Early childhood education and care (ECEC):
A focused review of reform impact studies

Accessible, Affordable and Quality ECEC

One first key lesson is that for ECEC to be effective, it needs to be accessible, affordable and of sufficient quality – sometimes called the “childcare triangle” (Gambaro, Stewart, and Waldfogel 2015). Yerkes and Javornik (2019) compare these aspects (alongside availability and flexibility) in detail from parents’ perspective. Based on their analysis, for ECEC to be accessible, not only do places need to be available, but other factors play a role as well – including eligibility rules, the minimum age at which children can be enrolled in ECEC, whether a place in ECEC is guaranteed, and whether the dominant mode of provision is public or market based (see also Brennan, Cass, Himmelweit, and Szebehely, 2012). To be de-facto accessible to parents, ECEC also needs to be sufficiently flexible to meet parents’ needs, for instance with respect to opening hours, adaptation to non-standard working hours (early, late, weekends or holidays), and overnight services for parents working night shifts, for example.

Affordability relates to the costs of ECEC, and what share of those costs are paid by the parents (out-of-pocket expenses). This depends on the subsidies parents receive, how the funding rules are set, for instance with supply-side financing, in which ECEC centres directly receive public funding, or demand-side financing in which parents receive vouchers or (partial) reimbursements of their childcare costs. Here, it is important to consider in detail how existing fee structures including the broader tax-benefit system (Van Lancker and Ghysels 2016) play out differently for families. For instance, although in many countries single parents pay lower fees for ECEC (in absolute terms) compared to couples, even these reduced fees form a larger burden on their lower household incomes (Maldonado and Nieuwenhuis 2020).

Quality of ECEC relates to aspects of structural quality (including maximum groups sizes, child-to-staff ratios and educational requirements of staff) and process quality (including the daily activities for children, the interactions among children and between children and staff, and the curriculum) (Motiejunaite et al. 2020; OECD 2018). Comparative data on structural quality are generally more widely available than data on process quality (OECD 2017), which can limit conclusions drawn regarding the relationship between process quality and outcomes, such as gender equality (Yerkes and Javornik 2019).

ECEC is associated with a wide range of positive outcomes

The second important lesson is that the benefits of ECEC can extend beyond maternal employment, a key focus of the focused literature review. Enrolment in ECEC is beneficial to children’s social and cognitive development, with particularly significant benefits for children from a disadvantaged background (Egert, Fukkink, and Eckhardt 2018; Gambaro et al. 2015; Van Huizen and Plantenga 2018). ECEC can also play a role in protecting children from the consequences of growing up in poverty (Leseman and Slot 2014), including short-term consequences such as food poverty as well as long-term consequences, such as learning disadvantages (Drange, Havnes, and Sandsor 2016). In terms of gender equality, childcare plays a role in supporting women’s employment and earnings, as well as more gender-equal divisions of time spent on care tasks (An and Peng 2016; Hook 2010), even though most mothers continue to take on more of the mental load of organizing childcare (Craig and van Tienoven 2019; Dean, Churchill, and Ruppanner 2022). Childcare also contributes to public attitudes being more in favour of maternal employment (Neimanns 2021). From a societal perspective, ECEC provision helps to protect families against poverty (Förster and Verbist 2012), to reduce gendered earnings inequality both within couples (with women earning larger shares of household incomes in countries with more extensive ECEC provision) as well as between couples (Nieuwenhuis, Need, and Van der Kolk 2019), and has the potential to boost economic growth (Cingano 2014).

Use of ECEC is socially unequal

A key third lesson is that the use of ECEC is socially unequal. Parents who have higher levels of education and income are more likely to enrol their children in ECEC (Ghysels and Van Lancker 2011; Van Lancker and Ghysels 2016). Importantly, the degree of inequality in childcare use varies across countries, which is a strong indication that the degree of inequality is in part affected by the institutional context. Merely increasing public expenditure on ECEC (spending that goes towards services, ECEC regulation, vouchers for parents or tax deductions) is not found to reduce this inequality (Van Lancker 2018). In contrast, Van Lancker and Ghysels (2016) and Nieuwenhuis et al. (2021) find that inequality in ECEC use (by parental education) is lower in countries in that have:

• Public or publicly-subsidized supply of ECEC services;
• Childcare guarantees (either guaranteed to parents or a guaranteed place for children in an ECEC centre);
• Lower out-of-pocket fees;
• High parental perceptions of ECEC quality.

The context in which ECEC is implemented

A fourth and final lesson is that the national context of ECEC matters for the effectiveness of ECEC policy and for how this effectiveness is evaluated. ECEC policy plays an important role in facilitating maternal employment and fostering gender equality, but its effectiveness is contingent to the policy context and structural conditions. Most notably, many countries have a ‘childcare gap’: a period in which (well-paid) parental leave has ended, but children have not yet reached the age at which they are entitled to ECEC provision (Eurydice 2019; Koslowski et al. 2021). The childcare gap example illustrates the importance of implementing interrelated family policies in a coordinated manner. Much less is known about the interplay between ECEC and other policies and contexts. However, studies suggest the returns to further increments in ECEC provision are in fact smaller in contexts where women’s employment is already high (Akgunduz and Plantenga 2018) and in contexts with other policies to support women’s employment, such as active labour market policies (Nieuwenhuis 2022). The explanation for these diminishing returns remains unclear, but likely has to do with the fact that in countries where maternal employment is common, those who do not work do so for specific reasons not addressed by ECEC.

The fact that the effectiveness of ECEC policies depends on the context in which these policies are implemented, also means that a reform proven to be effective in one country, may not necessarily be effective in another (Béland 2019). There is limited scholarly understanding of such policy interplay (complementarity and substitution), but clearly the implementation of ECEC policies and reforms is context dependent. Consequently, longer periods of time may be required before the full potential of ECEC policies is realised (e.g., as other policies need to be brought in alignment, and/or policy reforms need to be tailored to the local institutional context).

Step 1: Publication Search

A focused literature review of scientific research was performed relying on databases of high-quality, (predominantly) peer-reviewed scientific publications. Keeping the search within these parameters ensures the evidence base is of sufficient and consistent quality. Moreover, these databases allow for detailed searches (e.g., specifically in titles, abstracts, keywords) that drastically improved the relevance of the search results while maintaining an efficient workload. The following databases were searched:

1. Web of Science

2. Scopus

3. Ebsco

4. Worldcat

5. Jstor

6. Stockholm University library

These databases together cover journal articles, academic books, dissertations, and theses, that have undergone some form of peer review (e.g., either by external reviewers, or by an academic grading committee for theses). Google Scholar was not included as it is not a ‘curated’ database, which means it includes a substantial number of resources that are not peer-reviewed, and/or resources whose scientific quality cannot be verified.

The following search terms were used, searching within publication titles, abstracts and keywords:

1. “childcare policy” + reform

2. “ECEC policy” + reform

3. “childcare reform”

4. “ECEC reform”

5. “child care policy” + reform

6. “child care reform”

7. “child care” + reform⁵

The first four search terms in the list above were established based on the requirements of the tender. The latter three were later added in the process of the literature search because the original search terms were found to be too restrictive. Note that search terms pertaining to specific countries or groups of countries (e.g., EU-27, or specific member states) or specific outcomes (e.g., children’s participation, mothers’ employment) are not included as search terms, ensuring results on all countries and all outcomes were initially included.

Step 2: Publication selection

The search terms listed above yielded a total of 8031 publications (see Table 1 for an overview of the numbers at each step of the selection procedure). Bibliographic information for all 8031 publications was imported into reference management software ‘Zotero’. The further selection of publications relevant to the literature review took place in 3 sub-steps. First, 3022 duplicates were removed. Second, the titles, abstracts and keywords of the publications were individually reviewed using the following inclusion/exclusion criteria (this led to 4933 exclusions):

• It is an empirical study;
• The study examines the outcomes of policy reform. A reform means that a political decision was made to change the policy / rules regarding ECEC or childcare;
• The reform took place in an EU member state, Norway or Switzerland; it can be a national or a regional reform;
• The study is focused on outcomes (related to the reform) that pertain to:
  • women’s/maternal employment
  • gender equality
  • parents’ use of / children’s enrolment of children in childcare/ECEC;
• The study is published between 2000 and 2021.

Finally, the 76 remaining studies were read in full in a detailed review of literature to determine inclusion/exclusion, which resulted in 30 additional exclusions. Table 1 provides an overview of the total number of publications excluded at the subsequent steps.

Table 1 Publication overview at each stage of the selection procedure

Step 3: Database creation

Following this review and selection of publications, a total of 76 publications remained. These 76 publications were included in the third step of the literature review: the database creation.

At this stage, the 76 publications selected for inclusion in the database were downloaded, the bibliographic information was checked, and the full-text publications were read in detail. During this close reading of each publication, a further 30 publications were excluded, as it became apparent in the full text of the publication that these studies did not meet the strictest inclusion/exclusion criteria. Yet, these studies were still included and coded, as they may provide valuable contextual information.

Each of the 76 publications is included in a searchable database with separate tabs for publications included in the final set of relevant articles (46 publications), and a tab for 30 publications reviewed but excluded at this step but might include relevant contextual information. Table 2 provides an overview of information included in the database for each publication.

The database (in .xlsx format) is text searchable using the search function in Excel. In addition, several of the entries listed above have a selection menu (at the top of the spreadsheet) that allows for applying a filter to the database based on standardized categories. This helps to focus the search for specific types of reforms or outcomes. In Table 2, those entries are indicated as “Filter on categories?”. This is the case for the entries on country, region, type(s) of reform, and type(s) of outcome. The standardized categories will be used to organise the analysis of the database for the final report.

Table 2 Overview Coding Scheme ECEC Reform Study Database

Step 4: Analysis

The purpose of the analysis is to address the questions regarding which aspects of ECEC policy reform affect parents’ decision to use childcare (i.e., participation rates of children), and which aspects affect labour market participation for mothers as well as their working hours and earnings, to address a broader perspective on aspects relevant for gender equality. This focused literature review can only synthesize what has been published elsewhere in this regard, rather than providing a comprehensive answer to these questions.

The analysis of the studies included in the ECEC reform study database is organized based on a classification of the reform outcomes, and a classification of aspects of ECEC policy reform. The outcomes were classified as:

• ECEC enrolment / usage
• Maternal employment
• Maternal working hours
• Maternal earnings

The results are categorized based on these outcomes. Subsequently, for each outcome, studies are organized around the accessibility, affordability and quality of ECEC services and reforms. This provides a useful framework to organize the analysis of the reform studies included in the database. The three aspects of reforms are based on the expertise of the authors and a priori identification of potentially relevant aspects for ECEC services (informed by the lessons from comparative research summarised above) and can be conceptually distinguished. As noted in the ECEC literature however, aspects of ECEC accessibility, affordability, and quality are intrinsically intertwined in reality (e.g., Yerkes and Javornik 2019). Reforms of ECEC policy often – or even typically – affect multiple or even all of these aspects. It is the combination of these factors that determine whether parents will use ECEC: if parents looking for ECEC realize that the provided quality is inadequate, they may not enrol their children even if a place is accessible and affordable to them (Yerkes and Javornik 2019).

ECEC Enrolment / Usage

The following section focuses on childcare usage, an ECEC policy outcome emerging in 22 out of the 46 studies included in the ECEC Reform review. The policy reforms studied are primarily directed towards improving accessibility, a few towards affordability, and only three towards improving quality.

Accessibility

Based on their characteristics and outcomes, reforms related to accessibility can be separated into three groups. First, there are two types of reforms with a positive association with childcare enrolment: (i) clear reforms targeting accessibility (e.g., changing eligibility to make childcare a statutory right), and (ii) gradual expansions of childcare places and facilities rather than clear reforms. Secondly, there are trends where new eligibility requirements or childcare arrangements resulted in decreased childcare usage, as exemplified below in the Czech reform.

Reforms positively associated with childcare enrolment are exemplified well by several German laws that culminated with the Child and Youth Welfare Act (Kinderförderungsgesetz), a law passed in 2008. The law made it possible for all children aged 1 and older to legally claim a place in formal childcare by 2013. As a result, between 2006 and 2014, the average enrolment rate for 1-3-year-olds increased from 8 per cent to 27 per cent in West Germany, and from 40 per cent to 52 per cent in East Germany (Schuss & Azaouagh, 2021; Zoch & Schober, 2018; for a similar reform in Poland see Akgunduz et al., 2021).

The literature also shows examples of gradual expansions of childcare places, e.g., in Sweden and Belgium (Flanders more specifically) since the 1970s. In both countries, the number of places and thus also the take-up of childcare have slowly increased. Additionally, the gradual expansion of accessibility was combined with a reform that introduced a low-income tariff system (i.e., a reform targeting affordability) that led to different outcomes in the two countries. The new system resulted in a relatively equal distribution of childcare use across household-income groups in Sweden. In contrast, in Flanders, the low-income tariff was indirectly counteracted by childcare tax deductions benefitting high-income households 1.5 times more than the lowest income households (Van Lancker and Ghysels 2012). Van Lancker and Ghysels underline that Swedish governmental spending on childcare is substantially higher than in Flanders (only partially for demographic reasons).⁶ They also emphasize the importance of the difference in the number of accessible childcare slots, child-to-staff ratios, and staff salaries between the two countries in explaining the impact of the reforms targeting accessibility. Similarly unequal childcare take-up (based on differences in maternal education) resulted from German reforms targeting accessibility (Stahl and Schober 2018).

The literature suggests not all reforms increased accessibility. An example of reduced ECEC accessibility was reported in the Czechia, where childcare provision for under 3-year-olds was excluded from the pre-school system in 1991. This led to a considerable decrease in childcare facilities and childcare usage during the 1990s – enrolment rates in crèches fell from 18 per cent to 1 per cent from 1997 to 2002 (Hašková, 2014), although it is unclear whether the decrease in participation was an unintended consequence of the reform.

A decline in enrolment rates for 1-3-year-olds also followed a Norwegian 1998 cash-for-care (Kontantstøtteloven) policy reform that rearranged the ways of accessing and using childcare. The cash-voucher premium was granted to recipients who did not use formal childcare, regardless of whether the parent was employed or not. The voucher was approximately equal to the governmental subsidy for a childcare place at 3000 NOK (Chan and Liu 2018). The motivation for the cash-voucher premium was the desire to increase families’ freedom of choice, possibilities for family time, and offering greater variation in alternatives to subsidised formal childcare. As a result, formal childcare usage declined by 6.8 percentage points. Other policy reforms providing voucher premiums for households accessing informal childcare have had similar results in Finland, where the use of informal childcare increased significantly (Viitanen 2011), and Germany, where formal day-care attendance fell by 8 percentage points (Gathmann and Sass 2018).

Affordability

Affordability policy reforms affecting childcare usage often take the form of an increase in low-cost childcare, and as such are directly linked to accessibility expansions. Or, as in the Van Lancker and Ghysels (2012) study of Sweden and Flanders, a tariff system coincides with efforts targeting accessibility. Two examples of reforms that exclusively target affordability aspects of childcare usage are detailed below.

In Germany in 2006, a series of reforms removed fees for 6-year-old pre-school children. Pre-reform, families’ childcare out-of-pocket expenses varied from 100 to 400 euros per month. Post-reform, the number of households paying less than 25 euros a month had risen from 20 per cent to 60 per cent. Studying the outcome of the reform, the authors found no effect on childcare enrolment rates (whereby an average of 48% of 6-year-olds were enrolled in pre-school based on a sample of administrative data). However, hours in childcare increased, and the average 6-year-old was found to spend 0.7 hours (2.2%) more in day-care per week, which accounts for a 3.4 percentage point increase in all-day care, suggesting parents switched from half-day care to full-day care as a result of the reform (Huebener, Pape, and Spiess 2020).

A French reform in 2004 changed the provision of childcare subsidies received by families, increased the amount of the subsidy, and lowered the household-income bracket for receiving such subsidies. Post-reform, a family with eligible children could receive 350 euros more than before, depending on their income levels. The generosity of the benefit also differs marginally based on the number of children being cared for (de Muizon 2020). The effect of the outcome was also found to depend on household size, with a 1.6 percentage point increase in paid childcare usage for households with one child and a 1.2 percentage point increase for households with additional children (Givord and Marbot 2015).

Quality

All quality reforms included in the database were positively associated to the outcome of childcare usage. They were joint reforms, meaning that accessibility was reformed simultaneously with quality (and in the Norwegian case also affordability).

A major Norwegian reform of 1975 (Barnahageloven, The Kindergarten Act) led to a substantial expansion of subsidized childcare for 3 to 6-year-olds (Havnes and Mogstad 2011). Governmental funding more than doubled over a one-year period, with the aim of quadrupling the number of formal childcare places. Municipalities with the lowest usage rates were also targeted with the highest subsidies. Simultaneously, quality was regulated in relation to educational content, group size, and staff requirements. Childcare usage increased, more than doubling the levels recorded in 1975 from less than 10 per cent to 28 per cent. Since then, Norwegian childcare usage has continued to rise gradually, stagnating somewhat above 60 per cent in the mid-1990s (Havnes & Mogstad 2011).

A similar reform called LOGSE (La Ley Orgánica General del Sistema Educativo, General Law of the Education System) was adopted in Spain in 1990. It lowered the eligibility age for free universal pre-school from 4 years to 3 years of age and established full-day availability, while regulating educational content, group size, and staff requirements. As a result, childcare usage increased dramatically from 8.5 per cent in 1990 to 67.1 per cent in 2002 (van Huizen, Dumhs, and Plantenga 2019; Nollenberger and Rodriguez-Planas 2015).

In general, the results of studies related to childcare usage demonstrate that past childcare policy reforms have improved the use of/enrolment in public childcare by increasing the number of ECEC places and facilities. The Czech reform and cash-for-care reforms (e.g., Norway), where accessibility was reduced, were exceptions. While not all reforms targeting accessibility explicitly regulate affordability and quality simultaneously, these two aspects necessarily go hand-in-hand in explaining cross-country differences (e.g., Belgium and Sweden in Van Lancker & Ghysels, 2012).

Maternal Employment

The following section outlines research on how childcare policy reforms have affected maternal employment across EU member states. Many studies examine how successful reforms targeting accessibility have been aimed at bringing previously unemployed mothers into paid work, but some have also examined how making childcare more affordable and/or of higher quality impacts mothers’ employment decisions. Fathers’ employment will also be touched upon in brief.

Accessibility

Studies that examine the effects of reforming accessibility on maternal employment rates generally find that increased access to ECEC increases the maternal employment rate (Andresen and Havnes 2019; Bick 2016; Carta and Rizzica 2018; Dujardin, Fonder, and Lejeune 2018; Kunze and Liu 2019; Müller and Wrohlich 2020; Stahl and Schober 2018). Examples of how ECEC policy reforms have increased access include increasing ECEC provision, implementing a legal entitlement to an ECEC slot for children of a certain age, and extending eligibility to children who have unemployed parents. A few studies also examine how reforms targeting accessibility affect paternal employment. Typically, such reforms are found to have no significant impact on fathers’ employment rate (Andresen and Havnes 2019; Ravazzini 2018; Vikman 2013).

An important factor to consider in the context of employment is which labour market opportunities are available to mothers willing to enter employment. An Italian reform that extended subsidised childcare to cover two-year-olds was found to have an overall positive effect on maternal employment rates, but it increased employment the most in regions with a high employment vacancy rate (Carta and Rizzica 2018). This highlights that reform effects result from an interplay between different factors (e.g., ECEC and labour market policies), and consequently, the need to consider the broader social, political, and policy context when introducing or reforming policies in an attempt to increase maternal employment.

Several studies examine how reform effects vary between socioeconomic groups. They report that reforms of accessibility have a greater impact on high- and mid-educated women’s employment rates than those of low-educated women (Akgunduz et al. 2021; Carta and Rizzica 2018; Dujardin et al. 2018; Stahl and Schober 2018). This means that if high-educated mothers are more likely to enter paid employment than low-educated mothers as a result of the reform, then the employment gap between these groups will grow (Stahl and Schober 2018).

In 2009, a reform in Poland made all five-year-olds legally entitled to a childcare slot. This reform was implemented simultaneously with a lowering of the primary school age from seven to six. As six-year-olds moved into primary school, further childcare slots became available for those younger than six. Akgunduz et al. (2021) highlight that despite the introduction of the legal entitlement for five-year-olds, pre-school places may not be available to all five-year-olds in practice, due to municipality budgetary constraints. Nevertheless, the 2009 reform caused maternal employment rates to increase such that for each “10-percentage point increase in the ratio of preschool seats to preschool-aged children increases maternal employment rates by approximately 4.2 percentage points” (Akgunduz et al. 2021:1098).

In 2002, a reform in Sweden made children of unemployed parents entitled to attend at least 15 hours of childcare per week (available to children under the age of pre-primary school (Förskola)). Before the reform, only children of employed parents had been entitled to subsidised childcare. A study that examined the effects of the reform found that the probability of leaving unemployment increased by 16 per cent for mothers, while it had no significant effect on fathers’ probability of leaving unemployment (Vikman 2013).

Affordability

Scholars have also examined how reforms that reduce or remove childcare fees paid by parents affect maternal employment rates (Bettendorf, Jongen, and Muller 2015; Busse and Gathmann 2020; Givord and Marbot 2015; Hardoy and Schøne 2015; Lundin, Mork, and Ockert 2008; de Muizon 2020). As in the case of reforms targeting accessibility, those targeting affordability are generally found to increase maternal employment.

However, studies that compare maternal employment rates between socioeconomic groups have yielded mixed results. In France, a reform was introduced in 2003 that made all children born after 1 January 2004, as well as any siblings of such children, part of a new benefits system aimed at households with pre-school aged children. The reform consisted of two main parts: a new stay-at-home benefit for parents on parental leave that was not means-tested and childcare subsidies to cover some costs related to using a professional registered childminder. Parents are required to pay at least 15 per cent of childcare fees, but there is a cap on childcare costs. The subsidy varies according to three income brackets and the number of children cared for. The reform increased the amount of the subsidy by increasing the threshold and the generosity of the subsidy at each income bracket. Moreover, it was restricted to working parents. This reform was found to increase employment rates of middle-class and high-educated mothers with two children. Simultaneously, it reduced the employment rate of first-time mothers and low-educated women, due to a higher level of take-up of the parental leave benefit in these groups (de Muizon 2020). Note that none of these effects were observable until three years following the reform.

In contrast, Norway introduced the Child Care Centre Agreement (Barnehageforliket) reform in 2003. This reform increased ECEC service capacity and required municipalities to provide childcare for all parents of children aged one to five years. It also capped prices for full-time childcare slots. All parents with eligible children were entitled to the reduced childcare price. Examining effects of the reform, Hardoy & Schøne (2015) found that reduced childcare prices increased the maternal employment rate by approximately 4 percentage points, or approximately by 5%. Moreover, the reform was found to increase low-educated and low-income households’ employment rates more than it increased the employment rates of high-educated and high-income households (Hardoy and Schøne 2015). In other words, two similar reforms that introduced caps on childcare fees resulted in differing employment consequences for different socio-economic groups. Such contrasting results highlight the importance of consideration to specific aspects of the reform, as well as the broader policy context in which the reform is introduced.

Quality

The database does not include studies that solely examine the effect of reforms targeting quality on maternal employment rates, but a few studies have examined the effects of reforms that seek to improve quality alongside accessibility.

Two studies (van Huizen et al. 2019; Nollenberger and Rodriguez-Planas 2015) examine the effect of such a reform on maternal employment rates in Spain. The LOGSE (Ley Orgánica General del Sistema Educativo) reform was introduced in 1990 and lowered the eligibility age for universal public pre-school from four to five years of age to three to four years of age. The reform also introduced quality regulations related to the content of education (with a new focus on play-based education, group play, problem solving and critical thinking), group sizes (with the introduction of a maximum of 25 children in classes for 3-year-olds and a maximum of 20 children in classes for 4-5-year-olds) and educational requirements for staff (a college pedagogy degree requirement for all staff). Both studies found that the maternal employment rate increased as a result of the reform. Specifically, Nollenberger and Rodriguez-Planas (2015) found that for every 10 additional children entering childcare, 2 mothers entered employment. Van Huizen and colleagues (2019) found that in the short-term, the maternal employment rate increased by 0.2 per cent for every additional 3-year-old in preschool, and that in the long-term, the reform led mothers to reduce their employment interruptions.

Maternal Working Hours

In addition to studying the effect of childcare policy reforms on maternal employment rates, several studies also examine the effects of reforms on mothers’ working hours. While both outcomes concern maternal labour supply, employment rates relate to new participants joining the labour market, whereas working hours can refer to two different types of outcomes: the average hours worked by all employed mothers or the hours worked by mothers who were already in paid work pre-reform. In this section, we focus on the latter category.

Accessibility

The results of studies examining the effect of ECEC reforms on maternal working hours vary somewhat, but generally, it is found that work hours increase as a result of increased childcare accessibility (Ravazzini, 2018; Schuss & Azaouagh, 2021; Zoch & Hondralis, 2017). One study found that the positive effect of reforms is driven by those mothers already in employment increasing their probability of high part-time participation (Ravazzini, 2018). Another study finds that the increase in working hours was driven by mothers moving from part-time to full-time employment (Schuss and Azaouagh 2021).

To relate this finding back to maternal employment rates, some studies have found that both the maternal employment rate and hours worked by mothers increase (Bettendorf et al. 2015; Muller and Wrohlich 2016) as a result of reforms increasing ECEC accessibility. Others have found that only the working hours of mothers already in paid employment increase (Ravazzini, 2018). One study examining the effects of increased childcare provision for 1-2-year-olds reported only weak evidence of an increase in contracted hours of work, but an overall positive effect on actual hours worked by mothers (Kunze and Liu 2019).

Reforms on accessibility have also taken place in Germany. In 2005, Germany passed the Day Care Expansion Law (Tagesbetreuungsausbaugesetz, TAG), which aimed to expand subsidised childcare to cover 17 per cent of 1-3-year-olds in West Germany by 2010. In 2007, this target was raised to 35 per cent. The TAG reform was followed by a 2010 law that legally entitled children aged one and above to part-time childcare if they had a single parent, if both their parents had been employed before birth, if parents were job-searching or receiving unemployment assistance, or if their sibling was or had already attended public childcare. A 2008 law extended childcare eligibility to all children aged one and above by 2013. Prior to the TAG reform, the childcare coverage rate for this age group was almost inexistent, but by 2015 it had risen to 23.6 slots per 100 children (Schuss and Azaouagh 2021). Schuss and Azaouagh examined the effect of TAG on the maternal employment rate and hours worked by mothers. They found that using ECEC for children younger than three increases the likelihood of mothers working full-time by 13.2 percentage points. The increase was driven mainly by women already in paid work switching from part-time to full-time employment. The reform was found to have no employment effects on mothers with a migration background from outside the EU.

Affordability

Four studies have examined how reforms targeting affordability affect maternal working hours. All show at least some increase in maternal working hours in response to reforms that reduce the out-of-pocket expenses to parents (Brink, Nordblom, and Wahlberg 2007; Busse and Gathmann 2020; Hardoy and Schøne 2015; Huebener et al. 2020). Two studies report that single mothers are especially responsive to reforms aimed at improving affordability (Brink et al. 2007; Huebener et al. 2020), meaning that they are more likely to increase their working hours in response to lowered childcare costs than married or cohabiting mothers. Two studies report that mothers in low-income households increase their working hours more in relation to reduced childcare prices than higher-income households (Brink et al. 2007; Hardoy and Schøne 2015). Again, mothers’ employment rates and working hours are closely related, but not necessarily affected in the same way. In one German reform, mothers of toddlers increased their employment rate, whereas mothers of children in the last year of pre-school increased their working hours in response to the abolition of childcare fees (Busse and Gathmann 2020).

Sweden introduced a maximum fee (maxtaxa) for early childhood education and care in 2002, which aimed to improve the economic situation of families with younger children as well as to increase parental employment. Before the reform, childcare fees were determined at the municipal level, with fees varying greatly (between 0 and 4,160 SEK per month) across municipalities. Fees were mainly determined by parental income and the number of hours spent in childcare, with high-income households paying more than lower-income households. Following the 2002 reform, the fee is still based on household income, but fees are capped. The fee varies with the number of children; 3 per cent of the gross household income for the first child, 2 per cent for the second child, 1 per cent for the third child, and no fee for additional children.

Brink et al. (2007) report that the 2002 reform had a more positive impact on single mothers’ employment than on two-parent households. Single mothers increased their employment hours by 1.3 per cent. The impact was especially positive for low-income single mothers who increased their working hours by 16.5 per cent because of the reform. The increase in working hours was in part due to single mothers entering the labour market, and in part due to single mothers already in paid work increasing their working hours. The effect on working hours was modest for couples, but all except the highest-income bracket increased their labour market supply to some extent.

In Germany, states are responsible for organising and setting goals for public childcare. Between 2000 and 2016, several states removed parental childcare fees. Busse and Gathmann (2020) examined the effects of fee removal in nine states; six states removed childcare fees for children in their final year of pre-school, while three states removed fees for children aged 2 to 6 years old. They found that removing fees for children in their final year of pre-school led already employed mothers to increase their working hours. For mothers of younger children, aged two to three years, the reform led to an increase in their labour market attachment (Busse and Gathmann 2020).

Quality

The database does not include studies that solely examine the effect of reforms of ECEC quality on mothers’ working hours, but one study has examined the effects of increasing childcare provision, where receiving funding for expansion was conditional upon meeting relevant quality standards.

Ravazzini (2018) examines the effect of the gradual expansion of childcare provision that took place in Switzerland between 2002 and 2012. Due to a very high demand for childcare, the Swiss federal government granted subsidies of 200 million francs to early childcare and after-school facilities in 2003.⁷ The subsidy related to an initial 8-year-period, but was renewed twice thereafter, with a total expenditure of 300 million francs by 2016.⁸ Approximately half the amount was spent on early childcare services. In order to benefit from government funding, childcare services had to meet quality standards related to the ratio of trained educators to children, technical requirements, and opening hours.

Ravazzini found that the childcare expansion increased the hours worked in part-time employment (20-36 hours per week) by 2 percentage points for mothers of three-year-olds. This positive effect was mainly driven by an increase in working hours among married or cohabiting mothers with an upper-secondary degree and two children. It had no effect on maternal employment rates or fathers’ labour supply.

Maternal Earnings

The following section concerns maternal earnings, a policy outcome emerging in 6 out of the 46 studies in the focused literature review. The impact of ECEC reforms on maternal earnings varies depending on mothers’ educational and migration status. Generally, ECEC reforms usually enhance gaps between groups, leading to higher earnings among mothers without a migration background and higher educated mothers. At the same time, no effect, or even a negative effect on earnings, is found for mothers with a migration background and lower educated mothers. Only the Norwegian reform Barnahageloven stands out as an exception.

Accessibility

In 2000, the Swedish government decided upon a childcare reform that targeted unequal access to childcare. Pre-reform, childcare fees were gradually growing, making childcare less accessible and affordable – a development that affected some groups harder than others. Children of immigrant mothers were underrepresented in childcare usage in particular. The reform of 2000 therefore adjusted several things: (i) the fee system was regulated with cost restrictions capped at 3 per cent of family income (see also reforms targeting affordability below); (ii) unemployed parents became entitled to 15 hours of childcare per week; and (iii) a universal pre-school system was established for children between 4 and 5 years of age (Wikström, Kotyrlo, and Hanes 2015). The reform positively affected maternal earnings with a 10-21 per cent increase for mothers without a migrant background. In contrast, the effect on the earnings of mothers with a migrant background was non-existent (Wikström et al., 2015).

Two Norwegian childcare reforms that have been discussed childcare usage section have additional effects on maternal earnings. Firstly, the Barnahageloven reform of 1975 had an equalizing income effect on household earnings (Havnes & Mogstad, 2011), which is addressed further below under the quality section. Secondly, the Kontantstøtteloven – the cash-for-care reform of 2008 – established a home-care premium of up to 3000 NOK for households who abstained from using formal childcare. By comparing the outcomes of the reforms on maternal earnings between mothers of different educational levels and migration status, Naz (2010) demonstrates considerable variations.

For all Norwegian women, regardless of educational level, earnings dropped. Among native-born women, lower educated women were affected nearly twice as hard, witnessing a 12 per cent decrease in average earnings compared to a 5 per cent decrease for higher educated women. Among women with a migration background within the OECD, only lower educated women’s earnings dropped by 18 per cent. In contrast, for non-Western immigrant mothers, both lower and higher educated women were affected: a 41 per cent decrease was found for highly educated non-Western immigrants (Naz, 2010).

Austria also introduced reform to improve accessibility, aimed at a gradual expansion of subsidised childcare provision. Despite twenty years of childcare expansion between 1990 and 2010, maternal earnings have not increased. The author suggests that the absence of increased earnings is caused by the consistency in Austrian informal care availability and maternal care norms (Kleven et al., 2020).

Affordability

Few studies examined the relationship between affordability and maternal earnings. The abovementioned Swedish reform, for example, includes the introduction a limit to municipal childcare fees. The first part of the reform introduced a limit on municipal childcare fees proportional to household incomes (Maxtaxa) (for a complete description of the reform, see the section on Maternal Working Hours). If municipalities did not accept the fee ceiling, they lost approval for additional governmental funding on childcare. When focusing solely on the outcome of the fee restrictions, Brink et al. (2007) found a positive effect on maternal earnings for Swedish households in general. The gain was more prominent for higher-income households.

Quality

The Norwegian policy reform of 1975 (Barnahageloven; for a complete explanation of the reform, see the section on Childcare Usage) once again serves as an example, but with an emphasis on aspects related to the reform affecting quality (i.e., that regulated educational content, group size, and established staff requirements in Norwegian pre-schools (Havnes and Mogstad 2011). The paper does not, however, explicitly discuss the outcome of maternal earnings but rather household income levels. It is therefore not possible to assume that the effect found is due solely to increased maternal earnings. There is even the possibility that the rise in household income can be caused solely by fathers, but the paper does not make this explicit.

The 1975 Norwegian ECEC reform had a positive effect on the lower and middle part of the income distribution, reaching its peak at the 11^th percentile, where household earnings increased with $1.27 for every dollar invested in the policy reform. In the upper part of the income distribution, the reform negatively impacted household earnings with a -$1.15 decrease for every dollar invested. Accordingly, as the authors explain, the reform reduced income inequality between generations by mitigating the effect and advantage of being born in an affluent household, lowering the Gini coefficient by 2.9 per cent.

Research gaps in the database

The database and summary of outcomes provides an evidence base to inform future reforms and our initial analysis of this database has provided several general insights. Yet, the analysis of the database has also brought to light a number of limitations related to what is not covered in the (type of) research included in the database. We discuss five such research gaps, as they provide guidance on what can as well as what cannot be concluded based on the research collected here. Reflecting on these gaps may also give direction to future research.

First, the results of the studies investigating ECEC reforms are highly context-specific, meaning these results can and should not be unequivocally extrapolated to other country contexts. The research design of studies that evaluate reforms are inherently context-specific, as they seek to exclude the influence of contextual factors. Moreover, these studies are generally focused on a single country. These studies often seek to demonstrate the existence of a causal effect and are less focused on explaining why the reform (or specific features of the reform) is effective. The most straightforward implication of this focus is that there is no empirical evidence that what has worked in one country at a given moment in time, will work again in another country at any other point in time. This implication is particularly relevant because many EU-27 countries are not included in the database. The absence of countries in the database may be related to a lack of resources available for evaluating reforms in these countries (e.g., research expertise, data, priorities in the research agenda). It might also reflect an absence of ECEC reforms taking place (excluding gradual policy change). A secondary implication of the context specificity of the findings reported here is therefore that the evidence base might be the weakest for those countries that might benefit the most from ECEC reforms. In the absence of a strong evidence base for these countries, comparative research that relates (changes in) ECEC policy features to outcomes relevant for gender equality could supplement the research presented here.

Second, although the studies in the ECEC reform study database all studied one or more reforms related to accessibility, affordability, or quality, the potential interplay between these different aspects of ECEC policy were not studied. The absence of attention for the interplay of policy design can make it difficult to draw conclusions about the effectiveness of ECEC reforms. If accessibility expands and childcare coverage increases, while the quality of facilities and staff education remains low or even decreases, few parents would judge the newly provided ECEC as appropriate for their child – which might undermine the effectiveness of the reform and lower ECEC enrolment (Yerkes and Javornik 2019). Formal childcare usage depends, among other things, on ECEC accessibility and parents’ ability to compensate for the risk of low-quality care. If formal subsidised childcare is of lower quality or less accessible than childcare on the market, high-income households have incentives to resort to market options. Yet, if ECEC access is good, quality is high, and out-of-pocket expenses are generous, formal childcare compensates by giving low-income households greater purchasing power (Havnes & Mogstad, 2011). Accordingly, policy reforms supporting childcare usage are a joint effort where no type of reform does all the work alone.

Third, the way ECEC is used and can be effective to increase maternal employment, working hours, and earnings, depends on its interplay with other institutional and structural conditions. The existence of such interplays further increases the context-specificity of the findings as discussed above. The importance of such interplay was demonstrated in a limited number of studies in the database. One example was the reform in Italy, where an expansion of childcare policy was more effective in regions with greater (unmet) demands for labour (Carta and Rizzica 2018). Cash-for-care schemes in for instance Norway, Germany and Finland reduced the effectiveness of ECEC (Busse and Gathmann 2020; Chan and Liu 2018; Viitanen 2011).

Fourth, few studies have examined how long it takes for implemented policy reforms to have an effect. Inherent to the research design, most studies in the ECEC reform study database have examined the immediate and short-term effects of policy changes. Yet, there is a reasonable expectation that the effectiveness of ECEC policies change across time. For instance, as childcare usage becomes more common, public attitudes to maternal employment can change as exemplified by French reforms. The reform of the parental leave scheme as well as ECEC had “negligible effects in the short term”, but after three years, lower educated mothers used the extended parental leave rather than ECEC (de Muizon 2020:735). Similarly, in Germany, educational inequality in the use of ECEC and maternal employment were not found immediately, but emerged over time (Stahl and Schober 2018).

Fifth and finally, readers of this report and users of the ECEC reform study database should be aware of the existence of publication bias. In its most general form, publication bias refers to the fact that studies that do not find (statistically significant) results are less likely to be published. Indeed, almost all studies included here report that the reform was effective in achieving one or more outcomes. One of the few exceptions is a study on childcare expansion in Austria, which reported that maternal earnings had not increased more in municipalities that expanded childcare accessibility across time than in municipalities that had not (Kleven et al. 2020). Another exception is a study of a reform that capped childcare fees in Sweden, which found no reform effect on maternal employment (Lundin et al. 2008). Therefore, the database and the analysis presented here show what has been effective in increasing ECEC use and women’s employment, working hours and earnings. They may not show what has not been effective. A more specific form of publication bias is that the published studies focused more on some groups than others. Most notably, only a handful of studies examined fathers. The few studies that focused on fathers found little to no effect of childcare reforms on fathers’ employment (Andresen and Havnes 2019; Ravazzini 2018; Vikman 2013) or on fathers’ attitudes to maternal employment (Zoch and Schober 2018). One exception is a study of a Dutch reform, where fathers with a youngest child aged 0-3-years-old reduced their employment hours by 0.5 hours per week (Bettendorf et al. 2015).