This paper proposes a new functional definition of predistribution that is practically useful and unifies different predistributive policies. Despite its growing popularity in political science, economics and philosophy,Footnote 1 predistribution appears inherently elusive.Footnote 2 The source of confusion is the prefix ‘pre’, which denotes priority or prevention. It hints at a connection with the intuitively attractive idea that prevention is better than cure. John Rawls seems to appeal to this idea when he presents his vision of the just socio-economic institutions as having a function of pre-emptively dispersing material and human capitals (Rawls Reference Rawls, Kelly and Cambridge2001: 139). Yet, he remains notoriously vague about this. Some theorists claim that no principled reason exists for supporting the claim that prevention is better than cure in matters of distributive justice (Halliday and Thrasher Reference Halliday and Thrasher2020: 144–7).Footnote 3 Predistribution capitalizes on and substantiates Rawls’ underdeveloped thesis that prevention is better than cure.
So, what is predistribution? It’s not easy to find a clear answer. Predistribution allegedly shifts the focus of distributive policy from ex post distribution to ex ante distribution. But this explanation just invites another question: what is ex ante distribution? A more specific stipulation identifies predistribution as policies for fairer distributions of pre-tax income and of other economic powers and opportunities as opposed to the targeted tax-and-transfers of income. According to this view, the major components of predistribution include the expansion of public investments and universal public services as well as equitable regulations of the market (Hacker Reference Hacker2011). But even this specific exposition doesn’t pass muster.
An important challenge comes from Martin O’Neill. Despite having previously strongly advocated predistribution, O’Neill’s 2020 paper thoroughly criticizes predistribution understood as the preventive distributive policy that is distinct from redistribution. By defending the functional definition, I hope to rescue the account of predistribution as policies of preventive, ex ante distribution.
The central tasks of this paper are to explain the functional account of predistribution and to explore its justificatory grounds. Section 1 sets the scene by reviewing O’Neill’s criticisms and laying out my functional account of predistribution. Section 2 presents some notions of priority that help to distinguish predistribution from redistribution. Section 3 introduces three types of predistributive policies with different preventive functions. Section 4 considers reasons why policies of predistribution are desirable, with an emphasis on the underappreciated, instrumental reasons.
1. Definitional Debate: An Overview
1.1 O’Neill’s critique
Jacob Hacker kick-started the debate by proposing predistribution as a set of realistically feasible policies that address inequality at the level of pre-tax income.Footnote 4 Martin O’Neill initially advocated the idea but ultimately takes issue with its philosophical looseness. Let’s review O’Neill’s criticisms, which I will fully examine later.
The first criticism concerns the idea that predistribution precedes taxation.Footnote 5 Here, priority (or precedence) can be temporal or conceptual. Starting with the temporal priority, O’Neill argues that labelling some activities of the market ‘before’ and others ‘after’ is confusing and unsustainable (O’Neill Reference O’Neill2020: 70). Private economic activities and the government’s policies are constantly in motion, and both are intertwined in the whole process. O’Neill also rejects the conceptual priority. According to him, claiming that some distributive policies are conceptually prior to the tax-and-transfer policies involves an unfounded assumption. This assumption essentially commits the same error as ‘everyday libertarianism’, i.e. a mistaken belief that economic activities (can) exist prior to the government meddling with them (Murphy and Nagel Reference Murphy and Nagel2002).
Second, O’Neill criticizes the characterization of predistribution as a type of policy that shapes the market itself as opposed to ‘simply taxing and providing benefits’ (Hacker et al. Reference Hacker, Jackson and O’Neill2013). For O’Neill, this understanding is misleading because the market-shaping policy includes tax-and-transfer policies rather than being distinct from them (O’Neill Reference O’Neill2020: 76). Contrasting predistribution and redistribution as ‘distinct type[s] of policy’ would be conceptually and normatively arbitrary (O’Neill Reference O’Neill2020: 84).Footnote 6
1.2 Teleological versus functional account
These criticisms lead O’Neill to the teleological account of predistribution. He defines predistribution according to its aims, which are ‘reducing objectionable inequalities of power within market relationships, and giving individuals a secure standing outside of the market transactions’ (O’Neill Reference O’Neill2020: 85).Footnote 7 For O’Neill, predistribution is defined solely by the social egalitarian concerns about status and power in economic relations that go beyond narrowly distributive considerations (O’Neill Reference O’Neill2020: 88).
My initial dissatisfaction with this teleological account is that it does not offer a clear typology of predistribution and redistribution.Footnote 8 According to this account, social egalitarian aims define predistribution such that any policy supported or justified by social egalitarianism would be considered predistribution. I find this unhelpfully vague. Social egalitarianism concerns diverse types of social relations and seems insufficiently specific to determine what counts as predistribution. O’Neill explains the social egalitarian aims of predistribution as concerning equality within and outside the market. But this formulation isn’t too clear either, not least because ‘within and outside the market’ seems to simply mean just about everything. I believe the better way of stating the aims of predistribution is: changing the rules of the market and its background conditions that distribute opportunities and economic powers so that they are fairer and more equitable. I assume this is in line with what O’Neill has in mind too.
Additionally, and more fundamentally, I suspect that aims alone are often insufficiently determinate to provide stable typologies for policies, since a policy could be supported by diverse, distant aims.Footnote 9 If predistribution is solely defined by social egalitarian aims, it would include policies that concern matters that are not typically ‘distributive’ or ‘economic’, such as how parents treat their children or how faith-based communities are run, because social egalitarianism concerns the matters that are not narrowly ‘distributive’ or ‘economic’.Footnote 10 Surely, not all social egalitarian policies are predistribution? But that seems to be the logical conclusion of the teleological account. Perhaps such a view is still philosophically coherent, but it is too expansive and not very useful for practical policy discussions. To give a proper focus and stability to the concept, you need a functional typology. That is, to specify what predistribution is, you need to reference both its aims and functional features – i.e. policy mechanisms or instruments. A good account of predistribution should include, in addition to the aims, some functional features as its demarcation.
In that spirit, I propose an alternative, functional account of redistribution and predistribution. It focuses on the necessity of referencing the distribution of pre-tax income and the targeting of specific income groups:Footnote 11
Redistribution: a policy that shapes the distribution of income, wealth or other economic powers through directly providing cash or in-kind benefits to people of a specific income group. The policy’s constitutive features include explicit and essential reference to the distribution of pre-tax income.
Predistribution: a policy that shapes the distribution of income, wealth or other economic powers without directly providing cash or in-kind benefits to people of a specific income group. The policy’s constitutive features don’t include explicit and essential reference to the distribution of pre-tax income.
Two points of clarification are in order. First, the distribution of pre-tax income is used roughly synonymously with market outcomes in the sense of how the market process ends up rewarding different individuals. And, just to clarify, here I am talking about the pre-tax income and market outcomes of the relevant (i.e. current) cycle and not earlier periods. Second, I specify the relevant way of referencing the distribution of pre-tax income as explicit and essential. Such referencing is a constitutive feature of the policy in question, and the policy doesn’t make sense without it. For the typical policies of redistribution such as income support, reference to the distribution of pre-tax income is explicit and essential in this sense. This way of referencing stands in contrast to referencing for the purpose of funding or adjusting policy details based on who can pay how much, etc. (call it ‘referencing for fine-tuning’). The latter is trivially ubiquitous since basically all government policies need funding by taxation.Footnote 12 Referencing for fine-tuning is compatible with predistribution.
Key characteristics and implications of this formulation of predistribution, which I discuss in the following sections, can be cashed out as three points of contrast between my account and O’Neill’s. First, I defend as sustainable and useful the distinction of ex ante versus ex post distribution, which O’Neill rejects. Not only does the abandoning of the distinction fail to attain the alleged conceptual clarity, but it comes at the cost of the vagueness regarding what policies predistribution consists of. I aim to present a more practically useful definition of predistribution that retains the notion of ex ante distribution. Second, by tying predistribution exclusively to social egalitarianism, O’Neill leaves out strains of thought that can give normative bases for different functions of prevention relevant to distributive justice. My account of predistribution accommodates the diversity of what has been discussed as predistribution while clarifying the different functions. The third point of contention is whether priority (i.e. precedence) to taxation of income is a meaningful feature of predistribution. I hope to show there is a strategic value in focusing on equality at the level of pre-tax income, especially in the face of persistent ‘everyday libertarian’ attitudes. I argue that such instrumental thinking is not something the egalitarians should shy away from.
2. Defending Priority
This section explores different conceptions of priority and defends the distinction of ex ante versus ex post distributions. I offer a meaningful and robust notion of conceptual priority that I call procedural priority, which underpins how I understand predistribution.
2.1 Temporal priority
Let’s first discuss temporal priority.Footnote 13 O’Neill rejects thinking about the economy in terms of discrete periods: such thinking is ‘outlandish’ and ‘naïve’ because ‘the economy is an ongoing dynamic system, whereby the end of one period is always the beginning of another and where, as one might put it, one man’s ex post is another man’s ex ante’ (O’Neill Reference O’Neill, O’Neill and Williamson2012: 90).Footnote 14 On one level, O’Neill is obviously right. It is often inaccurate to describe some public policy as temporally prior to economic activities in a literal sense. The market process does not follow a strict time sequence, nor is it divided into discrete periods. Few people start their business at the beginning of a year, some taxes are paid as you earn, regulations and spending by the government may take place at any time, and so on. But that’s almost trivial.
The question is: Is it also problematic to apply frameworks of temporal sequence and periods loosely or figuratively? To what extent and in what way are such frameworks unsuited for understanding and ordering social phenomena? It seems absurd to say all such usage of a figurative temporal framework is unwarranted, since businesses and individuals operate according to such a framework of periodically divided time, like financial year. The periodic frameworks typically apply only to some aspects of our economic activities (e.g. bookkeeping, annual business planning and taxation). Nonetheless, they are real in the sense that they shape our socio-economic activities. In particular, income tax can be properly assessed only at the end of the financial year in question. Abandoning such rough periodized frameworks seems to make our grasp of economic phenomena more confusing rather than less so.
Moreover, few concepts used in policymaking or social science are rigid and binary. For example, policies about job security have a real effect on what we consider to be an issue of health (known as social determinants of health). But we still understand a policy for job security to be primarily about employment and only derivatively about health. For an example of periodic frames, think of electoral cycles. A cycle starts with candidates seeking a party nomination or otherwise nominating themselves, followed by the campaign phase when the candidates make a pledge and the electorate form their views, ending with ballots being cast and counted. This framework is rough and not completely accurate. Some politicians start preparing way in advance, some voters come already fully convinced which party to vote for even before the candidates are announced, other voters lobby candidates to influence the outcome prior to the voting, and sometimes we have a good idea of the outcome of the election before the first ballot is cast, and so on. Despite many such details not fitting the framework, the rough framework of the electoral cycle is reasonably useful for making sense of political phenomena. Abandoning the framework is unlikely to contribute to a clearer understanding of electoral politics. The same seems true of the periodized framework of economy.
Could the circularity of time sequence be a problem? Even if we can understand the economy according to a time sequence where some events follow others, the whole process is circular so that what looks like redistribution now can be seen as predistribution from a viewpoint of some later time, and vice versa. This observation, while correct, does not seem detrimental to a periodized and sequential understanding of social phenomena. Again, think of electoral cycles. Typically, how a candidate fares in one cycle of elections has some influence on how she fares in the next cycle. In a sense, the outcome of one election is also a campaign for the next election. But this fact does not mean we should abandon the understanding that a campaign phase is followed by a separate phase of voting, and one discrete cycle is followed by another. Again, we can say the same about the economy and economic policies.
So, we can meaningfully say, in some cases at least, an economic event or policy is prior to another in a loose, figurative sense of temporal priority.Footnote 15 Regarding distributive policies, we can understand some tax-and-transfer measures (e.g. income support) as primarily posterior to market outcomes, and other economic policies (e.g. corporate governance regulations) as primarily prior to market outcomes, in this loose sense. We can make informative policy categories based on such relative differences, which help distinguish policies based on their main features. There seems to be nothing confusing about that.
Moreover, if O’Neill rejects the figurative notion of temporal priority as being conceptually too loose, he doesn’t seem to live up to the same standard himself. Recall O’Neill’s stipulation of the aims of predistribution as ‘reducing objectionable inequalities of power within market relationships, and giving individuals a secure standing outside of the market transactions’ (O’Neill Reference O’Neill2020: 85).Footnote 16 He seems to be replacing a loose temporal notion of priority with similarly loose locational notions of outside and inside the market. For, the inside-outside distinction cannot be precise either. If ‘outside of the market’ means something like the public or political sphere, it overlaps significantly with the market so that political decisions inseparably interact with the economy. Alternatively, if what is meant is the intimate or private sphere, it is also difficult to think of it as completely separated from the market and (paid and unpaid) labour. My point is that useful distinctions can sometimes be vague and imprecise around the edges.
It seems most reasonable to understand that O’Neill’s criticism is about the strictly literal temporal priority. What can be established is that predistribution cannot be literally temporally prior to market outcomes and income taxation. If so, there is no case against the loose notion of temporal priority that I am interested in. At the very least, there is room for employing the loose temporal priority for picking out an important feature of predistributive policies.
2.2 Conceptual priority
Let’s turn to investigations about whether conceptual priority helps us make sense of predistribution. I start with O’Neill, who denies the conceptual priority of predistribution relative to taxation. He invokes Murphy and Nagel’s argument, which says: It is wrong to assume that your pre-tax income is something prior to and untainted by the government activities, because people’s incomes are shaped by the background institutions and infrastructure, which are funded by taxation. If you assume away the functions of government, your pre-tax income would be something entirely different from your current pre-tax income. As O’Neill puts it, taxation is ‘a constitutive part’ of the property system (O’Neill Reference O’Neill, O’Neill and Williamson2012: 71–2, cf. Murphy and Nagel Reference Murphy and Nagel2002: 33–7). I agree.
But O’Neill may be missing two important claims that Murphy and Nagel make about the logical priority of the government and the provision of extended public goods. Let’s focus here on logical priority (I will discuss the extended public goods in section 3). Murphy and Nagel argue that ‘[t]he logical order of priority between taxes and property rights is the reverse of that assumed by libertarianism’ (Murphy and Nagel Reference Murphy and Nagel2002: 33). So, the government and taxation are logically prior to property rights and market outcomes rather than, as O’Neill suggests, the former being merely ‘constitutive’ of the latter. Though they may often overlap, the two claims are still different. When one thing is constitutive of another, the two are inseparable, with no regard for an order between them. In contrast, when you say some activity or event is logically prior to something else, it obviously involves a claim of order where one precedes the other. Murphy and Nagel’s account involves the claim about the conceptual priority of taxation: It makes logical sense to say taxation and various socio-economic regulations by the government precede market outcomes. Here are two preliminary observations on this. First, you could say, in one sense, all governmental interventions are logically prior to market outcomes. Second, there may be logical orders among various governmental policies such that some policies are still prior to others. So, pace O’Neill, Murphy and Nagel may be rather sympathetic to the conceptual priority of predistribution.
So, O’Neill appears off the mark here, but we still need to examine whether logical priority is a sensible idea and especially whether it is sufficiently robust against objections from the pro-market theorists. Murphy and Nagel don’t seem to define logical priority, but I believe the following roughly captures what they mean:
Logical Priority: A is prior to B when reference to (or presupposition of) A is necessary for B, but not vice versa.Footnote 17
But this is still a bit ambiguous. It can be further specified at least in two ways:
Justificatory Priority: A is prior to B when reference to (or presupposition of) A is necessary to justify B, but not vice versa.
Ontological Priority: A is prior to B when the existence (or presupposition) of A is necessary for B to exist, but not vice versa.
Murphy and Nagel apparently conflate the two when they say the government and the holistic conception of justice are logically prior to private property and other market outcomes. However, at least initially, that claim looks plausible only for the priority in the order of justificatory logic. Any outcome can be considered just only if the society is organized in a just way. In this sense of justificatory priority, you can say even a pure tax-and-transfer policy has a clear logical priority to the market outcomes. The same cannot be said with equal plausibility with regard to the priority in the order of ontological logic. For, pro-market theorists would say the market requires only the minimal state. Some institutions and policies are ontologically prior to the market, but others are not. In particular, pure tax-and-transfer policies seem unnecessary for the market to exist.
However, this response is also unsatisfactory. For, presumably, the pro-market theorists do not want the bare minimal state with truly minimum function for the market to exist but an optimally minimal state, i.e. some appropriately extensive government that implements anti-trust laws, intellectual property laws, bankruptcy laws, monetary policies, etc.Footnote 18 The pro-market theorists also need to talk about justification of such governmental interventions that go beyond the truly minimal state. With the exception of the die-hard libertarians, even pro-market theorists must be committed to some moralized notion of a fair market. This, in turn, means that most pro-market theorists need some account of justificatory priority rather than simply sticking to the account of ontological priority.Footnote 19 Indeed, Murphy and Nagel may think that justificatory priority subsumes ontological priority (‘only just social institutions and practices are allowed to exist’), which would explain why they conflate the two. I am sympathetic to this supposed view of Murphy and Nagel, but there is room for reasonable disagreement here. The pro-market theorists could say: the ontological priority for the free and fair market requires only a purely procedural account of justice, which is only concerned with justice of process such that whatever outcomes result from the just procedure are also just. The logic of such an account is justificatory in nature but does not directly determine what outcomes are just. A pure procedural account of justice does not provide a reason for intervention in outcomes beyond what is necessary for making the procedure just. This position, which is not implausible, gives us an account of logical priority like so:
Procedural Priority: A is prior to B when reference to (or presupposition of) A is necessary for B to exist as an outcome or part of a fair procedure – and to be justified as such – but not vice versa.
Think of a fair game of gambling as an example of fair procedure. For a gamble to be fair, information about the rules of the play, rates, odds, etc. (or meta-rules about how they get decided) needs to be available to participants. It also requires some measures of reviewing the process, reporting violations, and making an appeal about penalties and other decisions taken by the house. It may also require, as in the game of poker, regulations about the amount of funds you start with or how you use them. A proper arrangement of these features is necessary for the procedure of the gamble to be fair and, accordingly, for the outcomes of the gamble to be fair. In this sense, these features are prior to the outcomes of the gamble. Additionally, some relative priority can be discerned among different features of the procedure: the existence of the rules must precede their publication, followed by measures of reviewing the process, reporting the violation and making an appeal, for example. Procedural priority capitalizes on the loose temporal priority: the fair procedure must be in place before the gamble takes place, even if some features of the procedure kick in along with or even after a game.
A fair procedure could be reasonably thick and substantive. Applied to distributive policies, some of the policies that may intuitively appear redistributive can be included in the procedure, to the extent that the policies are conceived as a part of the procedure.Footnote 20 For, again, procedural priority is only figuratively temporal.
Yet, with pure procedural justice, justifying redistributive policies that aim towards the justice of outcomes is less straightforward and faces greater difficulties compared with a more outcome-oriented or ‘telic’ account of justice.Footnote 21 In other words, for a purely procedural account of justice, it is harder to say that the redistributive policies are logically prior to fair market outcomes. Here, I do not examine what such a view of procedural justice looks like or whether it is ultimately sustainable.Footnote 22 In this paper, I simply accept this account of logical priority as the bar that is harder to clear for establishing that the government’s intervention in the market is logically prior to market outcomes. Recall that with regard to the (full) justificatory priority, we can say that all government activities are logically prior to market outcomes. Regarding the ontological priority, the radical libertarian may claim that only the bare minimal state is logically prior to market outcomes. The logic of pure procedural justice sits between the two and provides an account of logical priority with prima facie plausibility in relation to the government, distributive policies and the market. This account of logical priority is an appropriate and normatively interesting one to focus on because it seems to be the best case the (moderate) pro-market theorists can put forward in their favour.
Procedural priority underpins my functional account of predistribution. Recall,
Predistribution: a policy that shapes the distribution of income, wealth or other economic powers without directly providing cash or in-kind benefits to people of a specific income group. The policy’s constitutive features don’t include explicit and essential reference to the distribution of pre-tax income.
Predistribution does not make an explicit and essential reference to the distribution of pre-tax income. This functional feature reflects the aims of predistribution, which, recall, are: changing the rules and background conditions of the market that distribute various opportunities and economic powers so that they are fairer and more equitable.Footnote 23 Since you are focusing on the rules and the background conditions of the market, income distribution, which emerges after the market process, is not a central concern for predistribution. The typical mechanisms for referencing the distribution of pre-tax income are income testing and income targeting. As such, predistribution does not typically have income targeting as its constitutive feature. If you test the income of potential recipients before providing a benefit, the benefit will be too late to be predistributive. Instead, policies of predistribution tend to be universalistic.Footnote 24
Let’s take the public provisions of essential goods and services as an example. Such provisions are for preparing and supporting individuals for productive activities in the market.Footnote 25 Determining roughly the right kind and level of provisions of such goods and services does not require information about how the market process rewards different individuals differently. General information about society, excluding the income distribution, is sufficient for the purpose.Footnote 26 The policies can be conceived prior to and without reference to the distribution of pre-tax income.Footnote 27
The contrast here is with respect to redistribution, i.e. targeted tax-and-transfer policies such as income support. Reference to the distribution of pre-tax income is an integral part of such policies and cannot be just a matter of fine-tuning. Without information about the distribution of pre-tax income, the targeted tax-and-transfer policies cannot be conceived or designed because correcting market outcomes is exactly their point. Redistribution is an intervention that realizes (or approximates) a desired pattern of income distribution. To design a policy of redistribution even in a ballpark, you need to identify who needs additional resources for how much. Knowing the preceding distribution of pre-tax income is essential to the policy.Footnote 28
Predistribution is prior to the market outcomes in the sense of procedural priority, while redistribution isn’t. It means that predistribution is consistent with the idea of an appropriately extensive minimal state that most pro-market theorists are presumably committed to.
2.3 Objections
Let me address three objections to my definition of predistribution.
2.3.1 Not aiming at pre-tax income equality?
First, since predistribution aims to reduce the inequality of people’s pre-tax incomes, isn’t it contradictory not to reference the distribution of pre-tax income? In fact, in my view, equality of pre-tax income is not the direct aim of predistribution. Rather, the aim is to make the rules of the market and its background conditions meet the broad requirements of justice.Footnote 29 For example, the primary aims of capping executive salaries are to make corporate governance more aligned with the interests of ordinary workers and to reduce domination of employees by employers. Reducing the inequality of pre-tax income matters derivatively. Minimizing the inequality of pre-tax income may appear to be the central function for some predistributive policies such as minimum wage regulation. However, even minimum wage does not directly specify the distribution of pre-tax income. For one thing, access to income is conditional on people’s being employed and working for certain hours; even for people with the standard working hours, there are infinite variations to their income distribution because the minimum wage only determines the lower limit. More generally, when predistributive policies reduce the inequality of pre-tax income, they do so by targeting factors that determine people’s pre-tax income (such as access to basic services, regulations of the rules of the market) rather than pre-tax income per se.
2.3.2 Reference for assessment
Second, referencing the distribution of pre-tax income may seem essential for the assessment of the predistributive policies’ effectiveness.Footnote 30 Without knowing who is earning how much through work, you cannot tell if a policy promotes equality effectively. However, I don’t believe referencing the distribution of pre-tax income is necessary for assessing whether the core aims of predistributive policies are achieved. Let’s take the minimum wage as a hard case.Footnote 31 The core aims of minimum wages are: preventing abuse of employers’ wage-setting power to force employees to work for less than a decent wage; and guaranteeing people broad access to a decent income. The first aim can be assessed by checking the compliance of employers so that we need not reference the distribution of pre-tax incomes. The assessment of the second aim may appear to require knowledge of the distribution of pre-tax incomes. While we can (roughly) decide what level of minimum wage is in line with the decent minimum in society, that level could be too high for some employers, and, so the criticism goes, the minimum wage regulation may lead to greater unemployment, which defeats the aim of guaranteeing access to a decent income. However, for checking this potential negative effect of a minimum wage regulation, information about the distribution of pre-tax incomes is not essential (though it is useful); you only need information on the unemployment rate. This response is admittedly inconclusive, but I hope to have cast some doubt on the initial plausibility of the objection.Footnote 32
2.3.3 Too broad a category?
Finally, the reader may worry that almost all of what we know traditionally as redistribution or welfare policies become predistribution, excluding only the pure tax-and-transfer measures like income support. This is probably true, but it isn’t a problem unless we end up with an unhelpfully expansive category of policies. The next section introduces the categorization of different types of predistribution, which should allay this concern. Also, one of my criticisms of O’Neill’s definition of predistribution was that it cannot exclude policies that have little direct connection to the distribution of income, wealth or other economic powers. Though my definition is still broad, its scope is limited to the realm of the market and social cooperation, which concerns the production of economic and social values.
In this section, I have argued that both the temporal and conceptual priority are sustainable and useful as characterizations of predistribution. Temporal priority can be sustained if interpreted as a loose tendency to precede. The temporal priority underlies the conceptual priority in the sense of logical priority, which is central to my project of reframing predistribution. I have identified procedural priority as the appropriate notion of logical priority for characterizing some distributive policies as prior to market outcomes. Predistribution is ‘prior’ to market outcomes in this sense of conceptual priority.
3. Type-Distinctions
Predistribution includes diverse policies (see Appendix). They can be helpfully categorized into three types based on specific aims and functions of predistributive policies.
3.1 Three policy types
3.1.1 Pre-market distribution
The first policy type is ‘pre-market distribution’. It provides people with access to essential goods and services so that individuals can use them before they enter the market or other avenues of social cooperation. Extended public goods, which we flagged in the previous section, are a good example of this policy type. Murphy and Nagel suggest that they form a distinct class of distributive policy that is justified ‘not for purposes of redistribution’ (Murphy and Nagel Reference Murphy and Nagel2002: 86). In addition to the classic public goods,
there may be important aesthetic, social, and cultural goods that cannot be supplied privately. If we can ensure a decent level of education for all, independently of their ability to pay, the result will be a society that is much better for everyone to live in, and economically better for almost everyone, than a society with high levels of illiteracy and innumeracy. … [T]here is also a case, based on this type of efficiency consideration, for traditional social welfare policies guaranteeing a decent minimum standard of living, or decent minimum earnings, for everyone in the society. (Murphy and Nagel Reference Murphy and Nagel2002: 86–7)Footnote 33
The provisions of extended public goods are distributive policies that are not ‘redistributive in the usual sense’ (Murphy and Nagel Reference Murphy and Nagel2002: 87). Even (everyday) libertarians who think pre-tax income has a morally significant status may have pro tanto reason to support such public provisions. For one thing, they are like public goods and are more efficiently provided publicly than privately.Footnote 34 For another, wide access to such goods and services is beneficial to society at large. Equal access to things like education and healthcare is beneficial to the entire society because it puts everyone in a better position to engage in a cooperative activity, before they go into the market as it were. That is, much of the extended public goods should be available to people as a precondition for the fair and free market. More generally, pre-market distributions are measures that prepare people for social cooperation through access to human and financial capitals, including the provision of healthcare and support to care work.
3.1.2 In-market distribution
Another group of predistributive policies involve regulations and rules of the market that reduce inequality of power or domination over parties in disadvantaged positions.Footnote 35 Call it ‘in-market distribution’.Footnote 36 It includes regulations concerning minimum and maximum wages and corporate governance regulations to provide a meaningful representation of workers’ interests. Pre-market and in-market distributions are both (loosely) temporally prior to redistribution, which is post-market. But pre-market distribution is still prior to in-market distribution and the market process, while in-market distribution tends to be more intertwined with and contemporaneous with the market process.
3.1.3 Collective rent sharing
The third type of predistribution is what I call ‘collective rent sharing’. It is about fairly distributing valuable resources that legitimately belong to no particular group or individual. Such resources may include land, natural resources, airwaves, and material and immaterial wealth inherited from prior generations. Stuart White emphasizes how ‘[s]ome resources exist prior to the labour of current members of society’ (White Reference White, Lever and Poama2019: 362). Further, Gavin Kerr points out that values that are not produced by the labour of particular person(s) are economic rents and should be regarded as prior to the economic activity (Kerr Reference Kerr2017: 231). In somewhat different ways, both White and Kerr point to the economic value that exists prior to – or non-attributable to – economic activities of anyone alive today. For example, even though an individual human activity can add to the value of land or natural resources, their original value precedes economic activities of any person alive today (or any human activity whatsoever). Also, some creation (or appreciation) of value occurs only as a result of a large-scale change of circumstances so that any action of an agent cannot be considered as its cause. No one can rightfully claim the ownership of those kinds of economic value, i.e. rents.Footnote 37 Rents in this sense are procedurally prior – and typically ontologically prior too – to an economic activity of an agent who happens to own them. Rents are morally problematic because they lack justification – from the viewpoint of distributive fairness, desert or even efficiency. Collective rent sharing makes it right by distributing the economic value of rents by hypothetically going back to the (counterfactual) state where no member of the current generation laid claim on them. Here, the division of the economic value should follow a procedure that is distributionally fair, which typically means equal division for all.Footnote 38 To own the economic value of rents without such a fair procedure would be rent seeking. Collective rent sharing aims to prevent unfair appropriations of rent and accompanying inequality and exclusion of some people from access to common resources.
Collective rent sharing fits with my formulation of predistribution since policies of collective rent sharing need not reference the distribution of pre-tax income. O’Neill leaves out this group of policies, perhaps because it doesn’t quite fit with the social egalitarian aim of promoting status equality. The teleological account does not allow predistribution to have aims that do not conform to social egalitarianism. The exclusion of this type of policy may also have to do with the fact that these policies tend to be associated with the idea of predistribution formulated by the economist James Robertson. Either way, including this group of policies as a type of predistribution makes sense since they are often discussed as such, and affinity exists between the Hackerite idea and the Robertsonian idea. The Robertsonian predistribution too is about realizing a fairer distribution of resources and other economic powers that form the background of the ordinary market process.Footnote 39
3.2 Objection
Let’s consider an objection that the type-distinctions are arbitrary and incoherent.
3.2.1 Arbitrariness
O’Neill calls out arbitrariness in Hacker’s formulation of predistribution as ‘market-shaping’ policies that are distinct from pure tax-and-transfer policies (O’Neill Reference O’Neill2020: 80–1). O’Neill substantiates this objection by examining UBI and top-rate income tax to show that even the quintessential tax-and-transfer policies have the market-shaping, predistributive effect (O’Neill Reference O’Neill2020: 81–4).Footnote 40 Since UBI is universal and unconditional, it has a predistributive effect of providing people with some financial security outside of (or, as I would say, loosely prior to) the market, which may also translate to greater bargaining power within the labour market. Also, increasing the top marginal rates of income tax has a predistributive effect. It weakens high-income earners’ incentive to seek ever higher salaries because income in the top tax bracket will be heavily taxed anyway. Such a change in the incentive structure will likely reduce the inequality of pre-tax incomes. The type-distinction of market-shaping versus tax-and-transfer policies seems to lack consistency and philosophical substance. This objection is relevant to my project to the extent that I defend a similar distinction to Hacker’s.
I have three responses. My first, general response is: the fact that typically redistributive policies have some predistributive effect is not necessarily detrimental to the claim that predistribution and redistribution are two different types of policies; it is detrimental only if you are rigidly puritan about conceptual distinctions. We can simply say some policies are primarily predistributive and others are primarily redistributive. As long as we can make sense of the difference in emphasis between the different groups of policies and if the distinction is practically useful, the distinction doesn’t have to collapse.
You can usually tell whether a policy is primarily predistributive or redistributive by looking at the policy’s first-round effect.Footnote 41 If the policy’s immediate function is to provide cash or in-kind benefits to people of a certain income group, it is primarily redistributive, even if such provision comes to have the downstream, predistributive effect, i.e. indirectly influencing the pre-tax distribution of income, wealth or other economic powers. In contrast, the first-round effect of a predistributive policy is to affect the background conditions and rules of the market without directly providing cash or in-kind benefits to a group of people selected on the basis of their income level, even if it involves taxation of people’s income at some point.
Second, as shown in the previous section, my definition of predistribution is based on the idea of procedural priority. It illustrates the basic sense in which predistribution precedes the market (or prevents inequalities, poverty, etc.). As a fair procedure, predistribution should ideally be in place before the market process, even though some policies take effect through or even after the market process (I will elaborate on this shortly). As such, the definition is philosophically and normatively meaningful.
Third, O’Neill’s examples don’t present a puzzle to my classification. Given my definition of predistribution based on reference to the distribution of pre-tax income (and income targeting), neither UBI nor top-rate income tax qualifies as redistribution in a straightforward manner. UBI is a policy of predistribution because it distributes cash to everyone (as opposed to a specific income group), and designing it doesn’t require information about the distribution of pre-tax income. Top-rate income tax is less clear. We need to be careful because what is at issue is not distribution of taxed resources but taxation only. Obviously, taxation itself does not involve cash provision to anyone. Also, if policymakers were to set the top-rate income tax purely for disincentivizing excessively high incomes, it would be possible to design the tax rate without referencing the distribution of pre-tax income. More importantly, however, the primary aim of income taxation is usually to raise government revenue.Footnote 42 So, if we consider taxation in isolation from distributive aims, it is primarily a policy of revenue raising rather than predistribution or redistribution.Footnote 43 Presumably, the effect on pre-tax income distribution will always be present regardless of the intention of income taxation. It seems most straightforward to say that income tax is a policy of revenue raising, with some predistributive effect.Footnote 44
3.2.2 Coherence of the three types
A similar objection may be directed to the sub-categories of predistribution: are the three types of predistribution sufficiently meaningful and coherent? By picking out specific predistributive functions, each of the three types represents somewhat different meanings of prevention and priority. We can distinguish three relevant senses of prevention/priority: institutional, personal and value-generative.
First, as suggested earlier, all types of predistribution are concerned with the fairness of the market as a procedure for social cooperation. This makes them procedurally prior to the market in the sense that they form a part of fair procedure that needs to be in place for the market to be fair. As such, predistribution aims to prevent morally objectionable inequality, poverty and exploitation by setting up fair institutions of the market. This is the institutional sense in which predistribution is prior/preventive.
Second, pre-market distribution is prior/preventive in the second, personal sense, namely, in the sense of directly equipping individuals with access to resources for partaking in social cooperation in the market or other avenues of social cooperation on an equal footing with others.Footnote 45 From the perspective of an individual person, you can tell whether public provisions are ‘in time’ when they need them, which is before they take part in social cooperation.Footnote 46
Third, collective rent sharing is marked by value-generative priority as it concerns economic value that is generated before your economic activity (or the economic activities of anyone alive today). Temporally speaking, this priority may be long-term or short-term.Footnote 47 Either way, the economic value is generated prior to your economic activity so that you cannot legitimately claim to appropriate it. This is the third sense in which predistribution is prior/preventive.
You might wonder if the distinctions are really coherent. In particular, some policies of collective rent sharing, such as dividends from natural resources and inherited assets, could also be considered pre-market distribution. This overlap is contingent, and one category usually fits better than others. Though hybrids may exist, the type-distinctions are worth making because, as illustrated with different meanings of priority, pre-market distribution and collective rent sharing have different rationales. Collective rent sharing concerns how the value is generated and whom its rightful ownership belongs to, while pre-market distribution is primarily about equal (or sufficient) opportunities. The distinctions are meaningful and coherent enough to be useful.
4. What’s Good About Prevention?
I have proposed and elaborated on my functional account of predistribution and the underlying idea of procedural priority. What does my account of predistribution mean for its desirability? Instead of a full-fledged justification, this section mainly focuses on salient implications of my functional account: the availability of pluralistic grounds for fundamental justification and relevance of the instrumental justifications. Let’s consider fundamental and instrumentalFootnote 48 cases for predistribution in turn.
4.1 Fundamental justifications
When it comes to poverty and inequality, why are preventions better than curing? The broadly social egalitarian reasons seem important here. One is respect for persons. There is something disrespectful about reducing justice entirely to post facto compensations for disadvantages. I cannot erase injustice simply by providing you with compensation that is equal to the monetary value of the disadvantage you suffer – as if it is a matter of balancing an accounting ledger. Social egalitarians famously provide these lines of criticism against luck egalitarianism (Wolff Reference Wolff1998; Anderson Reference Anderson1999). Relatedly, when harm is cured post facto, the damage is already done, potentially unleashing various downstream effects.
Beyond these reasons, there may be broader grounds for the desirability of preventive economic policies. Let me offer a quick review of possible sources of normative justification for predistribution. I have proposed three policy types of predistribution. Each of them is congenial to, and potentially justified by, different normative values. First, pre-market distribution aims to prevent people from falling into poverty or exploitation by providing individuals with fair access to essential resources, including care, education and training. This type of predistribution has a strong intuitive link to the values of enabling or nurturing human capacities as well as fair equality of opportunity. Second, in-market distribution sets up institutions of the market so that they avoid problematic concentration of economic powers as well as inequality and poverty. This function has a strong affinity with theories of procedural justice, especially of the Rawlsian kind.Footnote 49 Also relevant to in-market distribution are theories of structural injustice that criticize how the unjust social structure perpetuates the oppression of the disadvantaged (Young Reference Young2011). Third, collective rent sharing aims to fairly distribute what legitimately belongs to no particular group or individual. This group of policies is likely to find support in the ideas of democratic ownership and in the unfairness of rent appropriation. While I have no space to offer an argument, these are some promising grounds for fundamental justifications of predistribution.
Finally, consistency with the broadly libertarian notion of economic justice would be another fundamental reason for supporting predistribution.Footnote 50 The fact that predistribution is characterized by the notion of procedural priority gives the pro-market theorist a reason to be in favour of predistribution. Let’s consider executive pay capping and control on the financial market as hard cases. The moderate pro-market theorists cannot categorically or easily reject them, though they are unlikely to fully endorse them either. Judgement about the desirability of the policies turns on their all-things-considered societal effects. To reject those policies, the pro-market theorists would need to explain why regulating the exorbitant executive pay is not desirable, while regulating pollution or exorbitant monopoly powers presumably is, and why regulating the financial market is counterproductive, while regulating the bankruptcy procedures or the intellectual property rights is beneficial. I suspect that the pro-market theorists who subscribe to the idea of procedural priority cannot categorically reject executive pay capping or controlling the financial market.Footnote 51 There is also a related yet conceptually separate question of how aggressive or generous you design a particular predistributive policy to be. On this question, the pro-market theorists are likely to disagree with the egalitarians. But this disagreement, though consequential, is not about the crush of conceptions of justice but about empirical facts about how cost-effective a policy is. As a matter of principled normative reason, the moderate libertarians should be favourably disposed to predistribution.Footnote 52 At the very least, predistribution is harder to reject than comparable measures of redistribution.
4.2 Instrumental justifications
There are instrumental (i.e. strategic) reasons for social democrats to opt for predistribution rather than redistribution, because predistribution is less likely to invite a right-wing backlash. According to Hacker, this is partially because some predistributive policies do not involve government spending and presumably also because predistributive measures tend to be less visible than typical redistributive policies. O’Neill expresses scepticism, arguing that if predistribution were to achieve substantive equality, it must go against the interests of the rich (O’Neill Reference O’Neill2020: 89). This scepticism seems only partially warranted so that the instrumental argument does not lose its force. While it is true that predistribution in its meaningfully egalitarian form would still involve substantive taxation (of income as well as assets), it is nonetheless plausible that predistribution is less likely to trigger resentment and backlash from the people on the right. There are three reasons for this. I will quickly review the first two and focus mainly on the third reason, which is about the psychology of how people evaluate distributive policies.
First, some policies of predistribution do have low requirements of taxation, especially on labour income. Two types of predistribution are relevant here. On one hand, in-market distribution is about changing the rules and structure of the market, for example, by regulating the floor and the ceiling for salaries and by improving the power balance between labour and employers. Such policies cost no more than a general enforcement of legal order. In a sense, they are in fact cheap, at least from the viewpoint of the taxpayer. On the other hand, collective rent sharing typically involves taxation of assets whose rent should be collectively shared. But such taxes are still different from the taxation of labour income. Both in-market distribution and collective rent sharing go against the economic interests of the super-rich; however, this group represents a very small portion of the population (the 1%, as is often called). Such policies are largely aligned with the economic interests of most citizens who might have an anti-tax instinct. Given these considerations, those policies are pro tanto more likely to gain democratic support.Footnote 53
Second, it is arguable that policies of predistribution contribute to efficiency in the sense of increasing gains from social cooperation or people’s welfare. More specifically, I focus on optimal allocation of resources and economies of scale as mechanisms for economic efficiency.Footnote 54 Starting with the allocation of resources, it will be more optimal if resources are distributed through various channels, including policies of pre-market and in-market distributions along with redistributive, tax-and-transfer measures, rather than using redistribution only. One reason, according to Zachary Liscow, is that due to the law of diminishing marginal return, the benefit of distributive policy tends to be better realized when it is pursued in many different channels than only through one type of policy.Footnote 55 Additionally, if you aim to realize the optimal allocation only through redistribution (i.e. tax-and-transfer), it will likely fall short of the optimal level due to the popular opposition to redistribution.Footnote 56 So, to promote the optimal allocation of resources in the real world, you’d better diversify the channels. Also, by minimizing the rent appropriation, policies of collective rent sharing may improve the liquidity of assets that might otherwise be held idly by a few hands.Footnote 57 Regarding economies of scale, policies of pre-market distribution achieve efficiency gains through public provisions of essential goods and services. Where people need relatively uniform goods and services, the government tends to be more efficient than a private actor.
Third, predistribution has an advantage over redistribution because it does not specifically target people’s pre-tax income, which people – albeit probably mistakenly – feel attached to. Debates about desirable distributive policies often get contaminated or hijacked by people’s excessive concern about protecting their pre-tax income. Predistribution is better at avoiding such contamination than redistribution is. The point is not that people’s attachment to their pre-tax income is legitimate but that it may be unavoidable in the current political climate of capitalist societies such as the UK and the USA. Given the well-known psychology of loss aversion (McCaffery and Baron Reference McCaffery and Baron2005; cf. LeBoeuf and Shafir Reference LeBoeuf, Shafir, Holyoak and Morrison2005) and given that people feel entitled or attached to their pre-tax income, there is a strategic case for not drawing people’s attention to pre-tax income as a tax base, at least to the extent that the circumvention is possible without compromising the substance of egalitarian policies. Philosophers tend to dismiss the merit of avoiding people’s apparently irrational reactions to tax-and-transfer policies. But people often hold tangled, conflicting preferences about distributive policies, and how you frame a policy may affect which preference they pay attention to and how rationally they deliberate. Well-documented evidence suggests that many people (1) support egalitarian policies other than redistribution (i.e. tax-and-transfer) more strongly than redistribution, (2) think economic inequality is too large while opposing (re)distribution, (3) support distributive policies including redistribution while opposing taxation or even (4) change their attitude towards the same policy simply based on whether it is called ‘redistribution’.Footnote 58 By virtue of not referencing the distribution of pre-tax income, policies of predistribution are more likely to be evaluated by people for their merits than redistributive tax-and-transfer measures that do target people’s pre-tax income.
The reader may worry that such strategic manoeuvring betrays the egalitarian aspiration of predistribution. Let me address two such worries. First, my view may look like acquiescence in small statism. That’s not my intention. In fact, predistribution is only marginally better than redistribution at deflecting the objection that it is costly. The objection of costliness is different from both the intuitive attachement to pre-tax income and the moral objection about the government’s intervention in market outcomes. Everyday libertarianism is a case of the latter, although the everyday libertarian may often raise objections about cost too. To review, everyday libertarianism places moral significance on people’s pre-tax income, based on the (unfounded) assumption ‘that the distribution of welfare produced by the market is presumptively just’ (Murphy and Nagel Reference Murphy and Nagel2002: 31). Everyday libertarians believe that ‘what we have earned belongs to us without qualification’ and any state intervention requires justification (Murphy and Nagel Reference Murphy and Nagel2002: 34–5). Because of how it is conceived, predistribution is better positioned to avoid that kind of resentment about taxation and transfer of income than redistribution is.Footnote 59 Predistribution faces one less objection, which appears persistent.
The second worry is that even if strategic manoeuvring succeeds in cutting back inequality, such manoeuvring may be problematically concessive to the unreasonable anti-egalitarian ethos of the rich (cf. Cohen Reference Cohen2008). The fear is that acknowledging everyday libertarian attitudes to taxation and trying to work around it means vindication of everyday libertarianism. I don’t believe it has to be so. For example, acknowledging that many people have unconscious, implicit biases towards certain race or gender that are hard to get rid of does not imply we give in to them;Footnote 60 we can and should seek ways to contain such biases so that they do not affect our actions and public policies. I take it that everyday libertarianism is a similar sort of obstacle to justice.
We have known about everyday libertarianism for at least a quarter of a century. Egalitarian philosophers and political actors have, so far, failed rather badly at removing this confused view from the public psyche. Though the effort of enlightenment should continue, we cannot simply wish it away.Footnote 61 Our public policies should be informed by the realistic recognition of the constraints of everyday libertarianism – or more broadly, anti-tax attitudes based on loosely libertarian grounds – to the extent that it is stubborn and difficult to eradicate in the short to medium term.
According to Murphy, everyday libertarianism is persistent because it tends to align with the dominant conventional morality about property. Our conventional morality of property takes the form of a simple deontological rule, which is exemplified by how we educate our children: do not steal, respect other people’s property and so on (Murphy Reference Murphy2020: 484). It may be philosophically the case that we should respect property rights only if the society is sufficiently just and/or doing so is socially beneficial. However, that is not what most of us tell our kids or even ourselves. Murphy observes that for most of us,
life is simpler, easier, and, in a positive sense of the word, more humble if we simply treat the rules of the practices as moral rules. Living according to the whole truth about property and contract seems to be, at best, complicated and dreary; at worst, it puts us in the position of some kind of impartial social planner to the apparent detriment of our own character and the decent treatment of those with whom we deal. (Murphy Reference Murphy2020: 484–5)Footnote 62
This analysis seems plausible. It explains why many people are drawn to viewing pre-tax income (not just of themselves but of others) as pro tanto inviolable without subscribing to philosophical libertarianism in any coherent way. In a sense, everyday libertarianism is a by-product of a useful heuristic for navigating daily life. It is certainly an obstacle to justice of social arrangements, yet it is understandable how, for many people, the experience of their daily lives takes precedence and spills over to the consideration of social justice.
Does this mean egalitarian theories of justice have an inherent tension with how people tend to think? Not necessarily. Everyday libertarianism doesn’t oppose distributive equality per se. If we can minimize income inequality at the level of pre-tax income, the popular obsession with pre-tax income no longer stands in the way of equality. Moreover, some evidence suggests that people who oppose tax-and-transfer redistribution may still support policies that, in effect, distribute income to the poor in other ways (Liscow Reference Liscow2022).
These considerations about the persistence of everyday libertarian attitudes make the instrumental case for predistribution more pertinent than its critics might think. My account of predistribution focuses on the lack of necessity to reference the distribution of pre-tax income. To the extent that people are sensitive to the government’s interventions in their pre-tax income, policies of predistribution hold a better chance of winning a democratic consensus in the real world of capitalism than the targeted tax-and-transfer policies of redistribution. My intention is not to downplay the importance of redistribution but to clarify the strengths and weaknesses of different policies. Prevention of inequality and poverty in the sense of minimizing inequality at the level of pre-tax income has a non-negligible strategic value and should be taken seriously.
This paper has attempted to provide a clear and practically useful account of predistribution. The functional definition of predistribution marked by the notion of procedural priority serves as a sufficiently clear demarcation of predistribution. Similarly, the distinction of three types of predistributive policies based on different meanings of priority/prevention should help normative and pragmatic discussions about predistribution, by categorizing policies around different functions of preventing inequality, poverty and economic domination.
Predistribution substantiates the prevention-is-better-than-cure thesis in four ways. First, all types of predistribution have institutional priority to the market. They aim to prevent morally objectionable inequality, poverty and economic domination by setting up fair institutions of the market. Second, pre-market distribution is preventive from the personal perspective as it equips individuals upfront with access to resources for partaking in social cooperation. Third, collective rent sharing capitalizes on the value-generative priority of some economic values. It aims to prevent unfair rent-seeking, which no member of the current generation is entitled to, as well as the accompanying inequality and exclusion from access to common resources. Finally, it has a strategic value, in the face of persistent everyday libertarianism, to prevent inequality and poverty by minimizing inequality before income taxation. These meanings of prevention sometimes overlap, but as long as we are clear about the primary feature of the policies, there should be little confusion. Instead, such overlap suggests a confluence of reasons for supporting predistribution.
Acknowledgements
For helpful discussions and feedback on this paper, the author is deeply grateful to Chikako Endo, Gen Fukushima, Takuto Kobayashi, Takeshi Miyai, Masaya Miyamoto, Martin O’Neill, Shin Osawa, Ken Oshitani, Tom Parr, Junichi Saito, Areti Theofilopoulou, Kate Vredenburgh, Stuart White, Jacob Watkins-Strand, Jo Wolff, and the participants at the Nuffield Political Theory Workshop and the MANCEPT Brave New World Conference. This paper also benefited from the extremely helpful comments of three anonymous reviewers for Economics and Philosophy.
Appendix. Representative policies of predistribution (non-exhaustive)
Dai Oba is a Junior Researcher and JSPS Postdoctoral Research Fellow at Waseda University. He is also an associate member at Nuffield College and the Department of Politics and International Relations, University of Oxford. His current scholarship focuses on theories of equality, philosophy and public policy, policies of social welfare and regulation of the market, and the role of normative theory in policy guidance in the real world.
Open access
