There is a well-known, longstanding gap between people’s actual and preferred work hours (e.g., Jacobs & Gerson, 2001; Knaus & Otterbach, 2019). According to Eurofound (2017) data based on 35 states in Europe, 30% of all employees would prefer to work fewer hours. Furthermore, the working hours mismatch has been studied extensively and has been found to be associated with many negative individual, couple, organizational, and societal consequences (e.g. Bell et al., 2012; Kim & Golden, 2022; Lepinteur, 2019; Swendener, 2021; Wang, 2016; Wunder & Guido, 2013). Although the literature is inclusive, several studies indicate that preferred work hours are gendered. Women seem to be somewhat more affected than men because of childcare responsibilities (Grund et al., 2023), but a recent literature review has concluded that, among heterosexual dual-earner couples, both spouses are more likely to desire a reduction than an increase in their working hours, especially if they are childless and pursue leisure activities (Antal et al., 2024).

These inconclusive and sometimes contradictory findings indicate that, whilst it has been more common for researchers to study hour mismatches at the individual level, a broader understanding of the relationship between gender and preferred work hours among dual-earner couples is needed. Current explanations of hour mismatch generally characterize them as a personally defined problem whose determinants vary with the person (e.g., gender, age, family status) and the job (e.g., work hours, job flexibility) (Bloch & Taylor, 2012; Knaus & Otterbac, 2019; Reynolds, 2003; Reynolds & McKinzie, 2019; Van Echtelt et al., 2006), with only a few exceptions (e.g., Clarkberg & Moen, 2001; Girtz, 2021; Reynolds, 2014; Schalembier et al., 2019; Stier & Lewin-Epstein, 2003; Wielers et al., 2014). The few attempts to gain insights into the preferred work hours of men and women at the couple level in dual-earner couples (Clarkberg & Moen, 2001; Waismel-Manor & Levanon, 2017; Wotschack et al., 2014) have not used a comparative perspective that looks at how societies can strengthen families by shaping the available options for work hour preferences through work-family policies. To advance this literature, we argue that a multilevel family-centered approach will offer a more integrated and contextualized analysis to the study of these work hour preferences among heterosexual dual-earner couples. With this approach, we assess preferred work hours within a family/couple level framework; that is, we investigate not only husbands' and wives' work hour preferences, but also their preferences for their spouses' work hours, together reflecting their preference for work hour arrangements at the couple-level context. This family-centered approach, coupled with insights from the social construction of gender, life-course perspective, and scholarship on welfare policy, will help us shape a more comprehensive theory that could explain gendered working time preferences and decisions.

Using this approach, our study makes two main contributions to the understanding of gendered work hour preferences and to scholarship on the connections between work and family. First, research has generally focused on describing the facts and figures regarding the gap between people’s actual and preferred working hours (e.g. Bielenski et al., 2002; Steiber & Haas, 2018), using the individual as the unit of analysis. Therefore, it has potentially underestimated the effects of spousal behaviors and preferences for work hours of men and women within couples. In addition, multi-level perspectives suggest that these behaviors and preferences should take into consideration country-level characteristics (e.g. Stier & Lewin-Epstein, 2003; Wielers et al., 2014). Thus, following the call to move beyond the 'work-family' frame to include other contexts such as social policy (e.g. Moen, 2011), and Hallgren and Risman’s (2022) recent call to use gender theory to argue for a multilevel analytical approach to research on work–life balance, our study adds to the literature by examining two separate streams of research–couples' adaptive strategies (Moen & Miller, 2021; Moen & Wethington, 1992) and a comparative perspective of work-family policies–that so far have not been analyzed together.

Second, while some studies have examined country-level economic characteristics and welfare policies as determinants of work-hour mismatches (Stier & Lewin-Epstein, 2003; Wielers et al., 2014), they have not investigated work-family policies as determinants of people’s preferences for reduced work hours. This study examines, for the first time, whether a specific set of work-family policies, known as defamilization policies (e.g., length of maternity leave, limitation of work hours, day care subsidies), are related to decisions about the number of hours women and men in dual-earner couples want to work, and the hours they want their spouses to work. As such, we evaluate decision-making at the couple-level, highlighting the strategies that families devise for themselves (as exemplified by their work-hours arrangements); not necessarily what husbands and wives would choose and prefer for themselves and for their families, but which strategies are feasible for dual-earner couples, given their nations' work-family policies, and the gendered structure of work opportunities and gender norms. We claim that nations' work-family policies, intended to help women juggle work and family commitments, offer another layer of help to both spouses in devising their couple-level work hour arrangements in order to reduce their experiences of a work-hour mismatch. In sum, this study proposes a multilevel family-centered approach to investigate the additive roles of couples' work-hour arrangements and nations' work-family policies in shaping men's and women's gendered preferences for reduced work hours for themselves and for their spouses in 19 countries.

A Gendered Perspective on Work Hour Preferences

Work hour preferences are gendered because work and family domains are structured differently for men and women. More couples in contemporary societies have two paid jobs rather than one, but the structure of the workplace and societal expectations often still assume that men are the breadwinners and women are the homemakers (e.g. Acker, 1990; Cha, 2010). Various perspectives suggest that the patterns of the division of paid and unpaid work are highly gendered (Risman, 2004; Shockley et al., 2021; Treas & Tai, 2016). The paradigm of the social construction of gender explains that various social processes, such as socialization, contribute to the asymmetry. Blair-Loy (2003) asserts that couples rarely question the idea that the wife rather than the husband should cut back on work to care for children because of internalized cultural schemas of women’s family devotion. Furthermore, as the time availability theoretical model asserts, while male breadwinners can work long hours because their spouses perform unpaid domestic labor, raise children, and care for elderly relatives, individuals in dual-earner couples face challenges because both spouses are employed (Jacobs & Gerson, 2004; Shockley & Shen, 2016). Given the failure of institutions to recognize and respond to the rise of dual-earner couples, it is not surprising that such couples experience a work hour mismatch and may adopt a full-time/part-time division of primary versus secondary paid work, based on the gendered breadwinner/ homemaker model, often termed as "neo-traditional' dual-earner couples (Moen, 2011).

Risman (2004) conceptualizes gender as a social structure encompassing three dimensions that are equally important: the individual level (e.g., socialization and internalization of cultural schemas), the interactional and cultural expectations (e.g., “doing gender” in social settings; West & Zimmerman, 1987), and the institutional domain (e.g., organizational practices and legal regulations). As Hallgren and Risman (2022) advocate, explicitly drawing on evidence from multiple dimensions is crucial in contextualizing the understanding and analysis of empirical research findings. Adopting this multidimensional structural model allows us to examine whether agentic women and men can effectively reject habitualized gender routines (Risman, 2004). Likewise, life course scholars also maintain that we must look beyond the individual and focus on family roles and relationships and their intersection with gender and the outdated socially constructed structure of time in determining people’s work hours (Kamp Dush et al., 2018; Moen, 2011; Moen & Miller, 2021). This approach underscores how spouses' circumstances, as well as the social structures in which they function, constitute important contextual considerations in understanding the work-hour preferences of those in dual-earner families.

This article borrows from life course scholarship two themes that are important in the context of individuals' lives. The first is the notion of “linked lives,” meaning the ways in which career decision-making is always a social, relational process (Moen & Miller, 2021). Couples choose their work hours and jobs, but these choices continue to be a gendered process (e.g., Ciccia & Bleijenbergh, 2014; Stone, 2007). Therefore, “choices” are often constrained by the spouses’ circumstances, particularly those of the husbands (Moen & Yu, 2000). The second theme of “adaptive strategies” refers to the ways in which spouses seek to maximize their ability to cope with the demands of work and family (Moen & Wethington, 1992), given the existing cultural and structural environment with its gender role expectations and constraints. Spouses jointly formulate adaptive work-family strategies that fit their lifestyles and optimize their effectiveness at work and at home (Moen, 2011; Voydanoff, 2005; Wotschack et al., 2014).

All the above perspectives suggest that these strategies are often gendered, taken for granted, and culturally prescribed responses. However, as the gender structure theory (Risman, 2004) and the life course perspective propose, some couples are deliberately devising their own mechanisms of adaptation (Becker & Moen, 1999; Scurry & Clarke, 2021; Shockley et al., 2021; Wotschack et al., 2014). Hence, the analysis below uses couples’ joint work hours as a measure of adaptive family strategies and preferences for work hours for themselves and for their spouses to evaluate which strategies are feasible for dual-earner couples, given their nations' work-family policies, and the gendered structure of work opportunities and gender norms.

Neo-Traditional Gendered Division of Labor and Work Hour Preferences at the Couple Level

The above discussion leads to an expectation that men and women in dual-earner couples will seek work-hour arrangements that most closely reproduce the neo-traditional gendered division of paid labor. Specifically, couples who follow the neo-traditional model, which prescribes a gendered division of paid labor, in which the husband works (at least) full-time, while the wife works fewer hours, are expected to be the least likely to prefer to reduce the work time for themselves and their spouses than those who adopt less conventional family adaptive strategies. Faced with lagging institutionalized gendered practices, those who adopt the neo-traditional model will experience a more limited work hours' mismatch. In other words, these couples will be the most satisfied with their existing work time arrangements. In contrast, those who have created arrangements that do not conform to this model will experience greater friction caused by the mismatch between their proposed model and the prevailing gendered cultural norms. As a result, there will be a greater disparity between their actual and preferred work hours.

A few studies have conceptualized work-hour preferences at the couple level. Examining a nationally representative sample of married American working couples, Clarkberg and Moen (2001) reported that husbands wanted to work full-time (rather than long hours), and wives wanted to work less than full-time hours. Neo-traditional couples (husbands working full-time [35–44 h per week] and wives 34 h or less), full-time working couples, as well as couples with a husband employed part-time, were least likely to experience work hours mismatch. Wotschack et al. (2014) obtained similar results using a Dutch sample. They reported finding that the mean gap between the spouses' actual hours and desired hours was larger for dual-earner couples (defined as both spouses working at least 33 h per week) than for one and a half earner couples (households in which one spouse worked at least 33 h per week and the other between 8 and 33 h). Waismel-Manor and Levanon (2017) also documented similar results for 14 European countries using data from the European Social Survey (see also Kanji & Samuel, 2017). In sum, these previous studies established that work-hour preferences are gendered and reflect the "ideal employee" work time culture. Couples who create work-hour arrangements that closely reflect the neo-traditional work-hour arrangements are the least likely to prefer reduced work hours.

As the above gendered discussion at the couple analytical level suggests, because of widely held gendered norms and institutional arrangements and different positions of men and women in the social structure, we expect men and women in dual-earner couples to report similar gendered preferred couple-level work hours' configurations (i.e., the neo-traditional work-hour arrangements). However, there may be meaningful gender differences in the strength of these relationships between actual and preferred family adaptive strategies. As Risman asserts (2004), “we must pay close attention to what men do to preserve their power and privilege” (p. 438). The bargaining/exchange theory, which predicts that a spouse who has more resources is better able to bargain for what he or she wants (Blood & Wolf, 1960), suggests that because husbands have culturally and materially more power than wives in decision-making processes within the family, they may negotiate a better match than wives between their actual and preferred family adaptive strategy. Thus, couples’ adaptive strategies may reflect men's preferences more closely than women's preferences.

Based on this research, we posit that:

  • H1: Individuals in dual-earner couples who follow the gendered division of paid labor in which the husband works (at least) full-time, while the wife works fewer hours, will be less likely to prefer to reduce the work time for themselves and their spouses than those who adopt less conventional family adaptive strategies (such as, husbands work part-time, both working long hours).

  • H1a: This relationship between work-hour arrangements and preferences will be stronger among men than among women.

National Work Family Policies and Dual-Earner Couples’ Work Hour Arrangements

Using an interdisciplinary life course approach, the current study also assesses individuals’ and couples' adaptive strategies and work-hour preferences as embedded in institutional and social contexts. While we expect that work-family policies will not nullify the association between couples' work-hour arrangements and preferences, as each country’s social context is different, we do expect that these policies will add an additional layer of influence onto their preferences (Reynolds, 2004). Cross-national studies have explained work-hour preferences using macro-level country characteristics such as the standard of living (GDP per capita), public social welfare efforts (measured by social expenditures), and the characteristics of the labor force (such as women’s labor force participation) (Sousa-Poza & Henneberger, 2002; Stier & Lewin-Epstein, 2003; Wielers et al., 2014).

However, the extant research has largely neglected to address the role of a nation’s work-family policies in shaping the available options for employment preferences for women and their spouses. Such policies seek to reduce people’s dependence on family relationships and narrow social and economic gender gaps by providing public childcare and paid maternity/parental leave, as well as regulations regarding the timing of work and the tax structure (Hegewisch & Gornick, 2011; Zagel & Lohmann, 2020). Countries with more developed work-family policies exhibit a greater share of dual full-time working couples, and a lower share of single earners families (Abendroth et al., 2012; Hook, 2015; Lewis et al., 2008). A single past study examined the role of these policies in shaping hour preferences only for women, documenting, for example, that public childcare is associated with smaller gaps in the preferred working hours between mothers and childless women (Pollmann-Schult, 2016). Expanding on prior research, we argue that these policies not only promote women's independence but also allow both spouses to work the hours they prefer, thus supporting gender equality within the family and offering more resources to the family. The main mechanism through which these policies impact the gap between actual and preferred work hours is through the reconciliation of the demands of the family and the workplace. In countries where such policies are underdeveloped, and hence the pressures of lagging institutionalized gendered practices are more pronounced, dual-earners couples are expected to have a stronger preference for fewer working hours.

A key topic in comparative literature on social policies is how to categorize countries into policy constellations, more commonly known in this literature as regimes, based on structural, institutional, political, and cultural aspects (e.g., Olsen, 2021; Pfau-Effinger et al., 2009). A prominent typology created by Gornick and Meyers (2004) combines extensive parental time for children with a gender-egalitarian division of labor (for similar configurations, see Budig et al., 2012; Ferragina, 2020; Javornik, 2014; Korpi et al., 2013; Lohmann & Zagel, 2016; Naldini et al., 2016). Social and labor market policies that support such a society are the most developed in the Nordic countries, with the conservative countries of Continental Europe, liberal Anglo-Saxon countries, as well as Mediterranean countries and most Eastern European (with the exception of Slovenia and Lithuania, which are not represented in our study), lagging substantially behind. It is crucial to note that for the latter group of countries economic necessity and the (declining role of the) legacy of the Soviet era, might suggest a preference for greater equality (e.g., Pascall & Manning, 2000).

A more recent typology, developed by Thévenon (2011), focuses on support provided to working parents with young children and on the generosity and length of parental leave entitlements. According to this framework, the Nordic countries are distinctive in providing continuous strong support for working parents of young children. Continental Europe, by contrast, provide high financial support for families but not services for dual earner families. Finally, Eastern and Southern countries generally allow for long leaves but afford limited support for families. We argue that such differences in state support for reconciling the demands of work and care can shape work hours preference, with greater support manifesting in a reduced preference for fewer working hours.

Therefore, we hypothesize that:

  • H2: The greater the development of work-family policies in a country, the less likely individuals in dual-earner couples are to prefer reduced work hours for themselves and their partners.

  • H2a: This relationship between work-family policies in a country and preferences will be stronger among women than among men.

Method

Participants and Dataset

To test our hypotheses, data from the 2010 European Social Survey (ESS) were used. The ESS is a cross-sectional survey that has been administered biannually in multiple European countries since 2002. This study used data from the fifth round of the ESS, which included a comprehensive module about attitudes toward family and work. While information for round 5 of the ESS was collected in 28 countries, the sample was limited to the 19 countries for which there was prior information on work-hour arrangements, preferences, and policies (c.f., Gornick & Meyers, 2004; Kanji & Samuel, 2017). While the surveyed population included everyone 15 and older, we limited our analysis to working age adults (i.e., ages 18–64). We further limited our analyses to working heterosexual salaried individuals (i.e., the unemployed, self-employed and people who were not in the labor force were excluded), who reported that they lived with a partner, regardless of their marital status. Considering the gender differences in work time preferences discussed above, the analysis was conducted separately for men and women. In addition, to reflect the target population, a post-stratification weight was used, which includes a design weight. The ESS recommends using such weights to reduce sampling errors and potential non-response bias. Taken together, the data include information on 4,381 women and 4,007 men.

Measures for Variables

Working Hour Preferences for Self and Partner

The calculation for the dependent variable – preferences for reduced hours – is based on the combination of responses to four questions. The first two original questions asked about the actual work hours of the respondent and the spouse. The remaining two original questions tallied the preferred work hours for the respondents and their preferences for the work hours of their spouse. Specifically, one question asked: “How many hours a week, if any, would you choose to work, bearing in mind that your earnings would go up or down according to how many hours you work?” The second question similarly asked: “If you could choose, how many hours a week, if any, would you like your partner to work, bearing in mind that your partner’s earning would go up or down according to how many hours s/he works?”.

The dependent variable was constructed with these variables by using a nominal scale that includes four categories: (1) those who would prefer to reduce only their own work time (“reduce only for self” category), (2) those who would prefer to reduce only their partner's work time (“reduce only for partner” category), (3) those who would prefer to reduce their work time and their partner’s (“reduce for both” category), and (4) those who would prefer not to reduce their work time or their partner’s (“not reduce” category, which serves as a reference category). The latter category reflects both those reporting that they would prefer to increase their own work hours and those of their spouse with those preferring not to increase or decrease their work hours (i.e., satisfied with work hours). Questions regarding the spouses' actual work hours and the preferences of the respondents regarding the work hours of their partners are unique for the data analyzed in this study. Hence, while the study lacks longitudinal information and does not cover both partners within each household, the inclusion of questions on one’s partner allows extending prior research to include couple-level dynamics, which is one of the main contributions of this study.

Couples’ Working Arrangements

The study focuses on couples' work-hour arrangements as the main micro-level independent variable. Based on working time policies (Gornick & Meyers, 2004), the variable first distinguishes between individuals working part-time (0–29 h a week), full-time (30–44) and over-time (45 and above), for both the respondents and their spouses. Based on Clarkberg and Moen (2001), couple-level work time strategies were identified by cross-classifying actual work-hour strategies for both the respondents and their spouses. While the total potential number of strategies was nine (3 × 3), we decided to combine some categories based on both their sample size and the similarity in the reports of the respondents in those groups regarding their preferences. Doing so resulted in six categories: (1) husband works part-time (combining the miniscule categories where the husband works part-time while the wife works part-time, full-time or long hours); (2) neo-traditional couples (i.e., husband works full-time and wife works part-time); (3) neo-traditional couples with long hours (i.e., husband works over-time and wife works part-time); (4) dual career (both work full-time – used as the reference category); (5) dual career long hours (husband works long hours and wife works either full-time or long hours); and (6) couples where the wife is the main provider (wife works long hours and husband works full-time). The relative share of each category for women and men in our sample is reported in Table 3 in the results section.

Control Variables

Several individual and couple-level control variables are also included, such as age (grand mean centered), education (measured in years, grand mean centered), presence of children in the household (two dummy variables for children ages 0–6 = yes and children ages 7–18 = yes; compared to those with no children or with older children, which are the reference categories), and household income (measured in deciles, grand mean centered).

An additional variable tapped gender ideology attitudes by calculating the mean of two 5-point scale variables (from 1 = strongly disagree to 5 = strongly agree): whether women should be prepared to cut down on paid work for the sake of the family and whether men should have more right to a job than women when jobs are scarce (the variable was grand mean centered).

An additional control variable was based on the occupation of the respondent and the spouse. Using 1-digit ISCO88 codes, it was first determined whether the respondent and spouse held a managerial or professional occupation. Based on these two variables, we distinguished among couples where (1) only the respondent held a professional or managerial occupation, (2) couples where only the partner held a professional or managerial occupation, (3) couples where both the respondent and spouse held a professional or managerial occupation, and (4) couples where neither of them held a professional or managerial occupation (the reference category). A correlation matrix for the relevant level-1 variables appears in Table 1.

Table 1 Correlation Matrix for (Continuous) Level-1 Study Variables
Full size table

We also included an indicator for work autonomy using responses on three 11-point scale variables: (1) whether the respondent was allowed to decide how his/her work was organized; (2) whether the respondent was allowed to influence decisions about the activities in the workplace; and (3) whether the respondent was allowed to change the pace of his/her work. Finally, the analysis accounts for two potential choices that might operate differently across the studied countries: choosing to live with a partner, and choosing to be employed. The two variables dealing with these choices are based on the inverse Mills ratios produced following a probit analysis predicting living with a partner OR employment (respectively).

Work-Family Policies

The main macro-level variable includes five groups of national work-family policies (Budig et al., 2012; Gornick & Meyers, 2004; Lohmann & Zagel, 2016), those in Anglo-Saxon countries (Ireland and Great Britain); Continental Europe (Belgium, France, Germany, Netherlands, and Switzerland); Eastern Europe (Bulgaria, the Czech Republic, Estonia, Hungary, Poland, and Russia); Mediterranean countries (Greece, Israel, and Spain); and Nordic countries (Denmark, Finland, Norway, and Sweden). For a review and assessment of the similarities and differences among these categorizations, see Lohmann and Zagel (2016). In order to account for differences across countries in potential confounding processes, models where we separately replaced the regime dummies with indictors for specific policies were also estimated (i.e., weeks of maternity leave, share of children ages 0–2 in public childcare, and the gender neutrality of the tax benefit system), as were models that separately included the following indicators: GDP per capita, unemployment rate, mean weekly work hours (calculated separately for men and women), female labor force participation rates, the gender wage ratio, percentage of public sector employment and percentage of men working part-time. Table 2 presents the definitions and sources for the macro-level variables.

Table 2 Sources for Country-Level Variables
Full size table

Method of Analysis

Given the nested structure of our data, the analyses employed multilevel multinomial models using the command GLLAMM available through the statistical software program STATA. The model was specified as follows (De Iaco et al., 2019):

$${\eta }_{ij}^{s}={\alpha }^{s}+{\beta }^{s}{x}_{ij}+{\varepsilon }_{j}^{s}+{\delta }_{ij}^{s}$$

where

$${\eta }_{ij}^{s}=\text{log}(\frac{{P}_{is}}{{}_{i1}})$$

And

$${P}_{is}=\frac{\text{exp}({\eta }_{ij}^{s})}{1+\sum_{r=6}^{S}({\eta }_{ij}^{s})}$$

P denotes the probability of response s for individual i, while J indexes the countries. In addition, α reflects the category-specific intercept, x and β are vectors of the independent variables and their coefficients, and ε and δ are vectors of random errors representing unobserved heterogeneity at the country and individual level (respectively). Bryan and Jenkins (2016) have recently suggested that the use of multilevel models with data containing large samples of individuals within countries but only a limited set of countries can result in unreliable estimates of the country effects. They suggest a threefold approach for addressing the problem, including descriptive analysis of measured country differences, bootstrapping, and the application of Bayesian methods of estimation and inference (see also Hamaker & Klugkist, 2011). Below, we provide information on all three strategies, specifically in Fig. 2

Results

Table 3 presents the descriptive statistics separately for men and women. Notably, for both women and men, many respondents said they wanted fewer work hours for themselves or for both partners. More men than women wanted to reduce their own work hours, mirroring the arrangement for actual work hours. In addition, 64% of women and 69% of men were currently in dual-career households, with or without long work hours. However, a sizable minority of couples still had an arrangement where the husband worked longer hours, as reflected in the neo-traditional arrangements. More men than women reported having a professional or managerial occupation, reflecting lingering patterns of gender segregation. The study also documented minor gender differences for the control variables.

Table 3 Percentages and Means (Standard Deviations) for Couple-Level Variables
Full size table

Figures 1 and 2 illustrate the association between the dependent variable and our primary individual and country-level variables. Figure 1 depicts the couples’ current workhour arrangements. The study excluded couples in which the wife was the breadwinner or where the husband worked part-time because these arrangements were relatively rare: 11% for women and 8% for men. Notably, men and women residing in arrangements where at least one person worked long hours were the ones reporting the greatest desire to reduce work hours. There was a consensus among men and women in neo-traditional arrangements in which one person worked long hours that men should reduce their work hours, evident in the responses of the women about their partners and the men about themselves. Interestingly, women in the neo-traditional arrangement and men in dual-career households expressed no desire to change their own work hours. One possible explanation is gender differences in care and household work, which remain unequal in most countries.

Fig. 1
figure 1

Distribution of Preferences for Reduced Work Hours by Couples’ Work Hour Arrangements

Full size image
Fig. 2
figure 2

Distribution of Preferences for Reduced Work Hours by Work-Family Policy Regimes

Full size image

Figure 2 highlights the differences in the preference for reduced work hours by work-family regime. Nordic couples were less interested in reducing their work hours than couples in Anglo-Saxon or Continental European countries, which potentially reflects the greater availability of policies aiming to reconcile work and family demands. For couples in Eastern Europe and the Mediterranean region, especially women in the former and men in the latter group, a large share also reported a preference for not changing their work schedules. However, in these cases, the preference may reflect anticipated economic necessity. Finally, the largest share of respondents reporting a desire to change their own work hours or those of both spouses, especially for men, was in the Continental European, where the gendered division of work remains highly uneven (Berghammer, 2014).

Table 4 presents the fit statistics of four groups of multilevel multinomial models predicting couples’ work time preferences. As Table 4 indicates, Model 2 provides the best fit with the data, based on the BIC index scores (Raftery, 1995). Therefore, we used the coefficients pertaining only to Model 2 in our analysis.

Table 4 Model Fit Statistics for Multilevel Multinomial Models Predicting Preferences for Reduced Work Hours
Full size table

Table 5 presents the odds ratios from Model 2, where the couples’ work time preferences were estimated separately for men and women, and the predictors included individual-level variables and the work-family regime dummies. Several important findings are presented in Table 5. First, men and women who worked long hours, as well as arrangements where the woman was the main provider, expressed the greatest desire to reduce the work hours of both spouses. For instance, men in dual-career households who worked long hours were almost 14 times more likely to indicate that they wanted both spouses to reduce their work hours than men in neo-traditional households. In addition, men and women in dual earner couples were less likely to prefer that men reduce their work hours than their counterparts in neo-traditional couples, but more likely to prefer that women reduce their work hours. As anticipated in hypothesis H1a, we documented gender differences in the strength of the associations between work-hour arrangements and preferences. For example, women in dual-career couples were 48% less likely to prefer reducing their spouse’s hours than women in neo-traditional couples. Men in dual-career households seemed to agree, as they were 58% less likely to report a preference for reducing their work hours. These findings indicate that long hours not only increase the desire to reduce one’s work hours and those of one’s spouse, but also a gendered desire for a neo-traditional work arrangement, supported more strongly by men. This goal is most clearly illustrated by the desire of both men and women in dual-career households who do not work long hours to reduce the wife's work hours, but not those of the husband. Therefore, these results support hypothesis H1 that women and men prefer a neo-traditional gendered arrangement, in which the wife works fewer hours than her husband. This is the case even among couples who do not work long hours.

Table 5 Odds Ratios (and Standard Errors) of a Multilevel Multinomial Model Predicting Preferences for Reduced Work Hours
Full size table

The results presented in Table 5 also allow us to investigate cross-national variations in couples’ worktime preferences. Men and women in Continental Europe and Anglo-Saxon countries reported a greater desire to reduce their work hours, compared to couples in the Nordic countries, where work-family policies are more developed. In line with Hypothesis 2a, in Continental Europe the gaps are stronger for women than for men. By contrast, couples in Eastern Europe were the least likely to report a desire to reduce the work hours of both men and women.

The results presented in Table 4 and Table 5 remained robust even after conducting various sensitivity tests and using several specifications (please see Table S1, S2, S3, S4, and S5 in the online supplement for these results). First, the model presented in Table 5 was replicated, while separately adding several country-level family policy and demographic factors that could potentially influence gender dynamics in work-hour preferences. Thus, country-level fertility rates, divorce rates and percentage of employees with a post-secondary education were included. BIC indices for all models were higher than the score for the initial results. Furthermore, the findings for the association between the couple’s work-hour arrangements and work-family policies, and preferences for reduced work hours remained the same. Similarly, the model presented in Table 5 was estimated while separately adding controls for economic indicators that could potentially influence work time preferences. Thus, the level of income inequality as measured by the GINI coefficient, union density and contemporaneous inflation rates are included. Again, the BIC scores were higher than those reported for Model 2 in Table 4. Furthermore, the overall findings regarding the couple’s work-hour arrangements and work-family policies remained the same. Finally, following the recommendation of Bryan and Jenkins (2016), we reestimated model 2 reported in Table 4, while using bootstrapping and Bayesian estimation. Results presented in Table S5 (please see in the online supplement for these results) remained similar for men and women in Continental Europe, Anglo-and Eastern Europe. The only change recorded was for the preference for reducing hours of work for oneself among men and women in the Mediterranean countries, which were significantly higher than for respondents in the Nordic regime only in the Bayesian estimation. This slight deviation, however, does not alter our main conclusions.

Discussion

Our analysis of work time preferences in contemporary European countries was centered on the multilayered family-related contexts within which individuals make decisions. Our goal was to investigate whether dynamics operating at the couple level can influence people’s work-hour preferences, specifically, their preferences for reduced work hours, for both themselves and their spouse. This was accomplished by measuring the couples’ work-hour arrangements and considering work-family policies at the country level. Given the lag that exists in most countries between traditional family responsibilities and identities and work-related roles and perceptions, the study examined whether men and women in dual-earner couples generally report preferences that conform to the traditional division of labor. The expectation motivating the study was that individuals in couples following the neo-traditional gendered division of paid labor would be the least likely to want to reduce their work time. In addition, the expectation was that people living in countries with more well-developed work-family policies would be less likely to want to reduce their work hours or those of their partners. We also expected to find gender differences in the strength of these relationships.

The study documented a pervasive preference for reduced work hours among both men and women in dual-earner couples in European countries. It was found that, while actual couple-level working combinations often deviate from the gendered division of paid labor, the preferences of both men and women, and across actual worktime arrangements, is for a traditional combination where the husband works longer hours. However, because men are overrepresented in jobs requiring long work hours, men and women alike expressed a desire for reducing the work hours of men in arrangements that require them to do so. A similar and more gendered preference for women to reduce their work hours whenever they work more than what is traditionally mandated by the male breadwinner model also commonly appeared for men and women. In line with the understanding of gender as a social structure, we documented gender gaps in the strength (but not in the direction) of the associations between couple level adaptive strategies and work hour preference, as well as the associations between work-family policies and work hour preference. Specifically, we found that the former association is stronger for men, while the latter association is stronger for women.

The picture that emerges from our analysis suggests striking agreement between men and women about the ideal couple-level work time arrangement, which typically combines a husband who works full-time, but not long hours, and a wife who works shorter hours, either in a part-time or full-time job. These results are similar to other studies that have used couples’ work-hour arrangements as indicators of couple-level adaptive strategies (e.g., Clarkberg & Moen, 2001; Reynolds, 2004; Waismel-Manor & Levanon, 2017). However, these studies have not concurrently compared the specific institutional, social, and economic environments in which the dual-earner couples are embedded. Focusing on the family, this study examined the difference that the work-family policy can make for individuals’ work-hour preferences. More well-developed work-family policies can combat gendered work-hour mismatches in a country, as individuals in dual-earner couples in these countries are less likely to prefer reduced work hours for themselves and their partners.

Two theoretical contributions of this study are particularly noteworthy. First, given that most studies focus on individuals, this study broadened the discussion of work-hour preferences by investigating couple-level work time and work-family policies simultaneously. By doing so, it demonstrated the importance of considering the family as an important multi-level theoretical issue. Although the life course approach highlights the importance of context, culture and institutions, studies often do not examine these factors empirically or in the context of couples’ adaptive strategies. Studying work-family policies is important because they provide resources that a family can rely on to accommodate their work-hour preferences and promote their adaptive strategies. Second, the study contributes to the literature on social policy by showing that generous work-family policies are effective in minimizing preferences for reduced work hours. Clearly, policies other than family related ones can be influential, as other studies have shown (Stier & Lewin-Epstein, 2003). However, the article highlighted the role of work-family policies, which have not been examined before as predictors of work-hour preferences.

Limitations and Future Research Directions

Like any research, our study has some limitations. First, given that our study is cross-sectional, we are unable to make inferences regarding the causal status of our variables. Particularly among dual-earner couples, where work and family are so intertwined, it is likely that the desired hours and the couples’ strategies have mutual influences on one another. Second, our analysis was limited to heterosexual couples who experience the structural lag most strongly. Future research is needed to expand the generalizability of our findings to other groups such as same-sex couples and single parents. Third, due to the small size of the “do not reduce” category we could not distinguish between employees who were satisfied with their work hours and those who preferred an increase in work hours. Future studies should try to examine them separately as they may be connected to different circumstances such as a low standard of living and countries with high rates of unemployment. Other adaptive strategies that can be investigated in the face of hours mismatch, would be for one partner (possibly the wife) to leave the workforce, and/or to change one or the others' hours, or if not possible, to change one's preferences. Fourth, while we borrow from the life course perspective in investigating couples’ adaptative strategies, we do not address the changing nature of these strategies and their implications for work hour preferences across different stages of the life course. This important endeavor remains a crucial potential next step.

Finally, in order to take a comparative perspective of not only people’s work hour preferences, but also their preferences for their spouses' work hours, we used data that dates back to 2010, in the aftermath, and in some countries during the Great Recession, at a time when unemployment was skyrocketing, and many households were facing economic insecurity and financial stress. To generalize our findings, and in light of the demand for remote working in the aftermath of the COVID-19 crisis, at least among professionals, scholars should include this factor in future studies of work-hour preferences, thus broadening the scope of adaptive strategies that couples can use.

Practice Implications

Our study contributes to a growing literature interested in understanding how gender, family, and employment interact and shape people's work and family lives. Our results highlight that, even among countries offering generous work-family policies, dual-earner couples in Europe still experience a structural lag. In a world geared to a strict gendered division of labor between home and work, couples respond to the challenge as best they can. Nevertheless, only those who use a modified version of the breadwinner/homemaker template in which the husband works full-time, but not long hours, and the wife works shorter hours than her husband are less likely to prefer reducing their work hours. It is important to recognize that this adaptive strategy also preserves gender inequalities at home and at work (McConnon et al., 2022; Moen & Yu, 2000; Waismel-Manor & Levanon, 2017). Our findings indicate that work-family policies can moderate these gendered inequities while at the same time extending their reach to the whole family unit, reducing the working hours mismatch not only for women but also for men.

Conclusion

Building on the gender as a social structure perspective and the life course perspective, particularly at the interactional and institutional levels, we find that gender is still a potent determinant of work hours preferences among dual-earner couples, and men and women have similar gendered preferences for themselves and their spouses. Women generally prefer to work fewer hours because of home obligations, and men prefer to work longer hours than their wives because of breadwinner obligations. People may want to reduce their work hours and those of their spouse, but their options are currently constrained. Hopefully, new welfare policies and organizational practices will be accompanied by changes in societal norms. Indeed, much scholarship shows that norms change when policies change (Knight & Brinton, 2017). Challenging the masculine "ideal worker" norm will benefit individuals and couples and help them find a better sense of fit between their work and the rest of their lives.