1. Introduction

When COVID-19 struck it changed the circumstances of the entire population. Countries had to quickly adapt and expand their support offer to their populations, transforming the recourse that individuals could make to family, state and market for resources and support. Children were especially affected with widespread closures of schools and early childhood education and care (ECEC) facilities, closures that in many countries lasted for a long time. This saw children largely reliant on the resources of their parents and families for all aspects of their material welfare as well as for their broader development and care. We still know too little about the details of the policies adopted in different countries for children and families as well as how to interpret them. Such relative ignorance is layered onto a more longstanding knowledge gap about how to understand the place of children in the welfare state. Reviewing the policy actions taken by the EU member states and the UK from the pandemic’s onset to December 2020, this article reveals the policy packages involved. Theoretically it builds on literature on child and family policy and engages debates and concepts on how we should conceptualise the field and assess how proactive and child-centred countries were. The resulting analysis sheds light on important questions about the priorities and preferred policy levers as well as cross-national patterning.

The article’s research questions probe the extent to which and how the measures taken during the pandemic prioritised children and what ‘child -centredness’ means in this context. It answers these questions from a policy database assembled specifically on child- and family-related policies for the 9 months of the pandemic in 2020. Five fields of policy are covered: ECEC, primary education, additional income support, food support and support for parental care-giving (e.g. through paid leave). The article is organised as follows. The first section sets out the background and analytic framework. The next and largest section considers the research questions in detail presenting quantitative and qualitative evidence on relevant developments and assessing countries’ responses in terms of their degree of child -centredness. A conclusion brings the article to a close.

2. Analytic and empirical framework

The literature on children and the welfare state is not large – much more meagre in comparison to that on family policy, for example. This means among other things that there is little or no country typologising based on child-related policy, with comparisons largely relying on geographic or linguistic classifications (Richardson, Reference Richardson, Irving, Fenger and Hudson2015). Given that children are still relatively new in comparative welfare state research, much scholarship reads the treatment of children from an analysis centred on parents and family policy. It is imperative that we move beyond this if we are to understand welfare state policies and practices regarding children and families.

It is generally agreed that the broad thrust of relevant policy development involves some movement from family as the main locus of support for children to policy focusing more directly on children. The growth of a children’s rights perspective has been very important in this context (Lundy et al., Reference Lundy, Kilkelly and Byrne2013). Thinking in universal terms, this perspective makes a case for children’s claims and entitlements by virtue of their status and personhood. The 1989 United Nations Convention on the Rights of the Child (UNCRC) is the master statement here (United Nations, 1989). As well as setting out a new international legal and governance framework, the Convention opened eyes to the child’s individual personhood, status and welfare as well as their membership in society. The rights enunciated in the Convention for children centre upon the child’s well-being and development (encompassing health, social security, a sufficient standard of living and education) as well as four rights as general principles to be taken into account in all matters aimed at children.Footnote ¹ One helpful way of understanding the Convention from a social policy perspective is to see it as emphasising 3 Ps for children: provision, protection and participation (Palme and Heimer, Reference Palme and Heimer2021). Of course, there are limits on the extent to which the welfare state can provide children directly with income, and their participation too is circumscribed – with age and stage of development very important in both regards (Lansdown, Reference Lansdown2005). Nonetheless, the drum beat of the children’s rights movement is the recognition and resourcing of children in their own right by policy. A counter perspective is to view children as requiring the protection of families and other collective institutions. Child protection is a prevalent notion here. This views children and childhood as a condition of vulnerability and its ‘ask’ of the welfare state is primarily to ensure children’s protection from abuse and harm (which tends to be defined as violence, neglect or maltreatment) (Tisdall, Reference Tisdall2015). While not universally positive about family as a place of protection for children, the notion of child protection tends to see children’s welfare in terms of child protective services that are family oriented and social work in nature (Skevik, Reference Skevik2003; Gilbert et al., Reference Gilbert, Parton and Skivenes2011).

Juxtaposing the two concepts allows us to appreciate tensions and even dilemmas for social policy – to target children in their own right or through parental and family measures?

Taken together, they frame a double challenge to existing scholarship – including comparative welfare state research – on the one hand to develop an approach that recognises children as subjects rather than objects of policy, and on the other to recognise the presence of both push and pull factors in a children’s rights and/or family direction (Skevik, Reference Skevik2003).

Even a cursory glance at policy confirms the kind of movement involved. As a field, policy on children and family has been growing in the EU and the UK – expenditure on child and family benefits as a percentage of GDP grew by 5.1% between 2014 and 2019 compared to a reduction in social expenditure as a whole of 4.3% (Eurostat, 2023). The child as the recipient has achieved greater attention in regard to access to ECEC, with the EU setting targets for children’s participation in ECEC (the Barcelona targets), and ECEC and education deemed essential entitlements or guarantees for ‘disadvantaged’ children in the EU Child Guarantee (Council of the European Union, 2021). There are strong elements of social investment thinking here, highlighting the welfare state’s significance in children’s human capital development (Daly and Ferragina, Reference Daly and Ferragina2018; Adema et al., Reference Adema, Clarke, Thévenon, Nieuwenhuis and Van Lancker2020). ECEC services especially have grown strongly, although there are many national and sub-national variations (Scheiwe and Willekens, Reference Scheiwe and Willekens2009; Blum and Dobrotić, Reference Blum and Dobrotić2021). Under the rubrics of work-life balance and/or parenting support, measures for parents have also grown. This has seen both an expansion and diversification of parental leaves and greater use of conditionality around parental behavior and family type in the channeling and design of income support (Stewart et al., Reference Stewart, Patrick and Reeves2023). Some developments – such as the growth in ECEC services and even parental leaves – seek to serve both children and parents, blurring and complicating policies and their analysis.

How are we to understand the pattern of developments involved? Palme and Heimer (Reference Palme and Heimer2021)Footnote ² and Daly (Reference Daly2020) both seek to develop a theoretically-informed conceptual framework for classifying child-related developments in European welfare states. Palme and Heimer (Reference Palme and Heimer2021) build on the child/family dynamic in policy, differentiating between the treatment of children as actors and as policy objects, the former authoring child investment-oriented policies and the latter measures that are family-oriented. This two-fold differentiation can be accessed by investigating whether policies engage directly or indirectly with children. This is to recognise that there are goods that can be directed specifically towards the child (such as services and in-kind goods like food); and there are other goods (such as income) that tend to be resources directed to parents or other adults. Daly (Reference Daly2020) develops a three-part framework which differentiates between policy that treats children as members of families, members of a generation, or is child-centred in the sense of treating children as subjects with entitlements and rights which are recognised and resourced through public policy. The most significant elements of policy design reside for her in the fine detail of eligibility conditions, especially the unit of entitlement, the universal versus targeted basis of provision and the conditions governing access for parents and children. While good on design, Daly failed sufficiently to develop the child-centred perspective. We consider the pandemic as a good opportunity to think this through further and take up that challenge in several steps oriented to developing an analytic perspective on child-centredness.

By definition, child-centredness starts with the child and assesses policy’s ‘closeness to the child’ and degree of its concern with children’s welfare. Here it is conceptualised as encompassing support for family as well as that for children. Hence, it is taken forward in two dimensions: the degree of material support for families with children (relative generosity); the relative prioritisation of children vis-à-vis adults for services (child-directedness). Relative generosity is measured by two indicators: the average generosity of social protection spending on children and family in 2020 and the scope or coverage of measures taken during the particular nine-month period to materially support children and families (especially whether countries gave additional income support to families, whether they provided food support and whether they specifically put in place measures to financially support parental care-giving for children). The second dimension of child-centredness as used here concerns the degree of prioritisation of the child or child-directedness. This picks up on the tensions identified earlier in regard to the direct or indirect focus on children vis-à-vis adults or families and explicitly examines whether children’s needs and well-being received priority during the pandemic, focusing only on measures that could have prioritised children over adults or where a clear focusing is possible. This leads us to ECEC and school opening and closure. For this purpose, we look at the opening and closure policy and in particular whether ECEC and primary schools based access on children’s needs and/or those of their parents (as key workers, for example). We also compare the duration of full school closures with full workplace closures as this too gives a sense of relative prioritisation.

Using these five indicators, an index was created by assigning scores to each dimension. The scoring scheme was designed in such a way that each country could score between 0 to 1 on each dimension using a four-point scale from 1 (full), 0.67 (high), 0.33 (low) and 0 (none). All the dimensions were assigned equal weighting. Table A1 in the Appendix summarises the dimensions and their relative scoring.

Two research questions guide the empirical analysis:

1. What responses did the pandemic trigger for children and families in terms of a) scale and b) policy substance?

2. What was the main national and cross-national comparative patterning in relative performance in regard to child-centredness?

As mentioned, the empirical analysis focuses on the period of the pandemic from mid-March to December 2020.Footnote ³ Although countries varied on the exact timing of onset and intensity of different waves, this was a period when all of the 28 countries were actively trying to counter the pandemic and its ill-effects and likely to view the pandemic as an emergency. With a focus on this period we are essentially making an assessment of the prioritisation of children in ‘shock-responsive’ policymaking conditions. The added advantage of coverage over an extended period is that it serves to increase the comparability of evidence across countries (by downplaying particularities of timing and the severity of both infection and measures to counter it). In identifying and classifying policy actions, we count only those that were put in place or extended in this period, and pay particular attention to additional measures (especially in the case of income support).

Children are defined as those aged up to 17 years. We recognise that this is a diverse age grouping but it reflects policy thinking and has the merit of encapsulating a range of policies. The fact that we cover five policy fields provides an opportunity to think holistically about the life of the child and bring together some of the key concepts outlined earlier. They do not, of course, exhaust the policies affecting children. The data are drawn from a database constructed at the University of Oxford, with part funding from UNICEF Office of Research – Innocenti. This CPC-19 Database records detailed relevant policies in 40 countries on the basis of available data (noting that some granular data is still not available).Footnote ⁴ The database was compiled on data drawn from a range of sources, especially international databases (as outlined in Online Appendix Table A1). Government and other national sources were also important for this purpose as were country reports from work undertaken under the auspices of the EU, OECD and UNICEF and other cross-national research activities.

3. Empirical Findings

To begin answering the first question on the type of response triggered, expenditure is a good place to start. Table 1 presents Eurostat data on per capita expenditure on family and child social protection in 2019 and 2020, measured in Purchasing Power Standards (PPS), and the percentage change between the two years. Looking at scale first, there was a 5.7% increase in relevant spending on child and family policy between 2019 and 2020, albeit from a small base (just 2% on average in EU and UK as a percentage of GDP). This scale of increase is larger than the five-year average increase between 2014 and 2019 in the 28 countries – 4.6%. There are some country- and regional-specific patterns though.

Table 1.

Social Protection Expenditure on Family and Children per Capita in PPS 2019, 2020 and Percentage Change

Source: Eurostat (2023); OECD (2023).

First, all the countries increased their social protection spending on families and children in 2020 apart from Bulgaria, Greece, Ireland, Malta and the UK. Second, there appears to be an inverse relationship between the extent of additional expenditure and the level of expenditure, suggesting that some low-spending existing systems could not cope with the pandemic. Protecting families during the pandemic required more spending than usual in the Eastern European countries, especially Czechia, Lithuania, Poland, Romania and Slovakia. The corollary is also the case: notably some of the high-spending countries did not increase their spending especially. For example, some typically high spenders either maintained the spending status quo (Sweden) or undertook only slight increases (Denmark, Finland, France). Two factors suggest themselves as a possible explanation. First, some of these countries did not institute a full lockdown of employment and/or schools/ECEC facilities – hence, reducing the need for compensatory policy; and, second, their already generous and wide-ranging income support and service systems may have been able to automatically stabilise for the COVID-19 upheaval. The Continental European countries – also typically high spenders – notably increased their family and child policy spending during 2020 (Austria, Belgium, Germany, Luxembourg, the Netherlands).

These data need to be put in context though. In terms of relative degree of welfare effort – which we know was huge with COVID-19 giving rise to the widest prevalence of cash transfers ever (Gentilini et al., Reference Gentilini, Almenfi and Dale2020) – the evidence indicates that family and child policy was not a primary locus of spending support for the pandemic (Eurostat, 2023). For example, the relative increase in child and family policy spending as a proportion of total expenditure was dwarfed by that of unemployment benefits – which grew by 136% as a share of overall total expenditure in 2020 in the 27 EU member states as compared with 2019, while the overall share occupied by family and child policy dropped by 3.6%. Although unemployment was far and away the main benefit that countries relied on during COVID-19, nations varied greatly in terms of the percentage increase in spending on unemployment benefits. The most important thing from our point of view, however, is that family and child policy was relatively de-emphasised as a response to the pandemic.

There is another way to assess volume or scale of effort as well – in terms of the number of policy fields of the five considered. Table 2 shows the overall scope of activity in terms of the number of fields engaged. While no country was completely inactive, only two (Cyprus and the Netherlands) took action in one field only, whereas France, Portugal and Sweden stand out as the only countries that adopted measures in all of the five policy areas. The other countries cross the spectrum in how many fields they focused their activities on.

Table 2.

Countries Compared in Terms of Number of Fields of Activity

But volume and distribution of expenditure are relatively crude as guides, not least because they fail to contextualise the spending changes in policy substance terms.

Supporting data in Table 3 show the relative popularity of the respective policy fields.

Table 3.

National Incidence of Activity in the Five Policy Fields

* No measures to protect in the event of closure or income loss.

Policies to enable parental care through paid parental leave or the modification of other leaves or benefits (such as sickness benefits) for this purpose was the most favoured policy field. Only five countries failed to give parents some additional compensation for home-based childcare during COVID-19 – Croatia, Ireland, the Netherlands and the UK made no changes to leave and Spain made leave possible for parents only on an unpaid basis. This was a minority occurrence, though, as 16 out of the 28 countries (across welfare state types and regions) introduced a new COVID-19 specific paid parental leave and 13 (sometimes in addition to the new measure) modified existing leave schemes to compensate parental care-giving during the pandemic (Online Appendix Table A2). Across countries, the exigencies that typically occasioned a response through parental-leave policy were either a child becoming ill with COVID-19 or a child/parent becoming unable to avail of a relevant education or care service through either exclusion of the child or the shutdown of the service. The details varied in practice. Age of the child was an important cross-national variation and in general, only dual-earner couples and lone parents were considered eligible for the measures. Generosity too varied, with Austria, France, Greece and Luxembourg having the highest wage replacement rate (Baptista et al. Reference Baptista, Marlier, Spasova, Peña-Casas, Fronteddu, Ghailani, Sabato and Regazzoni2021). Most widely, the leave was paid at a rate between 50 and 90% of pre-COVID-19 salary, with two countries specifying the minimum wage as a floor (e.g. Portugal and Slovenia) (ibid).

Support of parents with care-giving was followed in popularity by adjustments to ECEC services. The key issue here is the nature of the closure and whether children were prioritised for access (some version of which happened in 21 countries). The evidence (Online Appendix Table A3) indicates that, while seven countries fully closed the facilities, what really separates countries is whether they tried to protect certain sectors of the population from closure. A brief overview paints the broad picture. First, apart from Finland and Sweden, very few countries made decisions on opening or closing on the basis of children’s needs. Just three specified that (some) children should have continued access: Denmark, Ireland and the UK kept the facilities open for vulnerable children (the definition of which varied in each). There was, then, little or no direct prioritising of children. Rather, across countries parental need – in the sense especially of whether parents worked in key or essential sectors – was the dominant criterion for opening and access. Secondly, countries divide in terms of which parents they targeted for support (and indirectly the scale of opening). The broadest partial opening was when countries kept the facilities open for all working parents. This was the case in Austria, Belgium, Croatia, Denmark, Estonia, Greece, Hungary and Latvia. Prioritising access by key workers only (defined differently but usually including sectors such as health and care, police and food supply) was marginally more popular. The countries that adopted this approach were mainly from Continental or Eastern Europe: Czechia, France, Germany, Lithuania, Luxembourg, the Netherlands, Portugal, Slovakia and Slovenia. The remaining seven countries protected neither children nor their parents, closing down the ECEC facilities for at least four months between March and June 2020 without any quarter to child or parental need. These were all Eastern European or Mediterranean countries (Bulgaria, Cyprus, Italy, Malta, Poland, Romania and Spain) (Blum and Dobrotić, Reference Blum and Dobrotić2021).

The next most popular measure was food support, through such measures as ring-fenced grants for local authorities for food provision, food-related cash payments and benefits to enable and directly resource children’s nutrition by replacing school meals or giving vouchers/stamps. Table 3 shows this to be a popular field of activity with some action in 18 countries.Footnote ⁵ What makes a big difference to activity here is the pre-pandemic policy scenario. Indeed, no country introduced this form of assistance during the pandemic if it did not have it previously (Online Appendix Table A4). But existing provision did not guarantee continuation as five countries actually discontinued their pre-pandemic food provision. This was mainly due to the difficulty of finding an alternative distribution means to schools, or a lack of centralised guidance on food support. The latter picks up on the fact that food provision or support is frequently a municipal level function. This makes for variation in the types of food-related support (e.g. snacks, meal bags for delivery, takeaways, vouchers or cash) and the nature of priorities across municipalities (Gentilini et al., Reference Gentilini, Almenfi and Dale2020). Such devolution and variation also make data gathering more difficult and the data less reliable for countries where this is a devolved function.

Table 4.

Child-centredness in the 28 Countries during the Pandemic (Z-scores)

Fourthly, the data in Table 3 indicate that providing families with additional money through the cash transfer system was less favoured compared with other policies examined: 12 of the 28 countries failed to use their income support system to offer additional financial support to families with children (apart from paid leave) (Online Appendix Table A5). This is a very mixed set of countries in terms of overall welfare system, general approach to family policy and geographical location. Belgium, Cyprus, Denmark, Estonia, Finland, Hungary, Ireland, the Netherlands, Poland, Romania, Slovakia and the UK are all countries included together here. The other 16 countries were more responsive but varied in both the content and extent of their actions. There are two important observations to be made overall. First, a new, or additional COVID-19-specific, child-related payment was introduced in seven countries (Austria, Bulgaria, Croatia, France, Germany, Italy and Spain). This suggests capacity for innovation and political will, although the extent and generosity of the measures varied greatly. Second, rather than new measures, many countries opted for adjusting their existing income support policies to extend generosity or coverage. A common form of adjustment was the provision of a one-off, top-up or bonus for families (Online Appendix Table A6). Other than the type and conditions of accessing the measures, their targeting was another lever used during the pandemic. Targeting to the neediest groups of children was far more widespread than universal responses (only Austria, Germany, Lithuania and Slovenia were the exceptions here), with income conditionality being a common targeting criterion. Another such criterion was the child’s age; and family size was also used for targeting purposes.

When it came to primary schools 16 countries just shut down completely as against just seven fully closing ECEC. Sweden was the only country to keep primary schools fully open (Finland did so just for ECEC). Furthermore, the prioritisation of children for the purposes of the opening/closure was more limited as compared with ECEC, with just two countries (Austria and Croatia) keeping the schools open for children of all working parents (seven did so for ECEC) and only three (Czechia, Denmark and the UK) prioritising ‘vulnerable children’ (Online Appendix Table A6). Compensatory mechanisms and indirect support – such as the provision of laptops, tablets and help with internet for distance learning and/or additional learning support – were undertaken in 18 countries out of the 27 that completely or partially closed schools. But overall, school-aged children received less priority for services than the younger age group.

Turning to the second research question, we now use the index outlined earlier to examine the combinations of policy activity within and across countries and thereby investigate variations in regard to the overall orientation and extent of child-centredness in social policy during COVID-19. To recap, child-centredness is measured on the basis of two indicators: relative generosity in supporting families materially (expenditure on family and child support per head during 2020 and whether countries took action in income support, food support and financially supporting parental care-giving) and degree of child-directedness (in regard to ECEC and educational provision in primary as well as the differences in the duration of school and workplace closures). To combine these different indicators, all of the raw data and values on each indicator were standardised by conversion to z-scores, which are calculated on the basis of the number of standard deviations a given data or point lies above or below the mean. Z-scores can be positive or negative, with a positive value indicating the score is above the mean and a negative value indicating it is below the mean. Table 4 presents the data, with countries organised in rank order on overall scoring.

The data on relative expenditure make clear that Luxembourg, Germany and Denmark were the most generous spenders on child and family policy in 2020, with Luxembourg far in advance of the others. These countries are followed in relative generosity by Austria, Finland, Sweden, Poland and Belgium. France and the UK were also slightly above average. Most countries spent less than average, with the smallest spenders being Malta, followed most notably by its regional neighbours (Greece, Portugal, Cyprus, Italy and Spain) and Eastern European countries (e.g. Bulgaria, Latvia, Slovakia, Croatia and Romania).

If we look at what we are calling the scope (second data column), the best performing countries here are the nine that instituted measures in all three fields. These are a very mixed group: Bulgaria, France, Italy, Latvia, Lithuania, Malta, Portugal, Slovenia and Sweden. Somewhat surprisingly perhaps, the Netherlands makes up the rear here as it took no action in either additional income support for families, food support or paid parental care-giving. Other countries that score relatively poorly here are Belgium and Denmark (which of the three fields only acted to support parental care-giving), Ireland and the UK (which acted only on food support), Croatia and Cyprus (which took action on additional income support or support for parental care-giving).

Turning to child directedness in ECEC, as indicated earlier, continuation of ECEC facilities during the pandemic prioritised parental rather than children’s needs or rights. Finland and Sweden come out on top here (in never implementing a lockdown). Denmark, Ireland and the UK also score strongly because they prioritised access for ‘disadvantaged’ children. Countries that scored very low here are mainly drawn from the Eastern European and Mediterranean regions. Child-directness in school opening shows generally similar results in that a) total school closure was the blanket response in the primary education sector with little effort made to protect children and b) better performance by the Nordic countries and the UK. Overall, the lack of attention to and protection of the development and relational needs of primary-school children is notable.

The final dimension – lockdown duration – confirms this. Relative scores here were derived on the basis of relative differences between the duration of full workplace closures and those of schools at all levels with longer school closing interpreted as de-prioritising children vis-à-vis adults. For the 28 countries as a whole, schools (from pre-primary to secondary) were closed for an average of 37 days longer than workplaces (excluding summer holidays). Eight countries score top marks here either because school lockdowns were shorter than or (almost) equal to those of workplaces or were never implemented. These are Austria, Belgium, Finland, France, Ireland, Italy, Luxembourg and Sweden. The remaining countries had either relatively long school closures (at least 21 days in excess of workplace closures) or very long school closures (from 40 days to 144 days longer than workplace closures). The countries with the biggest negative gaps were Romania, Poland, Latvia and Greece.

The final column in Table 4 ranks countries in terms of their child-centredness across indicators. Sweden and Luxembourg stand out for being at the top here. Sweden scores relatively strongly on all dimensions whereas Luxembourg is boosted especially by high relative expenditure and keeping school closures relatively short. Other high scorers are Denmark, Finland, France, Austria, Ireland and the UK. While these achieved their scores in somewhat different ways, they rarely scored below average (although Ireland is somewhat of an exception in this regard). While the first four of these countries are either relatively generous welfare states, and/or strongly service oriented and/or tried to keep the facilities for children open, Ireland and the UK are here primarily because they prioritised ‘needy’ children for access to educational and care facilities and also because (certainly in the case of Ireland) they kept school lockdowns relatively shorter than other countries. The Netherlands (placed 26^th) is a strong exception to the continental European pattern in terms of much lower than expected child-centredness. Its relatively abstentionist approach to supporting families and children is arguably in line with past policy which has promoted the privatisation of responsibility for childcare with relatively weak state support (Knijn, Reference Knijn, Ostner and Schmitt2008). The Eastern European and Baltic nations were somewhat split in their responses, but generally tended to be placed in the lower half of the country rankings. Particular weaknesses here centred on relatively long school closures and relative failures to prioritise children for access to ECEC and school and targeting additional income support to families to cope with the pandemic. Czechia and Lithuania were somewhat exceptional to this pattern, adopting a generally more supportive stance that reached across policy areas and raising their child and family policy expenditure considerably during 2020 (especially Lithuania). Compared to other regions, Eastern European countries have shown diversities in their family policy constellations since the collapse of the communist-led regime, and it is therefore difficult to find one particular feature defining them as a whole (Szelewa, Reference Szelewa, Johnson, Fábián and Lazda2021; Bahle, Reference Bahle, Daly, Pfau-Effinger, Gilbert and Besharov2023). Portugal was the best performing of the Mediterranean nations, adjusting its existing policy system to widen its scope, increase spending and ease eligibility conditions. Spain in comparison scored poorly because it did not prioritise children’s access to services and offered parental leave only on an unpaid basis.

4. Overview and conclusion

Adopting a child-centred approach and looking at a swathe of relevant policies, this article has sought to pinpoint and make sense of the social policy activities of the member countries of the EU and the UK regarding children and families during the pandemic in 2020. It set out to answer two research questions which probed in turn the compensatory measures for children and their parents, on the basis of relative expenditure and the preferred modes of support, and the cross-national patterning in the overall degree of child-centredness. Both quantitative and qualitative evidence were used in an attempt to put expenditure and policy substance together and include cash support, food provision, leave for parents and a range of services in a single analysis. The focus has been on policy effort towards child-centredness rather than outcomes.

In pandemic conditions, the classic tripod of family state and market did not apply in the same way – with lockdowns in place, for example – and so the relationship between the family and the state was revealed much more clearly, as well as being elevated to a high level for the achievement of individual and collective welfare. The evidence reveals two relevant contemporaneous policy trends: a stripping back of ECEC and primary education services which had the effect of ‘returning’ children to their family in many instances; a strong favouring of policies resourcing parental and family-based care for children. Many countries tried to ‘soften’ ECEC and school closures, by enabling some access or providing alternative supports (as in the case of distance learning). In general though, parental and employment considerations dominated as criteria for service access, and the primary interpretation of children’s needs was in terms of resourcing parental care-giving and support. In this and other ways, it seems that what might be said to be traditional values of 'family-centredness' or the more recent focus on ‘parent-centredness’ prevailed over a child-centred approach, and that the disruptive potential of the pandemic was not utilised to significantly advance measures for children’s development or their autonomy and agency. Looked at as a whole, the pandemic saw no progress in the application or further development of a children’s rights perspective. Gaps in the timing of responses for children as against those for adults but also the use of conditionality for accessing services and the reversal of some previous guarantees for children are all pertinent evidence here. Significant also was children’s lacking visibility and participation in decisions affecting them.

Cross-country variation should, however, caution against blanket overview statements. Countries varied in the resources they directed at children and families, the package of policies they adopted, the degree to which the measures were explicitly directed at children and/or families, and the timeliness of the responses. But when we applied an index for a standardised comparison, we found quite strong patterning by region. These findings are generally as expected from what we know about welfare state models (although as mentioned the specificities of welfare states’ engagement with children do not yet have a firm foundation in welfare state typologising). The Scandinavian countries were among the most responsive and protective of children and their families. Some of the Continental European countries were also very strongly child- and family-centred, especially Austria, France, and Luxembourg. Ireland and the UK scored relatively well also, mainly because their long-standing targeted approach to children and families gave them an advantage when it came to prioritising ‘vulnerable’ children for access to ECEC and primary education during the pandemic. It is generally the case that many of the countries that were weak in child-centredness before the pandemic remained weak during it and that the more generous social spenders were better able to adapt to the pandemic than others (although we await final outcome data).

What explains this patterning? When the existing policy instruments are wide-ranging and available for use, they tend to reduce the need to find and embed new instruments (something of inherent difficulty in pandemic conditions that call for speedy action). This is particularly the case for services – countries that had a strong network of services for children and families prior to the pandemic were better able to keep these going during the pandemic whereas it proved practically impossible for countries to introduce new services (with food support as a classic example). This explains the high placing of some of the Scandinavian countries given their generous income benefits, prioritising of income continuity and security and well-developed network of ECEC and other services (Chzhen et al., Reference Chzhen, Gromada and Rees2019). Austria, France and Luxembourg are all historically strongly supportive of families with children and this fuelled pandemic responses that gave families greater financial assistance and some support through services. In regard to countries towards the lower end, many are classically low spenders on children and families and/or did not have a strong network of ECEC or other services for children and families that could be actioned to reduce the impact of the pandemic on children and families.

But a path dependency interpretation is challenged by another striking point about the reforms which was the degree to which pre-existing trends viewed in the round were disrupted. Such trends include guarantees for children to services, a move towards explicit target setting (as in anti-child poverty targets), a general commitment to early years support and the recognition that both income support and services are essential for children’s well-being (Richardson et al., Reference Richardson, Carraro, Cebotari, Gromada and Rees2020). One should add here also the suite of parent-differentiated and gender-differentiated leaves and supports which have sought to address both gender inequality and imbalances between work and family life. This (perhaps temporary) reversal can be interpreted in a number of ways. First, it may be that the existing measures were shallow and had not been strongly institutionalised prior to the pandemic (after all, many are stronger as EU goals than as national goals). Secondly, we know that many of these measures have multiple aims and foci and that, while they can be interpreted as child-centred, they are also intended to serve functions around parental employment and work-life balance. While there is no necessary opposition between the well-being of parents and children, there is a tension there and the question has to be asked whether focusing on parents or families – as happened during the pandemic in many countries – is a sufficient policy response to the well-being of children. Current thinking, the child-centredness measures developed here and the child rights perspective would suggest not (Tarshish, Reference Tarshish2019; UNICEF, 2020).

There are a number of qualifications to be lodged regarding the analysis offered. First, the data are not as detailed as they might be and in many cases we have had to rely on general indicators and broad features of policy and were not able to investigate patterns regarding broader system parameters. In addition, we could not fully match expenditure and policy design/decision. In addition, while broad, this analysis rests on a rather selective set of policy fields, omitting such policies as children’s access to health, to ‘voice’ and participation, housing and child protection and anti-violence measures. Reliable data is a continuing problem, with the quality of data on polices that are devolved to lower entities (e.g. food support, ECEC services) especially variable. Furthermore, as we concentrated on policy inputs, we cannot be sure whether and how the measures were implemented in practice and how many children and their families actually benefited from the pandemic-related provisions.

There is also a theoretical challenge in studying the pandemic. It was policy making in emergency conditions after all and it is not yet clear whether it left any lasting impact. It seems from the analysis carried out here, though, that short-term and reactive policy making prevailed. For example, when schools closed, home-based supports needed to be considered, whether for children’s education or as a substitute for wider support services often provided by schools (such as meal provision). This meant that child- or family-related measures were not planned carefully at the outset but were frequently the result of (negative) feedback and emerging gaps and holes in provision.

Limitations notwithstanding, this article has sought to develop and apply a framework in regard to children and families that can be used more widely. Differentiating between measures targeting children and/or adults and the degree of welfare effort both draws in and differentiates between domains of policy and types of policy approach and is a distinction that has proved penetrating for the analysis of policy for children during the pandemic.