept it. For this kind of view, empirical information about what satisfies basic desires would have much ethical significance.

One reason it is important to see that utilitarianism does not essentially concern either equality or scope of distribution is that a major objection to the view is that it cannot accommodate distributive justice. To build a distributive standard into it at the outset biases this question and may give utilitarianism an unearned advantage. The desirability of producing net good for more people rather than fewer is not—for

604 BUSINESS ETHICS QUARTERLY

utilitarians—necessary or intrinsic. Preference for wider distribution must earn its status by contributing to aggregate goodness (say, happiness, as for Bentham, Mill, and certain "welfarist" utilitarians). The same holds for equal treatment. If it is a good, it is an instrumental one.

If the aggregative reading of (classical) utilitarianism seems unduly restrictive,^' let me underscore what in large part makes the theory distinctive and (to many) attractive. First, it is empirical, in representing the good—hence the ground of right action—as empirical. Pleasure, pain, welfare as defined by various socioeconomic measures, and other candidates for utility, are in principle accessible to empirical determination (something that scientifically minded thinkers may find attractive in a decision standard). Second, pleasure, pain, and the other relevant candidates are (arguably) at least ordinally measurable; the theory is in this minimal sense quantitative. Third, given these points, together with the possibility of scientifically predicting the effects of actions, utilitarianism appears objective. Fourth, virtually all theories of the good that countenance intrinsic and not just instrumental goods include, as major cases, pleasure as good in itself and pain as bad in itself. This implies both that utilitarianism captures one major element in any plausible ethics and that no plausible ethical alternative is entirely free of the difficult measurement problems just described.

Justice, by contrast with pleasure and pain, is apparently not empirically mea- surable. Equality of distribution of material elements is empirically determinable, but such material equality may or may not be just, depending on what people have earned or on whether different distributions are required by promises or hy fairness. Utilitarians have commonly sought to derive these notions (or at least the evalua- tive standards they represent) from the principle of utility together with facts about people and the world: to "operationalize" justice, as it were. Apart from empirical information, they will refuse to endorse any criterion of just distribution.

V̂ Distributive Utilitarianism

We are now in a position to address a central question for this paper directly. Should distribution matter for utilitarianism—in standard versions, as we might call them? It should; the problem is to determine how it matters. In broad terms, for utilitarianism, taken to be a consequentialist view as its proponents intend it, distribution matters derivatively. The number of people affected by an act counts only insofar as how many affects how much. If we produce less overall good by helping ten people rather than five, then for standard (aggregative) utilitarianism we should help the five, even if the difference in quantity is small.

Thus, the answer to the question whether utilitarianism can be distributive is that it can be—derivatively. Someone might claim that in the "greatest good for the greatest number" formulation it is essentially distributive; but there are at least two reasons to resist this. First, it glosses over the ambiguity of the formulation and gives the impression that utilitarianism itself provides a way to solve the problems

CAN UTILITARIANISM B E DISTRIBUTIVE? 605

we have seen (such as tradeoff problems) for any unqualified inclusivist, aggrega- tive, or egalitarian interpretation. Second, calling utilitarianism distributive at all is likely to cause confusion if it is not accompanied by the kind of contrast I have drawn between the utilitarian views actually defended by proponents of the posi- tion and the distributive readings of the "greatest good for the greatest number" formulation.

Suppose, however, that one ignores these cautious and, taking advantage of the disambiguation in section III, speaks of distributive utilitarianism. This may have any of at least three basic forms (more sophisticated—and I think more plau- sible—forms may be constructed when tradeoff and measurement problems are taken into account, along lines suggested in section III). Each calls for maximization but restricts the conditions in which this is required. Inclusivist distributive utilitarianism requires that right acts affect the largest reachable number and only then calls for maximization of quantity of good. Aggregative distributive utilitarianism requires that right acts maximize aggregate good but calls for preference, given that result, for the widest distribution possible. Egalitarian distributive utilitarianism requires maximizing the greatest determinate good that can be brought about for all affected; right acts thus treat these people equally.

These views must still be developed in concert with a method of determining who would be affected by an act and, in the egalitarian case, an adequate measure of equality, whether equality is conceived hedonically, economically—as in some managerial decisions—or otherwise. But these are problems for any utilitarian theory—indeed, for the many ethical theories that give enhancement of pleasure and reduction of pain a major place—and the difficulties they imply for distribu- tive utilitarianism may still leave it more attractive than standard utilitarianism for those whose ethics is broadly utilitarian in inspiration. The three distributive views can also be combined in various ways. For instance, the inclusivist standard can be modified to have to allow either considerations of equality in a distribution or considerations of quantity that reach a certain threshold to offset a diminution in the total number to whom a good can be distributed.

This is not the place for an assessment of utilitarianism, and there is already a large literature for and against it.̂ " But our results bear on how it should be assessed. If read distributively, it tends to look more plausible than it is; for this reading makes a concern with distribution—intuitively, a kind approaching some standard of justice, however vague—appear to be intrinsic to the view when in fact it is not. My aim has been chiefly to clarify what it is that must be assessed and what are some of the revisions in standard versions of the view that may be attractive to many who think of it as committed to a specific distribution principle. If I have succeeded, the paper should facilitate clearer ethical thinking and better contribu- tions to solving ethical problems by (among others, I hope) those who have been

606 BUSINESS ETHICS QUARTERLY

characterizing utilitarianism as the claim that what we ought to do is produce the greatest good for the greatest number. This includes some who accept the view, many critics of it, many neutral expositors, and numerous others. My suggestion is that much reference to utilitarianism should be recast: the view should be described and explained primarily as an aggregative theory, and those who are attracted to a distributive version might, failing a better name, call it a distributive utilitarianism and indicate the difference. Distributive versions should not, in my view, be pre- sented as versions of utilitarianism without qualification, and the non-aggregative versions are best not considered forms of utilitarianism at all.

The research problems that come to the fore in the light of the distinction between utilitarianism as understood by major proponents and the various distributive prin- ciples discussed above go even beyond the difficulties of assessing utilitarianism. There are conceptual, normative, and empirical questions here. My concern has been mainly the first two kinds of questions, but utilitarianism in any formulation raises empirical questions that also deserve further inquiry. If the main points made here are correct, however, only a genuinely aggregative formulation of utilitarianism comports with the kind of theory of value that underlies utilitarianism: whether in act or rule formulations, it subordinates the right to the good and the achievement of justice to the promodon of goodness. For most ethical thinkers, giving justice a subordinate place goes strongly against the grain, and this gives impetus to conceiv- ing utilitarianism as essentially distributive. Many hybrid theories result from that interpretation, as we have seen. Some of these may be more plausible than utilitari- anism understood as its founders and leading proponents intend. But the possible merits of these theories are easily obscured by the ambiguity of the "greatest good for the greatest number" formulation, and utilitarianism itself may gain undeserved plausibility from the element of justice that formulation appears to promise.̂ ^

Notes

1. DesJardins (2005) presents a brief, balanced account of how both utilitarian and Kantian considerations can support a stakeholder approach. See esp. pp. 67-69.

2. DesJardins, e.g., says, "Utilitarianism is typically identified with the policy of 'maxi- mizing the overall good' or, in a slightly different version, of producing 'the greatest good for the greatest number'" (2005: 30). He does not explain the difference and discusses utilitarianism in relation to both characterizations. For earher recent examples of this usage see Dunfee et al. 1997, where utilitarianism is introduced as (though not only as) the view for which "The emphasis is on the greatest good for the greatest number" (p. 18); and Cohen (2001), who says, expositing Bentham, "A behavior is ethical if it results in the greatest good for the greatest number" (2001: 580). Cf. Beauchamp and Bowie 2001: 110. Distributive formulations are also used to represent utilitarianism by philosophers not in business schools. Maclntyre says of utilitarianism, "the position demands that some sense be given to the concept of the overall community of interests, to some notion of 'the greatest happiness of the greatest number'" (1964: 3-4), and uses the same formulation in stating utilitarianism in Maclntyre 1997: 219, and in Maclntyre 1984: 63-64; Rescher says of utilitarianism, "This doctrine is founded upon the principle of utility, which asserts that utility (or, if you like, the good things of this life) should be distributed according to

CAN UTILITARIANISM B E DISTRIBUTIVE? 607

the rule of 'the greatest good for the greatest number'" (1966: 8); and Scruton (2006), writing on Mill, speaks of "Utilitarianism ('that action is right which promotes the greatest happiness for the greatest number')" with no reference for the quotation (which is not from Utilitarianism if from Mill at all).

3. Bentham 1961: 164. The reference to the tendency an action "appears to have" suggests understanding the view in terms of (rationally) expected consequences rather than actual ones. This complication need not be introduced in the text, since my points apply to either probable- consequence or actual-consequence formulations of utilitarianism.

4. Bentham 1961: 164, emphasis added. I assume that, by "comes to the same thing," Bentham refers to having the same net effect on utility; he does not appear to be (implausibly) identifying reduction of pain as psychologically identical with enhancement of pleasure.

5. See Bentham 1843: sec. 6, pp. 5-6. There is at least one earlier formulation Bentham might have followed: Francis Hutcheson had said, "in equal degrees of happiness expected to proceed from the action the virtue is in proportion to the number of persons to whom the hap- piness shall extend (and here the dignity, or moral importance of persons, may compensate numbers); and in equal numbers the virtue is as the quantity of happiness . . . so that that action is best, which produces the greatest good for the greatest numbers." See Hutcheson 1990: 515. For extensive discussion and analysis of Bentham's use of "greatest happiness for the greatest number" see Burns 2005.

6. The question of what population utilitarian calculations should consider—all actual people, actual people and certain animals, actual and future people, e.g.—poses problems for utilitarianism—many brought out with clarity and force in Maclntyre 1977, but differences among the various possible solutions do not affect my arguments. For detailed discussion of Bentham's utilitarianism that is generally supportive of the interpretation sketched here, see Postema 2006.

7. Mill 1979: 7. 8. Moore 1903: chap. 5 (Dewey, Gramlich, and Loftsgordon 1961: 342). We again find

vagueness regarding the population in question (with Moore's cosmic phraseology giving the appearance of ascribing to us an obligation to populate the universe so long as aggregate hap- piness increases); but nothing in this paper turns on how the relevant population is delimited. Moore was not a hedonist in his theory of value, but this fact does not affect the point that, like Bentham and Mill, he uses an aggregate standard of goodness to determine our moral obligation. Cf. Parfit's point that for utilitarianism "All that matters are the amounts of happiness or suffer- ing, or of benefits and burdens. It makes no moral difference how these amounts are distributed as between different people" (Parfit 1984: 330).

9. Prominent defenses are offered by Smart and Williams (1973) and Brandt (1979), and a highly detailed, sympathetic presentation is provided by West (2004), who also devotes much attention to the question whether Mill was an act- or rule-utilitarian (pp. 74-95)—a matter on which this paper is neutral, since its points appiy to both kinds of utilitarianism. For a defense of a "cooperative utilitarianism" on which cooperation is combined with a maximizing strategy, see Regan 1980.

10. By "good as an end" Mill means intrinsically good, as opposed to instrumentally good. Largely in the light of this value claim, one writer has gone so far as to claim that "in Mill, the principle of Utility is the principle that happiness is the only thing desirable as an end" (Brown 1973; Lyons 1997: 9). Brown's case for this brings out much about Mill, but does not seem plausible given much else Mill says.

11. For critical discussion of a maximizing desire-satisfaction view of utilitarianism and a case for preferring a hedonistic view, see Brandt 1979. Much interpretive discussion relevant to business decisions is provided by Layard 2005.

608 BUSINESS ETHICS QUARTERLY

12. It should not be assumed that cost-benefit analysis has no significant ethical uses in non-utilitarian ethics. An indication of why this is so and of some of the problems it raises is provided in Audi forthcoming.

13. For a pessimistic view see Arrow 1963. He says, e.g., "interpersonal comparison of utilities has no meaning" (chap. 2). Cf. Maclntyre's stress on the incommensurability of goods in, e.g., Maclntyre 1984: 198-99.

14. The first two texts cited in note 2 are cases in point. In another case, the authors go to considerable lengths in characterizing utilitarianism as an aggregative (maximizing) consequen- tialism, yet close the relevant paragraph by citing a "greatest happiness for the greatest number" formulation (perhaps supposing that some distributive subprinciple is an appropriate route to the required maximization (Donaldson, Werhane, and Cordig 2002: 3).

15. To his credit, Frankena (1973) points out that the "greatest good for the greatest number" formulation is not entailed by utilitarianism; but he does not note its ambiguity. See esp. p. 42. Sidgwick, too, in his authoritative Outlines of the History of Ethics (1886), treats the formula- tion, which he cites in Hutcheson, as unambiguous. In their "Utilitarianism and Business Ethics," Snoeyenbos and Humber (1999: 106) note that the formulation differs from utilitarianism, but treat it is unambiguous (and indeed leave it indeterminate, as implying no priority ordering as between aggregate good and number of people affected). See pp. 203-04. Rescher (1966) notes the first two readings identified in this paper but not the egalitarian one, nor does it seem to occur to him—perhaps because he is writing on distributive justice—that he is considering formulations proponents of utilitarianism would not think intrinsic to their view.

16. Mill said, e.g., some kinds of pleasure are more desirable and more [intrinsically] valuable than others. . . . Of two pleasures, if there be one to which all or almost all who have experience of both give a decided preference, irrespective of any feeling of moral obligation to prefer it, that is the more desirable pleasure" (Mill 1979: 8). For history and analysis of this view see Gibbs 1986.

17. It is interesting to note here that (according to Layard 2005) people "hate a loss more than they value an equal gain. . . . Loss-aversion is contrary to standard economic theory, which implies that if the stakes are small enough, people do not care about risk" (Layard 2005- 141^2).

18. Dunfee 2006: 190. The context does not specify a time period during which this mini- mum should (apart from "financial exigency") be reached but the points made in the text will be unaffected by making the plausible assumption that the standard be applied annually.

19. The calculation procedure is the one commonly used in decision theory and cost-benefit analysis. Utilitarians need not, and do not in general, make the empirical assumption that ordinary moral decisions are reached by such calculations. Mill in particular emphatically argued that calculation is needed only when ordinary moral rules, such as the rule prohibiting lying, conflicts with another, for instance the rule prohibiting promise-breaking. See esp. Mill 1979: chap. 2. This emphasis is in part what leads to the idea that the proper application of utilitarian calculations is to (intemalizable) subsidiary rules and that individual acts are right or wrong on the basis of their conformity with these. For detailed discussion of such rule theories, with a critical treat- ment of (act) utilitarianism as carefully formulated, see Hooker 2000. (I ignore difficulties about what counts as an available alternative. Not just any remotely possible alternative counts. This difficulty is significant, but utilitarians have several ways of arguing that it is not crippling.)

20. Granted, applying one of these distribution principles could sometimes maximize ag- gregate utility, but that is a contingent matter. The theory as formulated by its major proponents entails no distribution principle—a point that is not controversial in major treatments of utilitarian- ism. See, e.g., Rawls 1971 and, for recent discussion of utilitarianism and related consequentialist theories. Hooker 2000.

CAN UTILITARIANISM B E DISTRIBUTIVE? 609

21. To avoid unnecessary complexity I am not addressing rule-utilitarianism here, though a case can be made for its being Mill's most considered position. In any event, even rule-utilitarians use aggregate good as what everyone's conformity to utility-maximizing rules must achieve. My points thus remain unaffected. See Hooker 2000 for an instructive discussion of how rule-theories differ from act-theories.

22. The phrase "benefits and burdens" is used by Rawls, who provides a sustained critique of utilitarianism (understood as I suggest it should be) in A Theory of Justice (1971).

23. I omit consideration of Mill's often debated point that some pleasures are intrinsically better than others (1979: 8-11). Many utilitarians do not accept it, but in any case, for calculation purposes, differences in quality can be represented as differences in quantity.

24. The literature on utilitarianism is vast and there is indeed now a professional jour- nal—Utilitas—devoted to the study of utilitarianism and related ethical matters. Among the general collections that carry important contributions, many critical of utilitarianism are Bayles 1968, Page 1968, Miller and Williams 1982, Glover 1990, and Lyons 1997. Maclntyre (1964, 1984), Bernard Williams's contribution to Smart and Williams 1973, and Hooker 2000 contain much plausible criticism. Quantitatively sophisticated contributions have been made in made papers by Amartya Sen, and Sen 1990 illustrates how "utility can be viewed as a vector" and indeed quantified in a way that supports my point (in the text) that it may be viewed as structurally analogous to cost-benefit analysis. For a non-utilitarian statement of his ethics in which utility is constrained by rights, see Sen 2004.

25. For helpful comments on earlier versions I particularly want to thank Edward Conlon, Norman E. Bowie, Thomas Donaldson, Georges Enderle, Rex Mixon, Patrick E. Murphy, Oliver Williams, and, especially, anonymous readers.

References

Aristotle. 2002. Nicomachean Ethics, trans. R. Crisp. Cambridge: Cambridge Univer- sity Press.

Arrow, K. J. 1963. Social Choice and Individual Values, 2nd ed. New York: John Wiley & Sons.

Audi, R. 2004. The Good in the Right: A Theory of Intuition and Intrinsic Value. Princ- eton, N.J.: Princeton University Press.

Forthcoming. "The Place of Cost-Benefit Analysis in Business and the Profes- sions." Business and Professional Ethics Joumal.

Bayles, M. D. 1968. Contemporary Utilitarianism. New York: Anchor. Beauchamp, T. L., and N. E. Bowie. 2001. Ethical Theory and Business, 6th ed. Engle-

wood Cliffs, N.J.: Prentice Hall. Bentham, J. 1843. The Works of Jeremy Bentham, ed. John Bowring. Edinburgh.

1961. An Introduction to the Principles of Morals and Legislation. In Dewey, Gramlich, and Loftsgordon 1961: 163-78.

Brandt, R. B. 1979. A Theory of the Good and the Right. Oxford: Oxford University Press.

Brown, D. G. 1973. "What is Mill's Principle of Utility?" Canadian Joumal of Phi- losophy 3: 1-12. Reprinted in Lyons 1997: 9-24.

Bums, J. H. 2005. "Happiness and Utility: Jeremy Bentham's Equation." Utilitas 17(1): 46-61.

610 BUSINESS E T H I C S QUARTERLY

Cohen, J. 2001. "Appreciating, Understanding and Applying Universal Moral Prin- ciples." Joumal of Consumer Marketing 18: 578-94.

DesJardins, J. 2005. An Introduction to Business Ethics, 2nd ed. New York: McGraw- Hill.

Dewey, R. E., F. W. Gramlich, and D. Loftsgordon, eds. 1961. Problems of Ethics. New York: Macmillan.

Donaldson, T , P. Werhane, and M. Cordig, eds. 2002. Ethical Issues in Business, 7th ed. Upper Saddle River, N.J.: Prentice Hall.

Dunfee, T. W. 2006. "Do Firms with Unique Competences for Rescuing Victims of Human Catastrophes Have Special Obligations? Corporate Responsibility and the case of AIDS Victims in Suh-Saharan Africa." Business Ethics Quarterly 11(2): 185-210.

Dunfee, T. W., F. F. Gibson, D. Whitman, D. B. Cohen, and B. A. Brennan. 1997. Modem Business Law: The Regulatory Environment, 3rd ed. New York: McGraw-Hill.

Frankena, W. K. 1973. Ethics, 2nd ed. Englewood Cliffs, N.J.: Prentice Hall. Frederick, R. E., ed. 1999. A Companion to Business Ethics. Oxford: Blackwell. Gibbs, B. 1986. "Higher and Lower Pleasures." Philosophy 61: 31-59. Glover, J. 1990. Utilitarianism and Its Critics. New York: Macmillan. Hollins, T. H. B. 1964. Aims in Education: The Philosophic Approach. Manchester:

Manchester University Press. Hooker, B. 2000. Ideal Code, Real World. Oxford: Oxford University Press. Hutcheson, F. 1990. The Original of Our Ideas of Beauty and Virtue. In Schneewind

1990: 505-23. Kant, Immanuel. 1948. Groundwork of the Metaphysics of Morals, trans. H. J. Paton.

New York: Harper and Row. Layard, R. 2005. Happiness: Lessons from a New Science. New York: Penguin Press. Lyons, D. 1997. Mill's Utilitarianism: Critical Essays. Lanham, Md.: Rowman and

Littlefield. Maclntyre, A. 1964. "Against Utilitarianism." In Hollins 1964: 1-23.

1977. "Utilitarianism and the Presuppositions of Cost-Benefit Analysis: An Essay on the Relevance of Moral Philosophy to the Theory of Bureauracy." In Sayre 1977: 216-32.

-. 1984. After Virtue, 2nd ed. Notre Dame, Ind.: University of Notre Dame Press.

Mill, J. S. 1979. Utilitarianism, ed. George Sher. Indianapolis: Hackett Publishing Co.

Miller, H. B., and W. H. Williams. 1982. The Limits of Utilitarianism. Minneapolis: University of Minnesota Press.

Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridge University Press. Page, A. N. 1968. Utility Theory: A Book of Readings. New York. John Wiley &

Sons. Parfit, D. 1984. Reasons and Persons. Oxford: Oxford University Press. Postema, G. J. 2006. "Interests, Universal and Particular: Jeremy Bentham's Utilitarian

Theory of Value." Utilitas 18(2): 109-33. Rawls, J. 1971. A Theory of Justice. Cambridge, Mass.: Harvard University Press.

C A N UTILITARIANISM B E D I S T R I B U T I V E ? 611

Regan, D. 1980. Utilitarianism and Co-Operation. Oxford: Oxford University Press. Rescher, N. 1966. Distributive Justice. Indianapolis: Bobbs-Merrill. Ross, W. D. 1930. The Right and the Good. Oxford: Oxford University Press. Sayre, Kenneth, ed. 1977. Values in the Electric Power Industry. Notre Dame, Ind.:

University of Notre Dame Press. Schneewind, J. B., ed. 1990. Moral Philosophy from Montaigne to Kant, vol 2. Cam-

bridge: Cambridge University Press. Scruton, R. 2006. "Thoroughly Modem Mill." Wall Street Journal (May 16): opinion

page. Sen, A. 1990. "Partial Utility." In Glover 1990: 78-83.

2004. "Elements of a Theory of Rights." Philosophy & Public Affairs 32: 315-56.

Sidgwick, H. 1886. Outlines of the History of Ethics. London: Macmillan. Smart, J. J. C , and B. Williams. 1973. Utilitarianism: For and Against. Cambridge:

Cambridge University Press. Snoeyenbos, M., and J. Humber. 1999. "Utilitarianism and Business Ethics." In Fred-

erick 1999: 17-29. Swanton, C. 2003. Virtue Ethics: A Pluralist View. Oxford: Oxford University Press. West, H. 2004. An Introduction to Mill's Utilitarian Ethics. Cambridge: Cambridge

University Press.