1 Introduction

The spread of information and communication technologies (ICTs) has facilitated greater flexibility in the timing and location of work (Messenger und Gschwind 2016)—a phenomenon that was accelerated by the COVID-19 pandemic (Abendroth et al. 2022). Some researchers have argued that digital communication serves the flexibility interests of supervisors by making employees more available even outside shared work locations and regular working hours and have underscored their arguments with terms such as “electronic leash” (Arnold 2003, p. 243; see also Duxbury et al. 2014; Piszczek 2017) and “constant connectivity” (Wajcman und Rose 2011, p. 959). By contrast, other researchers have argued that digital work communication serves the flexibility interests of employees, as it gives them more control over when and where to initiate and respond to work communications (Golden and Geisler 2007; Kossek 2016). Control over working time and place has been discussed as a resource for better aligning the demands of work and personal life (e.g. Chung und van der Lippe 2020; Kossek 2016). Empirical findings to date provide mixed evidence suggesting that flexible work arrangements, such as flexitime or flexiplace (for overviews, see Allen et al. 2015; Chung und van der Lippe 2020) and ICTs (Chesley 2014; Mazmanian et al. 2013; Piszczek 2017; Höge et al. 2016), can be both a resource and a demand contributing to more or fewer conflicts between life domains.

In this article, we ask whether the social preparation and consequences for work-to-life conflict of frequent written digital work communication (WDC) differ across European countries, depending on the respective family and employment protection policies. Despite EU-wide efforts to address the work–life balance challenges faced by parents and caregivers by way of the Work-life Balance Directive (European Commission 2019; Stoilova et al. 2020), and a general trend towards deregulation in globalised labour markets (Breen 1997; Gephart 2002), the welfare-state regimes of European countries continue to differ (Esping-Andersen 1999). These regimes likely shape not only supervisors’ and their staff members’ interest in using written digital communication, but also their opportunities for using this form of work-related communication for their own flexibility interests.

Family policies in European countries continue to encourage different gender and care regimes that shape gendered life-courses and consequently also employees’ interest in using WDC as a resource for flexibly aligning demands of work and personal life (e.g. Abendroth et al. 2012; Pfau-Effinger 2005). In addition, family policies likely shape employees’ opportunities for realising their flexibility interests, as work-facilitating family policies have been found to impose normative and economic pressures on employers to support employees in their efforts to balance their work and personal lives (e.g. Chung 2019; den Dulk 2001; den Dulk et al. 2012).

The degree to which labour market policies in European countries promote workers’ employment and financial security through employment protection legislation also varies. Empirical and theoretical applications of the power resource approach (Esser und Olsen 2012; Korpi 2006) suggest that this theoretical framework may also be useful in explaining employers’ opportunities for and interest in using WDC to make employees continually available for work.

To explore our research question, we rely on data from Round 10 of the European Social Survey (ESS) fielded between September 2020 and August 2022. The ESS is a biennial, cross-national household survey of attitudes and behaviour. The Round 10 questionnaire included a rotating module on “Digital Social Contacts in Work and Family Life” assessing, inter alia, the frequency of employees’ work-related communication with supervisors in person, via telephone, on screen and in writing (i.e. via text, email or messaging apps). Other questions related to respondents’ time- and strain-based work-to-life conflict. In our contribution, we focus on the frequency of written digital communication with supervisors via text, email or messaging apps, as this form of communication can be undertaken irrespective of place and time. Supplementing the ESS data with country-specific information on state policies provided, for example, by the Organisation for Economic Co-operation and Development (OECD 2023a, b, c, d) further allows us to investigate whether the frequency of work-related WDC with supervisors and its implications for work-to-life conflict depend on family and employment protection policies. Moreover, investigating the work-to-life-conflict-increasing or -reducing implications of frequent WDC with supervisors allows us, in addition, to establish whether they persist irrespective of whether WDC is embedded in different institutional contexts and thus subject to different social preparation. Previous comparative research on flexibility in digitalised working worlds has focused mainly on the implications of family policies for the availability, use and consequences of flexible work arrangements such as flexitime and flexiplace (Chung 2019; Chung and van der Horst 2018; den Dulk et al. 2012) or on the implications of labour market policies for new forms of digitally enabled flexible work such as crowd work, i.e., “paid work that is organised by an online platform” (Huws et al. 2016, p. 2; see also Krzywdzinski and Gerber 2020).

2 Theory

2.1 Digital Work Communication: Opportunities for the Realisation of the Flexibility Interests of Supervisors and Their Staff

Digital communication devices provide opportunities to initiate and respond to work-related communications more flexibly in time and space: for example, via text, email or messaging apps. These opportunities for flexible connectivity can serve the increasing flexibility interests of both supervisors and their staff.

By flexibility interests of supervisors, we mean their interest in making their staff more available for work even outside shared work locations and regular working hours. These flexibility interests have been attributed to the dynamics of globalisation and market volatility (Breen 1997; Wajcman 2015). They align with organisational norms of high work dedication, which include the expectation that employees should respond to high work intensity by working at home in the evenings or on the weekend (van der Lippe and Lippényi 2020; Williams et al. 2013). Existing research on flexitime and flexiplace arrangements indicates that they are indeed often used to get more work done (Schieman and Young 2010).

Thus, in line with norms of high work dedication, and in view of more unpredictable work demands due to globalised markets, employers may use WDC to make employees more available for work-related communication even outside shared work locations and regular working hours. In this connection Wajcman (2015, p. 941) pointed to the phenomenon of “constant connectivity”, whereby employees constantly check for incoming messages and quickly respond to them even at home and outside regular working hours. The term “electronic leash” (Arnold 2003, p. 243; see also Duxbury et al. 2014) further suggests that digital communication may serve as a new and more invasive form of control, thereby diminishing their autonomy (Mazmanian et al. 2013, p. 1345). The realisation of the flexibility interests of employers implies that work increasingly invades employees’ private sphere due to work intensification and reduced agency in the daily organisation of work, thereby increasing work-to-life conflict. The term work-to-life conflict refers to an inter-role conflict whereby demands in the domains of paid work and personal life interfere with each other (Greenhaus and Beutell 1985). Time- and strain-based work-to-life conflicts, which are the focus of the present study, occur when time constraints and strain in the work domain prevent workers from meeting their own expectations in the personal life domain (Adams and Golsch, 2021).

By flexibility interests of employees, we mean their interest in having more control over the time and location of their work in order to be able to flexibly balance work and personal commitments. This aligns with a resource perspective on flexible working. Work/family border theory (Clark 2000) and boundary management theory (Kossek et al. 2006) posit that having control over one’s schedule allows for flexible adaptation of the timing of work demands to personal commitments (see Chung 2019). In line with this perspective, digital work communication could give employees more control over when and where to initiate and respond to work communication with supervisors, in order to better align it with or flexibly adjust it to other personal commitments. This further implies less work-to-life conflict when digital work communication is used to flexibly connect with supervisors.

In the following sub-sections, we develop the argument that work-facilitating family policies and employment protection likely shape employees’ and employers’ interest in and opportunities for using digital work communication to realise their flexibility interests, with subsequent implications for (a) the extent to which employees communicate in writing with their supervisors via text, email or messaging apps, and (b) whether this means more or less work-to-life conflict.

2.2 Importance of Family Policies

Typologies that distinguish between different gender regimes imply that family policies facilitate different family models, such as the dual-earner/dual-caregiver, male-breadwinner, or one-and-a-half earner model (e.g. Lewis 2006; Sainsbury 1994). Policies, such as investments in (childcare) services and in-kind benefits for families, which foster female employment and the dual-earner/dual-caregiver model, have been described as “work-facilitating” (Chung 2019) or “de-familialising” policies (Esping-Andersen 1999). Previous research has shown that these policies do indeed increase female labour market participation and women’s opportunities to work more hours (Abendroth et al. 2012; Gornick und Meyers 2005). This implies that, as challenges in combining work and family increase, work-facilitating family policies contribute to employees’ interest in using digital work communication to flexibly align work-related communication with their supervisors and personal commitments.

In addition, rational choice and neo-institutional theories have been used to argue that work-facilitating family policies increase the economic and normative pressures on employers to invest in workplace arrangements that enable employees to better integrate their work and personal lives (e.g. den Dulk 2001; den Dulk et al. 2012). Following den Dulk (2001), neo-institutional theory suggests that the expectations of, for example, policymakers, the public, and the media create normative pressures on work organisations to offer family-friendly workplace arrangements, and these organisations may respond to these expectations in order to gain or secure legitimacy in their organisational environment. The application of rational choice theory and business case arguments (den Dulk 2001) refer to the economic pressures on work organisations to be more family-friendly, which may be a result of a more diverse workforce, with employees who increasingly need and want to align work and family demands. Here, employer investments in work–family-supportive workplace arrangements are an investment both in the employability of their workforce and in their own attractiveness as an employer (den Dulk 2001). In line with this, research based on employer survey data has shown the importance of family-friendly state policies for the availability of family-friendly workplace arrangements and for variation in the use of flexible workplace arrangements across countries (e.g. den Dulk 2001; den Dulk et al. 2012; van der Lippe and Lippényi 2020). This encouraged responsiveness to flexibility interests of employees further implies more opportunities for employees to use WDC to align work-related communication with their supervisors not only with their overall work schedules but also with their personal commitments—for example, by working from home or by flexibly adjusting starting and finishing times of work, thereby decreasing the opportunities for in-person communication. Following from this, we hypothesise:

H1A: The more the state invests in work-facilitating family policies, the more frequent work-related written digital communication with supervisors is.

H1B: Work-facilitating family policies increase the likelihood that frequent work-related written digital communication with supervisors will reduce work-to-life conflict.

2.3 Importance of Labour Market Policies

The power resource approach emphasizes that country specific institutions in liberal and coordinated economies shape power relations and conflicts in the exchange relationship between employers and employees in different ways (Esser und Olsen 2012; Korpi 2006). Applying this perspective to the question of the importance of labour market policies for negotiating digital work communication practices on the micro-level, it can be assumed that labour market policies likely shape employees’ and supervisors’ power to use work-related digital communication in their flexibility interest and consequently also the implications of frequent WDC with supervisors for employees’ work-to-life conflict.

More specifically, the power resources approach implies that weak employment protection policies enhance employers’ opportunities to realise their flexibility interests with the help of WDC. Weak employment protection means less bargaining power for employees and consequently more responsiveness to supervisors’ expectations that they will immediately respond to digital work communication irrespective of time and place. In the same vein, Lindbeck und Snower (2001, p. 168) pointed to the importance of labour market institutions for incumbent workers’ (i.e. “insiders’”) bargaining power, stating that “labor turnover costs are the ultimate source of insiders’ market power”. Rubery und Grimshaw (2001) further stressed the relevance of workers’ bargaining power in shaping the impact of ICT use on job quality.

On the other hand, a more balanced realisation of flexibility interests is likely when pronounced employment protection policies increase employees’ security and, as a consequence, their bargaining power. The more protected workers are against dismissal, the costlier their dismissal is. Therefore, employers might feel more inclined to accommodate workers’ flexibility interests, at least to some extent. Otherwise they risk the “crowding out of work effort” (Frey 1993, p. 664) due to constant connectivity. Crowding out refers here to reduced work effort when employees experience digital work communication as an indication of distrust and/or as unfair, which is especially likely when it serves—or is perceived—as an “electronic leash” (Arnold 2003, p. 243). Some research studies have shown that employment protection policies are indeed associated with greater job quality (Adăscăliței et al. 2022; Arranz et al. 2019; van der Wiel 2010). Consequently, employers and employees are more likely to coordinate their flexibility interests in line with the “gift exchange dynamic” described by Chung (2019, p. 25), whereby employers give employees opportunities to work more flexibly in time and place, but expect and receive responsiveness to flexible work demands in return. This would imply more frequent digital communication between supervisors and employees, not only because employees can work more flexibly in time and place during the regular workday to accommodate demands of work and personal life, but also because they are more responsive to flexible work demands (e.g. digital work communications from their supervisors). Moreover, when employment protection policies are pronounced, frequent digital work communication with supervisors is less likely to increase work-to-life conflict as a more balanced realization of flexibility interests is entailed. Following from the theoretical considerations and empirical evidence, we hypothesise:

H2A: The more the state invests in employment protection, the more frequent work-related written digital communication with supervisors is.

H2B: Employment protection policies decrease the likelihood that frequent written digital work communication with supervisors will increase work-to-life conflict.

3 Data and Methods

3.1 The European Social Survey

Our analyses are based on data from Round 10 of the European Social Survey (ESS) fielded between September 2020 and August 2022. To date, data have been released for 25 of the 32 participating countries. The Round 10 questionnaire included for the first time a module exploring the topic of digital social contacts in work and family life, which is central to the research topic addressed in the present study.

Our analytical sample comprised respondents aged 17–65 who were in paid employment (excluding self-employment or working in one’s own family business) in the 7 days preceding the survey, who worked under a supervisor, and who had no missing values for the variables included in the analyses. The latter requirement led to the exclusion of some participating countries (Bulgaria, Cyprus, Croatia, Iceland, North Macedonia, Montenegro, Serbia) due to missing information on the macro-level (either on family policies or employment protection). After listwise deletion (for a description of observations lost to missing values see Table A1), our final sample comprised 11,699 individuals from 19 countries.

3.2 Measures

3.2.1 Dependent Variables

Written digital communication (with supervisors about work)

To measure WDC with supervisors, we used responses to the question: “How often do you communicate about work in writing with your line manager via text, email or messaging apps?”, which was answered on a 7-point scale ranging from “never” (0), through “several times a month” (3) to “several times a day” (6). The average rating was 2.7, suggesting that, on average, respondents engaged in WDC with their supervisors several times a month (for a description of the sample and the average distribution of variables, see Appendix Table A1).

Work-to-life conflict (WLC)

Respondents’ work-to-life conflict was measured based on their subjective evaluation of (1) how often they felt too tired after work to enjoy the things they would like to do at home, (2) the extent to which their job prevented them from giving the time they want to their partner or family, and (3) whether their partner or family got fed up with the pressure of the respondent’s job. Respondents rated each of these questions on a 5-point scale ranging from 1 (Never) to 5 (Always). At 0.77, the size of the alpha coefficient for the variables was adequate to allow the indicators to be combined to form an average index for work-to-life (sum divided by number of items) conflict. For respondents who did not answer question 3 because they had no partner the index is based on only the first two items.

3.2.2 Individual-Level Control Variables

Besides general demographics (age, gender, living with partner) and family structure (number of children in household), we controlled for several job and organisational characteristics that likely influence workers’ access to and use of asynchronous work communication (e.g. contractual working hours, supervisory responsibility, type of contract). Moreover, to consider that job tasks also shape the opportunities for WDC, we integrated a measure on the characteristic of main job tasks differentiating between (1) Analytical non-routine tasks, (2) Interactive non-routine tasks, (3) Cognitive routine tasks, (4) Manual routine tasks, and (5) Manual non-routine following the definition of Dengler et al. (2014; for a description of the distribution of individual-level variables, see Appendix Table A1).

The COVID-19 pandemic

The survey was fielded between September 2020 and August 2022. Depending on the timing of the survey and the country-specific COVID-19 regulations, respondents’ may have either increased or decreased their in-person contact with supervisors and co-workers. Therefore, respondents were asked to rate the frequency of their online or mobile communication with people they worked with compared to before the pandemic. The response options were: “much more often now”, “a little more often now”, “about the same”, “a little less often now” or “much less often now”, and “I don’t have online or mobile communication with the people I work with”. We condensed the information into a dichotomous variable indicating whether online/mobile communication increased (1) or decreased/stayed the same (0). The latter category also included those who did not have online or mobile communication with the people they worked with.

3.2.3 Macro-Level Variables

Work-facilitating family policies

We used public social expenditure on services and in-kind benefits for families as a percentage of gross national product (GDP) as a proxy for work-facilitating policies (see also Chung 2019). This information is provided by the OECD for the year 2017 (OECD 2023a; for a description of the macro-level indicators by country, see Appendix Table A2).

Employment protection

Following Esser and Olsen (2012), both employment protection legislation and wage replacement payments for workers when unemployed protect workers against labour market insecurities. This information is provided by the OECD (OECD 2023b, c, d). The indicator for employment protection legislation allows us to differentiate between employment protection for regular contracts (EPRC) and employment protection for temporary contracts (EPTC). EPRC is concerned with the costs for employers of firing permanent workers, whereas EPTC captures the regulations for hiring temporary workers. Wage replacement payments, which are part of passive labour market programmes (PLMPs), were captured with the percentage of a country’s gross domestic product (GDP) expended on transfer payments to workers in case of unemployment (considering only OECD Main Category 80—“Out-of-work income maintenance and support”; OECD 2023d). For the UK the last available information dates back to 2011. In order to the keep the country in the analytical sample we used 2011 information for this country.

Macro-level control variables

To control for general differences in digital connectivity across countries, we relied on individual responses to the ESS Round 10 question as to whether respondents could access the Internet (a) from home, (b) at the workplace, and/or (c) on the move. After generating a sum index on the individual level, we aggregated this index on the macro-level (for a description of the macro-level indicators, see Appendix Table A2). We also controlled for countries’ GDP (OECD 2023e) in 2020 (measured in US dollars per head, current prices, current exchange rate) to account for economic cycles. For example, economic growth could increase organizations resources for investing in digital infrastructures, increasing overall WDC use. Alternatively, economic growth positively impacts employment rates (Ferreiro & Gomez 2019), which likely reduces pressures for employers and employees and could result in less experience of work-life conflicts, regardless of WDC use.

Work demands and resources

Finally, we considered work demands and resources that potentially mediate the implications of work-facilitating family policies and employment protection for (a) the degree of permeation of written digital work communication, and (b) the implications of WDC with supervisors for work-to-life conflict. Work demands were captured via respondents’ weekly overtime hours (top-coded at 40 additional hours), how often they were expected to work overtime at the workplace or at home, and whether they were expected to be responsive to work communications outside regular working hours. The latter two variables were rated on a 6-point scale ranging from 1 (Never) to 6 (Every day). Work resources were captured by (a) working-time flexibility rated on a 3-point scale ranging from 1 (Not at all) to 3 (Completely); (b) frequency of working from home, measured on a 6-point scale ranging from 1 (Never) to 6 (Every day); and (c) autonomy, measured with a question about the extent to which management permitted the respondent to decide how their daily work was organised, rated on an 11-point scale ranging from 0 (No control) to 10 (Absolute control).

4 Methods

Taking into account the clustered structure of the ESS data, we applied multilevel modelling to test our hypotheses. By adding macro-level indicators to the analyses, we could determine whether and to what extent specific macro-level characteristics (such as family and labour market policies) impacted individual-level use of digital technologies and work-to-life conflict.

First, we estimated a set of models on the frequency of WDC with supervisors; variables were added step-wise to the models. Our final model specification is depicted in Equation (1). WDCij refers to the frequency of written digital communication with supervisors, CVi represents the individual-level control variables. Work facilitation (work-facilitating family policies), EPRC (employment protection, regular workers), EPTC (employment protection, temporary workers), and PLMP (out-of-work income maintenance) represent the macro-level indicators, with CVj as macro-level control variables; \({\mu }_{j} { {\text{and}} \varepsilon }_{ij}\) represent the residuals on the individual and country levels.

$$\begin{aligned} WDC_{{ij}} & = ~\beta _{{0ij}} + ~\beta _{1} *CV_{i} + \beta _{2} *Work~facilitation_{j} + ~\beta _{3} *EPRC_{{j~}} \\ & \quad + \beta _{4} *EPTC_{{j~}} + ~\beta _{5} *PLMP_{{j~}} + ~\beta _{6} *CV_{{j~}} + ~\mu _{j} ~ + ~\varepsilon _{{ij}} \\ \end{aligned}$$
(1)

The same general procedure was applied to the second set of models on work-to-life conflict. Because we were interested in cross-national variation in whether WDC increased or decreased respondents’ work-to-life conflict, we integrated cross-level interactions between the individual frequency of engaging in WDC with supervisors and the macro-level characteristics. The model specification is shown in Equation (2).

$$\begin{aligned} WLC & = \beta _{{0ij}} + ~\beta _{1} *WDC_{i} ~ + \beta _{2} *CV_{i} ~ + \beta _{3} *(WDC_{i} *Work~facilitation_{j} ) \\ & \quad + \beta _{4} *(WDC_{i} *EPRC_{{j~}} ) + ~\beta _{5} *(WDC_{i} *EPTC_{{j~}} ) \\ & \quad + ~\beta _{6} *(WDC_{i} *PLMP_{{j~}} ) + \beta _{7} *CV_{j} + \mu _{j} ~ + ~\varepsilon _{{ij}} \\ \end{aligned}$$
(2)

5 Results

5.1 Permeation of Written Digital Communication with Supervisors

First, we examined the degree of permeation of workers’ WDC with their supervisors across countries. As shown in Fig. 1, there was substantial cross-national variation in the general use of WDC: the proportion of respondents who stated that they never engaged in WDC with their supervisors ranged from 15% in Israel to 61% in Greece (weighted averages).

Fig. 1
figure 1

Permeation and Frequency of Written Digital Communication with Supervisors. Source: ESS Round 10, own calculations, analyses weighted. N = 11,699

Full size image

Table 1 shows the results of our multilevel models on the frequency of WDC with supervisors. Model 1 is the empty model, which was used only for reference to calculate explained variances on each level. Integrating individual-level controls in Model 2 revealed that WDC with supervisors was positively associated with other forms of communications in which respondents and supervisors do not need to be in the same location (i.e. speaking on the phone and screen communication).

Table 1 Multilevel Models on the Permeation of Written Digital Communication with Supervisors—Including Macro-Level Indicators
Full size table

Turning to the macro-level indicators—our main interest in this study—Model 3 shows that work-facilitating policies were associated with more frequent WDC with supervisors (0.307, p < 0.05). This finding supports hypothesis H1A, which stated that the more the state invests in work-facilitating family policies, the more frequent work-related digital communication in writing with supervisors is. Model 3 further shows that higher public expenditure on out-of-work income maintenance and support as a percentage of GDP was associated with more frequent WDC with supervisors (0.182, p < 0.05). However, no significant association was found for employment protection policies. Thus, hypothesis H2A is supported only partly, as it stated that the more the state invests in employment protection, the more frequent WDC with supervisors is. Adding the macro-level indicators to Model 3 explained about 51,5% of the country-level variation in the degree of permeation of WDC with supervisors.

In Model 4, we additionally considered the mediating effects of work demands and resources. First, the expectation to be responsive to work communications outside working hours increased the frequency of WDC with supervisors. However, there was no statistically significant association between overtime hours per week or expectation to work long hours (the other indicators for the realisation of employers’ flexibility interests) and the frequency of WDC with supervisors. Second, except for autonomy, all work-resources indicators were associated with greater frequency of WDC with supervisors. Hence, it seems that WDC with supervisors is mainly part of work practices aimed at supporting employees in their work-life balance and autonomy, and only to a small degree at realising employers’ interest in greater availability on the part of their employees. Overall, the effect of work-facilitating policies remained robust after the integration of the mediator variables.

5.2 Work-to-Life Conflict and Written Digital Communication with Supervisors

We applied multilevel modelling also to examine whether work-facilitating family policies and employment protection policies moderated the implications of WDC with supervisors for work-to-life conflict (see Table 2).

Table 2 Multilevel Regression Models on Work–Family Conflict
Full size table

Again, Model 1 (Table 2) was used only for reference to calculate explained variances on each level. The addition of employee characteristics to Model 2 revealed that more frequent WDC with supervisors was associated with more work-to-life conflict (0.024; p < 0.01). By contrast, in-person communication with supervisors was associated with less conflict (−0,020; p < 0.01).

To examine how country-level policies mitigate the work-to-life-conflict-increasing nature of frequent WDC with supervisors, we included cross-level interactions between this type of communication and the macro-indicators in the models (see Table 2). First, we added the indicator for work-facilitating family-policies to the model (Model 3). The statistically insignificant interaction effect shows that these policies did not moderate the implications of WDC with supervisors for work-to-life conflict, despite their statistical significance for the frequency of WDC. Thus, no evidence is provided for H1B, which stated that work-facilitating family policies increase the likelihood that frequent digital work communication with supervisors would reduce work-to-life conflict.

Second, we added employment protection for workers with regular and temporary contracts (EPRC and EPTC) to Model 4 (Table 2). The results reveal that—for workers with regular contracts—having greater protection against layoffs was associated with less work-to-life conflict when WDC with supervisors was engaged in more frequently (Model 4; −0.020; p < 0.01). Hence, protecting permanent workers against dismissal can shield them from the otherwise work-to-life-conflict-increasing nature of WDC with supervisors. We did not find a similar result for either employment protection of temporary contracts or public expenditure on PLMPs (here: out-of-work income maintenance) as a percentage of GDP. Thus, only partial support is provided for hypothesis H2B, which stated that employment protection polices decrease the likelihood that frequent WDC with supervisors would increase work-to-life conflict. Because the PLMP indicator was only available for 2011 for the UK, we re-estimated our models excluding the UK (see Appendix Models A5 and A6). Results change only marginally when the UK is not included in models.

Model 5 (Table 2) shows that the negative interaction effect between employment protection and WDC with supervisors was resilient to the integration of work-facilitating policies and employment protection. Finally, we added work resources and demands as mediator variables to Model 6. All indicators point in expected directions: whereas work demands increased work-to-life conflict, work resources reduced it. In line with only marginal changes on the country level, mediating work demands and resources did not add to the percentage of explained country-level variance in work-to-life conflict explained; however, an additional 10% of the individual-level variance in work-to-life conflict was explained when these demands and resources were added to the model.

6 Discussion

Written digital communication increases the availability of employees and supervisors for work-related communication irrespective of time and place. This may serve the flexibility interests of employees and/or supervisors. Applying a cross-national comparative perspective, we examined whether the social preparation and implications for work-to-life conflict of frequent WDC with supervisors differ across European countries, depending on existing family and labour market policies.

We conclude, first, that the degree of permeation of frequent WDC with supervisors in the 19 European countries included in our analyses depends on family policies that support parental employment. These were found to be associated with more frequent WDC with supervisors. This is in line with the argument that work-facilitating state policies increase not only employees’ interest in flexible working, due to increased challenges in balancing work and family life, but also normative and economic pressures on employers to be responsive to these flexibility interests (den Dulk 2001; den Dulk et al. 2012).

In addition, WDC with supervisors was found to be more frequent in countries with a high level of EPRC and a high level of public expenditure on PLMPs (here out-of-work income maintenance). This is in line with arguments based on the power resources approach (Esser and Olsen 2012; Korpi 2006), suggesting that employment protection enhances employees’ power to negotiate their flexibility interests. Overall, these findings suggest that family and labour market policies promote employees’ flexibility interests and supervisors’ responsiveness to these interests, thereby contributing to the permeation of WDC. Furthermore, we conclude that the negative implications of frequent WDC with supervisors for work-to-life conflict persist despite the aforementioned social preparation of WDC by work-facilitating family policies and employment protection policies. Although the impact of WDC with supervisors on work-to-life conflict remained moderate, this finding supports the argument that work–life boundaries are likely to become blurred when work and personal life are not clearly separated but rather share similar times and locations (Clark 2000) or that employees are more likely to exploit themselves in more flexible work environments (Chung 2022). An additional explanation is provided by gift exchange dynamics where employees reciprocate for the gift of flexibility (Chung 2017), making themselves more available for work. Nevertheless, employment protection policies appear to limit this dynamic, as EPRC was found to reduce the negative implications of frequent WDC with supervisors for work-to-life conflict. Again, this points to the fact that employment protection policies restrict employers’ power to impose their flexibility interests with the help of WDC, which is in line with the power resources approach (Esser and Olsen 2012; Korpi 2006). Moreover, this finding suggests that self-perpetuating tendencies of increased work-life conflicts due to the permeation of WDC can partly be mitigated through political regulation.

Our study has some limitations. First, the cross-sectional design does not allow us to draw causal inferences. Moreover, further research is needed to provide an even more differentiated picture of the implications of frequent WDC with supervisors for different groups of workers (e.g. distinguishing between job tasks and qualifications, types of contract, occupational status or gender). Nevertheless, the present research points to the importance of studying the effects of the interplay of digital infrastructures and social policies on the dissemination of flexible working worlds and the subsequent transformation of the work–life interface.