Demographic projections unanimously suggest that the proportions of both elderly and long-term care recipients among the population will rise considerably in the foreseeable future. Informal care plays a vital role in meeting this increased demand for adult care. Growing female labour market participation rates and the political goal to raise pension ages however increase the potential for work–family conflict for traditional suppliers of informal care (Colombo et al. 2011).

Caregivers’ work–family conflict reduces their quality of life and impairs their career prospects. Costs may also accrue for the employer, as via family-to-work conflict, caregivers could be particularly prone to (partial) absenteeism and presenteeism (Hoskins 1993, p. 359; Smith 2004, p. 370), reduced productivity (Kossek and Ozeki 1999) and turnover. Finally, work–family conflict also entails impacts on the community, as incompatibility between work and family obligations will hamper engagement in informal care and thus raise utilization of formal care services (Fast et al. 1999). Consequently, investigating the interactions between informal care obligations and paid work and finding strategies to promote work–life balance is of utmost importance.

Two prominent strands of literature have discussed the interactions of caregiving and work-related outcomes. Role conflict literature addresses incompatibilities between demands originating from work and family roles (Greenhaus and Beutell 1985). Time allocation theory explains absences from work via utility-maximization of workers subject to environmental conditions, restraints and preferences (Allen 1981; Barmby 2002; Barmby et al. 1991; Dionne and Dostie 2007; Drago and Wooden 1992). While the situation of caregivers has been considered in both strands of literature (e.g., Barling et al. 1994; Johnson and Lo Sasso 2000; Kossek et al. 2001), little evidence exists on the complex interactions of workplace characteristics and adult care responsibilities.

This article adds to the literature by drawing from both theoretical approaches in order to investigate the effects of workplace characteristics on caregivers’ time-based work–family conflict. More specifically, our study uses the fourth European Working Conditions Survey (EWCS, Parent-Thirion et al. 2007) to analyse the level of perceived time-related work–family conflict and absenteeism as an outcome of family-to-work conflict. We test whether workers with informal adult care responsibilities experience higher levels of work–family conflict and absenteeism. We investigate whether the determinants of work–family conflict differ in contrast to non-caregivers, and what effect these determinants have on caregivers’ absenteeism. The EWCS’s detailed variables allow for a thorough consideration of workplace policiesFootnote 1 and work environment.

Our investigation of outcomes of work–family conflict is the first to use a well-established international and representative dataset. Our findings can serve to determine which measures best promote work–family balance for informal caregivers to adult dependents, which is of relevance to employers as well as policy makers.

Theoretical Background, Previous Studies and Research Hypotheses

Role Conflict Theory

Role conflict occurs if participation in one role makes it difficult to meet demands of another role. The source of this incompatibility may be the result of strain, time or behavioural requirements associateed with a role (Greenhaus and Beutell 1985). Time-related work–family conflict occurs if two roles, one located in the family sphere and the other in the sphere of paid work, compete for an individual’s time (a notion similar to the time allocation model). Adopting a role as caregiver to an adult dependent in the family context could hence conflict with the role fulfilled as an employee or co-worker.

The demands-and-resources approach (Voydanoff 2005a) sees individuals as juggling demands (expectations or structural or psychological claims associated with being engaged in a specific role) and resources (i.e., assets used to meet demands). Depending on whether resources are seen as adequate, the cognitive appraisal of the work–life fit can be positive or negative.

Empirical studies on work–family conflict have found determinants and outcomes of role conflict to depend on the directional nature of the conflict: While work-to-family conflict is determined mainly by work stressors and mainly affects family outcomes, family-to-work conflict depends mainly on family stressors and affects work outcomes (Frone et al. 1997). Consequently, research has found work domain variables to explain life satisfaction and non-work domain variables to explain work outcomes (e.g., Kossek and Ozeki 1998; Mauno and Rantanen 2012). However, recent research has identified direct and indirect effects of work domain variables on family-to-work conflict. For example, if family demands are extensive, it is more likely that work demands are seen as barriers to fulfilling them resulting in a positive relationship between the directional dimensions (Huang et al. 2004; Voydanoff 2005b).

Previous work also underlines the importance of the type of work demands and other workplace factors. While work in the private sector (Buelens and Van den Broeck 2007), work hours and, more importantly, work overload (Skinner and Pocock 2008), add to work-to-family conflict, flexible work time regimes (Barling et al. 1994) and climates for sharing concerns (Kossek et al. 2001) reduce caregivers’ role conflict. Any measures organizations take to reduce work–family conflict may pay off through decreased absenteeism and withdrawal behaviour (Hammer et al. 2003).

Time Allocation Theory and Rational Absenteeism

The standard model of time allocation features utility-maximizing individuals which are subject to budgetary and time constraints. Individuals maximize utility from income, leisure and—in the case of altruistic caregivers—from the care recipients’ well-being or health. In the optimum, time is allocated to work, leisure or home production activities in a way that equalizes the marginal utilities derived from these three activities. Extended to cover informal adult-care, the model hence posits that working caregivers balance the net marginal utility derived from paid work (which equals the difference between the wage rate and the marginal discomfort of labour), the marginal utility of leisure and the marginal utility gained from improving the care dependent’s well-being in allocating their time (Johnson and Lo Sasso 2000).

Time allocation theory provides an explanation for absences from work, which can be considered an adjustment mechanism in cases where contracts lack flexibility. In this perspective, imperfect and rigid labour markets however prevent workers from contracting exactly the number of work hours that maximizes utility. Therefore, absenteeism is the easiest option to adjust labour supply on a short-run, needs-based basis (Allen 1981). Furthermore, absenteeism is determined by the valuation of non-monetary utility of work and the fraction of remuneration forgone in the case of absences.

The potential loss in income in case of absenteeism is related to its sanctioning by employers. As absenteeism is costly to employers, firms set incentives which discourage absenteeism. Options to punish absenteeism include wage cuts, delaying promotions, denying otherwise granted rents or termination of the work contract (Drago and Wooden 1992). These incentives or disincentives will enter workers’ utility maximization calculus and could work to reduce voluntary (rational) absenteeism. They could also lead to self-selection of absenteeism-prone workers into absenteeism-tolerant jobs that are paid less well (Allen 1981, p. 80).

Especially for caregivers to the elderly, the key assumptions of this microeconomic model, full information and freedom of choice (Spieß and Schneider 2003), may not hold true. Besides planned absences, for example when accompanying care recipients to ambulatory doctor visits and participation in treatment, they often have no choice but to miss work in the unplanned event of an acute deterioration of the care recipient’s health status or other emergencies such as a fall (Arksey 2002, p. 154). These “involuntary” absences are beyond the control of the worker and might not be affected by workplace characteristics (Driver and Watson 1989; Hackett and Guion 1985).

Hypotheses

As caregivers face higher time demands in their family roles, work obligations are more likely to interfere with these and the opportunity costs of time rise, lowering the optimal level of labour supply. We thus hypothesize that there are higher levels of perceived work–family conflict (Hypothesis 1a) and absenteeism (Hypothesis 1b) among the subsample of employees with informal adult-care responsibilities. Furthermore, we hypothesize that conciliating factors such as positive work environments and flexible work time regimes are associated with reduced levels of caregivers’ perceived level of conflict (Hypothesis 2). Regarding possible differences in determinants for work–family conflict, role conflict theory suggests that caregivers to adults feature increased family demands but no fundamentally different relationship between work and family responsibilities. Thus, in this regard, our analysis is explorative.

In addition, we are interested in what effect measures that influence perceived work–family conflict have on absenteeism, an outcome of family-to-work conflict. In some cases, work–family conflict will result in absenteeism. In other cases, the feasibility to miss work in case of an emergency might constitute a resource that helps caregivers to meet family obligations, thus lowering the level of perceived conflict via decreased work-to-family interference. Regardless of whether such an option is part of the work contract or not, this could result in a trade-off between absenteeism and perceived role conflict via the degree to which the work environment allows for/tolerates absences, allowing the worker to better match labour supply to the preferred level. More precisely, we hypothesize that family domain variables, such as caring for an adult or having children, affect absenteeism and the perceived level of work–family conflict in the same direction (Hypothesis 3a), whereas variables capturing control over work schedules, a work domain resource, will raise absenteeism while lowering perceived levels of conflict (Hypothesis 3b).

Empirical Model and Data

Data

This study uses data from the Fourth European Working Conditions Survey (EWCS) carried out in 2005 (Parent-Thirion et al. 2007). It contains data from the 27 EU-countries plus Switzerland and Norway as well as the EU-candidate countries Croatia and Turkey. About 1,000 people were interviewed per country with the exception of Cyprus, Estonia, Luxembourg, Malta and Slovenia, where only about 600 entries exist.

The EWCS’s main focuses lie on the characteristics of the interviewees’ job, such as the sector it belongs to and the employment status (self-employed, full-time or part-time employed). Numerous job characteristics, such as machinery usage, work time regimes, etc. are thoroughly assessed. The survey puts particular emphasis on interpersonal relations at the workplace and attitudes of the workers towards their job. Additionally, the EWCS includes items that address work–family balance and outside-work commitments. This rich diversity of workplace-related variables is the main advantage of using the EWCS for studying working informal caregivers.

Missing values were imputed using the STATA module ›mi ice‹, which implements multiple imputation using chained equations (Royston 2004).Footnote 2 In cases of missing values for age, sex or caregiver status, the observation was dropped. Out of 29,680 entries 24,526 remained in the study sample; about 15 % provided informal care to an older person or disabled adult on at least weekly basis.

Measures

Informal caregivers to adult dependents, the comparison group of interest, were identified as those respondents who answered the question “How often are you involved in … [c]aring for elderly/disabled relatives?” with “Once or twice a week” or more often. This threshold represents a compromise between the attempt to include only workers whose care obligation is extensive enough that it could cause work–family imbalances and the effort to keep caregivers’ sample size per country high enough. The subsample includes all kinds of informal caregivers to adults irrespective of coping strategies employed. While this may not allow us to draw conclusions for the total population of caregivers irrespective of employment status, this is the representative sample of employed caregivers needed for assessing labour market outcomes and performing comparisons with workers who do not provide informal care.

Perceived work–family conflict, the main dependant variable of interest, is covered in the EWCS via the question “In general, do your working hours fit in with your family or social commitments outside work very well, well, not very well or not at all well?” This may capture time-based work-to-family and family-to-work conflict.Footnote 3

As regards the second dependant variable of interest, absenteeism, the EWCS asks whether respondents missed work during the last 12 months for various reasons.Footnote 4 This represents absenteeism occurrences in contrast to absenteeism duration. For the purpose of our study, we focused on absences due to family and health reasons, as these two categories are most likely to capture absences due to caregiving.Footnote 5

The EWCS offers a host of variables regarding job characteristics and job assessment, our explanatory variables of interest. The variables used in our analyses capture formal characteristics (contract type, work time regulations, work hours and days, tenure, etc.), informal characteristics (i.e., what workers do at work, such as whether they rotate tasks, have the possibility to take days off, etc.), interpersonal relations at the workplace (i.e., to what degree respondents feel at home at work, regard their colleagues as friends or can count on help if needed) and job assessment (i.e., whether respondents say that their job makes them sick, and if so, the number of health impairments mentioned and whether respondents are content with their work time arrangements).

Controls include sex, age and household composition (i.e., number of children and presence of spouse).Footnote 6 One of those variables, presence of a spouse in the same household, is of particular relevance, as it could constitute a family resource for coping with role conflict.

Estimation Method

Due to the nature of the dependent variables, we performed ordered logistic regression to examine the level of perceived work–family conflict and logistic regression to examine absenteeism. In a first step, we estimated the model of perceived work–family conflict and the absenteeism model based on data for the full sample. In this specification, the value of the adult caregiver identification dummy’s coefficient tells whether caregivers to adults differ from the full sample. Then we computed the same model on the subsample of carers. This gave first clues regarding potential differences in coefficients between workers with/without caregiving responsibilities. Variables with nonsignificant coefficients in both the regression on the full sample and the carer subsamble were removed from the regression equations. For each coefficient that features non-overlapping 95 % confidence intervals in these two regressions, we tested for significance of the difference by computing a third regression that includes interactions of these variables with the group identifier dummy using the full sample. If these cross-terms have significant coefficients, the respective factors have a significantly different effect on caregivers’ absenteeism/work–family conflict when compared to the full sample. As a number factors, such as legislation on leave arrangements, sick leave pay, the extent of care services provided, monetary support for families and established cultures and role models concerning care vary across Europe (Bettio and Plantenga 2004; Brandt et al. 2009; Deindl and Brandt 2011; Haberkern and Szydlik 2008; Johansson and Palme 2002), we employed a random effects approach to estimate multilevel models with the country as group identifier.Footnote 7

Results

Graph 1 compares caregivers and non-carers with respect to time-based work–family conflict and absenteeism. In both subsamples, about 48 % of respondents said their working hours fit “well” with their outside work commitment, however only 28.6 % of caregivers as opposed to 32.2 % of non-caregivers answered with “very well” while 23.4 % of the caregivers and only 20.2 % of non-carers reported that working hours fit “not very well” or “not well at all.” The Kruskal–Wallis rank sum test rejects the hypothesis of equal distribution with high significance (p < 0.001). In the full sample, 24.3 % of the respondents have been absent for health reasons, 12.4 % have been absent for family reasons and 31.8 % have been absent for health or family reasons. Among the caregiving subsample, these numbers are significantly (p < 0.001) higher (28, 14.8 and 36.7 %).

Graph 1
figure 1

Work-to-family conflict and occurrence of absences among carers and non-carers

Full size image

The results of the regression on time-based work–family conflict were presented in Table 1. Caregivers are roughly 12 % more likely to report a higher degree of conflict than workers without adult care obligations, lending support to Hypothesis 1a. The number of hours and days worked, the number of days worked at nights or at weekends and the number of days with more than 10 h of work have a large effect on this outcome. Work time regimes (flexitime and the possibility to set work times at will) also play a central role in explaining perceived time-based work–family conflict: All of those variables significantly reduce perceived work–life conflict for adult carers and the full population alike.

Table 1 Time-based work–family conflict and absenteeism regressionsa
Full size table

The results of the regression analysis of absences are also presented in Table 1. Most coefficients of personal and formal job characteristics were significant and had the expected sign. As assumed with Hypothesis 1b, caregivers were 28 % more likely to have missed due to family reasons and 19 % more likely to report a family- or health-related absence. Interestingly, of the three variables capturing flexible working contracts, two had a positive effect on absenteeism probability (possibility to take days off and flexitime regulations), while being able to set one’s own work hours reduced absence probability. Likewise, the two variables describing interpersonal relations at the workplace had opposite effects: Having good friends at work raises absenteeism, feeling at home at work lowers it.

In our investigation of driving factors associated with absenteeism, we found that most factors explaining absenteeism were equally applicable for caregivers and non-caregivers. Differences emerge in the coefficient of the gender and spouse dummy variables, which were nonsignificant for caregivers to adults.

As assumed with Hypothesis 2, generous working conditions and positive work climates, were found to reduce work–family conflict and hence constitute important aspects of workplace policies. We did not formulate any hypotheses concerning differences between caregivers and non-caregivers in this regard. However, we found several such differences, which might shed some light on different interactions between some of the explanatory variables and work–family conflict for the caregiving subsample. Firstly, caregivers did not experience more work–family conflict if they were married. Secondly, they featured a somewhat larger effect of having good friends at work. Finally, there was a significantly larger effect of work overload on caregivers’ work–family conflict.

What effect do factors that reduce the perceived level of role conflict have on absenteeism (Hypothesis 3)? Firstly, family domain stressors raised both absenteeism and the perceived level of work–family conflict, as assumed with Hypothesis 3a. Having children, caring for a sick or disabled adult and being married are examples. While all variables capturing control over work time featured a negative coefficient in the work–family conflict regression, only some of them raised absenteeism, offering partial support for Hypothesis 3b.

Discussion

The aim of this study was to explore the effect of informal caregiving and a range of job-related variables on perceived level of work-to-family conflict. We were interested both in whether caregivers perceive, on average, a higher level of work–family conflict and absenteeism, and whether they differ significantly from non-caregivers regarding the factors that influence this outcome. Furthermore, we investigated what effect factors relevant to caregivers’ perceived work–family conflict have on absenteeism.

Time allocation models and role-conflict theories served as theoretical background and were used to generate three research hypotheses. We hypothesized (1) that workers with informal adult care commitments perceive higher levels of absenteeism and time-based work–family conflict, (2) that positive work environments and flexible work schedules reduce adult carer’s perceived work-to-family conflict, and (3) that contrary to family domain variables, which affect absences and perceived level of work–family conflict, in the same direction, work-domain resources, such as control over work times and the possibility to miss work, reduce the level of perceived conflict even if they are related with higher absenteeism.

Using data of the fourth European Working Conditions Survey (Parent-Thirion et al. 2007), we regressed the reported level of work–family conflict and absenteeism occurrence on a number of explanatory factors, including household composition and detailed job characteristics. Our findings lend at least partial support to all research hypotheses.

A higher level of work–family conflict and absenteeism among caregivers comes as no surprise. On the one hand, caregivers to adults are engaged in a time-demanding family role. This might not be the case for a proportion of the non-caring subsample. Generous working conditions were found to reduce work–family-conflict, but they do so to no larger extent for carers’ than for non-caregivers (cf., Pavalko and Henderson 2006). However, we found significant differences between the full sample and adult carers with regard to marital status (no effect on caregivers as opposed to non-carers), having good friends at work (a stronger negative effect on work–family conflict for carers), and suffering from work overload (a stronger positive effect on work–family conflict for carers). The first two variables might constitute resources that help caregivers balance work and family roles. Presence of a spouse indicates that a worker is engaged in both work and family roles, causing possible work–family conflict. Informal caregivers however are engaged in a time-extensive family role whether they are married or not. The presence of a spouse could be easing pressures and strain related to care-giving if the spouse shared caregiving responsibilities (Kossek et al. 2001). Likewise, having good friends at work is a prerequisite for using informal ways to combine work and care (Arksey 2002). Furthermore, having good friends at work helps caregivers building extended care networks and provides opportunities to foster non care-related social contacts, which provides respite from care obligations. Finally, having good friends at work could express what Kossek et al. (2001) refer to as “work climates for sharing concerns:” Having someone to talk to considerably reduces work-to-life conflict.

Finally, the significantly larger effect of work overload on caregivers’ work–family conflict might be explained by work intensification. Informal caregivers have found to be less available for work in excess of the standards hours (working longer hours or extra hours/overtime) (Fast et al. 1999; Smith 2004). Thus, having not enough time to finish work is more likely to result in additional work-to-life conflict. A second explanation why, for carers, work overload is more likely to result in work-to-life conflict is related to care-management. Working caregivers need to coordinate care services or to check on the care recipient’s status while at work (Arksey and Glendinning 2008). This will increase time-pressure at work when schedules are already tight. Additionally, caregivers might react more strongly to work-related strain because they do not have enough time for recreation after work (Ponocny et al. 2010). Especially for caregivers, reducing work intensification could be an effective measure to improve work–family balance.

Hypothesis number three set out to investigate how determinants of work–family conflict affect absenteeism. With the exception of being married, which is nonsignificant for caregivers, all the other family domain variables (i.e., having children and caring for a sick/disabled adult) have the same effect on absenteeism and the perceived level of work–life conflict. Those variables represent family stressors and result in family-to-work conflict irrespective of workplace policy.

Absences are not only an outcome of work-to-family conflict; having the possibility to miss work also constitutes a work domain resource, if work time regimes allow for/tolerate it, whether it is part of the formal work contract or not. Increased occurrence of family-related absences but lowered work–family conflict associated with flexitime regulations and having the possibility to take days off thus indicate that workers utilize these resources to adjust their short-term labour supply in order to meet their family obligations, resulting in less work-to-family conflict. Despite of increased absenteeism, granting generous work time arrangements could therefore still pay off to employers via decreased work–family conflict, which in turn decreases fluctuation and increases job satisfaction (Anderson et al. 2002). Furthermore, by allowing workers to attend to family obligations, positive family-to-work spillover could occur (Pedersen et al. 2009).

Contrary to having the possibility to take days off and flexitime regulations, allowing workers to choose their work time freely is a work-related resource that is associated with lower instead of higher absenteeism. This degree of flexibility is obviously not possible for all organisations or jobs, although the positive effects of control over one’s work schedule on absenteeism and work–family conflict speak for themselves. Likewise, the overall positive influence of the “feeling at home at work” variable underlines the importance of work cultures and climates already pointed out by several other authors (Carmichael et al. 2005; Drago and Wooden 1992; Kossek and Ozeki 1998).

Limitations and Further Research

This study is not without its limitations. Using the EWCS dataset, which was neither primarily designed to analyse care nor absenteeism or work–life conflict, meant that our analysis lacked some variables which are usually included in such analyses, such as self-reported health status or more detailed questions on consequences of work–family conflict which would allow for better identification of directional effects. Since our absenteeism variable is binary, we could not test for partial absences or duration of absenteeism spells, where differences between care-induced absences and other absences could exist (Barling et al. 1994; Hepburn and Barling 1996). We have no data on the relation between the caregiver and the care recipient, whether the care is provided in their own home or elsewhere, and the nature and severity of the care recipient’s impairments.

Furthermore, a comparative study of European countries could not be performed due to the low number of caregivers per country in the dataset. Differences in policies directed at family care (Frericks and Pfau-Effinger 2011) and attitudes/role models have been found to account for considerable differences in various aspects of caregiving, such as the decision to support parents financially or by providing informal care (Bolin et al. 2008). Further research could add to the understanding of how institutional or cultural factors influence reconcilability of paid work and adult care.

Further research could also deepen the understanding of specific situations which subgroups of caregivers to adults face. Singling out “high intensity” caregivers, that is, employees who provide substantial time help to older persons, could be intriguing. A number of studies point to differences in the impact of caregiving on employment by intensity of caregiving. In their recent study on OECD countries, Colombo et al. (2011) found that the probability to give up paid employment depended on the intensity of care. Trukeschitz et al. (2010) revealed that the impact of elder care on work-related strain differed between employees caring for persons with and without cognitive problems. Additionally, literature suggests that absenteeism and work-to-family conflict are not the only possible outcome measures of time-related incompatibility of paid work and informal care of adult dependents. Caregivers have been found to be prone to presenteeism (Smith 2004, p. 380) or partial absenteeism (Barling et al. 1994) and have different levels of labour market attachment (Henz 2006); a study of the effect of workplace policies and interpersonal relations at the workplace on these outcomes would expand our understanding of the possibilities employers have to help combining work and care commitments.