Children or Migrants as Public Goods?

Family Policies

Family policies are public policies that ensure the ‘collective provision’ (Lewis, 2003) or the ‘socialisation’ of the morally required costs of raising children. These are the costs that must be incurred by someone (whether by parents ⁷ or society at large) in order to give children what they have a right to. ⁸ Examples of family policies include (1) publicly funded parental leave, (2) free or subsidised pre-school childcare and public schools and (3) child tax credits or family allowances. We assume that these policies ought to be generous enough to meet certain recognised benchmarks of children’s and parents’ needs (e.g. UNICEF, 2017). ⁹

Note that the family policies we focus on do not distribute certain further costs that are generated when people have and raise children. These further costs include non-financial costs, of which some may be morally required, for example, the cost of forgone free time which parents incur in order to care for their children (Folbre and Bittman, 2004; Gál et al., 2018; Rose, 2016), or the loss in personal autonomy parents incur in order to provide their children with continuity of care and stability (Alstott, 2005). Another further cost is the lifetime environmental cost of an added child, which figures prominently in debates over whether the choice to procreate accords with climate justice (Casal, 1999; Conly, 2016; Cripps, 2017; Heyward, 2012; Hickey et al., 2016; Meijers, 2016; Overall, 2012; Pinkert and Sticker, 2020; Young, 2001). While our discussion does not aim to settle questions concerning the measurement and distribution of these further costs, we briefly consider the environmental costs of children towards the end of this article, since – as mentioned earlier – a concern with environmental costs may strengthen the replacement migration challenge.

Our discussion also sets aside the question of what form, exactly, family policies should take. It does not examine whether family policies should aim at the defamilialisation of care (e.g. subsidised daycare for babies and toddlers) or supported familialism (e.g. longer and/or more generous parental leaves) (Saraceno, 2010). We also do not discuss further ways in which parents might be supported by the taxpayer – for example, through a pro-parenthood-adjusted public pension system (Gál et al., 2018; Sinn, 2005).

In sum, we focus on family policies that have an intended and substantial impact on the distribution of the financial costs that are morally required for adequate child-rearing between parents and non-parents. These policies currently divide a large proportion of those costs equally between all citizens, regardless of whether they are parents or not, and yet seem primarily to benefit parents and children. This asymmetry is of practical and theoretical significance. Practically, it makes it important to justify these policies in the face of a perceived ‘backlash’ against them from non-parents (Burkett, 2000; Fineman, 2004; Wolf et al., 2011). Theoretically, the justification of these policies raises a special challenge, because while many welfare state policies do not benefit all taxpayers equally, this is typically in order to meet needs people are not responsible for having (e.g. most medical needs). By contrast, it seems harder to justify policies which, while paid for by all citizens, appear to benefit only some of them and not in order to meet unchosen needs, thereby favouring conceptions of the good life that only some and not others embrace. For example, people disagree about how valuable it is to become a parent; about the value of large families; about how populated the planet should be. Family policies thus seem to be under pressure from what political theorists refer to as the challenges of personal responsibility and neutrality (Casal and Williams, 1995; Clayton, 2006; Rakowski, 1991).

The public goods argument is well placed to meet these challenges. Nancy Folbre (2001: 50) states the argument’s core idea as follows:

Parents who raise happy, healthy, and successful children create an especially important public good. Children themselves are not the only beneficiaries. Employers profit from access to productive workers. The elderly benefit from Social Security taxes paid by the younger generation . . . Fellow citizens gain from having productive and law-abiding neighbours. ¹⁰

By emphasising the benefits parents create, the public goods argument promises to meet head-on the challenges of personal responsibility and neutrality. In response to the former, it can say that people are required to bear the costs of their choices only if these would otherwise generate costs for others, as is the case, for example, when an imprudent motorcyclist prefers to ride without a helmet and risks creating avoidable medical care costs for others (Fleurbaey, 1995). Because parents’ having and raising children benefits society, the challenge from personal responsibility is mistargeted when directed against parents (Casal and Williams, 1995).

In response to the neutrality challenge, the public goods argument can say that because the basis for sharing child-rearing costs with parents is the contribution they make, cost-sharing does not presume any controversial judgement that parenting is an intrinsic part of a good life. The concern with neutrality certainly has implications regarding which of the various benefits parents create are relevant for justifying cost-sharing with them. Cost-sharing with parents could not be justified, compatibly with neutrality, if child-rearing generated benefits that are regarded as such only from the perspective of one or another controversial conception of the good life. It is important to observe, therefore, that child-rearing creates benefits which all or most members of society can reasonably be presumed to want (John Rawls (1999: 54) famously calls these ‘social primary goods’). One of the central benefits of demographic renewal – the maintenance of the tax base – is an uncontroversial benefit from the perspective of most conceptions of the good life; indeed, we could say that parents’ having and raising children is a socially necessary activity, that is, one without which society could not exist as an ongoing just scheme of social cooperation (Rawls, 2001: 162). ¹¹

The Challenge of Replacement Migration

Replacement migration can contribute some of the very same public goods that parents contribute by having and rearing children. This raises a challenge to the public goods argument: assuming that replacement migration does provide the same public goods that figure in the public goods argument for sharing the costs of children with parents, can that argument still justify family policies? This section explains the challenge in detail.

To begin with, it is important to clarify what is meant by ‘replacement migration’. ¹² It refers to immigration that makes up for a demographic deficit, where this is a deficit in a country’s population as a whole or in a particular group within its population (e.g. working-age adults). Different kinds of replacement migration are thus possible, depending on the demographic deficit it makes up for (Gosseries and Zwarthoed, 2017). We focus on replacement migration that aims to replace a deficit in skilled workers as this form of migration is especially attractive to host states for reasons we explain shortly, and potentially constitutes the most serious challenge to the public goods argument. To facilitate this kind of replacement migration, countries can adopt a points-based system that gives immigrants visas according to their skills (e.g. Australia) or an employer-based system that issues visas to immigrants who have accepted job-offers at certain occupations (e.g. United States). These policies can be adjusted in order to attract immigrants of certain age-groups, and can vary in their specific provisions, for example, whether they permit immigrants to bring relatives with them (Gosseries and Zwarthoed, 2017).

To think about how the availability of replacement migration might affect the public goods argument, a simple example will help. In 1996, some 254,000 children were born in Australia (Australia Bureau of Statistics (ABS), 1996: 10). In 2014, when this cohort turned 18 years of age, Australia admitted around 125,000 skilled immigrants (DIBP, 2014: 4). The Australian government could have influenced the sizes of these two groups of persons through various policies. It could have offered less public support for families prior to 1996 and it could have adjusted its immigration policy to increase the number of skilled immigrants entering in 2014. Stated more generally, the government could have shifted in various degrees from relying on procreation to relying on immigration as a source of human capital. This is, of course, a recurring possibility: the Australian government, like many other governments, can affect the balance between these two sources of human capital today and in the future. Now assume that the cost for Australian taxpayers of obtaining human capital via replacement migration is lower than the cost of obtaining human capital via procreation. This assumption is plausible: replacement migration is certainly not costless, but it is doubtful that the average public cost, in financial terms, ¹³ of integrating a skilled migrant into the labour force, plus compensating the migrants’ parents and/or home state, exceeds the average public cost involved in rearing a child to adulthood in the receiving state (Bou-Habib, 2019). This is the case, at least, if the sending state is considerably poorer than the receiving state, for the funds that will have been devoted to raising the migrant will likely have been smaller than those devoted to rearing a child to adulthood in the receiving state.

Granting this assumption, the public goods argument seems to imply that funding for family policies may be reduced. In particular, it seems to imply that taxpayers are obliged to bear only those costs they would need to bear if Australia relied on replacement migration instead. This conclusion rests on a deeper claim that appears plausible at first sight: for any public good, taxpayers owe its producers only the amount taxpayers would have to pay for the least expensive way of producing it (cf. Shields, 2021). If correct, the replacement migration challenge shows that the family policies the public goods argument is able to justify are far less generous than might otherwise be supposed. In particular, instead of justifying providing all local parents with the standard package of family policies mentioned at the start of this article, the public goods argument might justify only providing minimal support for a small number of local children, reflecting the limited net benefits which local parents produce.

The replacement migration challenge applies regardless of whether or not there are environmental concerns about overpopulation, but it is worth noting that these concerns reinforce it. As one writer notes, ‘[i]n an age of actual or prospective overpopulation….. the suggestion that an allowance for children is justified by parents “serving a societal function” is entirely unpersuasive’ (Brazer, 1977, cited in George, 1993: 216). Indeed, in such a context, parents may be said to produce, instead of a public good, a negative externality or a public bad (Casal, 1999). ¹⁴ We believe our defence of the public goods argument against the replacement migration challenge holds, even once that challenge is reinforced by concerns about overpopulation. We explain why in the last section of this article.

Three Versions of the Public Goods Argument

To assess the replacement migration challenge, we now distinguish between three different versions of the public goods argument which are generally collapsed in the literature. ¹⁵ Disaggregating the public goods argument is important because it enables us to determine whether it is vulnerable to the replacement migration challenge in any or all of its possible versions.

On the first, pro-natalist version of the public goods argument, family policies are justified insofar as they are necessary means for incentivising people to have children. If families produce the important public goods mentioned earlier by having children, and if it is true that, absent public support for the costs of child-rearing, people have too few children, then family policies can be justified by appealing to their role in boosting the fertility rate.

A second, pro-investment version of the public goods argument emphasises the importance of parental investment in children. Here, unlike the pro-natalist argument, it is assumed that adults will have enough children (fertility rates are not what is in question), both with and without family policies in place. What justifies such policies is, instead, the fact that they are necessary for ensuring that children are raised to become productive and law-abiding citizens. In short, the pro-investment version differs from the pro-natalist version in that its aim is to ensure effective upbringing rather than a higher fertility rate (see Heckman, 2013; Heckman and Masterov, 2007; Putnam, 2015).

Both the pro-natalist and the pro-investment versions of the public goods argument are forward-looking: they point to consequences that sharing the costs of children is predicted to produce, although they focus on different consequences (namely, the optimal quantity (pro-natalist) and optimal quality (pro-investment) of children). ¹⁶ By contrast, the third version of the public goods argument is backward-looking: it justifies sharing the costs of children by reference to what we owe parents. We call this the fairness version of the public goods argument. It states that family policies are justified as a way of achieving a fair distribution of the burdens and benefits of child-rearing between parents and others. The argument rests on two key premises. The first, now familiar, empirical premise is that by having and raising children parents generate benefits for all citizens. ¹⁷ The second, normative premise is a principle of fairness. ¹⁸ This demands that some of the burdens, as well as the benefits, of at least socially necessary activities, be shared between those who undertake them and their beneficiaries so that no unfair freeriding takes place. Here we do not elaborate or defend this normative premise; this brief description of the argument should serve our purposes. ¹⁹

We are now in a position to assess the replacement migration challenge. Consider, first, the two forward-looking versions of the public goods argument. (The fairness version of the argument requires a more detailed discussion, reserved for the following section)

It is relatively clear that the pro-natalist version of the public goods argument is undermined by the possibility of replacement migration. As Bou-Habib (2019) argues, if the inclusion in our societies of migrant adults, including those who come with children, can offset the need to increase the local fertility rate, the pro-natalist case for sharing the costs of child-rearing fails.

By contrast, the pro-investment version of the public goods argument survives in the face of replacement migration. This is so as long as three things are true: (1) local adults will continue to have at least some children, with or without the support of family policies; (2) there are reasons to ensure that children are raised in certain ways (e.g. as productive members of society); and (3) public support for parents is necessary for ensuring that most or all children are raised in these ways. ²⁰

While the pro-investment version of the public goods argument is not threatened by replacement migration, we believe it is also important to defend the fairness version of the public goods argument against that challenge. This is because the latter is needed to capture the underlying convictions that move many people to embrace the public goods argument. To see why, imagine that those who have children go to great lengths, even in the absence of family policies, to ensure that their children are cared for, have adequate healthcare and receive a good education; and that in so doing they help to produce essential public goods for others in society. (In other words, imagine that family policies are not necessary to induce parents to have or invest in children) A society in which parents are made to pay mostly or fully for the costs of children, while at the same time ensuring that those public goods are shared among everyone, seems to distribute the burdens and benefits of social cooperation unfairly, to the detriment of parents (Olsaretti, 2013). It is thus important to consider whether the replacement migration challenge threatens the fairness version of the public goods argument.

Fairness and Public Goods: A Question of Baselines

As defenders of the fairness version of the public goods argument (or ‘the fairness argument’) emphasise, for beneficiaries to be obligated to share costs with benefits-producers, it is not enough that they obtain a benefit: it is also necessary that they obtain benefits that are ‘worth their costs’ (Arneson, 1982: 623). ²¹ In this section, we show that how we understand this condition matters crucially for how we interpret the fairness argument, and bears directly on whether it is threatened by the availability of replacement migration: when replacement migration can provide the same public goods that local parents provide, but more cost-effectively, it could be said that the benefits which local parents produce are not worth their costs and that, therefore, fairness does not require sharing the costs of children with local parents. As Rolf George (1993: 216) (who opposes this line of thinking) puts it, in a deliberately blunt way: ‘Why should one pay for locally produced children if they can be obtained so much more cheaply by import?’

In what follows we argue that this reasoning is flawed. We proceed in three steps. First, we explain that the fairness argument seems vulnerable to the replacement migration challenge only if we adopt one particular view of when the benefits of productive activity are worth their costs, that is by reference to an optimal baseline. Second, we introduce an alternative possible view of when benefits are worth their costs, which uses what we call a no-production baseline. If the fairness argument adopted the no-production baseline, it would offer almost unconstrained support for family policies for local parents. In a third step, we suggest that neither the optimal nor the no-production baseline is defensible and propose a third alternative, which we call the interest-sensitive baseline for determining when benefits are worth their costs. A fairness argument that adopts this baseline supports family policies for local parents even if replacement migration is a less costly source of human capital, but – it bears emphasising – the obligations it justifies are constrained, in ways we identify. What emerges from this discussion is both a defence of the fairness argument against the replacement migration challenge and a more nuanced understanding of public goods arguments.

We begin by unpacking a central claim of the public goods argument, namely, that those who receive benefits acquire obligations to share costs of production with producers only if the benefits are worth their costs. Merely obtaining benefits, regardless of what amount of cost beneficiaries are asked to share, is not enough to justify fairness obligations. As a well-known critic of the fairness argument, Robert Nozick (1974: 93), remarks, even though you may derive some benefit from your neighbours’ taking turns at running a local radio station, you are not obligated to do your turn when the time comes if ‘. . .you would rather not have any of it and not give up a day than have it all and spend one of your days at it’.

Note, however, that what exactly it means for benefits to be worth their costs is something that can be specified in different ways, with different implications for the demands of the fairness argument. In particular, it can have implications regarding how much of the public good beneficiaries have obligations to share the costs, and at what price, since these factors affect the cost-benefit relation. ²² Different ways of specifying the claim that benefits must be worth their costs rely on adopting different baselines. ²³ The challenge from replacement migration assumes that the relevant baseline is an optimal baseline: Here, a productive activity only provides benefits that are worth their costs if it brings the beneficiaries closer to the point at which they can optimally benefit from it. This is the point beyond which the cost of obtaining an extra unit of the relevant public good is greater than the benefit that the extra unit contributes. ²⁴ The optimal baseline view appears intuitively plausible in some cases. Consider a paradigmatic case of public good provision: national defence. If weapons factories were to produce military equipment in excess of the optimal amount needed by a society, it seems right that taxpayers do not have an obligation to share the costs of producing the oversupply. ²⁵

With regard to children, a fairness argument that adopts the optimal baseline would say the following: suppose that N_o is the optimal number of children for a given society, in the sense that, for any child over and above N_o, this child’s contribution in terms of public goods would be smaller than the cost of provision for that child. (Accordingly, the net total public goods yielded by parents’ having more than children N_o children is smaller than the net total public goods yielded by parents’ having N_o children.) A fairness argument that adopts the optimal baseline would say that the benefits of children in excess of N_o are not worth their costs and conclude that funding for family policies is not owed to parents for that excess number of children. Consider, again, our earlier example of Australia. Suppose that the number of new working-age persons added to Australia in 2014 – that is 379,000 persons – was optimal in the sense mentioned above. (These new persons comprised 254,000 local children born in 1996, and 125,000 skilled immigrants admitted in 2014.) Now suppose that, back in 1996, the Australian government knew it would be able to admit 225,000 rather than 125,000 skilled migrants in 2014, and that the financial costs of admitting the extra 100,000 migrants would be smaller than the costs of family policies supporting 100,000 local children. Under these circumstances, the benefits which Australian parents would contribute by having those 100,000 extra children would not be worth their cost according to the optimal baseline. The fairness argument would then imply that Australian parents in 1996 were entitled to family policies that shared the costs of only 154,000, not 254,000, children. As a result, either fewer Australian families should receive support from family policies or all Australian families should receive less generous support. Either way, replacement migration challenges the funding of standard family policies for all Australian parents if the fairness argument uses the optimal baseline.

Notice, next, that the replacement migration challenge is unlikely to arise if we adopt the no-production baseline – that is, if we maintain that benefits are worth their costs so long as the beneficiaries gain, by receiving and paying for the benefits, compared to a scenario in which no benefits are produced at all. This is true in the scenario we believe is realistic for most states in which replacement migration cannot fully supplant a society’s reliance on local child-rearing. ²⁶ Under these circumstances, within a certain range, local production of children is better for non-parents than no-production. So, if we adopt the no-production baseline, non-parents will have obligations of fairness to share the costs of children within that range, that is, up to the point when, if more local children were produced, non-parents would be worse off than if zero local children had been produced.

The no-production baseline is a rather undemanding standard: it is satisfied whenever the total benefits of children outweigh their total costs, and regardless of whether there are feasible and permissible alternative schemes for producing greater net total benefits (and relatedly, regardless of whether the marginal benefit of each additional child is greater or smaller than the marginal cost of that child). ²⁷ Take again the example of Australia: assuming that parents’ having 254,000 children in 1996, or even many more, made non-parents better off than if there had been no local children born in 1996, the benefits produced by all 254,000 or more children would be worth their costs, and family policies supporting all of them would be justified.

Which of these two baselines should the fairness argument use – the optimal or the no-production baseline? We believe that neither is defensible, and for the same reason: both are unreasonably one-sided in the manner in which they balance the claims of the producers and beneficiaries of public goods. The optimal baseline view favours beneficiaries unreasonably, whereas the no-production view favours producers unreasonably. After explaining this, we propose an alternative baseline view that we believe properly balances the claims of beneficiaries and producers. ²⁸

To see why the optimal baseline view is one-sided, recall that it says that beneficiaries must share costs with the producers only if the producers produce in a manner that approaches or achieves the optimal degree to which the beneficiaries can benefit from their production. This, however, overlooks the relevance of the fact that the manner in which public goods are produced affects also the interests of the producers. By way of illustration, suppose that a public good can be produced either through work performed during night-time or during the day, and that the costs of night-time production are lower, for example, because transport that no one uses at night-time can be used by the public goods producers. The beneficiaries and producers of this public good would benefit, even once the costs are taken into account, both if the public good were produced at night and if it were produced during the day; however they would obtain greater net benefits if the public good were produced at night, as they would have to share lower production costs for the same amount of public good produced. Yet night-time production would deprive those who produce the public good of a range of fundamental goods which are accessible only on a day-time schedule, such as opportunities to share quality time with friends and family. Suppose that, for this reason, those who produce the public good decide to produce it during the day rather than during the night. The optimal baseline view tells us that the beneficiaries in this case do not have a duty to share the costs of daytime transportation with the producers. This is because an alternative manner of producing the public good is available that could leave the beneficiaries still better off: the public goods could be produced during the night. Given this, beneficiaries have a duty to share the costs of producing this public good only to the extent that they, the beneficiaries, end up as well as off as they would have been had the producers produced it at night. Another way of putting this would be to say that the optimal baseline view makes producers exclusively liable for the shortfall in total net benefits they produce that results from their decision to work during day-time as opposed to night-time. This, we submit, is unreasonably one-sided: the optimal baseline view settles the question of which costs of producing a public good must be shared only from the perspective of beneficiaries; it fails to give appropriate weight to the interests that producers have in the manner in which a public good is produced.

Consider, next, the no-production baseline view. It maintains that beneficiaries must share costs with producers equally so long at the beneficiaries are made better off than they would have been in the absence of production. This view would allow producers to shift onto beneficiaries costs which the producers choose to incur in producing a public good that many will regard as unjustified. ²⁹ The no-production baseline view implies, for example, that beneficiaries must share equally with producers the costs of their choosing to travel to work by private cars, for example, even when more inexpensive and convenient public transportation is available, at least assuming that, despite these additional costs, the beneficiaries are still made better off, overall, than had the public good not been produced at all. Just as the optimal baseline view is insensitive to the interests of producers, the no-production baseline is insensitive to the interests of beneficiaries, making beneficiaries liable for the shortfall in optimal production regardless of whether the producers can offer the beneficiaries an adequate justification for that shortfall. (We assume that ‘this is how we like to produce the good’ is no justification). One could say that the no-production baseline view grants producers too much liberty in determining what the costs are that should be shared. ³⁰

The view we propose attempts to avoid these problems by invoking, as a yardstick for settling which costs should be shared, an idea familiar from Rawls’ theory of justice. This is the idea that certain interests are of central importance to all citizens despite their having different conceptions of the good life. We will call these compelling interests. The interests that Rawls identifies – including freedom of speech, conscience, and occupational choice – are ones that all individuals have in virtue of being ‘moral persons’, that is, persons with the capacity to develop and pursue their own plans of life and a sense of justice (Rawls, 1996: 293). Crucially, for our purposes, we contend that the interest in procreative choice counts as a compelling interest by the same standard Rawls adopts: to be able to realise any conception of the good we must be able to decide whether or not to procreate. ³¹ This seems defensible once we see this as the interest in being able to choose whether or not to have a child and is not reducible to an interest in procreating, because it protects the option of not procreating as well. That interest thus counts as exercised when one chooses to not procreate as well as when one chooses to procreate. ³²

Appealing to the idea of compelling interests, the alternative we propose to the optimal baseline view and the no production view is what call the interest-sensitive baseline view. It maintains that benefits are worth the costs of production, and hence that beneficiaries have an obligation to share costs with producers, if (1) beneficiaries are benefitted relative to the absence of the productive activity and (2) beneficiaries are brought as close as possible to the optimal baseline as is compatible with securing the producers’ compelling interests. If beneficiaries do not derive any net benefit relative to the absence of that activity, fairness does not generate an obligation to share costs with producers at all; if the productive activity departs from the optimal baseline in a way that is not necessary for securing a compelling interest, there is no fairness obligation to share in the costs of the suboptimal production.

To see what adopting the interest-sensitive baseline implies, consider the example of Australia one last time. Suppose that the compelling interest in procreative choice protects the choice to have more children per family than Australian parents in fact had in 1996, and assume that even with the non-optimal ‘extra’ 100,000 local children added, non-parents still benefit compared to a situation in which no new local children were added. In that case, the fairness argument would justify funding family policies also for the non-optimal ‘extra’ 100,000 children, or, in other words, for all 254,000 children.

We believe the interest-sensitive baseline will justify undiminished family policies, as in the Australia case just mentioned, despite the availability of replacement migration, under most even if not under all circumstances. There are two exceptions, in particular, that we should note.

First, in all states, large and small, there is no obligation for taxpayers to share the costs of child-rearing with parents who have a very large number of children because, we assume, the compelling interest in procreative choice does not protect the choice to have just any number of children. To be sure, whether we have a compelling interest to choose to have more than one child is a complex question. Compelling interests are especially weighty interests that everyone has equally. Given today’s global population size and concerns about climate justice, it is implausible that we have a compelling interest to choose to have as many children as we like. Accordingly, it is likely that obligations of fairness to share the costs of children in wealthy states, where individuals’ lifetime environmental impact is exceedingly high, run out after the first child. ³³ We return to the issue of the environmental costs of children later, but we would like to emphasise, for now, that it is compatible with our view to say that, in situations in which some people have children over and above the number that would be justified by their compelling interest, or where not everyone can exercise their compelling interests compatibly with our collective climate justice obligations, those who choose to not procreate in order to not contribute to the climate crisis may themselves have claims to some form of compensation. ³⁴

Second, under some circumstances, in very small states, it is possible that taxpayers are not benefitted by local parents’ having children relative to the no-production baseline, that is compared to an alternative arrangement when no new local children are created. It is possible that a state like San Marino, for example, could rely fully on immigration in order to meet its needs for human capital. ³⁵ Suppose, however, that local adults in San Marino still choose to have children. Unless it can be successfully argued that at least some local procreation is necessary to produce a distinctive public good, that is, to maintain the cultural continuity necessary for just local institutions to continue functioning, then the case for sharing the costs of local children cannot be grounded in fairness owed to parents. Instead, the case would have to be made on the basis of a different, purely interest-based argument.

To see how an interest-based argument differs from the interest-sensitive fairness argument, note that what grounds the obligation to fund family policies, according to an interest-based argument, is the fact that the fulfilment of the interest in procreative choice benefits the parent. To be sure, on a plausible version of the interest-based argument, how costly it is to fulfil parents’ interests matters for whether there is an obligation to subsidise that interest: such an obligation is easier to justify if it is less costly to fulfil. In the case we are considering, parenting produces some benefits for others. Helping to fulfil the interest in parenting is, for that reason, less costly, on the whole, than it would be if parenting did not produce any benefits for others. This makes it easier to justify the obligation to promote parents’ interests. But depending on the strength of the interest in parenting, and the costs of fulfilling it, the interest-based argument might also require non-parents to accept a net loss in order to support parents. On our interest-sensitive fairness argument, by contrast, what grounds the obligation to fund family policies is the fact that it is unfair to obtain benefits that others produce without bearing a fair share of the costs; the appeal to compelling interests only enters into the argument by constraining how much the beneficiaries may expect to gain from the production of others. By way of rough comparison, consider the difference between the argument that we should pay someone for his job as an engineer on the ground that he is making a socially productive contribution, on the one hand, and the argument that we should pay him on the ground that he has an interest in engaging in complex work, on the other. Just as it would be wrong, when people work at socially productive jobs, to claim that the only or main reason they should be paid is as a form of subsidy for the interests which those jobs help them realise, so it would be wrong, when parents produce public goods, to view sharing the costs of children as equivalent to a subsidy for an expensive project they have an interest in pursuing. ³⁶

In conclusion, we would like to consider the suggestion that it would be more accurate to characterise the argument we defend as a ‘hybrid’ argument, rather than as a pure public goods argument. It might be suggested, that is, that our argument combines two different and independent sets of considerations, one regarding fairness obligations (which public goods arguments appeal to), and the other regarding an obligation to help fulfil each other’s compelling interests (which an interests-based argument appeals to). In reply: we do not find the distinction between a hybrid and a pure public goods argument enlightening. This is because, we submit, no plausible public goods argument is ‘pure’ in the sense of being unconstrained by respect for the compelling interests of producers. Think again about our military equipment example. People lack a compelling interest to choose how many weapons to manufacture, and accordingly, it is plausible to think that taxpayers do not have an obligation to fund an over-supply of weapons just because it is the result of that choice. But people who work in weapons factories do have other compelling interests (e.g. their interest in being able to take some time off work) and these do constrain how much we may expect to benefit from them. We maintain, then, that the correct baseline relative to which beneficiaries must gain before they owe obligations to producers is always, in the end, an interest-sensitive baseline. Accordingly, we see no reason for characterising our argument as a hybrid argument, rather than as the most plausible interpretation of the public goods argument.

The Environmental Costs of Children

The replacement migration challenge seems especially pressing once we take into account concerns about the environmental costs of local procreation. Local procreation contributes to human capital by adding a new person to the global population. Replacement migration, by contrast, admits to the state of a person who already exists. While the migration of persons from lower-emitting to higher-emitting states may increase carbon emissions (Cafaro, 2009), it is plausible to assume that local procreation in higher-emitting states increases carbon emissions even more, because it adds a whole lifetime’s worth of especially high carbon emissions. ³⁷ Considerations of climate justice thus seem to support replacement migration. ³⁸

While we cannot fully explore these considerations, we now briefly explain why our version of the public goods argument remains defensible even when taking into account the environmental costs of children. To see this, notice that there are two distinct limits that climate justice imposes on the public goods argument.

First, the public goods argument cannot justify a taxpayer obligation to share costs with local parents if the carbon emissions of local procreation cause impermissible harm to third parties, or, in other words, if these emission push the state beyond its ‘sustainability frontier’ (Caney, 2020). Proponents of the fairness argument fully acknowledge that there can be no obligation to share the costs of productive activities if these are morally impermissible. ³⁹ The question, therefore, is whether there is nevertheless room for some amount of local procreation within the state’s sustainability frontier. That is not only an empirical question but a normative one because different theories of climate justice will identify the sustainability frontier in more or less accommodating ways. However, we believe an affirmative answer to this question is plausible – that is, it is plausible that climate justice does not require that higher-emitting states cease local procreation altogether. ⁴⁰

There is a second limit that climate justice may impose on the public goods argument. Even if local parents have children within the sustainability frontier, it may be that the emissions caused by local procreation threaten to render it a net cost to non-parents (e.g. because those emissions need to be offset by reducing consumption). In reply to this, we repeat the points made in the previous section. So long as parents produce benefits for non-parents that are worth their costs by reference to the interest-sensitive baseline, non-parents have an obligation to share the costs of children, including their environmental costs. This will mean, for example, that non-parents lack a complaint against sharing the costs of policies that invest in greener technologies in order to make individuals’ per-capita lifetime environmental footprint smaller, so that a somewhat larger population can live within the sustainability frontier, as opposed to policies that foster economic growth and more per capita consumption that would be compatible with the sustainability frontier if a smaller population existed. It seems realistic to assume that a limited amount of local procreation (e.g. one child per family) is better for non-parents than no local procreation, even taking into account the environmental costs of children that non-parents will have to share. The fact that non-parents could benefit even more if there were fewer local children constitutes no objection if the compelling interest in procreative choice protects the option of having one child.

Conclusion

The public goods argument is the most influential argument in the debate over whether the costs of child-rearing should be shared between the taxpayer and the family. Yet it has not been examined against the challenge that replacement migration is available as an alternative source of the public goods child-rearing provides. We have argued, through an in-depth analysis of that argument, that importantly different versions of it can be formulated, and that some plausible ones remain robust against this challenge. While improving our understanding of the normative underpinnings of family policies, this conclusion improves our understanding of the nature of the public goods argument itself, and may thus advance our ability to resolve other policy controversies in which it plays a role.