The sins of the parents: Conceptualizing adult-oriented reforms to family benefits

Social policies for children: a typology

In a 2020 article, Mary Daly calls for better classifications and a more nuanced understanding of policy approaches to children. She notes that child-focused policy in Europe has been motivated by both commitments around children’s human rights and a social investment rationale (Morel and Palier, 2011), suggesting a range of motivations underpinning social policy for children (Daly, 2020). Daly’s proposed typology distinguishes three broad approaches: children-oriented, family-oriented and childhood-oriented policy (see Supplemental Appendix Table A1). Of most relevance here is the distinction between children- and family-oriented policy.

Children-oriented policies recognize children as individuals with human rights and entitlements (Skevik, 2003). These policies confer recognition and entitlements on children in distinction to adults (Daly, 2010). Within this category, Daly identifies a sub-category of child-centred policy which sees children not only as holding rights but as subjects capable of identifying and articulating their needs; child-centred policy therefore aims to secure children’s participation, providing resources to this end. This is the ‘highest and hardest level to achieve’ (Daly, 2020: 356) and is rare within Europe, thereby serving an important normative function within the typology. More common within the children-oriented group are child-focused approaches which provide children with recognition and rights while treating them as objects rather than subjects of policy (Daly, 2020). The goal of child-focused policies is the recognition and resourcing of children. In terms of social security policy, we consider child benefit to be a children-oriented (child-focused) policy, if paid at the same rate per child of a given age regardless of family income, size or household structure.

In contrast to children-oriented policy, family-oriented approaches locate children within the familial context. The focus of family-oriented policies is on the adult world rather than children (Daly, 2020). Sufficiency of family income is the core goal; children’s welfare is addressed by ‘resourcing the family or parents and/or seeking to affect parental behaviour and institutions’ (Daly, 2020: 356). In relation to social security policy, an obvious example is a means-tested child benefit for which eligibility and/or payment size depends on household resources.

Finally, childhood-oriented policies consider children as an age group; children are resourced ‘as members of a generation, rather than as having value and claims as persons’ (Daly, 2020: 356). The emphasis is on the temporal construct of childhood, aligning with social investment arguments for spending on the early years of life (Esping-Andersen, 2002). Daly sees childhood-oriented policy as being primarily concerned with investment in services, such as education (Kuitto, 2016), making this category less relevant to our analysis. Nonetheless, childhood-oriented policy could be an important lens when considering, for example, the relative allocation of social security spending to children compared to spending on pensioners or working-age adults.

Daly’s typology is one of the most recent and significant attempts to make sense of different approaches to social policies for children (see also Palme and Heimer, 2021). We deploy it as our starting point to explore in further detail the rationale behind differences in the treatment of children within social security systems in different European contexts. We begin in the next section by testing its usefulness in relation to social security through a consideration of recent developments in the UK, chosen because of our in-depth knowledge of this system and because recent policy changes make it an interesting case. How far do the children-oriented and family-oriented policy categories help us categorize this system and make sense of these changes? Our focus throughout is on child-contingent financial support: support that the child or her parents receive in the form of financial transfers because of the child’s presence in the household.

An application of Daly’s typology: child-contingent support in the UK

Financial support for children has contained a means-tested element in the UK for many years, with the importance and reach of means-testing growing from the late 1990s onwards (Hills, 2014). Presently, there are two main aspects to child-contingent support: Child Benefit and the ‘child element’ of Universal Credit/tax credits. There are also some policies that vary across the four UK nations, as well as additional streams of support targeted at families with very young children.

Child Benefit is payable for all children in the UK but is taxed back from higher rate taxpayers; from 2024 families with one earner on at least £80,000 lose the full amount. The rate is higher for first-borns. As Child Benefit is dependent on family income and on parity it fits the family-oriented policy category.

The child element of Universal Credit (UC) (previously Working Tax Credit and Child Tax Credit, now being phased out) is the main system of means-tested support for households with no or low earnings in the UK, and core to financial support for low-income families; in 2024 the maximum per-child amount (which is a flat rate) was worth over 2.5 times the value of Child Benefit for a first child and nearly four times the value for a second child. Because it is income-contingent, the child element also fits within the category of family-oriented policy. So does the new Scottish Child Payment (SCP), introduced by the devolved Scottish administration in 2021 and paid per child for all children under 16 in receipt of UC or tax credits in Scotland.

The 2017 introduction of a two-child limit in UC/tax credits means that the child element is now only paid for two children in the family: third and subsequent children born after April 2017 have no entitlement (Sefton et al., 2019). Support is therefore conditioned not only on household means but on the number of older siblings a child has. This policy marks a significant break with the basic principles of the post-war Beveridgean welfare state: that of linking entitlement to assessed need (Stewart et al., 2023). Despite this shift, child-contingent support paid through UC and tax credits remains family-oriented within Daly’s typology.

Education Maintenance Allowances are paid weekly in Scotland, Wales and Northern Ireland to young people aged 16–19 who are in full-time non advanced education (Middleton et al., 2004). These allowances are contingent on family income but, because they are paid directly to the young person, we consider them children-oriented.

Across the UK some additional support for babies and young children is available to families during pregnancy and early childhood. The support combines cash payments (e.g. the Sure Start Maternity Grant) with a voucher or payment card system restricted to healthy foods (e.g. Healthy Start Vouchers for pregnant mothers and very young children). Details vary across the devolved nations (with equivalent policies in Scotland often being more generous), but in all cases payments go to parents/carers in families in receipt of other means-tested benefits. Under Daly’s typology, all aspects of this support across the four nations can be classified as family-oriented. Payments are dependent on family circumstances (always on household income and sometimes on the presence of other children, as with the maternity grant, which is paid only for first-borns in most of the UK and is more generous for first-borns in Scotland). Nothing here meets the definition of children-oriented support.

Overall, with the exception of the EMA, child-contingent support in the UK falls squarely under the family-oriented category in Daly’s framework. Yet this seems unsatisfactory. Intuitively, there are substantive differences in policy orientation encapsulated by the policies presented, but these are obscured within the typology. For example, there is no difference in the categorization of the Scottish Child Payment (a flat-rate means-tested payment for all children aged under 16) and the child element of Universal Credit since 2017 (flat-rate and means-tested but only covering the first two children in the family). Both are conditioned on household means and paid to adults, so are grouped together as family-oriented. But the fact that one is restricted to two children per family represents a marked departure in approach to support for children which is important to capture. There are also questions about whether Healthy Start Vouchers – restricted to certain purchases and paid via vouchers or payment card – and the Sure Start Maternity Grant – paid in cash – should be in the same category.

The UK system clearly falls short of Daly’s ideal of policy that centres children’s rights, and Daly’s framework is valuable in helping us to identify this. But, as currently conceptualized, the family-oriented category covers a diverse and wide range of policies, underpinned by distinctive and sometimes contrasting implicit goals. With this in mind, and embracing the overarching utility of Daly’s framework as a starting point, we propose a refinement.

Refining Daly’s framework: diversity within family-oriented policy approaches

Family-oriented policies locate children within the familial context, in contrast to children-oriented policy in which the child is the focal point. Daly argues that family-oriented policies focus on the adult world rather than directly on children (Daly, 2020). A core aim of the distinction is to show how little social policy succeeds in truly centring the child, or even aims to do so. But while this point is important, the typology leaves the family-oriented category doing a lot of work. Policy can treat children as situated within the family for a variety of reasons.

As resources are known to be shared within the family (albeit imperfectly), social policies that aim to ensure children’s needs are met do well to take account of family circumstances. Targeting household income is an efficient way to reduce poverty using limited resources (Marx et al., 2013). While not children-oriented in Daly’s framing, policy that is primarily concerned with ensuring families have adequate resources to provide for their children may still be argued to be promoting and protecting a child’s right to a sufficient standard of living.

We start our refinement of the family-oriented category by labelling such policies as ‘needs-oriented’ (see Table 1). We consider policies to be needs-oriented if they take account of household needs and/or resources in determining either eligibility or the size of the payment. This includes benefits that are means-tested and/or that pay higher amounts for children in lone parent families. We also include benefits where support per child increases with family size (albeit with some exceptions described below). There is strong empirical evidence that the risk of poverty rises in larger families, due to both additional needs and greater employment constraints (Redmond, 2000; Bradshaw, 2006; Stewart et al., 2023). Increasing per-child benefit payments in larger families is therefore an effective way to address the greater needs of these families. While there are potentially economies of scale that mean earlier children need greater support – clothes and equipment can be handed down – such scale economies are dwarfed by the additional demands and constraints on the household (Hirsch et al., 2021). We make an exception for benefit structures that offer a higher amount to the first-born child, with flat-rate per-child benefits thereafter; we allow this to count as needs-oriented, on the grounds that the cost of transition to a first child may reasonably be considered more expensive than to a second or third child. There is also some evidence, e.g. regarding the decision to increase Child Benefit for first-born children in the UK in 1991, that raising rates for first-borns can be used as a way to provide some additional support for all families within tight budget constraints (Bennett and Dornan, 2006). In sum, then, where benefit payment structures appear to respond to increased needs within the household because of the ordering of the child – encompassing both greater payments to first-born children, and greater payments to second and subsequent children – we classify these policies as needs-orientated (though see below for one further exception). The logic for each approach is distinctive but the material consequences of designing the policy in these ways is to respond to need.

OPEN IN VIEWER

Table 1.

An extended typology of social security support for children, building on Daly (2020).

	Family-oriented		Child-oriented
	Needs-oriented	Adult-behaviour- oriented
Primary focus	Child & adults	Adults	Child
Engagement with children	Indirect, child in the foreground	Indirect, child in the background	Direct – recognition of children as a distinct group with needs
The entitlement	Income support adjusted to family needs	Income support conditional on adult behaviour	Resources for the child
The desired outcome	Sufficiency of family income	Changing adult behaviour & decisions	Recognition and resourcing of children
Social security policy	Child benefits vary according to household needs in one or more of these ways: • Same value per child but means-tested	Child benefits are linked to one or more of these: • Employment (conditional on work or job search)	Child benefits are universal and independent of family structure and size
	• Higher per child in larger families	• Partnership (e.g. higher if married)	Benefits paid directly to older children are also included here, even if means-tested
	• Higher in lone parent families	• Fertility (e.g. support capped by number of children, or policy explicitly pro-natalist)
	Weak needs-oriented policy:	Weak adult-oriented policy:
	• Tax allowances (unless support falls per child/capped)	• Benefits fall per child with family size (though not actually capped)
	• Social insurance	• Benefits paid as vouchers not cash
		• Child related conditions

The two-child limit is very clearly not a needs-oriented policy. This policy is focused on the adult world in a different way to other policies in the family-oriented category. It is not designed to ensure that a family has adequate resources; indeed it explicitly introduces a sizeable wedge between needs and support. The two-child limit is designed with adult decision-making in mind: it seeks to support them in some choices (having one or two children) but not in others.

We therefore create a new category of ‘adult-behaviour-oriented’ policy to capture this: support that is child-contingent (like all policies considered in this article) but is at the same time oriented towards influencing adult behaviour. Where this occurs, child benefits can be said to be open to manipulation in the interests of influencing adults’ behaviour or in order to achieve other policy goals beyond the immediate protection of children’s standard of living and well-being.

We identify three main types of adult decision that child-contingent policies may seek to influence: employment, partnership and fertility. First, we classify children’s benefits as adult-oriented if they are conditional on employment or job-seeking behaviour (e.g. if benefits are higher or only available for working parents, or if they are subject to withdrawal if parents do not comply with work search conditions). Second, we include benefits that reward partnership or marriage. Finally, we include benefits explicitly designed either to discourage or penalize fertility, like the two-child limit, or to encourage or reward fertility. We face a practical difficulty in relation to the latter category: how to distinguish policy designed to meet the additional needs of larger families from policy intended to incentivize more births? This is particularly challenging given that a detailed examination of policymakers’ underlying motivation is beyond the scope of the article. Our approach is to err on the side of needs-orientation, classifying benefits as pro-natalist only if they are labelled as such, or if the level of additional support seems clearly to go beyond a needs-based criterion. In practice, the rough rule of thumb is that if support for a third or subsequent child is at least double what a second child receives this is grouped as pro-natalist. We acknowledge that we are treading a fine line here and we advocate further research, including discourse analysis able to explore the motivation behind changes to benefit structures. We hope that our typology can promote and aid future analyses to this end.

Our revised classification raises several additional questions about the categorization of particular forms of support. The first regards the treatment of social insurance benefits and benefits restricted to taxpayers through tax allowances or non-refundable tax credits. The classic understanding of such benefits is in terms of horizontal equality – redistribution to households with relatively high current needs from others without dependent children at that time. As such, we classify them as weakly needs-oriented, as the support may not reach those on the lowest incomes.

The second question regards child-related conditionality: benefits made conditional on human capital conditions centred on children, such as school attendance. This type of conditionality has dominated debates about cash transfers for children, as conditional cash transfer (CCT) programmes have become more widespread in lower- and middle-income countries since the 1990s. The focus of these conditions is the child, setting them apart from benefits focused on adult behaviour with the child in the background. Despite this, adults are generally responsible for ensuring conditions are fulfilled and family resources suffer if this does not happen. We therefore add child-related conditions as a category under adult-behaviour-oriented policy but keep them distinct from other such policies. Part of our core argument is that conditionality is being extended to affect child benefits in ways that are not linked directly to children’s development in the way that CCT conditions are, and that this needs attention; thus the nature of conditions is important.

Finally, there is the question of support provided as restricted vouchers rather than cash, like England’s Healthy Start Vouchers. Such support attempts to shape adult decision-making by ensuring resources are spent on particular items such as food and milk. We argue that vouchers must therefore be classified as adult-behaviour-oriented policy. For parsimony we group them as a form of child-related conditionality.

Table 1 summarizes our new distinction within the family-oriented category and the type of policies included in each category.

While adult-behaviour-oriented policy is not, by definition, immediately concerned with children’s welfare, it may be argued to be taking an alternative route to the same goal. Indeed, policies rooted in behavioural change objectives are frequently justified on grounds that changing behaviours will improve outcomes; e.g. that having parents not in paid work is bad for children (Centre for Social Justice, 2006). But the primary focus is on adult behaviour, and it is accepted (albeit often implicitly) that children may be required to live in hardship if the adults in the household do not fulfil the expected conditions. Skevik (2003) writes that policy approaches towards lone parents in the UK in the late twentieth century wrestled with ‘the tension between securing need on the one hand and regulating behaviour on the other’, often favouring regulating behaviour: ‘The parent’s behaviour is in the spotlight, and the child is hidden behind the parents’ (p. 425). Our adult-oriented category helps us tease out policies where children are hidden or have lower visibility, situated in the background not the foreground when policies that affect them are designed.

It will be noted that we have so far used gender-neutral language, referring to ‘adults’ and to ‘parents’. This reflects our core interest in this article in how children are recognized and treated within benefit systems, and how far this is conditional on what the adults in their household do. However, much of what we are calling adult-behaviour-orientation is in reality deeply gendered. Frequently it is women’s fertility and employment that is scrutinized and problematized as requiring corrective action, and therefore in the vast majority of cases adult-behaviour-orientated policies could actually be described as ‘women’s-behaviour-orientated’ policies. In part because of the gendered allocation of paid and unpaid work, women are highly over-represented among lone parents, who have a much greater likelihood of needing financial support than coupled families. Thus the shift we identify in the UK in the treatment of children can also be viewed through a gender lens as a shift away from a benefit system that provided lone mothers with a minimum standard of living for themselves and their families towards one that penalizes lone mothers for their dependency (Gazso and McDaniel, 2010; Ruspini, 1999). More generally, the shape of social policies that affect children are deeply intertwined with assumptions about gender roles and obligations (Saraceno, 2012). We recognize this and see great potential for a future gendered exploration of the adult-oriented category.

Revisiting child-contingent support in the UK using the extended typology

Our extension to Daly’s framework is based on the premise that structuring benefits with the aim of ensuring that children have adequate resources, regardless of their family circumstances, is conceptually distinct from structuring benefits to change, reward or punish adult behaviour. We believe this is a vital adaptation to Daly’s typology. We now test out how this extended typology looks in relation to recent UK policy development. Over the period since the birth of the British welfare state we find markedly little in the way of child-oriented policy (Timmins, 2017). Indeed, we identify just three policies in the last 75 years that fall into this category, all of them short-lived: Child Benefit (1979–1991, the period when it was universal and flat-rate); the Health in Pregnancy Grant (2009–2011); and the Education Maintenance Grant, discussed above (1999–2011, though continuing in Scotland, Wales and Northern Ireland).

By contrast, there is much that qualifies as needs-oriented. The original design of post-war Family Allowances fits this category as allowances channelled more support to larger families, paying benefits only for second and subsequent children until they were replaced with universal Child Benefit in 1979. New Labour’s child tax credit system, built on the principle of ‘progressive universalism’, was also clearly needs-oriented, along with a number of other smaller policies including Sure Start Maternity Grant.

Some adult-behaviour-oriented policy can be found historically: for example, the Family Income Supplement (FIS) was introduced in 1971 as a means-tested wage supplement available only to those in work (later scrapped as part of child tax credit reforms).

Overall, the system as it stood in 2010 could be classified as a mix of child- and needs-oriented policy, with needs-orientation dominant. Over the following decade, however, we see a very clear shift towards adult-behaviour-orientation. Indeed, remarkably, by 2020, we suggest that all UK-wide social security policies for children can be categorized as adult-behaviour-oriented. Existing child- and needs-oriented policies were either scrapped, changed or restricted in such a way that they became adult-behaviour-oriented. Scotland, as already highlighted, remains an exception. We identify three key aspects to this shift.

First and most obvious is the 2017 introduction of the two-child limit. The extent to which this policy reflects a focus on adult’s (and particularly women’s) decision-making, with children’s needs pushed into the background, is encapsulated by the so-called ‘rape clause’, which provides an exemption for children conceived as a result of rape or coercion, if the mother has a social worker or medical professional testify to this fact, and if she no longer lives with the perpetrator (Hobson, 2022). Similarly, the limit is only applied to one baby in a multiple birth.

Second, an overall benefit cap was introduced in 2013, placing a limit on the total support available to households without an adult working enough to earn the equivalent of 16 h per week at minimum wage and not in receipt of disability benefits (Kennedy, 2023). The cap was lowered in 2013 and has only been uprated with inflation once across the decade, meaning growing numbers of households are affected, generally pushed into the cap by support with housing costs. Nearly nine in ten affected households have children; two-thirds are lone parents. The cap means that children in families without paid work can be denied the child element or Child Benefit to which they would otherwise be entitled. For example, a new baby born as a third child into a capped household will in effect receive no financial support unless and until an adult enters work.

Beyond the operation of the benefit cap, children’s benefits are not subject to adult work-related conditionality. However, the replacement of six means-tested benefits, including child tax credit and unemployment and disability benefits, with Universal Credit, has arguably reduced the visibility of children within the UK benefits system in a way which blurs the line between needs- and adult-orientation. Penalties linked to failure to meet conditions are calculated based only on adult not child allowances within UC. But the total payment is treated as one, and the child element is not explicitly protected; if entitlement is to a low amount of support because of other earnings, a sanction would not discriminate based on whether that amount was intended for children.

Finally, the only form of means-tested support available equally to all children in eligible families is paid not in cash but as restricted vouchers for milk, fruit and vegetables. Healthy Start Vouchers paid during pregnancy and the first few years of life are the only part of the system not subject to either the two-child limit or benefit cap. Yet the approach is clearly rooted in assumptions about adult behaviour; hence we classify these too as adult-behaviour-oriented policy.

Our analysis suggests that the two-child limit is part of a wider shift in the treatment of child-contingent support in social security policy by the UK government. Children have become less visible within the policy framework, with social security payments for children increasingly open to manipulation to achieve policy goals focused on adult’s (often women’s) decision-making. This important shift is made visible by our refinement to Daly’s framework: the separation we propose within Daly’s family-oriented category enables a more finely-grained analysis of social security policies affecting children.

The extended typology in comparative perspective

Having explored the utility of our refined typology within the UK, we turn to look at child benefits in other European countries. Is the growing domination of adult-behaviour-orientation a more general phenomenon in social security policy for children, or does it reflect UK exceptionalism?

To conduct this comparative analysis, we draw primarily on EUROMOD country reports (EUROMOD, 2022). EUROMOD is a tax-benefit microsimulation model for the European Union which brings together coding of the policy systems of all Member States with national microdata. While intended primarily as a resource for EUROMOD users, the country reports are invaluable in providing a comprehensive, consistent and up-to-date source on social security policies in each country. We use the 2018–2021 reports that accompanied the release of EUROMOD I14.0+, extracting details for 2021.

We include all benefits which are child-contingent, that is, they are dependent on the presence of children in the household. Maternity, paternity and parental leave benefits are excluded. Social assistance and unemployment benefit are excluded from our main results, but to ensure we are covering the different ways that countries structure financial support for families we discuss them briefly in a separate analysis.

Table 2 provides a broad picture of the orientation of child benefit systems in the 27 EU Member States plus the UK (with Scotland shown separately). Regular child benefits or family allowances (X) are represented separately from birth grants (BG) and from one-off or annual payments to support school expenses (SG). The columns show where countries have policies that fit the categorization developed above, such that reading across the rows conveys a broad story about any given country’s child benefits system. For example, we observe that Austria pays universal child benefits equal in size regardless of position in the family (though increasing as children age) and there is a universal school grant. Austria also pays means-tested supplements for young children in the household, ¹ while two further supplements offer additional support for later-born children. The Austrian system therefore combines elements of child-oriented and needs-oriented policy, with no examples of adult-behaviour-oriented policy.

OPEN IN VIEWER

Table 2.

Classifying child benefits by orientation, EU countries plus UK.

Country	Child-oriented	Needs-oriented								Adult behaviour-oriented
	Universal per-child payment independent of family circumstances	Higher for first-born	Higher for later-born	Higher for lone parents	Means-tested	Means-tested & higher for first-borns	Means-tested & higher for later-born	Means-tested & higher for lone parents	Tax allowances or social insurance benefits	Anti-natal (favours smaller families)	Pro-natal (e.g. generous large family bonus)	Favours working families	Benefits subject to adult conditionality	Benefits subject to child-focused conditionality	Overall classification of main benefits
Austria	X, SG		X		X(1–3s)		X								C + N
Belgium	X	BG			SG		X							SG	C + N
Bulgaria	SG				SG					X(f2), BG(f2), TA(c3)				X, SG	A
Croatia	BG				X			X	TA		X				N + A
Cyprus				BG			X	X		X(f4 c4)		BG			N + A
Czechia									TA	BG(f1 c2)		X			N + A
Denmark	X			X	X										C + N
England/UK						X, BG(c1)				X(c2)		X	X	V(U4s)	A
Estonia	BG		X, BG	X					TA		X				N + A
Finland			X					X							N
France				X	X(U3s), BG, SG		X		TA						N
Germany			X		X										N
Greece					BG		X		SI						N
Hungary			X		X			X						BG	N
Ireland	X				SG			X	TA			X			C + N + A
Italy				BG	X		X								N
Latvia	BG		X						TA		X				N + A
Lithuania	X, BG		X		X										C + N
Luxembourg	X, SG														C
Malta	X, BG				X										C + N
Netherlands	X					X									C + N
Poland	SG				X, BG, SG		X	X			X				N + A
Portugal					X		X	X	TA					SG	N
Romania	X									X(c4), TA(c4)				X	C + A
Scotland					X, SG	BG									N
Slovak Republic	X, SG				X (secondary school), SG					BG(f3)				SG	C + N
Slovenia	BG		X				X		TA(f1)		X				N + A
Spain							BG	BG	TA	X(c3/4)					N + A
Sweden	X		X												C + N

Further down the table we see a very different picture in Czechia. There the main child allowance is means-tested, but it also increases in value (by close to 50%) if at least one person in the household has some income from employment (at least equal to the Minimum Living Standard of a single person). This makes it an adult-behaviour-oriented policy. However, Czechia also has a refundable tax credit which rises per child up to three children and is then worth the same per child. Overall, Czechia is classified as combining needs-oriented and adult-behaviour-oriented policy. A third example, Latvia, has been classified as combining needs- and adult-behaviour-orientation because of a pro-natalist benefit design. Compared to a one-child family, a two-child family in Latvia receives four times the amount of family benefit, a three-child family 11 times and a four-child family 16 times the amount. This is an example of a policy we classify as going beyond needs-orientation in favouring larger families.

As a summary, Figure 1 shows the overall categorization of each country (29 in all, with Scotland treated separately). We include here the main child benefits, not birth grants or school grants. Only one country, Luxembourg, has all benefits falling into the child-oriented category: all child benefits in Luxembourg are paid for each child irrespective of wider family structure or circumstances. There are eight countries where all child benefits are classified as needs-oriented and a further eight which combine child- and needs-orientation. That leaves 12 countries classified as having some element of adult-behaviour-orientation in their policies. The shading on the figure shows that five of these countries are classified in this way because they combine needs-oriented with pro-natalist policies; each provides particularly generous support for larger families. These are mainly countries in Eastern Europe or the Balkans (Croatia, Estonia, Latvia, Poland and Greece). We distinguish this group because the consequences for children are potentially very different than for other forms of adult-oriented policy: a policy may not be motivated by children’s rights or needs, but children are likely nonetheless to be beneficiaries.OPEN IN VIEWER

Besides the pro-natalist group, seven other countries show some element of adult-behaviour-orientation. In most cases, these policies are just one aspect of child support, alongside other policies that are needs- or child-oriented or both. We identify only two countries – the UK (excluding Scotland) and Bulgaria – where all child benefit policies fall into the adult-behaviour-oriented category. The UK, then, is highly unusual in having a system of financial support for children that is exclusively built on an adult-behaviour-orientation. Bulgaria caps tax allowances at three and has a means-tested child benefit system that is most generous to the second child with support per child then falling, especially from the fourth child onwards. Child-related (human capital) conditionality is also attached to all the main child benefits in Bulgaria.

Adult-behaviour-oriented policies across Europe

Setting aside pro-natal benefit design, the most common reason policies are classified as adult-behaviour-oriented is their less generous treatment of later-born children in the family. This is true of four countries other than the UK: Cyprus, Spain, Romania and Bulgaria. In all four of these countries, support is only paid for a certain number of children, as with the two-child limit, although nowhere else is the limit set at two. In Cyprus and Spain the limit applies to the main form of child benefit available but both are somewhat complex cases. In Cyprus the main means-tested child benefit provides no extra support for children beyond four, though the income eligibility threshold continues to rise with additional children, and the benefit is most generous to third and fourth children. In Spain, there is no longer a system of child benefit or family allowances, but there is a Guaranteed Minimum Income. This is paid for up to four dependents in the family – meaning it is effectively capped at three children in coupled households and four in lone parent households. Regional governments in Spain also have their own additional schemes, which in most cases are capped at six family members or fewer. In Romania means-tested child benefits are also capped at four children, but there is a universal child benefit which is more than double the value of means-tested support. In addition, tax allowances in Bulgaria are capped at three children, while child benefit is means-tested and most generous to the second child, then falling in value.

What of the other adult-behaviour-oriented categories? Two countries have benefits that favour working families. Czechia was discussed above. In addition, alongside universal child benefits and means-tested lone parent support, Ireland provides a working family payment to families with children working at least 38 h/fortnight (couples can combine their hours). While there is also support for children within Jobseekers Benefit, the rates per child in the Working Family Payment are more generous – meaning low-income working families are eligible for more support per child than families without work.

Finally, the UK appears to be the only country in which child benefits are effectively subject (via the benefit cap) to adult work-related conditionality. However, we do observe human capital-focused conditionality in Romania and Bulgaria. In Romania children must be attending school to receive means-tested support allowance. In Bulgaria, child benefit can be stopped if a pre-school child misses more than 3 days or a school-age child more than five school hours without valid reasons.

In focusing on the main form of social security support for children, the analysis so far excludes social assistance benefits, often the last resort safety-net for those on the lowest incomes. If we include these benefits, the picture changes somewhat, and in interesting ways. Many countries now show some element of adult-orientation. This is mostly in the shape of work requirements, as might be expected (Watts and Fitzpatrick, 2018). But in addition, social assistance varies in its treatment of extra children, indicating a different approach to children in families with the greatest level of needs, compared to the approach to children in general. Table 3 sums up the treatment by family size for each country across different types of benefit – universal, means-tested and social assistance (for the full picture see Supplemental Appendix Table A2). This reveals a group of countries in which children are less visible within social assistance support than within other, more generalized, forms of provision. These countries provide universal and/or means-tested benefits which are the same or rising in family size, but support for children in social assistance falls or is capped. Strikingly, universal benefits are never capped. Standard means-tested benefits are capped in four countries and fall in size in two others. But social assistance benefits are effectively capped in nine countries in addition to Spain and the UK and fall with family size in one other. Three countries – Belgium, the Netherlands and Poland – appear to provide no recognition of children, or only of one child, within social assistance while having universal and/or means-tested provision for all. This could be because these latter benefits are available to help with the costs of children. But families requiring social assistance are by definition assessed as having particular needs, related to very low levels of household income.

OPEN IN VIEWER

Table 3.

Support per child as family size grows, by type of benefit.

	Universal benefits	Means-tested benefits	Social assistance
Austria	SAME	RISES	SAME
Bulgaria	--	RISES & FALLS	SAME
Croatia	--	RISES	SAME
Cyprus	--	RISES & CAPPED (4)	SAME
Czechia	--	SAME	SAME
Estonia	RISES	--	SAME
France	--	RISES	SAME
Germany	RISES	RISES	SAME
Hungary	RISES	SAME	SAME
Ireland	SAME	RISES	SAME
Latvia	RISES	--	SAME
Luxembourg	SAME	--	SAME
Malta	SAME	SAME	SAME
Portugal	--	RISES	SAME
Slovenia	RISES	RISES	SAME
Sweden	RISES	--	SAME
Belgium	SAME	RISES	CAPPED (1)
Denmark	SAME	SAME	CAPPED (Complex)
Finland	RISES	--	CAPPED (3)
Greece	--	RISES	CAPPED (5)
Italy	--	RISES	CAPPED (3.5/5.5)
Lithuania	SAME	RISES	CAPPED (3)
Netherlands	SAME	SAMEx	CAPPED (0)
Poland	SAME	RISES	CAPPED (0)
Romania	SAME	CAPPED (4)	FALLS
Slovak Republic	SAME	SAME	CAPPED (5)
Spain	--	CAPPED (3/4)	<--
UK	SAME*x	CAPPED (2)	<--

This analysis reveals the extent and nature of adult-behaviour-oriented policies pursued across diverse European contexts. Ours is the first attempt to explore this orientation, both within UK recent social security history and in comparative perspective. We propose that more research is needed, both to better understand the motivations behind an adult-behaviour-oriented approach, and also to document its consequences – in particular the impact on affected children.