There is nobody in this country who got rich on his own – nobody. You built a factory out there? Good for you. But I want to be clear. You moved your goods to market on the roads the rest of us paid for. You hired workers the rest of us paid to educate. You were safe in your factory because of police-forces and fire-forces that the rest of us paid for … Now look, you built a factory and it turned into something terrific … God bless – keep a big hunk of it. But part of the underlying social contract is, you take a hunk of that and pay forward for the next kid who comes along.
Elizabeth WarrenFootnote 1The benefit principle … is inconsistent with every significant theory of social and economic justice.
Murphy and Nagel (Reference Murphy and Nagel2002: 19)1. Introduction
The benefit principle occupies a peculiar place in tax theory. On the one hand, it enjoys widespread acceptance as an account of the justice of taxation for the funding of public (i.e. non-rival and non-excludable) goods such as security, a clean environment and basic infrastructure (Vogel Reference Vogel1988a: 24–33; Cooper Reference Cooper1994: 430–458). The principle is also regularly held up as one of the foundational principles of the modern international tax regime (Avi-Yonah Reference Avi-Yonah2000: 1573; Kemmeren Reference Kemmeren2006: 430; Rigoni Reference Rigoni2017: 205). Moreover, as Elizabeth Warren’s allusion to it illustrates, its motivating idea that you ought to pay something for what you get out of social cooperationFootnote 2 not only packs considerable rhetorical force (Weinzierl Reference Weinzierl2017: 54; but also see Duncan et al. (Reference Duncan2014: 120), who find that support for mileage fees is lower than that for fuel taxes), but also identifies an area of overlap between political polar opposites, from Warren’s progressive politics to libertarianism (Fried Reference Fried1999: 158; Duff Reference Duff2005: 31).Footnote 3 On the other hand, there is wide theoretical consensus to the effect that these virtues cannot be scaled up to a plausible general principle of tax justice. In his celebrated treatise on personal income taxation, Henry Simons heaps scorn on the benefit principle as a ‘slogan [that], of course, has little more than emotive content’ and ‘leads nowhere at all’ (Simons Reference Simons1938: 4; Shaviro Reference Shaviro2013: 20), and as ‘a significant element in a reactionary social philosophy, constructed from the gratuitous implications of laissez-faire economics’ (Simons Reference Simons1938: 34; Stewart Reference Stewart2022: 49). Similarly, Murphy and Nagel do not mince their words when they conclude that the benefit principle is ‘inconsistent with every significant theory of social and economic justice’ (Murphy and Nagel 2002: 16–17). It is no wonder that, reviewing more recent literature on optimal taxation, Richard Scherf and Matthew Weinzierl note that ‘benefit-based taxation has been sidelined’ in modern theoretical debates (Scherf and Weinzierl Reference Scherf and Weinzierl2020: 387).
The indictment has several counts. One concerns quantification. Estimates of how much a person benefits from social cooperation require a pre-benefit baseline and an idea of what counts as a benefit. Let that baseline be some state of nature in which social cooperation is minimal (Murphy and Nagel Reference Murphy and Nagel2002: 17). Also let the quantity of the benefit that person receives from a given good be equal to what that person would be willing to pay, under certain ideal conditions, for that good to be produced. Determining what counts as a benefit for that person will still be a challenge. A person can benefit from a good in several ways, some more proximal (a new highway allows them to move their goods to market), some more distal (the highway removes traffic from the country road that passes outside their house, making the area more pleasant for them to live in). Moreover, some of those benefits involve goods that are not public in the economic sense, such as education, the internet, corporate structure, or various forms of state subsidy (for examples of public investment creating such goods see Mazzucato Reference Mazzucato2013: Chs 5–7). To determine how much a person ought to pay for a given package of goods, we would need to settle not only what each category of goods includes and how to quantify the benefits falling within it, but also how to distribute the cost of producing the relevant good between proximal and distal beneficiaries.Footnote 4 The problem is serious because distal benefits will sometimes constitute the larger and more important part of the total benefit that a person derives from a given public good. People benefit clearly and greatly from the existence of roads they have never travelled on.
A related problem concerns valuation. Under anything but the most paternalistic accounts of value, to say that a good creates a benefit for a particular person, we must pay some heed to that person’s preferences and valuations. However, people differ in what they value. Some value state education, others value home-schooling. Some value highways, others value unspoilt pastures. Even when two persons value a good and prefer to have the same amount of it, they may still derive significantly different benefits from that good because of income effects. That is conspicuously true of money itself, the marginal utility of which declines sharply as one’s income rises (Murphy and Nagel Reference Murphy and Nagel2002: 17–18; Fried Reference Fried, O’Neill and Orr2018: 166).Footnote 5 Some aspects of this problem may prove more tractable than others (US Office of Management and Budget 2023: 65–66 surveying the literature and concluding that ‘1.4 is a reasonable estimate of the income elasticity of marginal utility for use in regulatory analyses’), but any defence of the benefit principle needs to offer a credible solution to it.
A third problem is that a tax system designed in accordance with the benefit principle would seem to have some strongly counter-intuitive distributive implications. The claim that taxation is payment for benefit assumes that people are entitled to their pre-benefit holdings, or else the relevant transaction would not count as an exchange. Moreover, if equal benefit justifies equal tax liability, the benefit principle seems inconsistent with progressive taxation, and would even seem to entail that those who rely most on public goods, including some of the weakest persons in our societies, should bear the heaviest tax burdens. These implications beg some of the major questions of distributive justice, so reliance upon the principle would appear to be ‘an argumentative poison-pill’ for anyone with egalitarian political commitments (Verbeek Reference Verbeek and Gaisbauer2015: 70).Footnote 6 Classical applications of the principle to public goods recognize the problem when they stipulate fair distribution as a background assumption (Wicksell Reference Wicksell, Musgrave and Peacock1958: 79; Lindahl Reference Lindahl, Musgrave and Peacock1958a: 173). Once that background condition is relaxed, the principle’s normative force seems to wane very quickly. If justice requires that I have ten apples, the distribution gives me two, and state activity supplies me with one more, it is not obvious that I owe the state payment for that one apple (Lindahl Reference Lindahl, Musgrave and Peacock1958b: 227: ‘justice in taxation is inextricably linked with justice in the distribution of property, since it would obviously be nonsense to speak of “a just portion of an unjust whole”’).
By and large, this torrent of criticism has put tax and public finance theorists off the benefit principle. In the words of a colleague, the principle looks too strong to dismiss as an insight, but too weak to take seriously as a theory.Footnote 7 At the same time, the principle retains its strong appeal not only in political rhetoric but also in modern debates in public finance. For example, it continues to animate several modern accounts of the funding of public goods (Brennan Reference Brennan1976; Moulin Reference Moulin1987; Hines Reference Hines2000), and recent public finance literature has sought to integrate its insights into theoretically more robust accounts of income taxation (Weinzierl Reference Weinzierl2018; Scherf and Weinzierl Reference Scherf and Weinzierl2020). Moreover, a number of modern theorists find normative or justificatory merit in the principle (Fried Reference Fried, O’Neill and Orr2018; Stewart Reference Stewart2022; for a similarly motivated normative account of desert, see Sheffrin Reference Sheffrin2017).
The present article joins this effort to rehabilitate the benefit principle by trying to identify more clearly the moral insight that underlies its continued political and theoretical vitality. More specifically, I propose an account that explains why and when it is morally permissible to levy tax in return for benefit conferred. One might think that the answer is blindingly obvious: when you derive a benefit from another person’s activities, you ought to pay for it; quid pro quo. However, we already know that the obvious answer is mistaken. Conferral of a benefit gives rise to a duty to pay something in return only under certain background conditions. A defence of the moral permissibility of levying tax for benefit conferred must therefore be able to specify those conditions in more detail. I argue that we can begin to meet this challenge by tracing the force of the benefit principle to five moral ideas: individual justification, causal feedback, reciprocity, opposable valuation, and non-objectionable baseline. I use these ideas to propose a canonical formulation of the moral basis of the benefit principle, and to explain how that formulation addresses the problems of quantification, valuation, and the significance of the background distribution.
Needless to say, the aim of moral argument is to set the ballpark of permissible tax design, not to generate a fully specified policy standard. My discussion does not, therefore, compete with familiar formulations of the benefit principle in the theory of public finance, though I try to show how the moral case for the benefit principle is connected to those formulations, and to identify some further directions for economic research.
The paper proceeds in two steps. Section 2 distinguishes three different versions of the benefit principle in tax theory. The first concerns the funding of public goods (the ‘public goods’ version), a second treats the benefit principle as a general standard of just taxation (the ‘general’ version), and a third tries to integrate the idea of benefit into otherwise non-benefit-based accounts of tax justice (the ‘integrated’ version). Section 3 uses a well-known illustration from Adam Smith to trace the normative force of those versions to the aforementioned five moral ideas, and to explain how these supply what I will call a ‘distribution-sensitive’ but ‘justice-independent’ moral basis for the benefit principle.
As the paper focuses on the general normative question of what counts as benefit and why benefit matters morally, I will not take a position on downstream questions about the appropriate base (labour, capital, consumption etc.) for assessing individual benefit for tax purposes, or the appropriate rate of taxation. For simplicity, I will also set aside international tax justice,Footnote 8 though I believe that the general ideas I will identify in section 3 capture much of the appeal of the benefit principle in that context too.
2. Three Versions of the Benefit Principle
Reference to the benefit principle creates the misleading impression that there is only one such principle to take or leave. In fact, the benefit principle appears in the literature in at least three different versions. Here I try to distinguish the normative commitments of each version, with particular emphasis on their respective conceptions of benefit and their relation to the justice of the background distribution.
2.1 The public goods version
We can look at taxation as payment for certain non-rival and non-excludable goods that a person receives as a result of state-mediated social cooperation, e.g. clean air, security, basic infrastructure, a stable economy etc.Footnote 9 We can also look at it as a tool for promoting a certain pattern of distribution across society (egalitarian, prioritarian, welfare-maximizing etc.). When we consider these views of taxation as expressions of preferences, it seems clear that endorsing one view does not commit us to endorsing or rejecting the other. As Musgrave and Musgrave (Reference Musgrave and Musgrave1980) put it:
People’s attitudes toward redistribution need not coincide with their preferences for social goods. A person who wants public services should not have to oppose them because he dislikes redistribution, and vice versa. (Musgrave and Musgrave Reference Musgrave and Musgrave1980: 18)
One implication of the separability of those preferences is that fusing them into a single fiscal function may lead to socially sub-optimal outcomes. If we prefer an egalitarian distribution and try to improve the position of the worst-off through the provision of public goods, we may end up with an over- or under-supply of those goods, depending on how much inequality there is in our society. The better policy, on this approach, would be to try to keep the two functions separate, i.e. to determine our appetite for public goods and fix who should pay for them and in which proportion assuming something close to our preferred distribution, and to work towards that distribution using direct transfers instead.Footnote 10
This version of the benefit principle speaks to the former function only, and solves a problem specific to the nature of public goods (I leave aside club goods, where the problem appears in somewhat attenuated form). The problem is that people have an incentive to under-report how much they would be willing to pay for the production of non-excludable goods. The principle addresses that problem by identifying the optimal provision level of a given public good and setting individual tax contributions to its cost. The ideal level of provision depends on whether there is marginal social benefit in producing another unit of that good rather than leaving the required resources in private hands. The level of individual contributions depends on an assessment of the benefit that each individual derives from the public good in question (unlike markets for private goods, where the price per unit is uniform and the level of consumption varies, markets for public goods feature a uniform level of consumption at differentiated prices). So conceived, the benefit principle is committed to an individualized approach to the valuation of a given public good, but it aims to leave the question of distribution open, at least insofar as distributive preferences do not affect the production or allocation of that good (I return to this point below).
Tax as a distributive device and tax as payment for benefit from public goods need not conflict even when we look at distribution as a question of justice or fairness, rather than preference-satisfaction. For example, one may reject the benefit principle as a norm of tax justice and still endorse it as a norm for the optimal provision of public goods. In fact, the two views of taxation may not stand in conflict even if we read them both as claims of justice. It is plausible to think that justice permits us both to charge people tax for the benefit they receive from public goods and to bring about a certain pattern of resource distribution. Attending to both goals will require some practical trade-offs, but the task seems normatively coherent. That position counts Knut Wicksell, Hans Lindahl and Richard Musgrave among its advocates (Wicksell Reference Wicksell, Musgrave and Peacock1958: 79; Lindahl Reference Lindahl, Musgrave and Peacock1958b: 214; Musgrave Reference Musgrave1959). It is also the position Murphy and Nagel take in The Myth of Ownership (Murphy and Nagel Reference Murphy and Nagel2002: 85: ‘Against the background of a distribution which is assumed to be just, the funding of public goods may best be arranged according to some standard of proportionate benefit’). See also Rawls Reference Rawls1999: 246, though Rawls notes at 247 that its justification depends on his two principles of justice, rather than on maxims like benefit received or ability to pay). Deployed in this way, the benefit principle claims to state a necessary condition for the justice of taxation only in the case of public goods. And since this version of the principle maintains the separation of the fiscal functions, it can be attached as a relatively self-standing module to one’s preferred account of distribution.
2.2 The general version
At the other end of the spectrum lies a version that treats the principle as a standard governing the justice of all taxation; call this the ‘general’ version. That version is the most obviously problematic. For a start, it is clear that the general version would attach potentially unacceptable conditions to redistributive taxation. ‘No taxation without benefit’ leaves no space for distribution from the richest to the poorest unless that policy produces a benefit for the former (note that there is no symmetry here; ‘no taxation without distribution’ is not a remotely plausible principle). A similar conclusion seems to follow when we state the principle in positive form. ‘Pay tax for every benefit’ suggests that the weakest persons in our societies, who rely more on state institutions, should pay more in tax. Either way, the justification of policies that strike many people as necessary features of modern society, e.g. basic income support, or progressive taxation, emerges as contingent on whether those policies confer benefits to those that fund them. If one’s preferred theory of justice rejects that condition, e.g. because it justifies those policies on egalitarian or prioritarian grounds, the general form of the benefit principle emerges as flatly inconsistent with just distribution (Murphy and Nagel Reference Murphy and Nagel2002: 18: ‘The benefit principle is, in fact, incompatible, as a matter of political morality, with every account of social justice that requires government to provide any kind of income support or welfare provision whatsoever to the destitute (let alone more strongly egalitarian distributive aims)’.)
The conflict between this general version of the benefit principle and distributive intuitions may turn out to be less radical than it looks. Consider progressive taxation. The benefit principle does not allow the state to charge persons at the top end of the distribution higher rates of tax simply because those persons have a higher income than everyone else. Rather, it sees the justification of such taxation as turning on whether those persons derive greater benefits from social cooperation than others do. But if we can identify specific benefits produced by social cooperation and enjoyed only or predominantly by those at the top, the benefit principle can deliver a higher tax burden for those persons without resort to a progressive tax rate.
Despite these possibilities of rapprochement, Murphy and Nagel argue that the benefit principle remains radically incompatible with fair distribution. One aspect of their critique is that the principle is ‘myopic’ because it fails to see the justification of taxation and the justification of spending as two sides of the same coin (Murphy and Nagel Reference Murphy and Nagel2002: 17: ‘[The benefit principle] takes expenditure as given, and allocates taxes in proportion to the resulting benefit. That is an example of what we mean by myopia’; Kaplow Reference Kaplow2008: 210: ‘when there is a system of redistributive taxation in place, which itself may be freely adjusted, the purpose of isolating the benefit tax component is not apparent’.) This is a mistake. The benefit principle is, in fact, the paradigm of a spending principle. Its central ambition is to determine whether spending on a given set of public goods is worth the cost for every person who will be asked to bear it, and to distribute that cost in proportion to benefit received. In that regard, the benefit principle is clearly superior to the ability-to-pay principle, which has no answer to questions of spending.Footnote 11 So while it is true that the benefit principle may conflict with some intuitions about spending, namely those that treat some expenditures as justified regardless of whether they benefit every payer, it is not true that the principle generally fails to see the connection between taxation and spending.
The real point of Murphy and Nagel’s critique is that the insight underlying the benefit principle, namely the fairness of exchanging payment for benefit, is not morally basic at all. Rather, that insight owes its force to the fact that the exchange takes place under certain conditions, including conditions about who is entitled to what before that exchange takes place (Murphy and Nagel Reference Murphy and Nagel2002: 8). If it is unjust for me to have all the apples and for you to have none, it is by no means clear that I may require you to pay me for the benefit of an apple. The benefit is no less real than it would have been under a just distribution, but its moral significance varies depending on the background conditions against which it is conferred. Similarly, the fact that a person receives a certain benefit in the form of certain public or private goods does not dictate any particular answer as to whether justice requires that this person pay tax in return. That answer will depend on the qualities of the institutional scheme that delivers the good in question, e.g. on the resources, opportunities and entitlements that the scheme affords to that person. In that respect, benefit turns out to be very similar to desert: an intuitively plausible idea whose strength nevertheless turns out to be both limited and highly dependent on the shape of certain institutional arrangements (Scanlon Reference Scanlon2018: Ch. 8, arguing that claims of desert do not fix any particular level of sacrifice that others ought to accept for the sake of the deserving; Sheffrin Reference Sheffrin2017: 143–148). If this is correct, there can be no justice-independent standards for deciding what counts as a benefit for the purposes of the principle, because there can be no justice-independent way of justifying why benefit matters in the first place.
Note that this critique need not rely on Murphy and Nagel’s stronger claim that pre-tax entitlements are a ‘myth’. As several comments on The Myth of Ownership have noted, a right to tax is compatible with persons having a prima facie moral claim to their pre-tax earnings (Brennan Reference Brennan2005: 129–130; Marmor Reference Marmor2005: 71–73).Footnote 12 It will still be the case, however, that the exchange of some of those earnings for benefit will be justified only against certain background conditions. Murphy and Nagel’s most fundamental charge against the benefit principle is that there is no justice-independent way of specifying those conditions.
2.3 The integrated version
The benefit principle also appears in the literature in a third and somewhat less often-noted guise, namely as an insight integrated into otherwise non-benefit-based accounts of tax justice; call this the ‘integrated’ version. Adam Smith gestured towards that approach by treating a person’s income-earning ability or endowment as a benefit that they derive from state activity. His first maxim of taxation reads:
The subjects of every state ought to contribute toward the support of the government, as near as possible, in proportion to their respective abilities; that is in proportion to the revenue which they respectively enjoy under the protection of the state. (Smith Reference Smith1776: Book 5, Ch. 2)
The maxim uses the idea of benefit as a general justification of why the state may legitimately take income-earning ability as the fundamental tax base (Musgrave Reference Musgrave1959: 42; Stewart Reference Stewart2022: 50; Musgrave and Musgrave Reference Musgrave and Musgrave1980: 219: ‘Benefits are here viewed in terms of protection received and are thus related to income which, in turn, is also a measure of ability to pay’). In contemporary public finance, Matthew Weinzierl has deployed Smith’s insight in a more focused way, as a corrective to the tendency of mainstream optimal tax theory (an exception is Stancheva Reference Stancheva2017: 1931) to take individual income-earning endowments as a natural given (Weinzierl Reference Weinzierl2018: F39; Buchholz and Peters (Reference Buchholz and Peters2007: 157) combining the benefit-based Lindahl equilibrium with an equal sacrifice postulate). Weinzierl argues that a more accurate estimate of a person’s endowment should incorporate the effect on it of the consumption of public goods. In joint work with Robert Scherf, he claims that this move is sensible because
it is implausible to assume that the level and mix of public goods provided by a government are irrelevant for the distribution of income-earning abilities of its taxpayers or, therefore, those taxpayers’ preferences over public goods (Scherf and Weinzierl Reference Scherf and Weinzierl2020: 387).
Why is that assumption implausible? Drawing on Smith’s insight, Weinzierl argues that
an individual’s ability [is] a function of the activities of the state to which that individual contributes. Any endowments are moot unless individuals cooperate to fund public goods that set the stage for those abilities to develop (Weinzierl Reference Weinzierl2018: F39).
Weinzierl’s account exemplifies the integrated deployment of the idea of benefit in two ways. First, its appeal to the beneficial effect of public goods consumption changes the measurement of individual endowments and, therefore, the distributions that the optimal tax framework would otherwise produce.Footnote 13 More specifically, Weinzierl shows that amending that framework in this way allows it to generate distributions that have a reasonably close match with actual policy, and especially progressive taxation (Weinzierl Reference Weinzierl2018: F41). Second, the benefit principle is not simply one of several principles that could justify this amendment.Footnote 14 Inasmuch as the justification for taking account of the effect of public good consumption on individual endowments turns on the idea that individuals whose position improves as a result of social cooperation ought to pay something in return, the benefit principle is the most natural candidate for the job (Stewart Reference Stewart2022: 50: ‘we need only observe that the critique of the benefit theory concerning distribution of the tax burden does not undermine the theory as a justification for taxes to finance the necessary collective endeavour of the state’).Footnote 15
Note, however, that unlike the public goods version, this integrated version does not assume the justice of the background distribution. To that extent, it remains exposed to Murphy and Nagel’s critique. More specifically, while Weinzierl does not set out to defend the benefit principle on moral or normative grounds (Weinzierl Reference Weinzierl2018: F39: ‘this paper is not intended to defend classical benefit-based taxation as a normative principle. Other principles, including utilitarianism, may well be preferable from the perspective of moral and political philosophy’), the ‘stage-setting’ relation that he describes assumes that causing an improvement in a person’s income-earning capacity justifies requiring that person to pay something in exchange. That claim is plausible, but it does not reflect a general moral truth. A new shopping mall may increase the turnover of other local businesses without entitling its owners to recoup that windfall. We therefore need to ask which other conditions must be satisfied for contributions to another person’s wellbeing to have that normative consequence. And if Murphy and Nagel are right, those conditions cannot be discussed independently of questions about the justice of the background distribution.
An account of those conditions would not only help justify integrating public good consumption into estimates of income-earning ability. It would also help us determine how far, if at all, the same might apply in respect of non-public goods. The opening quote by Warren suggests that private goods distributed by the state, e.g. education, corporate structure, or the internet, have a similar enhancing effect on a person’s natural endowments. Perhaps the same goes for various forms of ‘social capital’ that emerge without any state involvement. As Barbara Fried puts it:
Many of the reasons people value living in a particular community have to do with its built social and physical environment: norms of civility and trust, cultural institutions, good restaurants, job opportunities, transportation and housing infrastructure, architecture, good teachers, skilled doctors, well-functioning hospitals, and all the other working parts that go into making a city, state, or country a place that works, a place in which people want to live … That is why the wealthy … [are] willing to pay a very high (tax) price to live in Manhattan, Chicago, London, Paris, and so on. Why shouldn’t the members of a polity who have collectively generated and maintained its social capital be entitled to charge for the right to enjoy it, along with the explicit public goods that the jurisdiction supplies? (Fried Reference Fried, O’Neill and Orr2018: 164–165)Footnote 16
Suppose that we could write out functions that produce reasonable estimates of the beneficial contribution of social capital to income-earning ability.Footnote 17 Would we be morally justified in taxing people more for some or all of those benefits?
3. Why Benefit Matters
We need an account of the conditions under which the receipt of a benefit grounds a duty for the recipient to pay something in return. One aspect of that challenge is to determine whether those conditions are justice-dependent, as Murphy and Nagel argue. Another is to determine which of the many ways that social cooperation benefits individuals can ground a legitimate tax demand. I will try to derive an answer to these questions using an example that, in my view, brings out the moral ideas that give the benefit principle its appeal in all three versions discussed in the previous section. The example comes from Adam Smith’s discussion of tolls, where he notes:
When the carriages which pass over a highway or a bridge, and the lighters which sail upon a navigable canal, pay toll in proportion to their weight or their tonnage, they pay for the maintenance of those public works exactly in proportion to the wear and tear which they occasion of them. It seems scarce possible to invent a more equitable way of maintaining such works. (Smith Reference Smith1776: Book 5, Chapter 1(III))
The goods that Smith discusses in this passage are excludable and, to a certain extent, rival, so they lie somewhat outside the comfort zone of the benefit principle. That makes the example particularly useful for my discussion.
If Murphy and Nagel are right, those who think that the system of tolls that Smith describes could be fair assume that the exchange of toll for service takes place under certain background conditions, including conditions about property ownership and just distribution. Classical applications of the benefit principle to public goods concede Murphy and Nagel’s claim by treating just distribution as a background assumption. My intuition is that this concession is unwarranted, and that receipt of benefit can generate a duty to pay tax even in conditions of distributive injustice.
More specifically, I want to suggest that the benefit principle draws its moral force from the conjunction of five ideas, which I will label:
-
Individual justification.
-
Causal feedback.
-
Reciprocity.
-
Opposable valuation.
-
Non-objectionable baseline.
Here is how these ideas coalesce in Smith’s example. The justification for the imposition of tolls is addressed to each and every individual payer, and does not rely on an appeal to the common good (individual justification). That justification is based on two causal claims: the use of the bridge is causally connected to an improvement in the payer’s position, and the payment of tolls is causally connected to the building and maintenance of the bridge by the relevant institution (causal feedback). The demand for the toll is based on a reciprocal relation at two distinct levels: the toll buys the use of the bridge, and it also buys the commitment of other persons to the institution that produces the relevant benefit for all (reciprocity). The benefit in question is something that the payer can avoid having to pay for if they do not use the bridge, depending on their valuation of the relevant good (opposable valuation). Finally, the use of the bridge is not something that the user was already entitled to and does not otherwise violate an important moral interest of that person (non-objectionable baseline).
I will make some brief observations about each of those five ideas and their relation, and then suggest how their conjunction yields a ‘distribution-sensitive’ but ‘justice-independent’ moral argument for the benefit principle.
3.1 Individual justification
The benefit principle aims to justify taxation to each and every person who may be required to bear a tax burden. Smith’s argument exemplifies that idea because it speaks directly to the users of the bridge, just as Warren’s argument for taxation is addressed directly to the rich entrepreneur. Individual address has a good claim to being a fundamental aspect of moral justification. According to moral contractualism, the philosophical tradition that takes that idea to heart, a person is entitled to a justification as to why and when others may require them to bear burdens, and to have that justification framed in terms that they cannot reasonably object to (from a vast literature see Scanlon Reference Scanlon1998; Darwall Reference Darwall2006). By contrast, utilitarian justifications of taxation are not clearly committed to justifying taxation to each individual bearer of tax burdens, because they focus on social welfare and therefore aggregate individual preferences and benefits.
Within a contractualist moral justification, the fact that a person has received a benefit from social cooperation serves to block or defeat objections that this person would otherwise have against a principle that required them to pay tax (‘someone who does not benefit from the principle could object to it, but you do benefit, so you cannot’). Knut Wicksell appealed avant la lettre to this objection-blocking function of benefit as a principle of justification in one of its earliest statements in modern tax theory:
No-one can complain if he secures a benefit which he himself considers to be (greater or at least) as great as the price he has to pay. (Wicksell Reference Wicksell, Musgrave and Peacock1958: 79. See also Buchanan Reference Buchanan1975; Johnson Reference Johnson2010: 187 on the link between social justice and economic theory in Wicksell’s account)
Wicksell used this idea to propose that decisions about public goods should be decided as a package with proposals for their funding through taxation. He also believed that individual justification required unanimity in the relevant vote, on the ground that such a rule would make it rational for people to reveal their true valuations of the public good, lest they voted down proposals that match their actual preferences (Wicksell Reference Wicksell, Musgrave and Peacock1958: 89). Later literature has criticized the unanimity requirement as too demanding (Musgrave Reference Musgrave1986: 94; Coleman Reference Coleman2002: Ch. 11 considers three arguments in favour of the unanimity rule (Pareto efficiency; liberty of exchange; and moral scepticism) and discusses why none of them supports the rule), but I believe that it is possible to reinterpret that requirement as making a broadly similar demand to modern contractualist accounts of moral justification. The only major difference between them concerns the grounds of objection to a proposed principle or decision, a point to which I turn in the discussion of opposable valuation below.
3.2 Causal feedback
All appeals to the benefit principle are appeals to a causal argument. In Smith’s bridge example, the bridge causes an improvement to the position of its users. Weinzierl points to the causal contribution of public goods on a person’s income-earning capacity. Warren points to the causal impact of a mix of public and private goods to the entrepreneur’s success. Fried draws attention to the causal impact of less formal or state-centric cooperative schemes on the quality of life of taxpayers.
In all but one of those arguments, the causal relation between the benefit and tax payment goes both ways. Without tolls, the bridge could not be maintained, and without the protection of the state, individuals could not enjoy securely the revenue that they do, grow their natural endowments to the same extent, have favourable circumstances for developing their entrepreneurial ideas, and so on. The odd one out is Fried’s argument. The various forms of social capital that she describes do confer benefits on individuals, and require a level of support from formal and informal social structures, but there is no clear causal link between the payment of tax and their creation and maintenance. That indicates that the moral case for levying tax in exchange for this particular benefit must be couched in different terms.
3.3 Reciprocity
Equally central to all invocations of the principle is the idea of reciprocity. In the most direct and obvious sense, this involves the exchange of benefit for payment. But there is an institutional sense at work too. In participating in the scheme or institution that produces the benefit, the persons who receive it are also committing themselves to support the scheme in question, and obtain the commitment of other beneficiaries to do the same (see e.g. Irish Commission on Taxation and Welfare 2022: 42 citing reciprocity as one of its guiding principles). That is the sense in which Oliver Wendell Holmes invokes the benefit principle when he says that ‘taxes are what we pay for civilized society’ (Compania General de Tabacos de Filipinas v Collector of Internal Revenue (1927) 275 US 87: 100; a more familiar, if apocryphal, version of the quote is: ‘I like to pay taxes. With them I buy civilization’). The idea of reciprocity at the institutional level can explain why payment entitles every user of the bridge to complain in case any other user fails to pay the toll, even though the user receives the benefit by someone else, namely the institution that operates and maintains the bridge (I discuss exemptions from payment due to a person’s position in the distribution in section 3.5 below). Such relations of reciprocity become more complex when the institution in question extends over time and generations, as conferral of benefit and payment of contribution will typically be asynchronous, e.g. a person receives most of their schooling before they pay any tax, they pay into pension funds long before they enjoy their pension etc. (Otsuka Reference Otsuka2023: 2 defending a reciprocity-based account of collective pensions ‘as a form of cooperation between persons which is to the expected benefit of each’). However, those schemes remain reciprocal insofar as they extend the same bundle of duties and benefits to everyone, under the same terms.
Note that, here too, the social capital that Fried describes cannot be fully accounted for in terms of reciprocity in either sense. The professionals and artists who make the city attractive, and those who participate in the non-institutional practices that create social norms of civility and trust do so for the intrinsic value of those practices, rather than for their exchange value. It would therefore be a moral mistake to treat any tax extracted from the beneficiaries of those practices as payment for benefit.
3.4 Opposable valuation
One point of contention in early debates in public finance was whether the value of a benefit should be assessed ‘subjectively’ or ‘objectively’. Another was how far we should assume that all persons derive the same amount of objective benefit from a given public good (Brennan Reference Brennan1976; Aaron and McGuire Reference Aaron and McGuire1976; Hines Reference Hines2000). A third concerned whether the appropriate measure for that purpose should be marginal or inframarginal (Moulin Reference Moulin1987; Hines Reference Hines2000; Scherf and Weinzierl Reference Scherf and Weinzierl2020: 392 note that the focus on inframarginal benefit draws attention to the fact that the aggregate of inframarginal benefits for all persons will typically exceed the cost of the relevant public good, thereby raising a normative question of distribution). The first debate matters most for my present purpose, so I leave the other two aside.
From a contractualist perspective, the move from the idea of individual justification to that of a subjective valuation (Wicksell Reference Wicksell, Musgrave and Peacock1958: 79: ‘if he secures a benefit which he himself considers to be (greater or at least) as great’; Lindahl Reference Lindahl, Musgrave and Peacock1958a: 173) looks natural. In a system that aims to be sensitive to every person’s valuation, it would be presumptively objectionable to charge the same tax to persons who attach very different valuations to the same good (Buchholz and Peters Reference Buchholz and Peters2007: 419; Nozick Reference Nozick1974). That idea is common to all three versions of the benefit principle as defended in the literature, though it is most prominent in the case of public goods.Footnote 18
Of course, sometimes it will be practically difficult or impossible to obtain accurate information about individual valuations. The cost of detecting who has breathed more clean air in our city will be typically prohibitive, so in the absence of some clever proxy, a policy that assumed that members of our community attach equal value to that good would not be objectionable to any one of them. Epistemic limitations will sometimes go hand in hand with political ones. As George Klosko has shown, it will often be practically difficult to ‘unbundle’ public goods that every member of the community needs (e.g. security or basic infrastructure) from those that are discretionary (Klosko Reference Klosko1990: 200–205; Miller Reference Miller and Dowding2004: 133–134).
However, I want to suggest that, as a moral matter, benefit-based taxation need not always track a person subjective valuation, even when we have solved the problems of preference-revelation and bundling. Sometimes a public good may be worth more to a person than it is to others, yet it would be morally permissible to allow that person to pay only as much as anyone else. Similarly, sometimes that good may be worth less to a person, yet it would be permissible to require them to pay as much as others. Wicksell’s account does not make sufficient allowance for these situations because it assumes that every valuational difference has the same moral significance. That account is, in that respect, insensitive to the reasons for which a person attaches a particular valuation to that good.Footnote 19
Suppose that the bridge in Smith’s illustration is of especially high value to two persons. One is an entrepreneur whose trucks use the bridge to transport goods across the country. The other is a disabled person, who finds that the bridge is the only way across that allows them to use their mobility vehicle. The entrepreneur attaches greater value to the good because it helps their business prosper. The disabled person attaches greater value to it because their condition has put them in a situation of great and lifelong need for it. The question is whether this difference matters when we consider how the cost of that good should be distributed. In contractualist terms, this involves asking whether the disabled person could reasonably reject a principle that required them to pay more by some increment and, conversely, whether others could reasonably reject a principle that required them to socialize that increment. Three features seem to me to stand out in that regard. One is that, on certain assumptions about the incidence of serious disability in the community, the cost to everyone else of socializing the increment will be very small. Another is that the cost to the disabled person of having to pay more for the good across their lifetime could be serious. A third is that the disabled person’s greater need for the good is due to brute bad luck. Contractualism holds that these features affect whether a given distribution of the cost of the good is reasonably rejectable insofar as they speak to the strength of the reasons that each person has in the situation (Scanlon Reference Scanlon1998: 231 et seq. refers to this method as involving ‘pairwise comparisons’; Nagel Reference Nagel1979: 122–127; Kamm Reference Kamm1993: 87). This assessment will be complex, as the strength of the disabled person’s reasons will depend on the existence of other forms of social support and compensation, but it does not seem plausible to think that those reasons will never be sufficiently strong to allow the disabled person to object to being required to pay more. If that is the case, it will sometimes be permissible to diverge from subjective valuation and distribute the increment across all payers.
The more general point is that while subjective valuation is legitimate as a default, the moral significance of subjective differences in valuation is not fixed. The real question is not simply whether a particular person’s valuation of a good differs from those of others, but whether that valuation is morally opposable to them, i.e. whether it is based on reasons that others ought to treat as affecting what they may require of that person in relation to that good. Subjective valuations will typically be opposable in this sense, as Wicksell and Lindahl anticipated. However, the case for this will depend not on the mere fact that they are subjective, but on the strength of the reasons that underlie them (I discuss the relation between the idea of opposability and benefit-based arguments in Voyiakis Reference Voyiakis2025).
3.5 Non-objectionable baseline
If I steal someone’s wood planks, take away their land, build a bridge on that land, and try to charge them tolls for its use, that person could legitimately refuse to pay. The very idea that they receive a benefit from the use of the bridge is misplaced because the background conditions of the exchange of payment for the use of the bridge are wrong. Writ large, this is Murphy and Nagel’s deeper critique: without a fair distributive baseline, the moral force of the benefit principle wanes. The question is whether that critique is justified.
I think we should distinguish between two ways in which the conditions that give the benefit principle its moral force could be connected to distribution. Call those conditions justice-dependent if we cannot identify them without making assumptions about the justice of a particular distributive pattern. Call them distribution-sensitive if they allow the normative force of the principle to vary under certain distributions, whether or not those distributions meet the demands of justice. Whereas a justice-dependent argument claims that, absent a just distribution, benefits from social cooperation do not create a duty to pay tax, a distribution-sensitive argument holds that a benefit-based tax demand can be justified whether or not the distribution falls short of justice, insofar as that tax demand is not objectionable in certain other ways.
Some objections will concern personal liberty. In Smith’s example, say that the institution that builds and maintains the bridge also blocks access to an existing side road in order to force everyone onto the toll-charging alternative. Users of the side road could object to a principle that allowed the institution to appeal to this as a reason for restricting their transit options. The stolen wood and land variation illustrates a more complex version of the theme. An intuitive way to frame the relevant objection would be to say that a benefit-based toll demand is unjustified insofar as the production of the relevant benefit violates that person’s property rights over the land and the planks. That framing is, however, unduly narrow, as that person’s claim need not appeal to a fully determined system of property rights. It can find enough of a basis in that person’s interest in having, as T.M. Scanlon puts it, a measure of ‘control over objects that are needed to provide for one’s life, and in having the stable control over these possessions over time that is needed to plan, and to carry out one’s projects’ (Scanlon Reference Scanlon2018: 106). It is basic interests of that kind that help explain why that person could object to systems of property rights that allowed me to take those objects away at will and without compensation.
In other cases, a benefit-based tax demand will be objectionable not because it interferes with property or other personal liberty interests, but because it is excessively burdensome for a person because their position in the distribution makes it very difficult for them to meet the relevant tax demand. Moreover, as we have seen, sometimes the tax demand will be objectionable because the benefit and the payment are not connected in the right way, e.g. because the payment is not causally necessary for the production of the benefit, or the cost of producing that benefit can be lowered for everyone without compromising the quality of the good, e.g. lower tolls could still cover the cost of maintaining the bridge.
All of these objections are connected to distribution, but none of them is premised on the claim that a given distribution fails to meet the demands of justice. That is, the core of the relevant objections to a given benefit-based tax demand is not ‘the background distribution is unjust’ but more like ‘the demand has an objectionably adverse impact on me’, while allowing that the moral significance of that impact will sometimes be a function of one’s position in the distribution (Scanlon Reference Scanlon2018: 157 noting that the relevant justification appeals ‘to the reasons individuals have for accepting such institutions based on how their lives would be affected’).
3.6 The ideas combined
These five ideas support the following formulation of the moral basis of the benefit principle:
A person whose well-being improves as a result of the activities of an institution cannot object to being required to pay for that benefit, when such payment is necessary for the maintenance of the institution that produces the benefit for all, the institution does not infringe an important liberty interest of that person, the valuation of that benefit is opposable to them, and payment is not excessively burdensome given their position in the distribution.
I believe that this formulation captures the basic intuitions of those who find merit in the benefit principle in the three versions I described in the previous section. It explains why Weinzierl is right to treat the increased earning ability that flows from the consumption of public goods as a ground for requiring beneficiaries to pay more in tax, and why Warren is right to insist that the argument can extend to the consumption of state-produced private goods too. At the other end of the spectrum, the formulation can explain why conceptions of benefit that focus on certain ‘social goods’ lie outside the scope of the principle and need to be justified in different terms.
The proposed formulation does not state a generally necessary condition for the justice of taxation. It addresses only the moral permissibility of benefit-based tax demands, and so does not preclude demands justified on different grounds. In particular, it does not exclude redistributive taxation. By that token, however, it suggests that the general version of the benefit principle must be abandoned as insufficiently justified.
One might argue the account I have provided still suffers from the ‘myopia’ that Murphy and Nagel caution against. If the moral case for the benefit principle reflects only one moral idea within a much larger scheme of principles of justice, we should expect its justification to be bound up with that of that overall scheme (see e.g. Rawls Reference Rawls1999: 247). I think that the point is fair but innocuous. Most practical principles must be embedded in larger justificatory schemes and may require qualification in specific contexts. The challenge is to identify those contexts, and to explain who might object to a given principle of tax justice and why. Murphy and Nagel’s critique of the general version of the principle hits its target because it draws attention to the objections of victims of unjust distribution of property and vulnerable persons who are likely to need more extensive state support than others. By the same standard, the formulation I have offered counts as an improvement insofar as it manages to take adequate account of such objections.
4. Conclusion
The benefit principle has fallen out of favour as a general account of tax justice, but it has not lost its intuitive appeal. My aim in this paper has been to account for its moral force. I approached the task by distinguishing between the different ways that the benefit principle has been invoked in the literature. I argued that, even though each version has different normative commitments, they all draw their moral force from an account of the background conditions against which the conferral of a benefit grounds a duty to pay something in return. I then constructed an account of those conditions by identifying five moral ideas (individual justification, causal feedback, reciprocity, opposable valuation and a non-objectionable baseline) common to all versions of the principle, and used them to fashion an argument for the permissibility of levying a tax charge in exchange for benefit.
As I noted in the Introduction, my account is not in competition with the formal versions of the benefit principle discussed in the public finance literature, but to identify the moral ballpark within which such versions may move. I hope to have shown that this ballpark is broader than critics of the benefit principle have supposed. A tax demand for benefit conferred may be justified even if the background distribution is not just, and the amount of that tax need not always track a person’s subjective valuation of the benefit. Moreover, it is permissible for the state to levy tax for more distal benefits flowing from the consumption of both public and some private goods, as long as these enhance a person’s income-earning ability. Capturing these effects in formal terms is a task for economists, and further research in that direction would be welcome.
Finally, the argument I have outlined does not purport to show that the benefit principle is a necessary condition of the justice of all forms of taxation. If anything, the fact that the principle performs very well when applied to public goods and not well at all when applied as a general standard of tax justice should be enough to suggest that taxation is not a normatively unified domain: different principles account for different aspects of it. My point has been that the benefit principle deserves better press than it has been getting, and that the ideas to which the principle owes its force track moral requirements that every theory of taxation should be sensitive to.
Acknowledgements
I am grateful to Reuven Avi-Yonah, Eduardo Baistrocchi, Dominic de Cogan, Sebastian Gazmuri Barker, David Kershaw, Ian Roxan, Andy Summers, Ziming Xia, participants in the LSE Tax Seminar Series (which hosted the paper in Lent 2023), and the Journal’s referees for their comments and suggestions. I am particularly indebted to Matthew Weinzierl, whose generous feedback helped me shape the argument more clearly at several junctures.
Emmanuel Voyiakis is a Professor at LSE Law School, UK. He is the author of Private Law and the Value of Choice (Hart Publishing, 2017). His current scholarship focuses on the variety and significance of benefit-based arguments as grounds of moral and political obligation. URL: https://www.lse.ac.uk/law/people/academic-staff/emmanuel-voyiakis
Open access
