A liberal theory of externalities?

Abstract

Unlike exploitative exchanges, exchanges featuring externalities have never seemed to pose particular problems to liberal theories of justice. State interference with exchanges featuring externalities seems permissible, like it is for coercive or deceptive exchanges. This is because exchanges featuring negative externalities seem to be clear cases of the two exchanging parties harming a third one via the exchange—and thus of conduct violating the harm principle. This essay aims to put this idea into question. I will argue that exchanges featuring negative externalities are not unjust in this straightforward way, i.e. because they would constitute an instance of wrongfully causing or risking a bodily or material harm. In fact, unless we are subscribing to particularly demanding variants of liberalism—e.g. perfectionist liberalism—or unless we are exclusively focusing on borderline cases of externalities—i.e. of effects of exchanges hardly to be called externalities—there is no liberal theory of how exchanges featuring externalities are unjust.

1 Introduction

Unlike exploitative exchanges, exchanges featuring externalities have never seemed to pose particular problems to liberal theories of justice. State interference with exchanges featuring externalities seems permissible, like it is for coercive or deceptive exchanges. This is because exchanges featuring negative externalities seem to be clear cases of the two exchanging parties harming a third one via the exchange—and thus of conduct violating the harm principle. The argument which follows aims to put this idea into question. I will argue that exchanges featuring negative externalities are not unjust in this straightforward way, i.e. because they would constitute an instance of wrongfully causing or risking a bodily or material harm. In fact, unless we are subscribing to particularly demanding variants of liberalism—e.g. a perfectionist liberalism as proposed by Raz (1986)—or unless we are exclusively focusing on borderline cases of externalities—i.e. of effects of exchanges hardly to be called externalities—there is no liberal theory of how exchanges featuring externalities are unjust.

I shall first define externalities and argue what would have to be true for exchanges featuring externalities to generally be unjust from a liberal perspective (Sect. 2). Namely, as other standard routes are blocked, they would have to constitute conduct violating the harm principle. However, I find that exchanges do not relate to externalities in the right way to meet this criterion. They neither wrongfully cause nor risk externalities (Sect. 3), nor wrongfully enable them (Sect. 4), which would be sufficient for violating a broader version of the harm principle. Section 5 highlights consequences of the lack of a liberal theory of externalities. If exchanges featuring externalities are not unjust, this impairs a liberal state’s ability to interfere with these exchanges. There might still be other ways, though, to interfere with the emergence of externalities.

2 Externalities and harm

Externalities are first and foremost an economic concept. They are one of the classic cases of market failure—i.e. when externalities are present, voluntary exchanges do not give rise to an efficient market equilibrium. Economists speak of an externality in situations in which the production or consumption of a certain good by an agent either confers benefits or imposes costs on others which are not accounted for in the market price of the good. More specifically, production (consumption) externalities are present when the production (consumption) of a certain good has external effects. Negative (positive) externalities occur when the production or consumption imposes costs (confers benefits) on others. “Externality” hereby refers to the costs or benefits created, not to the actions of consumption or production. This definition is widely agreed on (e.g. Mankiw 2004, 11; Pindyck and Rubinfeld 2013, 306; Varian 2006, 329). In the following, I will focus on cases of negative externalities as these seem to be more pressing with respect to questions of justice.

A classic example of a negative production externality is that of a pharmaceutical company dumping its polluted effluents in a nearby river without having to pay for this, thereby negatively affecting the quality and quantity of fish some fishermen catch downstream. While the pharmaceutical company and its customers exchange the company’s goods, e.g. a certain drug, on the market—with the customers compensating the company for its production costs (not including any disposal costs) in the form of paying the drug’s price—the fishermen incur costs that nobody compensates them for. Both aspects are essential for externalities. That they (i) impose costs on people external to an exchange which (ii) are not accounted for in the price of the good exchanged.

Note that many costs to third parties are included in an exchanged good’s price. Say the production of the drug consumes certain raw materials. Then no third party is able to use these very raw materials for alternative purposes. But in that case third parties, notably the prior owner of those raw materials, are in fact compensated for this (opportunity) cost, because the pharmaceutical company pays a price for acquiring the raw materials. These costs are paid by the company (as part of the production costs), and eventually by the buyers of the drug (as the market price includes and covers production costs in market equilibrium).

However, when the pharmaceutical company calculates its costs of production, and sets the price of its goods accordingly, it typically only takes into account those costs it itself actually incurs; i.e. its private or internal costs. What the pharmaceutical company does not consider when making its pricing decisions are the costs others have to bear because of its production, but for which it does not have to pay anything. Because the disposal of the polluted effluents does not cost anything in our scenario, these costs are not reflected in the producer’s decision of how much to supply to the market—and thus they are not reflected in the market price either.

Figure 1 illustrates the effect a production externality has on a market. It is partly a typical supply-and-demand diagram featuring a market equilibrium at an equilibrium price (p*) and an equilibrium quantity traded (q*) at the point where demand D and the pharmaceutical company’s supply curve S intersect.

Fig. 1

A negative production externality

However, Fig. 1 also depicts what the supply curve of the pharmaceutical company would look like, if it also took into account the external costs imposed on the fishermen (S _ext). With respect to S, S _ext is shifted upwards. This is because with S _ext we try to depict a situation in which everything stays the same, but the producer faces higher costs of production (because he also compensates the fishermen for the water consumed, i.e. polluted, in the production process). In this case he ceteris paribus would charge a higher price for each unit of output, which is just what an upward shift of the supply curve illustrates. From the intersection of D and S _ext we can derive the market price and the quantity traded which would obtain if the externality were accounted for (p _ext and q _ext). Generally speaking, we see that whenever negative production externalities are involved, the uncorrected market equilibrium features too high a quantity of the good traded at too low a price.

Given this understanding of externalities as the uncompensated-for costs certain exchanges impose on third parties, there seems to be an intuitive way for liberals to account for how they are unjust. For if matters of justice are those which are by definition enforceable (Mill [1861] 1991b, Utilitarianism, chap. 5, para. 13–40; Skorupski 2010, 351–359)—i.e. those particular moral matters the state (or a similar force-wielding political organization) is permitted to call on force to address them—then we can rely on the harm principle in order to argue why we are permitted to call on force to address exchanges featuring externalities, i.e. why they are unjust. Intuitively, we can unproblematically conceive of the uncompensated-for costs as harms. Think of the material harm of foregone revenue the fishermen incur if they catch less fish. Or think of the bodily harm they incur if they eat the polluted fish themselves. Thus, if the exchanges of drugs for money between the pharmaceutical company and its customers wrongfully cause or risk these harms, it seems like these exchanges qualify as conduct which violates the liberal tenet of the harm principle in its traditional or modern form (cf. Mill [1859] 1991a, On Liberty, chap. 1, para. 9; Feinberg 1984, 105–106, 1988, 176–210). The focus on harms done and causation of this proposed account bears witness to the fact that we are concerned with evaluating the justice of human conduct, namely exchanges featuring externalities, rather than of end states. This is not to say that externalities like, for example, global warming might not also constitute a problem from the standpoint of distributive justice.

Importantly, the two most prominent considerations for why such alleged violations of the harm principle by exchanges featuring externalities might not be wrongful—and thus not truly violations (just like the harming of a patient by a doctor)—do not apply. On the one hand, exchanges featuring externalities are not what one might call market harms (Thomson 1986, 160). It is true that many harms which arise out of competition, e.g. of not making a deal because a competitor undercuts you, might constitute real harms for the defeated competitor. Even if the loss of the competitor is only an external effect of the winner’s winning, a harm is indeed inflicted. However, such market harms are not instances of wrongful harming but a classic example of harms that do not warrant interference by the society—as long as the competition was fair—because they are a natural consequence of competition itself (Mill [1859] 1991a, On Liberty, chap. 5, para. 3). On the other hand, the maxim volenti non fit iniuria does not apply, i.e. the idea that one is not wronged by conduct to which one consented. Those on whom externalities are imposed typically do not consent to this harm. Often, they are not even asked.

Therefore, instead of dealing with market harms or with harms which are subject to the Volenti maxim, we are dealing with cases in which we can consider the conduct causing externalities to be wrongful precisely because it is harmful (Kagan 2011, 113–132). That is, the wrongfulness constraint of the harm principle is met (Feinberg 1984, chap. 3). The use of force is permissible, and we may consider such conduct unjust.¹

Note that another characteristically liberal way of accounting for the injustice of externalities—namely by focusing on rights violations—seems blocked; at least if we aim for a general theory of why exchanges featuring externalities are unjust. This is because, whereas externalities as uncompensated-for external costs of exchange have a quasi-definitional relationship to harm done, the same is not true for rights violations. That is, there is no inherent relationship between externalities and rights violations. It is true that, if the fishermen had a right to unpolluted water, then the later exchange of the pharmaceutical company’s drugs might be tainted by this prior rights violation in the same way that exchanges of stolen organs are. In both cases we could say that the exchanges are unjust from a liberal perspective because there was a prior rights violation. That is, we could rely on the same strategy Steiner (1984) uses when proposing his liberal theory of exploitation. But not all instances of externalities are like this. Notably, we still face an externality if the fishermen have no right to unpolluted water but suffer from the pollution.² So the suggested approach of arguing that exchanges featuring externalities are unjust as they violate the harm principle remains the more promising and notably more general one.

Now, as mentioned above, I do not think this approach succeeds. Exchanges do not relate to externalities in the right way to make it work. But before I turn to my argument for why that is, it is important to outline what my argument in the following will not rest on. Notably, it will not rest on the fact that many of today’s most pressing externalities seem to be cases of threshold harms or collective harms.³ In the case of threshold harms or collective harms it is more complex to evaluate an individual agent’s actions—and one might come up with an argument that such cases actually do not constitute violations of the harm principle. I do not go down this route for two reasons. First, because questions of harm thresholds and collective harms are not inherently linked to question of externalities. Some externalities are best considered threshold or collective harms, but not all of them are. So the question of how we should conceive of the injustice of externalities to some extent must be answered without pointing to such cases. Second, because I am confident that even in such cases there are ways to conceive of the individual agent’s actions as wrongfully harmful or wrongfully risking a harm. Kagan (2011), for one, gives a powerful argument to this effect.

3 Exchanges, causes, and enabling conditions

Even if we can qualify conduct creating externalities as wrongfully harmful for a wide range of different kinds of externalities, I shall devote the next two sections to arguing that, as exchanges do not wrongfully cause or risk externalities but merely enable them, exchanges featuring externalities should not be deemed unjust. I will first establish the distinction between causes and enabling conditions, then mention three objections to it—the most pressing of which I will address separately in the next section.

Consider that regulating market exchanges by means of introducing a tax can be an effective means to combat externalities. Looking at Fig. 2, we can see how, say, a Pigovian sales tax on deodorants containing CFCs (which destroy the ozone layer) can alleviate the problem of market prices being “too low” and of the quantity traded being “too high”.

Fig. 2

The effect of a Pigovian tax

By implementing a Pigovian tax of size t, i.e. interfering with the terms of exchange, we can shift the unregulated supply curve S upwards, until it is equal to S_ext, thus reducing overall CFC emissions. The market would reach a new equilibrium at price p _ext and quantity q _ext which reflects the existence of the externality. Notably, if the tax is set at the right level, we might thus ensure that the harm-threshold for the stock externality of ozone depletion is not exceeded.⁴ There is no denying the effectivity of such measures.

However, even if regulating market exchanges allows us to combat externalities, this does not mean we have found an answer to our question of whether exchanges featuring externalities should be judged unjust. Just because we can use regulated market exchanges in order to prevent the emergence of some externalities by dictating the price at which the exchanging takes place (and combat the effect of other externalities by compensating those affected with the tax revenues), this does not mean that the exchanges were responsible for causing or risking the externality in the first place. But this question of the cause of the harm is important. For unless it is the exchanges which wrongfully cause or risk the externality, we cannot say that the exchanges violate the harm principle and, thus, are unjust in the proposed way.

In fact, it is not the buying and selling (short: the exchanging) of goods or services which causes or risks externalities. It is the production or consumption of certain goods or services. What causes the consumption externality of CFC-emissions, for example, is not the fact that deodorants containing CFCs are sold too cheaply. Rather, it is the actual use of these deodorants by the consumers. It is true that standard economic theory predicts that when the price of a good or service is comparably lower, as it is the case for the unregulated exchanges, more of that good or service will be demanded. But we should not confuse the demand for a certain good with the effects its consumption has.

As for production externalities, consider again the case of the pharmaceutical company. The important thing to note is that at the time when the drug is exchanged on the market, the external costs have already been imposed on the fishermen. It would be odd to argue that the exchanging is what causes or risks the externality, for it only happens after the externality has already been created. Notice that even if the company did not sell its drugs eventually, the externality would have already been created. What causes the externality is not the exchanging of the drug for money, but the pollution associated with its production.

This reasoning about why exchanges do not cause or risk the emergence of externalities applies to all consumption and production externalities connected to exchanges of goods. For such externalities it is always true either that the harmful effect has already occurred at the time of the exchange (production externalities), or that the harmful effect only emerges when the good is consumed, i.e. after the good with external effects has been purchased in a first step (consumption externalities).⁵ Now, I think that the idea that exchanges cannot be the causes of production externalities because of temporal considerations is fairly uncontroversial.⁶ So I will focus on consumption externalities to address the natural question: if exchanges connected to the emergence of consumption externalities are not causes with respect to them, what are they?

I think the correct way to conceive of the relation between exchanges and consumption externalities is to hold that the selling and buying of, say, deodorants containing CFCs is an enabling condition (Dancy 2000, 128) for the emergence of the consumption externality. Following Dancy, there is

a difference between a consideration that is a proper part of an explanation [i.e. a cause], and a consideration that is required for the explanation to go through, but which is not itself a part of that explanation [i.e. an enabling condition]. (Dancy 2000, 128)

Consider the following example for the distinction of enabling conditions and causes. If I want to causally explain why a grain I planted three months ago now has developed into a flower, the formation of the sun four billion years ago surely is a consideration that is required for my explanation to go through. But unlike the consideration that the sun has been shining in the last three months, it hardly seems to be a proper part of the explanation. In much the same way, if we want to causally explain ozone depletion in the upper atmosphere, the fact that people in modern Western societies were buying and selling deodorants containing CFCs (in contrast to manufacturing them themselves or otherwise acquiring them) is a consideration that will figure in most explanations. But it is not a part of the explanation proper. In the explanation proper, only the excessive use of such deodorants would figure.⁷

There are several objections one might want to make against my claim that exchanges do not cause or risk but enable externalities—or against its alleged implication of exchanges featuring externalities not violating the harm principle. First, one might admit that we can distinguish causes and enabling conditions in explanatory contexts, but doubt whether the distinction meaningfully transfers to our context of looking for a liberal theory of justice able to deal with exchanges featuring externalities.

The answer to this objection is straightforward. For our purpose of questioning whether the state may permissibly interfere with a certain conduct, the most relevant context is a legal one. And in the legal context there is a distinction which is largely analogous to the one between enabling conditions and causes. I am thinking of the concept of an intervening cause—with the metaphor of an intervening cause being likewise derived from explanatory contexts (Honoré 2010, sec. 3.3). An intervening cause cuts off existing causal chains by “intervening” between agency and outcome. Following Honoré the idea is that “an agency will not be regarded as the cause of an outcome when that outcome is [instead] explained by a later … action [, event,] or a deliberate intervention” (Honoré 2010, sec. 3.3). The theory of intervening causes mirrors the distinction between enabling conditions and causes, at least on non-causal intervention accounts (e.g. Kadish 1985). For such accounts, whereas the intervening cause is a cause proper, everything up to this point does not have causal status with respect to the outcome. In our case, the use of the deodorant would be the intervening cause, and everything leading up to this intervention would not be causally relevant. Hart and Honoré (1985, 28–32) claim that the idea of intervening causes is “the common sense view” of causation in legal matters.

A second objection to the distinction between enablers and causes is that it is metaphysically unfounded and vague. Moore (2007, 408–412), for one, argues that although intervening causes purportedly are nonscalar, thorough metaphysical analysis shows that this is not the case. For him, all the actions and events that lead up to the intervention of the intervening cause might be causes with respect to the outcome in a “minor”, “secondary”, or “lesser” sense, but they are still causes.

Now this is not the right place to discuss the metaphysics of causation. In reply to this objection I only want to highlight the practical need for distinguishing between enabling conditions and causes on a day-to-day basis in explanatory and legal contexts. We cannot include every causal factor in the explanation of a phenomenon lest the explanation become unwieldy. For the same reason, we cannot factor in every causal aspect that contributed to a harm when administering justice. At some point we have to agree on a cut-off point for factors to consider and to ignore, and this is just what we do. We seemingly can satisfactorily distinguish between causes and enabling conditions. Although the distinction between causes and enabling conditions might be metaphysically shaky, we commonly make good use of it. All I am proposing is a further area in which it can usefully be applied.

A third objection is that the distinction between enabling conditions and causes might be possible, but that our judgment of whether exchanges featuring externalities are unjust should not hinge on it. The idea is that, as soon as we realize that exchanges featuring externalities neither cause nor risk externalities, we should actually reconsider our approach of relying on the harm principle as the way to deal with how externalities might be unjust from a liberal perspective. After all, there seems to be something normatively problematic about exchanges enabling a harm. And if our original approach simply ignores this aspect, it might have been the wrong one to start with.

This objection becomes even more plausible if one points to Duff and Marshall’s (2015) argument that both Mill’s classic discussion of the harm principle as well as Feinberg’s influential modern treatment oscillate between two different understandings. On the one hand, there is what Duff and Marshall call the Harmful Conduct Principle (2015, 135). It roughly holds that the state is permitted to call on force if a certain conduct wrongfully causes or risks harm. This is the way in which we have understood the harm principle until now (cf. Mill [1859] 1991a, On Liberty, chap. 1, para. 9; Feinberg 1984, 105–106, Feinberg 1988, 176–210). On the other hand, there is the more encompassing Harm Prevention Principle which roughly holds that the state may also call on force with respect to some conduct enabling harm, if by interfering with this conduct we can efficiently prevent harm (Duff and Marshall 2015, 134–135; cf. Mill [1859] 1991a, On Liberty, chap. 1, para. 9; Feinberg 1984, 26). So there seemingly is a way to keep violations of the harm principle as our straightforward approach of how exchanges featuring externalities might be unjust from a liberal perspective—and to account for their merely enabling character. This important objection calls for a detailed reply.

4 Enabling externalities

If we want conduct enabling harm to constitute a violation of the harm principle, then we need to say something as to what kind of enabling conduct we want to consider. Given that the chain of events leading up to a certain harm can be quite long, we probably would not want to propose that just any kind of exchange enabling a harm should be considered a violation of the harm principle and thus unjust. For instance, to also judge the exchange between the supplier of plastic caps for deodorants and the producer of the deodorant containing CFCs to be unjust in light of the externality would seem excessive. But the position that at least exchanges which wrongfully enable a harm should be considered unjust is intuitively appealing—especially as it is analogous to the cases of conduct wrongfully causing or risking a harm.

Now, if we want to endorse this position, we face the problem that we can no longer simply use the enabling exchange’s harmfulness as the reason for why it is wrongful. Whereas consequentialist approaches to externalities like that of Kagan (2011) are able to qualify conduct causing (or risking) externalities as wrongful because of its (potentially) harmful consequences, this is not an option we have with respect to conduct enabling harms. For it is the defining characteristic of conduct enabling harms that it is not itself (potentially) harmful. Yet, there seem to be two other promising ways in which we can ground an enabling exchange’s wrongfulness.

First, one might argue that we have some kind of associative duty to not only attend to the harms we cause or risk, but also to the harms we enable other people to cause or risk (Dworkin 1986, 195–216). Take, for example, the case of a host who refills her guest’s wine glass, although the guest has mentioned that he intends to drive home, and although she is aware that another drink will put the guest over the legal alcohol limit (cf. Duff 1990, 175). By doing so, the host enables her guest’s later drink-driving. In such a scenario, arguably the host does not simply have the duty to generously serve her guest whatever he likes, but also the associative duty to attend to the effects of her hostly conduct. And if she violates the latter duty, she may be said to wrongfully enable her guest’s drink-driving.

In a similar vein, one might then argue that sellers have an associative duty to attend to what the buyers of their products proceed to do with these products. Think of a gun shop owner. It seems like people selling guns have a special duty because of their role as providers of dangerous goods. And arguably such duties also extend to sellers of deodorants containing CFCs. Thus, in light of sellers’ associative duties, one might consider market exchanges enabling conduct that causes or risks externalities to be instances of wrongfully enabling exchanges.

Second, one might ground an enabling exchange’s wrongfulness by pointing to an analogy with the legal doctrine of aiding and abetting. Consider the case of a gun shop owner again. Suppose that, when selling a gun, the owner knows that the customer intends to use this gun to kill the president. Since the owner despises the president just as much as the would-be assassin, she advises him as to which gun best suits his needs. Two weeks later, the buyer of the gun shoots the president. In this scenario, the gun shop owner would be guilty of aiding and abetting a crime (Duff 1990, 168–174). Thus, just because the shop owner does not cause or risk the harm herself, this does not necessarily mean that her actions are not wrongful. (In fact, aiding and abetting even is a criminal offense in most jurisdictions.) Her normative involvement with the deed is decisive in this respect. Because of her normative involvement, her enabling exchange of selling the gun might be considered an instance of wrongful enabling.⁸

Even in light of these two ways to conceive of wrongfully enabling exchanges, I do not think that exchanges enabling externalities are unjust. I do think that enabling an externality is normatively problematic. It certainly feels like blaming the exchanging parties is not completely out of place. But not every enabling exchange which is blameworthy or even wrongful also is unjust. On the one hand, associative duties strike me as too shaky or unsuitable a basis for founding the judgment that exchanges enabling harms in violation of such duties are unjust—or even as an altogether incoherent basis for such a judgment. On the other hand, exchanges enabling externalities typically do not feature a normative involvement of either the seller or the buyer comparable to cases of aiding and abetting—and thus are not wrongful instances of enabling exchanges in the first place. Let me first elaborate on associative duties.

As regards associative duties, there is the following problem. Whenever we are trying to determine whether exchanges enabling a harm are wrongfully doing so based on the violation of such duties, we realize with Duff and Marshall (2015, 145) that it is “radically unclear, or controversial, just what the scope of those duties should be”. Take the example of the generous host who keeps refilling her guest’s glass. In Duff 1990, stated that many people would think that

[t]he fact that another drink would put [him] over the legal limit is not something to which she need attend (though it is something to which [he] should attend) as a reason against giving [him] another drink: she is not responsible for such effects of her hostly conduct. (Duff 1990, 175)

Social norms seem to have changed quite a bit since then. Relatedly, one might ask: does only a host have an associative duty to attend to effects of her hostly conduct, or also a bartender selling drinks to a guest whose car she knows to be parked outside? Or: do sellers of all kinds of products (guns, airline tickets, cars, gasoline, deodorants, etc.) have the same kind of duties? Overall, associative duties shape up as too changeable and contested a basis for general judgments about which kind of exchanges wrongfully enable harms.

In any case, they seem to be a bad basis for founding the more far-reaching judgment that exchanges enabling harms in violation of such duties are unjust. Note that to call a certain kind of conduct unjust is quite a harsh and momentous judgment—irrespective of the fact of whether we are dealing with conduct causing, risking, or enabling harm. Matters of justice are enforceable, i.e. the state may permissibly interfere with respect to them and they properly fall not into the private but the public domain. At the same time, arguments for the existence and strength of associative duties seem to get weaker as we move from mostly private relationships and people we regularly engage with to our duties towards strangers and one-shot, public interactions like market exchanges. So if we want to argue that conduct enabling harm in violation of such duties is unjust, this leads to two undesirable effects. First, we make a case for why the state may permissibly interfere in genuinely private affairs. Because associative duties are comparably clear and strong as regards the private domain, that case will be a strong one. Ceteris paribus this is a case liberals will be rather reluctant to make. Second, the case we actually want to make, namely that the state may permissibly interfere with high volume, anonymous market exchanges featuring externalities, remains weak. This is because associative duties are comparably unclear and weak as regards the public domain. To choose such shaky a fundament to legitimize state interference strikes me as problematic.

Apart from that, I feel like there is something self-defeating in appealing to associative duties for determining when a certain enabling exchange is unjust. It seems to be the very nature of associative duties to be moral rather than legal duties. They seem to cover conduct we deem morally important, desirable, maybe even admirable. But if we said that enabling exchanges violating such duties are unjust, these duties would become legally enforceable. And that would strike me as running counter to what these duties are. Again, even if such duties might serve as a basis for calling a certain conduct blameworthy, they seem to be an incoherent basis for calling it unjust.

All of this is not to say that there is no way to rely on associative duties in order to render state interference with exchanges enabling externalities in violation of such duties permissible. Consider, for example, a perfectionist political system championing the idea of citizens being their brother’s keepers—as a manifest means of bringing more people closer to leading a good life. And consider the deodorant case again; notably the seller’s selling.

From the seller’s point of view, whether the harm caused by the use of the deodorant actually occurs depends not only on her choice, but on what the buyer does as a result of what she does. There will be no harm done or risked unless the buyer actually proceeds to use the deodorant. In von Hirsch’s terms, what we are dealing with here is a remote harm, “remote in the sense that [it] involve[s] certain kinds of contingencies” (1996, 259). Notably, there is the contingency of the buyer’s intervening choice. The harm clearly transcends the individual sphere of the seller.

Yet, in a perfectionist system championing the idea that sellers are their buyer’s keepers, we might still choose to impute the harm done by the buyer on using the deodorant to the seller. It is true, one might argue, that the buyer still has the choice not to use the deodorant. But as the seller wrongfully enables the buyer’s harmful conduct, i.e. as she is not her buyer’s keeper with respect to an exchange featuring a good (potentially) harming a third party, her conduct might permissibly be interfered with by the state. In order to justify this interference, we might, for example, rely on Raz’s (1986, 412–413) argument that “[p]roviding, preserving or protecting bad options”—as enabling others to cause or risk harms—is not something a sound conception of liberty and autonomy calls for. Thus, he argues, sometimes a liberal state may permissibly interfere with conduct violating associative duties although this intrudes in the seller’s individual sphere of liberty.

To conclude, I admit that if we rely on more demanding understandings of liberalism, as for example Raz’s perfectionist liberalism, then arguably we can provide a theory of why exchanges featuring externalities are unjust. But it is an open question whether we can still call such a theory which relies on perfectionist reasons to argue why exchanges featuring externalities are unjust a liberal theory without any reservation.

If we instead argue for an enabling exchange’s wrongfulness along the lines of the exchanging parties’ normative involvement, we run into other problems. For there is an important disanalogy between cases of aiding and abetting a harm (like the gun shop case) and cases of enabling an externality (like the deodorant case). In the gun shop case, both the seller and the buyer exhibit a high level of mens rea. Having a mens rea is, in turn, a necessary condition for the offense of aiding and abetting. Following Duff (1990, 168–176), there are two clear scenarios in which we can affirm a mens rea for sellers.

(1):

The seller intends that the crime be committed

For example, if the shop owner selling the gun to the assassin is a co-conspirator in the killing of the president.

(2):

The shop owner does not intend that the crime be committed, but instead intends to assist in the successful commission of the crime

Duff’s example is that of a shop owner running a store specialized in selling tools for crimes, like untraceable guns (1990, 169–170). In this case, the owner might not care whether the president is dead or alive. But she surely has an interest that the assassination will not fail because of faulty equipment, as this would negatively affect her business prospects.⁹

What distinguishes exchanges which aid and abet a harm from exchanges which enable a harm in the form of enabling an externality precisely is the lack of mens rea for both the seller and the buyer in the latter case. Typically, neither the buyer nor the seller exhibit any kind of normative involvement with respect to the creation of the externality.

In the deodorant case, the buyer does not intend to create an externality. What he intends is, say, to feel refreshed or to smell good. As for the seller, she neither intends that the ozone layer be depleted, nor to assist in its depletion. In many cases, the seller does not even intend that the buyer uses the good he bought which features a consumption externality. Think of airline tickets. The airline certainly wants to be paid, but is indifferent with respect to whether the buyer actually boards the plane. (If anything, it seems preferable if the buyer does not, as this leads to less weight and a corresponding lower use of kerosene, a lowered likelihood of overbooking, etc.) Even in cases in which the seller hopes for the buyer to return once the good has been used, all she intends is that the good be used, in the sense of “consumed” or “used up”, in order to initiate repurchases. But she has no intention as to how it should be used. To her, an activist who repeatedly buys deodorants and destroys them in a non-harmful way (so that nobody else buys them and actually uses them in their intended use) is as welcome a customer as any.

To be sure, in cases in which the good has an intended use which causes a negative consumption externality, the seller foresees that the good most likely will be used in this way. But this only brings us back to the finding that a seller foreseeing that a buyer will commit a crime can only be considered to have a mens rea if we also affirm a strong role of associative duties. In a non-perfectionist state, the use the buyer actually puts the good to is beyond the things we want to impute to the seller. A good having an intended use only affects the probabilities with which a harm may ensue, but does not affect the normative involvement of the seller. This is because a good having an intended use does not mean that, necessarily, the seller intends the buyer to use the good in this particular way. It is a functional characteristic of the good. It basically means that if the buyer uses the good in accordance with its intended use, then it will serve its purpose, it will function well. The seller only is normatively committed to the extent that, if the product is put to its intended use, he promises that it will function properly in the advertised way. But typically, for all the seller cares, the buyer may use the good in whichever way he wants.¹⁰

Cases in which the seller or the buyer intend to contribute to an externality seem to be far-fetched—if only because the individual contribution to the most pressing externalities (global warming, water pollution, etc.) is small, and the exchanging parties are themselves negatively affected. That is, the reason why we consider aiding a harm a wrongful kind of enabling conduct—the involved parties’ normative involvement—does not easily transfer to cases of exchanges enabling externalities.

Note that just as it would be odd to attribute the intention to deplete the ozone layer to the buyer or the seller in the deodorant case, it would be odd to consider the killing of the president an externality. One would not reasonably suggest that guns should be taxed in order to internalize the externality of assassinations of presidents. Yet again, it would not be odd to consider a high rate of accidents involving guns an externality of cheap and easily available weapons, and to call for a high sales tax on guns in order to combat this problem. In this latter case, again neither the gun shop owners nor the buyers intend these accidents in any way.

What we see here is a subtle line of demarcation between exchanges aiding harms and exchanges enabling harms by enabling externalities. As outlined above, when we say that a certain exchange features an externality, this is to say that the market price of the good exchanged is distorted because of some external costs not being factored in. As this is a statement about the price all the people exchanging that good face, we are not concerned with individual instances of an externally costly use (or abuse) of a certain good. That is, it typically is not individual cases (read: crimes) which underlie our judgment that an externality is present. Rather, we are making a judgment about exchanges which systematically enable the emergence of a foreseen harm, but not intentionally so.

Therefore, even if exchanges aiding a harm—i.e. exchanges wrongfully enabling it in light of the exchanging parties’ normative involvement—were unjust, exchanges enabling harms by enabling externalities would not necessarily be. Because there is a difference as regards the normative involvement of the exchanging parties between exchanges aiding a harm and exchanges enabling externalities, the latter are not wrongful and thus could not be unjust.¹¹

Whenever externalities are intended either by sellers or buyers, we are no longer dealing with a “plain vanilla” externality scenario. Some other normative concern than just the price of the good exchanged being too low and the quantity exchanged being too high because of uncompensated-for costs of people external to the exchange seems to be present. In such cases, it is an open question whether it is indeed the externality which is normatively troublesome or some altogether different phenomenon—i.e. of whether we are really giving a theory of how exchanges featuring externalities are unjust.¹²

5 Externalities, justice, and interference

There is no general liberal theory of how exchanges featuring externalities are unjust—at least with respect to the most common meanings of the terms “liberal” and “externality”. We might be able to reasonably widen these meanings in our context. For example, we might be able to identify some cases of non-standard or quasi-externalities which are unjust according to the criteria proposed here. But an account of why some exchanges featuring non-standard or quasi-externalities are unjust, or why “plain vanilla” externalities are unjust according to a more-demanding-yet-broadly-speaking liberal theory, is hardly what we are looking for when we are seeking a liberal theory of externalities.

I do not claim that a liberal state may never legitimately interfere with the emergence of externalities. One can at the same time hold that we have no liberal theory of why exchanges featuring externalities are unjust and that a liberal state may sometimes permissibly interfere with externalities. This is important because fighting externalities is a public matter. Unconcerted efforts typically will not be enough to overcome many externalities.

Before considering the possibilities a liberal state has, however, note that another route for legitimate interference with exchanges of goods featuring externalities seems blocked. Namely, the route via regulation and mala prohibita. Suppose a legislator aims to prevent the depletion of the ozone layer. She realizes that interfering with the exchanges of deodorants containing CFCs is an effective way to do so. So she goes ahead and bans their sale and purchase.¹³ That is, she introduces regulation which makes selling and buying such deodorants a malum prohibitum, a kind of conduct whose wrongfulness depends essentially on its illegality (Duff 2007, chap. 4.4). How could she justify such regulation? I have just argued that she could not justify it by pointing to a violation of the harm principle and corresponding reasons of injustice. But she might appeal to other reasons. For example, she might argue that regulations of the kind she proposes are justified whenever we can prevent a harm without this regulation being unduly burdensome for those regulated. And she might then further argue that a liberal state cannot only interfere with exchanges on grounds of justice (as backed by the harm principle), but also on legal grounds once the respective regulations are in place.

Regulations dealing with conduct enabling harm in order to prevent harm are widespread. Duff and Marshall (2015) discuss the case of regulating drink-driving. They argue that whatever legal alcohol limit a rational legislator might set in order to prevent the harms connected to drink-driving, there will be some people who can drink more than that amount and still drive perfectly safely—e.g. because of their special bodily constitution. That is, whatever regulation the legislator passes, it will rule out some conduct which neither causes nor risks a harm, let alone wrongfully so, but merely enables a harm. Each regulation will be over-inclusive in the sense of ruling out conduct which is not unjust. Still, some think that these kinds of regulations are warranted, precisely because they effectively prevent harm by interfering with conduct enabling harm. (And we could call conduct violating such regulations “unjust” in the sense that, once the regulation is in place, it is illegal.)

Now, this general possibility of legitimately interfering with conduct enabling a harm once we have turned it into a malum prohibitum should not lead us to the conclusion that this conduct was unjust to begin with, i.e. a malum in se as judged by the harm principle, or that we may easily do so in a liberal state. Only if there is a valid liberal principle allowing for regulation creating mala prohibita, the conduct may be legitimately interfered with in a second step. But that is a big if. What is crucial for our context is that we cannot rely on the harm principle to this effect. As argued in Sects. 3 and 4, exchanges of goods featuring externalities do not violate either understanding of the harm principle. Thus, this route for legitimate interference by a liberal state is blocked. We cannot easily find a principle other than the harm principle which is both weaker and robustly liberal, and which justifies turning exchanges of goods featuring externalities into a malum prohibitum.¹⁴ If we honestly present our reasons for regulation, these reasons may not contain any talk of injustice, for such regulation rules out exchanges over and above those ruled out by the harm principle.

At this point one might ask why we should even bother to have a liberal theory of how exchanges featuring externalities are unjust. After all, my findings do not really seem to detract from a state’s ability to legitimately engage with externalities. I argue that it would be wrong for the state to interfere with exchanges featuring externalities, e.g. by levying a Pigovian sales tax. But if we chose to interfere in some way in order to prevent externalities (which we probably should), we could still target consumption or production rather than exchange.¹⁵ I do not deny that a liberal state may interfere on grounds of justice with the production and consumption of goods featuring externalities.

The finding that exchanges featuring externalities do not share the same normative status as exploitative, coercive, or deceptive exchanges in a liberal state still is important. Firstly, simply because somewhat surprisingly there is such a difference in status. While typically even libertarians agree that coercive and deceptive exchanges are unjust (e.g. Nozick 1977, 150–158; but cf. Child 1994), and while there is an established liberal theory of exploitation (Steiner 1984), there is no liberal theory of externalities—although there is virtually unanimous agreement that exchanges featuring externalities are normatively problematic. In discussions about how markets and exchanges are just or unjust, the issue of externalities—unlike the issues of exploitation, coercion, or deception—often is more or less brushed aside, e.g. in footnotes which read: “I assume no externalities to be present”. Seemingly they constitute a widespread but ultimately simple complication of just markets and exchanges. But externalities are a surprisingly complex normative issue.

Furthermore, whereas interference with the production or consumption of goods connected to externalities is permissible, such interference ceteris paribus seems less desirable than interference with exchange for practical reasons. The consumption of goods is notoriously difficult to regulate, as it often happens in private. And interference with production, e.g. in the form of production quotas, runs a comparably high risk of creating an inefficient market equilibrium. That is, a particularly practical way of dealing with externalities is normatively blocked because of my finding.