Abstract
Since the millennium, the labour market participation of women and mothers is increasing across European countries. Several work/care policy measures underlie this evolution. At the same time, the labour market behaviour of fathers, as well as their involvement in care work, is relatively unchanging, meaning that employed mothers are facing an increased burden with respect to gainful employment and providing care. We propose a family working-time model that incentivizes fathers and mothers to both work in extended part-time employment. It provides a benefit in the form of a lump-sum transfer or income replacement for each parent if, and only if, both parents work 30 hours/week. Thus, it explicitly addresses fathers and – contrary to most conventional family policies – actively promotes the dual earner/dual carer paradigm. Combining microsimulation and labour supply estimation, we empirically analyse the potential of the family working-time model in the German context. The relatively small share of families already choosing the symmetric distribution of about 30 working hours would increase by 60 percent. By showing that a lump-sum transfer especially benefits low-income families, we contribute to the debate about redistributive implications of family policies. The basic principles of the model generalize to other European countries where families increasingly desire an equal distribution of employment and care. In order to enhance the impact of such a policy, employers’ norms and workplace culture as well as the supply of high-quality childcare must catch up with changing workforce preferences.
Introduction
Since the start of the millennium, many European countries are promoting policies aimed at ‘reconciling work and family’ (Ferragina and Seeleib-Kaiser, 2015; Lewis, 2009). In most cases, these policies seek to enable women to enhance their labour market participation or to work longer hours. It is high on political agendas, not least due to the demographic change and the shortage of skilled labour (Gagnon, 2014). Policies in this context include a variety of measures, such as increasing the availability of public or publicly subsidized childcare, facilitating access to part-time employment, and parental leave programmes (OECD, 2011). Some of these policies do indeed enhance women’s labour market activities (Boeckmann et al., 2015; Budig et al., 2012; Keck and Saraceno, 2013; Korpi et al., 2013). They contribute to the decline of the male breadwinner model which is replaced with the rise of the one-and-a-half earner model in Europe (Lewis et al., 2008; Trappe et al., 2015; Pollmann-Schult, 2016; for Germany, Ciccia and Bleijenbergh, 2014). This is strengthening women’s economic independence and career prospects.
Despite their success regarding women’s labour market attachment, these policies designed to ‘reconcile work and family’ have barely influenced paternal behaviour. The labour force participation rate and the employment rate of women increased from, respectively, around 55 and 50 percent at the beginning of the 1990s to about 72 and 66 percent in 2014 (EU-28 average, Figure 1). However, the labour force participation rate, like the employment rate, of men remained fairly constant at around 85 and 78 percent over the same time period. Although intentionally formulated in a gender-neutral way, these policies have not succeeded in activating paternal commitment to childcare as the available evidence shows. 1
Development of male and female labour force participation rate, 1990–2014 in EU-28 countries (age group 25–64).
Given that men have barely changed their working behaviour, women’s heightened labour market attachment entails potentially negative consequences for families and, in particular, women with respect to time and care squeeze (Lewis et al., 2008). Despite evidence stressing the link between more equality in employment and domestic work, women are still responsible for the largest part of non-market, mostly care, work (Bühlmann et al., 2010; Van der Lippe et al., 2011). The gender gap in unpaid work is often argued to be one of the underlying reasons why gender equality in various dimensions of the labour market, for example, earnings and career opportunities, remains a goal unachieved (Pascall and Lewis, 2004; Pettit and Hook, 2009).
Scholars have pointed out that both the move from the male breadwinner model to the dual earner model or to the one-and-a-half breadwinner model entail disadvantages (Crompton, 1999; Gornick and Meyers, 2003; Morgan, 2008; Pfau-Effinger, 1999). In the case of the dual earner model, it is the above-cited time and care squeeze, whereas in the case of the one-and-a-half earner model, it is the perpetuation of traditional gender stereotypes and their consequences such as segregation on the labour market, the gender wage gap and low old-age incomes of women, among others. The dual earner/dual carer model (Gornick and Meyers, 2003) or the universal caregiver model (Fraser, 1994) has been suggested as a model of time allocation within couples that could achieve greater gender equality in the labour market and in the distribution of care work. These scholarly concepts have also found their way to the political debate. Several European countries are debating policies directed at increasing the men’s share of care work. The most prominent example is the introduction of ‘daddy months’ or a ‘daddy quota’ in parental leave schemes. A chunk of the paid parental leave time is granted exclusively to fathers and cannot be transferred to mothers, thus generating financial incentives for men to provide more care work. These regulations are part of parental leave schemes in many European countries (Ekberg et al., 2013; Huerta et al., 2013; OECD, 2011). However, the impact of these schemes on paternal childcare involvement is generally very limited. 2
Thus, the aim of this article is to put forward the family working-time model (details below) for discussion as a new, complementary and more radical concept to promote and incentivize the dual earner/dual carer model. It can be seen as a policy template that could stimulate the transition from a one-and-a-half breadwinner model towards a dual earner/dual carer model. Originally developed for Germany, we argue that the underlying principles generalize to other contexts: at the heart of the model lies the idea that fathers and mothers who share market and care work equally are rewarded financially. In these families, both parents individually generate an income that ensures a minimum standard of living. In other words, both partners should be able to devote the necessary amount of time to gainful employment while their respective working hours should be short enough to avoid a time (care) squeeze that otherwise two full-time working parents face.
The family working-time model presupposes that fathers and mothers engage in market work of about 30 hours/week which seems to match working hours preferences (the second section of this article). This implies a shift from the one-and-a-half earner model prevalent in (West) Germany and many other European countries to the two times three-quarters earner model (Pascall and Lewis, 2004). Parents who choose this model receive both a financial benefit that incentivizes them to adjust their working hours accordingly. The benefit also provides financial support to parents with low earnings who could otherwise not afford this working-time combination. It is paid individually to mothers and fathers if, and only if, both partners choose to work about 30 hours/week.
Subsidizing this symmetric parental working-time arrangement is intended to promote gender equality on the labour market. This model only works under the prerequisite that childcare facilities of good quality are readily available for parents. Nevertheless, it values private care work highly, emphasizing that childcare cannot solely be a responsibility shared by mothers and external childcare providers (be it state or private institutions). Fathers must also do their share of care work if gender equality is actually going to be achieved.
A first contribution of this article is to introduce the family working-time model as an alternative approach into the broader debate about policies promoting equality in gainful employment and housework. Eligibility requirements and the design of financial incentives are discussed in detail. We distinguish two alternatives regarding the configuration of the subsidy: a benefit replacing (parts of) forgone earnings versus a lump-sum transfer. As a second contribution, we address the question of which behavioural reactions the introduction of the family working-time model would induce. We specify and estimate a structural labour supply model that is used to simulate the effects of this policy. According to our simulations, the overall share of families opting for such a working-time arrangement would increase from a moderate level by about 60 percent. A third contribution is the special focus of our empirical analysis on distributional issues. Total effects are decomposed by region, quartiles and number of children. Although implying similar overall effects, we show that a lump-sum transfer is more favourable for low-income households than an income-related benefit. Thereby, different versions of the family working-time model are related to the emerging literature examining gender and class inequality (Hook, 2015).
The remainder of the article is structured as follows. After a brief review of other policies promoting the dual earner/dual carer model, the following (second) section provides descriptive evidence on actual and desired working hours across Europe. The family working-time model is introduced in the third section. The micro-econometric analysis of the expected behavioural effects is detailed in the fourth section including a discussion of the results from various policy simulations. The final section concludes by putting the model and empirical findings into a more general perspective, relating it to the policy debate in Europe and highlighting limits of the analysis as well as allies for future research.
The dual earner/dual carer model in European countries: policies and time preferences
As noted in the ‘Introduction’ section, policies seeking to ‘balance labour market and family life’, even if formulated in a gender-neutral way, mostly affect mothers. Paternal behaviour remains unaffected. Paid parental leave schemes are a striking example: all European countries provide some form of paid parental leave for parents with young children (OECD, 2011). In most cases, there is a short period of maternity leave, that is, paid time off that is reserved for mothers around childbirth. After this, many countries provide paid leave entitlements that can be used by fathers or mothers. Due to prevailing social norms of mothers as primary care givers (Boeckmann et al., 2015; Steiber and Haas, 2012), as well as economic reasons (on average, mothers have lower earnings than their male partners, for example, Aisenbrey and Bruckner, 2008), the vast majority of this leave is taken by mothers (OECD, 2011). This applies even to countries that provide income-replacement benefits.
Thus, some countries (e.g. Sweden, Norway, Finland, Denmark, Austria and Germany) have introduced leave entitlements that can only be used by fathers (Dearing, 2016; Ray et al., 2010). These so-called ‘daddy months’ cannot be transferred to mothers. The overall period of father-specific leave is relatively short (usually 1 or 2 months) compared to the total amount of available parental leave. Empirical evidence shows that fathers are more inclined to take parental leave in countries that have parental leave entitlements specifically reserved for fathers (e.g. Bünning and Pollmann-Schult, 2016). Only a very small share of fathers, however, takes leave that is longer than the time explicitly reserved for fathers (Huebener et al., 2016, for the German example, or Ekberg et al., 2013, for Sweden).
In addition to the ‘daddy months’, in 2008 Sweden introduced a gender equality bonus in the form of tax credits to parents who shared parental leave equally. Evaluation studies (Duvander and Johansson, 2012) show, however, that this bonus did not lead to an increase in the duration of fathers’ leave.
In the Netherlands, since the 1990s, part-time employment has been promoted as part of the ‘combination model’ adopted by the Dutch government as a main guideline for policies in the field of market work and care (Plantenga, 2002). Policies included the abolishment of minimum hours requirements for the entitlement to the minimum wage, minimum holiday pay and social security benefits as well as a legal entitlement to part-time work. However, since no financial incentives were attached to the equal sharing of market and care work between men and women, these policies led to an increase in the labour force participation of women, mostly in part-time work, whereas weekly hours of work of men hardly changed. These policies thus facilitated the transition from the one-breadwinner model to the one-and-a-half earner model in the Netherlands, but they failed to establish the dual earner/dual carer model as the societal norm (Plantenga, 2002).
Up to now, the dual earner/dual carer model is not the common model in Europe (Ciccia and Verloo, 2012), not even in the Nordic countries. In Sweden, which introduced ‘daddy months’ within parental leave in the 1980s, women still do 1.6 times as much care work as men, but only 0.7 times as much market work. Care work is split even more unequally in European countries where family policies do not explicitly promote gender equality (Hipp and Leuze, 2015). In Germany, Italy and Spain, for example, women do 1.9, 2.3, and 1.7 times as much care work as men, respectively, while they engage in much less market work (0.6 times the hours of men in Germany and in Spain, 0.5 times in Italy; OECD, n.d.).
As of 2016, the prevalent mode of time allocation in many European countries is still the one-and-a-half breadwinner model. Multiple surveys show that this reality does not match young parents’ wishes. Using the 2005 European Social Survey (ESS), Lewis et al. (2008) report that more than 60 percent of working fathers across 13 European countries wish to reduce their working hours, while mothers mostly want ‘long part-time’ employment. This basic pattern still holds using more recent data (Steiber and Haas, 2015). Moreover, attitudes towards the distribution of paid and unpaid work vary across and even within welfare regimes (Edlund and Öun, 2016; Pollmann-Schult, 2016).
A closer look at 2010 ESS data for a sample of cohabiting couples with at least one child within the household confirms these findings in principle, but also conveys a more nuanced picture (ESS, 2010; Table A1 in the Online Appendix). First, it shows that employed women and men exhibit distinctly different hours patterns across all countries in the sample: about 49 percent of women have jobs up to 35 hours/week, while only about 45 percent are above this threshold (6 percent do not work). The share of full-time working women is below-average in Great Britain or the Netherlands but above-average in most Southern (conditional on a lower overall participation rate here), Nordic and Eastern European countries. By contrast, only about 20 percent of men work at most 35 hours and about 80 percent are in full-time employment. The male full-time share (defined as more than 35 hours/week) is higher in almost all countries, with France a notable exception at about 45 percent.
Second, the difference between actual and desired working hours exhibits common patterns for men and women in virtually all countries: the higher the share of people preferring to increase their hours, the lower their actual working hours are. Conversely, the share of people wishing to reduce their work hours increases the more hours they actually work. Significant shares of the population regard themselves as underemployed or overemployed. However, the overall share of fathers who would like to extend their working hours is larger in all hours categories than the fraction of mothers and vice versa. This gender difference exists in all countries.
Third, given the contrast in actual working hours, a larger share of mothers would like to expand their hours of work. On the contrary, a larger share of fathers would rather reduce their working hours. At the same time, full-time employed mothers and part-time employed fathers would increase and decrease their hours, respectively. Lewis et al.’s (2008) conclusion that a gender equal distribution of employment and care is not achieved but desired thus still holds in 2010. An increasingly preferred arrangement seems to be the two times three-quarter employment model. The degree this arrangement is desired varies across the countries in our sample.
The same data set also provides information on the distribution of house and care work. Individuals are asked about their own hours spent on these activities as well as their respective partners’. Answers are quite consistent across mothers and fathers: women shoulder about 2.5 times the workload in the household compared to their male partners. There are considerable differences in mothers’ total hours spent on house and care work with longer hours worked in Southern and Eastern European countries and shorter hours in Nordic countries, France and Great Britain. Another finding is that the volume of housework shouldered by employed mothers barely varies with their working hours. Full-time employed mothers face almost the same amount of house and care work as mothers working in part-time jobs with low working hours. These findings underline the problem of the time squeeze for mothers with full-time jobs. 3
The family working-time model: financial incentives for the dual earner/dual carer model
In recent years, the discrepancy between actual working hours of parents with young children and their wishes regarding working-time arrangements is increasingly discussed by the public, not least as an important dimension of the gender pay gap (Ponthieux and Meurs, 2015). Consequently, the basic principles of the ‘family working-time model’ (‘Familienarbeitszeit’) evolved from discussions with practitioners initiated by the Friedrich-Ebert-Stiftung (FES) and the Hans Böckler Stiftung (HBS) (Müller et al., 2013). This process also caught the attention of the Federal Ministry for Family Affairs, Senior Citizens, Women and Youth. Elements from the model were integrated in the 2014 reform of parental leave regulations in Germany (see Geyer and Krause, 2016). The concept will also be part of the 2017 Social Democratic Party (SDP) campaign platform.
The core idea of the family working-time model is to pay financial benefits to couples if, and only if, both partners work about 30 hours/week. Should only one partner choose this amount of working time and the other one works either more (e.g. full-time at 40 hours) or less (e.g. part-time at 20 hours or less), neither is granted the benefit. By conditioning the financial subsidy on the positive (observable) market work hours of both partners, it sets strong financial incentives for the dual earner/dual carer model. The benefit is tied to a specific working-time arrangement where both parents work equal amounts. At the same time, it equips parents with sufficient financial resources to provide care work. Unlike other cash-for-care benefits, for example, the ‘Betreuungsgeld’ in Germany (Müller and Wrohlich, 2016), the subsidy does not imply negative work incentives for secondary earners with zero or low working hours, who are usually identical with mothers.
This concept is, in principle, universally applicable to various family and care work settings. However, to keep fiscal costs reasonable, in this article, we assume that the benefit could be drawn by parents for up to three years after parental leave benefits have expired. In the German case, this period used to start when the child is either 12 or 14 months old (depending on the utilization of daddy months). After 2014 parental leave reform, the period could start as late as when the youngest child is 28 months old because parental leave benefits can be stretched out when both parents work part-time (Geyer and Krause, 2016). 4
We distinguish two versions of the family working-time model (in the following referred to as scenarios A and B) in terms of the benefit design. In scenario A, each parent receives a subsidy that replaces part (for mean incomes about 65%) of the difference in net earnings between working full-time and an extended part-time position of 30 hours. When, for example, a full-time working individual who earns a middle income decides to reduce his or her working hours from 40 hours to 30 hours/week, 65 percent of his or her net earnings forgone after the reduction of hours are compensated by the subsidy. The simpler scenario B consists of a lump-sum benefit amounting to €250 per month for each parent as long as both are eligible according to the working-time condition. In both scenarios, the transfers are not means-tested. The implied incentives for different incomes can be visualized by replacement rates drawn as a function of the net earnings (Figure 2).
Replacement rate as a function of net earnings.
In scenario A, the benefit replaces a portion of the difference in net earnings between full-time and extended part-time employment of about 30 hours/week. The replacement rate amounts to 65 percent for those with middle incomes and increases at 0.05 percentage points per euro for parents with low incomes (net earnings below €1300 per month, corresponding to the 30th percentile). The benefit is capped at €360 per month and per parent for high-income households. This implies a decrease in the replacement rate that starts at a net income of €2750 per month. 5 Note that the benefit replaces the difference between potential net earnings at 30 hours of market work and full-time employment. All persons are entitled to this benefit, regardless of their prior employment status. 6
In scenario B, each parent is paid a lump-sum transfer of €250 per month as long as both parents meet the working-time condition. This amount is roughly equal to the average amount of the benefit under scenario A. Scenarios A and B imply more or less the same short-term fiscal costs, thus allowing for direct comparisons of any behavioural reactions induced by them. Although the lump-sum benefit does not depend on monthly earnings, there is an implicit replacement rate that can be compared to scenario A (Figure 1).
A benefit that replaces a significant proportion of foregone earnings, as in scenario A, is more appealing to middle- and high-income parents, in particular fathers. Empirical evidence from the evaluation of parental leave schemes shows that fathers respond more to benefits that replace a high proportion of their earnings rather than to a lower lump-sum rate (Geyer et al., 2015). On the contrary, a lump-sum benefit, as in scenario B, is more favourable to low-income families, thus potentially reducing overall inequality. Moreover, it comes at a much lower bureaucratic cost, both for families and for the administration authorities compared to scenario A.
The family working-time model is meant to set incentives for a specific allocation of market and care work within couples. Thus, it is not straightforward how to extend the concept to single parents. Different possibilities are under discussion. 7 We define single parents here as adult parents who live in the same household with their dependent children but without other adults and who are granted the single parents’ tax allowance. These single parents are also eligible for the benefit up to three years after parental leave. The benefit is conditional on their working hours, that is, they have to work about 30 hours/week. The working time of the other parent not living in the household is not relevant. We do not include results for singles in the empirical section of this article because the focus here is on the distribution of market and household work within couples. 8 Nevertheless, the family working-time model is designed as a comprehensive approach to promote gender equality and, thus, includes single parents.
Empirical analysis
The empirical analysis of behavioural effects that the family working-time model potentially induces is based on a representative sample of families in Germany and a microsimulation model. First, using a tax-transfer simulation model, net household incomes are computed under the status quo legislation and under the two-family working-time model scenarios. Second, taking the simulated incomes in the status quo, we estimate a micro-econometric household labour supply model. The preferences estimated in this model in combination with simulated net incomes under the policy reform scenarios are then used to predict labour supply reactions of mothers and fathers to the policy reform, that is, the implementation of the family working-time model.
Data and sample
The microsimulation model is based on data from the German Socio-Economic Panel (SOEP) (see Wagner et al., 2007) and the SOEP supplement ‘Families in Germany’ (FiD) (Schröder et al., 2013). 9 We restrict the sample to families with children aged 1–3 years for the empirical analysis. The SOEP and FiD combination yields a sample of about 1900 households per year, which represent roughly 1.7 million families in Germany. Since we pool three waves of the data between 2010 and 2012, we end up with a sample size of roughly 5500 observations (Table 1). We then select couples with flexible labour supply (i.e. individuals are not (fully) disabled, self-employed or pensioners) for the estimation. Individuals who are self-employed, in school, provide vocational training or retired are excluded from the empirical analysis. Thus, the final estimation sample consists of 3355 observations. Some of the families appear in multiple years. In total, the record entails observations of 2064 unique couples.
Description of the sample.
Source: SOEP, waves 2010, 2011, 2012 (Wagner et al., 2007); FiD, waves 2010, 2011, 2012 (Schröder et al., 2013); own calculations.
SOEP: Socio-Economic Panel; FiD: Families in Germany.
Between 2010 and 2012, only 1.15 percent of all families with children aged 1–3 years chose a working-time arrangement in which both partners worked around 30 hours/week (Table 2). This share is much larger in East Germany (4.2%) than in West Germany (0.4%). Other working hours arrangements with equal market work are also relatively rare. Both parents work full-time in only 4.4 percent of all couples with children in this age group. However, the share increases to almost 10 percent when overtime is also subsumed as full-time.
Distribution of families across working hours categories in couple households.
Source: SOEP, waves 2010, 2011, 2012 (Wagner et al., 2007); FiD, waves 2010, 2011, 2012 (Schröder et al., 2013); own calculations.
SOEP: Socio-Economic Panel; FiD: Families in Germany.
Marginal employment: 1–14 hours; small part-time (mothers): 15–24 hours; small part-time (fathers): 1–24 hours; large part-time: 25–35 hours; full-time: 36–40 hours; overtime: >40 hours.
Again, there are large differences between East and West Germany: while the dual breadwinner model, with both parents working full-time, is not common in West Germany (about 6.5%, including overtime), it is more widespread in East Germany (24.2% when overtime is included). The largest share of couples with children aged 1–3 years in Germany still choose the male breadwinner model: in more than 37 percent of the families, the mother does not engage in market work, while the father works either full-time (17%) or even overtime (20%). Almost 25 percent stick with the one-and-a-half breadwinner model, where the mother works short part-time hours and the father works full-time or overtime.
Methodology
We employ a microsimulation model for the empirical analysis that accounts for behavioural responses of fathers and mothers in terms of their labour supply. The first part of the microsimulation model consists of the tax-transfer model STSM (‘Steuer-Transfer-Simulations-Modell’; Steiner et al., 2012) that calculates net incomes for all households of the micro data set based on comprehensive information on all sources of income and household characteristics. It incorporates all relevant elements of the German tax and transfer system, in particular the income tax, social security contributions and social transfers. Net household incomes are calculated for the status quo legislation as well as for the two-family working-time benefit scenarios previously introduced. Moreover, in preparation for the labour supply estimation part, net household incomes are calculated not just for the observed working hours of the parents but also for other hypothetical working hours categories (Table 2). 10
The second element of the microsimulation model consists of a micro-econometric labour supply model. We use a household labour supply model depicting the joint employment decision of fathers and mothers among discrete choices (Van Soest, 1995): a couple selects among several alternatives consisting of combinations of both partners’ working hours, corresponding leisure time and associated net income. Note that total time available is only split between market work and leisure time (see below). In our case, these choice alternatives correspond to the working-time categories described in Table 2. The model is rooted in basic utility theory (Stigler, 1950a, 1950b) that is, both partners jointly maximize household utility which is determined by household income and the available leisure time of both partners.
To estimate this model, the observed combination of net household income and working times is interpreted as the choice category that yields the highest utility for the couple. Based on this, we can estimate so-called preference parameters that describe the relation between net household income and leisure of both partners. In order to simulate labour supply under a changed policy, we assume that preferences stay constant and predict the couples’ choices given the changed net household incomes on the basis of the estimated parameters. More information on the model, its assumptions and the estimation is provided in Table A2 in the Online Appendix. Although several assumptions of the model seem very restrictive, the model has been validated in previous studies. A comparable structural model predicts behavioural responses to actual policy reforms reasonably well (Geyer et al., 2015).
This standard utility model, in particular the assumption that spouses jointly maximize household utility, has been criticized. First, more flexible approaches, such as collective or bargaining models (Bourguignon and Chiappori, 1992), have been put forward that allow each individual to maximize his or her own utility. These models, however, have been shown to be very difficult to implement empirically (Blundell and MaCurdy, 1999). Second, our model does not differentiate between leisure time and household production such as, for example, care work (Chiappori, 1997). Total time is only divided between market work and non-market work, which is interpreted as leisure. Unfortunately, modelling household production is not possible with the SOEP data set that contains reliable data on time allocation and market wages.
Results
Simulation results for scenario A show that a family working-time benefit that replaces about 65 percent of foregone net earnings for middle-income families compared to full-time work would increase the share of couples where both partners choose to work around 30 hours/week from originally 1.5 percent by 0.9 percentage points (Table 3). While this seems like a very small effect in absolute terms, it is actually a rather large effect in relative terms: it translates to an increase in the share of couples choosing this working-time arrangement by more than 60 percent. The average working time of mothers would increase by 0.5 percent, while fathers’ working time would decrease by 0.2 percent.
Simulation results for scenario A (income replacement).
Source: SOEP, waves 2010, 2011, 2012 (Wagner et al., 2007); FiD, waves 2010, 2011, 2012 (Schröder et al., 2013); own calculations.
SOEP: Socio-Economic Panel; FiD: Families in Germany.
We also simulated the effects for several subgroups (Table 3). A comparison of couples in East and West Germany shows that this policy would lead to much larger behavioural responses in East Germany than in West Germany, at least in absolute terms. While in West Germany, the share of couples choosing this particular working-time arrangement would increase by less than 0.8 percentage points (starting from roughly 1.3%), in East Germany the increase amounts to more than 1.5 percentage points (starting from 2.35%). It seems the social norms for mothers and fathers regarding working and caring still differ strongly between the two parts of the country (Blome, 2016a, 2016b; Rosenfeld et al., 2004).
Moreover, we find that under scenario A (income replacement) the simulated behavioural changes increase with family income. While the share of couples choosing the family working-time model in the first income quartile only increases by 0.7 percentage points (starting from 1.3%), the effect amounts to 1 percentage point for couples in the highest income quartile (starting from almost 2%). In this income group, the share of couples choosing this particular working-time arrangement is highest – more than 3 percent under a scenario A benefit.
The average results from scenario B, the lump-sum benefit, are very similar to those from scenario A. This is not surprising since we chose the monthly amount of the lump-sum benefit (€250 per month) to roughly equal the average amount of the income-replacement benefit in scenario A. According to our simulation, the share of couples choosing the family working-time model would increase by almost 1 percentage point or 66 percent (Table 4).
Simulation results for scenario B (lump-sum transfer).
Source: SOEP, waves 2010, 2011, 2012 (Wagner et al., 2007); FiD, waves 2010, 2011, 2012 (Schröder et al., 2013); own calculations.
SOEP: Socio-Economic Panel; FiD: Families in Germany.
The difference between East and West Germany is also quite similar to what we found for scenario A. However, we see a difference in the pattern of behavioural changes across income groups. In contrast to scenario A, which affected families with higher incomes more strongly than families with lower incomes, the simulated behavioural changes are very similar across all income groups: the share of couples choosing this particular working hours arrangement would increase by roughly 1 percentage point in all income quartiles. This can be explained by the fact that the income-replacement benefit in scenario A is more appealing to parents with higher earnings, while the lump-sum benefit is more favourable to low-income families. In quantitative terms, the difference between the two scenarios, however, is not very large: the income-replacement benefit (scenario A) would lead to an increase of 1.3 percentage points, the lump-sum benefit (scenario B) to an increase of 1 percentage points in the share of couples in the upper income quartile who chose an egalitarian working-time allocation of about 30 hours, respectively. Given that bureaucratic costs should be considerably lower in the lump-sum benefit scenario (both for applicants and the administration) and that it would be more appealing to lower income couples, the lump-sum benefit might be the better choice.
It should be noted, however, that the results obtained from the econometric model must be interpreted with caution. One basic assumption in the behavioural model is that preferences do not change in case of a policy change, that is, the introduction of the family working-time model in our case. The simulated behavioural change is obtained from the adjustment of financial incentives holding all other factors constant. However, preferences and social norms change over time (Blome, 2016a). More importantly, the introduction of such a policy reform might, per se, change social norms even further (Brooks and Manza, 2006, 2007; Kenworthy, 2009; Kenworthy and McCall, 2008; Unterhofer and Wrohlich, 2017). Moreover, existing restrictions, for example, regarding the supply of subsidized childcare, which are not modelled explicitly here, might be reduced. Thus, our results should only be interpreted as short-run effects of these potential policy reforms and likely serve as lower bound estimates. If social norms regarding the gender division of market and care work change, effects might be higher in the medium and long run.
Discussion and policy conclusions
The labour market engagement of women is steadily increasing, but the gender gap in family and care work has barely been affected by this development. Although many European countries have incorporated specific elements into existing or freshly introduced family policies that incentivize fathers to engage in childcare (e.g. ‘daddy quota’ in parental leave schemes), gender equality in the distribution of market work and care work is far from being achieved. At the same time, social value surveys show that norms and preferences are changing: in particular, young parents demand a more equal sharing of gainful employment and family care between fathers and mothers. A significant share of young fathers state that they are not able to reduce their working hours according to their preferences for financial reasons.
In response, the proposed family working-time model is part of the public debate in Germany. It presents a way to promote the dual earner/dual carer model among parents with young children. The core idea of the model is that parents are paid a financial benefit when both partners decide to work about 30 hours/week. If they both do so, each parent individually receives a financial transfer. Should only one partner opt for this working time and the other one works more (e.g. full-time) or less (e.g. short part-time hours), then neither parent receives the benefit. The family working-time model thus explicitly incentivizes the dual earner/dual carer model.
Our empirical analysis of the family working-time model is applied to families with children aged 1–3 years in Germany. The model could increase the share of parents choosing this particular working-time model by 1 percentage point. Given that the share of couples currently opting for this time allocation is low (about 1.5 percent), this is an increase of about 60 percent. Our simulations show that in the short run, the costs for this benefit would amount to about €130 million per year, which is very low, given that the parental leave benefit amounts to almost €5 billion per year.
The comparison of an income-replacing benefit with a lump-sum transfer sheds light on the distributional implications of the family working-time model. Average effects are strikingly similar in both scenarios. A lump-sum transfer, however, would be more beneficial to parents with low incomes by evoking stronger behavioural reactions among them. Thus, this variant of the model offers a great opportunity for a more gender-egalitarian division of labour for families in the lower parts of the income distribution, a group traditionally lagging behind medium- and high-income families. Observing the current political debate in Germany, this alternative of the model is thus also more likely to be implemented.
Although providing strong financial incentives, the behavioural changes induced by the model are relatively small, at least considered in absolute terms. According to our simulation results, the family working-time model would not be able to close the gap between actual and preferred working arrangements in the short run. Solely improving financial incentives through a family working-time benefit will not suffice to achieve that goal. Several additional areas have to be addressed: first, besides financial reasons, fathers often state that they are not able to decrease their working hours because of their workload or because of restrictions set by their employers. This points to the fact that cultural norms regarding gender stereotypes in the workplace, but also in terms of how employers define a ‘standard’ employment relationship or whether they insist on a ‘culture of physical presence’ at all times need to evolve as well. Second, the family working-time model also needs to be complemented by readily available, affordable, and high-quality childcare.
How do the results from our policy simulations for Germany generalize to other institutional contexts? Our descriptive analysis shows that the discrepancy between actual and desired working hours spans national boundaries and the different patterns for men and women are qualitatively similar. Differences are mostly driven by the current integration of women (and mothers) in the labour market and their actual hours of work. A regional comparison within Germany illustrates the likely effect of the family working-time model in different contexts: the gender division of market and care work is currently more equal in East than in West Germany. Thus, parents in East Germany would respond more strongly to the family working-time model than those in West Germany. This is probably due to, first, social norms that are more favourable towards gender equality and, second, a historically much better childcare infrastructure for children under the age of 3. Thus, policies that set financial incentives similar to the family working-time model can be expected to have greater effects in societies that already achieved a certain level of gender equality.
Several countries have already tried to foster the dual carer/dual earner model by other policy measures. So far, these strategies have not been successful. The Netherlands, for example, have put forward the ‘combination model’ and tried to make part-time work more attractive for men and women by expanding legal rights of part-time workers. Unlike in the family working-time model, however, no explicit financial incentives were provided for part-time working men. Thus, this strategy failed to stimulate the dual earner/dual carer model in the Netherlands (Plantenga, 2002). A policy that is more similar to the family working-time model and has been introduced as part of the parental leave scheme is the Swedish ‘gender equality bonus’. However, the total maximum amount of this bonus is rather limited compared to the family working-time model, which might explain why it has not proven to be successful in Sweden (Duvander and Johansson, 2012).
Results from our simulation exercise suggest that families in Germany would indeed react to the introduction of the family working-time model. However, it should be noted that from our simulation-based approach, we can only infer the short-run effects of such a policy change. An underlying assumption in the behavioural microsimulation model is that preferences stay constant as financial incentives are varied. Changes in social norms over time and, in particular, the change in social norms induced by a policy reform itself are not captured by this approach. It is therefore not suited to identify medium- or long-term effects. To what extent such a policy would be able to impact attitudes and norms among employees and – even more importantly – among employers remains a question for future research.
Footnotes
Acknowledgements
We thank Agnes Blome for many valuable comments and suggestions on an earlier version of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We gratefully acknowledge funding by the Friedrich-Ebert-Stiftung as well as the Hans Böckler Foundation. These foundations played a pivotal role in developing the family working-time model that is introduced and analysed in this article.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
