Should market harms be an exception to the Harm Principle?

Abstract Many proponents of the Harm Principle seem to implicitly assume that the principle is compatible with permitting the free exchange of goods and services, even if such exchanges generate so-called market harms. I argue that, as a result, proponents of the Harm Principle face a dilemma: either the Harm Principle’s domain cannot include a large number of non-market harm cases or market harms must be treated on par with non-market harms. I then go on to discuss three alternative arguments defending the status of market harms as exceptions to the Harm Principle and discuss why these arguments also fail.


Introduction
A question at the heart of political philosophy is under which conditions a state is justified in interfering with its citizens' conduct. Liberal theory answers this question with the Harm Principle (HP). HP provides a necessary, albeit not sufficient condition for justified state intervention: If A exposes B to harm or a significant risk of harm, then there is a pro tanto reason for the state to intervene with A's conduct.
The principle has been and remains widely endorsed by liberal authors (Raz 1986;Feinberg 1987). 1 Liberal reasoning in line with HP has also influenced policymakers to consider production-or consumption-related externalities, such as tariffs or climate change, through the lens of harm imposition. Prominent examples include the World Trade Organization's Agreement on Subsidies and Countervailing Measures (World Trade Organization 1994: Annex 1A) and the United Nations Framework Convention on Climate Change (United Nations 1994: Preamble, Recital 8). But often, liberal reasoning adheres to a particularly influential exception of HP: market harms. Market harms are severe welfare losses mediated by price shifts in a competitive market and borne by the losers of market competition. 2 Authors such as Sen (1985), Thomson (1986), Hausman (1992) and Olsaretti (2004) have highlighted some problematic implications of market harms, but to the best of my knowledge, the literature offers no comprehensive discussion of the issue as of yet. This article aims at filling this gap and going beyond the contributions made in the debate thus far.
Market harms are often considered to be an exception to HP: if A suffers a harm due to B's and C's voluntary market transactions, the state is not justified in intervening with B's and C's conduct. The strongest and most common justification for this exception is based on efficiency concerns: promoting the efficient usage of scarce resources in an economy overrides the pro tanto reason in favour of state intervention provided by HP. In other words, market harms are a necessary by-product of efficient markets. They are the 'destructive' aspect of Schumpeter's dictum that capitalist markets operate via a process of 'creative destruction' (Schumpeter 1994(Schumpeter [1942).
In this paper, I argue that it is implausible that market harms are an exception to HP. The result I present is novel insofar as it demonstrates that proponents of HP cannot endorse HP and simultaneously endorse market harms as an exception to HP due to efficiency-related concerns. The preservation of efficiency is likely the strongest argument in favour of conferring an exceptional status to market harms. Without it, proponents of HP have reason to treat market harms on par with non-market harms. Due to constraints of space, I will unfortunately not be able to discuss potential policy implications in detail in this article.
Another shortcoming of this article is that I abstract from the fact that markets are legal constructs. In real markets, both sellers and buyers are provided with legal rights to protect their interests. 3 Consequently, by deciding on the rules that govern their markets, societies also decide on who stands to be harmed through other market participants' transactions. Even though I do not share the economistic view that markets exists outside and apart from legal constraints, it is helpful for my purposes here. In order to illustrate how HP relates to the notion of market harms, I hence abstract from facts of legal construction in what follows.
The article is structured as follows: in the first section, I discuss the link between HP and efficiency. I do so by introducing two examples which illustrate the claim that market harms are treated as an exception to HP in order to promote efficient markets. The second section defines market harms and contrasts them with the textbook case of an externality. I use the two examples to show how market harms, contrary to non-market harms, do not affect efficiency. If the state interferes with market harms in the same manner it interferes with non-market harms, efficiency is threatened. Hence, some authors argue that it is warranted to treat market harms as an exception to HP. In the third section, I introduce a 2 Strictly speaking, these price shifts are typically the result of shifts in the demand or supply of a particular good or service. challenge to the claim that market harms ought to be an exception to HP. The challenge demonstrates that if market harms were an exception to HP, state intervention would only be justified under very specific circumstances even in standard non-market harm cases. As a result, a dilemma emerges: either HP's domain cannot include a large number of non-market harms or market harms must be treated on par with non-market harms under HP. Proponents of HP have good reason to embrace the second horn of the dilemma. In either case, market harms are not an exception to HP.
In the remaining sections, I discuss three popular alternative arguments in favour of treating market harms as an exception to HP. The first argument states that the right to impose market harms is an extension of one's property rights. The second argument states that market harms are implicitly consented to by market participants. A last argument states that market harms are mitigated by welfare state programmes and thus do not raise any questions of justice. While I am unable to produce any arguments to conclusively defy the exceptional status of market harms here, I hope to shift the onus of proof onto those who want to defend market harms as an exception to HP.

The Harm Principle and efficiency
Markets are often considered to have special moral status. This special moral status is, for example, expressed in the claim that even if market transactions generate harms, they are still permissible and do not warrant state interference. However, state interference into harmful conduct outside of the market can be straightforwardly justified by HP.
According to the original formulation of John Stuart Mill, HP states 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' (Mill 1977(Mill [1859: 9). HP thus provides the conditions under which a state can coercively interfere with the conduct of its citizens. Contrary to Mill's suggestion, I propose that there might be other reasons that allow the state to intervene with its citizens' conduct. In order to account for this, the principle can be restated as follows: The Harm Principle (HP): The state is pro tanto justified in coercively interfering with a citizen's conduct if her conduct is harmful or likely to be harmful to other citizens.
A number of different versions of HP have been proposed in the literature. 4 The one I am promoting here is most similar to Feinberg's version of HP, according to which 'the need to prevent harm (private or public) to parties other than the actor is always an appropriate reason for legal coercion' (Feinberg 1987: 11). Conduct that harms others or is likely to harm others provides the state with a strong reason to interfere with such conduct, even if such a reason might be outweighed by other considerations. HP hence justifies state intervention in cases such as the following: Pollution: Pharmaceutical company 'PollMed Inc.' produces several kinds of drugs in their factory complex, which is located close to a river that is not owned by the company. PollMed Inc. manages to keep production costs low by simply dumping the toxic waste generated in the production process into the river. Due to the pollution of the river, the residents of the downstream village, who drink water from the river, suffer considerable liver damage.
Pollution is a classic example of a non-market harm case. Most of us would agree that state interference is warranted in this case. 5 The villagers are not explicitly forbidden from using the river, even if they do not have any explicit property rights over it. PollMed Inc. is either harming the villagers by restricting their use of the river or by poisoning them, should they drink the water from the river. 6 However, HP purportedly does not apply as straightforwardly when market transactions generate harmful consequences, i.e. market harms. Consider this case: Ann's Apples: Ann sells apples at the weekly market. It is her only source of income. One day, CheapPears Inc. opens up a stand right next to her, offering high-quality pears at a much lower price than Ann's apples. Consumers now exclusively buy pears from CheapPears Inc., while Ann is forced to shut down her apple stand and has no means to maintain her income.
Contrary to non-market harm cases like Pollution, our intuitions are less clear in a market harm case like Ann's Apples. It is clear that CheapPears Inc. generates a benefit for those customers who prefer pears to apples at the reduced price. But CheapPears Inc. is also imposing a harm onto Ann merely by competing with her. Harms which involve coercion, fraud or deception in the marketplace straightforwardly fall under the purview of HP. 7 However, this is not the case here, since Ann is being harmed merely due to price fluctuations emerging from market transactions. The harmful consequences resulting from these price fluctuations are the kind of market harms that I am concerned with in this article.
Authors across a broad political spectrum seem to support the thesis that market harms are morally unproblematic. More specifically, many support the claim that once so-called externalities have been eradicated, there is no strong reason for the state to interfere with market transactions even if the resulting price fluctuations generate outcomes that are harmful to some. 8 In this vein, Gauthier speaks of 5 For a dissenting view, see Mildenberger (2018). I am unable to address Mildenberger's arguments here, because they crucially rely on a distinction between causal and enabling conditions, which is beyond the scope of this paper. For a general endorsement of HP as a guide for the regulation of externalities, see Claassen (2016). 6 Feinberg, for example, argues in favour of employing HP in pollution cases when he states that 'HP lends legitimacy to legislative efforts to solve the multidimensional problems of air and water pollution', even if 'in its bare formulation without supplement, it offers no guide to policy' (Feinberg 1987: 232). 7 Mill calls this the Doctrine of Free Trade, which is essentially establishing market harms as exceptions to HP: according of the Doctrine of Free Trade, HP can only be invoked if competitors resort to 'fraud or treachery, and force' (Mill 1977(Mill [1859: 293). 8 It is worth noticing that it is very common for both philosophers and economists not to scrutinize between different kinds of externalities. Typically, talk of externalities refers to non-pecuniary markets as 'moral free zones' (Gauthier 1986), Dworkin argues that markets ought to be at 'the center of any attractive theoretical development of equality of resources' (Dworkin 1981: 284) and Heath defends the position that if 'all companies fully internalized all costs : : : it would be impossible to make the case for any further "social responsibility"' imposed onto market participants (Heath 2006: 551). In On Liberty, Mill himself also seems to suggest that market harms constitute an exception to HP: 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. (Mill 1977(Mill [1859: 105) This leads us to an important question: why would we advocate state interference in Pollution, but not in Ann's Apples, even though persons are being harmed by others in both cases? Mill argues that the state ought not interfere with the success of a market participant to the benefit of her losing competitor, because doing so would undermine what is 'better for the general interest of mankind'. 9 The most uncontroversial and common way for contemporary liberals to cash Mill's idea out is that market economies, unless interfered with, are more likely to result in Pareto-efficient allocations of goods and services than non-market economies (Waldron 1987: 148;Freeman 2011: 36). An allocation of goods and services is Pareto-efficient if there is no alternative allocation available under which an individual could be made better off without making any other individual worse off. 10 This complements Mill's view: reasons for state interference in accordance with HP are outweighed by efficiency concerns in the case of market harms. In other words: when it comes to market harms, efficiency concerns outweigh the pro tanto reason of harm prevention provided by HP.
A principle which captures this special status of market transactions can be stated as follows: Efficiency Exception Principle (EEP): If a voluntary market transaction between two or more individuals is a member in a set of transactions that generates an externalities. I will discuss the link between the relevant kinds of externalities and market harms in detail in the next section. 9 Interestingly, Mill, according to his own testimony an avid supporter of laisser-faire, promotes state interference in markets far beyond what is justifiable via HP (Mill 1965(Mill [1848: Bk. 5, Ch. 3). 10 In what follows, I employ the terms 'efficient' and 'efficiency' to refer to Pareto-efficiency, unless otherwise specified. I refrain from a discussion of Kaldor-Hicks-efficiency in what follows, since Kaldor-Hicks-efficiency will typically not be compatible with HP in a straightforward manner. For a detailed discussion of these concepts in economics, see Mas-Colell et al. (1995: 311-315). For a philosophically more intricate discussion, see Buchanan (1985: 4-13). efficient allocation of goods and services, the state is unjustified in interfering with the transaction in a manner that upsets the efficiency of the allocation.
EEP prohibits the state from interfering with voluntary market transactions only if these transactions would have otherwise promoted efficient allocations of goods and services. EEP is a fairly weak principle: A proponent of EEP can endorse state interventions such as minimum wages, subsidies, welfare programmes etc., insofar as they do not interfere with individual market transactions that promote efficiency. 11 Pro tanto reasons for state interference provided by HP will only be outweighed according to EEP if such interference would hinder a voluntary market transaction that promotes efficiency. This includes cases in which market harms emerge. Hence, EEP implicitly provides us with a concrete reason for why Ann's Apples and Pollution do not both justify state intervention, even if both cases involve harm: State intervention in Ann's Apples will lead to an inefficient outcome, i.e. a market failure, while state intervention (e.g. via optimal taxation) in Pollution will not upset efficiency. According to EEP, distorting an efficient allocation resulting from voluntary exchanges is unjustified. In order to understand why precisely EEP justifies state intervention only in Ann's Apples, but not in Pollution, it is necessary to explain how both examples are discussed in economics.

Market harms and non-pecuniary externalities
Both Ann's Apples and Pollution are standard examples of what economists refer to as negative externalities. 12 Externalities are unintended spillover effects of one agent's actions on another agent's level of welfare (Hausman 1992: 96). Formally, the presence of an externality can be described as follows: Where w i denotes the welfare level of individual i and activities (x 1 ; x 2 ; x 3 . . .) are entirely under i's control, whereas activity y j , with y j ≠ 0, is under j's control, but has an impact on i's level of welfare. 13 Economists typically contrast non-market harms (such as in Pollution), i.e. 'non-pecuniary externalities', with market harms (such as in Ann's Apples). 14 11 For example, this will be the case if such state interventions promote, rather hinder an efficient allocation of goods and services. 12 'Negative', since both externalities create welfare losses, rather than gains. As such, these externalities are prima facie relevant from a perspective of justice. 13 The standard formalization of externalities as presented here is taken from Buchanan and Stubblebine (1962). Notice that the formulation does in no way indicate that the impact of y j on w i is of unintentional nature. For a discussion of the issue, see Baumol andOates (1988 [1975]: 17). 14 Economists typically refer to market harms as 'pecuniary externalities', hence the distinction to nonpecuniary externalities. Pecuniary externalities are called externalities due to the fact they are external to the transaction between two parties, not because they are external to the market in its entirety. Cf. Mas-Collel et al. (1995: 352).
Non-pecuniary externalities are externalities that are mediated 'outside of the market'. In Pollution, PollMed Inc. is creating a non-pecuniary externality for the residents of the village. The polluting activity of PollMed Inc. is not a market transaction, but rather an action performed 'outside of the market'. PollMed Inc. has internal costs, such as acquiring raw materials to produce their drugs, paying wages to their workers etc. The costs that PollMed Inc. does not incur are costs for waste processing that the firm shifts onto the villagers, thereby using the river effectively as a costless resource. This reduction in costs for waste processing, in turn, enables the firm to generate a higher output of drugs at no cost to themselves. In other words: PollMed Inc. does not internalize the social cost of their production.
Market harms are spillover effects that are mitigated via the price system, i.e. 'inside of the market'. A market harm is a severe loss of welfare imposed onto a transacting or third party due to a price shift. Arguably, not all externalities resulting from price shifts constitute market harms. Often, price shifts impact third parties adversely, but they do not generate harms that would justify state interference according to HP and can therefore not be classified as harms in the sense that is relevant to our purposes here. But Ann's Apples exemplifies an externality that amounts to a market harm. A shift in consumption preferences decreases the market price for apples. The exchanges between CheapPears Inc. and their customers affect Ann, who is external to their exchanges. Because customers prefer pears at the reduced price to apples, Ann suffers a severe welfare loss, since the decrease in demand for apples constitutes a loss of income for her.
The crux is that the presence of a market harm does not affect efficiency in a perfectly competitive market, while the presence of non-pecuniary externalities does. In a perfectly competitive market, every market harm is exactly offset by a countervailing benefit. However much profit sellers of a good lose when the price of the good declines is weakly less than what the buyers of the good gain due to the price reduction. 15 Since perfectly competitive markets generate efficient allocations, efficiency remains unaffected by market harms. Changes in price levels only have distributional effects. 16 Contrary to market harms, nonpecuniary externalities constitute inefficiencies which divert resources away from their socially optimal, i.e. efficient, use. In Pollution, PollMed Inc. did not internalize the entirety of their production costs, but instead shifted some of their costs over onto the villagers.
According to the classical, so-called Pigouvian analysis, if private costs and socially optimal costs are not aligned, this typically leads to over-or underproduction. More specifically, in the case of Pollution, PollMed Inc.'s drugs will generally be supplied at a price that is lower than and a quantity that is 15 It should be noted that a price shift in a complete market, which imposes a market harm on some market participants, does not mark a shift to a pareto-superior allocation, but merely a shift to another pareto-efficient allocation. The respective allocations are pareto-incomparable, because some are made worse off while others are being made better off. 16 Scitovsky (1954) is often credited with providing proof of this claim. See also Meade (1952), Loong and Zeckhäuser (1982: 171), Greenwald and Stiglitz (1986: 229) and Mas-Colell et al. (1995: 352).
higher than the one that would have resulted in a Pareto-efficient allocation. Pigou's discussion of inefficiencies generated by non-pecuniary externalities concluded that the only way to restore efficiency in such cases is by forcing the externalityproducing firm (or consumer) to internalize their costs via state interventions (Pigou 1960(Pigou [1920). 17 In line with the Pigouvian conclusion, EEP tells us that because PollMed Inc.'s waste dumping generates an inefficiency, state intervention (such as optimal taxation of PollMed Inc.'s products) can be justified. In Ann's Apples, promoters of EEP will argue that taxing CheapPears Inc. is unjustified, since it distorts the price mechanism and thus generates an inefficiency. In general, because market harms do not distort efficiency, but their prevention would, EEP thus establishes that the state is not justified to intervene with transactions generating market harms in otherwise complete markets.
The important take-away message here is that if the state interfered to prevent (or sanction) market harms in the way that it interferes to prevent (or sanction) nonmarket harms, market efficiency would be undermined. This grounds the exceptional status of market harms. In the next section, I argue that those who endorse EEP face a dilemma which forces them to choose between either promoting HP or promoting market efficiency.

Coase's challenge
Until Coase's famous essay 'The Problem of Social Cost' was published, economists sided with Pigou (and thereby HP) in cases involving non-pecuniary externalities such as Pollution (Coase 1960). Contrary to Pigou's analysis, the so-called 'Coase Theorem' stated that in the absence of transaction costs, which disincentivize market participants from engaging in voluntary exchanges with each other, externalities will be internalized efficiently via the market itself. 18 Thus, some economists classify the presence of externalities as an 'absence of markets' via which externalities can be bargained away (Arrow 1969). In what follows, I show how this crucial insight permits transforming many non-market harm cases into market harm cases. This, in turn, will have an impact on the range of cases in which state intervention is justified according to EEP.
Coase demonstrated that, given that transaction costs are low and property rights are well-defined, non-pecuniary externalities can be transformed into market harms. 19 In other words: we can alter Pollution (under the conditions of the 17 The accuracy of Pigou's analysis remains disputed. As Dahlman (1979) argued, asserting the presence of an externality requires the presumption that an attainable, pareto-superior allocation is known to be available. Whether an externality is present or not is a judgement that has to be made in relation to knowledge about alternative allocations, rather than a mere matter of fact, as Pigou's analysis suggests. 18 It is well worth mentioning that Coase himself lamented the reception of the Coase Theorem. Coase himself, as is doubtlessly correct, assumed that positive transaction costs were always present in real world transactions. In summarizing the results and reception of his 1960 article, Coase states (Coase 1990: 15): 'What my argument does suggest is the need to introduce positive transaction costs explicitly into economic analysis so that we can study the world that exists. This has not been the effect of my article.' 19 A short note on the impact of transaction costs is in order here. Transaction costs are almost always present in real world transactions. However, the absence of transaction costs is a crucial assumption in the Coase theorem) such that the harm the villagers incur merely constitutes a market harm. Consider the following case: Pollution*: PollMed Inc. is dumping waste into the nearby river and affecting the villagers like in Pollution. Assume now that PollMed Inc. has been assigned the right to dump waste into the river by a court. The court was highly partial in assigning the property right to PollMed Inc., since the presiding judge is the cousin of PollMEd Inc.'s CEO. Assume that PollMed Inc. values their property right to dump waste in the river at $500. The villagers value an unpolluted river at $600. In this case, the villagers and PollMed Inc. can strike a Pareto-efficient bargain: The villagers simply pay PollMed Inc. $501 in order to stop their pollution of the river. This way, PollMed Inc. gains $1 and the villagers gain a benefit equivalent to $99.
Our initial judgement in Pollution was that the state would, all things equal, be justified in intervening with PollMed Inc.'s pollution activity. The outcome that would have ensued if HP granted the villagers the right not to suffer from pollution would have been that PollMed Inc., in one way or another, would have been either forced or strongly incentivized to stop polluting at no cost to the villagers.
In Pollution*, two new elements have been introduced to the original example: First, an unjustly allocated property right to pollute and second, the possibility for a pareto-efficient bargain. 20 In this section, I will focus on the latter and postpone the discussion of property rights to the next section. It suffices to say here that unjustly allocated property rights should clearly not be accepted as a reason not to intervene with PollMed Inc.'s pollution.
According to EEP, once the villagers and PollMed Inc. can bargain the externality away, state intervention is unjustifiable. This is so because the bargain struck between the villagers and PollMed Inc. is purely voluntary (in the sense that fraud, deception and coercion are superficially absent) and the resulting allocation is efficient, since the externality has been eradicated: A market for pollution has been created. Hence, there is a market price for eradicating the externality. This price is $501, exactly the amount that the villagers are paying for a non-polluted river. The non-pecuniary externality in Pollution has been dealt with by transforming it into a market harm in Pollution*. This, however, entails that the villagers also suffer a market harm by having their purchasing power reduced by $501. so-called First Fundamental Welfare Theorem, which establishes that competitive markets generate Paretoefficient allocations. The presence of transaction costs therefore diminishes credibility in real markets' capacity to generate Pareto-efficient outcomes. Thus, the assumption of zero transaction costs is a concession to my opponent, the market harm exceptionalist. I do not doubt that transaction costs themselves might affect the implementation of HP, but this discussion is beyond the scope of my article. 20 Interestingly, efficiency would have been preserved even if we assigned the property rights to the villagers, since PollMed Inc. would have been forced by the state to stop polluting the river. This is known as the Invariance Hypothesis, which is one implication of the Coase Theorem: the efficiency of the bargaining solution is invariant with respect to the initial assignment of property rights. See Frech (1979: 254) and Zerbe (1980: 84).
The mere fact that the background conditions (assigning property rights in some arbitrary manner and introducing low transaction costs) to deal with the externality have been set in place should have no bearing on the employment of HP. None of the facts that justified state intervention via HP in Pollution have been changed in Pollution*. It is still the case that the villagers' health is impacted by the waste dumping of PollMed Inc. in Pollution* as it was in the original case.
The significant problem that emerges from our discussion of Pollution* is that if the possibility to strike an efficient bargain outweighs considerations of harm, as suggested by EEP, state intervention will be unjustifiable in a large number of cases that are structurally identical to Pollution*.
Some readers might not be convinced that there is anything objectionable about the fact that the villagers ought to pay PollMed Inc. to stop the pollution of the river. Perhaps the following case, which exhibits the same structure as Pollution and Pollution*, is more compelling to them: Breaking Legs: Bob is out on a late night walk through his neighbourhood when he meets Alice. Alice's hobby is breaking legs. She values the opportunity to break Bob's legs at $50. Bob values his unbroken legs at $60. They could thus strike a bargain: Bob pays Alice $51 and gets to return home with his legs intact. Bob accepts the deal.
Should the state not intervene in order to protect Bob? Proponents of EEP have to conclude that the state is not justified in intervening with Alice's behaviour, irrespective of how credible her threat of breaking Bob's legs is. The only fact that matters is that Bob and Alice can strike a Pareto-efficient bargain. While it is unclear how EEP could justify state intervention in Breaking Legs, HP readily delivers an answer: If Alice credibly threatens to break Bob's legs, the state is justified in interfering with her conduct, irrespective of whether an efficient bargaining solution is available. 21 The challenge ultimately resulting from the Coase Theorem is this: given the right background conditions, every non-pecuniary externality can be transformed into a pecuniary externality, i.e. every non-market harm can be transformed into a market harm. According to EEP, the state is not justified in interfering with any market harms, since such interference would generate inefficiencies. However, blocking state intervention into all market harm cases in this way will prevent state intervention even in cases like Pollution* or Breaking Legs.
Proponents of EEP hence face a dilemma. Call its horns 'opening the flood gates' and 'no exception', respectively. If they stick with EEP and argue that market harm cases do not justify state interference, then they are opening the floodgates for noninterference even in cases such as Pollution* and Breaking Legs. Whenever efficient bargaining is possible, the state is unjustified in interfering in order to prevent harm to the relevant party, i.e. the villagers or Bob. This entails effectively abandoning HP 21 Coase himself promotes the principle according to which liability for damages due to an externality should be assigned to the party that can avoid the damages at the lowest cost. Calabresi (1970) hence refers to Coase's solution as the 'cheapest cost avoider' principle. In some cases, this will be the polluter, in others, it will be the victim. Coase's own suggestion is therefore not compatible with HP.
in favour of an efficiency-based principle. The state will then only intervene in nonmarket harm cases if such interference does not threaten the possibility for efficient bargains. Hence, market harms are not an exception to HP because HP is effectively abandoned.
Alternatively, proponents of EEP can abandon EEP itself and argue that there is no exception for market harms. State interference is then pro tanto justified in virtue of a harm imposition and this pro tanto justification cannot be outweighed by efficiency concerns alone. This would entail that the state is justified in intervening in Pollution* and Breaking Legs. But it also entails that the state is justified in intervening in Ann's Apples, because Ann is harmed by CheapPears Inc. and its customers, even if the harm is mediated via the price system. The upshot is that market harms are then not an exception to HP either, because they are treated on par with non-market harms.
At the beginning of this article I stated that proponents of HP have good reason to treat market harms on par with non-market harms. This is so because embracing the horn I called 'opening the flood gates' entails abandoning HP. But giving up on market harms as an exception to HP is an acceptable price to pay, since the possibility for efficient bargaining is not a fringe phenomenon, as is indicated by cases like Breaking Legs. It is worth pointing out that this result is not as shocking as it might appear at first glance for two reasons: first, HP only provides justification for state interference if natural persons stand at a risk of harm. Unless natural persons are significantly threatened by harm, the state is not justified in interfering with transactions between firms. Second, not every adverse price shift will constitute a market harm proper. While foregoing purchasing power is a nuisance to every market participant, not every instance of losing purchasing power counts as a harm that is sufficient to stimulate HP. Cases like Ann's Apples, however, demonstrate that the loss of purchasing power (e.g. by losing one's future income stream) can be equivalent to a significant harm. 22 In conclusion, whether a harm is mediated via the price system or not should not make a difference to whether HP justifies state intervention or not. What ultimately matters is only whether an agent causes another harm, which is the case in Pollution, Pollution*, Breaking Legs and also Ann's Apples. Those who think that efficiency concerns can trump the pro tanto justification provided by HP must give up on HP altogether.
However, some might object that state interference in Ann's Apples is unjustified for reasons that are unrelated to efficiency concerns. The efficiency-based argument discussed so far constitutes only one possible (albeit the most plausible) argument in favour of treating market harms as an exception to HP. In what follows, I consider three commonly invoked, alternative arguments against state interference in market harm cases. 22 This points to a central problem in Coase's own analysis of the reciprocal nature of externalities: 'To avoid the harm to B would inflict harm on A' (Coase 1960: 2). Coase has a much more permissive understanding of what harm amounts to than proponents of HP. From their perspective, forcing CheapPears Inc. to produce at lower quantities or tax their products need not harm either of them, ceteris paribus. But if Ann starved due to a lack of income, this would clearly constitute a harm.

Property rights
Even if efficiency concerns are insufficient to establish that market harms are an exception to HP, alternative routes of justifying the exceptional status of market harms are conceivable. 23 Consider the following argument by Thomson: There is a kind of harm that, when inflicted by one person on another, does not infringe a right. I have in mind what might be called 'market harms'. Suppose, for example, that you make lace and now have a cupboardful ready to bring to the market tomorrow. This afternoon, I invent a way of making lace cheaply, by machine, which only an expert can tell from handmade lace like yours. I do this without damaging your lace or dirtying it, without in any way touching it. It seems plain enough that I infringe no right of yours in doing so. (Thomson 1986: 160) Thomson acknowledges that I do cause you harm, but nonetheless, I do not infringe any right of yours in doing so, since I do not tamper with your produce. 24 Thus, the harm that I inflict upon you by competing with you in the market place is insufficient to justify a state intervention on your behalf. In other words: every market participant has a right to inflict market harms onto other market participants. This right is entailed by their property rights over their goods.
In Ann's Apples, both Ann and CheapPears Inc. have property rights over their apples or, respectively, pears. Assume that neither stole or otherwise acquired their goods in a manner that violates justice. Hence, both prima facie justly acquired their goods. (Nozick 1974: 178-179) Furthermore, let us assume that their property rights over their apples and pears also include, in Honoré's terminology, the 'incidence of transmissibility', i.e. a right to sell their goods (Honoré 1961). 25 What the statement by Thomson then seems to imply is that state interference with the exchanges between CheapPears Inc. and their customers to Ann's favour would violate CheapPears Inc.'s (and their customers) property rights, since both parties have a right to freely exchange their goods, insofar as they have acquired them justly. Prima facie, the condition of just acquisition is fulfilled in Ann's Apples.
Contrary to this, neither PollMed Inc. nor Alice have acquired the goods that they use or attempt to exchange in Pollution, Pollution* or Breaking Legs. In Pollution, PollMed Inc. has no property right over the river and ergo neither the right to pollute it nor to exchange the right to pollution of the river. In Pollution*, PollMed Inc. has a legal property right over the river, but the right has by assumption not been justly acquired and thus does not constitute a moral right. In Breaking Legs, Alice has no property rights over Bob's body. Thus, she 23 Interestingly, Demsetz (1967) provides a link between efficiency and property rights by arguing that private property rights' primary function is the mitigation of externalities. But Demsetz's theory of property rights is not concerned with just distributions of property rights, and therefore it has no bearing on the argument presented here. 24 Notice here that the property right in question is not a legal, but a moral one. 25 Whether a property right over a specific good includes the right to sell the good is a much debated question. For discussion, see Cheney (1977) and Fried (1995). cannot demand money for refraining from violating Bob's right to bodily integrity. All of the non-market harm cases exhibit an unjust assignment of exchangeable property rights. Since the property rights in these cases have not been justly acquired, the state is justified in interfering in Pollution, Pollution* and Breaking Legs, but not in Ann's Apples. We could thus stipulate the following principle: Property Rights Exception Principle (PEP): If an initial allocation of property rights is just, any coercive state interference changing the allocation resulting from the voluntary transaction of these property rights is unjustifiable.
Given that the initial assignment of property rights is just, it is unjustifiable to interfere with market harms resulting from exchanges of these property rights. A principle like PEP seems to provide a way to justify state intervention only in non-market harm cases, while prohibiting state interference in market harm cases. CheapPears Inc. and their customers prima facie justly acquired the property rights over their goods. If CheapPears Inc. and their customers are entitled to their property and exchange it, they are also entitled to the result of that exchange. PEP entails that state intervention on Ann's behalf is unjust, even if Ann is exposed to a market harm as a result.
However, PEP leaves open the question which property rights assignments are just and which are not. 26 Those who wish to argue that there is a relevant difference between cases of market harms like Ann's Apples and cases of non-market harms like Pollution, Pollution* and Breaking Legs must provide a criterion according to which the initial assignment of property rights was only unjust in Breaking Legs and Pollution. 27 But any criterion of just initial property rights assignments (such as the Lockean proviso) allowing for the exchange of property ought to be sensitive to severe market harms. Even if CheapPears Inc. and their customers have justly acquired their property and retain the right to trade it freely, market harms can leave third parties significantly worse off. Consider for example a variation of Ann's Apples, Ann's Apples*, in which Ann starves as a consequence of her having to shut down her apple stand. Our assessment that a particular assignment of property rights is unjust depends at least partially on such harms. Treating harm-related considerations as unproblematic exclusively in market harm cases requires arguments beyond the mere stipulation that the initial assignment of property rights which allowed for harms to come about was just. As Hausmann argues: Negative pecuniary externalities (i.e. market harms) can force people on pain of starvation to leave their homes, occupations, families, countries, and cultures; indeed, they can force people to starvation itself. Can one plausibly accept the view that human actions (i.e. market exchanges) with such 26 In reality, it is doubtful that legal property rights (including bankruptcy law, liability rules, collateral law etc.) normally clear the high bar of justification required from assignments of moral property rights. overwhelming impact on the lives of other people raise no questions of justice? (Hausman 1992: 104, parentheses added) Severe market harms give us reason to doubt whether an initial assignment of property rights was just to begin with. Without offering further arguments (independent of efficiency considerations) which outweigh harm-related considerations, no property rights assignment that allows for individuals to end up severely harmed should be accepted as just. 28

Consent
Another way to argue that state intervention is not justified in market harm cases is related to the notion of consent: Market participants consent to being exposed to market harms by transacting voluntarily with others. Hayek makes the following statement with regard to suffering market harms as a market participant: Once we have agreed to play the game and profited from its results, it is a moral obligation on us to abide by the results, even if they turn against us. (Hayek 1966: 611) According to Hayek, market harms are harms that market participants have consented to, because they have 'agreed to play the game'. Inflicting harm on a victim that consented to being exposed to the harm does not warrant state interference via HP. This principle is also known as the maxim volenti non fit injuria (Feinberg 1987: 115). In the example cases, the reason why it is justifiable for the state to interfere in Pollution on the villagers' behalf is that the villagers have not consented to being exposed to liver damage. Neither did Bob consent to Alice breaking his legs. The reason why it is not justifiable for the state to intervene in Ann's Apples is that Ann tacitly consented to potentially lose to her competition by participating the market. Prima facie, since Ann's harm resulted from voluntary transactions between consenting market participants, state intervention is not justified. In what follows, I assume in accordance with Hayek that consent to market participation is sufficient to establish consent to suffering market harms. Nonetheless, I argue that the 28 I am indebted to an anonymous reviewer for the following objection: in some cases, economists talk about other social institutions as if they were markets. Consider for example Becker's marriage market model (Becker 1973). In such a market, the imposition of harms is justified due to the fact that we accept these harms in order to protect certain liberties. If both A and B want to marry C and C chooses A, then B might be harmed in a relevant sense. Similar liberties might justify the imposition of market harms: the customers of Cheap Pear Inc. have the liberty to buy the product they prefer and Cheap Inc. has the liberty to sell its pears. My response is as follows: the liberty to marry whomever one wants to and the liberty to sell and buy whatever one wants to are, at least prima facie, not morally on par. We would severely limit C's autonomy if we forced her to marry B. It is not clear to me that we would, in a similar manner, restrict Pear Inc.'s autonomy (whatever that precisely might turn out to be) if we limited the firm's ability to sell and buy whatever it chooses to. Furthermore, some of the liberties to buy and sell are derivative of an underlying assignment of property rights, which I have already discussed at length in this section. relevant kind of consent to market participation cannot be established in Ann's Apples.
Authors such as Olsaretti, Hausmann and Wollner argue that this appeal to consent is far-fetched. Market participants typically have no reasonable alternative to participating in the market, because opting out of the market is unreasonably costly (Hausman 1992;Olsaretti 2004;Wollner 2013). 29 However, while Ann might not be consenting to participate in the market in general, she presumably consents to participating specifically in the apple market. Ann expresses her consent to being impacted by market harms by choosing to be a seller in the apple market, rather than a seller in the market for lemons, for example. Thus, Ann consents to being harmed as a seller in the apple market, since she is not changing professions.
The problem with this argument is not only that the availability of reasonably costly alternatives is an insufficient condition for consent in the market place. 30 Worse yet, if the costs associated with changing professions (or, more generally, markets) are positive, it fails to establish the required asymmetry between Pollution and Ann's Apples. Suppose that Ann can choose between selling apples and selling lemons. Changing from selling apples to selling lemons is not costless for Ann. Even in an entirely unrestricted market, she would have to buy lemons first in order to sell them. This implies that the costs of transitioning to another market are effectively imposed onto Ann.
The crux is that, quite similarly, it is conceivable that the villagers in Pollution could move sufficiently far away to avoid PollMed Inc.'s polluting activity, even if this is costly to them, too. But, as HP dictates, the villagers ought not bear any costs for PollMed Inc.'s pollution at all. Even if the cost of changing location was minimal to the villagers, we would not read their refusal to move as them consenting to the pollution. However, in Ann's case, precisely this is entailed by deriving Ann's consent from her choice not to leave the apple market. If we concluded that Ann is consenting to being harmed qua apple seller because she does not transition to another market, we would have to conclude that the villagers are consenting to living nearby a polluted river, because they are not abandoning their homes. If choices associated with positive costs cannot be understood as expressing consent in the manner presented above, establishing an asymmetry in Ann's Apples and Pollution fails. If the villagers' refusal to move cannot establish their consent to being harmed via pollution, neither can Ann's refusal to change markets establish her consent to being exposed to market harms.
In the quote above, Hayek implicitly points to an alternative source of consent, namely that Ann benefits from participating in the market. However, merely conferring a benefit onto Ann is insufficient to establish her consent. A more defensible position would claim that Ann would consent to enter the market if she could rationally expect to gain a net benefit from participation. This argument is being hinted at, for example, in Friedman's claim that the 29 For a dissenting view on consent in market exchanges, see Posner (1981: Ch. 4). 30 As Peter argues, the availability of reasonably costly alternatives does not establish consent to market participation because the set of available alternatives itself is typically not chosen by the agent herself (Peter 2004).
'possibility of coordination through voluntary cooperation rests on the elementaryyet frequently deniedproposition that both parties to an economic transition benefit from it' (Friedman 1962: 13).
For the sake of argument, let us assume that consent can be established via the net benefit criterion. 31 Clearly, neither Bob nor the villagers gain a net benefit in their respective cases and thus cannot consent to being exposed to harm. The criterion thus gets Pollution and Breaking Legs right.
The problematic cases in which consent is hard to establish in this way are cases such as Ann's Apples*, in which Ann starves. Imagine that before starvation, Ann had to decide whether she would consent to participating in the market. Ann's ex ante knowledge that she could starve due to being impacted by market harms must figure into the cost-benefit analysis required to establish her consent. Outcomes such as starvation, even taking into account the probability with which they will materialize, weigh heavily on the cost side of the calculation. The task of those who wish to argue that consent is nonetheless present in Ann's Apples*, is to explain why such outcomes were ex ante sufficiently unlikely to undermine the prospect of an expected net benefit. As a general rule, the probability of being affected by excessive market harms crucially depends on the wealth available to the individual in question. After all, if Ann had sufficient wealth, she would be able to avoid starvation irrespective of her success as an apple seller. All things equal, the likelihood of Ann gaining a net benefit in the market, and ultimately her consent to participating in the market, thus depends significantly on her private wealth at the time she is supposed to give her consent. Rational individuals will not agree to participate in market procedures that are ex ante likely to harm them (MacLean 1982: 65). If an individual's private wealth is insufficient for her to rationally expect that she will avoid significantly harmful outcomes due to market harms, she cannot consent to her participation in the market. Consequently, some individuals will be able to consent to their participation, while others will not. Given that the latter do not enjoy a sufficient level of wealth to cushion harsh market harms, the claim that it is necessarily rational for every individual to participate in the market is implausible. For those individuals for whom consent cannot be established, exposition to market harms would be morally on par with exposing them to ordinary non-market harms and hence warrants state interference according to HP. Without any redistribution of wealth allowing for the protection of market participants from severe market harms, consent is unlikely to be established.

Welfare state
As has been pointed out in the previous section, consent to market participation crucially depends on how likely it is that severe market harms impact individuals. The previous section established that the redistribution of wealth plays a crucial role in mitigating the impact of market harms. Welfare states 31 Notice that this condition is very weak. Contractualist accounts, for example, state that consent can only be established if none of the affected persons has a reasonable complaint (Scanlon 1998). See Meyer (2018) for a related discussion on the contractualist justification of private property. provide institutional mechanisms for redistribution and provide assistance to those impacted by market harms.
The core of a welfare state are programmes designed to provide 'general social assistance to those in desperate circumstances' (Goodin 1988: 5). For example, if a citizen ends up unemployed in a well-functioning welfare state, the harm constituted by the loss of their future income stream, which might otherwise have caused severe harm, will be cushioned by generous unemployment benefits. Some authors even argue that welfare state programmes are a precondition for the legitimacy of markets (Goodin 1988: Ch. 6). Thus understood, welfare state programms are an ex ante measure for mitigating market harms. Call this the ex ante perspective. Alternatively, welfare state programmes can be understood as a form of compensation for market harms. The latter interprets welfare state programmes as an ex post measure of mitigating market harms. Call this the ex post perspective. Notice that the ex ante perspective is more ambitious than the ex post perspective. It implies that without a functioning welfare state, no market transaction, including mutually beneficial ones, can be justified. While I am sympathetic to this perspective, I do not aim to defend it here and thus discuss only the ex post perspective, as depicted above.
By embracing the ex post perspective, one could argue that appropriate welfare state programmes preempt HP. Consider the following principle: Welfare State Exception Principle (WSEP): If a citizen incurs a market harm and is compensated by the state for it, interfering with the conduct of those who generated the market harm is unjustified.
Market harms only raise issues of justice because they impose severe welfare losses onto some market participants. Welfare state programmes prevent such losses from affecting market participants. Imagine a variation of Ann's Apples*, in which Ann receives generous unemployment benefits that prevent her from starving. WSEP effectively rules out further interference into CheapPears Inc.'s exchanges only under the condition that Ann does not suffer any severe market harms. Prima facie, it follows that market harms in a properly functioning welfare state do not raise any issues of justice. Market harms are then an exceptional kind of harm not subject to HP, but rather subject to welfare state compensation.
This line of reasoning faces a challenge: because WSEP allows welfare state programmes to take precedence over HP, the burden of compensating for market harms is assigned to those who fund welfare state programmes, rather than those who impose market harms onto others.
Welfare state programmes constitute a kind of social insurance. The risk premiums are typically paid for by citizens and firms via taxation. 32 As stated at the beginning of this section, the welfare state's core function is to protect its citizens from adversity. This is achieved via social insurance. The term 'adversity' encompasses various sources of harms, such as harms that are not 32 This picture of how the welfare state is funded is overly simplified, but suffices for our purposes here. For a more accurate and detailed characterization, see Barr (2001). man-made (e.g. harms resulting from natural catastrophes) and harms that are man-made, but self-regarding (e.g. harms inflicted upon oneself due to a mental disorder). Contrary to this, HP is designed merely to prevent or rectify otherregarding, man-made harms. But under special circumstances, the welfare state might also protect its citizens from man-made, other-regarding sources of harm, such as lung cancer contracted from second-hand smoke. Depending on one's conception of the welfare state, HP and the welfare state can share overlapping jurisdictions. WSEP explicitly advises that welfare state programmes ought to take precedence over state interventions via HP in the case of market harms.
Many sources of other-regarding, man-made harm not covered by this social insurance fall under HP's domain. If Alice breaks Bob's legs, HP dictates that she ought to be punished for her conduct. Proponents of HP can justify such punishment either in order to deter her and others from future harmful conduct or because Alice's wrongdoing requires retribution (Edwards 2014: 256). 33 But WSEP does not require Alice to be punished. It just requires that Bob be compensated for the harm he incurred. Such compensation imposes burdens onto taxpaying citizens who pay into the social insurance scheme. But for Alice to be punished, the burden imposed onto her must be disproportionately higher than the burden imposed onto any taxpaying citizen who was not complicit in harming Bob. This is precisely what WSEP denies in market harm cases. If CheapPears Inc. and its customers harm Ann, all taxpaying citizens pay to compensate Ann. No disproportional burden is placed onto CheapPears Inc. and its customers.
Those who wish to promote both HP and WSEP propose that only in market harm cases, HP's requirement to punish those who harm others be alleviated. Thus, Alice would be punished, but CheapPears Inc. would not. But justifying this disparity in treatment requires that the proponent of WSEP explains what the morally relevant difference between market and non-market harms is. Simply arbitrarily stipulating such a difference is not enough. From the earlier sections in this paper, it should be clear that this difference cannot be derived from either efficiency or property rights considerations.
If proponents of WSEP fail to provide a justification for treating market harms and non-market harms differently, appropriate state interventions ought to be implemented. If the burden of rectification imposed onto Alice in Breaking Legs ought to reflect her wrongdoing, then so should the burden of rectification imposed onto CheapPears Inc. and their customers for harming Ann. 34 The state can accommodate this requirement via various policy measures, ranging from progressive taxation reflecting one's causal involvement in (or likelihood of) generating significant market harms to restricting the exercise of the right to exchange one's property freely, e.g. via the introduction of quotas. 33 See also Mill (1977Mill ( [1859: 292). 34 Notice that the burden of rectification imposed on Alice need by no means be equal to the harm incurred by Bob. By the same token, if the state forced CheapPears Inc. and their customers to compensate Ann, it need not be the case that they are required to provide full compensation on their own. The point I am arguing for here is merely that they ought to pay more than taxpayers that did not take part in the transactions.
Without an independent justification of the asymmetrical treatment of market harm and non-market harm cases, proponents of WSEP merely postpone providing an answer to the challenge why market harms ought to be treated differently from non-market harms.

Conclusion
In this article, I argued that market harms should not be considered an exceptional type of harm. Just as other kinds of harm, market harms generate reasons for coercive state interference. Efficiency-based reasons in favour of treating market harms as an exception to HP fail, because non-market harms can be transformed into market harms if transaction costs are low and property rights well-defined. Hence, if market harms ought to be treated as an exception to HP due to efficiency-based reasons, it follows that the same must be true for all non-market harms that can efficiently be bargained away. As I have argued, this is implausible. Proponents of HP would effectively have to abandon HP in order to treat market harms as an exception. Furthermore, I considered three alternative arguments from property rights, consent and welfare state programmes in favour of treating market harms as an exception to HP. I concluded that even these alternative arguments failed. Yet, the discussion here has not exhausted all possible arguments in favour of treating market harms as an exception to HP. Despite the problems I discussed in this article, the argument from efficiency found wide acceptance in liberal political philosophy. But if efficiency considerations are insufficient to justify an asymmetry in the treatment of market and non-market harm impositions, we have good reason to believe that other forthcoming arguments might be even less plausible. Ultimately, the onus of proof is on those who want to claim that market harms ought to be considered exceptional in a morally relevant manner.