When Jack and Jill Make a Deal

In ordinary circumstances human actions have a myriad of unintended and often unforeseen consequences for the lives of other people. Problems of pollution are serious examples, but spillovers and side effects are the rule not the exception. Who knows what consequences this essay may have?

This essay is concerned with the problems of justice created by spillovers. After characterizing such spillovers more precisely and relating the concept to the economist's notion of an externality, I shall then consider the moral conclusions concerning spillovers that issue from a natural rights perspective and from the perspective of welfare economics supplemented with theories of distributive justice. I shall argue that these perspectives go badly awry in taking spillovers to be the exception rather than the rule in human interactions.

1. Externalities

Economists have discussed spillovers under the heading of "externalities." To say this is not very helpful, since there is so much disagreement concerning both the definition and significance of externalities. As Arrow notes, "Yet, surprisingly enough, nowhere in the literature does there appear to be a clear general definition of this concept [of public goods] or the more general one of "externality."<1> In the appendix I review some definitions of externalities and justify the assertions in the paragraphs that follow.

The crucial features of externalities that are common to most accounts are (a) missing or imperfect markets<2> (b) interdependencies in "objective functions" (preferences or production functions)<3> and (c) problems of exclusion from individual transactions.<4> The driving concern behind these apparently disparate accounts is the impact of externalities on efficiency (Pareto optimality), and it is not hard to see that, despite the wide differences in the different accounts, there is a single kind of phenomenon at issue. These accounts are descriptions of different parts of the same animal. Consider, for example, the case of a fast-food restaurant, the smell of whose rancid oil disturbs nearby residents. The market in odorless air is missing or imperfect, there is an interdependency between the restaurant's production and cost functions and the neighbors' utility functions, and none of the neighbors can be excluded from enjoying the benefits or suffering the costs of decisions concerning the release of smells.

It is easy to see how such interdependencies, market failures and problems of exclusion can give rise to problems of efficiency, and indeed some economists and philosophers build inefficiency into the definition of an externality.<5> But whether one does so is merely a matter of terminological convenience. The heart of the notion is that some costs or benefits of a decision do not bear on the decision-maker, and the efficiency of voluntary transactions is thus not guaranteed. The focus is on efficiency and on the decision-maker A.

But what is crucial about spillovers from the perspective of justice is that A's actions have effects on individuals besides those engaged in any voluntary transaction with A. The focus is usually on those affected by A's actions. Economists' definitions of externalities do not focus on this feature, and none seems entirely suitable for an inquiry into the justice of market outcomes.

I suggest that the properties of what economists have called "externalities" that are relevant to questions about the justice of the outcomes of voluntary transactions are that they involve an unintended (side) effect of A on B to which A or B have not consented.<6> Although it may be misleading to think of this as a definition of the term, "externality," which might be better left as a concept with which to address questions of efficiency, I think it is worth doing so. For most of what economists have called externalities are unintended effects of A on B to which A and B have not both consented. Intuitively one might maintain that intended harms and benefits can also be externalities, but I think this intuition is questionable. How, except via intention, could one distinguish an externality, such as water pollution, from the intentional poisoning of a neighbor's ground water?

So I shall take an externality (or, if one objects to my appropriation of the term, the analogous concept that is relevant to considering the justice of voluntary interactions) to be an unintended and not fully voluntary effect of some agent's actions on others.

The distinction between effects that are intended and effects that may be merely foreseen should remind one of the doctrine of the double effect, about which more will be said below in section ??. This echo is important, for, as we shall see below, the deontological rights theorist's best option for dealing with some of the difficulties posed by externalities is to rely heavily on the doctrine of double effect. Unintentional rights violations, even if foreseen, may be permissible when intentional rights violations are not.

The effect of my bad breath on another bus rider, the inspiration of Gandhi's example on Martin Luther King, the madness of Nineteenth-Century hatters, and the displacement of type-setters by computers are all externalities. Notice that externalities may be beneficial (called "positive") or harmful ("negative"). They may result from the action of only one individual or from the action of some group. They may result from market activities or from activities that have nothing to do with economics. And their causal consequences may be mediated via markets (in which case they are called "pecuniary externalities"), outside of markets (in which case they are called "non-pecuniary externalities") or partly via markets and partly outside of markets. The damage caused to the products of nineteenth-century hand weavers by the smoke from the steam-powered looms would be a negative non-pecuniary externality, while the damage caused to the human capital and livelihood of the handweavers by the competition of machine-woven cloth would be an example of a pecuniary externality. The increased nuisance of rodents living in crumbling housing abandoned by bankrupted hand weavers would be a negative external effect of the introduction of power looms that is neither purely pecuniary nor purely non-pecuniary. I shall focus on negative externalities, because they raise more pointed questions of justice, but many of the same issues arise with respect to positive externalities.

Conceived of in this general way, externalities are unavoidable and ubiquitous. When I buy the last six-pack of beer at the corner convenience store, the causal effects on next customer, who may be made miserable or saved from alcoholism by my action, are unintended and have no effect on my action. The results may consequently be suboptimal. The next customer and I might, for example, both prefer an outcome in which she gets the beer and I receive some compensation. It is impossible to imagine a world of any complexity without externalities at every turn, and some of these are of moral significance.

2. Externalities Cannot Be Ignored

It is obvious that externalities can lead to rights violations and injustices. If the storage tanks leak at the neighborhood gas station, poison my wells and cause serious illness or death, every moral theory would recognize that a moral injury has apparently occurred. While this much is obvious, many seem not to have appreciated the significance of such spillovers for an extremely common vision of market and moral life. For example, Robert Nozick maintains that

If the world were wholly just, the following inductive definition would exhaustively cover the subject of justice in holdings:

1. A person who acquires a holding in accordance with the principle of justice in acquisition is entitled to that holding.
2. A person who acquires a holding in accordance with the principle of justice in transfer, from someone else entitled to the holding, is entitled to the holding.
3. No one is entitled to a holding except by (repeated) applications of 1 and 2.<7>

In the actual world one must also be concerned about the rectification of past injustices. Just acquisition is appropriation of unowned things that satisfies Nozick's so-called "Lockean proviso" (which basically requires that the appropriation not make others worse off). Just transfer is voluntary transfer of what one is entitled to, provided the transfer also satisfies the proviso.<8> But, as Nozick hints in a footnote and makes explicit later, voluntary transfers can run afoul of the Lockean proviso. The Lockean proviso would, for example, forbid transactions that permit someone to acquire rights to all drinking water (Anarchy, State and Utopia, p. 179). But Nozick readily forgets his own qualification. In considering the thought experiment of individuals selling rights to control various of their activities, he writes for example, "Since this very extensive domination of some persons by others arises by a series of legitimate steps, via voluntary exchanges, from an initial situation that is not unjust, it itself is not unjust." (Anarchy, State and Utopia, p. 283). One should also note that Nozick believes that "the free operation of a market system will not actually run afoul of the Lockean proviso." (Anarchy, State and Utopia, p. 182).)

If we ignore rectification, force and fraud, and the Lockean proviso, Nozick's entitlement theory of justice apparently implies that market outcomes are automatically just. If Jack and Jill are entitled to what they have and make a deal, then they are entitled to what they get as a result. And if Sally had no claim on their holdings before the deal, she has no claim on their holdings afterwards, either.

This intuition concerning justice parallels the intuitions of economists concerning efficiency. If something results from voluntary exchange, then it must be efficient; and (with the right background), it must be just. But such intuitions about the efficiency and justice of voluntary exchange suppose that individuals are affected only by those with whom they voluntarily interact. The image is of autonomous farmers behind their separate fences, who are unaffected by others, except when they choose to interact for their mutual benefit.

But the same complications that sometimes make the results of voluntary exchanges inefficient sometimes make them unjust.<9> Protecting individual rights requires more than preventing force or fraud. The actions and interactions of individuals typically have unintended and often unforeseen spillover effects on the interests of others. Since "capitalist acts between consenting adults"<10> can have unintended consequences that violate the rights of others, Nozick is going to have to restrict some of these. But more important than the particular revisions is the inappropriateness of the entire image. If one looks at the interactions in a just human society as it were through a low resolution telescope, one does not see only voluntary transactions. Externalities are not minor complications, exceptions or qualifications. This whole image of market life and social justice is hopelessly distorted.

3. Externalities and Deontological Rights Theories

For any rights theorist, including a deontological rights theorist such as Nozick, externalities raise questions of justice whenever they infringe on rights. I call Nozick a deontological rights theorist because he takes rights to be absolute side constraints. Rather than taking a rights violation to be a bad consequence of an action that enters with its own finite weight along with other consequences into an assessment of an action, policy or rule, rights violations are simply forbidden regardless of the consequences.<11> Nozick might also justly be called a deontological rights theorist because his derivation of rights is not consequentialist. Rights are not instrumentalities for achieving goods such as autonomy, self-respect, equality or happiness. For if they were, it is hard to see how one could justify taking them to be side constraints.

Externalities pose severe problems for deontological rights theories. According to Nozick, an action is unjust if and only if it violates some right. Both positive and negative externalities may do so, for even though I may enjoy the loud music my neighbor George plays, his playing it without my consent may violate my right to determine whether there will be quiet in my home. But negative externalities will be the main concern. Leaking gasoline storage tanks at the corner gas station that violate rights call for restrictions, punishments or at least compensation.<12> Acceptable deontological rights theories must say what to do about externalities.

There are two serious difficulties that confront such theories. First, as argued effectively by Railton, a deontological rights theory would appear to be too restrictive.<13> Since rights are in Nozick's view "side constraints", which may not be violated in order to secure benefits or even in order to minimize overall rights violations, there would appear to be little I would be free to do even on my own property. Can I build a fire if soot land will next door? Can I use fertilizer if any will leach into the ground water and affect your wells or lakes? Can I build a house, if doing so will limit your sunlight? Can I plant sunflowers if the pollen will make you sneeze? It would seem that I would have to get the consent of my many neighbors at every turn.<14> Real markets will generate conventions to resolve these difficulties, but such conventions cannot set inviolable boundaries on "side constraints," which must be heeded, regardless of welfare costs.

One way to lessen this stringency and to make a natural rights view more plausible would be to insist, as defenders of the doctrine of double effect do, that rights violations are absolutely forbidden only when they are intended. For example, many have held that intentionally killing an innocent person is absolutely morally impermissible. Yet a driver who swerves and kills one pedestrian to avoid killing a whole group may have done nothing wrong, for the killing, although foreseen, was not intended. According to the doctrine of double effect, one may permissibly perform actions such as killing or violating other rights when the following four conditions are met:

1. The act is good in itself or at least indifferent
2. Only the good consequences of the act are intended
3. The good consequences are not the effect of the evil
4. The good consequences are commensurate with the evil consequences.<15>

Defenders of the doctrine of double-effect regard an action as intended when it is either a goal of an agent or a means to an

agent's goal. When a rights violation is merely a side effect of an action, which may or may not be foreseen, but which is not intended, then, depending on factors such as the goodness or badness of the consequences, the action may be permissible. Yet one can continue to regard intentional rights violations as absolutely impermissible.

To invoke the doctrine of double effect raises, however, a mass of difficulties, which I shall not attempt to resolve, since my main objection to a rights approach is not that it is overly stringent. Not only is the doctrine of double-effect itself problematic,<16> but it obviously does not save the view that market outcomes (in the absence of force, fraud and violations of the proviso) are automatically just.<17>

A second difficulty for a deontological rights theorist such as Nozick comes in generating the precise property rights that are needed to regulate people's conduct toward one another. If one sets aside consequentialist accounts of rights, then one is left with an account of natural rights as prerequisites of moral agency or as implications of the equal respect due to moral agents.<18> One might argue, for example, that people cannot be moral agents at all without life, control over their bodies, the ability to move about, and some property.

In order to address the problems raised by externalities within such a framework, one must somehow use the general notions of agency, autonomy, equality and respect to derive detailed and precise specifications of property rights, or one must show how property rights can be legitimately legislated by some legal authority within the constraints of natural rights. For Locke, actual property rights are constrained by natural rights, but are social constructions arising from agreement.<19> Nozick does not and cannot accept this option, for it grants states more authority than Nozick thinks they can legitimately have,<20> and it undermines a purely deontological view of rights. If property rights can be legislated (within the constraints of natural rights), then their content and force depend in part on their consequences. Nozick's insistence that only the history of human interactions, not the resulting patterns or consequences is relevant to justice would collapse. Property rights can only be true side constraints if they are entailed by natural rights.

A deontological rights theory provides guidance concerning when externalities raise questions of justice only if it can make reasonably clear what natural rights to property encompass. I have no compelling argument showing that this task cannot be accomplished. But I cannot see how considerations of moral autonomy will decide what air or mineral rights individuals have or how such a perspective will either prevent or adjudicate rights conflicts.<21>

So it seems unlikely that a deontological rights theory can cope with the problems externalities raise. Such theories seem overly restrictive and unable to specify adequately what the rights are.

4. Property Rights and Pecuniary Externalities

Rights theories that permit the consequences of various rights assignments to influence what rights agents are granted can avoid both of the above difficulties that confront a deontological rights theory. Rights that seem unduly to restrict the freedom of others can be weakened and the problem of deriving property rights from notions of moral agency regardless of consequences never arises. But standard views of property rights run into a different problem.<22> For, it seems that questions of justice can arise even when no rights are violated.

Consider pecuniary externalities. These result from market transactions via the market mechanism itself, and on virtually all theories of property rights violate no rights. So if externalities raise questions of justice only if they violate rights, pecuniary externalities that arise on an otherwise untainted market never raise questions of justice. But why should only non-pecuniary externalities be of moral concern?<23> For pecuniary externalities can totally transform people's lives. It is odd, if not monstrous, to suppose that the power looms that destroyed virtually the whole of the handweaver's capital (both human and otherwise) did those weavers morally significant harm only insofar as the smoke they gave off smudged the hand-woven cloth. Positive pecuniary externalities can eliminate the greatest privations of the past and transform the human condition. Negative pecuniary externalities can force people on pain of starvation to leave their homes, occupations, families, countries and cultures, or indeed they can force people to starvation itself.<24> Can one plausibly accept the view that human actions with such overwhelming impact on the lives of other people raise no questions of justice?<25>

One might try to make this remarkable implication plausible by arguing that by engaging in market interactions, people have consented to provide positive pecuniary externalities and to suffer negative ones. (So by the definition of externality endorsed above, the notion of a pecuniary externality is a contradiction in terms.) It might be contended that however greatly one may be harmed by a pecuniary externality, one has already agreed to play by the rules of the market, and in seeking the benefits of market transactions, one has waived one's right to be protected from the harms markets sometimes cause their participants.<26> Formerly affluent blacksmiths, who were later bankrupted, suffered great harm from the market. But having set themselves up in business for the advantages that market life might bring, they consented to suffer the harms it brought instead. Yet individuals can only be regarded as voluntarily engaged in market relations if they could choose not to be part of market life. Since this is not a real option, the smith's decision to participate in markets seems little more voluntary than is his decision to hand over his wallet to an armed robber.<27>

It seems to me that both pecuniary and non-pecuniary externalities are of moral concern, and thus that the moral questions that externalities raise cannot be resolved merely by the specification of an appropriate set of property rights.

5. Welfare Economics and Externalities

Welfare economists have been concerned almost exclusively with the inefficiencies rather than the injustices to which externalities may give rise. Given externalities, rational self-interested individual action may lead to suboptimal actions. Most economists would maintain, however, that the suboptimalities arise from transaction costs (costs of locating those responsible for and affected by the spillovers and negotiating and enforcing a settlement), not from externalities per se.<28> For example, the two players in a prisoner's dilemma reach a suboptimal outcome because of the transaction costs involved in reaching and enforcing agreement to play the cooperate strategy. Cooperation on the part of each individual results in a positive externality to the other. If there were no transaction costs, these externalities would be "internalized" and the conflict between individual and social good would be avoided.

In the real world there are transaction costs, and the welfare economist is concerned to lower them as far as possible. Within the limited ethics of welfare economics, transaction costs are "the root of all evil."<29> Exactly what to do about transaction costs is not obvious; and, as economists of this last generation in particular have eloquently demonstrated, the remedies of government intervention and regulation may be worse than the maladies.

When there are no transaction costs, then (by Coase's theorem<30>) the outcome is efficient or optimal, and externalities cease to be a normative concern for most economists.<31> The only remaining moral concerns are distributional. Some economists would go on to argue that distributional problems are of no concern in a theory of justice either because they accept some sort of entitlement theory or because they believe that attempting to alleviate distributional problems always has bad consequences on the whole. As is implicit in the discussion of pecuniary externalities in the last section, I do not believe that these views take the interdependencies among human actions seriously enough. Even when there are no missing markets and no inefficiencies, the consequences of externalities can be so grave or wonderful that they must perforce be morally significant.

Most welfare economists would readily concede that society has a legitimate interest in distributive justice, though they would argue that these legitimate concerns about justice or distribution do not require one to interfere with the day-to-day operations of the market. If a certain distribution of income is morally preferrable, it can be achieved either by adjusting the structure of property rights or by transfer payments. For example, in the case of pecuniary externalities on a perfectly competitive economy, efficiency already obtains. The new producers drive out the old because that is, by implication, what the individuals who are related by the market (when weighted by their dollars) want. The change in income distribution and the short-run dislocations may be morally undesirable, but the cure is non-market transfers rather than economic planning or centralized control.

How then should welfare economics be supplemented to deal with these distributional problems? When, as a matter of justice, ought society to compensate those providing positive externalities or those suffering negative externalities? I cannot fully answer this question, but I shall offer the following remarks. Again I shall focus on negative externalities.

1. Since this is, by assumption, a question of justice, not efficiency, one will have to rely on some theory of justice to answer it. Standard rights theories are, as argued above, inadequate to the task. We have no choice, I believe, but to rely on some sort of consequentialist theory, although it may be one in which rights are given significant weight.

2. Any difference in compensation for pecuniary, as opposed to non-pecuniary externalities is in need of justification. There is no general consent argument against compensating for pecuniary externalities (see section ?? above), and the harms and benefits of the one sort of externality may be just as great as those of the other kind. Different responses to the harms of pecuniary as opposed to non-pecuniary externalities need justification in terms (for example) of the different costs and benefits of the compensation plans.

3. Ordinary moral intuitions are divided, if not inconsistent on these matters. Much seems to depend on how much we sympathize with the activities and ways of life of the agents involved. We seem to start with a general sympathy for anyone who loses his or her job and suffers economic and psychological hardship, not as a result of incompetence or misbehavior, but from general market forces. In addition, we admire certain ways of life and skills and support compensation for family farmers or for skilled artisans who are deprived of their living and whose human capital is made valueless. Our intuitions are less clear when the skills are less clear or less clearly valued and when non-human capital rather than human capital is at stake. We feel more sympathy for fisherman driven out of work by water pollution than for owners of fish canneries (perhaps because the former may not be voluntarily playing the market game). One might regret the demise of locally owned banks on the grounds that larger financial institutions care less about the local community, but few feel that the owners of the local banks should be compensated for their competitive failures. Whether to compensate the owners of the local downtown pizzaria driven into bankruptcy by the new Pizza Hut in the Mall seems to have more to do with one's attitude toward malls and restaurant chains than with any general assessment of the justice of compensation for negative externalities.

It is hard to see any definite moral principles at work here, although perhaps there is the following underlying attitude. "Pure capitalists" who are simply attempting to profit by financial dealing or buying and selling (without much care as to what it is they buy and sell) are playing the market game. And if they lose because somebody else is luckier or plays it better, well that is just how the game works. (And they probably will not be completely impoverished in any case.) But workers and self-employed artisans and producers who put themselves into their products may be involved in market relations only because they have no choice. If they suffer badly as a result of market forces, then they deserve some compensation.

4. From a consequentialist perspective the crucial factors determining when the benefits and harms resulting from externalities deserve compensation are the consequences of arranging such compensation. And there are many costs associated with any compensation scheme: administration costs, incentives to engage in inefficient practices (and not to shift to more efficient ones), and detailed inequities resulting from inevitable imperfections in compensation methods.<32> On the other hand, both pecuniary and non-pecuniary externalities can cause great suffering, and the greater that suffering the more individuals will resist economic changes and the slower the pace of technological progress.<33> Although hardly a rigorous argument, these considerations constitute a plausible consequentialist case for welfare-state policies of high employment, generous unemployment insurance, and good job retraining programs rather than attempting to compensate on a case-by-case basis. Such a policy could also be supported by contractualist theories.

6. The Deeper Challenge of Externalities

An apparently plausible mainstream view of the moral challenges raised by externalities thus runs roughly as follows. Both via the market and apart from the market the actions and decisions of individuals and groups have broad consequences for others. Some of those consequences create no costs or rewards for the decision makers and may lead to inefficient outcomes. With a well designed system of property rights, many of these non-pecuniary externalities can be "internalized", but, given the ubiquity of transaction costs, many remain, and in any case pecuniary externalities will always be with us. The harms and benefits associated with externalities are of prima facie moral significance. Suboptimal outcomes that result from transaction costs may justify government regulation and interference, and the distributional consequences of externalities may justify income transfers. Given the costs and benefits of the externalities themselves, the costs of compensation, and the benefits of reducing resistance to technological innovation, the frictional problems caused by externalities are best handled (if at all) by standard welfare-state policies.

Yet it seems to me that this construction, although it grudgingly recognizes that the consequences of individual exchanges may be morally problematic, evades the problem. The real problem with all the above moral perspectives lies in the vision of market life (and perhaps even social life as a whole) as nothing but the means by which individuals pursue their own objectives. For through market life we collectively shape human existence. We not only make expensive, difficult or impossible the satisfaction of some given preferences (and cheap, easy or inevitable the satisfaction of others), but over time we remake our own and our children's and grandchildren's preferences. For example, the political process in our nation and the citizens who participate in it are very different from those of a century or two earlier, and the differences are largely unintended consequences of market activities of individuals. In the absence of transaction costs and of endogenous preference changes, it could be argued that Americans are getting the changes in their political system Americans (weighted by their dollars) want, but only in fairyland are there no endogenous preference changes and no transaction costs. And there is also no reason to believe that the political decisions implied by dollar-weighted and relatively unreflective consumption choices of televisions, automobiles, or snow mobiles coincide with the political decisions Americans would make after careful and explicit reflection on the political alternatives. Similarly when some individuals buy and sell blood, they may convert a blood donation from a gift of something literally priceless--a gift that can be regarded as the gift of life itself--into the gift of the equivalent of $75. The results may be blood shortages, increased spread of diseases, higher medical costs and a general undermining of altruism and solidarity.<34> Optimistically one can hope that all is for the best, or pessimistically one can lament that human frailty combined with the problems of uncertainty and the difficulties of collective decision making would be sure to do even worse. But the standard view does not even recognize that there is anything to be optimistic or pessimistic, confident or worried about.<35>

What is needed is a view of market processes that sees voluntary exchange among individuals as an integral part of a series of complex social linkages. One must reject the picture of an ideal market in which individuals always or even typically bump up against one another only by choice. One must not ignore spillovers, and one should not wall them off as externalities, from which one typically abstracts and which (apart from transactions costs) one deflates to merely frictional and distributional issues. Externalities should be at the center of the picture in any but narrow short-run analyses. Externalities are the normal case. Pure voluntary exchanges without spillovers are the exception.

All of the perspectives we have examined are deficient, because they give center stage to the exceptional case of independent action, free of all spillovers. In addition the deontological rights theorist cannot provide the needed specification of property rights and cannot find room for the sort of individual freedom that motivates the rights theorist's whole enterprise. More reasonable rights theorists avoid these difficulties but must deny that harms, no matter how grave they might be, raise problems of justice if no rights are violated. The defender of standard welfare economics avoids these difficulties, but at the cost of exaggerating the importance of questions of efficiency. But none of these perspectives bring to center stage the interdependencies of human choice and action and the stark reality that whether we want to or not, we are making the future, and we can collectively attempt to do so consciously and rationally.

Perhaps the pessimist is right and self-consciousness about the way we are shaping ourselves and our futures will make life worse, but without a more adequate treatment of the generalized consequences of market decisions, even this pessimism cannot be justified. And, in any event, there is reason to seek the truth, even when the truth is painful and dangerous. When Jack and Jill make a deal, we all may come tumbling after; and I'd rather know what is coming and have a voice in some of their machinations than pretend that only their interests are at stake in their deals.

Appendix: Externalities

In Marshall's classic early discussion of "external economies", he stressed how the development of the economy could affect the prospects of firms by making available better information, better trained workers, new machinery and techniques, better transportation, and so forth.<36> There are also external "diseconomies". For example, increased demand by other firms for firm F's inputs might increase F's costs. External economies and diseconomies are decreases or increases in F's costs, which are caused by agents outside F either through the market or via some other causal route. Generalizing, one can take an externality to be any market action by A that affects firm F, which is not a voluntary interaction between A and F.<37> are the external economies experienced by consumers.) Externalities in this sense are a significant economic factor, for A's actions create a cost or benefit that has no influence on A's choice. Pursuing what is best for himself or herself and eschewing force and fraud, A may still be causing harm.

A second view of externalities finds classic expression in Part 2, ch. 8 of A. C. Pigou's The Economics of Welfare,<38> although Pigou does not use the words "externality" or "external". Pigou is concerned about circumstances in which there is a divergence between "marginal social net product" and "marginal private net product." In competitive circumstances, the source of this divergence is, in Pigou's language "that, in some occupations, a part of the product of a unit of resources consists of something, which, instead of being sold by the investor, is transferred, without gain or loss to him, for the benefit or damage of other people."<39> This definition limits the perpetrators of externalities to producers, but otherwise seems to capture the same broad class of effects that Marshall touched on. But Pigou is not interested in the whole broad class of externalities or external effects. He is only concerned with cases in which marginal social and private net product diverge and in which, consequently, there is a prima facie case for government intervention. Thus, he excludes cases that Marshall would have included and that fit Pigou's own definition, such as the introduction of new production processes that depreciates old machinery. For, Pigou argues, the benefits that such innovations bring to consumers compensates for the losses that they bring to owners of the old machinery. But, even if the injury is compensated by benefits elsewhere, there is no question that owners of the old machinery are injured.

If one consults contemporary works for a definition of externalities, one will find many. In the introduction to a recent collection, Theory and Measurement of Economic Externalities, Steven Lin writes, "Externalities arise whenever the value of an objective function, for example the profits of a firm or the happiness of an individual, depends upon the unintended or incidental by-products of some activity of others."<40> Notice that according to this definition, externalities must be "unintended or incidental by-products."

In the first essay in Lin's anthology, however, Walter Heller and David Starrett point out that Lin's definition would make exchanges in a barter economy externalities. Heller and Starrett suggest instead, "We define an externality to be a situation in which the private economy lacks sufficient incentives to create a potential market in some good and the non-existence of this market results in losses in Pareto efficiency."<41> So according to Heller and Starrett's definition, unlike Lin's or Pigou's or Marshall's, a necessary condition for the existence of an externality is suboptimality or inefficiency.

Like Heller and Starrett, Jules Coleman stipulates that externalities involve inefficiencies, but like Lin he stresses that externalities involve byproducts and interdependencies of objective functions, "External effects are byproducts of an activity that influence the production of other goods or the welfare (or utility) of other individuals....Externalities are inefficient external effects--social costs or benefits that result in inefficient production or nonoptimal distributions of welfare."<42>

Kenneth Arrow in "The Organization of Economic Activity," and Harold Demsetz in "Towards a Theory of Property Rights" define externalities, like Lin and so many others, as involving interdependencies in objective functions, but they also insist, like Heller and Starrett, that externalities involve missing or imperfect markets. They do not, however, make it part of their definitions that externalities lead to inefficiencies. For Demsetz externalities are more or less coextensive with limitations in markets:

What converts a harmful or beneficial effect into an externality is that the cost of bringing the effect to bear on the decisions of one or more of the interacting persons is too high to make it worthwhile, and this is what the term shall mean here.<43>

Notice that Arrow's and Demsetz's view is not equivalent to Pigou's insistence on a divergence between marginal private and social net product, although it is a necessary condition for it. For in the case of an innovator who depreciates the machinery currently in use (which would be an externality on Arrow's and Demsetz's views) neither the harm to the owners of the old machinery nor the consumer's surpluses are brought to bear on the decision, but since these cancel, there is no divergence between marginal social and private net product.

For most economists, the notion of an externality has been important in the analysis of ways in which markets can be inefficient, and some economists have consequently insisted that an externality must, by definition, lead to inefficiency. But what is crucial about spillovers from the perspective of justice is that A's actions on markets (and off) have effects on individuals besides those engaged in any voluntary transaction with A. The focus is on those affected by A's actions. To insist that all externalities involve inefficiencies as Coleman, Heller and Starrett and many others do, would exclude cases that appear to be of moral concern. For example, according to their definitions, a case of a firm polluting a stream and poisoning a community involves no externality, if all the feasible outcomes without the pollution would make the owners of the firm worse off. Insisting that externalities involve suboptimality also raises the tricky question of the benchmark against which the actual outcome is judged to be suboptimal.<45> Nor does it seem desirable to insist that externalities always involve divergence between marginal private and social product. For the definition of an externality should not itself determine that so-called "pecuniary" externalities, such as the depreciation of old machinery caused by the introduction of new technologies are never of moral significance.

Arrow's and Demsetz's definitions are better suited for purposes of this paper, for difficulties in bringing the costs to B of A's actions to bear on A arise in roughly the same class of cases as those in which A's actions affect B even though A and B are not engaged in a fully voluntary transaction. But neither their definitions nor the simpler definition of an externality as any effect of agent A on agent B which is not voluntarily consented to by both A and B will do. For these definitions make feeding or conceiving my two-year-old or arresting a suspected murderer externalities. Perhaps these difficulties could be addressed by insisting that externalities apply only to market activities, but in the context of studying the implications for justice of spillovers, this seems arbitrary. Alternatively, one might respond to this difficulty by adding a clause to the definition: An externality is any effect of agent A on agent B which is neither explicitly permitted by some legitimate social norm nor voluntarily consented to by both A and B. But the addition is a disaster, since it makes the notion of an externality depend on a prior determination of what is morally permissible and implies that paradigm cases of externalities, such as overgrazing a commons or polluting a stream, are not externalities at all if they are explicitly permitted by a legitimate social norm.

As I argue in the text, the notion that is of particular interest with respect to questions about the justice of voluntary transactions seems to be of an unintended effect of A on B to which A or B have not consented.

Notes

* This paper arose out of an interesting discussion in August, 1989 with members of the philosophy department at the University of Sydney, for which I am very grateful. Maurice Lagueux and Michael McPherson provided essential assistance in the early stages of drafting this essay. Andrew Levine and Julius Sensat provided invaluable criticisms of earlier drafts. Comments on the penultimate draft by Richard Arneson, David Copp, Gregory Kavka and Ellen Paul were particularly helpful. Back to text

1. Kenneth Arrow "The Organization of Economic Activity: Issues Pertinent to the Choice of Market versus Nonmarket Allocation," Collected Papers, vol. 2, General Equilibrium (Cambridge, MA: Harvard University Press, 1983), p. 133. Back to text

2. Harold Demsetz, "Toward a Theory of Property Rights," American Economic Review, vol. 57 (1967), pp. 347-59. Back to text

3. Steven Lin, ed., Theory and Measurement of Economic Externalities (New York: Academic Press, 1976); T. Bergstrom, "The Use of Markets to Control Pollution," Recherches Economique de Louvain, (1973); James Buchanan and C. Stubblebine, "Externality," Economica, New Series, vol. 29 (1962), pp. 371-84; S. Cheung, S, "Transaction Costs, Risk Aversion and the Choice of Contractual Arrangements," Journal of Law and Economics, vol. 12 (1969), pp. 23-42; Ronald Coase, "The Problem of Social Cost," Journal of Law and Economics, vol. 3 (1960), pp. 1-44; Harold Demsetz, "The Exchange and Enforcement of Property Rights," Journal of Law and Economics, vol. 7 (1964), rpt. in Tyler Cowen, ed., The Theory of Market Failure: A Critical Examination (Fairfax, VA: George Mason University Press, 1988), pp. 127-45 and "Some Aspects of Property Rights," Journal of Law and Economics, vol. 9 (1966), pp. 61-70; A. Freeman, Robert Haveman, and A. Kneese, The Economics of Environmental Policy (New York: Wiley, 1973); Ezra Mishan, "Reflections on Recent Developments in the Concept of External Effects," Canadian Journal of Political Economy, vol. 31 (1965), pp. 3-34; David Starrett, "Fundamental Nonconvexities in the Theory of Externalities," Journal of Economic Theory, vol. 4 (1972), pp. 180-99. Back to text

4. Kenneth Arrow, "The Organization of Economic Activity...," pp. 133-55 Back to text

5. For example, Jules Coleman, "Efficiency, Auction and Exchange," rpt. in Markets, Morals and the Law (Cambridge: Cambridge University Press, 1988), pp. 67-94. Back to text

6. Charles Fried, "Difficulties in the Economic Analysis of Rights," in Gerald Dworkin, B. Gordon, and Peter Brown, eds., Markets and Morals (New York: Wiley, 1977), p. 192 equates externalities with "unintended impositions incidental to the pursuit of some other end." See also the quotations in the appendix from Lin and Coleman. Back to text

7. Robert Nozick, Anarchy, State and Utopia (New York: Basic Books, 1974), p. 151. Back to text

8. Nozick sometimes suggests that all voluntary transfers of what one is entitled to are just, as when he asks the rhetorical question (with respect to his famous Wilt Chamberlin example), "By what process could such a [voluntary] transfer among two persons give rise to a legitimate claim of distributive justice on a portion of what was transferred, by a third party who had no claim of justice on any holding of the others before the tranfer?" (Anarchy, State and Utopia, pp. 161-62 Back to text

9. See Thomas Scanlon, "Liberty Contract, and Contribution," in Dworkin et al., Markets and Morals, p. 44. Back to text

10. Anarchy, State and Utopia, p. 163. Back to text

11. Anarchy, State and Utopia, pp. 28f. Back to text

12. Such leaks are actionable at law, and if the owners of the perhaps defunct gas station are solvent, alive and locatable, then our system of tort law may provide proper compensation. But the point remains either that actions involving no force or fraud can violate (or can risk violating) rights or that spillovers will typically involve a pervasive "force and fraud". Back to text

13. Peter Railton, "Locke, Stock and Peril: Natural Property Rights, Pollution and Risk," in Mary Gibson, ed. To Breathe Freely (Totowa, NJ: Rowman and Allenheld, 1985), pp. 89-123. Railton also raises serious difficulties concerning risks that I shall not discuss. Judith Thomson in "Some Ruminations on Rights," in Jeffrey Paul, ed., Reading Nozick: Essays on Anarchy, State and Utopia (Totowa, NJ: Rowman and Littlefield, 1981), pp. 130-47 raises serious questions about whether Nozick can really be committed to such an implausibly stringent view of rights (see esp. pp. 136-38). Back to text

14. I would fully endorse Railton's conclusion,

...if we take seriously the fact that we find ourselves situated in and connected through an environment, we are soon impressed with the inaptness of a conception of morality that pictures individuals as set apart by propertylike boundaries, having their effect upon one another largely through intentional action, limiting their intercourse by choice, and free to act as they please within their boundaries, although absolutely constrained by them. ("Locke, Stock and Peril...," p. 119) Back to text

15. James Sterba, "Abortion and Euthanasia: Basic Concepts," in James Sterba, ed., Morality in Practice, 2nd. ed. (Belmont, CA: Wadsworth Publishing, 1988), p. 135. See also Elizabeth Anscombe, "War and Murder" and "Mr. Truman's Degree" in her Ethics, Religion and Politics(Minneapolis: University of Minnesota Press, 1981), pp. 51-71; Elizabeth Anscombe, "Medalist's Address: Action, Intention, and 'Double Effect'," Proceedings of the American Catholic Philosophic Association, vol. 56 (1982), pp. 12-25; Joseph Boyle, Jr., "Toward an Understanding of the Principle of Double Effect in Ethics," Ethics, vol. 90 (1980), pp. 527-37; and Warren Quinn, "Actions, Intentions, and Consequences: The Doctrine of Double Effect," Philosophy and Public Affairs, vol 18 (1989), pp. 334-51. Back to text

16. I regard it as an ad hoc manuever designed to reconcile the conviction that there are absolute moral prohibitions on actions such as killing people with the fact that, depending on the circumstances, all of these supposedly prohibited actions may be morally permissible after all. For the most plausible recent defense, see Warren Quinn, "Actions, Intentions, and Consequences." I am here stretching the doctrine of double effect, which is designed to permit one to violate absolute prohibitions in exceptional circumstances when some great harm might be avoided or some great good achieved. So it may be that a deontological rights theorist would reject this way of answering the objection that a deontological rights theory is too restrictive. But these are problems for deontological rights theorists such as Nozick, not for me. Back to text

17. The doctrine of double effect might help to make sense of Nozick's puzzling remarks concerning risk and compensation in Ch. 4 of Anarchy, State and Utopia (see Eric Mack, "Nozick on Unproductivity: The Unintended Consequences," in Reading Nozick, pp. 169-90). Consider also the following brief discussion of pollution, which seems inconsistent with Nozick's insistence that rights are side-constraints:

Since it would exclude too much to forbid all polluting activities, how might a society (socialist or capitalist) decide which polluting activities to forbid and which to permit? Presumably, it should permit those polluting activities whose benefits are greater than their costs, including within their costs their polluting effects....A society must have some way to determine whether the benefits do outweigh the costs. Secondly, it must decide how the costs are to be allocated. It can let them fall where they happen to fall: in our example, on the local homeowners [harmed by airport noise]. Or it can try to spread the cost throughout the society. Or it can place it on those who benefit from the activity: in our example, airports, airlines, and ultimately the air passenger. The last, if feasible, seems fairest. (Anarchy, State and Utopia, pp. 79-80)

That property rights are inviolable seems to have been forgotten. Indeed it is not clear whether rights have any special weight here at all. One would have thought that if homeowners have a right to quiet and to the airspace over their homes, then an airport could be operated only with the unanimous consent of (and consequently full compensation to) the homeowners whose airspace is violated and whose quiet is disturbed. Notice that airspace violations are a means to the end of air travel, not a side-effect, and cannot be defended via the doctrine of double effect. Back to text

18. or as protections against various practical threats to moral agency or autonomy. Such practicality grounds may or may not be available to a natural rights theorist such as Nozick. I am here influenced by the excellent discussion of the basis of rights in James Griffin, Well-Being: Its Meaning, Measurement and Moral Importance (Oxford: Oxford University Press, 1986), ch. 11. See also Charles Fried, "Difficulties in the Economic Analysis of Rights." Back to text

19. See John Christman, "Can Ownership be Justified by Natural Rights?" Philosophy and Public Affairs, vol. 15 (1986), pp. 156-77, who emphasizes the importance for Locke of the consent involved in the use of money in enlarging property rights. Back to text

20. Yet Nozick does allow that considerations of welfare may have a role in determining "otherwise arbitrary features" of "a statute" (Anarchy, State and Utopia, p. 153n), and he at one point writes, "Perhaps the precise contour of the bundle of property rights is shaped by considerations about how externalities may be most efficiently internalized..." (Anarchy, State and Utopia, p. 280). Back to text

21. In his notorious essay, "The Problem of Social Cost," Coase was, among other things, concerned to show that responsibility for an externality is always shared and that the positions of the parties is fully symmetrical. Thus Carl Dahlman in "The Problem of Externality," Journal of Law and Economics, vol. 22 (1979), pp. 141-62; rpt. in The Theory of Market Failure, pp. 209-34, for example, writes, "Perhaps the real significance of the court cases cited by Coase is that the distinction between emittor and recipient of an externality is irrelevant: what matters is whether we achieve a higher-valued output by putting the liability on one or the other of the parties involved, and not who is the "source" of the externality." (p. 230) This is one point upon which "conservative" or "neo-liberal" economists, who stand opposed to government interference in the market are deeply at odds with libertarians such as Nozick. See also Charles Fried, "Difficulties in the Economic Analysis of Rights." Back to text

22. I am not maintaining that the following is a problem for all possible property rights theories, for I suspect that the conclusions of virtually any acceptable moral theory can be echoed in some sort of rights theory. Back to text

23. Harold Demsetz, in "The Exchange and Enforcement of Property Rights," p. 144 also questions why economists should have a different attitude toward pecuniary externalities, though he is concerned to argue for the moral demotion of non-pecuniary externalities, rather than the importance of pecuniary externalities. Tibor Scitovsky in "Two Concepts of External Economies," Journal of Political Economy, vol. 62 (1954), pp. 70-82 and and Martin Shubik Shubik, "Pecuniary Externalities: A Game Theoretic Analysis," American Economic Review, vol. 61 (1971), pp. 713-18 explain the lack of interest in pecuniary externalities by showing how they disappear in static general equilibrium, which is the standard theoretical perspective employed by economists. Back to text

24. See Amartya Sen, Poverty and Famines: An Essay on Entitlement and Deprivation (Oxford: Oxford University Press, 1981). Back to text

25. A critical reader objected that I should argue that mass starvation raises questions of morality and justice rather than just appealing to intuition. But in my view one is very close to the foundations of morality here. I doubt that there are premises to be had that are more secure than is this "intuition." Back to text

26. Just as, in Locke's view, the consent involved in the use of money justifies inequalities in riches. See his Second Treatise of Government, ch. 5. Back to text

27. On Nozick's view of what is voluntary (Anarchy, State and Utopia, esp. p. 263), the smith's decision would be voluntary. For on Nozick's view, a choice is voluntary provided that no rights are violated in limiting the alternatives. But the definition assumes that there can be no injustice done unless rights are violated, and it thus begs the question at issue here. Back to text

28. This claim is probably more a definition of transaction costs than a substantive truth. It is also dubious. See Kenneth Arrow, "The Organization of Economic Activity," pp. 141-42 and Donald Regan, "The Problem of Social Cost Revisited," Journal of Law and Economics, vol. 15 (1972), pp. 427-38. Back to text

29. Carl Dahlman, "The Problem of Externality," p. 210. Back to text

30. See Ronald Coase, "The Problem of Social Cost." Back to text

31. "Proportional transaction costs do not generate Pareto-relevant externalities, but only the trivial Pareto-irrelevant variety." (Carl Dahlman, "The Problem of Externality," p. 214). "If it costs smog breathers too much to set up the exchange that induces the emittors to reduce the outpour of pollutants, then it follows necessarily that, the funny smell notwithstanding, the optimal level of pollution has been achieved." (Carl Dahlman, "The Problem of Externality," p. 216). The last "the" is indefensible. Back to text

32. John Stuart Mill, On Liberty (rpt. Indianapolis: Bobbs-Merrill, 1965), pp. 114-15 makes the following remarks about a subset of these harms:

In many cases an individual, in pursuing a legitimate object, necessarily and therefore legitimately causes pain or loss to others, or intercepts a good which they had a reasonable hope of obtaining....Whoever succeeds in an overcrowded profession...reaps benefit from the loss of others, from their wasted exertion and their disappointment. But it is, by common admission, better for the general interest of mankind that persons should pursue their objects undeterred by this sort of consequences. In other words, society admits no right, either legal or moral, in the disappointed competitors to immunity from this kind of suffering,..." Back to text

33. Thus, other things being equal, labor-saving technological change should be more rapid in those countries with better benefits for those whose jobs are threatened by such changes. One might regard this implication as refuted by the abysmal rate of technological innovation in Eastern Europe prior to 1989. But other things were not equal. In support of the claim, on the other hand, one might point to the coupling of job security and the rapid and relatively conflictless introduction of labor-saving technological changes in Japan. But other things are not equal there, either. The purported link between compensation to those harmed by technological progress and the rate of such progress is as plausible as most generalizations economists offer, but it is neither proven nor indisputable. Back to text

34. See Richard Titmuss, The Gift Relationship: From Human Blood to Social Policy (London: Allen and Unwin, 1971), Kenneth Arrow, "Gifts and Exchanges," Philosophy and Public Affairs, vol. 1 (1972), pp. 343-62 and Peter Singer, "Altruism and Commerce: A Defense of Titmuss against Arrow," Philosophy and Public Affairs, vol 2. (1973), pp. 312-20. Back to text

35. See also Kenneth Arrow, "The Organization of Economic Activity," p. 152; Thomas Scanlon, "Liberty Contract, and Contribution." p. 45; and especially Albert Hirschman, "Against Parsimony: Three Easy Ways of Complicating Some Categories of Economic Discourse," Economics and Philosophy, vol. 1 (1985), pp. 16-19. Back to text

36. Alfred Marshall, Principles of Economics, 8th. ed. (London: Macmillan, 1920; rpt. 1982), esp. pp. 237-39. Back to text

37. Consumers are affected, too, although they do not, in Marshall's terminology, suffer external effects. "Consumer surpluses" (Principles of Economics, Bk. III. ch. VI. Back to text

38. 2nd. ed. (London: Macmillan, 1924). Back to text

39. The Economics of Welfare, p. 152. Back to text

40. p. 1. Lin then cites the references mentioned above in footnote ??. Back to text

41. Walter Heller and David Starrett, "On the Nature of Externalities," in Steven Lin, ed., Theory and Measurement of Economic Externalities, p. 10. Back to text

42. Markets, Morals and the Law, p. 76. At this point Coleman includes a footnote maintaining that "The distinction between externalities and external effects is widely misunderstood even in elementary economics texts,..." (note 18, p. 353). Maurice Lagueux in "Neoclassicism and Neoliberalism on Externalities," (forthcoming) remarks more justly that "All of the following expressions (or equivalent expressions) were widely used to define externalities: "lack of appropriate property rights," "market failure to reach optimality," "interdependencies among consumption or production functions," "side (or third party, or spillover) effects," impossibility of exclusion," "presence of (wholly or partially) unpriced inputs or of uncompensated services," "situation leaving room for a free rider (or free loader) effect," "absence of appropriate negotiations due to excessive transaction costs.""(p. 1) Back to text

43. "Towards a Theory of Property Rights," p. 343. Back to text

44. "The Organization of Economic Activity," pp. 145-48. Back to text

45. See Guido Calabresi, "Transaction Costs, Resource Allocation and Liability Rules--A Comment," Journal of Law and Economics, vol. 11 (1968), p. 69. Carl Dahlman in "The Problem of Externality," p. 222, argues that the fact that the actual situation is suboptimal relative to a situation without transaction costs is no more relevant than is the suboptimality of a world with costly apples relative to a possible state of affairs in which apples could be produced costlessly. In his view there are no such externalities and that the whole concept depends on the moral judgment (which he clearly does not share!) that government regulation is better than market interaction.