Theoretical Model
In order to understand the simultaneity of people’s decisions for a specific childcare arrangement and the hours they spend in the labor market, we relied on a standard labor supply model (Creedy and Kalb 2005). Labor supply decisions comprise the choice whether to take part in the labor market and, if yes, how many hours to work. The theory maintains that labor market participation and the number of hours people choose to work are the result of an evaluation of the expected monetary and non-monetary benefits and costs associated with a certain employment for one thing; and the time (and quality thereof) that remains for one’s children and family for another (Becker 1965, 1974, 1991; Connelly 1992). Likewise, the availability and quality of different care arrangements for one’s children further shapes the decision for or against an employment and the hours spent in paid employment.
Following the theory, people evaluate the utility of taking up an employment and of supplying a specific amount of labor based on an evaluation of consumed goods, the quality of childcare, and time remaining for leisure and to be spent with one’s children, respectively (Connelly 1992; Kornstad and Thoresen 2007). More specifically, excluding the possibility of a potential feedback effect of subsidies on people’s wage rates, we concentrated on the resulting monetary costs and benefits (Connelly 1992; Kornstad and Thoresen 2007; Viitanen 2005). Given that the revenues from employment must cover consumption and childcare expenditures, we expected subsidies to affect labor supply by reducing the costs of childcare. However, subsidies could have an effect in two opposing directions. On the one hand, people might reduce the number of hours they are willing to work since subsidies reduce the costs of childcare (income effect). However, this implies that consumption is fixed. Based on a rational action framework, the utility of a certain labor supply depends on the predicted net earnings after taking costs for childcare into account or, in other words, the money that can be used for consumption (Aaberge and Colombino 2014; Becker 1965; Herbst 2010). On the other hand, subsidies could encourage higher levels of labor supply by increasing the monetary payoff of working more hours and thus generating higher opportunity costs of working less hours (substitution effect). Under the assumption that time for leisure is a “normal good”, the relative strengths of income and substitution effects will determine the amount of supplied labor (Mincer 1962).
Based on these broad assumptions, people’s labor supply depends on the wages they receive from their jobs as well as on the amount of income from their paid labor and other sources, especially from other household members who are engaged in wage labor (Chiappori 1992; Haan and Wrohlich 2011; Powell 2002; Ribar 1995). Furthermore, as far as there are differences in people’s preferences for various commodities, their response to subsidies will likely also differ across social groups, such as by social status, migration background, and gender (e.g., Baker et al. 2008; Blau et al. 2006; Nollenberger and Rodríguez-Planas 2011). This expected heterogeneity is discussed further below. First, however, we outline the context of the present study and examine existing evidence on the effect of childcare subsidies in general.
Context
The present study was set in Bern, the capital city of Switzerland, and with 133,798 inhabitants (Federal Statistical Office 2019a) the fifth largest city. In Switzerland, children enter compulsory school at the age of six, after usually having attended kindergarten for 2 years (in some cantons kindergarten is not compulsory or only compulsory for 1 year, but even in these cases the vast majority of children attend kindergarten for 2 years). Day care centers always provide care for preschool-aged children, and sometimes accept kindergarten children as well. In Bern, about 85% of parents regularly resort to extrafamilial childcare and 62% of the children involved receive care in day care centers. Of those children, the vast majority spend either 2 (49%) or 3 days (28%) per week at the childcare center. Next to center-based care, care provided by relatives (e.g., grandparents) and friends is also of great importance (38% of all parents make use of such care arrangements), whereas only 3% of children in Bern experience care by a childminder or nanny (Ecoplan 2016). There is a high concentration of childcare centers in Bern and they are largely subsidized either directly with childcare slots at income-based prices or with childcare vouchers. On the whole, this makes Bern a context with readily available center-based care.
Concerning the labor market in Switzerland, the employment rate among permanent residents aged 15 years and older is 68%. Differentiated by gender, 74% of men are in paid employment while women’s employment rate amounts to 63% (Federal Statistical Office 2020b). However, on closer inspection it becomes apparent that Switzerland is still a male-breadwinner country (e.g., Wise and Zangger 2017): While 82% of the employed men work full-time, only 40% of the women hold full-time jobs (Federal Statistical Office, 2020a). Consequently, a majority of women hold part-time jobs. Although about four-fifths (79%) of mothers with a child of 15 years or younger living in the same household are employed, 81% of them work part-time. At the same time, having a child hardly leads to a reduction of working days for men, who predominantly work full time (in the Swiss context, full-time refers to working 38–42 h per week; Federal Statistical Office 2016). For most people, deciding in favor of a full-time employment means cutting down on family time on weekdays. Unsurprisingly, the most important reason for working part-time in Switzerland is childcare (Federal Statistical Office 2019b).
Review of Relevant Literature
Because subsidizing childcare results in reduced childcare costs and higher net wages of parents, we address both the childcare subsidy literature as well as studies dealing with the costs of formal childcare. Although most of the recent literature finds that decreased childcare costs and the presence of more subsidized slots have a positive effect on parental labor supply, the results are nevertheless somewhat mixed. Differences in data quality and methods are one potential explanation for this heterogeneity; differences in policy design, maternal labor force participation rates as well as gender-specific and family related social norm differences across countries are further reasons (Brewer et al. 2016; Cattan 2016). It is also important to keep in mind that there is a qualitative and not only a quantitative difference when it comes to participating in the labor market and the hours worked. Due to a substitution effect, childcare subsidies increase labor market participation rates of mothers because they raise net wages. However, for mothers already engaged in paid labor, subsidies do not necessarily lead to more hours worked because of an income effect, which might lead mothers to choose “leisure”, or time at home, over more work (e.g., Boeri and Ours 2008; Cattan 2016). Furthermore, the intended effects of subsidies can be offset by crowding-out effects, causing dead-weight costs (Cattan 2016). Crowding-out happens when a policy change just leads parents to substitute their informal childcare arrangements, typically provided by a relative, such as the grandmother, for formal paid childcare, without adjusting the number of hours they work (e.g., Asai et al. 2015; Blau et al. 2006; Brewer et al. 2016; Heckman 1974).
The Context in Which the Policy was Introduced
Research on the effects of childcare policies has yielded different findings depending on the context of the policy at stake. In their analysis using data from the US, Blau and Currie (2006) found higher maternal employment rates among subsidy recipients compared to non-recipients.Footnote 1 However, the average hours they engaged in paid labor was similar among both groups. This finding can partly be explained by a crowding-out effect (Blau and Currie 2006). Further evidence from the US confirms that maternal employment rates increase with lower childcare costs and higher childcare subsidies (Anderson and Levine 1999; Berger and Black 1992; Connelly and Kimmel 2003; Herbst 2010; Kimmel 1998; Leibowitz et al. 1992; Tekin 2004, 2007). It should be noted here, that with average net childcare costs of about 40% of net earnings, in international comparison, the US has one of the highest formal childcare costsFootnote 2 (Cascio et al. 2015). Another such high-cost context is the UK, with a market-oriented childcare provision setting, that, failing to meet the demands of parents, results in queues for childcare. Subsidized care is mostly reserved for families in need such as low-income families, those with a child with disabilities or single parents (Viitanen 2005). Under these conditions, using subsidy price simulations, Viitanen (2005) showed that the price of childcare had only small effects on labor market participation and the use of paid childcare. Even with a 50% subsidy, informal care was favored (Viitanen 2005), presumably because formal childcare remained unaffordable or unattractive for low-income families.
For Québec, Canada, a context with a high demand for formal childcare in which newly introduced subsidies lowered the costs of formal childcare substantially, evidence points to a positive effect on parental labor supply: treating the introduction of a new childcare policy as a natural experiment, several studies found positive short- as well as long-term effects of reduced childcare fees on maternal labor supply (Baker et al. 2008; Lefebvre and Merrigan 2008; Lefebvre et al. 2009). A similar approach followed by Viitanen (2011), who exploited the introduction of vouchers for private childcare in 33 municipalities in Finland, somewhat contrasts with the findings from the Canadian context. While the vouchers increased the use of private childcare (especially in areas with an excess demand for childcare), they had no effects on the use of public care and no effects on labor force participation. However, since Finland was already providing high-quality and low-cost public childcare prior to the change in policy, this might explain the lack of effects on labor market participation (Viitanen 2011). This is in line with a finding from Sweden, as another provider of readily available and affordable center-based childcare, where, using simulations, Lundin, Mörk, and Öckert (2008) found that setting a cap on the price of childcare would have no further effects on the already high female labor market participation. For Norway, with a similarly well-developed affordable childcare system, Kornstad and Thoresen (2006) estimated that for a scenario that does not quite cover the demand for formal child care, maternal labor supply would increase by about 2 h per week. They estimated a larger increase of about 3.5 h per week for an alternate scenario with no queues for center-based childcare (Kornstad and Thoresen 2006). To the contrary, also for Norway, Havnes and Mogstad (2011) found no effects of expanding the subsidized childcare system further on maternal labor supply. Using difference-in-differences estimates, they found that the new policy, an expansion of subsidized formal childcare, only crowded out existing forms of informal care arrangements. On the other hand, Hardoy and Schøne (2015), also Norway, found a positive effect of reduced childcare costs. However, because the reform in question was accompanied by an increased supply of day care slots, the finding might have been induced by the latter circumstance or a combination of both factors (Hardoy and Schøne 2015).
On the other hand, countries with a less well-developed formal childcare system (smaller capacity and less government funding), exhibit a high potential for raising maternal labor supply with reforms. The evidence for the Netherlands (Graafland 2000), the UK (Viitanen 2005), Austria (Mahringer and Zulehner 2015), Spain (Nollenberger and Rodríguez-Planas 2011), and Russia (Lokshin 2004) suggests, that higher subsidies and lower childcare costs are important determinants of the maternal employment rates as well as hours worked. For Southern Italy, Del Boca and Vuri (2007) also found that both childcare availability and price of childcare were important determinants of maternal labor market participation. Price was especially important for mothers in areas with a higher availability of childcare slots (Del Boca and Vuri 2007). For Australia, Guest and Parr (2013) also found a significant positive effect of family benefits and childcare subsidies on the hours couples work. For Germany, subsidized childcare was estimated to have only a small effect on maternal employment decisions compared to other countries such as the US, Canada, or the UK (Haan and Wrohlich 2011; Wrohlich 2004). However, the authors claimed that the reforms would especially affect the subgroup of highly educated women as well as those having their first child, raising their labor supply as well as fertility rates.
Heterogeneous Responses to Subsidies
Apart from the broader policy context, various groups of people might also differ in their response to childcare subsidies, be it due to individual differences in resources or beliefs about childrearing. Because men still have a higher employment rate than women in Switzerland and men still mainly work full-time (e.g., Wise and Zangger 2017), there is a difference in the potential that policies can be expected to unleash. While men hardly have leeway to work more hours, women have more room to react. Due to this ceiling effect for men, it can be assumed that women react stronger to subsidies than their male counterparts. Nevertheless, the hypothetical nature of the factorial survey experiment also allowed us to investigate men’s preferences when childcare is subsidized. Furthermore, it has repeatedly been shown for mothers that those with a higher educational degree have a higher probability of being employed and of working more hours than those with lower educational degrees (Andrén 2003; Chiuri 2000; e.g. Connelly 1992; Del Boca and Vuri 2007; Guest and Parr 2013). This raises the question of whether parents respond differently to subsidies depending on their socio-economic status.
Existing research on the variation in parental responsiveness to different subsidy levels and childcare costs by social status (e.g., education, household income) has yielded mixed results. Regarding the educational background of parents, evidence is available for the US, Canada, Germany, and Spain. While in the US and in Spain those who were the least skilled were more affected by childcare costs (Anderson and Levine 1999; Haan and Wrohlich 2011; Nollenberger and Rodríguez-Planas 2011), in Canada, the effects did not vary by educational background (Lefebvre and Merrigan 2008), and in Germany, the responsiveness was highest for highly educated women (Haan and Wrohlich 2011). For the US, Herbst (2010) also found heterogeneous effects of childcare subsidies and effects of Earned Income Tax Credit (EITC), but contrary to the German case, they primarily raised the labor supply for subgroups of single mothers, with low-income mothers responding most to EITC. Although Michalopoulos (2010) and Michalopoulos, Lundquist, and Castells (2010) found no overall effect of childcare subsidies on parental labor force participation, they also highlighted that the high-income families that predominantly constituted their sample had little room to increase their labor supply since they were already steadily and in high employment prior to the program.
Finally, evidence suggests that the migration background or ethnic descendance of parents also seems to affect labor market and childcare decisions. In the American context, Michalopoulos and Robins (2000), for example, showed that Black American mothers were more likely, while Hispanic/Latin Americans were less likely to work full time than White Americans. For the UK, results indicated a higher employment probability for White mothers compared to non-Whites (Viitanen 2005). However, to our knowledge, previous studies have not investigated whether or not the reactions to childcare costs and subsidies vary with ethnic or migration background. Nevertheless, given the position of migrants in the host country, potentially different values regarding childrearing, and differences in terms of resources and networks (e.g., absence of relatives who could provide informal care), they are nevertheless expected to vary in their response to subsidies.
What We Expect
Generally speaking, the evidence strongly suggests that the context in which subsidies are introduced greatly determines the effects the policy will have. In countries with a well-developed formal childcare system and high maternal labor supply, subsidies will only provoke small increases in maternal labor market participation and hours worked. Subsidies seem to have the highest impact in situations in which there is a higher availability of slots since costs are a factor parents can only seriously consider if they have a choice. In the present context of the City of Bern, Switzerland, there was a policy-related increase in the availability of day care centers shortly before this study was conducted and availability was no longer a major issue.
Because childcare slots were widely available in the present context, we expected positive effects of childcare subsidies on (maternal) labor market participation and labor supply (Hypothesis 1). However, subsidies have been shown to be particularly effective in raising the labor supply of parents in contexts with comparatively low employment rates and lower levels of formal childcare use, and especially when there is minimal crowd-out (Brewer et al. 2016). Since men and women’s employment rates differ considerably in the Swiss context, with men predominantly working full time (Wise and Zangger 2017), we further expected women to adjust their working hours more strongly in response to the amount of childcare subsidies than men (Hypothesis 2). With regard to heterogeneous responses due to varying resources, the expectation was less clear. On the one hand, mothers, for example, with more available non-labor income have a higher reservation wage. Consequently, we would expect them to exhibit higher levels of non-employment, even in the presence of higher subsidies (Cleveland et al. 1996; Viitanen 2005). On the other hand, women with the highest educational backgrounds can also expect the highest returns to labor, meaning they have higher opportunity costs of non-employment. However, this is why these women have higher levels of labor market participation and attachment in the first place and are not as dependent on childcare subsidies (e.g., Cattan 2016; Nollenberger and Rodríguez-Planas 2011). For this reason we expected parents with a lower social status and with less financial means from other sources to be more sensitive to subsidies, because for them working more would increases their household income significantly once childcare was subsidized (Hypothesis 3).
Finally, due to the near inevitable loss of social capital and the geographical distance to potential caregivers (e.g., grandparents), an overall higher reliance on formal childcare was expected for parents with a migration background. At the same time, many immigrants might also perceive a stronger economic necessity to work more due to a lower endowment with resources in the host country. It can thus be expected that those who were born abroad were somewhat more likely to start employment and to increase their labor supply with increasing subsidies (Hypothesis 4). Summing up, we expected the probability of parents’ choosing to go to work and of using center-based care to rise with subsidy levels. In the context of Bern, with relatively widely available good-quality care, we expected only moderate effects. In regard to the number of working hours, we expected the effects to differ with socio-economic status and financial resources, migration background, and gender. We expected people with lower social status and available income from other sources, people with a migration background and women to react stronger to subsidies.