Keywords

The majority of fathers in Germany wish to spend more time with their families (Väter gGmbH 2012; Allmendinger and Haarbrücker 2013; Gründler et al. 2013) and want to work fewer hours (Institut für Demoskopie Allensbach 2013; Statistisches Bundesamt 2015; Bernhardt et al. 2016). Yet despite these preferences, many continue to work full time, a phenomenon that is not well researched, especially in the German context. The Act on Part-Time and Fixed-Term Employment of 2001 created a framework that grants all employees the right to work part time. But unlike mothers, fathers have thus far rarely made use of this right. In 2012, just 5.5% of fathers worked part-time, while 69.1% of mothers did so (Keller and Haustein 2013).

Qualitative studies show that cultural and structural conditions in the workplace can have a decisive influence on whether fathers put their working-time preferences into practice or hesitate to do so because of possible negative implications for their careers (Epstein et al. 1999; Gesterkamp 2007; Hobson et al. 2011; Richter 2011; Sallee 2012; Galea et al. 2014). A series of predominantly experimental studies on the so-called “flexibility stigma” (Williams et al. 2013) indicates that these concerns are justified: like mothers, fathers who make their care obligations visible, for example, by reducing their working hours, are penalised at the workplace (Wayne and Cordeiro 2003; Leslie et al. 2012; Vinkenburg et al. 2012; Rudman and Mescher 2013; Vandello et al. 2013). This research discusses deviations from the ideal worker norm, which expects workers to prioritise work demands over all outside obligations, as the driving force of the flexibility stigma and as a potential cultural barrier that limits the effectiveness of flexible workplace arrangements.

However, robust empirical evidence on the relationship between workplace conditions and fathers’ working hours is still lacking. Workplace case studies and qualitative studies on parents indicate that the ideal worker norm is by no means equally prevalent in all companies; instead, it is related to other workplace culture factors that can encourage or prevent fathers from making use of their right to reduce their working hours (Botsch et al. 2007; Holter 2007; van Echtelt et al. 2008; Den Dulk et al. 2011; Hobson et al. 2011; Brinton and Mun 2016; Bernhardt et al. 2016; Väter gGmbH 2012). In addition, studies on working time policies offered by companies (Fagan and Walthery 2011) and on the use of parental leave (Bygren and Duvander 2006; Lappegård 2012) suggest that fathers’ working hours are also linked to workplace structures such as establishment size or sector.

This article seeks to identify what workplace conditions enable fathers in Germany to work shorter hours. Based on a representative survey of parenting couples from 2015, we take a closer look at four facets of workplace culture: first, the extent to which there are formal and transparent workplace policies that fathers feel entitled to use; second, the extent to which fathers perceive an ideal worker norm; third, the extent to which they feel supported by their supervisors when it comes to balancing family and career; and fourth, how easily they can find a substitute among their colleagues at short notice. With regard to workplace structures, we take into account the establishment size, the sector, whether it is part of the public sector and whether there is a works council or a staff representative council.

We examine the working hours of fathers using three different indicators: fathers’ self-reported employment status (distinguishing part-time and full-time employment), their contractual working hours, and their actual working hours. We decided to use all three indicators because weekly working-time policies vary according to organisation and industry, so it is not possible to clearly separate part-time and full-time employees using one uniform value. Furthermore, if part-time employment is used as the standard for shorter working hours, this may be setting the bar too high (Cyprian 2005). If a particular organisation has a strong full-time norm, it may be difficult for fathers to request part-time employment. But more minimal adjustments to working hours can already help fathers to organise their family life, for example, if they can slightly reduce their contractual working hours or ensure that they can complete their workload within their regular working hours, thus avoiding overtime or at least compensating for it through leisure time (Dommermuth and Kitterød 2009; Björk 2013).

1 Workplace Conditions and Fathers’ Working Hours

Fathers’ working hours are not solely determined by the negotiation processes with their partners. The workplace is another central location where work-family balance is negotiated. Workplaces are where decisions are made on whether employees face professional consequences for pursuing their working-time preferences and what these consequences are. Particularly for fathers, the decision to avail of their legal right to part-time work or workplace work-life balance policies seems to be dependent on organisational cultural factors.

Based on Amartya Sen’s capability approach, Hobson and colleagues argue (Hobson and Fahlen 2009; Hobson 2011; Hobson et al. 2011) that asserting and making use of legal rights, for example, to work part time, requires a sense of entitlement that is often lacking in fathers. According to the authors, this is largely due to the persistence of traditional gender norms in organisations, which limit the possibilities for fathers to work shorter hours and become more involved in care work. Gender norms manifest themselves in various dimensions of organisational culture.

In the growing number of predominantly qualitative and experimental studies, four workplace culture factors have emerged that could be particularly relevant for fathers’ working hours: the commitment of organisations to formal, universal and transparent policies for balancing family and working life, informal expectations regarding worker availability (the ideal worker norm), support from supervisors and the availability of substitutes (Lewis 1997; Williams 2000; Botsch et al. 2007; Pfahl and Reuyß 2009; Lewis and Den Dulk 2010; Hobson et al. 2011; Gärtner 2012; Sallee 2012; Vinkenburg et al. 2012; Rudman and Mescher 2013; Vandello et al. 2013; von Alemann and Oechsle 2015; Bernhardt et al. 2016). The decisive factor here is fathers’ subjective perceptions: their knowledge of workplace policies, their sense of entitlement to use them and their feeling of being supported by the working environment without having a guilty conscience or worrying about negative professional consequences (for an overview, see Booth and Matthews 2012).

In addition, fathers’ working hours could also differ according to the structural characteristics of organisations. For instance, it could be easier or more difficult to put employees’ desired working times into practice depending on establishment size or industry sector, as suggested by some studies on the contextual factors in workplaces that influence working times and parental leave (Döge et al. 2005; Bygren and Duvander 2006; Fagan and Walthery 2011; Escot et al. 2012). Another possibility is that the cultural workplace conditions with regard to fathers’ working hours are also related to the structural characteristics of organisations (Fagan and Walthery 2011).

For the first time, we now want to examine these factors and interrelationships for Germany using quantitative data. The following sections outline the arguments and hypotheses for the subsequent empirical analyses.

1.1 Theory and Hypotheses on the Relationship Between Workplace Culture and Fathers’ Working Hours

Family-friendly workplace policies can be an important resource for fathers to demand and legitimise shorter working hours. However, as previous studies suggest, the specific design of these measures is important. An important prerequisite is the establishment of binding and transparent rules that fathers can refer to and that cannot be withdrawn at will (Russel and Hwang 2004; Botsch et al. 2007). If workplace policies offer too much scope for individual negotiation, fathers, according to a Norwegian study, make use of them less often than they would if such policies are clearly regulated (Brandth and Kvande 2001). The authors ascribe this to the fact that a clear, formalised regulation provides normative orientation on what kind of behaviour is appropriate; they thus reduce the risk of negative sanctions.

Furthermore, fathers also need to feel that they are a target group for existing workplace policies (Hobson and Fahlen 2009). If such policies are aimed at all employees and not only at mothers with small children, fathers have less of a feeling that they are putting their masculinity at risk when they make use of them (Lewis 1997; Russel and Hwang 2004; Klenner et al. 2010; Escot et al. 2012). Directly addressing fathers via various information channels can increase their awareness of policies and convey a sense of entitlement (Väter gGmbH 2012).

If fathers have the sense that there are clear rules for reducing working hours in their organisation that apply equally to all employees, this should increase their likelihood to ask for and actually work their desired working times. Our first hypothesis is therefore:

H1

Fathers have shorter working hours the more they perceive clear, universally valid and transparent policies on work-family balance in their workplaces.

However, the effectiveness of formal policies in the workplace may be restricted by informal expectations regarding worker availability, which make it more difficult to take advantage of workplace policies and even legal rights (Hobson 2011). As Acker (1990) has shown, organisational structures are not gender neutral, but are instead shaped by expectations, policies and practices that are typically oriented towards a male breadwinner. For decades, an ideal worker norm (Williams 2000) has become established in organisations. An ideal employee works full-time and is always available to the organisation without any time restrictions for family or private reasons. This extensive availability goes hand in hand with the assumption that fathers have a partner at home who takes care of the children and the household. This norm is incompatible with a gender-egalitarian division of labour and makes it difficult for fathers to become more committed to their families (Sallee 2012).

Empirical studies assume three mechanisms by which the ideal worker norm influences the behaviour of fathers. First, fathers are under pressure to comply with the ideal worker norm because performance is often equated with availability. As data on working hours in management positions suggest, long working hours are a prerequisite for career advancement (Klenner et al. 2010). Fathers who reduce their working hours must fear that they will damage their careers because they are seen as less efficient (Rudman and Mescher 2013; Vandello et al. 2013). Second, Williams et al. (2013) argue that the ideal worker norm also has a work-ethic dimension. Those who deviate from the usual availability standards risk having their professional commitment, motivation and willingness to work questioned (Vinkenburg et al. 2012). Third, fathers who deviate from the ideal worker norm risk being devalued because their behaviour is perceived as unmanly (Rudman and Mescher 2013; Vandello et al. 2013).

If fathers have the impression that the ideal worker norm is strongly established in their organisation, they may therefore be afraid of reducing their working hours so that they are not punished as “time deviants” (Epstein et al. 1999) and will instead work even longer than contractually agreed (Bernhardt et al. 2016). Our second hypothesis is accordingly:

H2

Fathers have longer working hours the more they perceive availability expectations in line with the ideal worker norm in their organisation.

Independent of formal policies and the general working-time culture in the organisation, specific workplace actors can also support or hinder fathers in putting their working time preferences into practice (for an overview, see e.g. von Alemann and Oechsle 2015). Many studies indicate that social support plays a central role in reconciling family and career (Abendroth and Den Dulk 2011). Supervisors have a key role to play here. As a study by Gärtner (2012) shows, supervisors are usually the first point of contact when fathers want to reduce their working hours; they act as gatekeepers. Through their influence on fathers’ professional advancement, they can, on the one hand, counteract existing workplace policies and make it difficult for fathers to reduce their working hours or to stick to official working hours. On the other hand, if they are sympathetic to fathers’ concerns, they can also find informal solutions, even if there is not any workplace policy to justify it (Döge et al. 2005; Botsch et al. 2007; Bernhardt et al. 2016).

Another factor that plays a role here is how fathers assess their supervisor’s family orientation and gender conception and the kind of behaviour the supervisor exemplifies in this regard. If a father’s supervisor works late into the evening and is in the office for an almost unlimited amount of time, he will expect less sympathy for his situation than if his supervisor leaves the office early to pick up his/her children from daycare. If fathers expect a negative reaction, they even hesitate to bring their working time preferences up in the first place (Döge et al. 2005; Pfahl and Reuyß 2009; Gärtner 2012; von Alemann and Oechsle 2015). We therefore expect fathers to more often express and put into practice their desire to work fewer hours if they feel that their supervisor supports them. Our third hypothesis is therefore:

H3

Fathers have shorter working hours the more they feel supported by their supervisor.

We also expect fathers’ working hours to depend on the support they receive from their colleagues. If they are prepared to stand in for and substitute for each other, it could be easier for fathers to come to work later if necessary, to go home early or to leave the workplace in between and thus achieve shorter (actual) working hours. A contractually agreed reduction of working hours also depends on how it is being practised in the workplace. Without additional staff, any reduction in working hours must be compensated for by additional work—by employees doing the same amount of work in a shorter time or by colleagues doing some of the work. Under such circumstances, fathers may decide not to reduce their working hours because they do not want to burden their colleagues or they expect team conflicts (Döge et al. 2005; Holter 2007; Pfahl and Reuyß 2009; Lewis and Den Dulk 2010; Gärtner 2012; von Alemann and Oechsle 2015). Shorter working hours should therefore be more likely to be observed among fathers who can easily find colleagues to substitute at short notice. Our fourth hypothesis is therefore:

H4

Fathers who can easily find a substitute if necessary have shorter working hours than fathers who find it difficult or impossible to find a substitute at short notice.

1.2 Theory and Hypotheses on the Relationship Between Workplace Structure and Fathers’ Working Hours

It is plausible that differences in fathers’ working hours are not just related to the cultural characteristics of their organisations; they may also be linked to the structural conditions of their workplace. Although little is known in quantitative terms about how this plays out in Germany, qualitative and international studies suggest that certain structural conditions in organisations could make it easier or more difficult for fathers to work shorter hours: the size of the organisation, the sector, whether the organisation in question is part of the public sector and whether it has a works council.

It has been argued in previous studies that larger organisations have more scope for manpower planning, as they can reallocate work more easily and more easily find staff to substitute because of the higher number of employees. They are also more likely to have developed routines to deal with such concerns (Fagan and Walthery 2011). This suggests that workers in larger organisations are more likely to be able to put their working-time preferences into practice than in smaller ones (Bygren and Duvander 2006; Lappegård 2012). This assumption is further substantiated by the fact that small organisations are exempted from the Act on Part-Time Work and Fixed-Term Contracts; a legal entitlement to part-time work exists only in organisations with at least 15 employees. However, small organisations might be able to react more flexibly due to a lower degree of formalisation (Buchmann et al. 2010; Hipp and Stuth 2013). A survey study shows, for example, that employees of small organisations rate their workplaces as more family friendly than employees of medium-sized and larger organisations (Klenner and Schmidt 2007). In line with these two alternative hypotheses from the literature, we expect the following relationship between organisation size and fathers’ working hours:

H5

Fathers in large and small organisations have shorter working hours than fathers in medium-sized organisations.

The possibilities for shorter working hours may also differ between different sectors. Previous studies (Fagan and Walthery 2011; Hobson et al. 2011; Escot et al. 2012) indicate that fathers in sectors with a high proportion of women may find it easier to balance family and working life than fathers in male-dominated sectors. Since being able to successfully balance family and working life has long been regarded primarily as a women’s issue, organisations in these sectors are more likely to have established family-friendly policies and to have experience in dealing with part-time preferences than organisations in sectors with few women. This makes it easier for fathers in these industries to put their desire to work reduced hours into practice. Areas with a high proportion of women include administration, healthcare and social services (Escot et al. 2012). Therefore, we expect the following hypothesis:

H6

In the administrative, healthcare and social service sectors, fathers have shorter working hours than in other sectors.

In addition, previous studies on the uptake of family friendly policies suggest that there are differences between the public and private sectors. Since the public sector does not aim to maximise profits, economic considerations may be a less important argument against reducing working hours. Another argument is that the public sector offers more secure jobs, meaning that employees who reduce their working hours have fewer career consequences to fear (Bygren and Duvander 2006; Escot et al. 2012; Lappegård 2012). This gives rise to the following hypothesis:

H7

Fathers in the public sector have shorter working hours than fathers in the private sector.

Another relevant criterion could be whether the organisation has a works council or staff representative council. In accordance with the Works Constitution Act, works councils are tasked with promoting a balance in work and family life. Works councils can take action in different ways, for example, by supporting the introduction and implementation of working time arrangements. For example, Fagan and Walthery (2011) find, based on European data, that organisations with works councils are more likely to facilitate switching between full-time and part-time work options. According to the German Employee Representation Act, applications for shorter working hours in public administration can only be rejected with the approval of the staff representative council (Scheiwe 2007). This provides a legal basis for employee representative bodies to support fathers who want to reduce their working hours. Employee representative bodies are also expected to ensure that existing policies are addressed to all employees and not only to certain groups—such as mothers (Brinkmann and Fehre 2009). We therefore expect:

H8

Fathers in organisations with works councils or staff representative councils have shorter working hours than fathers in organisations without employee representative bodies.

In summary, we expect that fathers should have shorter working hours if they work in organisations with a high degree of regulation regarding fatherhood-related work-family policies, a paternity-sensitive working time culture and a high degree of support from their supervisors and colleagues. Shorter working hours should also be more feasible in large and small organisations, in administrative, healthcare and social service organisations, in the public sector and in organisations with works councils or staff representative councils. In addition, the remarks above on the relevant structural features point to possible connections with cultural conditions. For example, the degree of family-friendly workplace regulation, including the fathers’ feelings of entitlement to avail of such policies, could be more pronounced in large organisations, in the public sector and in organisations with works councils.

In addition, we expect to see different relationships when using different indicators to measure working hours. As explained at the beginning, the hurdle for fathers to work part time seems to be so high that we also want to take into account minor adjustments to working hours within the scope of full-time employment. For this reason, we investigate the relationships between part-time work, contractual working hours and actual working hours. In the following sections, we address these questions empirically.

2 Data and Methods

2.1 Sample

This study is based on an individual telephone survey of 878 parenting couples in 2015. The data were collected by the Berlin Social Science Center (WZB) in 2015 as an add-on to the German AID:A—survey (“Growing up in Germany”) which is collected on a regular basis by the German Youth Institute.Footnote 1 The aim of the survey was to close an important data gap in current research on the influence of workplace conditions on working hours and parental leave for mothers and fathers. To this end, cohabiting couples were asked about the structural and cultural characteristics of their workplaces.

The target population consisted of different-sex couples who, at the time of the survey, (a) lived with at least one child under 13 years of age in a shared household and (b) had at least one parent in dependent employment. In order to be able to study issues such as part-time work of fathers, rare employment constellations (both parents part time, both full time, women full time/man part time or not employed) were oversampled compared to their actual occurrence in Germany. Within these individual constellations, a random selection was performed. By using weighting factors based on the 2013 Microcensus, we can obtain a sample that is representative with regard to the following characteristics: the couples’ employment constellation, marital status, number and age of children in the household, state of residence, gender, age and level of education.

The data set consists of 763 fathers in dependent employment. Of these, 36 cases (4.78%) were excluded from the analyses because the fathers stated that they did not have contractually agreed working hours. Another five cases were excluded because they lacked information on actual working hours. In addition, we excluded eleven fathers who stated they did not have a supervisor. This leaves 711 cases for the analyses. Due to stratification, the unweighted sample contains a comparatively high proportion of fathers in part-time employment (14%); the average contractual working time of 37.5 h per week and the actual working time of 42.6 h are, however, close to the German average (Institut für Arbeitsmarkt- und Berufsforschung 2014).Footnote 2

In 108 (15.2%) of the 711 cases fathers had missing values on at least one explanatory variable. In particular, the items relating to workplace culture had a higher number of missing values; this occurred in up to 53 cases (3–7%; see Table 1). These were imputed in Stata using multiple imputation by chained equations (MICE) (Azur et al. 2011). As recommended by White et al. (2011), 15 imputed data sets were generated, since 15% of the cases have missing values. Missing values were imputed for the individual items on the structural characteristics of the workplace and the scales were created after the imputation to make optimal use of all available information (Gottschall et al. 2012). In addition to the model variables, we used additional workplace cultural characteristics, information on conflicts between work and family life, net income and the Big Five personality traits for the imputation in order to estimate the missing values as precisely as possible.Footnote 3

Table 1 Overview of variables, descriptive statistics and number of imputed values
Full size table

2.2 Variables

Table 1 provides an overview of the variables used, descriptive statistics and sample size. As dependent variables, we use (1) fathers’ self-assessments as to whether they work part time or full time, (2) their contractual weekly working times in hours, and (3) their information on the hours they actually work on average per week, including overtime.Footnote 4

The four workplace cultural conditions—the degree of family-friendly workplace regulation, the ideal worker norm, support from supervisors and arrangements regarding substitution—were measured as follows (see also Table 1): For the degree of family-friendly workplace regulation, we generated a scale based on the company case studies by Botsch et al. (2007) consisting of five items on management’s commitment, formalisation, scope and communication of workplace policies (Cronbach’s Alpha: 0.79). The perception of an ideal worker norm is also represented by a scale, consisting of four items adapted from Booth and Matthews (2012) (Cronbach’s Alpha: 0.83). We operationalise supervisors’ support following various studies (Hammer et al. 2009; Harrington et al. 2011; Pfahl et al. 2014) on a scale of four items (Cronbach’s Alpha: 0.80). To capture existing arrangements regarding access to substitutes at short notice, we use a dummy variable to determine whether the interviewed fathers can easily find a substitute if they cannot get to work at short notice or have to leave early.

As workplace structural characteristics, we include the sector (administration/health/social services, retail, scientific services, production, other), establishment size (small: maximum 10 employees, medium: 11–249 employees, large: 250 and more employees), whether the organisation is in the public sector (dummy variable) and whether it has a works council (dummy variable). In the multivariate analyses, we also control for various work-related and socio-demographic characteristics of the fathers as well as for their partners’ characteristics (respondent-provided information from the mothers’ interviews) and information on the household context (Hobson 2011).

2.3 Methods

In the multivariate analyses we used binary logistic regressions for the part-time work calculations and OLS regressions for the calculations on contractual and actual working hours. In the logistic regression models, we calculated average marginal effects.Footnote 5 All models used Huber-White standard errors to account for possible violations of model assumptions by using more conservative confidence intervals. A number of alternative specifications (see Sect. 3.3) indicate, in fact, a high robustness of the results.Footnote 6

3 Results

3.1 Multivariate Analyses on the Relationship Between Workplace Culture and Fathers’ Working Hours

The results of the multivariate analyses are shown in Table 2. The first model shows the average marginal effects (see above) of the model variables on fathers’ probability of working part time; the second and third models show the non-standardised coefficients from the OLS regressions for contractual or actual working times of fathers.

Table 2 Results of the multivariate analyses of relationships between workplace cultural and structural characteristics and the working hours of fathers
Full size table

Across all three models, it is evident that fathers work less if they perceive clear workplace policies on work-family balance that apply equally to all employees and are communicated regularly. With each unit increase in the 10-point scale of family-friendly workplace regulation, fathers’ probability of working part time increases by two percentage points; their contractual working hours are on average 20 min shorter (−0.34 × 60) and their actual working hours are 30 min shorter. The magnitude of those coefficients is considerable. Fathers with the maximum value (10) on the scale measuring family-friendly workplace regulation have a 20 percentage points higher probability (0.02 × 10 × 100) of working part time than fathers with the lowest value on the scale (0); their contractual working hours are on average 3 h and 24 min shorter (−0.34 × 10) and their actual working hours are even 5 h shorter (−0.50 × 10). The results thus support Hypothesis 1.

For the other indicators of workplace culture, there are differences between the various indicators for the working hours of fathers. Fathers who perceive a strong ideal worker norm have longer actual working hours than fathers who do not perceive a pronounced ideal worker norm in their workplace. Fathers who assign the highest value on the scale work about three hours more per week than fathers who assign the lowest value. Yet there is no evidence in the models that a strong ideal worker norm goes hand in hand with longer contractual working hours or a lower likelihood of working part time. The results thus only partially support Hypothesis 2.

Contrary to our expectations, there is also no evidence that fathers work less if they are supported by their supervisor. In fact, compared to fathers who do not feel supported by their supervisor (scale value 0), fathers who feel strongly supported (scale value 10) even have an average 20 percentage points lower probability of working part time and have contractual working times that are 4:12 h longer. Yet the relationship with fathers’ actual working hours is small and not statistically significant. Hypothesis 3 can therefore not be confirmed.

A good availability of substitutes at short notice, on the other hand, goes hand in hand with shorter actual working hours for fathers, but, like the ideal worker norm, it is not related to contractual working hours or the probability of working part time. Fathers who state that they can easily find a substitute work on average an hour and a half less per week than fathers who have difficulty finding a substitute. Thus, Hypothesis 4 is also only partially supported by the data.

All in all, it is evident that the four workplace culture factors are relevant in different ways for whether fathers work part time, how many working hours they have contractually agreed upon and how many hours they actually work. While formal reductions in working hours for fathers (part-time work, lower contractual working hours) are more likely in organisations with a high degree of formalisation of family-oriented policies, informal customs relating to working time culture (the ideal worker norm) only play a role for informal working time adjustments (actual working hours). Thus, the results suggest that there is a need for more transparent formal arrangements to allow fathers to adapt their working status or contractual working hours, while actual working hours are also linked to perceived informal resources or barriers.

3.2 Multivariate Analyses of the Relationship Between Workplace Structural Characteristics and the Working Hours of Fathers

Contrary to our expectations, we do not find any empirical evidence for the partial hypothesis that fathers in large organisations have shorter working hours because such organisations are expected to have greater scope for manpower planning (Hypothesis 5). Fathers in small organisations, on the other hand, have shorter contractual working hours (approx. 2 h) and have a 15 percentage points higher probability of working part time compared to those in medium-sized organisations. The results thus support Hypothesis 5 in that it could be easier for fathers to reduce their formal working hours in small organisations with more flexible structures and flatter hierarchies. However, there is no correlation between the size of the organisation and the actual working hours of fathers. Especially in small organisations, it is therefore questionable to what extent part-time work is actually voluntary and helps fathers to increase their time with their family.

In line with our expectations, we do indeed find connections between the sector and the working hours of fathers. Fathers working in administration, healthcare and social services tend to have shorter contractual and actual working hours and are more likely to work part time than fathers in all other sectors. This may be attributed to the fact that organisations in these sectors have a high proportion of women and have therefore developed more family-friendly working time arrangements than in other sectors (Escot et al. 2012). The differences in contractual working hours between the other sectors are rather small. However, there are noticeable differences in the actual working hours, with particularly long working hours in the retail sector (approx. 3:30 h longer than in administration/healthcare/social services) and in the area of scientific services (approx. 3:20 h longer). The results thus support Hypothesis 6.

There is, however, no link between employment in the public sector or the presence of a works council or staff representative council and fathers’ working hours. Hypotheses 7 and 8 are therefore not supported by the data. This finding tends to point more in the direction of studies that conclude that works councils and staff representative councils are not yet making sufficient use of their scope to facilitate work-family balance (Döge et al. 2005; Brinkmann and Fehre 2009).

3.3 Robustness Tests

In a series of alternative specifications, we tested the relationships between workplace cultural and structural factors and the working hours of fathers for their robustness.

Table 3 shows the results of a reduced model that does not contain information on workplace structural characteristics. Unlike the fully specified model, this model allows us to investigate the extent to which the relationships between an organisation’s cultural characteristics and fathers’ working hours are robust in relation to the influence of the workplace’s structural factors.

Table 3 Relationships between workplace cultural characteristics and the working hours of fathers (without controlling for workplace structural characteristics)
Full size table

Overall, the results of these reduced models reveal very similar relationships to those from the complete model described above. However, there are two additional correlations between an organisation’s cultural characteristics and fathers’ working hours that were not statistically significant in the full model. First, fathers who can easily find a substitute if necessary have a five percentage points higher probability of working part time. Second, fathers who perceive a very strong ideal worker norm in their workplace (scale value of 10) have a contractual working time of nearly 2 h longer than fathers who do not perceive an ideal worker norm in their workplace (scale value of 0). Using interaction models we checked whether the two effects of the substitution situation and the ideal worker norm, which disappear when controlling for workplace structural characteristics, only occur in certain industries, size structures, etc. We find no empirical evidence for this.

This suggests the possibility of mediation effects (instead of moderation effects): particular industry-specific and workplace-specific structural conditions may support particular working cultures. Descriptive analyses of the relationships between an organisation’s cultural and structural characteristics partially support this hypothesis. In particular, the organisation’s degree of family-friendly regulation is significantly higher in the public sector, in large organisations and in those with works councils, whereas it is lower in the retail sector than in other sectors.Footnote 7

In Table 4, we therefore also examine the question of correlations between workplace structural characteristics and the working hours of fathers (without controlling for workplace cultural characteristics). The results are very stable compared to the overall model; the coefficients are only slightly larger. This robustness test also indicates that cultural and structural relationships with fathers’ working hours are relatively independent of each other.

Table 4 Relationships between workplace structural characteristics and the working hours of fathers (without controlling for workplace cultural characteristics)
Full size table

Furthermore, there are no robust interactions among the variables of workplace culture, for example, as suggested by the hypothesis that the effectiveness of formal policies depends on a family-friendly culture in the workplace (Lewis 1997). A significant interaction would suggest that the lower the ideal worker norm in the organisation is, the stronger the correlation between the degree of formalised, universal and transparent regulation and fathers’ working hours. However, we do not find any empirical evidence for this.

Nor do we find any evidence for selection effects in relation to fathers with egalitarian attitudes. Fathers with egalitarian attitudes could make family-friendliness and the possibility of part-time work a decisive selection criterion when looking for a job. In the analyses, we would then have to observe a stronger relationship between workplace characteristics and working hours for these fathers. This is not the case. However, since we only have one indicator of gender role attitudes, further research is needed to back up this finding.

4 Discussion and Conclusion

Many fathers in Germany want shorter working hours but still work full time and often more hours than contractually agreed. A growing number of primarily qualitative and experimental studies indicate that workplace conditions are an important obstacle for fathers who wish to put their preferences for working shorter hours into practice. However, we still lacked evidence from quantitative studies due to a lack of data. Based on original data on couples from 2015, the present study aimed to contribute to closing this important research gap and to investigate how the working hours of employed fathers in Germany are related to the cultural and structural characteristics of their workplaces.

In particular, the results show a robust correlation between the extent of formal, universally applicable and transparent rules in the workplace and fathers’ working hours. The stronger the fathers’ confidence is that there are family-oriented workplace policies they are entitled to use and that are regularly communicated and supported by the management, the lower the contractual and actual working hours of fathers and the greater the probability that they will work part time. It should be noted that our indicator for workplace policies goes far beyond merely offering family-friendly policies. It is essential that fathers also feel that the organisation’s management is behind it and that, as a part of the target group for family-friendly policies, they are being addressed and regularly informed.

At the same time, the results indicate that the other cultural workplace conditions considered in this study are of particular importance for the actual working hours of fathers, but less so for adjustments to formal working time. A low prevalence of the ideal worker norm and good access to substitutes at short notice are associated with shorter actual working hours. Thus, these workplace culture factors could explain to what extent fathers succeed in minimising overtime and adhering to their contractual working hours. This finding reinforces previous research suggesting that in a paternity-sensitive working time culture, fathers feel less under pressure to demonstrate performance and motivation for career advancement by working long hours (Vinkenburg et al. 2012; Rudman and Mescher 2013; Vandello et al. 2013). Surprisingly, however, fathers’ contractual hours are not related to their perceptions of an ideal worker norm in their workplace. This finding contradicts our hypothesis that fears of being punished as “time deviants” (Epstein et al. 1999) may discourage fathers from officially working reduced hours.

These findings offer new insights for the conceptual and empirical research grounded in Amartya Sen’s capability approach, which argues that asserting and making use of the legal right to work shorter hours requires a sense of entitlement (Hobson and Fahlen 2009; Hobson 2011; Hobson et al. 2011). Our study indicates that workplace cultures influence this perceived entitlement and that formal and informal rules of workplace support for fathers matter in different ways for their formal and actual working times: In the absence of a strong ideal worker norm, fathers feel entitled to limit their actual working hours to contractually agreed hours. However, it needs proactive workplace policies to convey a sense of binding entitlement for fathers to demand, legitimise and realise shorter contractual working hours or part-time work.

Our results also contribute to a better understanding of the relationship between workplace policies and working time norms. We find only limited evidence for the argument that the effectiveness of formal workplace policies may be restricted by the informal working-time culture in organisations (Lewis 1997; Hobson 2011). Even though the results indeed show that a family-unfriendly working-time culture partially offsets the limiting impact of workplace regulation on fathers to work long hours, both factors seem to operate independently from each other. We find no evidence that a strong ideal worker norm weakens the supporting effect of workplace regulation on fathers’ working hours.

The results for supervisor support are somewhat unexpected. They show that fathers have longer contractual working hours and are less often in part-time work if they feel supported by their supervisors. Roeters et al. (2012) come to similar conclusions and assume that the motivation to work increases when fathers feel supported by their supervisors, with the result that they work more instead of less. Another possible explanation is that understanding supervisors enable fathers to adapt their working hours flexibly to their needs and leave early or work at home if necessary. Thus, these fathers may find sufficient time for their family in spite of longer contractual working hours. This interpretation is supported by the fact that their actual working hours do not differ from those of fathers who receive less support from their supervisors. As a result, these fathers may even work less unpaid overtime. Based on our data, however, we cannot rule out the possibility that there is an inverse causal relationship and that supervisors support those fathers who comply with the full-time norm.

The analyses also provide initial indications that regulations and practices that support shorter working hours for fathers are possible under various workplace structural conditions. It is true that fathers in the administrative, healthcare and social services sectors have shorter working hours than fathers in other sectors. In small organisations, too, fathers have formally (but not actually) shorter working hours. However, there is no connection between employment in the public sector or in an organisation with a works council or staff representative council and fathers’ working hours. In addition, the interrelationships between workplace culture characteristics and the working hours of fathers have proven to be robust and independent of the influence of workplace structural features. Overall, the results therefore suggest that it is more about a commitment to creating a family-friendly work environment within organisations than structural barriers such as sector characteristics or a lack of resources in smaller organisations.

However, it should be noted that the data do not allow us to distinguish between different forms of overtime based on the available data when there are deviations between contractual and actual working hours. Unpaid overtime, unlike overtime with time off as compensation, could occur more frequently in organisations with a low degree of family-oriented regulation, a strong ideal worker norm, a lack of support from supervisors, or a lack of substitution options. The payment of overtime, in turn, could provide countervailing incentives to family-friendly conditions by encouraging more work. Furthermore, it is not possible to use the data to determine the causal direction of the correlations found between workplace conditions and fathers’ working hours. It is alternatively conceivable that the relationships could be due to selection effects: fathers who consider it important to get involved in the family and therefore want to work less often may choose organisations in which these wishes can be put into practice more easily. In this study, we could not find any evidence that fathers with egalitarian gender role attitudes work in disproportionately large numbers in certain organisations. However, the question of how workplace policies and cultural practices influence fathers (and mothers) in their employment and working time decisions can only be investigated in longitudinal studies. This would also make it possible to investigate to what extent family-friendly conditions in workplaces open up potential for mothers to work more extensive hours. In the long term, integrating workplace culture characteristics into existing panel surveys would have the potential to close the research gap on the role played by workplaces in gender inequalities in the labour market and in the family.