2. Externalities in economic theory

It is surprisingly hard to give a precise definition of the concept. Several elements are regularly included in the definition of externalities, but each of these elements is subject to interpretation (Hausman, Reference Hausman1992):

1. (1) An externality refers to a – positive or negative – effect of an action by a person A on a third party C;

2. (2) The effect is imposed on the third party without his/her consent;

3. (3) The externality marks a Pareto-inefficient situation.

The first characteristic highlights the fact that an externality is a spillover, an incidental by-product of an activity of A, which has an effect on a third party. A's action may or may not be part of a market transaction with a person B. Some externalities are not spillovers from a market transaction, but spillovers from an essentially solipsistic action by A. The crucial thing is that there is no market between A and C. This is what the second element highlights. The third party has not voluntarily consented to receiving the effect. This does not necessarily mean that he would not welcome the effect; after all a positive externality does make the receiver better off. This however does not take away the absence of consent as a distinguishing mark of an externality. Finally, the externality leads to an inefficiency. As we will see, this is often cashed out in terms of missing property rights and missing markets. The third party lacks a clear property right which he could claim as a defense when he is subjected to a negative externality.

In economics, there are two traditions in thinking about externalities (for an overview, see Barnett and Yandle (Reference Barnett and Yandle2009), Demsetz (Reference Demsetz1996), Lagueux (Reference Lagueux2010) and Medema (Reference Medema1999). The first one is the Pigovian school, which was the dominant one for much of the 20th century. It dates back to Pigou's analysis in the Economics of Welfare (Pigou, Reference Pigou1960).Footnote ² Its main thrust is that externalities represent a divergence between marginal private cost and marginal social cost. A classical example of a negative externality is that of a factory emitting smoke or dumping pollution in a river, in both cases damaging the property of the surrounding population. In such a case, it seems clear that those harmed should be protected or compensated against the polluter. Inefficiency arises because the factory doesn't take into account the total costs of its behavior. These costs should not only include the private costs of production but also the social costs to others. State regulation serves to force the factory to internalize these costs, by taxing or subsidizing the producer. The Pigovian tradition, then, has an optimistic view of the state's capacities to solve externality problems.

In reaction to this analysis, the Coasean school emerged. Ronald Coase argued in his classic paper ‘The Problem of Social Cost’ (1960) that Pigou's analysis neglects the existence of transaction costs: the costs of searching, bargaining and enforcing contracts. In a fictitious world without transaction costs, Coase argued that private parties could always bargain about an optimal solution to the externality problem. If preventing an externality is more valuable to the recipient than having it is to the producer, then the former can pay the latter to refrain from his actions. Externalities would be non-existent as a result, and the situation is Pareto-optimal. In the real world, there are always positive transaction costs. In that world, Coase argued, externalities might persist because bargaining is prohibitively costly. To reach an efficient outcome, this situation calls for an initial assignment of the property rights to the party which values them most. Note that in the transaction cost-less world, it does not matter who is assigned the initial rights, since bargaining will make sure that they will end up in the hands of the party valuing them most. In both cases, then, the state need not resort to taxes or subsidies. It should confine itself to assigning and enforcing property rights in the (former) external effect. The newly created market can do the rest (Coase, Reference Coase1960).

For purposes of this paper, it is useful to highlight three features of economic externality analyses from this history. The first two stress important differences between the Pigovian and Coasean tradition, the third feature is shared by both traditions.

First, the Coasean analysis shows that the relation between the producer and the recipient of an externality is of a reciprocal nature. In this, it radically diverges from the Pigovian tradition which held the producer responsible for the externality – he is the one who is expected to adapt his actions. Coase convincingly showed that this assumption cannot be sustained. If by smoking a cigarette I pollute my neighbor's air, we cannot non-controversially say that I am imposing an externality on him. In fact if I were forced to refrain from smoking, then he would be imposing an externality on me. This is an important insight. Our intuitions may make us see one party as the perpetrator and another one as the victim, and lead us to identify the externality in such a way that the perpetrator is imposing costs on the victim. All of that, however, is the result of implicit moral analysis which needs to be made explicit to see if we really want to assign responsibility in that way. Whichever way we assign the responsibility for incurring the costs cannot be decided by the concept of externality itself. The concept merely highlights a boundary crossing, the existence of an interaction effect for which there is no market. It cannot tell us who should be held responsible for that effect. In Coasean economics, the decision is made by allocating property rights to the party who most values them. This is itself a normative criterion.

The second difference is that the Coasean analysis restricts the scope of application of the externality concept. Transaction costs are not in principle different from other costs of production. It does not really matter whether the costs of solving a case of pollution are too high for them to be internalized due to high production costs (i.e. the costs of installing technologically advanced filters to prevent the pollution), or due to high transactions costs between the factory and its neighbors. In both cases, as long as these costs are higher than the value of preventing the externality to the neighbors, then the externality won't be internalized. This has led some to claim that there is no ‘policy-relevant’ externality. The constraints of transaction costs are real-world facts we have to accept. Everybody is as happy as he could possibly be in this constrained world, since there is no relevant world to which we have access in which he could be happier (Dahlman, Reference Dahlman1979). In sum, there are no externalities in a transaction cost-less world, but neither are there any in a transaction cost-constrained world. The concept loses its application. The Coasean analysis betrays a ‘panglossian’ worldview (Lagueux, Reference Lagueux2010).

However, this conclusion is too quick, for it takes current transaction costs as fixed. However, there may be situations in which it is possible to lower transaction costs either through efforts of the parties themselves or by state intervention. If the costs of that intervention are lower than the gains that it brings about, there is a policy-relevant externality worth internalizing. This restores a role for the state beyond assigning and enforcing property rights. The state can take an enduring externality as a sign that private parties themselves are unable to lower transaction costs. It can then investigate whether intervention to lower transaction costs would lead to welfare benefits. A Coasean view, then, requires a comparative institutional analysis of the relative virtues of markets and states in overcoming transaction costs. This can explain why the externality literature after Coase has flourished as much as before his seminal paper (Demsetz, Reference Demsetz1996: 577). While it is true that the Coasean analysis restricts the scope of the externality concept (by insisting that the presence of transaction costs makes some externalities not worth internalizing), there will still be plenty of situations in which internalization can be presented as optimizing welfare.

A third and final point is the market-making nature of externality analyses in economics, both in the Pigovian and in the Coasean tradition. Internalizing an externality is a form of extending the scope of the relevant market. An effect formerly escaping market-governed interaction now becomes part of that market. When one presents this as an efficiency-improvement one presumes that marketization of that effect is itself (morally) unproblematic. This may be the case for many externalities but need not be the case for all of them. An example is the internalization of environmental externalities. The latter presume that we can correctly value nature by putting a (shadow) price on natural assets (wild parks, landscapes, biodiversity), and that this is something we should morally do. By focusing on the strategy of creating private property rights for the external effect, economic analyses tend to neglect alternative solutions such as assigning collective property rights or keeping the effect in the informal sphere.Footnote ³ This market-making aspect explains why the externality-concept has been enthusiastically embraced and criticized from both sides of the political spectrum (Arnold, Reference Arnold2009: 28–36). In the eyes of conservatives, the concept is prone to abuse by state-enthusiasts who want to identify market failures everywhere in economic life and interfere on that basis. In the eyes of liberals, the concept extends the scope of the market to domains where it doesn't belong. This political ambivalence can be traced to an ambiguity that is inherent in the use of the externality concept. It does extend the scope of the market, which is a price liberals will sometimes want to pay to protect vulnerable non-market interests, such as social or ecological values. At the same time externalities do involve state intervention, but this is a price conservatives sometimes will want to pay to create new markets.

The second feature – the need for a comparative institutional analysis – defines the task of externality analyses. The third feature – its market-making character – points to the politically charged waters in which externality analyses have to sail. The first feature – symmetry – explains why this is the case: assigning responsibility for boundary transgressions is a normative task. This feature leads us to political philosophy.

3. Philosophical criticisms of the externality concept

The few philosophers who have written about externalities have criticized the normative thrust of standard economic analyses. The following passage by Debra Satz conveys the essential points:

I agree with Satz that the concept is invoked selectively. Indeed, given the ubiquity of externalities in daily life, any sound analysis must make selective use of the concept. I also agree with Satz that selective use require a moral theory not implicit in the concept itself. However, I would disagree with the suggestion that such a moral theory is absent in economic reasoning itself and imported from elsewhere. In most economic reasoning, there is a moral theory to identify externalities: utilitarianism. Utility in turn is normally understood as the satisfaction of subjective preferences.Footnote ⁴ Even if economists use the concept selectively, economic theory can be applied non-selectively by using its utilitarian calculus. My aim is this section is first, to show this utilitarian nature of the analysis at work, and second, the problematic character of relying on that theory. I will do so by looking at three types of controversial externalities.

First, let's look at what we can call moralistic externalities. These are externalities arising from preferences about other people's behavior. One example is dress code. If I hate men with long hair, then any confrontation on the street with such men imposes an externality on me. Another example is sexual imagery. Pornography offends many women (and some men) and thus producers of pornography impose an externality on them (Trebilcock, Reference Trebilcock1993: 65). Another example is sexual orientation. Homosexuals by having sex – even in private – offend the sensibilities of some orthodox religious people (Herzog, Reference Herzog2000: 911–912). All these examples are cases in which the externality concept could be put to use. These situations, after all, involve spillovers of actions that are imposed on others and lead to inefficient outcomes. A Pigovian analysis would be absurd here, since it would automatically lead the government to impose taxes on the long-hair wearing men, producers of pornography and practicing homosexuals. This shows the value of the point introduced in the Coasean analysis: that the situation is reciprocal and responsibility need not rest on the one we identify intuitively as ‘the producer’ of the externality. Instead, the Coasean analysis approaches this situation, as we have seen, by introducing the rule that there only is an externality when there is a mutually beneficial gain to be made between the affected parties. The parties’ preferences thus become decisive, in combination with the initial assignment of property rights. Does that deliver a satisfactory analysis?

If I value being freed from having to watch men with long hair at 60 euros, and the only man with long hair in the village values the opportunity to wear long hair at 50 euros there is space for a mutually advantageous bargain. If initially the man has the right to decide on his own hair dress, then I can bribe him for, say, 55 euros, to cut his hair. If I have the right to be freed from having to watch long hair, no bargain needs to be made. In both cases, I end up having it my way. If the man values his hair dress at 60 euros and I value being freed from it at 50 euros, then he will end up wearing his hair long (with the need for bargaining and the distributive consequences again differing depending on the initial assignment of property rights). Similarly, if women feel offended by pornography, but feel less strongly about it than the producers of pornography, pornography will be allowed. If homosexuals feel less strong about having sex then the religious believers who will be outweighed, then the state should prohibit their sexual practices. The calculus depends on the numbers of individuals involved as well as on the intensity of the preferences. Big groups of religious believers, even with medium-intensity preferences, will outweigh small groups of homosexuals with high-intensity preferences.Footnote ⁵

Most people will have strong intuitions that this analysis leads to the wrong outcomes. Liberals will most often argue that long hair, homosexual practices and pornography should be allowed, whatever the preferences of certain groups. These are a matter of individual freedom.Footnote ⁶ This position depends on a non-utilitarian view, which can be systematized using the harm principle. I will discuss this hereafter. However, even a conservative who would endorse the opposite outcomes, although diverging from the liberal in his substantive conclusions, will run up to the same problem in economic analysis: that it leaves the outcomes up to the preferences of individuals. So whatever one's position on the issues, the problem is not so much that economic analysis requires importing normative theories from elsewhere. The theory's rule of allocating to whomever values the property right already consists in the application of a normative principle. The point is that the outcomes do not match our intuitions (an economist may bite the bullet and argue that any counter-intuitive conclusions are a reason to revise our intuitions. The only way of refuting that argument is to show that another theory leads to less counterintuitive results).

A second example of controversy is the category of pecuniary externalities. Due to a change in consumer preferences, a technological innovation or a price-cutting strategy from a competitor, some producers will face a decline in demand of their products. A perfectly competitive market thus generates a cost for losers in a competition, which is imposed on them by consumers or their competitors. Can this be an instance in which, depending on the calculus, negative externalities require internalization? The standard economic analysis denies this because in a general equilibrium analysis, the costs to the losers are more than offset by the benefits to the winners of the competition and to consumers who benefit from the enhanced performance of a competitive market. It is best overall to have the competition. However, Daniel Hausman has forcefully argued that pecuniary externalities still raise questions of justice. The losers may be severely harmed in their existence, sometimes to the point of starvation. The reply that they have consented to playing the game does not help, since this is a case of forced consent. Not playing the market game for most people is simply not an option (Hausman, Reference Hausman1992: 103–105). In this area too, we see how a utilitarian analysis omits an important part of what is morally relevant.

Third, let's take positional externalities, a species of negative consumer externalities. Economist Robert Frank has argued that these should be internalized through a consumption tax. Positional externalities arise where consumers lose welfare because they compare themselves to others. The utility they derive from a consumption item partly depends on their place in the relative distribution of that good. When some buy more expensive goods and move up the hierarchy, they harm others, or impose a negative externality upon them. These will retaliate by also trying to move up, resulting in a ‘positional arms race’ in which all parties are worse off (Frank, Reference Frank2008: 1778). Frank sets his argument explicitly in a Coasean framework, balancing the costs to both parties who are symmetrically placed. On his view, the costs outweigh the benefits. He argues there are psychological costs of those who lose out in a positional competition (the feeling of ‘lagging behind’ average expenditure patterns) and more tangible costs.Footnote ⁷ Frank leaves it very unclear how to calculate the costs, and whether these are larger than the benefits of positional competition. The welfare calculus seems highly indeterminate.Footnote ⁸ This is all the more so since he includes things like psychological costs. Any utilitarian calculus, however, needs to first determine which costs can enter into the calculation. And this requires normative decisions that utilitarianism cannot make itself. All in all, it therefore remains highly questionable whether those who impose status losses on others should be held responsible for doing so.Footnote ⁹

These cases point to different problems with a purely utilitarian calculus: it ignores issues of individual freedom (moralistic externalities) and justice (pecuniary externalities); and the calculus itself is highly indeterminate (positional externalities). However, philosophers thus far have been stronger at criticizing economic externality analyses than at providing an alternative. Can we find a more solid ground for a normative analysis of externalities? In the following, I will follow the suggestion made by Satz and others (Herzog, Reference Herzog2000: 912; Trebilcock, Reference Trebilcock1993: 61) and look at the harm principle.

4. From the harm principle to basic interests

John Stuart Mill introduced the harm principle as a criterion for deciding when the state would be allowed to interfere in the private life of individuals: ‘the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant’ (Mill, Reference Mill and Gray1991: 14). In this section, I will briefly discuss the analogy between this principle and the concept of externalities, and propose an interpretation which links it to a theory of basic interests; this interpretation offers, so I claim, a convincing non-utilitarian way of operationalizing the concept of externalities.Footnote ¹⁰

Harming another person is analogous in important ways to imposing an externality, as can be seen when going back to the three main elements of the externality concept (see section 2). The first element was that of an incidental effect of one person's actions on another person. This is also part of the harm principle, with one important exception. The externality concept applies to positive and negative effects, while the harm principle only covers negative effects. The second element was that an external effect is imposed without the consent of the effected person. This too is part of the harm principle. It is widely accepted in the literature on the harm principle that a consented harm does not count as a harm for which state interference is warranted (Feinberg, Reference Feinberg1984: 115). Third, the harm principle at first sight seems to lack anything analogous to the third element of the definition of externalities: a state of inefficiency. However, this appearance is deceptive. The qualification of a case as containing ‘harm’ implies that one has identified a suboptimal state, compared to some baseline. The harm principle does not necessarily operationalize this in terms of utility (subjective preference satisfaction), but then again, there is nothing in the concept of ‘inefficiency’ that dictates such a normative criterion either. All in all, then, the harm principle aims to mark off a sphere of self-regarding behavior where one is to be left alone. This is analogous to the market sphere in economics, where consenting adults who impose no harm on third parties are to be left alone by the government.Footnote ¹¹

Let's now turn to the crucial point. The philosophical discussion on the harm principle suffers from the same problem as the economic discussion on externalities. In both cases, the principle on closer inspection proves to be largely empty. It requires a normative theory to decide what qualifies as a harm. Once such a normative theory is formulated, it is doing the actual work, instead of the concept of harm itself (the latter merely ‘passes the bucket’). This conclusion is exactly the same as the one I reached in the case of externalities in the previous section.Footnote ¹² So we have to ask: which normative theories are available to give substance to the harm principle? In the recent literature, two prominent normative interpretations of the harm principle have been given: in terms of welfare interests, or in terms of personal autonomy (Holtug, Reference Holtug2002).

The welfare-analysis is proposed by Joel Feinberg. In his view, a harm is a setback of our interests. This pushes us from the question of harm to the question of what our interests are. Here, Feinberg distinguishes between interests of different types: mere passing desires (such as a desire for ice-cream), instrumental wants (such as ‘to forgo dessert’), welfare interests (such as ‘physical health’ or ‘economic sufficiency’) and ulterior or focal aims (such as ‘writing a book’ or ‘acquiring political power’) (Feinberg, Reference Feinberg1984: 56–60). He argues that the state should not protect us if we are harmed in terms of a setback of our passing desires or instrumental wants, since these are either not sufficiently durable, or instrumental to our more ulterior goals. The state however should also not protect our ulterior goals directly, since achievement of these should be in large parts up to us. What it should do is to protect the necessary conditions for reaching our more ulterior goals, i.e. our basic welfare interests: ‘achieving and maintaining that minimum level of physical and mental health, material resources, economic assets, and political liberty that is necessary if we are to have any chance at all of achieving our higher good or well-being, as determined by our more ulterior goals’ (Feinberg, Reference Feinberg1984: 57). Thus, Feinberg associates the harm principle with basic welfare interests. These are independent of the specific wants or interests that anyone may have: they are what everyone has an interest in, in virtue of having other, more specific and personal (higher or merely passing) wants and interests (Feinberg, Reference Feinberg1984: 62).

The other major alternative interpretation of the harm principle is from Joseph Raz, who has argued that we should understand the harm principle as protecting the conditions of our personal autonomy. A person who is autonomous is the author of his own life. The conditions for being autonomous, according to Raz, include not being manipulated or coerced by others, having sufficient mental abilities to make autonomous decisions, and having an adequate range of options to choose from (Raz, Reference Raz1986, Reference Raz and Gavison1987). Feinberg's basic welfare interests turn out to be structurally analogous to Raz's conditions for personal autonomy. Both relate to a set of necessary conditions for (1) being able to choose one's own aims (Raz) or (2) achieving one's own more ulterior goals (Feinberg). The difference here is more one of words than of substance.Footnote ¹³ This convergence between Feinberg and Raz on a set of objective conditions for autonomy/ulterior goals is markedly different from the economic analysis in which subjective preferences are the normative criterion. The economic analysis accepts all of Feinberg's four types of interests as input into the calculus as long as these are held as preferences by those individuals whose situation is the object of the analysis. Feinberg convincingly shows how this would make the analysis of harm/externalities hostage to fleeting convictions and intangible goals.

The structural convergence of Raz's and Feinberg's theories of harm is on what is often called an ‘objective list theory of the good’: a list of basic interests which make a life go well. Raz's and Feinberg's proposals are by no means the only ones in political philosophy which defend such a theory. John Rawls's theory of justice, for example, includes at its core a list of what he calls ‘social primary goods’. These are defined, like Feinberg's welfare interests, as goods a person needs whatever his more specific goals in life are (Rawls, Reference Rawls1999). Similarly, Martha Nussbaum has proposed a list of basic capabilities, with a similar aim (Nussbaum, Reference Nussbaum2000). Others have produced theories of basic human rights (Gewirth, Reference Gewirth1978; Griffin, Reference Griffin2008) or basic human goods (Finnis, Reference Finnis2011). All of these theories either present (or pre-suppose the presentation of) a list of basic human goods or interests. Citizens are to have rights to these goods and the state (or whatever form the political community takes) is to protect these rights for its citizens against violations from others. Despite the enormous differences between these theories, I will therefore here refer to them together as basic interest theories. Any of these, not just Feinberg's or Raz's proposed theories, could be used to fill in the harm principle. A harm is a violation of a basic interest. For purposes of this paper, one need not choose a particular theory within this category. What is needed is some theory of basic goods, and Raz's and Feinberg's theories are the examples of this type of theory which are most closely engaging with the harm principle. But the reader can simply take his/her favorite theory by way of example, and think of it when reflecting on the harm principle.Footnote ¹⁴

Basic interest theories may help generate more plausible responses to the three controversial types of externalities discussed in the previous section. Most of these theories would judge that moralistic externalities do not provide a reason for internalization because unusual dress codes, pornography or homosexual acts do not setback the conditions for an autonomous life of those who feel offended by it. It may set back their ulterior goals (if these include other-regarding wants), but that kind of interest is, as we have seen, not part of a basic interest interpretation of the harm principle (for good reason). So here the principle sides with the liberal philosophers who have protested to the economic analysis of these externalities.Footnote ¹⁵ The basic interest interpretation of the harm principle would judge pecuniary effects reason for interference to the extent that those subject to market-based losses of competition are set back below the threshold level of their basic interests. This means it shares Hausman's intuition that sometimes there is a justice-based case for worrying about pecuniary externalities.Footnote ¹⁶ Whether this is done by interfering directly with the competitive process (so that this harm does not occur) or by allowing the harm to happen but then compensating the losers of the competition afterwards, is an open question. Finally, the basic interest interpretation of the harm principle would judge positional externalities too on their potential for undermining the objective conditions for an autonomous life. This would mean that Robert Frank's argument about psychological costs would have to be converted into an argument about harm to the necessary conditions for autonomous living (I am inclined to judge that these psychological costs will not qualify as harm, but shall leave it to the reader to make up her own mind on this issue).

The proposal to adopt some basic interest theory for operationalizing externality analyses might be subject to several objections. Let me briefly discuss three of them and show how they can be answered. This will help to further elucidate essential features of the proposal.

First, one might object that the concept of harm cannot deal with cases of positive externalities (and similarly, since public goods are essentially large-scale externalities, one might claim the proposal can only deal with public bads, not public goods). One may think this is a severe restriction compared to economic externality analyses. However, it should be emphasized that in the proposal the concept of harm, as an intermediate term, drops out of the equation. What matters is the analysis of externalities (i.e. any interdependencies in interaction, whether positive or negative) in terms of basic interests. Indeed, a case for public goods can be made on the grounds that a good which represents a basic interest for every citizen (say national defense) can only be delivered by the state due to technical problems with market provision (excludability, free riding). Similarly, only positive externalities that contribute to reach a threshold level of basic interests need to be a reason for state subsidies or other measures. When a spillover does not represent such an interest, it may be ignored.

A second objection may be that a basic interest analysis cannot capture all relevant harms. Imagine two neighbors, both of whom have their basic interests met. Next, imagine that one imposes loud noise or any other inconvenience on the other. Can the second person complain about a negative externality imposed on him? The basic interest view must remain silent on the issue, since presumably all their interests are met. Against such an objection, I would emphasize two points. First, the objection pre-supposes that all basic interests are met, but we do not have to accept that set-up. Noise, for example, when it endures and is sufficiently loud, leads to sleep deprivation and other serious ills, which do represent a basic interest. Second, basic interest analyses normally make use of a threshold. A basic interest in free speech does not guarantee every citizen maximal police protection at all moments, but may for example require that one accepts limitations in the moments and places one can march on the streets to voice one's cause. If in this case the threshold of protection against noise is reached, then indeed remaining inconveniences need to be swallowed (Shiffrin, Reference Shiffrin2000).

Third, one may object that it is unclear how we can come to a non-controversial list of basic interests. Economic analysis, it is often thought, by simply working with individuals’ own preferences as input, avoids that kind of controversy. In response we need to note, first, that the examples above (and the example in the next section as well) show that relying on utility information will also be controversial. There simply is no non-controversial way to draw boundaries around persons and their interactions. It is better to confront rather than avoid these issues. Second, basic interest theories are philosophical constructions, but in the end political actors (parliaments, judges) decide upon these issues. The procedural legitimation political actors have in a democracy derives from their promise to deal with controversy in a way that is acceptable to all interests involved. In processes of political will formation, of course, citizen preferences also play a role as input. A democratic process, however, offers opportunities for will formation that include intelligent deliberation and go beyond mere aggregation of pre-existing preferences (Lewinsohn-Zamir, Reference Lewinsohn-Zamir1998). There is a division of labor in this respect between philosophical theory and actual democratic processes. This leads to a familiar problem when democratic processes lead to outcomes which the theorist may deem philosophically unacceptable, e.g. because they violate basic interests (or, in terms of externalities: because they lead to paternalistic or moralistic interventions). Here one's philosophical commitments to a theory of justice and to a theory of democracy clash. This problem, much discussed in democratic theory (Claassen, Reference Claassen2011; Michelman, Reference Michelman, Bohman and Rehg1997; Van Parijs, Reference Van Parijs2011; Waldron, Reference Waldron1999; Walzer, Reference Walzer1981), must be confronted elsewhere.

While more needs to be done to defend this proposal, these remarks hopefully suffice to show the coherence and attractiveness of basic-interest theories in the realm of externality analyses.Footnote ¹⁷ In the next section, I will take a more in-depth look at one recent court case to show how in legal reality this type of reasoning is already present when making decisions about externalities.

5. Externalities in the health care market

In June 2012, the US Supreme Court delivered its long-awaited verdict on the health care reforms of the Obama administration (the ‘Affordable Care Act’).Footnote ¹⁸ The core question was whether individual citizens could be required to purchase health insurance, the so-called ‘individual mandate’. Legally, the question required an interpretation of the constitutional clause that Congress has the power ‘to regulate commerce (. . .) among the several states’ (art. I, sect. 8, cl. 3).

According to the opponents, the mandate was unconstitutional, since it obliged citizens to buy something, while every citizen should be free to decide whether or not to engage in market transactions. The Supreme Court's majority opinion endorsed this view and judged the mandate unconstitutional.Footnote ¹⁹ The majority opinion relied on the view that the non-insured are ‘outside the market’. It essentially gave two arguments for this view, an ontological one and a normative one. The ontological argument is that non-insured individuals are not ‘active’ on the market: ‘The individual mandate, however, does not regulate existing commercial activity. It instead compels individuals to become active in commerce by purchasing a product’ (Opinion of Roberts, p. 20).Footnote ²⁰ The fact that even the uninsured consume health care when they fall ill, is not a reason to require them to purchase insurance now. This ontological argument about the definition of the market's boundaries is meant to stand on its own as an interpretation of what it is to ‘regulate commerce’. But it derives its convincingness from a further, normative argument: that if harmful inactivity is treated as a reason for governmental coercion the freedom of American citizens is in danger:

This came to be known as the ‘broccoli argument’. If government could impose the purchase of health insurance why can't it impose the purchase of healthy food like broccoli? What is most remarkable about this argument is that it sets aside all externalities arising from inaction, as a class, because of the fear that allowing them would lead to vast new powers for the state. Basically, this is a slippery slope argument: one starts with health insurance and ends with broccoli. But the normative justification in terms of individual freedom refers back to the ontological distinction: interferences with omissions are judged to be a priori more coercive than actions. The externalities arising from actions, but not from omissions, are therefore a justification for government intervention. Forcing someone to do what he was not planning to do himself should be considered unconstitutional.Footnote ²¹

In the administration's view, the mandate was essential to the legislation as a whole (together with the obligation for insurers to accept all citizens, irrespective of prior ‘existing conditions’). Only if all individuals would insure themselves, the pool would contain good and bad risks, and health care would become affordable. If young and healthy people would be able to opt out, the pool would contain too many bad risks and prices would go up. The minority opinion (written by Ginsburg) expressed this view. With respect to the ontological argument, it acknowledged that the uninsured are not presently active in the market. However, they become active predictably: ‘Unlike the market for almost any other product or service, the market for medical care is one in which all individuals inevitably participate. Virtually every person residing in the United States, sooner or later, will visit a doctor or other health care professional’ (Opinion of Ginsburg, p. 3).

The prospect of future participation doesn't render one a participant per se, but Ginsburg adduced evidence that future participation really is quite nearby and predictable when she writes that ‘more than 60% of those without insurance visit a hospital or doctor's office each year (. . .) Nearly 90% will within five years’ (ibid, p. 19). She argued that therefore it is reasonable to have one's definition of the market ‘encompassing all transactions virtually certain to occur over the next decade (. . .) not just those occurring here and now’. (ibid, p. 20). But however predictable, these figures still do not show ‘activity’ at the point at which people decide not to insure. Where Ginsburg thinks that the ontological argument can stand on its own, it seems to me that to make it effective she needed to bring in a normative argument as well, based on the consequences of non-insurance:

This statement contains the core of Ginsburg's normative argument: non-insurance leads to an unfairness, given the inevitable participation of the uninsured. In the economists’ argument, the uninsured represent an externality. Of course, Ginsburg does not morally blame the uninsured, as if they are free riding parasites on those who do pay. After all, most of them are uninsured because of poverty, and this background provides an unfairness which cancels the first one. Bracketing this background injustice for our purposes here, it is crucial that Ginsburg subsequently reads the first unfairness problem back into the ontological dispute: ‘Given these far-reaching effects on interstate commerce, the decision to forgo insurance is hardly inconsequential or equivalent to “doing nothing” (. . .) it is, instead, an economic decision Congress has the authority to address under the Commerce Clause’ (ibid, p. 16–7). For Ginsburg, then, the uninsured are active in the market. This forms the ontological basis for her normative argument in favor of inclusion of the uninsured.

This case illustrates all the themes I have stressed throughout this paper in exemplary fashion. Let's focus on three of the them: the market-making nature of externalities, the symmetrical nature of the parties in externality cases and the deeper normative ideals behind the economics which lead the court to take up the case in terms of basic interests.

First, the market-making nature of externalities analyses. The presentation of the arguments shows that on both sides of the dispute the parties present ontological arguments about the market and what it is to be active on a market. Whether or not one sees an externality, depends on one's view of what the relevant market is. However, these ontological arguments are deeply interwoven with normative arguments about the desirable boundaries of the health care market. Majority and minority defend fundamentally different normative ideals (negative freedom and fairness, respectively). The majority opinion accepts the excess costs of leaving people uninsured and thus does not see a relevant externality. The case is an excellent illustration of liberal enthusiasm about internalizing externalities to protect the interests of vulnerable parties, and conservative reluctance to do so because this establishes a supposed ‘market failure’ too quickly (see section 2). Now this presupposes that the externality analysis is unequivocally ‘on the side’ of the liberals and the Court's minority opinion. That would violate the Coasian idea that things are not so easy: the parties in an externality case are symmetrically placed.

Second, the symmetrical nature of externalities. The language of externalities was most clearly invoked on the side of the minority opinion. The quote above from Justice Ginsburg mentions externalities implicitly with its reference to freeriding. More explicitly, the analysis of a team of economists who sent an Amici Curiae Economic Scholars brief to the Court invoked the language of externalities:

The minority side does not have a monopoly on invoking the externality concept, however. Its externality analysis suggests that the uninsured are harming the interests of the insured. But are they? The insured receive health care treatments under the terms of their insurance, let's call this ‘level 1-care’. Their insurance premiums are higher than they could be, as a result of having to pay for ‘level 2-care’ given to the uninsured in the system (let's assume that level 2-care leads to lower health quality, since the uninsured typically appeal to the system only in a later stage, when their health conditions are deteriorated and they have to receive emergency care). But if the insured, tired of paying for others, succeed in enforcing that the uninsured buy health insurance, the latter could complain that this compromises their freedom to act in the market place. This was the majority opinion's argument, and its basis is the value of negative freedom for those who wish to forgo insurance. Thus, internalizing the externality that the minority opinion cares about relieves the insured but now imposes an externality on the uninsured. This is Coase's symmetry between both parties. We must therefore choose between these harms – only balancing which harm weighs heavier can resolve the case.

Third, this balancing exercise leads us to conceive of the matter in terms of basic interests. On the one hand, every person has a basic welfare interest in having his or her health maintained at some threshold level that is high enough for autonomous functioning. This is a very basic intuition, which can be cashed out in various terms, for example by qualifying health as a primary good or basic capability (Daniels, Reference Daniels1981; Reference Daniels, Brighouse and Robeyns2010). On the other hand, the freedom to make economic decisions is itself, arguably, a condition for autonomous living, which qualifies as a basic interest for Feinberg, Raz and others. The question is how to weigh these interests. The majority side's reasoning might be reconstructed as follows. Obamacare curtails consumers’ freedom of choice and this weighs heavily. Without Obamacare, the insured only face the inconvenience of higher prices, but no violation of an essential interest. While the uninsured do raise the premiums of the insured, they arguably do not lower their level of care (this would be different, perhaps, if the number of uninsured would be massively higher, putting a larger drain on the system). Hence, the conclusion is clear.

This position can be shown to be inconsistent, I believe. For it tacitly relies on one essential feature of the case: the social and professional norm which mandates health treatment of uninsured people (at least in case of emergency, i.e. ‘level-2 care’). Absent this norm, the uninsured would not be treated at all, since that is what normally happens to everyone who asks for a product without being able and/or willing to pay for it. Absent this norm, then, there would be no externality at all. Now some may willing to bite this bullet, abjure this norm and accept the consequences. High-risk taking young people may say they want to live without insurance and will forgo treatment when they fall ill. Also, some insured people may be willing to bite the bullet, take an unforgiven stance on the uninsured and be able to impose this stance on health providers. In general however, both the uninsured will (maybe with some feelings of shame, but still) be happy to be helped when necessary and the insured will generally not be so harsh as to seriously contemplate abolition of this norm. It is telling that those opposing Obamacare never argued in this direction. Apparently, they do believe this norm has to be upheld.

The implication of this acceptance is that providing a sufficient level of health care, because it is a basic interest, is a collective responsibility. Accepting the norm of treatment implies that one justifies the existence of a system (ultimately backed up by the state) that everyone shall receive treatment; even when health care is privately organized. But if something is a collective interest, then the usual way of providing for it is to have every citizen pay its share. So this implies that the state can coerce every citizen to pay for health care, since it has been previously established to be in every citizen's interest to receive health care (when needed). Health care is a public good and the ‘freedom to purchase’ only applies after all justified public goods have been identified that need to be provided collectively. If this analysis is correct, then the majority opinion's argument is inconsistent. It should either accept leaving uninsured citizens without treatment, or accept the implications of the moral norm that those who are ill have to be treated.

Much of the trouble about this case comes about because health care is not a public good in the technical economic sense (non-excludability). It can be provided as a discrete good by markets. Therefore, there is no strict need for direct public provision by the state. The tension is between the demands of a shared basic interest of all citizens and the logic of the market where freedom to buy is paramount. Indeed, if the state would have decided to provide health care as a national service, then the argument about coerced payment could not have been made. That would simply have been a matter of taxation.Footnote ²⁴ Health care markets are an example of quasi-markets (Le Grand, Reference Le Grand2007): markets for which the state creates a competition between providers, but buyers are subsidized so that everyone is able to afford the product. It is the quasi-aspect that causes the trouble. While the uninsured are subsidized under Obamacare if they cannot afford to pay themselves, they are also forced to insure. This is not so different, in the end, from subsidized schooling, where one is simultaneously forced to buy the product (given the social norm that minors should go to school and be educated as citizens).

In the end, then, I would argue that a basic interest-analysis support the minority side of the dispute. Whatever of that, the larger point is that the analysis of externalities needs to be made in terms of basic interests. The Supreme Court's reasoning reveals this feature and so helps us to remind that we need to fill in economic concepts such as externalities with substantial concepts whose justification presupposes a philosophical theory of basic human interests.