From homemakers to breadwinners? How mandatory kindergarten affects maternal labour market outcomes

Gangl, Selina; Huber, Martin Article — Published Version From homemakers to breadwinners? How mandatory kindergarten affects maternal labour market outcomes Journal of Population Economics Provided in Cooperation with: Springer Nature Suggested Citation: Gangl, Selina; Huber, Martin (2025) : From homemakers to breadwinners? How mandatory kindergarten affects maternal labour market outcomes, Journal of Population Economics, ISSN 1432-1475, Springer, Berlin, Heidelberg, Vol. 38, Iss. 2, https://doi.org/10.1007/s00148-025-01099-y This Version is available at: https://hdl.handle.net/10419/323188 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Journal of Population Economics (2025) 38:44 https://doi.org/10.1007/s00148-025-01099-y ORIGINAL PAPER From homemakers to breadwinners? How mandatory kindergarten affects maternal labour market outcomes Selina Gangl1·Martin Huber2 Received: 18 October 2023 / Accepted: 26 March 2025 / Published online: 16 April 2025 © The Author(s) 2025 Abstract The majority of Swiss children attend mandatory and cost-free kindergarten at age four. We examine the effect of this policy on maternal labour market outcomes. Using administrative data from Switzerland, we exploit the birthday cut-off for kindergarten entry in the same or in the following year and apply a non-parametric regression discontinuity design (RDD). We find that mandatory kindergarten has a statistically significant positive effect on the labour market attachment of previously non-employed mothers, increasing their employment probability by 4 percentage points. In contrast, there are no significant effects on other groups or in the total sample of mothers. Keywords Mandatory kindergarten ·Maternal employment ·Regression discontinuity design JEL Classification H40 ·J13 ·J18 ·J21 ·J22 1 Introduction The labour force attachment of mothers has been receiving much attention in scientific and public discussions, given the large employment drops observed in many countries after giving birth (see, for instance, Kleven et al. (2019)). In OECD countries, the average maternal employment rate was 71% in 2019, with Iceland, Slovenia, and Responsible editor: Kompal Sinha BSelina Gangl [email protected] Martin Huber [email protected] 1Department of Economics, Friedrich-Alexander-Universität Erlangen-Nürnberg, Lange Gasse 20, 90403 Nürnberg, Germany 2Department of Economics, University of Fribourg, Bd. de Pérolles 90, 1700 Fribourg, Switzerland 123 44 Page 2 of 23 S. Gangl, M. Huber Sweden at the top with rates exceeding 85%.1In Switzerland, which is the country consideredinthisstudy, mothers’ employmentratedeclinesafterchildbirth(Herrmann and Murier 2016), but it remains high at almost 78%, on average.2Although this figure is higher than the OECD average, most of it is due to part-time work (63%), rather than full-time work (15%).3 A policy tool to encourage mothers’ labour market outcomes is the provision of childcare. For example, Krapf et al. (2020) show that the opportunity to place the first child in childcare increases the maternal labour force participation in the canton Bern by 5.1 percentage points in the long term. In addition to the extensive margin, the birth of a child also affects the intensive margin of labour supply: mothers are more likely than women without children to work part-time (Herrmann and Murier 2016). The provision of childcare influences also mothers’ working hours positively: An expansion of childcare facilities raises maternal part-time hours (Ravazzini 2018), while after-school care increases mothers’ full-time employment (Felfe et al. 2016). Childcare availability may also increase maternal earnings, but not the earnings of the total household (Krapf et al. 2020). An unanswered question in this context is how mothers’ labour market responses react when childcare is mandatory. While kindergarten for four-year-olds has always been cost-free, it has recently become mandatory in most Swiss cantons. This paper examines the causal effect of mandatory kindergarten for four-year-old children on maternal labour behaviour and earnings in Switzerland. The aim of the policy, with regard to mothers, is to promote the compatibility of family and work. We exploit this novel setting and use the birthday cut-off to identify the causal effect of the mandatory kindergarten. The paper uses a large administrative dataset about the population of mothers with four-year-old children living in Switzerland. We link the Population and Households Statistics (STATPOP) with data about the Old-Age and Survivors’ Insurance (OASI). Moreover, we have collected data about the implementation year of the policy, the birthday cut-off dates per year, and the requirements regarding how many years of kindergarten must be offered. These data stem from the cantonal departments of education, the Swiss Conference of Cantonal Departments of Education (EDK), and the cantonal laws. The State Secretariat for Economic Affairs (SECO) provides information about the cantonal unemployment rate on their homepage. Again we link this cantonal information to the administrative data. We restrict the sample to mothers whose youngest child is four-year-old and who are living in a canton with mandatory kindergarten.4Methodologically, we exploit the random assignment at the birthday cut-off for entering the kindergarten in the same year (= treatment) versus the following year (= control) and apply a Regression Discontinuity Design (RDD). Since parents may postpone the kindergarten entry of their child by one year and we cannot observe the actual kindergarten entry, we rather assess the intention-to-treat (ITT) effect than the average treatment effect (ATE) at the threshold. 1Reported by OECD. https://www.oecd.org/els/family/database.htm, retrieved 2021-08-24. 2Reported by OECD. https://www.oecd.org/els/family/database.htm, retrieved 2021-08-24. 3Reported by OECD. https://www.oecd.org/els/family/database.htm, retrieved 2021-08-24. 4See Sect.4Data for details. 123 From homemakers to breadwinners? How mandatory… Page 3 of 23 44 We find that mandatory kindergarten boosts the labour market outcomes of previously non-employed mothers. In this subgroup, the probability of maternal employment increases by 4 percentage points on average across our outcome periods from 2010 to 2017 (statistically significant at the 5% level). Likewise, the annual income from dependent employment increases by 1030 CHF on average (statistically significant at the 10% level) and the probability of earning the income from dependent employment increases by 4 percentage points (statistically significant at the 5% level). Our results suggest that mandatory kindergarten has no significant effect on the labour market outcomes in the total sample of mothers. We argue that the high participation rate of mothers in the labour force, the half-day care in mandatory kindergarten, and the support of grandparents for childcare are likely reasons for this finding. Our paper fits into the literature on school policies and its impact on the family (see, e.g. Landersø et al. (2020)). In particular, our paper belongs to a group of studies focusing on maternal labour market outcomes. For example, Gelbach (2002) investigates the effect of school enrolment and Barua (2014) analyses the variation in school entry laws on maternal labour supply in the U.S. Finseraas et al. (2017) examine a decrease in the school starting age from seven to six years on maternal labour supply in Norway. The common result of these studies is that the school entry of the child raises mothers’ labour supply. In line with that, Zhu and Bradbury (2015) find that the birth of a child shortly after the birthday cut-off leads to a negative impact on mothers’ labour supply. Whereas the aforementioned studies focus on children who are starting school at age six or seven, our setting is unique because all four-year-olds are required to enter school, or in our case, mandatory kindergarten. Therefore, this paper contributes to the existing literature by examining the impetus for maternal reintegration into the workforce two years earlier due to mandatory kindergarten attendance of their child. Thisisparticularlyrelevant because industrialised countries are currently facing labour shortages that are expected to increase due to an ageing population and might be mitigated by working mothers (European Commission 2023). Moreover, an earlier reintegration in the labour force might decrease the risk of poverty in old age that mothers face due to interrupted working lives (United Nations 2022). This paper relates also to the family policy literature and its impact on maternal labour market behaviour (see, e.g. Kleven et al. (2020)). Childcare, specifically, may facilitate mothers’ participation in the workforce. Hence, one strand of literature examines whether the introduction of subsidised childcare like in Austria (Kleven et al. 2020), Canada (Lefebvre and Merrigan 2008), Norway (Hardoy and Schøne, 2015), Sweden (Lundin et al. 2008), and the Netherlands (Bettendorf et al. 2015) impacts maternal employment. Further research analyses the expansion of childcare places as done in Spain (Nollenberger and Rodríguez-Planas 2015), Italy (Carta and Rizzica 2018), Germany (Geyer et al. 2015), and Norway (Kunze and Liu 2019; Eckhoff Andresen and Havnes 2019), or the provision of a legal entitlement to childcare, as in Germany (Bauernschuster and Schlotter 2015). Evaluations of these policies suggest a significantly positive effect on mothers’ labour force participation, but only under the condition of a low labour force participation rate and a small extent of childcare prior to the reform. However, the availability of (subsidised) childcare can lead to a crowding out of private childcare, which may offset any increase in overall childcare 123 44 Page 4 of 23 S. Gangl, M. Huber provision and, as a result, fail to improve maternal labour market outcomes (Havnes and Mogstad 2011). For this reason, Goux and Maurin (2010) find only an effect of childcare availability on maternal labour supply for the subgroup of single mothers in France. Whilethis literature analyses the impact of childcare on the labour marketbehaviour of a self-selected group of mothers who choose to enrol their children in childcare, it does not encompass the effect on all mothers. Our paper addresses this inquiry by examining whether mandatory kindergarten incentivizes all mothers to participate in the labour market and the resulting impact on their wages. This information seems to be crucial for policy makers seeking to promote maternal labour force participation. The findings of our study align with those of Havnes and Mogstad (2011), Goux and Maurin (2010), and Lundin et al. (2008), indicating hardly no impact of childcare on maternal labour market outcomes. In comparison, Berlinski and Galiani (2007) and Lefebvre and Merrigan (2008) report significant positive effects. These divergent conclusions may partly stem from variations in the socio-economic settings of the studied countries. For instance, a high rate of maternal employment, such as that observed in Sweden, could elucidate the near-negligible effect of childcare policies Lundin et al. 2008. Similarly, in Switzerland, the focus of our study, where maternal employment rates hover around 80%, the presence of mandatory kindergarten shows only discernible effects on labour market outcomes of previously non-employed mothers. Another feature of the Swiss setting is that working parents mainly use the support of grandparents for childcare (Jacobs Foundation 2018). The insignificant impact of mandatorykindergarten onlabourmarketresponseinthetotalsampleofmotherscould potentially be attributed to a crowding out effect of informal childcare, as suggested by Havnes and Mogstad (2011). Different policy designs may also explain the mixed results in the studies. For example, Lefebvre and Merrigan (2008) analyse the effect of full-time childcare and find significant and large effects. The mandatory kindergarten is only half-day care in most Swiss cantons, so mothers are often faced with the challenge of balancing their work around the limited kindergarten hours or finding another (costly) care option. This circumstance could be a possible explanation for the fact that mandatory kindergarten does not have a far-reaching impact on labour market outcomes in the total sample of mothers. The remainder of this paper is organised as follows: In the next section, we provide information about the preschool policy in Switzerland. Thereafter, we present the data source, descriptive statistics, and define the sub-sample. Then, we discuss the empirical strategy. In the following section, we report the results and the robustness checks. Finally, we draw a conclusion from the empirical analysis. 2 Institutional background Thissectionprovidesanoverviewofthemandatoryand cost-freekindergarten forfouryear-old children in Switzerland. Education policy in Switzerland is decentralised, with each of the 26 cantons having autonomy over the quality and structure of education. This decentralisation leads to variations in aspects such as the school curriculum, the age of entry into kindergarten, and the number of school years (SWI 2006). Hence, 123 From homemakers to breadwinners? How mandatory… Page 5 of 23 44 each canton decides about the implementation of the mandatory kindergarten. Table A.1 in the Online Appendix shows that Basel-Stadt was the first canton to introduce mandatory kindergarten for four-year-olds in 2005. In 2006, a national referendum about an educational reform took place, which was accepted by 86% of the voters. One year later, an inter-cantonal “HarmoS” concordat was established, 15 cantons entered into the concordat, while seven cantons rejected it and four cantons are still indecisive (as of 2019). The main goal of the concordat is to harmonise the mandatory school education among the cantons. This agreement also includes an age decrease for mandatory school attendance. All children turning four before the first of August must enter kindergarten (Schweizerische Konferenz der kantonalen Erziehungsdirektoren 2007). Therefore, other cantons followed Basel-Stadt and introduced mandatory kindergartens. But also in this case, neither the implementation year nor the shift of the birthday cut-off date is regulated on a central level, see Table A.1 in the Online Appendix for an overview. In 2017, almost two-thirds of the cantons had enforced mandatory kindergarten attendance for four-year-olds.5In these cantons, children who turn four before or on the cutoff date must enter kindergarten in the same year, while four-year-olds born after the cutoff date do not enter kindergarten until the following year. Table A.1 in the Online Appendix shows that these birthday cut-off dates vary from year to year and canton to canton and settle on 31.7. The reason for this shift is the joint decision by the cantons to determine a uniform cut-off date. Most cantons have decided to gradually shift the cut-off date so that the rush to kindergartens remains roughly the same during the transition period and does not increase particularly sharply in one year and overload the infrastructure. The twofold aim of the mandatory kindergarten is to promote the development of children, but also to encourage the compatibility of family and work (Schweizerische Konferenz der kantonalen Erziehungsdirektoren 2014). In this paper, we focus on the latter and examine whether the obligation to attend kindergarten at the age of four affects mothers’ work behaviour. Hence, the length of the mandatory kindergarten may be crucial for the maternal work decision. The kindergarten hours differ from canton to canton: At the lower bound is the canton Valais, where four-year-old children are supposed to attend 12 lessons per week,6each lesson is 45 minutes, i.e. the children spend in sum 9hours per week in the kindergarten. At the upper bound of kindergarten hours is the canton Ticino, where children attend the kindergarten on four days from the morning till the afternoon and an additional half-day. The mandatory kindergarten is part of early childhood education, which takes place prior to primary education (OECD 2015) and corresponds to level 0 in the International Standard Classification of Education (ISCED) (Eurydice 2016/17). In the Germanand Italian-speaking parts of Switzerland, kindergarten is rather a separate institution, whereas in the Frenchspeaking part, kindergarten is represented as two additional classes in primary school (Schweizerische Konferenz der kantonalen Erziehungsdirektoren 2016). 5Aargau, Basel-Land, Basel-Stadt, Bern, Fribourg, Geneva, Glarus, Jura, Neuchâtel, St Gallen, Schaffhausen, Solothurn, Thurgau, Ticino, Vaud, Valais, and Zurich. 6EDK Kantonsumfrage 2016/17. 123 44 Page 6 of 23 S. Gangl, M. Huber Red-shirting behaviour of the parents, i.e. deferring a four-year-old child from kindergarten, is possible, but is associated with hurdles that vary from canton to canton. In some cantons, an application by the parents is sufficient, in other cantons a medical certificate is required which proves that the admission of a child to the kindergarten would not be reasonable even with other educational assistance.7 3Hypotheses As the literature review reveals, childcare reforms as well as a lower school starting age affect mothers’ work behaviour. In the former case, a subsidy either reduces the financial burden or provides more childcare places. In the latter case, the effect comes from the obligation to attend school at a younger age. In contrast, we evaluate the impact of the mandatory kindergarten on mothers’ work behaviour. Therefore, the analysis combines features from both literature strands: The kindergarten is an obligation and is free of charge like in the school starting age literature. On the other side, we evaluate a policy that affects children in the childcare age. This chapter presents the hypothesis and the possible channels. We evaluate the effect of the mandatory kindergarten on maternal labour market behaviour within a specific age window: We compare mothers with slightly older children who were born prior to a specific birthday cut-off and are supposed to enter kindergarten in the same year (= treated) with mothers whose children are slightly younger, born after a specific birthday cut-off and who are supposed to enter kindergarten in the following year (= control).8Mothers with a slightly younger child may spend their time by taking care of their child or must organise a (costly) childcare to allocate their time to work. Contrarily, mothers, whose child is supposed to enter kindergarten, can spend their time on other activities, for example, labour market participation. Consequently, a mother who would take care of her child in the absence of thereform can rearrange her time: Following the model ofthe allocation of time Becker (1965), the mother can spend more time in market work and/or leisure activities. Swart et al. (2019) find that mothers increase their labour supply after the school enrolment of their youngest child, given that they take care of their children prior to the school start. Therefore, we expect in our setting that only one subgroup of mothers increases their labour supply: Mothers whose children are supposed to enter kindergarten, but who would take care of them if no policy is implemented. All in all, we hypothesise a positive, but small effect because the effect is zero for mothers who would send their child to kindergarten anyway. 4 Data Our evaluation strategy is a non-parametric regression discontinuity design (RDD) to examine the effect of the reform at the birthday cut-off for entering the kindergarten 7https://www.swissmom.ch/de/kind/kindergarten-und-schule/mit-4-jahren-schon-in-den-kindergarten19063. 8One exception is the canton of Ticino that offers kindergarten for three-year-old children. 123 From homemakers to breadwinners? How mandatory… Page 7 of 23 44 in the same versus in the following year. It relies on a large administrative dataset including the universe of mothers with four-year-old children living in Switzerland, the so-called Population and Households Statistics (STATPOP). We link the latter with data coming from the old-age and survivors’ insurance (OASI) in Switzerland by a unique identifier. STATPOP provides information on the stock and structure of the Swiss resident population from 2010 to 2018. This enables us to examine the population of mothers whose youngest child is four-year-old and living in a canton with mandatory kindergarten. Furthermore, the data contain personal characteristics from the mother, their children, and the father of the four-year-old child. Two variables are central for our regression discontinuity approach (RDD): Firstly, the canton of residence to determine whether the four-year-old lives in a canton with mandatory kindergarten in the current year. Secondly, the exact date of the birth of the four-yearold determines whether a child was born prior (= treated) or after (= control) a specific birthday cut-off date. Additionally, other personal characteristics such as the marital status, country of birth, resident permit of the mother, father, and the children are also contained in the dataset. The OASI data contain information about the income, unemployment benefits, disability indemnities for every insured person in Switzerland from 2010 to 2017. Based on these data, we calculated the “Income from work” as the total income minus transfer payments (for instance, unemployment insurance and welfare benefits) and minus other indemnities (like the compensation for military service) for the following subgroups: (i) employees whose employer resides in Switzerland, (ii) voluntary insured employees whose employer is not liable for contributions in Switzerland, and (iii) selfemployed persons.9Subsequently, we generated the variable “In labour force”, which is 1 if a person earns income from work or if she receives unemployment benefits and 0 otherwise. We calculated the dummy “Out of labour force” as 1 minus the value of “In labour force”. Furthermore, we constructed the dummy variable “Employed”, which turns 1 if the person earns an income from work and participates in the labour market and is 0 otherwise. Vice versa, a person is indicated as “Unemployed” if she participates in the labour force but is not employed. The variable “Income from dependent employment” gives the sum of the income for mothers working in dependent employment, whereas the variable “Income from dependent employment (binary)” indicates whether a mother works in dependent employment. In the last step, we matched this database with cantonal information about the unemployment rates, the yearly birthday cut-off date of the mandatory kindergarten, and an indicator of whether a canton joined the “HarmoS” concordat. The “HarmoS” concordat is an association of some cantons to harmonise mandatory education like the uniform school entry age.10 The database contains a total of 687,749 mothers with at least one four-year-old child between 2010 and 2018. Since the last period in the OASI data is 2017, we drop the STATPOP observations from 2018 such that 606,434 observations remain. Another data restriction is based on the empirical literature, because childcare seems to not affect the mothers’ labour market outcomes when there is a younger sibling of the four-year-old child in the household (see, for example, Berlinski and Galiani (2007), 9excluding agriculture. 10 More information about the institutional background can be found in Online Appendix B.1. 123 44 Page 8 of 23 S. Gangl, M. Huber Fig. 1 Timeline of measured variables Note: The timeline shows the time points at which the variables were measured relative to the treatment Fitzpatrick (2012), and Gelbach (2002)). We overcome this concern by restricting the subsample to mothers whose youngest child is four years old, which decreases the sample size to 360,234 mothers. Three cantons with mandatory kindergarten did not determine a birthday cut-off date, hence we exclude observations from Aargau, Fribourg prior to 2013, and all the municipalities of the canton Bern except the federal city of Bern, after that 289,728 observations are in the sample. Furthermore, we drop observations with negative income entries11 in the OASI data and keep 289,631 observations. We exclude observations, which live in cantons without a mandatory kindergarten for four years old. Since two sources12 reported two consecutive days as birthday cut-off dates in Basel-Stadt, we took the later cut-off date and excluded observations on the previous day from our analysis.13 Finally, 201,993 observations remain in our evaluation dataset. Figure1providesatimeline for the measurement of the keyvariables in our analysis, with t denoting a specific year. The running variable will be henceforth denoted by D and depicts whether the child turns four prior (D= 1) or after the cut-off date (D= 0), measured in the baseline period (t= 0). The outcome periods start in the year the child turns four (t= 0) and end five years later (t= 5). Personal characteristics (e.g. number of children), are measured one year before the child turns four (t=−1). We check the balance of labour market relevant covariates as suggested in Lee (2008), by running several RDD estimations where each covariate serves as an outcome. Table 1reports the sample mean, the balance checks, and the relative number of missings for each variable at the threshold. Mothers’, fathers’, and cantonal covariates are generally well balanced, which points to a quasi-random assignment of mandatory kindergarten at the threshold. We also see from the table that mothers close to the threshold have on average 1.87 children, are predominately born in Switzerland, and live in a relationship. Furthermore, 73% of these mothers are employed, earn on average 32,760 Swiss francs (CHF) per year, which mainly stem from dependent employment. Fathers labour market characteristics show with 90% a very high employmentrate,anannualmeangross income of 94,342CHF,and86%of this income comes from dependent employment. The cantonal characteristics reveal a rather low mean unemployment rate (3.6%) and a high probability of participating in the HarmoS concordat (94%). Table 2reports the balance checks for cantonal dummies at the threshold. Due to the different birthday cut-off dates in the Germanand the French-speaking part of the canton Valais, we analyse Upper Valais and Central/Lower Valais separately. 11 for example due to adjustment entries. 12 see “Regierungsratsbeschluss betreffend Stichtag für die Einschulung für die Schuljahre 2011/12 bis 2015/16 (§56 Abs. 1 Schulgesetz)” and “EDK/IDES-Kantonsumfrage” for the years 2011 to 2016. 13 to run a Donut-RDD for this canton. 123 From homemakers to breadwinners? How mandatory… Page 15 of 23 44 care of their children (Jacobs Foundation 2018). The mandatory kindergarten may have reduced the necessity of informal childcare options, but this did not affect the overall impact on mothers’ labour supply. The effect of mandatory kindergarten might be also diminished by the presence of formal childcare services. However, daycare in Switzerland is known to be expensive compared to other countries (Stern and Felfe 2015), and there is a substantial demand surplus for childcare places (Bundesamt für Sozialversicherungen 2006). These factors contribute to a lower daycare attendance rate, with only 36% of children under the age of four currently attending daycare (Jacobs Foundation 2018). Finally, the labour force supply of mothers in Switzerland is already high, which might also explain the lack of a sensible effect of mandatory kindergarten on maternal labour market participation in the total sample of mothers. This finding is in line with other studies that have found either no effect or only a modest effect of childcare on maternal employment in the context of high maternal labor force participation (see, e.g. Lundin et al. (2008) and Havnes and Mogstad (2011)). An important factor to consider when examining the lack of impact on maternal income is the limited flexibility provided by half-day mandatory kindergarten. With the majority of cantons offering half-day care, mothers are often faced with the challenge of balancing their work schedules around the limited hours of kindergarten. This constraint can significantly hinder their ability to pursue full-time employment opportunities. Furthermore, the reliance on alternative childcare arrangements for the rest of the day, such as private daycare, can come at an additional cost to mothers. These expenses, combined with the logistical challenges of coordinating multiple care providers, can create financial and logistical burdens that further discourage mothers from pursuing full-time employment. Therefore, the lack of access to full-day care following mandatory half-day kindergarten not only restricts the hours available for maternal employment but also introduces financial and logistical barriers that impede the potential increase in maternal income. We check the robustness of our ITT effects by estimating the RDD approach without covariates (Table A.2 in the Online Appendix) and find rather similar point estimates, yet the precision decreases. Another check consists of varying the optimal bandwidth, we multiplied this by 1.5 and 2/3 (Tables A.3 and A.4 in the Online Appendix). Our findings indicate that the effect sizes remain relatively stable, with the exception of the effects on “Annual income from dependent employment”, as shown in Tables A.3 and A.4 in the Online Appendix, and “Annual work income” as presented in Table A.4 in the same Appendix. 6.2 Effects over years In the next step, we analyse the effects in separate and consecutive outcome periods, relative to the year the child turns four. Figure3shows the results of the different outcomes, starting in period 0 (i.e. the year the child turns four) and ending in period 5. The dots depict the ITT at the cut-off and the bands represent the point-wise 95% confidence intervals based on robust standard errors. We do not find any statistically significant effect for the outcomes in comparison to the pooled estimates. This might be caused by lower power due to a decrease in the number of observations. Moreover, we find unstable point estimates, both the sign and the effect size change over periods. 123 44 Page 16 of 23 S. Gangl, M. Huber Fig. 3 RDD: Effects over years with covariates Note: This figure shows the effect of mandatory kindergarten on each outcome variable separately and over time. Dots represent ITTs at the threshold, bands correspond to 95% confidence intervals. The following control variables are included: Number of children, born in Switzerland, resident permit, age, relationship status, labour market characteristics of the mother and the father of the four-year-old child, cantonal characteristics, and dummies. Data stems from STATPOP (2010– 2017) and OASI (2010–2017), calculations are done by the authors. Income is deflated, the base year is 2011. The official currency in Switzerland is the Swiss Franc (CHF), which had an average exchange rate of 1.04 USD/CHF in the last decade The plots reveal further that the effects are, especially in the later periods, imprecisely estimated because of the decrease in the number of observations. Figure A.7 in the Online Appendix depicts the plots without covariates and shows similar results in terms of significance, point estimates, and precision. 6.3 Heterogeneous effects From a policy perspective, it is interesting to know whether the effect of mandatory kindergarten differs for specific subgroups. Non-employed mothers may react differently to the kindergarten policy than employed mothers. Therefore, we split the sample according to the maternal labour force status one year before the youngest child turns four. 123 From homemakers to breadwinners? How mandatory… Page 17 of 23 44 First, we start with the presentation of the tests for the RDD assumptions. Tables A.5 and A.6 in the Online Appendix show, similar as for the entire sample in Table 1, that the labour market relevant covariates are balanced for both sub-samples. Analogous to the balance checks for cantonal dummies of the entire sample in Table 2, we conclude from Tables A.7 and A.8 that in most cases, our tests do not reject the null hypothesis that the cantonal dummies are balanced around the birthday cut-off. In contrast to the result of the entire sample in Fig.2, the results of the Frandsen’s test in Figs. A.9 and A.10 in the Online Appendix suggests evidence for manipulation at the cut-off if we assume that the distribution around the cut-off must be precisely linear (k=0), but does not point to manipulation if we allow for non-linearities (k= 0.1 and k= 0.2). We follow the argumentation of Oosterbeek et al. (2021), which states that the RDD becomes invalidated by manipulation only if there exist systematic disparities between mothers on either side of the birthday cut-off. However, Tables A.5 and A.6 in the Online Appendix reveal no discernible association between being born slightly after versus before the birthday cut-off and the predetermined covariates. Second, we report the estimation results. Table 5represents the results for employed mothers, whereas Table 6reports the effects for non-employed mothers. The results suggest that the effect is driven by non-employed mothers who start working when their youngest, one-year-old child is required to attend kindergarten. Three effects in Table 6remain statistically significant at the 10% level after applying the BenjaminiHochberg procedure.17 To further validate the robustness of our findings, we conduct an RDD with a fake cut-off set 30 days earlier than the actual threshold, and the results in Table A.9 in the Online Appendix indicate that the effects are insignificant. Additionally, we employ a Donut-RDD approach in Table A.10 in the Online Appendix, which excludes observations with running variables close to the cut-off (values between −2 and 1). The results obtained under this design confirm that all estimated effects maintain the same direction as in the main analysis. However, the statistical significance of the effects is generally reduced. Specifically, only the effects on employment, out of the labour force, and income-dependent employment exhibit p-values close to 10%, while the remaining effects show p-values that are considerably beyond any conventional levels of statistical significance. Third, we provide numbers on the effect of the kindergarten policy on reducing the labour shortage. As shown in Table 6, the effect of mandatory kindergarten on employment is an increase of 4 percentage points for the subgroup of previously nonemployed mothers. If we extrapolate this local effect to all 179,895 non-employed mothers with four-year-olds in our sample, mandatory kindergarten results in 7196 additional employed mothers. When this number is compared to the total number of employees in Switzerland in the first quarter of 201018 it corresponds to an increase in aggregate employment of 0.002 percentage points. Comparing these 7196 additional employed mothers to the job vacancies in Switzerland in the first quarter of 201019 we 17 We inspect also the heterogeneous effects for older versus younger mothers as well as divorced mothers and do not find statistically significant effects. 18 4,530,917; Federal Statistical Office: https://www.bfs.admin.ch/bfs/en/home/statistics/industryservices/businesses-employment/jobs-statistics/jobs.assetdetail.32008419.html, 19 51,162; Federal Statistical Office: https://www.bfs.admin.ch/bfs/en/home/statistics/industry-services/ businesses-employment/jobs-statistics/vacancies.assetdetail.32008426.html, 123 44 Page 18 of 23 S. Gangl, M. Huber Table 5 RDD: heterogeneous effects: employed mothers Sample mean Coefficient Standard error p-value Bandwidth Observations within bandwidth Out of labour force (binary) 0.06 0.00 0.01 0.58 67.03 183,521 Employed (binary) 0.93 −0.00 0.01 0.72 67.66 183,521 Unemployed (binary) 0.01 −0.00 0.00 0.54 51.19 138,858 Annual work income (in CHF) 46,460.42 −531.87 698.65 0.45 72.22 196,930 Income dependent employment (binary) 0.90 0.00 0.01 0.96 61.60 167,217 Annual income from dependent employment (in CHF) 44,347.23 −606.55 650.57 0.35 62.65 170,025 Note: This table reports the local linear estimates of Eq.5.2 only for previously employed mothers. Bandwidth shows the MSE-optimal bandwidth chosen by (Calonico et al. 2021). The following control variables are included: Number of children, born in Switzerland, resident permit, age, relationship status, labour market characteristics of the mother and the father of the 4-year-old child, cantonal characteristics, and dummies. The standard errors are clustered at the individual level. Data stems from STATPOP (2010–2017) and OASI (2010–2017), calculations are done by the authors. Income is deflated, the base year is 2011. The official currency in Switzerland is the Swiss Franc (CHF), which had an average exchange rate of 1.04 USD/CHF in the last decade. Number of observations in total: 484,153 123 From homemakers to breadwinners? How mandatory… Page 19 of 23 44 Table 6 RDD: heterogeneous effects: non-employed mothers Sample mean Coefficient Standard error p-value Bandwidth Observations within bandwidth Out of labour force (binary) 0.70 −0.04 0.02 0.05 59.46 59,463 Employed (binary) 0.30 0.04 0.02 0.04 58.26 58,459 Unemployed (binary) 0.01 −0.00 0.00 0.39 59.39 59,463 Annual work income (in CHF) 5582.01 936.74 652.42 0.15 55.40 55,452 Income dependent employment (binary) 0.28 0.04 0.02 0.03 57.77 57,634 Annual income from dependent employment (in CHF) 4789.60 1030.41 596.34 0.08 54.96 54,394 Note: This table reports the local linear estimates of Eq.5.2 only for previously non-employed mothers. Bandwidth shows the MSE-optimal bandwidth chosen by (Calonico et al. 2021). The following control variables are included: Number of children, born in Switzerland, resident permit, age, relationship status, labour market characteristics of the mother and the father of the 4-year-old child, cantonal characteristics, and dummies. The standard errors are clustered at the individual level. Data stems from STATPOP (2010–2017) and OASI (2010–2017), calculations are done by the authors. Income is deflated, the base year is 2011. The official currency in Switzerland is the Swiss Franc (CHF), which had an average exchange rate of 1.04 USD/CHF in the last decade. Number of observations in total: 179,895 123 44 Page 20 of 23 S. Gangl, M. Huber seeareduction in vacantpositions of 0.14 percentage points. We assumethatthe effects can be extrapolated. Therefore, it should be noted that this is only an approximation subject to errors. It’s also worth noting that the administrative dataset lacks information on working hours or whether mothers work fullor part-time, making it impossible to calculate the employment increase in full-time equivalents. Additionally, we exploit the spatial variation and the staggered implementation of the mandatory kindergarten in a Difference-in-Differences Approach. The description of the data, method, and results are presented in the Online Appendix. We find that labour market participation, employment, and the part-time rate of mothers increase because of mandatory kindergarten. However, the precision of these results is low due to a small sample size. 7 Conclusion This paper analyses the impact of mandatory and cost-free kindergarten for four-yearold children on maternal labour market behaviour using Swiss administrative data. Our identification strategy exploits birthday cut-off dates as random assignment of the mandatory kindergarten and compares labour market responses for mothers whose children enter kindergarten in the current year versus the following year. We apply a non-parametric regression discontinuity design (RDD) and find statistically significant effects on the employment and earnings of previously non-employed mothers, but not of the total sample of mothers. Our findings contribute to the literature on childcare and school entry by analysing the mandatory and free of charge kindergarten. Current literature suggests that attending childcare or school affects maternal labour force participation positively in countries with low maternal labour force participation and low childcare attendance rate (see, e.g. Nollenberger and Rodríguez-Planas (2015) or Bauernschuster and Schlotter (2015)). In Switzerland, maternal labour force participation exceeds the OECD average. Our study’s result that mandatory kindergarten only sensibly affects labour market decisions of previously non-employed mothers, but not of other groups, is in line with no or only modest effects of childcare on maternal employment found in countries with a high maternal labour force participation (see, e.g. Lundin et al. (2008) and Havnes and Mogstad (2011)). Furthermore, working parents mainly use the support of grandparents for childcare in Switzerland (Jacobs Foundation 2018). The insignificant impact of mandatory kindergarten on labour market response in the total sample of mothers might also be explained by a crowding out effect of informal childcare (see, e.g. Havnes and Mogstad (2011)). Since mandatory kindergarten is half-day care in most cantons, mothers are often faced with the challenge of balancing their work around the limited kindergarten hours or finding another (costly) care option. This circumstance might potentially account for the absence of a widespread impact of mandatory kindergarten on the earnings in the total sample of mothers. For future research, it would be interesting 123 From homemakers to breadwinners? How mandatory… Page 21 of 23 44 to examine whether a mandatory full-day kindergarten impacts maternal labour force participation and earnings. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1007/s00148-025-01099-y. Acknowledgements We thank the editor, Kompal Sinha, and three anonymous referees for helpful comments and suggestions. The study has been realised using the data from the Federal Statistical Office (FSO) and the Central Compensation Office (CCO). We are grateful for the data provision as well as the linking and anonymizing of the data conducted by the FOS. Furthermore, we thank Maryna Ivets, Laura Ravazzini, Regina T. Riphahn, Jürg Schweri, and Andreas Steinmayr for their helpful suggestions on a prior version of this paper. We have benefited from comments by seminar participants in Fribourg, the Ski & Labour Seminar in Engelberg-Titlis, the 9th ifo Dresden Workshop on Labour Economics and Social Policy, the 10th International Conference of Panel Data Users in Switzerland in Lausanne, the Swiss Society of Economics and Statistics Annual Meeting in Geneva, the EEA in Manchester, the 3rd International BFH Conference on Discrimination in the Labor Market in Bern, the Verein für Socialpolitik - Annual Conference 2019, the 1st and 4th Workshop of the Swiss Network on Public Economics (SNOPE), and the Lunch-time Meeting at the University of Innsbruck 2021, COMPIE 2022, European Association of Labour Economists (EALE) 2024, 37th Annual Conference of the European Society for Population Economics (ESPE), Workshop on Labour Economics 2024, Trier, Germany and the Preand Postdoc Seminar at the Friedrich-Alexander-University Erlangen-Nürnberg. Funding Open Access funding enabled and organized by Projekt DEAL. Data availability This paper uses confidential data from the Federal Statistical Office (FSO) and Old-Age and Survivor’s Insurance (OASI) in Switzerland. The data can be obtained by filing a request directly with the FSO (https://www.bfs.admin.ch/bfs/de/home/dienstleistungen/datenverknuepfungen/fuer-dritte.html). Declarations Conflict of interest The authors declare no competing interests. Open Access This articleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. References Barua R (2014) Intertemporal substitution in maternal labor supply: Evidence using state school entrance age laws. Labour Econ 31:129–140. http://www.sciencedirect.com/science/article/pii/ S0927537114000852,https://doi.org/10.1016/j.labeco.2014.07.002 Bauernschuster S, Schlotter M (2015) Public child care and mothers’ labor supply–evidence from two quasi-experiments. J Pub Econ 123:1–16. https://www.sciencedirect.com/science/article/pii/ S004727271500002X,https://doi.org/10.1016/j.jpubeco.2014.12.013 Becker GS (1965) A theory of the allocation of time. Econ J 75:493–517. http://www.jstor.org/stable/ 2228949 Berlinski S, Galiani S (2007) The effect of a large expansion of pre-primary school facilities on preschool attendance and maternal employment. Labour Econ 14:665–680 123 44 Page 22 of 23 S. Gangl, M. Huber Bettendorf LJ, Jongen EL, Muller P (2015) Childcare subsidies and labour supply — evidence from a large Dutch reform. Labour Econ 36:112–123 Bundesamt für Sozialversicherungen 2006 Botschaft zum Bundesbeschluss über Finanzhilfen für familienergänzende Kinderbetreuung. https://www.fedlex.admin.ch/eli/fga/2006/357/de Calonico SD, CM, Farrell MH, Titiunik R (2021) Robust data-driven statistical inference in regressiondiscontinuity designs. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web& cd=&cad=rja&uact=8&ved=2ahUKEwiywffQiqvyAhWG_rsIHRdnA60QFnoECAgQAQ& url=https%3A%2F%2Fcran.r-project.org%2Fweb%2Fpackages%2Frdrobust%2Frdrobust.pdf& usg=AOvVaw1D9Rf0BUIsPFdj2akTOvar Carta F, Rizzica L (2018) Early kindergarten, maternal labor supply and children’s outcomes: evidence from italy. J Public Econ 158:79–102. http://www.sciencedirect.com/science/article/pii/ S0047272717302141,https://doi.org/10.1016/j.jpubeco.2017.12.012 Eckhoff Andresen M, Havnes T (2019) Child care, parental labor supply and tax revenue. Labour Econ 61:101762. https://www.sciencedirect.com/science/article/pii/S0927537119300880,https://doi.org/ 10.1016/j.labeco.2019.101762 European Commission 2023 Commission report finds labour and skills shortages persist and looks at possible ways to tackle them. https://ec.europa.eu/social/main.jsp?langId=en&catId=89& furtherNews=yes&newsId=10619#navItem-3. Accessed on 05 Oct 2023 Eurydice (2016/17) Compulsory education in Europe – 2016/17. Eurydice facts and figures. https://publications.europa.eu/en/publication-detail/-/publication/2f15cd79-9a83-11e6-9bca01aa75ed71a1/language-en. Accessed on 13 Aug 2019 Federal Statistical Office, 2022. Umzugsquote der ständigen wohnbevölkerung nach kanton. https:// www.bfs.admin.ch/bfs/en/home/statistics/construction-housing/dwellings/home-moves.html, note=Accessed on 14 Feb 2024 Felfe C, Lechner M, Thiemann P (2016) After-school care and parents’ labor supply. Labour Econ 42:64–75. http://www.sciencedirect.com/science/article/pii/S0927537116300616 Finseraas H, Hardoy I, Schøne P (2017) School enrolment and mothers’ labor supply: evidence from a regression discontinuity approach. Rev Econ Household 15:621–638 Fitzpatrick MD (2012) Revising our thinking about the relationship between maternal labor supply and preschool. J Hum Resourc 47:583–612 Foundation Jacobs (2018) Whitepaper zur Vereinbarkeit von Familie und Beruf: Zwischen Wunsch und Realität. Technical Report, Jacobs Foundation Frandsen B (2017) Party bias in union representation elections: Testing for manipulation in the regression discontinuity design when the running variable is discrete. Adv Econ 38:281–315. https://doi.org/10. 1108/S0731-905320170000038012 Gelbach JB (2002) Public schooling for young children and maternal labor supply. Am Econ Rev 92:307– 322. http://www.jstor.org/stable/3083335 Gelman A, Imbens G (2019) Why high-order polynomials should not be used in regression discontinuity designs. J Bus Econ Stat 37:447–456. https://doi.org/10.1080/07350015.2017.1366909 Geyer J, Haan P, Wrohlich K (2015) The effects of family policy on maternal labor supply: combining evidence from a structural model and a quasi-experimental approach. Labour Econ 36:84– 98. http://www.sciencedirect.com/science/article/pii/S0927537115000755,https://doi.org/10.1016/j. labeco.2015.07.001 Goux D, Maurin E (2010) Public school availability for two-year olds and mothers’ labour supply. Labour Econ 17:951–962. http://www.sciencedirect.com/science/article/pii/S0927537110000576, https://doi.org/10.1016/j.labeco.2010.04.012 Hahn J, Todd P, der Klaauw WV (2001) Identification and estimation of treatment effects with a regressiondiscontinuity design. Econometrica 69:201–209. http://www.jstor.org/stable/2692190 Hardoy I, Schøne P (2015) Enticing even higher female labor supply: the impact of cheaper day care. Rev Econ Household 13:815–836 HavnesT,Mogstad M(2011) Moneyfornothing? Universalchildcare andmaternal employment.JPubEcon 95:1455–1465. https://www.sciencedirect.com/science/article/pii/S0047272711000880,https://doi. org/10.1016/j.jpubeco.2011.05.016. special Issue: International Seminar for Public Economics on Normative Tax Theory Herrmann AB, Murier T (2016) Schweizerische Arbeitskräfteerhebung - Mütter auf dem Arbeitsmarkt. Bundesamt für Statistik, Neuchâtel https://www.bfs.admin.ch/bfs/de/home/statistiken/katalogedatenbanken/publikationen.assetdetail.1061095.html 123 From homemakers to breadwinners? How mandatory… Page 23 of 23 44 Kleven H, Landais C, Posch J, Steinhauer A, Zweimüller J (2019) Child penalties across countries: evidence and explanations. AEA Papers Proc 109:122–26. https://www.aeaweb.org/articles?id=10.1257/pandp. 20191078,https://doi.org/10.1257/pandp.20191078 Kleven H, Landais C, Posch J, Steinhauer A, Zweimüller J (2020) Do family policies reduce gender inequality? Evidence from 60 years of policy experimentation. Working Paper 28082. National Bureau of Economic Research. http://www.nber.org/papers/w28082,https://doi.org/10.3386/w28082 Krapf M, Roth A, Slotwinski M (2020) The effect of childcare on parental earnings trajectories. https://www.cesifo.org/en/publikationen/2020/working-paper/effect-childcare-parental-earningstrajectories. cESifo Working Papers Kunze A, Liu X (2019) Universal childcare for the youngest and the maternal labour supply. Technical Report. CESifo Working Paper No. 7509 Landersø RK, Nielsen HS, Simonsen M (2020) Effects of school starting age on the family. J Hum Resourc 55:1258–1286. https://jhr.uwpress.org/content/55/4/1258,https://doi.org/10.3368/jhr.55.4. 1117-9174R1 Lee DS, Lemieux T (2010) Regression discontinuity designs in economics. J Econ Lit 48:281–355. https:// www.aeaweb.org/articles?id=10.1257/jel.48.2.281,https://doi.org/10.1257/jel.48.2.281 Lee DS (2008) Randomized experiments from non-random selection in U.S. house elections. J Econometrics 142:675–697. https://www.sciencedirect.com/science/article/pii/S0304407607001121,https:// doi.org/10.1016/j.jeconom.2007.05.004. the regression discontinuity design: Theory and applications Lefebvre P, Merrigan P (2008) Child-care policy and the labor supply of mothers with young children: a natural experiment from Canada. J Law Econ 26:519–548 Lundin D, Mörk E, Öckert B (2008) How far can reduced childcare prices push female labour supply? Labour Econ 15:647–659 Nollenberger N, Rodríguez-Planas N (2015) Full-time universal childcare in a context of low maternal employment: Quasi-experimental evidence from spain. Labour Econ 36:124– 136. http://www.sciencedirect.com/science/article/pii/S0927537115000238,https://doi.org/10.1016/ j.labeco.2015.02.008 OECD (2015) Isced 2011 level 0: early childhood education, in isced 2011 operational manual: guidelines for classifying national education programmes and related qualifications. https://www.oecd-ilibrary. org/docserver/9789264228368-4-en.pdf?expires=1565703850&id=id&accname=ocid56025002& checksum=FABD3CBC97D44F307409C640E5CCAC30. Accessed on 13 Aug 2020 Oosterbeek H, ter Meulen S, van der Klaauw B (2021) Long-term effects of school-starting-age rules. Econ Educ Rev 84:102144. https://www.sciencedirect.com/science/article/pii/S0272775721000637, https://doi.org/10.1016/j.econedurev.2021.102144 RavazziniL(2018) Childcare andmaternalpart-time employment:a natural experimentusing swisscantons. Swiss J Econ Stat 154:15. https://doi.org/10.1186/s41937-017-0003-x Schweizerische Konferenz der kantonalen Erziehungsdirektoren (2007) Interkantonale Vereinbarung über dieHarmonisierungderobligatorischenSchule(HarmoS-Konkordat)https://edudoc.ch/record/24711/ files/HarmoS_d.pdf, Accessed on 30 Jan 2019 Schweizerische Konferenz der kantonalen Erziehungsdirektoren (2014) Obligatorische schule: Schuleintritt und erste jahre. https://edudoc.ch/record/111988/files/schuleintritt_d.pdf. Accessed on 30 Jan 2019 Schweizerische Konferenz der kantonalen Erziehungsdirektoren (2016) Kurz-info: Obligatorische schule: Schulstufen, zählweise der schuljahre. https://www.edudoc.ch/static/web/arbeiten/sprach_unterr/ kurzinfo_zaehlweise_d.pdf. Accessed on 30 Jan 2019 Stern S, Felfe C (2015) Krippenkosten und -finanzierung im internationalen Vergleich. Technical Report, Soziale Sicherheit Swart L, van den Berge W, van der Wiel K (2019) Do parents work more when children start school? Evidence from the Netherlands. Technical Report, CPB Discussion Paper SWI (2006) Wuchtiges Ja für die Bildungsverfassung. https://www.swissinfo.ch/ger/wuchtiges-ja-fuerdie-bildungsverfassung/5206776. Accessed on 01 Dec 2018 United Nations (2022) Old-age poverty has a woman’s face. https://www.un.org/development/desa/dspd/ 2022/11/old-age-poverty/. Accessed on 06 Oct 2023 Zhu A, Bradbury B (2015) Delaying school entry: shortand longer-term effects on mothers’ employment. Econ Record 91:233–246 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 123