National family policies and the association between flexible working arrangements and work-to-family conflict across Europe

Objective: This paper explores how national family policies moderate the association between flexible working arrangements and work-to-family conflict across countries. Background: Although flexible working is provided to enhance work-family integration, studies show that it can in fact increase work-to-family conflict. However, certain policy contexts can help moderate this association by introducing contexts that enable workers to use of flexible working arrangements to better meet their family and other life demands. Method: The paper uses the European Working Conditions Survey of 2015 including data from workers with caring responsibilities from across 30 European countries. It uses a multilevel cross-level interaction model to examine how family policies, such as childcare and parental leave policies, can explain the cross-national variation in the association between flexible working arrangements, that is flexitime, working-time autonomy, and teleworking, and work-to-family conflict. Results: At the European average, flexible working was associated with higher levels of work-to-family conflict for workers, with working-time-autonomy being worse for men’s, and teleworking being worse for women’s conflict levels. In countries with generous childcare policies, flexitime was associated with lower levels of work-to-family conflict, especially for women. However, in countries with long mother’s leave, working-time-autonomy was associated with even higher levels of work-to-family conflict for men. Conclusion: The results of this paper evidence how flexible working arrangements need to be introduced in a more holistic manner with possible reforms of wider range of family policies in order for flexible working to meet worker’s work-family integration demands.


Introduction
Increasing number of workers are getting access to Flexible Working Arrangements (FWAs) that give workers more control over when and where they work -such as flexitime and tele/homeworking (Chung, 2024;Chung & Van der Lippe, 2020;Lyness et al., 2012).Especially after the mass scale use of homeworking implemented during the COVID-19 lockdown periods (Eurofound, 2020), FWAs are expected to be widespread in the future (ONS, 2022).Although FWAs are generally provided to enhance work-family integration of workers (European Commission, 2019), a number of studies show that they can in fact increase worker's feeling of work-to-family conflict or not have any real impact (Allen et al., 2013;Bernhardt & Recksiedler, 2022;Michel et al., 2011;van der Lippe & Lippényi, 2020;Yucel & Chung, 2023).This is because FWAs can make workers work harder and longer, increase domestic burden of workers, and increase boundaryblurring between work and family (Chung & Van der Lippe, 2020;Glass & Noonan, 2016;Kurowska, 2020;Lott, 2023;Lott & Chung, 2016).
However, this association can vary across countries.Institutional and cultural contexts shape the extent to which workers are better able to make use of family-friendly arrangements for work-life balance needs (Hobson, 2011;Moreno-Mínguez et al., 2022).Previous studies have evidenced how FWAs can increase or decrease workers' feeling of work-to-family conflict depending on the arrangements in question, how it is used (Abendroth, 2022;Bernhardt & Recksiedler, 2022), the company context (Abendroth, 2022;van der Lippe & Lippényi, 2020) or the national context (e.g., Kurowska, 2020;Lott, 2015) in which it is used.This paper contributes to these studies and studies carried out examining the antecedents of work-family conflict (for example, see the Journal of Family Research Special Issue on Work-family conflict from the perspective of the family (Schulz & Reimann, 2022).It does this by examining the role national family policies, namely childcare provision, mother's and father's leaves, have in shaping the outcomes of FWAs.This is done using the European Working Conditions Survey of 2015 and a multilevel modelling approach.Although this data is from before the pandemic, it is the most recent high-quality cross-national comparative data set that includes both teleworking and flexible schedules, alongside a wide range of work-to-family conflict and other variables of interest to this study.What is more, many of the outcomes of flexible working has not changed, despite the shift in homeworking practices during the pandemic, due to the enduring gender and work norms (Chung, 2022;Lyttelton et al., 2022b).Therefore, the findings of this paper are expected to be applicable for the 'post-pandemic' labour markets.
The main population this paper focuses on are those with caring responsibilities, including parents and those with other care responsibilities, such as caring for the elderly.This is because the demand for and use of FWAs is different for this group of workers compared to others (Singley & Hynes, 2005).However, the results for this population will be compared with the findings for the general population as a robustness check.The gender variation in the association between FWAs and work-to-family conflict, and how the moderating role of family policies differs for men and women is also examined.The results show that different FWAs relate to work-to-family conflict differently for men and women, with teleworking having stronger association with women's, and working-time autonomy with men's work-to-family conflict.In countries with generous childcare policies, the association between flexitime and work-to-family conflict is weaker, especially for women.In countries with lengthy mother's leave, the association between workingtime autonomy and work-to-family conflict is even stronger, especially for men.Further evidence indicates that the role of policies are partially carried out through changing norms around work, work-life balance, and gender.The results of this study help us understand how the policy context is crucial in understanding the efficacy of work-family integration policies.More specifically, the paper shows what additional policy interventions are needed to ensure that flexible working arrangements can meet the goals of enhancing workers' work-family integration.

Definitions
Work-family conflict is the conflict coming from the opposing pressures of work roles and family roles that are mutually incompatible in some respects (Greenhaus & Beutell, 1985).This paper focuses on work-to-family conflict, i.e., when the participation in a work role inhibits the participation in a family role.The reason for this is because very few workers often experience family-to-work conflict (about 3% of our sample of carers, which is similar to what was found in other surveys such as the European Social Survey) while many often experience work-to-family conflict (ranging from 12-22% depending on the type of conflict of our sample of carers).Work-to-family conflict can include time-based and strain-based conflict.The former is when the time spent on the job makes it difficult to tend to family responsibilities.The latter is when the stress coming from the job produces both physical and mental strain symptoms affecting workers' role within the family.Both types of conflicts are included in this paper.Another element included in this paper is work-to-family spill-over, i.e., the feeling that work is spilling over to family life (Schieman et al., 2009), which for FWAs is highly relevant due to its boundary blurring potentials (Clark, 2000).
In this paper, I focus on FWAs that provide workers more control over when or where they work.More specifically, flexitime is having control over the timing of one's work -e.g., to alternate the starting and ending times.Working-time-autonomy gives workers greater freedom to control their work schedule and the number of hours they work, where in many cases constraints such as core hours (e.g., 10am to 4pm) or specified hours of work (e.g.37 hours per week) do not exist (Chung & Van der Lippe, 2020;Lott, 2020).Teleworking allows workers to work outside of their normal work/employer's premises, for example, from home.In this paper, the access to flexitime or working-time-autonomy and the use of teleworking is examined, largely due to the way the data is gathered (More details in the Data section).Although access and use are different, previous studies (Chung & Van der Horst, 2018;Eaton, 2003) have shown the importance of access to FWAs on work-family outcomes, especially when it is genuine access which this data is capturing.

Flexible working and work-to-family conflict
Job demands and resources theory (Bakker & Geurts, 2004;Voydanoff, 2005) suggests that having control over one's work can help workers reduce feelings of work-to-family conflict.Flexitime and working-timeautonomy provide workers with the flexibility and control over the time boundaries between work and family life, enabling workers to shift the time borders between work and family time, reducing a conflict between the two domains (Clark, 2000).Teleworking/homeworking allows workers to meet work and family demands by possibly blending/integrating the two spheres (Clark, 2000).Work-to-family conflict could potentially be reduced due to the time and energy saved from commuting, especially for workers with long commutes (Allen et al., 2015).This explains why some studies provide evidence of the negative association (e.g., Abendroth, 2022;Kelly et al., 2014) or a weak association between FWAs and work-to-family conflict (Allen et al., 2013;Michel et al., 2011).
However, based on recent studies across Europe (Lott, 2020;van der Lippe & Lippényi, 2020;Yucel & Chung, 2023), FWAs, especially teleworking and working-time-autonomy, is expected to be positively associated with work-to-family conflict.There are several reasons for this.Firstly, flexible workers tend to work longer and harder (Glass & Noonan, 2016;Lott, 2023;Lott & Chung, 2016;Wanger & Zapf, 2022), and both working hours and work intensity are key factors explaining work-to-family conflict.Secondly, during periods of increased care demands -e.g.becoming parents, or when informal care demands arise -workers with FWAs were less likely to reduce the amount of paid work they carry out, compared to workers on fixed schedules or working in the office (Chung & Van der Horst, 2018).This results in workers doing more hours of work, both paid and unpaid, again increasing the potential for feeling work-to-family conflict.This may especially be true as workers, especially women, who work flexibly may feel pressured to do more housework and childcare than they would have otherwise when not working flexibly (Hilbrecht et al., 2008;Kim, 2020;Sullivan & Lewis, 2001).This again is likely to increase feelings of work-to-family conflict.Finally, FWAs can increase spill-over of work to the family spheres or increase multi-tasking (Lyttelton et al., 2022a) all of which can increase feelings of work-to-family conflict (Yucel & Chung, 2023).

H1:
The access to or use of FWAs is associated with higher levels of work-to-family conflict.

Variation across gender and arrangements
Teleworking and working-time-autonomy may be worse for work-to-family conflict.This is because they are more likely to increase boundary blurring, (e.g., Abendroth, 2022;Abendroth & Reimann, 2018;Lott, 2020;Lott, 2023;Yucel & Chung, 2023), which increases their potential for increased working hours or overtime, as well as multi-tasking, compared to flexitime where time and place boundaries are clearer (Abendroth, 2022;Chung & Van der Horst, 2020;Glass & Noonan, 2016;Lott & Chung, 2016;Wanger & Zapf, 2022).However, a gender variation in this relationship is expected due to societal gender norms where women are perceived as care-givers and men as breadwinners (Knight & Brinton, 2017).Flexible working enables couples to 'do gender' (Clawson & Gerstel, 2014;West & Zimmerman, 1987) with different arrangements enabling this to different degrees (Chung & Booker, 2023).Women are expected to be able to do more housework and care, and multi-task more often, when working from home as the physical boundaries of work and home are blurred (Lyttelton et al., 2022a;Sullivan & Lewis, 2001).The expansion of paid-work hours, that is enabled by homeworking for women (Chung & Van der Horst, 2018), may also result in a higher likelihood of feeling of work-to-family conflict (Hilbrecht et al., 2008).This is why teleworking may be more likely to be linked with work-to-family conflict for women with caring responsibilities.Men end up expanding their overtime hours especially when having access to working-time-autonomy, as the time boundaries are blurred with this arrangements (Chung & Van der Horst, 2020;Lott & Chung, 2016;Wanger & Zapf, 2022), allowing men to further live up to their 'ideal worker' status (see also, Lott, 2023;Williams, 1999).This is especially true for fathers, whose breadwinning roles are emphasised (Hodges & Budig, 2010).This is why working-time-autonomy is expected to increase work-to-family conflict especially for men.
H2-1: The access to or use of working-time-autonomy and teleworking is more likely to be associated with higher levels of work-to-family conflict compared to flexitime.H2-2: The access to or use of working-time-autonomy is more likely to be associated with higher levels of workto-family conflict especially for men.H2-3: The access to or use of teleworking is more likely to be associated with higher levels of work-to-family conflict especially for women.

National variation in the outcomes of flexible working
Flexible working is not used in a vacuum and the contexts in which it is used matters.Social policy scholars have emphasised the importance of institutional and cultural contexts in shaping the extent to which family-friendly arrangements can truly enable workers' work-life balance (Moreno-Mínguez et al., 2022).Policy and cultural contexts can limit capabilities and 'real' choices of workers even when arrangements are made available at the policy level, by not only limiting access to the arrangements, but also influencing the extent to which the use of the arrangements can truly benefit workers (Hobson, 2011).Similarly, although FWAs can be seen as important resources to enhance workers' capacity to better integrate work with family life (Voydanoff, 2005), under certain contexts, they can rather result in increasing worker's feelings of workfamily conflict, as discussed above.I argue that national policy contexts matter in determining this association.

H3: The association between FWAs and work-to-family conflict varies across countries.
The main focus of this paper is to examine how national family policies -specifically childcare, mother's and father's leave, shape the outcomes of flexible working arrangements.There are two ways in which family policies can help shape the way in which FWAs can help support or deteriorate worker's workfamily/private life integration.Firstly, through providing additional or auxiliary support for parents and carers that allow for a better work-family integration, they enable FWAs to work better for work-family integration.Secondly, through enabling cultural changes in the way people think about work, work-life balance, and gender roles, national policies can also help shape the outcomes of FWAs.
Firstly, family policies can provide additional or auxiliary support enabling workers to use FWA to better to support their work-family integration.One key reason why parents, especially mothers, are likely to experience work-to-family conflict when working flexibly, is due to the double burden they face from both work and family when they work flexibly (Sullivan & Lewis, 2001).Flexible working enables workers with caring responsibilities to meet both work and family demands without sacrificing one or the other (Chung & Van der Horst, 2018;Hilbrecht et al., 2008).Or in fact, FWAs can be used as a way for parents to address childcare gaps while ensuring maximum working hours of parents (Chung & Booker, 2023;Powell & Craig, 2015).However, this also means that flexible working may result in carers doing more of (or feeling the need to do more of) both paid and unpaid work (Kurowska, 2020;Lyttelton et al., 2022b), resulting in higher levels of conflict felt between the two spheres.In countries where parents are less likely to face childcare gaps due to better public childcare provision, such as in countries like Denmark or Norway, it can be expected that FWAs are less likely to be used to fill in the gaps of childcare.In these countries, therefore, FWAs are less likely to lead to workers feeling conflict between work and family.This would especially be true for women as they are generally the ones who are thought to be responsible for childcare.Longer father's leave, found in countries like France and Luxembourg, when done right, can help relieve women's domestic burdens by enabling more fathers to take an active part in the household not only during but well beyond the leave period (Wanger & Zapf, 2022).Under such contexts, flexible working is less likely to result in work-to-family conflict, especially for women (Kurowska, 2020), as their perceived household responsibilities will be smaller.Finally, longer and more generous mother's leave, like the ones found in Slovakia, Estonia, Finland, Hungary etc., reduce periods where families need to balance work and childcare demandsas mothers can focus on care work during the early periods of a child's life.However, the impact of mother's leave on shaping gender norms may be stronger, which is something explored in the next section.
National family policies also shape norms around work, work-life balance, and gender norms which can indirectly shape how FWAs are associated with work-to-family conflict.The extent to which prevailing norms around work adheres to the ideal worker norm, where workers are expected to work long-hours and prioritise work above all else in life, is a driving factor explaining why workers end up working harder and longer when they gain freedom or autonomy at work -namely through FWAs -rather than expand their leisure time (Chung, 2022;Putnam et al., 2014).In fact, the expansion of paid work of flexible workers has been particularly observed in countries (e.g.US, UK) and occupations (e.g.STEM) that adheres more to the ideal-worker culture (e.g., Cech & Blair-Loy, 2014;Glass & Noonan, 2016).In work centric contexts, workers may have to identify more with their work domain, and expand work when there are flexible boundaries between work and private life (Clark, 2000), as failure to do so may threaten their professional status (Mazmanian et al., 2013).This can strengthen the association between FWAs and work-to-family conflict.Similarly, in the ideal worker norm, when workers use FWAs to meet work-family demands, they are more likely to be stigmatised as someone who is not motivated, committed, or productive, and more likely to experience negative career outcomes (Cech & Blair-Loy, 2014;Williams et al., 2013).To overcompensate for this, workers may have no other options but to work harder and longer (Chung & Seo, 2023;Kelliher & Anderson, 2010), again increasing the likelihood of feeling work-to-family conflict.
On the contrary, in countries with a more balanced notion of work and family life, as is the case in countries such as Sweden and the Netherlands, the association between FWAs and work-to-family conflict is expected to be weaker.Workers will not feel pressured to expand their work hours, nor fear negative repercussions when they use FWAs to address work-family demands (Kelly et al., 2014;van der Lippe & Lippényi, 2020).National family policies can change what is considered to be acceptable work-family reconciliation practices (Bünning & Hipp, 2022;Hobson, 2011).Generous family policies at the national level can change the norms around what companies should provide as family-friendly benefits (den Dulk et al., 2013) encouraging the idea that family-friendly FWAs are a part of the general terms of employment, a norm, rather than 'a gift' provided to workers that needs to be reciprocated (Been et al., 2017).There is evidence of this shift in norms coming from generous national family policy contexts.Family policies have been linked to increased company provision of and worker's access to FWAs (Chung, 2019;den Dulk et al., 2013) and reduce stigmatised views against those who use FWAs for care purposes (Chung & Seo, 2023).In sum, generous family policies at the national level shape national norms around what is an appropriate work-life balance practice (Bünning & Hipp, 2022), ensuring it is not at odds with productivity (Been et al., 2017).This can then help ensure that FWAs are used by workers to better balance work with private/family demands.Of the different family policies, 'work-facilitating' policies such as generous childcare policies (Misra et al., 2011) may be most effective, as such policies emphasise the fact that both men and women are in the labour market and in need of more work-family integration.
Family-policies can also shape gender norms that moderate the association between FWAs and work-tofamily conflict.Long-hours based ideal worker culture is inevitably linked to the (hegemonic) masculine work culture and traditional gendered division of labour (Acker, 1990;Berdahl et al., 2018).The 'ideal' worker implicitly is a heterosexual coupled male, because on the one hand, it is only men that are able to work long hours, devote themselves to and only to work, as their female partners carry out all their reproductive work (Williams, 1999).Similarly, women are assumed to not be able to take on this ideal worker norm due to their responsibilities at home.In other words, women are assumed to be limited in their abil-ity to take part in paid-labour as it requires full-commitment to work, which women with caring and domestic responsibilities cannot afford.
Well paid generous ear-marked leaves for fathers and second parents, i.e., father's leave, can shape gender norms so both parents are expected to provide care as well as engage in breadwinning (Bünning, 2015;Wanger & Zapf, 2022).When both men and women are expected to be involved in childcare and housework, the need to balance work with responsibilities outside of work will be considered universal for all workers.In such cultures, the positive association between FWAs and work-to-family conflict is expected to be weaker for both men and women.On the other hand, in countries with very long mother's leave, maternity and parental leaves available for mothers, there are greater expectations towards women to be carrying out majority of the care and housework (Mandel & Semyonov, 2006) and men are generally expected to be responsible for breadwinning.In such contexts, men's flexible working is more likely to result in men working harder and longer, to ensure that they live up to the male breadwinner image when working flexibly.Women in such contexts are likely to be pressured to do more domestic work when working flexibly, especially when working from home (Kurowska, 2020).In this case, both men and women's work-to-family conflict will increase when working flexibly but for different reasons.
In sum, the following hypotheses are expected.

H3-1: In countries with generous family policies, the positive association between FWAs and work-to-family conflict is weaker. H3-2:
In countries with more generous childcare provision, the positive association between FWAs and work-tofamily conflict is weaker, especially for women.H3-3: In countries with generous father's leave, the positive association between FWAs and work-to-family conflict is weaker, especially for women.H3-4: In countries with long and generous mother's leave, the positive association between FWAs and work-tofamily conflict to be stronger, for both men and women.

Data
The paper utilises the 6th wave of the European Working Conditions Survey (EWCS) of 2015 gathered by the European Foundation providing information on several dimensions of working conditions of workers across Europe.Although this data is from the pre-pandemic period, the more recent version, EWCS 2021, does not include information on flexible schedules or wide ranges of work-to-family conflict outcomes which restricts our analysis significantly.Data from 30 countries are used for this paper -namely, the EU28 then member states including the UK, plus Norway and Switzerland.A random stratified sampling procedure was used to gather a representative sample of those aged 15 or over and in employment (minimum 1 hour a week) at the time of the survey and was conducted through face-to-face interviews.Approximately 1000 cases are included per country with varying response rates.Of the total sample, for the analysis, the sample is further limited to those employed by an employer, and with some caring responsibilities.This group of workers have different levels of family demands compared to other workers and are likely to use flexible working to maintain performance and labour market engagement during periods of high family commitments (Chung & Van der Horst, 2018;Singley & Hynes, 2005;Sullivan & Lewis, 2001).Workers with caring responsibilities are defined as those who are currently living with a child, or caring for (grand)children, or an elderly/disabled relative at least several times a month.Those in the armed forces, and in agriculture/fishery are further excluded due to the specific nature of these jobs, and workers over the retirement age of 65 are also excluded from the analysis.Further excluding cases with a missing response in any one of the variables included in the model resulted in 14,200 cases.Please see Appendix A-5 with regards to the reduction of case numbers used for the analysis of this paper.Some additional cases are removed due to the lack of national level context variables -namely, mother's and father's leave data comes from the OECD database, which excludes Bulgaria, Croatia, Cyprus, Malta, and Romania.Norway is further excluded for models with father's leave.For more information about the EWCS data see: https://www.eurofound.europa.eu/surveys/european-working-conditions-surveys,see the online appendix for the descriptive statistics of all variables including missing cases.

Variables
Following previous studies (e.g., Kelly et al., 2014), five work-family outcome variables are used to construct a single work-to-family conflict factor (see Appendix A-1 for details).This includes items such as how often the respondent has in the past 12 months "...keep worrying about work problems when you are not working", "...feel too tired after work to do some of the household jobs which need to be done".All variables are coded so a higher score entails a higher level of work-to-family conflict.A principal components analysis of these variables resulted in a single factor solution, indicating that the variables can be considered to represent one underlying latent factor.The Cronbach alpha was 0.70 indicating internal consistency.
The key independent variable of this paper is the use and access to FWAs (see Appendix A-2 for details).First, arrangements that provide workers control over when they work are included by using the question "How are your working time arrangements set?" where the workers can answer "They are set by the company/organisation with no possibility for changes" (fixed hours=reference group), "You can choose between several fixed working schedules determined by the company/organisation" (employer-flexibility), "You can adapt your working hours within certain limits (e.g.flexitime)" (flexitime), and "Your working hours are entirely determined by yourself" (working-time-autonomy).Although very strictly speaking, these variables measure access not use of flexible schedules, this can be distinguished from access information, for example, based on company policy data.The way the question is posed (e.g."you can", "are entirely determined by yourself"), it is expected that only those who can easily use the arrangements when the need occurs would have chosen the response options.Given this, it is expected that workers who have chosen the flexitime or working-time-autonomy option will regularly work in these ways, especially as there are no variables in the questionnaire that measures use more directly (see also, Lott et al., 2022).Teleworking is measured using the following question."How often have you worked in each location [during the last 12 months in your main paid job]."The answer categories included employer's premises, client's premises, a car or other vehicle, an outside site, your own home, and public space (such as coffee shops, airports etc.).Those who have answered that they work at their own home (14%) or at a public space (9%) at least several times a month are considered to have teleworked.Note that those who work from home could also respond that they worked in a public space.Although not at the core of our interest, another FWA included is time-off work, i.e., the ability of workers to take a couple of hours off work to tend to personal issues, which has been shown to increase workers' work-life balance outcomes (Chung, 2019;Eurofound, 2016).These variables are examined separately as previous studies have shown how these arrangements, despite their similarities, have different outcomes (Abendroth, 2022;Lott et al., 2022;Wanger & Zapf, 2022).However, as a robustness check, the interaction between flexible schedules and teleworking is also considered.
The model includes job and family demands and resources, individual and workplace characteristics included in previous studies examining worker's work-to-family conflict (e.g., Allen et al., 2013;Michel et al., 2011;Reimann et al., 2022), and access to FWAs (e.g., Chung, 2019;Wiß, 2017).Family resource variables include, whether the worker cohabits with a partner, and the partner's employment status.Family demand variables include household income insecurity, whether the respondent lives with a young child (less than 6 years old), the number of children they have, and other care responsibilities.Job demands variables include work intensity, i.e., whether the respondents need to work at a high speed, on a tight deadline, and their hours worked; and their perceived job insecurity.Job resources variables include career advancement prospects of the job, whether the worker is well-paid, and management support.Occupational level/higher status shapes workers' work-to-family conflict, due to the stress of a higher status (Schieman et al., 2009).This is measured through ISCO 1-digit occupational categories, and a dummy indicating whether the respondent is in a supervisory role is included.Other work characteristics include whether the respondent works in the public sector, whether an employee representative is present at the workplace, whether the respondent has a female boss, the gender composition of the workplace (majority women, majority men, or equally represented (reference group), whether there is individual performance related pay, company size and sectorbased on the NACE R.2 1-digit divisions.Other individual level characteristics include gender, age of the respondent, and their perceived health status.
Family policies are multi-dimensional with very different labour market outcomes, especially for women (Misra et al., 2011).Following previous studies, this paper looks at one overall indicator measuring the general generosity of family policies, and three distinct types of family policies.First, the overall generosity is measured through public expenditure on family policies as a % of the GDP.Second, the proportion of children using formal childcare for age group 0-3 years is used to measure 'work-facilitating' policies, which was found to be key in explaining access to (family-friendly) FWAs (see also, Chung, 2019).All data are for 2015 or closest year available, and from EUROSTAT.Third, mother's and father's leave are included, as previous studies have shown how they can change the gender norms around whose responsibility it is to care (Hobson & Fahlén, 2009).Father's leave is measured as the length of paid paternity, parental, and home care leave available to fathers for the year 2015.Mother's leave is the equivalent available to mothers in weeks and is derived from the OECD Family Policy data base.Note that models including father's and mother's leave do not include all countries, as data were not available.All context variables have been centred and standardized in the model, allowing us to compare the coefficient sizes.For more details on the operationalisation and descriptive analysis of the data please see the Appendix A.

Modelling method
Two-level random-slope multilevel regression models are used.Multilevel modelling methods are used when data is clustered and assumes that the lower-level sample -here individuals -is subject to the influences of groupings/contexts (Hox, 2002), such as country context in this paper's case.First, to examine Hypothesis 1, the model examines the association between the use and access to FWAs and work-to-family conflict without and with all individual and company-level controls.To test Hypotheses 2, an interaction term of FWAs with gender is included, to see whether the association between the access to/use of FWAs and work-to-family conflict vary by gender.To test Hypotheses 3, on how national contexts moderate the association between the access to/use of FWAs and work-to-family conflict, cross-level interaction terms of the national level context variables and FWA variables are included.Finally, gender difference in the moderating role of institutions is tested by interacting gender to the cross-level interaction terms supplemented by analysis separating the male and female sample as a robustness check.STATA 15.1 mixed function (multilevel regression model) is used for the analyses.

Flexible working and work-to-family conflict
Table 1 examines the association between FWAs and work-to-family conflict among caregivers, with (model 1-2) and without (model 1-1) controlling for individual and company-level controls.Confirming hypothesis 1, all FWAs except for time-off work is associated with higher levels of work-to-family conflict.As the coefficients show, the associations between working-time-autonomy and work-to-family conflict, and teleworking and work-to-family conflict are stronger than that for flexitime, yet the difference is only statistically significant in the case of teleworking, partly confirming Hypothesis 2-1.The associations vary by gender.There are significantly different levels of work-to-family conflict between men with vs. without working-timeautonomy, however, for women this difference is barely noticeable (model 1-3; figure 1 left) partly confirming Hypothesis 2-2.Also, compared to those who do not, men and women who telework regularly have higher levels of work-to-family conflict, yet this association is stronger for women (model 1-3; figure 1 right) with a much clearer gender difference in work-to-family conflict between men and women who telework confirming Hypothesis 2-3.
The results, however, need to be taken with a bit of caution.Workers may be taking up FWAs (or have chosen jobs with such access) because they experience work-to-family conflict, indicating a reverse causality.Furthermore, the way teleworking is measured in the data may not necessarily be capturing the opportunities workers have to work from home or outside of the office during normal working hours.It may be capturing the phenomenon of having to work from home/elsewhere in addition to working in the office, to catch up with work demands.In this case, the association between teleworking with higher levels of workto-family conflict is not surprising, although a weaker association could have been expected, given that the model controls for working hours and work intensity.-19060.227 -19088.57 -19060.227Note: *** = p < 0.001, ** = p < 0.010, * = p < 0.050, + = p < 0.100; N level 1= 14,200, N level2=30

Cross-national variations
Table 2, 3 and 4 report how family policies moderate the association between FWAs and work-to-family conflict.Model 2-2 shows that in countries where there are generous childcare policies, the positive association between flexitime and work-to-family conflict is weaker.Model 3-4 shows that in countries where there is longer mother's leave, the positive association between working-time-autonomy and work-to-family conflict is stronger, although it is only significant at the 0.081 level.Each association confirms our hypothesis 3 that there are variations across countries in the way FWAs relate to work-to-family conflict.More specifically, hypothesis 3-2 can be partially accepted in that in countries with generous childcare policies, the association between FWAs and work-to-family conflict is weaker, and hypothesis 3-4 that in countries with generous mother's leave the association is stronger.Hypotheses 3-2 expected that the moderating role of childcare provision will be stronger for women.Model 2-5 and Figure 2 show that childcare coverage moderates the association between flexitime and workto-family conflict, but only for women.Further analyses examining the male and female sub-samples separately confirm this result (Appendix Table D).In countries where public childcare is not widespread, access to flexitime is associated with higher levels of work-to-family conflict for women.In countries with widespread public/formal childcare, women's flexitime is not associated with feelings of work-to-family conflict, or rather it is associated with lower levels of conflict.This could be because in these countries, there are less of an ideal worker norm and generally more support for work-life balance demands for all workers, which allows workers -especially women -to use flexible working in a more family-friendly way.This could also be because generous childcare provisions allow women to have reduced childcare responsibilities allowing them to better use FWAs in a more balanced way, rather than having to use it to meet both high levels of childcare and work demands simultaneously.This could reduce the feelings of conflict between the two.As models 3-5 and 4-5 show, similar patterns emerge where significant gender differences are found in the way family policy expenditure moderates the association between working-time-autonomy and work-tofamily conflict, and how childcare coverage moderates the association between telework and work-to-family conflict.In both cases, the direction of the relationship shows that in countries with generous family policies, women's flexible working is more likely to result in positive outcomes for work-life balance, whereas policy contexts did not matter as much for men, or even exacerbates the association somewhat.However, the significance level is outside of the traditional 0.05 level and further sub-sample analyses show insignificant interaction term for both men and women (available upon request). -

Adjusted Predictions of flexitime
Finally, model 3-6 and Figure 3 show how mother's leave moderates the association between workingtime-autonomy and work-to-family conflict, but only for men (see Appendix Table D2-4).As shown in Figure 3, for women, there is very little difference in work-to-family conflict between those with and without working-time-autonomy (sub sample analysis shows insignificant result for working-time-autonomy).This is especially true in countries where mother's leave is short.In countries where there is long mother's leave, work-to-family conflict levels are lower for women, but the gap between those with and without workingtime-autonomy is larger.This indicates that in countries with long mother's leave, women may be pressured to use the flexibility at work to meet higher demands both at work and at home.There is a large crossnational variance in the work-to-family conflict men with and without working-time-autonomy feel.Men with more autonomy over their working hours feel a much higher level of work-to-family conflict in countries where mother's leave is long compared to when they use the arrangement in countries where mother's leave is short.Contrarily, among men with fixed schedules, work-to-family conflict levels are lower in countries where mother's leave is long compared to where mother's leave is short.This may be because men in countries with long mother's leave may especially feel the need to use the autonomy over their working hours to extend their paid work, as it is likely that their breadwinner role is emphasised (Lott, 2023).On the other hand, in countries where mother's leave is short, there may be less pressures to do so, allowing a somewhat better use of working time autonomy for men.For men with fixed schedules, in countries where mother's leave is long, the expectations towards men with regards to their involvement in domestic life may be lower, and they may be better able to keep clearer boundaries between work and private lives which can reduce the feelings of work-to-family conflict.In sum, hypotheses 3-2 and 3-4 are partially accepted regarding the gender variation in the moderating role of policies.-1.1 -.9 -.7 -.5 -.3 -.1 .1 .3.5 .7 .91.1 1.3 1.5 1.7 1.9 2.1 2.3 -1.1 -.9 -.7 -.5 -.3 -.1 .1 .3.5 .7 .91.1 1.3 1.5 1.7 1.9 2.1 2.3 men women no access wta yes access wta Linear Prediction, Fixed Portion Maternity leave Adjusted Predictions of working time autonomy

Robustness checks
A series of robustness checks were carried out to ensure the reliability of the results of this paper.Firstly, the paper examined analysis results for men and women separately, rather than testing the gender variance in the moderating role of policies using three-way interaction terms (as used in Table 2-4).The results (Appendix Table D) show that firstly, just looking at the female sample, employer's flexible working and working-time-autonomy at the European average do not have a significant association with work-to-family conflict, and flexitime only has a small marginal (p<0.10)association.In comparison, in the male sample, all arrangements have a significant association, although, as mentioned above, teleworking's association is weaker.Table D-2 shows that childcare coverage significantly moderates the association between flexitime and work-to-family conflict only for women (Model D2-1), and mother's leave moderates the association between working-time-autonomy and work-to-family conflict only for men .This confirms what was found in the previous models/tables.
Secondly, although this paper only examined workers with caring responsibilities, studies show how other workers also have demands for work-life balance or better work-family integration (Kelliher et al., 2019;Wilkinson et al., 2017).In addition, when family policies generally influence flexible working outcomes indirectly through the normative changes they make (Hobson, 2011), then all workers, not only parents and carers, could potentially benefit from it.Therefore, additional analyses were carried for the total working-age sample, which is presented in Appendix E, including the male, female sub-sample analyses.
The individual/company level results remain stable in that FWAs are generally associated with higher levels of work-to-family conflict, and that teleworking's association with work-to-family conflict is stronger for women, while working-time-autonomy's association with work-to-family conflict is stronger for men.There were slight differences in the national level cross-level interaction analysis results.Childcare coverage significantly moderates the association between flexitime and work-to-family conflict (Model E2-2), but in this case for both men and women, with a weaker gender variation compared to the parent/carer model (Model E2-6).Having said that, the sub-sample analysis for men shows that it is only significant at the 10% level (Model E5-4).Childcare coverage is also significant in moderating the association between working-timeautonomy and work-to-family conflict (Model E3-2), but again only for women (Model E3-6, E5-5, E5-6).Working-time-autonomy is likely to be associated with lower work-to-family conflict levels for women, unlike for men, especially in countries with greater childcare coverage.In other words, generous childcare coverage may indicate a more dual-earner supportive policies especially facilitating women's employment, and with it, a cultural shift in the way people think about work-life balance and gender roles.Such contexts enable women to use flexible schedules -both flexitime and working-time-autonomy to avoid conflict between work and family life.Working time autonomy, given its work expanding nature for men, does not seem to have the same effect, even in such supportive contexts.Interestingly, family policy expenditure is also found to significantly moderate the association between flexitime and work-to-family conflict, but here only for men (Model E2-1, E2-5, E5-2).Although the negative association exists for women, it is not statistically significant.Next, the moderating role found for mother's leave on the association between workingtime-autonomy and work-to-family conflict does not exist when examining the total population sample (Model E3-7, E5-7, E5-8).In other words, the influence long mother's leave has on shaping father's breadwinning roles, potentially pressuring them to work longer hours when having workingtime-autonomy, does not seem to be true when examining all men in the sample.This makes sense given what is known with regards to men's roles and labour market patterns of fathers compared to non-fathers (Abendroth & Lükemann, 2023;Hodges & Budig, 2010).Finally, the gender varying association found for childcare coverage moderating the association between teleworking and work-to-family conflict also disappears when the total population is examined.Although the direction of the associations remains stable, it is no longer statistically significant even at the 10% level.
Thirdly, national cultural norms such as work-centrality and gender norms are examined directly to test if they moderate the association between FWAs and work-to-family conflict.As results in Appendix Table F-3 show, both work centrality and gender norms are significant in explaining the cross-national variance in the association between flexitime and work-to-family conflict.In countries with work centric views, flexitime's positive association with work-to-family conflict is stronger, while where gender norms are more progressive, the relationship is weaker.Examining the male/female sample separately, it can be seen that this is especially true for women.For women, work centrality also seems to moderate the association between working-time-autonomy and work-to-family conflict.In other words, women's flexitime and working-time-autonomy can potentially help reduce their work-to-family conflict, but only in countries where work is not seen as central to one's lives and where gender norms are more progressive.Examining norms alongside policy variables, in many cases the norm variables remain significant while policy variables become insignificant (see model F3-1 as an example) indicating a potential mediating relationship.However, as the outcomes in Table F-6 shows, this is not necessarily the case when male/female samples are separated out, indicating that policies may have a role in moderating the relationship between FWAs and conflict above and beyond the normative changes they make.However, given that the norm variables (e.g.work centrality) are not necessarily a perfect measure of the extent to which work-life balance is the norm, because the multi-dimensions of gender norms (Knight & Brinton, 2017) could not be included, and because the inclusion of norm variables in the model reduces many country cases, it is hard to make further conclusions.
As a final robustness check, flexible schedule variables were interacted with teleworking to see if using two arrangements simultaneously may help or exacerbate problems with regards to work-to-family conflict.There is little evidence of this (Appendix G), with most showing insignificant results.However, for women (Table G-1), using flexitime and teleworking together may help moderate the positive association both arrangements have with work-to-family conflict.

Conclusions
This paper examined how family policy contexts moderated the association between flexible working and work-to-family conflict.At the European average, the use of flexible working arrangements was positively associated with workers' feeling that work conflicts with family life, with working-time-autonomy and teleworking being generally worse than flexitime.This confirms previous studies (Abendroth, 2022;Abendroth & Reimann, 2018;Glass & Noonan, 2016;Lott & Chung, 2016;Wanger & Zapf, 2022) indicating that arrangements that allow for greater boundary blurring may be more problematic in relation to work-to-family conflict.The association between using working-time-autonomy and feeling work-to-family conflict was stronger for men, where for women at the European average, using working-time-autonomy was not associated with stronger work-to-family conflict.This may be because working-time-autonomy tends to increase working hours and work intensity for men more than for women, resulting in feelings of work-to-family conflict (Lott, 2023;Lott & Chung, 2016;Wanger & Zapf, 2022).The association between teleworking and feeling work-to-family conflict was stronger for women (see also, Kurowska, 2020;van der Lippe & Lippényi, 2020), although in this case the association was positive and significant for men as well.This may be because teleworking can increase the housework and childcare responsibility women feel (Chung & Booker, 2023;Sullivan & Lewis, 2001), resulting in higher levels of work-to-family conflict (Yucel & Chung, 2023).
As the 2022 Special issue in this journal on work-family conflict (Schulz & Reimann, 2022) noted, family and workplace contexts shape how family/job demands and resources relate to individual's work-family conflict.Similarly, the results of this paper has shown that family policy contexts matter in shaping the association between FWAs and work-to-family conflict.Countries where there were generous policies especially supporting dual-earner systems, such as public childcare, seem to weaken the positive association between FWAs and work-to-family conflict, especially for women.This may be due to the cultural changes generous family policies have in shaping norms around work-life balance, or due to a more direct impact of having a combination of generous public childcare and flexible working arrangements on worker's, especially mother's capacity to combine work with family life.On the other hand, in countries with long mother's leave, working-time-autonomy seems to be associated with an even stronger feeling of work-to-family conflict, especially among men.This may be because in these countries, long leaves for mothers may reinforce traditional gender norms around whose role it is to care and to do the breadwinning (Korpi et al., 2013;Mandel & Semyonov, 2006), pushing men with working-time-autonomy to further become workdevoted breadwinners, increasing their working hours and overtime (Chung, 2022;Lott, 2023, Lott & Chung, 2016).Having explored the influence of work and gender norms more directly, evidence of the importance of norms was found.Yet, policies seem to have some influence above and beyond the normative changes they can make.
There are some limitations to this study.First, the paper is limited in its cross-sectional nature, which means the direction of the relationships is not certain.For example, people who feel conflict between work and family life may be the ones who are actively seeking to work flexibly -explaining the positive associa-tion between the two.When possible, future studies should examine longitudinal data from broader range of countries or possibly using a field experimental approach to tease out causality.Second, future studies should also look at the importance of contexts at the organisational level, to see how organisational culture and policies can change the outcomes of flexible working (see for an example, Abendroth, 2022).Third, more investigation is needed to examine other types of contexts that may help shape flexible working outcomes, such as other policies like the right to disconnect (Eurofound, 2023), or socio-economic contexts (Moreno-Mínguez et al., 2022).Finally, as the data used for this paper was collected before the pandemic, a question arises whether the results of this study are applicable in the 'post-lockdown' labour markets.Indeed, flexible working has become more normalised with more workers having greater access to teleworking practices (Eurofound, 2022).Such normalisation can be useful in ensuring that FWAs do not lead to negative work-family outcomes (Abendroth, 2022;van der Lippe & Lippényi, 2020).However, in recent times, more employers are asking workers to return-to-office, and stigmatised views around flexible working are returning (Li, 2022).What is more, evidence gathered even during the peak of the pandemic (e.g., Chung et al., 2021;Lyttelton et al., 2022b) shows that many of the negative outcomes of flexible working observed during the pre-pandemic times largely remained or even exacerbated, due to many of the policy and cultural contexts not having changed during these times.In other words, much of what was found in this paper, even though the data is from pre-pandemic times, is likely to be applicable in the 'post-lockdown' labour markets.However, when possible, further analysis using post-lockdown data would be useful.
Despite these limitations, this paper provides an important contribution showing how developing supportive family policy contexts can help ensure flexible working provides positive outcomes for workers' work-family/private life integration.Some of the negative outcomes of flexible working, such as the encroachment of family life, or blurring of boundaries that have been observed in previous studies (e.g., Chung & Van der Horst, 2020;Glass & Noonan, 2016;Lott, 2023 ;Lott & Chung, 2016), are not inevitable.By providing family-friendly policies that support parents' employment, that does not entrench traditional gender norms and support a more balanced notion of work and work-life balance, flexible working can lead to more positive outcomes for workers.
Flexible working is gaining increased attention due to its potential to enhance workers' well-being, work-family integration, and through that to promote gender equality both at home and in the labour market (Chung, 2024;European Commission, 2019).However, it is not used in a vacuum.The policy context in which flexible working arrangements are used shapes what it can mean for workers.This is why the introduction of flexible working policies alone is not enough to meet these policy goals.A more holistic approach using a wide range of policy instruments is necessary to ensure that flexible working can meet work-life balance and gender equality goals.Without such an approach, flexible working can lead to unintended and damaging outcomes not just for the individual and their families, but ultimately for companies and society as a whole.https://europeanvaluesstudy.eu/methodology-data-documentation/survey-2017/full-release-evs2017/accessible via GESIS.

Figure 1 :
Figure 1:Predicted work-family conflict levels of men and women using working time autonomy/teleworking across 30 European countries

Figure 2 :
Figure 2:Graph depicting the association between flexitime and work-family conflict across countries with different levels of childcare coverage (0-3) for men and women (having controlled for several covariates in Model 2-5)

Table 1 :
Flexible working and work-family conflict for those with care responsibilities across 30 European countries in 2015

Table 2 :
National family policy moderating the association between flexitime and work family conflict for those with care responsibilities (for men and women) across 30 European countries in 2015 Note: model controls for a variety of control variables listed in model 1 of Table1; ***=p < 0.001, ** = p < 0.01, *= p < 0.05, + = p < 0.10

Table 3 :
Country level factors moderating the association between working time autonomy and work family conflict for those with care responsibilities (for men and women) across 30 European countries in 2015 Note: model controls for a variety of control variables listed in model 1 of Table1; ***=p < 0.001, ** = p < 0.01, *= p < 0.05, + = p < 0.10

Table 4 :
Country level factors moderating the association between teleworking work-family conflict for those with care responsibilities across 30 European countries in 2015 Note: model controls for a variety of control variables listed in model 1 in Table1; ***=p < 0.001, ** = p < 0.01, *= p < 0.05, + = p < 0.10