II Children as Public Goods

According to the fair play principle, when some individual or group intentionally incurs costs to produce a non-excludable and non-rivalrous benefit that is enjoyed by non-producers, those non-producers incur an obligation to contribute a fair share of the costs incurred by the producer(s) in the productive act(s) (Hart Reference Hart and Waldron1984[1955]). To refuse to do so would be to wrongly ‘free ride’ on the efforts of the producers. In the case of children, parents are seen as the producers and non-parents are seen as the beneficiaries.

Three reasons explain why the fair play principle is an attractive approach to justice in the distribution of the costs of children. The first is that the principle promises to establish an enforceable rather than unenforceable obligation. An enforceable obligation is necessary to ground the sorts of tax and transfer policies entailed by public subsidy for the provision of education and childcare as part of a welfare state. The second reason is that the fair play principle does not rely on any perfectionist premises about the contribution that parenting makes to the flourishing of parents. For example, arguments for cost-sharing based on the contribution that parenting makes to the flourishing of some individuals could be reasonably rejected by those who deny that parenting makes such a contribution. It would be better, at least, to have an argument that did not rely on such controversial claims, and could be the basis for political agreement among reasonable people who disagree about the correct view of human flourishing. The third reason is that the fair play principle offers us a normative argument rather than a prudential one. This is important because, although non-parents have prudential reasons to contribute to the costs of children to incentivize their production, prudential arguments can rarely justify enforceable obligations on their own. These three reasons provide a strong case for viewing the fair play principle as the best hope for grounding the Standard View which, as I noted earlier, is the view that parents and non-parents must share the costs of all children’s entitlements equally (Olsaretti Reference Olsaretti2013, Olsaretti Reference Olsaretti2017).Footnote ⁵

Serena Olsaretti explains that one possible reason for the neglect of the question of Who? is that theories of justice have typically assumed the truth of the Standard View.Footnote ⁶ Olsaretti’s reasoning applies at least as strongly to the question of How Much?

Other than Olsaretti’s brief statement above, there is not much available in terms of an answer to our question. We are in unchartered territory.

I take it that the Standard View consists of two main claims. First, the costs of each child’s entitlements should be shared and, second, these costs should be shared equally between parents and non-parents. Furthermore, if we suppose that children’s entitlements to education and care go beyond what is provided in our currently unjust world, then the total to be shared could be very large indeed since those amounts are already large. Below are just two examples.

In the United States, Child Tax Credit entitles working parents to refunds of up to $3,600 per child per year (US Treasury 2022). According to the Education Data Initiative, US Public Education Spending Statistics funding for K-12 education spending is $16,390 per pupil (Education Data Initiative 2023). The UK government was forecast to spend £12.6 billion on child benefit by 2023–24, representing 1% of all public spending (Office for Budget Responsibility 2023). In 2023–24, UK government spending on Education for under 16s in total was £57.3 billion, which works out at £7,460 per pupil (UK Government Statistics Service 2023).

While the Standard View seems favoured by many theories of justice it cannot simply be assumed. It must be argued for. Since the three reasons given above lead us to conclude that the fair play principle is the best hope for justifying cost sharing, we should consider how that principle limits answers to the question of How Much?

III The Ambiguity of the Fair Play Principle

According to the fair play principle, it is wrongful to refuse to contribute to the costs of public goods from which you benefit. But does the principle of fair play hold that it is wrongful if beneficiaries stand willing to contribute only up to some limit? If so, where is that limit? The fair play principle generates duties for beneficiaries based on their benefitting from the productive efforts of others. If there is no production, no effort, or no benefit, the principle simply does not apply. We can draw out four plausible claims about which costs can and cannot be shared from the logic of the fair play principle.

1. 1) Morally Required: the costs to be shared by non-producers must be the costs of meeting entitlements generated to create the public good. Costs of non-entitlements may be incurred or not by producers since it is not a matter of duty whether those costs are paid at all. The Standard View holds that these are the costs of children’s entitlements.

2. 2) Incurred Costs: the costs to be shared must be ones that the producer in fact incurs. For instance, if a wizard is able to produce a bridge by merely casting a spell, one is not required to pay the wizard a fair share of the market price of building that bridge.

3. 3) Related Costs: the costs to be shared must be costs of production of the public good. One cannot be required to share the costs that are unrelated to the production of a public good. For instance, the costs of other projects the producer is involved in. There must be some salient connection between the costs and the productive activity that the non-producers benefit from.

The above are general limits to the costs that can be shared by non-producers. But there are also some costs that, by their nature, beneficiaries cannot refuse to share.

1. 4) Rights Violations: the costs to be shared must be the costs of production of a public good if it is to be produced without violating anyone’s rights. Non-producers cannot refuse to share a cost because the producer could have stolen some item necessary for production or else refused to pay the workers. The beneficiary cannot expect the producer to reduce their costs by taking part in wrongdoing.

These considerations lead us to what I call The Costs Claim.

The Costs Claim takes us beyond what has already been said in relation to the question of How Much? but it is ambiguous in two important respects. First, the nature of the relationship between the entitlements and public good production, identified in ‘Related Costs’ above, is unclear. Does it include all children’s entitlements, as the Standard View states, or only those entitlements that are necessary to produce a public good? Second, the size or type of public good that is created can vary and so do the entitlements arising from it. Do beneficiaries have to pay the related costs of any public good that is created or is there a limit to how large the good can be?

In the next two sections, I resolve these ambiguities to arrive at a more determinate account of the extent of non-producer cost-sharing that differs from the Standard View. But first, let’s consider why a much simpler approach is not available to us. It may appear that we can resolve both ambiguities within the Costs Claim by appealing to an independent account of a ‘fair share’. Here we must keep in mind the difference between the total costs to be shared and the size of individual shares. There are two ways that a beneficiary’s share can increase. First, it can increase because their share of the same total increases; for example, a beneficiary pays principle would hold that some should pay more because they benefit more than others.Footnote ⁷ Second, it can increase because the total that is to be shared is larger; for instance, because the public good is produced in a more expensive way.Footnote ⁸ The answer to the question of How Much? here identifies the total costs that are to be shared fairly by the group of beneficiaries, not what an individual’s fair share of that total is.Footnote ⁹ We have to identify the total costs before we can work out the size of anyone’s fair share, which will require the application of the principle of distributive justice. As we have seen, some costs simply do not factor into the total to be fairly shared. These include: morally optional costs, costs that are not incurred by producers, and costs that are unrelated or tangentially related to production. What costs should and should not be shared is distinct from the question of how to fairly share what can be shared. This is the difference between which costs can be billed and how to share the bill between individuals. For this reason, we cannot apply an existing account of fairness to arrive at our answer.

IV Which Entitlements?

The first ambiguity within the Costs Claim concerns the relationship between entitlements created and the production of the public good. There are many different entitlements that may be created in the creation of children. The Standard View focuses on the costs of meeting each child’s entitlements, and not more broadly on entitlements created in order to produce the public good without violating anyone’s rights. In discussions of the costs of children, scholars have distinguished between the costs of meeting children’s just entitlements, including meeting their need for care and their right to a fair share of resources, and more extravagant costs, such as organic cotton clothes (Olsaretti Reference Olsaretti2013, 230). While the latter are not entitlements of the child, they are entitlements created in the creation of the public good that children produce. A payment is owed to the organic cotton clothes merchant. Because they are extravagant it is plausible to assume that there is no duty for non-parents to share these costs. So, how do we distinguish those costs that must be shared from those that need not?

There are two possible reasons why the costs of organic cotton clothes should not be shared. It could be because they are not entitlements of the child, as the Standard View would hold. This view supports the following restriction to cost sharing.

But I will argue that the reason why the costs of the clothes should not be shared is that they are not necessary to produce the public good. I will defend a different restriction on cost sharing.

To compare the two ways of drawing the distinction, we must consider whether it is plausible to include necessary entitlements that are not the child’s and the child’s entitlements that are not necessary to produce the public good without violating rights.

We can now consider a case involving Necessary Entitlements that are not Children’s Entitlements.

Pain relief is a clear example of a necessary entitlement, but no child is entitled to be born to someone who is on pain relief. The public good of children, whether in terms of workforce or pension contributions, cannot be produced without violating anyone’s rights without pain relief. It is intuitive that non-parents should share the costs of pain relief in this example, just as it is intuitive that beneficiaries of national defence must share the costs of building tanks and paying soldiers and that beneficiaries of clean water must share the costs of building water treatment facilities.

Now consider an example of an entitlement that is not the child’s entitlement. Consider the recent history of the world, where fertility levels surpass what is required to produce adequate levels of public goods (Roser Reference Roser2014). Some individuals use Assisted Reproductive Technology to procreate, but no child has the right to be created through the use of ART, so the entitlements created in the commission of ART are not those of the child and they are not necessary to produce the public good. When there is a plentiful supply of children to produce public goods, the costs of producing additional children in a more expensive way cannot be shared using the fair play principle.Footnote ¹¹ Viewing things through the lens of the fair play principle, our focus should be solely on the production of the public good. Our sympathy for those who experience difficulty procreating and the disadvantage they face may provide good normative reasons for sharing their costs, but such reasons do not stem from the fair play principle. Nor should we expect the fair play principle to do all the work of a theory of justice or a theory of rectification of injustice. This is not to dispute that the costs of ART should be covered as a matter of justice for other reasons. However, these arguments would not be supported by all three reasons noted in the introduction for focusing on the fair play principle.

With much of the world failing to reach replacement rates of fertility currently, ART may become increasingly necessary for minimal levels of procreation for the production of public goods (Bhattacharjee et al. Reference Bhattacharjee2024). Then the costs of ART would become shareable. A beneficiary could not insist that the public good was created without ART because the good they enjoy could not have been produced without it.Footnote ¹² This supports Necessary Entitlements as the correct determinant of whether a cost can be shared.

Before we move on to the second ambiguity, I consider one further example that suggests the Necessary Entitlements are implausible because too restrictive. Paul Bou-Habib and Serena Olsaretti consider an example of producing a public good more cheaply by workers working at night rather than during the day (Reference Bou-Habib and Olsaretti2023, 10). They believe that it is counter-intuitive to hold that beneficiaries must share only the costs of nighttime production. The reason they give is that the workers have a compelling interest not to work at nighttime. They argue by analogy between parents and the workers and claim that parents have a compelling interest in choosing their family size. They then conclude that beneficiaries must share the costs of additional children. In brief, the objection to Necessary Entitlements is that there may be unnecessary costs that should be shared in order not to compromise a compelling interest of producers.Footnote ¹³

Necessary Entitlements can accommodate the intuition in this case if the compelling interest grounds a right. If workers have a right against working in the evenings, as seems plausible, then we should not index the costs to be shared to evening work. If, however, workers lack that right or were willing to waive it in order to earn overtime payments, these lower costs would be all that non-producers could be required to share based on the fair play principle. In this variation, the compelling interests view seems to imply that beneficiaries should pay the higher costs that daytime work entails. This seems implausible and since they are not necessary, the Necessary Entitlements view should be favoured.

In the case of the workers, the size of the goods produced is the same whenever the work takes place. Only the costs are different. The analogous case with families would be that parents have additional children, increasing the costs, but producing the same public good. So, for the family size case, these additional costs are strictly unnecessary to produce the public good of whatever size. The fair play principle cannot support cost sharing because there is no contribution to the public good by the action. This entitlement is like that of organic cotton clothes. A parent could complain by appealing to their compelling interest in determining family size. The compelling interest in family size is most forceful when it protects procreative liberty; for example, when governments make it illegal to have more than one child. But procreative liberty is not at stake in the case of cost sharing. Indeed, Bou-Habib and Olsaretti accept limits to cost sharing based on the size of the family (Reference Bou-Habib and Olsaretti2023, 11). I am not arguing that adults act impermissibly by having more or fewer children. What is at stake is the interest in ‘having children and not paying the full costs of their entitlements but, rather, sharing them with others who didn’t make the same choice’. Non-parents’ countervailing interests will be similarly compelling and include interests in all the other things they could do with those contributions. In fact, each interest is a type of more general interest in pursuing one’s important projects while sharing their costs. To avoid perfectionism, it would appear the interest must be formulated more generally: the compelling interest in ‘having one’s especially important project’s costs shared’.

The arbitration of conflicts of interests of this kind, about which there can be reasonable disagreement, cannot be made by the fair play principle. The arbitration of such conflicts would make the fair play principle deeply controversial and arguably perfectionist. The compelling interests objection, then, does not undermine the Necessary Entitlements view.

Where family size becomes relevant to cost sharing is where parents exercise their freedom to procreate by having additional children, and where doing so contributes to a larger public good that non-parents enjoy. When these additional children are necessary to create that larger good, Necessary Entitlements do not disallow cost sharing for larger public goods.

V At What Cost?

It is possible to create public goods of varying sizes at varying costs. In the case of children, a larger more-skilled generation will likely cost more than a smaller less-skilled generation, but it will also likely confer larger benefits on beneficiaries. The costs claim is ambiguous about whether there are any limits to the size of public good that can be created and its necessary costs be shared. In this section, I consider three possible answers to this ambiguity: the Default Answer, the Cheapest Answer, and the Good Value Answer. I defend the Cheapest Answer.

First, consider the Default Answer that I associate with the Standard View because it holds that the costs of all children that are born are included.

According to this answer, there are no limits based on how large or small the public good can be and its necessary costs shared. If a public good of three sizes could be created: large for large costs, medium for medium costs, and small for small costs, the default answer tells us that the costs to be shared are those linked to the good that is produced. If a medium public good is produced, medium costs are to be shared even though a larger benefit could be created, or a cheaper benefit could be created instead. It doesn’t matter if it is poor value for the benefit compared to the others; for example, if a slightly smaller benefit would be produced at a much lower cost or if a much larger benefit can be created at very little additional cost.

What is objectionable about this answer is the absence of some independent principled limit to cost foisting. The unilateral imposition of a duty is what motivates objections to the fair play principle more generally, and these objections are not without force. Whatever else we can say about public goods production, the decision about the level of public good provision and the attending levels of necessary cost sharing cannot be left to the producer’s actions alone. Even if parents retain a right to produce a good of any size or of the kind they wish, such as determining family size, they cannot be granted the power to determine the extent of duties to share the costs of what they produce. But we must be clear to distinguish this objection from an objection that would apply with equal force to any version of the fair play principle.

Critics of the fair play argument, such as Nozick, argue that foisting any sort of benefit on someone else cannot generate such obligations.Footnote ¹⁴ Such arguments cannot be reconciled with any cost-sharing and thus appear inconsistent with the fair play principle.

But the objection I am making is importantly different. I do not deny that there can be duties to share the costs of foisted benefits. I hold instead that choosing to foist more demanding rather than less demanding duties to share costs is objectionable. The argument thus retains what is forceful in Nozick’s objection without undermining the principle of fair play. Foisting goods and requesting payment generates a strong argumentative burden, though not an insurmountable one. To understand the fair play principle differently is to grant producers not only the power to create duties unilaterally, which must be a power they possess if the fair play principle is sound, but also the power to increase the demandingness of those duties in line with the producers’ choices about the level of public goods provision. I have defended a constraint on the extent of costs not on the idea of cost sharing altogether and so my defence does not beg the question against the fair play argument. Nozick’s stance appears too protective of the beneficiaries and not enough of the producers. The Default Answer appears to favour the producers at the expense of beneficiaries. What I am suggesting strikes a balance between the claims of producers to have the costs shared and the claims of beneficiaries to have some principled limits to those costs. To avoid the problem with the Default Answer, we must set costs independently of what good was in fact produced. This means that parents may overproduce and not be entitled to have non-parents share with them the full costs of all children, at least not on the basis of the fair play principle alone.Footnote ¹⁵

The independent standard I will defend is the Cheapest Answer.

The Cheapest Answer is the lowest cost that would be justified if the fair play principle is sound since it is the least objectionable liability that non-parents could have. If non-parents can’t even be required to share the cheapest costs, then no cost sharing could be justified. Since I am assuming the truth of the public goods argument, this position must be accepted and our question then becomes whether any deviations from the cheapest answer are justifiable. An alternative independent standard that is worth considering would require non-parents to share the costs of the public good that represents good value.

Non-producers may have prudential reasons to pay more if they enjoy a larger benefit for relatively small additional costs. For example, recruiting more and more military personnel will increase national defence marginally. The same can be said about clean air and public transportation systems. On the Good Value Answer, non-producers would be required to share only the costs that would be the best value for a good of that kind, up to the costs incurred by the producer.Footnote ¹⁶

But the Good Value Answer is problematic. While it may be narrowly rational to prefer a benefit of 10 for the price of 5 than a benefit of 6 for the price of 4, it may not be rational in a broader sense because public goods are not the only things people care about. Paying a price for 5 rather than 4 deprives the non-producer of 1 that can be used elsewhere, perhaps much more efficiently given their conception of the good or welfare function. Even a strictly prudential argument doesn’t, therefore, always support the Good Value Answer, and even so, a prudential argument would not be of the right kind to justify enforceable cost sharing, as discussed in the introduction. A conflict of interest arises around what is best for any individual, as was noted above regarding the compelling interest in choosing family size. There will be reasonable disagreement about this. The idea of good value is itself hard to determine and cannot be determined by the fair play principle alone. Here we run into major problems in attempting to define an independent standard for cost sharing that is justifiable.

The Good Value Answer assumes that providing a larger benefit for better value must be at least as justifiable as a smaller benefit for worse value. But the point of public policy is not generally to maximize any particular net benefits people receive in terms of public goods. It is easy to see how this can get out of control when a very large but great-value public good improvement is on offer. Perhaps national defence can be improved 100 times by only increasing costs by 4 times. That looks like a good deal, but the increase in costs can have huge opportunity costs for some, maybe even all. The fair play principle justifies the foisting of benefits and retrieval of costs only in a narrow set of cases, acknowledging the special burdens that face any attempt to foist.

In this section, I have argued that non-producers lack a duty grounded in the fair play principle to share any costs greater than those necessary to create the cheapest public good. A willingness to pay the cheapest costs but not more is all that is required for non-parents to avoid free riding. While larger goods can be created (after all my argument does not propose limiting the freedom of producers), the corresponding larger costs cannot be shared consistent with the fair play argument.

VI Upshot

I have argued that the fair play principle requires that the necessary cheapest costs of children as public goods are shared. Non-parents who pay their fair share of those costs are not guilty of free-riding. In this section, I clarify the upshot of this answer. First, I explain what I mean by the cheapest necessary costs. Second, I highlight three ways in which this level of cost-sharing will be different from that supposed by the Standard View.

The cheapest necessary costs are the costs of meeting the necessary entitlements arising from the procreation and nurture of a set of children that will produce the cheapest substantial public good. The costs will include those that are practically as well as morally required to produce a substantial public good without violating anyone’s rights. We should note that, in different places and at different times, meeting a child’s entitlements was more and less expensive. My focus is on the costs of meeting the entitlements at the time they should be met, but even so, in different places, the costs of meeting children’s entitlements vary dramatically based on where they are raised. This variation can have an important implication, for example, in a case that involves skilled immigration. Suppose that a state can hypothetically benefit from admitting skilled immigrants. Furthermore, suppose that the costs that the parents of those skilled immigrants would have to bear in raising them are lower than the costs of raising local children in the receiving state. In that case, the Cheapest Answer may well imply that non-parents in the receiving state cannot be expected to share with local parents all of the costs of raising local children, but only those costs that match the lower amount it would have cost to raise skilled immigrants.Footnote ¹⁷

There are three ways that the Cheapest Necessary Costs View will differ from the Standard View. Overall, the view supports far less cost-sharing than has been supposed. First, the Cheapest Necessary Costs View permits that there are additional entitlements whose costs must be shared over and above the child’s entitlements. This follows from the rejection of Children’s Entitlements in favour of the Necessary Entitlements. The cost of entitlements of non-children, such as costs of pain relief during childbirth that are necessarily created to create the public good without violating rights must be shared. This will drive cost sharing upwards. But there are two further ways the costs can be driven down.

Second, the Cheapest Necessary Costs View permits that the entitlements of some children will not be shared. This is because it was possible to create the public good without violating anyone’s rights and without creating those children’s entitlements. Only the set of children that it is necessary to create in order to create the cheapest public good without violating anyone’s rights can be shared. This follows from rejecting the Default Answer and Children’s Entitlements in favour of Necessary Entitlements and Cheapest Answer. Unless the production of children corresponds exactly with the cheapest way of producing public goods, without violating anyone’s rights, the total amount that can be fairly shared by non-parents will be less than it costs to cover all of the entitlements of all the children that are born. Moreover, many children who are created do not contribute to the public good.Footnote ¹⁸ So, their entitlements cannot be shared on the fair play principle either. This significantly brings down the total costs that can be shared.

Third, and what is an implication of the other two, the view does not require equal sharing between parents and non-parents, as the Standard View holds. The Cheapest Necessary Costs View permits inequality in sharing between parents and non-parents. Although the debate about the costs of children has focused on the way it generates duties for non-parents, it also generates duties for parents. That is because the argument generates duties for beneficiaries and not merely non-producer beneficiaries such as non-parents. The fair play principle will sometimes justify very little cost-sharing by non-parents, and less than its proponents have supposed. It doesn’t mean those parents should receive nothing. That turns on how we distribute the revenue. But since the revenue, when divided, won’t cover all children’s entitlements, each parent may get less than it takes to cover the costs, though there are different ways of dividing the revenue, including sharing equally across all parents.

This leaves proponents of the fair play principle for the costs of children with a choice. Either they must supplement fair play with another argument, though the three reasons outlined at the start for thinking that approach is attractive seem almost unique to it, or they must accept that the extent of cost-sharing that the Standard View adopts is incorrect and likely overstates the extent of cost-sharing that is required.Footnote ¹⁹