ESHET Conference 7-9 June 2018-05-08 Complutense University of Madrid

John Stuart Mill and the concept of externality: what counts as a harm

Elodie BERTRAND1

Draft version

The history of the concept of externalities is necessarily also a history of the ethics of externalities. In this perspective, the philosophy of John Stuart Mill is inescapable. His principle limiting private liberty to what does not harm others is regularly referred to in the analyses of externalities (e.g., Chauvier 2013). Even more, this principle is sometimes interpreted as pointing at externalities (e.g., Medema 2009). The present paper aims at assessing the role of Mill’s harm principle in the definition of the concept of externality. It will interpret On Liberty (Mill 1859) and The Principles of Political Economy (1871, V.11) with the lens of the modern definition of externalities. Mill’s principle does not aim at involuntary harm, hence does not aim at externalities. However, it is true that On Liberty is crucial to the economists’ modern definition of externalities, but for another reason, insufficiently explored. Other types of externalities can be found in Mill: indirect harms that, in his view, do not justify public interference, and that correspond to external effects that are excluded from the standard definition of externalities. These are “moral” externalities, these effects of our actions that harm others by injuring their morality. I conclude that Mill does not refer to externalities as market failures, but that he gave welfare economists the philosophical background that justifies them in excluding moral externalities while their view of welfare as preferences satisfaction should lead to this concept. What counts as an externality rests on the liberal harm principle. The economists’ concept of externalities is grounded in Mill, but not in the sense Mill would write about externalities, rather because economists exclude from externalities the same moral effects that Mill excludes from the harm principle.

1 CNRS, Institut des Sciences Juridique et Philosophique de la Sorbonne, UMR 8103 (CNRS & University Paris 1 Panthéon-Sorbonne), [email protected] I thank Sophie Jallais, Nathalie Sigot and Regis Servant for their comments and suggestions.

1 1. Introduction Externalities pose specific problems to ethics. Moral philosophers agree that there is a duty not to harm, a principle that John Stuart Mill developed particularly in On Liberty (1859). Externalities, however, cause a peculiar harm, which is involuntary and indirect: it arises from the conjunction between an otherwise legitimate action and a specific state of the world that makes this action cause this harm to others. Does this harm become legitimate from the moment it is involuntary? Or does the harm to others taint the action that causes it with its illegitimacy? In his Ethique sans visage (Ethics without Face, 2013), the philosopher Stéphane Chauvier poses the problem in these terms, which bring to light two elements. First, the history of the concept of externalities is necessarily also a history of the ethics of externalities: thinking about externalities necessarily leads to think about its legitimacy or about the legitimacy of forbidding the action that causes it. Second, when reflecting about externalities, and therefore about their ethics, the philosophy of John Stuart Mill is inescapable. It so happens that his principle limiting private liberty to what does not harm others is regularly referred to in the analyses of externalities. Chauvier (2013) applies some parts of On Liberty to build his ethics of externalities. The jurist Michael Trebilcock parallels the delineation of Mill’s harm to others and the delineation of externalities: “What should count as an externality in welfare terms, or what should count as a harm to others in liberal terms?” (1993, 58-59). As for economics, Kenneth Arrow (1969) and Ezra Mishan (1971a; 1974), two important authors for the definition of the concept of externalities, made passing references to Mill. Regarding now the historical and philosophical analyses of the concept, the link between Mill and the concept of externalities is rarely developed. Donald Herzog (2000), professor in political theory and public law, suggested in a book review the important intuition that the borders of the externalities that economists take into account – technological but not moral – were grounded in Mill’s liberalism. Moreover, the sociologist James Coleman interprets Mill’s principle as applying to externalities (1990, 343). In his Hesitant Hand (2009), Steven Medema asserts that Mill’s harm principle directly aims at externalities: they would be included into the “harm to others” that public interference is legitimate in interfering with. Medema adds that Mill would also discuss a list of positive and negative externalities in the chapter of his Principles that delineates a legitimate sphere of economic intervention for the State.2 The present paper therefore aims at assessing the role of Mill’s harm principle (as expanded in On Liberty, 1859) in the definition of the concept of externality. This will lead to challenge a part of Medema’s interpretation and develop Herzog’s intuition. Analysing this role also contributes to the history of the concept (both in philosophy and in economics), and to the analysis of some philosophical presuppositions of the economic concept. In On Liberty (I, 9), Mill states from the start that the object of his book is to establish what will be later known as the “harm principle” (or “liberal principle”), which allows limiting liberty in case of harm to others (and only in this case):

2 In what follows, that public intervention is legitimate does not necessarily implies that it is necessary or that it is the only solution. Mill’s harm principle is not a sufficient principle; and when Mill opens a domain for public intervention in the Principles, he also adds that in each case, one has to prove that it is more efficient (see references below and Medema 2009, chapter 2).

2 The object of this Essay is to assert one very simple principle, as entitled to govern absolutely the dealings of society with the individual in the way of compulsion and control, whether the means used be physical force in the form of legal penalties, or the moral coercion of public opinion. That principle is, that the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number, is self-protection. That 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. In accordance with the modern and standard definition of technological externalities, the New Palgrave defines them as “the indirect effect of a consumption activity or a production activity on the consumption set of a consumer, the utility function of a consumer or the production function of a producer. By indirect, we mean that the effect concerns an agent other than the one exerting this economic activity and that this effect does not work through the price system” (Laffont 1987, 263). The modern definition excludes pecuniary externalities and other utility interdependences (like envy). To proceed, this paper interprets Mill’s On Liberty (1859, OL herein) and book V (precisely ch. 11) of The Principles of Political Economy (1871 for the 7th edition, PPE herein) with the lens of the modern definition of externalities. On the one hand, Mill’s harm principle does not aim at involuntary harm, hence does not aim at externalities. Examining other passages from his works that are sometimes interpreted as referring to externalities, I will argue that once put in their contexts, it must be recognized that they do not aim at externalities. On the other hand, however, it is true that On Liberty is crucial to the economists’ modern definition of externalities, but for another reason, insufficiently explored. Other types of externalities can be found in Mill: indirect harms that, in his view, do not justify public interference, and that correspond to external effects that are excluded from the standard definition of externalities. These are “moral” externalities, these effects of our actions that harm others by injuring their morality. They were excluded while welfare economists restrained the concept of externalities to technological externalities. But if economists were consistent in their view of welfare as satisfaction of preferences, they should take into account these moral externalities (which diminish utility). The implicit (and sometimes explicit) reason for excluding moral externalities from the analysis was, precisely, that, in chapter IV of On Liberty, Mill eliminates them from the domain of legitimate intervention. Consequently, there is no evidence that On Liberty concerns phenomena denoted today by “technological externalities”. But the economists’ exclusion of moral externalities from the domain of externalities relevant to welfare originates from their liberalism, anchored in Mill’s discussion of the harm principle. What counts as an externality rests on the liberal harm principle. Mill is important for the construction of the concept, but not (or not only) for the reasons invoked by Medema; it is rather for what he excludes from harms to others legitimating liberty limitations than by what he includes in these harms.

3 2. Externalities are not included in Mill’s harm to others

The domain of externalities relevant to welfare economics The discussion on externalities took off in the 1950s and focused on the disparity between general equilibrium and Pareto-optimum that they cause, with Meade 1952, Scitovsky 1954, and Bator 1958 (see Mishan 1971b, Medema 2017, Berta 2017). Economists first considered this concept theoretically, as a source of suboptimality, then they realized that the concept had a real counterpart, namely pollution. The history of the concept of externality from the 1950s to the 1970s is thus a story of its restriction, under a double constraint. On the one hand, the concept was increasingly associated with a specific phenomenon, that of pollution (Papandreou 1994, Berta 2017, Medema 2017). On the other, since it was associated with sub-optimality and hence public intervention, it was urgent to circumscribe the concept that could otherwise formally encompass almost every interaction (Berta 2017). A relative consensus identifies the following features for characterizing externalities (Berta 2017): The concept was first limited to technological externalities, that is, unpriced and hence uncontrollable effects. This excludes so-called pecuniary externalities (Viner 31; Scitovsky 54). We find the unpriced feature of externalities in most of the classic articles (Meade 1952, Scitovsky 1954, Bator 1958, Mishan 1965) until Arrow’s formalization of externality as an exogenous variable (agent A’s utility) in agent B’s utility function (Arrow 1969): it is external to price and hence external to her optimization program (see Berta 2017). This means that externalities are suboptimal. Indeed, when the conditions of optimality of general equilibrium prices were formalized, externalities were proven to cause inefficiencies (Medema 2009, 72). Mishan could then argue that “externalities today provide the standard exception to the equation of optimality with universal perfect competition” (1971b, 1). The formalisation of externalities as interdependences between production and/or utility functions, however, may lead to a confusion with other concepts that are excluded from the phenomenological definition. To exclude malevolence, this definition must add a second feature, that externalities are involuntary. Mishan notes: “What the notation alone does not succeed in conveying, however, is that the essential feature of the concept of an external effect is that the effect produced is not a deliberate creation but an unintended or incidental by-product of some otherwise legitimate activity” (1971b, 2; see also Mishan 1969). I will spend more time on the last feature. The formal definition does not either translate that economists focus on technological externalities, which excludes from the domain of relevant externalities other-regarding preferences, that is other forms of utility interdependences, like envy. Envy and other Veblen effects were soon excluded from the analysis, while encompassed in the formal definition (Medema 2017).3 For example, they are excluded from Mishan’s analysis because, he argued, there is a social consensus that public intervention is not legitimate in that case (see also Berta 2017, 312-3): Although the interdependent-utility case may be expressed in the same functional form as external effects on consumption, it may be expedient to

3 On envy as an externality, see Hammond 1987b and Pignol 2012. In the same line, altruism can also be seen as an externality (Hammond 1987a), but this is less often mentioned.

4 exclude its classification under external effects. … In proposing [welfare] criteria economists have drawn on basic values that are apparently widely acceptable in the Western communities for which such criteria are fashioned. One can well imagine that the costs of the more tangible external diseconomies, such as obstruction, noise, and smoke, are readily accepted as such by a consensus. Indeed, anyone who is able to rationalize the working of the market cannot consistently ignore external effects. If, on the other hand, a man were bold enough to complain that the mere fact of others becoming better off saddened him considerably, we may have some sneaking sympathy with him. But it is hardly likely that practical measures would be contemplated which were calculated to impoverish others in the hope of restoring his spirits. A widely acceptable social welfare function, that is, could quite conceivably take cognizance of all the Pigovian external effects while excluding the pangs of envy or ill-feeling induced by the knowledge of other people’s good fortune. … It might be universally agreed that such inescapable differences did give rise to feelings of resentment, and no less universally agreed that this was a vice and not a virtue and should not therefore be included in a social welfare function. (Mishan 1965, 7-8) Mishan hence excludes what he calls elsewhere these “imaginary external effects”, caused by, for example, envy (1969, 340, fn 9). And he focuses on “tangible effects”, those that are suboptimal according to a consensual welfare criterion, which therefore puts some preferences aside for the evaluation of welfare. The same exclusion by consensus applies to moral externalities: “If the mere sight of a person, or the mere knowledge of his existence or behaviour, offends me, it is not likely that society would support my claim to damage” (Mishan 1971a, 119, fn 1). Moral externalities refer to a specific type of involuntary and unpriced effect of a contract on a third party: a moral effect, an embarrassment, a trouble, an offense. Some persons are offended (their utility is diminished) as a result of someone else’s behaviour. This kind of externality is mentioned by Bator alongside other consumption externalities that he excludes: “Bator was also rather unconcerned that his model had ruled out consumer-side external effects – including ‘such phenomena as Y tossing in sleepless fury due to X’s “consumption” of midnight television shows; or X’s temperance sensibilities being outraged by Y’s quiet and solitary consumption of Scotch’” (Medema 2017, 26 quoting Bator 1957, 42, my emphasis). To sum up, the standard definition of externalities as relevant to welfare economics refers to them as unpriced, suboptimal, involuntary and physical (or technological, or physiological).

Mill’s place in the history of the concept The history of the concept of externality generally begins with Arthur Cecil Pigou, who characterized this way the second type of divergence between private and social costs that he identifies: “[O]ne person A, in the course of rendering some service, for which payment is made, to a second person B, incidentally also renders services or disservices to other persons (not producers of like services), of such a sort that payment cannot be exacted from benefited

5 parties or compensation enforced on behalf of the injured parties” (1932, 183).4 Several features of the future concept of externalities are already present: a contract causes an effect (positive or negative) on a third party; this effect is unpriced (“uncompensated” or “uncharged”, 183, 185) and involuntary (“incidental”). The examples given by Pigou of this type of divergence encompass, however, other phenomena than proper externalities, like public goods and Veblen effects. He cites, among others: lighthouses, smoke prevention, scientific research, the building of a factory in a residential area, or the production and sale of alcohol (extra cost in policemen and prison) (184-188). Because of these divergences between private cost and social cost, the national dividend is not maximized, hence “there is a prima facie case for public intervention” (331), through regulation, taxation and subvention. Since government intervention entails its own difficulties, “[t]he case, however, cannot become more than a prima facie one” (331-332).5 This is the usual story, but in his Hesitant Hand, Medema draws a line from Mill to Pigou through Henry Sidgwick and Alfred Marshall. He argues that Mill clearly identifies a list of instances where the system of natural liberty fails, but that he is still close to classical political economics in his view of the defects of state intervention. Sidgwick, still in Medema’s view, will suggest an even more extensive list, but with more optimism in the state’s ability to correct these failures. Finally, Pigou, influenced by Sidgwick and Marshall, will propose the first neo-classical theory of market failures. More precisely, Medema asserts that “Mill was clearly arguing here for freedom of individual action where spillovers – or externalities are absent” (2009, 35). He takes support from different formulations of the harm principle, for example this one, extracted from the famous first formulation of the harm principle already quoted: “the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number, is self-protection. That 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” (OL I, 9, quoted by Medema 35, fn 25). Medema also uses formulations of the harm principle and definition of private sphere that one can find in Mill’s Principles (PPE V, 11, 4-5, see Medema 2009, 34 and 35, fn 26). The harm principle there limits authoritative interference to free trade, that is interference through obligation or ban. In his recent article on the “rise, fall, and rebirth” of the economic theory of externalities, Medema begins with Mill (as well as Sidgwick and Pigou), characterizing his approach in the PPE as “phenomenological” in the sense that it would associate “externalities with specific real phenomena” (Medema 2017, 30). The present paper will not look for a filiation between Mill and Pigou, which exists and is historically evidenced, and which is visible in the proximity of some arguments (prima facie intervention) and examples (lighthouses, scientific research, alcohol, see below). What I am looking for is the equivalent in Mill of the modern concept of externality, as defined above.

4 The fourth edition of The Economics of Welfare (Pigou 1932) reiterates his 1912 typology of three kinds of divergence between private and social costs. The other types are: agency problems (typically land tenancy contracts), and increasing and decreasing returns (Marshallian “external economies”). The term “externalities” appears for the first time in print in Bator 1957 (see Medema 2017, 19). 5 On the problems with public intervention that Pigou identifies, see Medema 2009, Aslanbeigui and Medema 1998, Bertrand 2010.

6 Does Mill’s harm principle aim at externalities? I have reminded the first formulation of Mill’s harm principle in On Liberty. This paragraph continues: His own good, either physical or moral, is not a sufficient warrant. He cannot rightfully be compelled to do or forbear because it will be better for him to do so, because it will make him happier, because, in the opinions of others, to do so would be wise, or even right. These are good reasons for remonstrating with him, or reasoning with him, or persuading him, or entreating him, but not for compelling him, or visiting him with any evil in case he do otherwise. To justify that, the conduct from which it is desired to deter him, must be calculated to produce evil to some one else. The only part of the conduct of any one, for which he is amenable to society, is that which concerns others. In the part which merely concerns himself, his independence is, of right, absolute. Over himself, over his own body and mind, the individual is sovereign. (OL I, 9) With this principle, Mill aims at defending three types of liberty (OL I, 12): of conscience and expression (see OL II), of tastes and pursuits (see OL III and IV), and of association. His harm principle does not apply – which means liberty can be restricted for other reasons – to children and some populations (“barbarians”) themselves considered as in their infancy (OL I, 10).6 Mill protects these liberties against other justifications for public interference, which it has became usual to name, after Feinberg four-volumes The Moral Limits of Criminal Law (1984, 1985, 1986, 1988): offense principle (prevent an offense, as opposed to injury or harm7), legal paternalism (prevent harm to the person acting) and legal moralism (prevent immoral conduct without harm or offense). Mill authorizes interference only when there is harm done to someone else (harm principle).8 The application of the harm principle defines a private sphere: there is a sphere of action in which society, as distinguished from the individual, has, if any, only an indirect interest; comprehending all that portion of a person’s life and conduct which affects only himself, or if it also affects others, only with their free, voluntary, and undeceived consent and participation. When I say only himself, I mean directly, and in the first instance: for whatever affects himself, may affect others through himself; and the objection which may be grounded on this contingency, will receive

6 Note that harm may also be understood as a lack of good: “There are also many positive acts for the benefit of others, which he may rightfully be compelled to perform; such as, to give evidence in a court of justice; to bear his fair share in the common defence, or in any other joint work necessary to the interest of the society of which he enjoys the protection; and to perform certain acts of individual beneficence, such as saving a fellow-creature’s life, or interposing to protect the defenceless against ill-usage, things which whenever it is obviously a man’s duty to do, he may rightfully be made responsible to society for not doing. A person may cause evil to others not only by his actions but by his inaction, and in either case he is justly accountable to them for the injury” (OL I, 11). 7 Feinberg’s offense principle defines an offense as a disliked mental state caused by the wrongful conduct of another person. 8 There is a debate on whether Mill also accepts some version of the offence principle: see Feinberg 1985, Trebilcock 1993, Brink 2014.

7 consideration in the sequel. This, then, is the appropriate region of human liberty. (OL I, 12) This private sphere includes what affects others with their consent: Mill adheres to the volenti non fit injuria maxim (regarding the affection of others, it will be dealt with in section 3). Mill rejects paternalism and moralism for utilitarian reasons: “The only freedom which deserves the name, is that of pursuing our own good in our own way, so long as we do not attempt to deprive others of theirs, or impede their efforts to obtain it. Each is the proper guardian of his own health, whether bodily, or mental and spiritual. Mankind are greater gainers by suffering each other to live as seems good to themselves, than by compelling each to live as seems good to the rest” (OL I, 13). It is, however, a perfectionist utilitarianism (Brink 2014), in which utility includes a person’s interest in the development of her capacities: “It is proper to state that I forego any advantage which could be derived to my argument from the idea of abstract right, as a thing independent of utility. I regard utility as the ultimate appeal on all ethical questions; but it must be utility in the largest sense, grounded on the permanent interests of man as a progressive being. Those interests, I contend, authorize the subjection of individual spontaneity to external control, only in respect to those actions of each, which concern the interest of other people” (OL I, 11). This utilitarian criterion implies that the harm principle defines a prima facie case for intervention, but only a prima facie one: “If any one does an act hurtful to others, there is a prima facie case for punishing him, by law, or, where legal penalties are not safely applicable, by general disapprobation” (OL I, 11). Harm is necessary but not sufficient to justify public interference, which has to be proven causing more benefits than harms. The principle does not apply when “the attempt to exercise control would produce other evils, greater than those which it would prevent” (OL I, 11). Does Mill’s harm principle concern externalities, i.e. involuntary harm done to others? The harm principle does not aim at unintended harms to third parties such as those invoked by the concept of externality. One of its formulations (already quoted) makes explicit the voluntariness of the harm that is aimed at: “To justify that, the conduct from which it is desired to deter him, must be calculated to produce evil to some one else” (OL I, 9, my emphasis). For example, the liberty of expression does not extend to the expression of opinions that is directed at causing harming actions: “An opinion that corn-dealers are starvers of the poor, or that private property is robbery, ought to be unmolested when simply circulated through the press, but may justly incur punishment when delivered orally to an excited mob assembled before the house of a corn-dealer” (OL III, 1).9 Here the effect (harming action) is clearly intended while externalities are involuntary. Chapter IV draws a list of acts “injurious to others”, which again refer to direct and voluntary acts: “Encroachment on their rights; infliction on them of any loss or damage not justified by his own rights; falsehood or duplicity in dealing with them; unfair or ungenerous use of advantages over them; even selfish abstinence from defending them against injury” (OL IV, 6).

9 This context sheds another light to the sentence of the same passage quoted by Medema (2009, 35): “The liberty of the individual must be thus far limited; he must not make himself a nuisance to other people” (OL III, 1). It does not apply to involuntary and indirect effects, but concerns liberty of expression when expression is directed at causing harm.

8 What about involuntary and unavoidable private interferences? Another passage of On Liberty is partially quoted by Medema and Chauvier as invoking externalities. In this passage, Mill insists that the harm principle is not sufficient to restrain liberty since, in some cases, harm does not justify public interference: [I]t must by no means be supposed, because damage, or probability of damage, to the interests of others, can alone justify the interference of society, that therefore it always does justify such interference. In many cases, an individual, in pursuing a legitimate object, necessarily and therefore legitimately causes pain or loss to others, or intercepts a good which they had a reasonable hope of obtaining. Such oppositions of interest between individuals often arise from bad social institutions, but are unavoidable while those institutions last; and some would be unavoidable under any institutions. (OL V, 3) Medema (2009, 36) uses the second sentence of this quote in support of his interpretation that “Mill was cognizant of the ubiquity of these external effects”, and the first to insist that, in Mill’s view, public intervention is not necessarily called for when externalities are present. In Medema’s interpretation, therefore, externalities are both in the harm principle and in this passage. Chauvier only retains that “[i]n many cases, an individual, in pursuing a legitimate object, necessarily and therefore legitimately causes pain or loss to others” (quoted in French in Chauvier 2013, 82). In his view, externalities are not included in the harm principle which forbids any harm (and we cannot forbid every externalities), but they are included in this passage, which legitimates some unavoidable nuisances. He argues that the involuntary aspect (“external” in Chauvier’s words) of the nuisance implies that we cannot apply the harm principle in this case, and that we have to develop a specific and systematic ethics of external nuisances. However, the private interferences that are at the core of this passage are not externalities. Mill’s paragraph, because of the interpretations that are made of it, deserves to be quoted entirely. I have already quoted the first part. It goes on like this: Whoever succeeds in an overcrowded profession, or in a competitive examination; whoever is preferred to another in any contest for an object which both desire, reaps benefit from the loss of others, from their wasted exertion and their disappointment. But it is, by common admission, better for the general interest of mankind, that persons should pursue their objects undeterred by this sort of consequences. In other words, society admits no right, either legal or moral, in the disappointed competitors, to immunity from this kind of suffering; and feels called on to interfere, only when means of success have been employed which it is contrary to the general interest to permit—namely, fraud or treachery, and force. (OL V, 3)10

10 Medema quotes this passage to illustrate that “[o]ne class of spillovers that Mill ruled out of bounds for government on expediency grounds is those that naturally and inevitably occur within the context of a competitive market system. Life, according to Mill, is often a zero-sum game, as in, for example, when A gets the job and B does not. But, he said, society correctly ‘admits no right… force’” (Medema 2009, 37, fn 29).

9 Once in their context, it becomes clear that the first sentences of V, 3 do not refer to externalities. The negative effects referred to here are indeed involuntary, but they result from the normal functioning of the institution. Even more, the institution in question (competition, contest) is designed to create this effect, here to select winners, and hence create losers, whose interests are necessarily harmed. In this case, a person A may cause involuntary nuisances to another person B, but this private interference with B’s liberty in the exercise of A’s own does not legitimate public interference with A’s liberty.11 This legitimate private interference is the logical equivalent to: A buys a property right that B, hence, can no longer buy. And this may be why Mill carries on his discussion with commerce, which suspends the application of the harm principle (OL V, 4). Harm is not necessary to justify public interference because “trade is a social act”, and a seller’s action hence “comes within the jurisdiction of society”. Free Trade has been proven to be more efficient (in the sense that competition permits cheapness and quality of commodities), and “restrictions on trade” such as “prevention of fraud” or “arrangement to protect workpeople” are morally legitimate, but are judged according to their results. Mill then discusses the regulation of some peculiar markets (poisons, alcohol, gambling, sex, slavery) (OL V, 8-10). In this discussion, some elements refer to the effects of contracts on third parties. Mill adds that some contracts entail duties towards the third parties on whom they have some consequences: “if the relation between two contracting parties has been followed by consequences to others; if it has placed third parties in any peculiar position, or, as in the case of marriage, has even called third parties into existence, obligations arise on the part of both the contracting parties towards those third persons, the fulfilment of which, or at all events the mode of fulfilment, must be greatly affected by the continuance or disruption of the relation between the original parties to the contract” (OL V, 11). It is not clear, however, whether these consequences are involuntary. Unfortunately the only example Mill gives is that of child as third parties to the marriage contract, of which, even if we cannot say that they are totally voluntary at that period, we cannot negate that they were supposed to be the intention of the contract in question. However, Mill does evoke the effects of the education of these children on society, which are unintended at the individual level: “to bring a child into existence without a fair prospect of being able, not only to provide food for its body, but instruction and training for its mind, is a moral crime, both against the unfortunate offspring and against society” (V, 12). The mention of a crime against society, and not only against the child, looks like a positive externality of education of one’s children on entire society. But this is developed in Mill’s Principles, and I now turn to this argument.

Are they positive externalities in Mill’s Principles? In chapter 11 of book V of his Principles, Mill gives a list of cases where the pursuit of self- interest does not lead to general interest, and in which public intervention is legitimate (if efficient). Medema interprets this list as a list of “the limitations of natural liberty – market failures, if you will”, which “fall into both negative externality and positive externality (including public goods) categories” (Medema 2009, 40, fn 34). Brink (2014) also recognizes

11 Mill could also have invoked the volenti maxim: losers in a competition or a contest have accepted the possibility of losing when participating in (Feinberg 1984, 220).

10 in this list different types of market failures, including positive externalities: Mill “thinks that there are other goods for which market provision will lead to underproduction, presumably because of positive externalities, which is why he thinks that the state should subsidize scientific research and the arts.” In my view, however, the problems mentioned by Mill fall into different categories including, indeed, public goods, but not externalities in and by themselves.12 Mill distinguished two types of state intervention: authoritative interference (obligations or bans) and agency interference (when the state directly provides some goods or services, or provides information and advices) (V, 11.3). The former is limited by the harm principle, as evoked above (V, 11.4-5). With the latter, “there is no infringement of liberty” (V, 11.6), but other problems emerge: this intervention of the state is financed by taxation (6), it represents “an increase of its power” (7), and a new “occupation imposed upon a body already overcharged with duties” (8), but it also implies doing at the person’s place, which does not help her: “The business of life is an essential part of the practical education of a people” (13). Mill’s main argument is nevertheless the following: “the great majority of things are worse done by the intervention of government, than the individuals most interested in the matter would do them, or cause them to be done, if left to themselves. The grounds of this truth are expressed with tolerable exactness in the popular dictum, that people understand their own business and their own interests better, and care for them more, than the government does, or can be expected to do” (11). Because of these information and incentive problems, public interference should be restricted to what is absolutely necessary, and “the burthen of making out a strong case” should lie on “those who recommend, government interference. Laisser-faire, in short, should be the general practice: every departure from it, unless required by some great good, is a certain evil” (16).13 A first exception is mentioned already at that stage: But if the workman is generally the best selector of means, can it be affirmed with the same universality, that the consumer, or person served, is the most competent judge of the end? Is the buyer always qualified to judge of the commodity? If not, the presumption in favour of the competition of the market does not apply to the case; and if the commodity be one, in the quality of which society has much at stake, the balance of advantages may be in favour of some mode and degree of intervention, by the authorized representatives of the collective interest of the state. (22) Education is the typical example, for parents who do not educate their children “commit a double breach of duty, towards the children themselves, and towards the members of the

12 Externalities are implicitly linked to public goods since Samuelson 1958 (see Berta 2017, 293-294; Medema 2017, 21; Cornes and Sandler 1986). The two problems are different. For example, at the optimum, there are no longer externalities (since they are no longer suboptimal), but public goods are produced. 13 Or again: “We have observed that, as a general rule, the business of life is better performed when those who have an immediate interest in it are left to take their own course, uncontrolled either by the mandate of the law or by the meddling of any public functionary. The persons, or some of the persons, who do the work, are likely to be better judges than the government, of the means of attaining the particular end at which they aim. Were we to suppose, what is not very probable, that the government has possessed itself of the best knowledge which had been acquired up to a given time by the persons most skilled in the occupation; even then, the individual agents have so much stronger and more direct an interest in the result, that the means are far more likely to be improved and perfected if left to their uncontrolled choice” (PPE V, 11.22).

11 community generally, who are all liable to suffer seriously from the consequences of ignorance and want of education in their fellow-citizens” (24). Medema (2009, 39) argues that this argument refers to the (absence of) positive externalities of education. And it is correct that Mill here justifies intervention because harm (here breach of duty) is sometimes caused by inaction (see fn 6). Mill then repeats the general rule: “The ground of the practical principle of non- interference must here be, that most persons take a juster and more intelligent view of their own interest, and of the means of promoting it, than can either be prescribed to them by a general enactment of the legislature, or pointed out in the particular case by a public functionary” (PPE V, 11.29). This rule admits “very large and conspicuous exceptions” (29), when its ground is not verified, i.e. when individuals are not “the best judges of their own interest” (33), or do not have the means to promote it, or to make it effective. The complete list of Mill’s exclusions is as follows: 1) child, idiots, insane persons, slaves, lower animals (30-31), all cases where “[t]he person most interested is not the best judge of the matter, nor a competent judge at all” (30); 2) irrevocable long-term contracts, in which case the volenti maxim does not suffice to make the contract legitimate (33). 3) When the concern is managed by delegated agency, government management may be more efficient than joint-stock management (34). It is particularly the case for infrastructures, and what would be called today natural monopolies: There are many cases in which the agency, of whatever nature, by which a service is performed, is certain, from the nature of the case, to be virtually single; in which a practical monopoly, with all the power it confers of taxing the community, cannot be prevented from existing. … [T]he community needs some other security for the fit performance of the service than the interest of the managers; and it is the part of the government, either to subject the business to reasonable conditions for the general advantage, or to retain such power over it, that the profits of the monopoly may at least be obtained for the public. This applies to the case of a road, a canal, or a railway. (36) 4) Public interference is also needed to maintain collective agreements despite opposed individual interests; the argument is similar to the free rider problem for financing public goods or enforcing non individually rational agreements: “There are matters in which the interference of law is required, not to overrule the judgment of individuals respecting their own interest, but to give effect to that judgment: they being unable to give effect to it except by concert, which concert again cannot be effectual unless it receives validity and sanction from the law” (37). Mill gives two examples: a limitation of working hours, and a limitation of the surface of lands that can be appropriated in new colonies: “It is the interest of each to do what is good for all, but only if others will do likewise” (39). 5) Finally, the general rule cannot either “apply to the very large class of cases, in which those acts of individuals with which the government claims to interfere, are not done by those individuals for their own interest, but for the interest of other people” (41). Mill distinguishes

12 three sub-cases: charities (41)14, political effects on the nation of economic individual decision – like colonization (48)15, and public goods. The last sub-case deserves to be detailed here, “in which important public services are to be performed, while yet there is no individual specially interested in performing them, nor would any adequate remuneration naturally or spontaneously attend their performance” (57). Examples are given to exploration missions, lighthouses (its light being non-excludable)16, scientific research “of great value to a nation and to mankind, requiring assiduous devotion of time and labour, and not unfrequently great expense, by persons who can obtain a high price for their services in other ways” (57). Mill here clearly refers to a general interest that may be different from self-interest: “anything which it is desirable should be done for the general interests of mankind or of future generations, or for the present interests of those members of the community who require external aid, but which is not of a nature to remunerate individuals or associations for undertaking it, is in itself a suitable thing to be undertaken by government” (59). This list is not only an “extensive list of divergences between private and social interests” (Medema 2009, 40); but the crux of the matter is a difference between, on the one hand, self-interest as perceived by the agent and, on the other, either a true, long-term self- interest of the person as a member of the community or, indeed, the community’s interest. Some externalities are mentioned (e.g. of education), but they tend more towards the problems of public goods (non rivalry, non excludability, problem of free rider), which are clearly expressed.

To conclude this section, and as said above, I am not questioning the intellectual filiation between Mill and Pigou about their identification of the problems of the system of natural liberty. I am not denying that by his influence on Sidgwick, and through him, on Marshall and Pigou, Mill did not contribute to the emergence of the concept of externalities. But the concept of externality is not included in the harms to others that, in Mill’s liberalism, legitimate public interference. The closest problem Mill examines is that posed by the (absence of) positive effects of some public goods. Although the harm principle can be applied to the ethics of externalities17, spillovers are not the harms to others that Mill aims at

14 “Though individuals should, in general, be left to do for themselves whatever it can reasonably be expected that they should be capable of doing, yet when they are at any rate not to be left to themselves, but to be helped by other people, the question arises whether it is better that they should receive this help exclusively from individuals, and therefore uncertainly and casually, or by systematic arrangements, in which society acts through its organ, the state” (PPE V, 11.41). Follows a discussion of the Poor Laws. 15 “[T]he acts done by individuals, though intended solely for their own benefit, involve consequences extending indefinitely beyond them, to interests of the nation or of posterity, for which society in its collective capacity is alone able, and alone bound, to provide. One of these cases is that of Colonization. If it is desirable, as no one will deny it to be, that the planting of colonies should be conducted, not with an exclusive view to the private interests of the first founders, but with a deliberate regard to the permanent welfare of the nations afterwards to arise from these small beginnings; such regard can only be secured by placing the enterprise, from its commencement, under regulations constructed with the foresight and enlarged views of philosophical legislators; and the government alone has power either to frame such regulations, or to enforce their observance” (PPE V, 11.48). Mill then discusses the benefits of colonization. 16 On this famous example, see Coase 1974 ; Bertrand 2006. On the controversy that issued, see Bertrand 2017. 17 This is what Mishan does, for example, to rebut Coase’s (1960) reciprocity thesis: “But a situation may be Pareto symmetrical without being ethically symmetrical. In other words, with respect to the mutuality of conflict it is symmetrical, but it is not symmetrical with respect to the equity. In accordance with the liberal maxim, the freedom of any man to smoke what he chooses, when he chooses, and where he chooses, would indeed be

13 with this principle. By contrast, there is a moral harm that Mill excludes from the application of the harm principle that is similar to the moral effects that the economists exclude from their externalities.

3. (A part of) what is excluded from externalities rests on Mill

The exclusion of moral externalities from the domain of externalities relevant to welfare Remember that welfare economists have had a tendency to restrict the concept of externality and exclude interdependences of utilities, and with them moral externalities. They faced a problem with the concept of externality that could be extended with no limits: “As social life is obviously full of ‘unpriced’ individual interactions, there is no limit to the application of externality. This is why Baumol and Oates, as well as Mishan, call upon a more traditional meaning, close to environmental effects whose importance was more and more recognised in the 1960s. They try to limit these various interpretations to unintended effects, in order to avoid the fact that externality could encompass every unpriced individual interaction” (Berta 2017, 314). Economists had to restrict the contours of this concept, but on which basis? Moral externalities actually are unpriced, uncontrollable, and involuntary interactions, hence why not consider them as relevant externalities? Mishan’s solution to exclude other-regarding preferences eventually rested on a social consensus, of what counts or not. The relevance of the utility interdependency lays on the legitimacy of the offense felt, a legitimacy that is judged by moral standards. Mishan’s 1965 exclusion of interdependent utilities rested on “basic values that are widely acceptable”, “consensus”, “a widely acceptable social welfare function”. But this is not consistent with the view of welfare as satisfaction of preferences. The existence of preferences regarding others is a main objection to this view of welfare (see Hausman and McPherson 2009). By excluding this type of externalities, welfare economists recognize that there are preferences whose satisfaction does not increase well-being. It is inconsistent because they do not explain how they distinguish preferences that are acceptable or not, that enter or not the evaluation of welfare. If economists were consistent in their view of welfare, there would be no way of distinguishing between moral and physical externalities: in both cases, B’s utility is unintentionally diminished by A’s behaviour. Herzog also raises this inconsistency. He takes the example of a woman “horrified at the very thought that the lesbians down the hall are engaging in oral sex”, and assert that “as far as utility and preference go, here again we have externalities literally indistinguishable from those generated by … [a] polluting municipality” (Herzog 2000, 911). A coherent economist should deal with them with the same tools: “If the only conceptual resource we have on offer for thinking about such matters is preference, these are externalities” (Herzog 2000, 912).

conceded, but with the critical proviso that his smoking take place in circumstances which do not reduce the freedom or welfare of others. In so far as it does reduce the freedom or welfare of others, the freedom of the smoker to smoke is not symmetrical with the freedom desired by the non-smoker, since the freedom of the latter does not go beyond the breathing of fresh air. Unlike the freedom of the smoker, it does not reduce the amenity enjoyed by others” (Mishan 1974, 79).

14 Social choice theorists, as for them, because they discuss the content of welfare, face the question of which preferences are to be included in welfare, and hence face moral externalities. Typically, the other-regarding preferences that explain Sen’s paradox of a Paretian liberal (Sen 1970) can be interpreted as moral externalities: Lewd’s reading of Lady’s Chatterley Lover offenses Prude. 18 Peter Hammond puts physical and moral externalities on the same level: “the example of Lady Chatterley’s Lover which drives Sen’s… classic libertarian paradox illustrates how Pareto inefficiency arises from giving individuals the right to create externalities... One example of this important conflict between rights and externalities is driving private cars, especially on congested roads. Other examples that are more remote from economics include the abuse of religious freedom by those who preach intolerance, or the abuse of freedom of speech by those who disseminate hate” (1998, wp 2). But this is not how the standard analysis of externalities proceeds. Taking moral externalities into account offends (if I may) the majority of economists: putting moral and physical externalities on the same level (and hence opening the door to intervention in both cases) puts very different harms on the same level. At the moment we recognize that these are different harms, other categories enter into the analysis; which categories? The answer is given by Herzog when he remarks that economists do not invoke externalities (and their solutions) everywhere, and “that’s because economists are opportunistic about invoking externalities. They do so not whenever we find people with preferences about others’ preferences and actions; they actually do so in ways closely tracking the traditional harm principle of liberal theory… So it looks like the economists’ invocations of externalities are invidiously opportunistic, that the concept is secretly parasitic on the liberal harm principle” (Herzog 2000, 912, 914).19 It is correct that their will to distinguish between moral and physical externalities, and exclude the first from the analysis – and from the domain of public intervention, implicitly rests on Mill’s delineation of harms to others that justify interference with liberty.

Mill’s exclusion of moral externalities from the domain of legitimate intervention Chapter IV of On Liberty aims at defining the extent of the private sphere. Mill deals with case of a person harming herself, and completes his chapter III’s argument against paternalism. But he slips to the case of a person perceived as harming herself because perceived as acting contrary to morality, and here the argument slips against moralism. The chapter hence begins by the case where a person A does some harm to himself, and this affects other persons B. Against paternalism, and to defend that “neither one person, nor any number of persons, is warranted in saying to another human creature of ripe years,

18 For an interpretation of Sen’s paradox in terms of negative externalities, see Igersheim 2005 and the references therein. 19 The philosopher Debra Satz took up Herzog’s idea: “But nothing in economic analysis generates or supports this particular interpretation of costs or harm; the economic argument for identifying inefficiencies in the case of only certain externalities – pollution but not intolerance of religious diversity – feeds off moral theory done elsewhere. That’s not necessarily a problem, as long as we attend to the moral theory and make it explicit in our understanding of inefficiency” (2010, 32).

15 that he shall not do with his life for his own benefit what he chooses to do with it” (OL IV, 4), Mill mobilizes several types of argument.20 First, the diminution of B’s utility is weaker than the rise of A’s utility: “He is the person most interested in his own well-being: the interest which any other person, except in cases of strong personal attachment, can have in it, is trifling, compared with that which he himself has; the interest which society has in him individually (except as to his conduct to others) is fractional, and altogether indirect” (OL IV, 4). Second, society cannot have sufficient information to decide what is good for A: with respect to his own feelings and circumstances, the most ordinary man or woman has means of knowledge immeasurably surpassing those that can be possessed by any one else. The interference of society to overrule his judgment and purposes in what only regards himself, must be grounded on general presumptions; which may be altogether wrong, and even if right, are as likely as not to be misapplied to individual cases, by persons no better acquainted with the circumstances of such cases than those are who look at them merely from without (OL IV, 4). Third, if we constrain private behaviour in private sphere, this would globally diminish the utility of all in terms of harm to their liberty: “All errors which he is likely to commit against advice and warning, are far outweighed by the evil of allowing others to constrain him to what they deem his good” (OL IV, 4). Mill rejects paternalism in the name of liberty, but liberty in that it permits the person’s accomplishment: “The term duty to oneself, when it means anything more than prudence, means self-respect or self-development; and for none of these is any one accountable to his fellow creatures, because for none of them is it for the good of mankind that he be held accountable to them” (OL IV, 6). The person’s liberty therefore encloses the liberty of harming herself, and this becomes condemnable only when there is harm done to another’s rights: “They may be proofs of any amount of folly, or want of personal dignity and self-respect; but they are only a subject of moral reprobation when they involve a breach of duty to others, for whose sake the individual is bound to have care for himself” (OL IV, 6). Mill recognizes that a person’s private conduct necessarily causes some effects on close persons and on society: “I fully admit that the mischief which a person does to himself may seriously affect, both through their sympathies and their interests, those nearly connected with him, and in a minor degree, society at large” (OL IV, 10). But these negative effects have to be true harms, breaches of duty, or wrongs to rights, and not only harms to minor interests.21 And these breaches of duty may be individual or social:

20 In chapter III, the arguments against paternalism are based on the same arguments that defend the liberty of expression: a variety of modes of action is necessary and profitable for all. 21 Mill cites harms done to interests when citing a possible objection to his non-interference principle for private matters: “No person is an entirely isolated being; it is impossible for a person to do anything seriously or permanently hurtful to himself, without mischief reaching at least to his near connexions, and often far beyond them. If he injures his property, he does harm to those who directly or indirectly derived support from it, and usually diminishes, by a greater or less amount, the general resources of the community. If he deteriorates his bodily or mental faculties, he not only brings evil upon all who depended on him for any portion of their happiness, but disqualifies himself for rendering the services which he owes to his fellow-creatures generally; perhaps becomes a burthen on their affection or benevolence; and if such conduct were very frequent, hardly any

16 When, by conduct of this sort, a person is led to violate a distinct and assignable obligation to any other person or persons, the case is taken out of the self-regarding class, and becomes amenable to moral disapprobation in the proper sense of the term. If, for example, a man, through intemperance or extravagance, becomes unable to pay his debts, or, having undertaken the moral responsibility of a family, becomes from the same cause incapable of supporting or educating them, he is deservedly reprobated, and might be justly punished; but it is for the breach of duty to his family or creditors, not for the extravagance… In like manner, when a person disables himself, by conduct purely self-regarding, from the performance of some definite duty incumbent on him to the public, he is guilty of a social offence. No person ought to be punished simply for being drunk; but a soldier or a policeman should be punished for being drunk on duty. Whenever, in short, there is a definite damage, or a definite risk of damage, either to an individual or to the public, the case is taken out of the province of liberty, and placed in that of morality or law. (OL IV, 10; see also V, 6)22 Mill concludes by reminding his third, utilitarian, argument when there is no harm to others: “with regard to the merely contingent, or, as it may be called, constructive injury which a person causes to society, by conduct which neither violates any specific duty to the public, nor occasions perceptible hurt to any assignable individual except himself; the inconvenience is one which society can afford to bear, for the sake of the greater good of human freedom” (OL IV, 11). The mention of an injury to society switches to the next argument, which more precisely opposes moralism. Mill indeed adds a fourth argument, “the strongest of all”, noting that we generally judge that some behaviours hurt the person acting because we judge these behaviours immoral. This leads to legislate in accordance with the legislator’s morality or preferences: “[T]he opinion of a similar majority, imposed as a law on the minority, on questions of self- regarding conduct, is quite as likely to be wrong as right; for in these cases public opinion means, at the best, some people’s opinion of what is good or bad for other people; while very often it does not even mean that; the public, with the most perfect indifference, passing over the pleasure or convenience of those whose conduct they censure, and considering only their own preference” (OL IV, 12). In other words, legal paternalism generally comes down to a legal moralism. The second part of chapter IV hence concerns behaviours that violate the moral judgements of the legislating class (and for this reason are judged self-harming). Against legal moralism, Mill argues that the moral offense of some does not justify interference with the others’ liberty. We have here a clear rejection of moral externalities as legitimating public interference.

offence that is committed would detract more from the general sum of good” (OL IV, 8). However this is a kind of harm that, precisely, does not legitimate interference. 22 This last sentence is quoted by Medema (2009, 35) in support of his interpretation of Mill as arguing that liberty stops “where spillovers do exist”. Nevertheless, it is to be noted that we have here just another formulation of the harm principle, with no implicit reference to something like externalities. The sentence applies to specific harms to others (direct breaches of duty) provoked by specific behaviours (harmful to self), like a father who does not educate his children because of his intemperance or a drunk policeman.

17 Mill refers to a legal moralism that could be justified by the harm principle when A’s misconduct causes a harm to B’s sense of morality (a moral externality): “There are many who consider as an injury to themselves any conduct which they have a distaste for, and resent it as an outrage to their feelings; as a religious bigot, when charged with disregarding the religious feelings of others, has been known to retort that they disregard his feelings, by persisting in their abominable worship or creed” (OL IV, 12). Against the application of the harm principle in this case, Mill reminds his first argument against paternalism: “there is no parity between the feeling of a person for his own opinion, and the feeling of another who is offended at his holding it; no more than between the desire of a thief to take a purse, and the desire of the right owner to keep it” (OL IV, 12). The link with the arguments against paternalism also appears in the sentence that follows, which parallels liberty of expression and liberty of taste (see fn 20): “And a person’s taste is as much his own peculiar concern as his opinion or his purse” (OL IV, 12). But the main ground of his opposition is that legal moralism comes down to the class who detains the power legislating in accordance with its own preferences, masking its preferences as universal morals: “In its interferences with personal conduct it [the public] is seldom thinking of anything but the enormity of acting or feeling differently from itself; and this standard of judgment, thinly disguised, is held up to mankind as the dictate of religion and philosophy, by nine-tenths of all moralists and speculative writers” (OL IV, 12). The public becomes a “moral police” when it “improperly invests its own preferences with the character of moral laws” (OL IV, 13), as it happened in several instances examined by Mill. He develops several examples, religious for the most part of them, where some persons impose their practises to others, sometimes with arms. Mill brings to light a slippery slope: if we begin to legislate according to the preferences of some, hence we do not stop and it leads to legislate on all the aspects of life (moral, intellectual, even physical) and to religious intolerance until war. Mill’s examples go as follows: First, Muslims resent it as an offense that Christians eat pork. But the ban on eating pork meat must be condemned since “with the personal tastes and self-regarding concerns of individuals the public has no business to interfere” (OL IV, 14). The second example is also associated to religious intolerance: Spanish majority considered as an offense other religions than Catholicism, and made them unlawful in Spain. Mill insists that recognizing a legitimacy to interfere with private liberty is a slippery slope towards this kind of religious intolerance, which cannot but appear unjust to his Protestant readers: What do Protestants think of these perfectly sincere feelings, and of the attempt to enforce them against non-Catholics? Yet, if mankind are justified in interfering with each other’s liberty in things which do not concern the interests of others, on what principle is it possible consistently to exclude these cases? … No stronger case can be shown for prohibiting anything which is regarded as a personal immorality, than is made out for suppressing these practices in the eyes of those who regard them as impieties; and unless we are willing to adopt the logic of persecutors, and to say that we may persecute others because we are right, and that they must not persecute us because they are wrong, we must beware of admitting a principle of which we should resent as a gross injustice the application to ourselves. (OL IV, 15)

18 Mill’s third example with religion is the interdiction of amusements by Puritans (OL IV, 16). The following instance concerns the “popular disapprobation” that accompanies very large income and spending in some parts of the United States, which, even if exaggerated, suggests “a probable result of democratic feeling, combined with the notion that the public has a right to a veto on the manner in which individuals shall spend their incomes” (OL IV, 17). About the ban of alcohol, fifth, Mill insists again on the danger of accepting a principle that would lead to control each and every part of the private sphere, the danger of a “doctrine [that] ascribes to all mankind a vested interest in each other’s moral, intellectual, and even physical perfection, to be defined by each claimant according to his own standard” (OL IV, 19). Mill then comes back to religious duties, which may forbid amusements on Sunday, “a motive of legislation which never can be too earnestly protested against… The notion that it is one man’s duty that another should be religious, was the foundation of all the religious persecutions ever perpetrated, and if admitted, would fully justify them. … It is a determination not to tolerate others in doing what is permitted by their religion, because it is not permitted by the persecutor’s religion” (OL IV, 20). Mill finally opposes the temptation of a crusade against Mormons, already exiled in an American desert, a crusade justified by their polygamy: “I am not aware that any community has a right to force another to be civilized. So long as the sufferers by the bad law do not invoke assistance from other communities, I cannot admit that persons entirely unconnected with them ought to step in and require that a condition of things with which all who are directly interested appear to be satisfied, should be put an end to because it is a scandal to persons some thousands of miles distant, who have no part or concern in it” (OL IV, 21). Mill’s chapter hence argues that acts harming the person acting (or judged as) are not to be interfered with, except in cases of breach of duty to others.23

Coming back to economics As Herzog suggested, the exclusion of moral externalities from the domain of externalities relevant to welfare, and hence from public intervention, parallels the exclusion by Mill of moral externalities from the domain of harms that legitimate public interference. Two elements confirm this link. Mishan was calling on a social consensus, but it is in fact a consensus forged in Mill’s moral philosophy. In his ironic criticism of Coase’s negotiation solution (1960), he opposes Coase’s reciprocity thesis by explicitly invoking Mill’s harm principle. “According to the libertarian philosophy of John Stuart Mill, the freedom of a man to do as he wishes must not go so far as to infringe the freedom of his fellows. If George, in his addiction to motorized gardening, shatters Jim’s week-ends, the case in equity is not symmetrical. The freedom that

23 Another exception is mentioned later: “Again, there are many acts which, being directly injurious only to the agents themselves, ought not to be legally interdicted, but which, if done publicly, are a violation of good manners, and coming thus within the category of offences against others, may rightfully be prohibited. Of this kind are offences against decency; on which it is unnecessary to dwell, the rather as they are only connected indirectly with our subject, the objection to publicity being equally strong in the case of many actions not in themselves condemnable, nor supposed to be so” (OL V, 7). This quote justifies Feinberg’s assertion that liberalism can accept an offense principle, which he details in his volume 2 (Feinberg 1985). Feinberg’ offense principle is amenable to the criticism that determining what causes an offense when done in public comes down to a form of legal moralism (Trebilcock 1993).

19 George would claim certainly harms Jim. In contrast, Jim’s claim for quiet and clear air does not, of itself, harm George” (Mishan 1971a, 118-119). And it is in the footnote to this sentence that he excludes moral externalities according to an “ethical consensus”: “This distinction between the ‘active’ and ‘passive’ agent in a conflict of interest is tenable even though tenuous cases may be resolved only by recourse to an ethical consensus. If the mere sight of a person, or the mere knowledge of his existence or behaviour, offends me, it is not likely that society would support my claim to damages” (Mishan 1971a, 119, fn1). The proximity of this appeal to Mill when excluding moral externalities gives some weight to Herzog’s thesis. This is more explicit in Arrow’s 1969 paper about externalities, which makes the link between the concept and Mill’s harm principle less secret than Herzog implies. Arrow invokes Mill’s definition of the private sphere when mentioning and implicitly rejecting from his analysis these direct effects on utility caused by “crimes without victims” such as homosexuality: “Do we have to extend the concept of externality to all matters that an individual cares about? Or, in the spirit of John Stuart Mill, is there a second-order value judgment which excludes some of these preferences from the formation of social policy as being illegitimate infringements of individual freedom?” (1969, 148). As already explained, the exclusion of moral externalities from the domain of externalities relevant to welfare contradicts the view of welfare as satisfaction of preferences. As Arrow makes it explicit, this exclusion implies a selection of the preferences that enter in the evaluation of social welfare. And this selection rests on philosophical presuppositions, as it happens, Mill’s harm principle. It is not by mere chance that Arrow is the most explicit about the presupposition of his definition of externalities, since he comes from social choice theory. Finally, we are back to Sen’s paradox. The Paretian has to take into account moral externalities as determining welfare, but the Liberal has to exclude them as a motive for legislating. Sen concludes: “If someone takes the Pareto principle seriously, as economists seem to do, then he has to face problems of consistency in cherishing liberal values, even very mild ones” (1970, 157). And he adds in the footnote: “What is at issue here is the acceptability of Pareto optimality as an objective in the context of liberal values, given certain types of externalities” (157, fn 6, emphasis omitted). The history of the concept of externalities suggests that welfare economists were conscious of this problem, before it was formalized by Sen. This would explain why they put aside moral externalities: they kept the Pareto criterion, but excluded some preferences from its application.

4. Concluding remarks This analysis of Mill’s On Liberty and book V of his Principles led to identify several harms that could be interpreted as externalities. They are summed up in Table 1. Once in their contexts, it has to be recognized, though, that they do not correspond to externalities, as economists understand them today. The closest concept examined by Mill is that of public goods, indeed linked to externalities, but a different concept nevertheless.

20 Public Nature of the harm Example Externalities? interference? Voluntary Classic harm to others Legitimate Does not concern and direct Molestation OL I,9 (harm principle) externalities harm Unavoidable private interference (direct I obtain a job someone Does not concern Not legitimate effect of the system) else was desiring externalities OL V,3 Social positive effects of some individual Children education, Legitimate if more Public good Involuntary behaviours scientific research efficient harm to PPE V,11 third party Harm to self with The drunken Legitimate Does not concern breach of duty policeman (harm principle) externalities OL, IV (Perceived as) Not legitimate Moral externalities immoral behaviour, Practice of another (against legal excluded by welfare in private religion moralism) economics OL, IV Figure 1 - Classification of some harms identified in Mill's works (OL and PPE)

With the concept of externalities, which justified public intervention, economists had opened a Pandora box that it was urgent to close. If enlarged to all indirect harms to others, the domain of legitimate intervention could cover all aspects of life. Economists had to restrain the concept to the phenomenon they wanted to analyse, pollution. And they excluded moral externalities from this domain. This exclusion was, however, inconsistent with their view of welfare as satisfaction of preferences (when there is no other mention of the kinds of preferences that are considered). It is henceforth based on another philosophical presupposition, precisely Mill’s liberalism. It has to be concluded that Mill gave economists the philosophical background that justifies them in excluding moral externalities while all their theory should lead to this concept. The economists’ concept of externality is grounded in Mill, but not in the sense Mill would write about externalities, rather because economists exclude from externalities the same moral effects that Mill excludes from the harm principle. It remains that On Liberty, and its exegeses like Feinberg’s, are a wonderful basis to approach the ethics of externalities. For the concept of externality poses the same questions as the harm principle: what constitutes a harm? Where does the limit between the private sphere and the social sphere lie? Do actions that really do not touch anyone exist? Finally, the discussion of moral externalities raises again that the preferences- satisfaction view of welfare and Mill’s liberalism are not compatible unless we a priori define a private sphere on which others’ preferences do not count.24 It is not a problem in itself; it

24 “Opening a space between their preferences and their welfare might seem an unholy recipe for paternalism run rampant or vanguard politics. … But this is to dodge a complex argument that we really have to engage on the merits. We need to be able to show that in many cases people have a right to do what they like, even if they don’t act in their own interests, because the right is itself constitutive of autonomy and autonomy is partly constitutive of human welfare, of a life worth living. If we dodge this admittedly ambitious and difficult argument, we really have no alternative but to concede that R [the woman who does not comb her hair] and the polluter and the

21 just raises once again the impossibility for economists of not engaging the debates on the difference between rights, needs, and wants; on which Mill did not give the least interesting insight. In other words, the economists’ definition of externalities rests on Mill, but not enough.

References Arrow K. J. 1969. “The organization of economic activity: Issues pertinent to the choice of market versus non market allocation”, in Joint Economic Committee (ed.), The Analysis and Evaluation of Public Expenditures: The PPB System, vol. 1. Washington, DC: Government Printing Office, 47-64. Repr. in Collected Papers of Kenneth J. Arrow, vol. 2: General Equilibrium. Oxford: Basil Blackwell, 1983, 133-55. Aslanbeigui, N. and Medema, S.G. 1998. “Beyond the dark clouds: Pigou and Coase on social cost”, History of Political Economy, 30(4), 601-25. Bator F. M. 1957. “The simple analytics of welfare maximization”, American Economic Review, 47(1), 22-59. Bator F. M. 1958. “The anatomy of market failure”, Quarterly Journal of Economics, 72(3), 351-79. Berta N. 2017. “On the definition of externality as a missing market”, European Journal of the History of Economic Thought, 24(2), 287-318. Bertrand E. 2006. “The Coasean analysis of lighthouse financing: myths and realities”, Cambridge Journal of Economics, 30(3), 389-402. Bertrand E. 2010. “The three roles of the ‘Coase theorem’ in Coase’s works”, The European Journal of the History of Economic Thought, 17(4), 975-1000. Bertrand E. 2017. “Lighthouses”, in A. Marciano and G. Ramello (eds), Encyclopedia of Law and Economics, New York: Springer, DOI: 10.1007/978-1-4614-7883-6_629-1 Brink D. 2014. “Mill’s Moral and Political Philosophy”, Stanford Encyclopedia of Philosophy, online, first published 2007; substantive revision 2014, https://plato.stanford.edu/entries/mill-moral-political/ Chauvier S. 2013. Ethique sans visage, Paris: Seuil. Coase R. H. 1960. “The problem of social cost”, Journal of Law and Economics, 3, 1-44. Coase R. H. 1974. “The lighthouse in economics”, Journal of Law and Economics, 17(2), 357-376. Coleman J. S. 1990. Foundations of Social Theory, Cambridge, MA: Harvard University Press. Cornes R. and Sandler T. 1986. The theory of externalities, public goods and club goods, 2nd ed., Cambridge: Cambridge University Press. Feinberg, J. 1984. The Moral Limits of the Criminal Law. Vol. 1: Harm to Others, New York: Oxford University Press. Feinberg, J. 1985. The Moral Limits of the Criminal Law. Vol. 2: Offense to Others, New York: Oxford University Press. Feinberg, J. 1986. The Moral Limits of the Criminal Law. Vol. 3: Harm to Self, New York: Oxford University Press. lesbians and the person daring to order vanilla all, in principle, should be treated precisely the same way. And that can’t be right. Put differently, that the organizing categories of economic analysis provide no way to distinguish these cases gives us no justification for refusing to distinguish them” (Herzog 2000, 914).

22 Feinberg, J. 1988. The Moral Limits of the Criminal Law. Vol. 4: Harmless Wrong-Doing, New York: Oxford University Press. Hammond P. J. 1987a. “Altruism”, in J. M. Eatwell, M. Milgate and P. Newman (eds), The New Palgrave: A Dictionary of Economics, online. Hammond P. J. 1987b. “Envy”, in J. M. Eatwell, M. Milgate and P. Newman (eds), The New Palgrave: A Dictionary of Economics, online. Hammond P. J. 1998, “Rights, free exchange, and widespread externalities”, in J.-F. Laslier, M. Fleurbaey, N. Gravel, and A. Trannoy (eds), Freedom in Economics: New Perspectives in Normative Analysis. London: Routledge. Hausman D. and McPherson M. S. 2009. “Preference satisfaction and welfare economics”, Economics and Philosophy, 25(1), 1-25. Herzog D. 2000. “Externalities and other parasites”, University of Chicago Law Review, 67(3), 895-923. Igersheim H. 2005. “Libéralisme de la liberté ‘versus’ libéralisme du bonheur : Le cas du paradoxe libéral-parétien”, Revue Economique, 57(3), 389-98. Laffont J.-J. 1987. “Externalities”, in J. M. Eatwell, M. Milgate and P. Newman (eds), The New Palgrave: A Dictionary of Economics, vol. 2, London: MacMillan, 263-5. Marshall A. 1890. Principles of Economics. London: Macmillan. Meade J. E. 1952. “External economies and diseconomies in a competitive situation”, Economic Journal, 62(245), 54-67. Medema S. G. 2009. The Hesitant Hand: Taming Self-Interest in the History of Economic Ideas. Princeton: Princeton University Press. Medema S. G. 2017. “’Exceptional and unimportant’? The rise, fall and rebirth of externalities in economic analysis”, mimeo. Mill J. S. 1859. On Liberty. London: J.W. Parker and Son. Mill J.S. 1871 [1909] Principles of political economy with some of their applications to social philosophy, London: Longmans, Green and co., 7th edition. Mishan E. J. 1965. “Reflections on recent developments in the concept of external effects”, The Canadian Journal of Economics and Political Science, 31(1), 3-34. Mishan E. J. 1969. “The relationship between joint products, collective goods, and external effects”, Journal of Political Economy, 77(3), 329-48. Mishan E. J. 1971a. “Pangloss on pollution”, The Swedish Journal of Economics, 73(1), 113- 20. Mishan E. J. 1971b. “The postwar literature on externalities: An interpretative essay”, Journal of Economic Literature, 9(1), 1-28. Mishan E. J. 1974. “The economics of disamenity”, Natural Resources Journal, 14(1), 55-86. Papandreou, A. 1994. Externality and Institutions. New-York: Oxford University Press. Pignol C. 2012. “Rousseau’ notion of envy: A comparison with modern economic theory”, The European Journal of the History of Economic Thought, 19(4), 529-49. Pigou A. C. 1912. Wealth and Welfare. London: Macmillan. Pigou A. C. 1932. The Economics of Welfare (4th ed.). London: Macmillan, 1948. Samuelson P. A. 1958. “Aspects of public expenditure theories”, Review of economics and statistics, 40(4), 332-338. Satz D. 2010. Why Some Things Should Not Be For Sale. The Moral Limits of Markets. Oxford: Oxford University Press.

23 Scitovsky T. 1954. “Two concepts of external economies”, Journal of Political Economy, 62(2), 143-51. Sen A. 1970. “The impossibility of a Paretian liberal”, Journal of Political Economy, 78(1), 152-57. Trebilcock, M. J. 1993. The Limits of Freedom of Contract. Cambridge, MA: Harvard University Press. Viner J. 1931. “Cost Curves and Supply Curves”, Zeitschrift für Nationalökonomie, 3, 23-46.

24