Introduction
The care-giving demands associated with an ageing population have intensified significantly across Europe and Israel, driven by increased life expectancy, demographic ageing and shifting socio-economic conditions such as increased female labour force participation and changing family structures (European Commission 2023; Knickman and Snell Reference Knickman and Snell2002). These demographic shifts create considerable challenges for both families and welfare systems, as the informal care-giving responsibilities often fall disproportionately on adult children, who must balance these duties alongside their own employment and family roles (Bolin et al. Reference Bolin, Lindgren and Lundborg2008; Da Roit and Le Bihan Reference Da Roit and Le Bihan2019).
Previous research indicates that informal care-giving provided by adult children varies significantly based on gender and labour force participation, with daughters frequently offering higher levels and more intensive types of support compared to sons (Brandt Reference Brandt2013; Evandrou et al. Reference Evandrou, Falkingham, Gomez-Leon and Vlachantoni2018; Grigoryeva Reference Grigoryeva2017). Furthermore, employment status substantially influences adult children’s ability and willingness to engage in care-giving tasks, highlighting an important trade-off between employment and informal care commitments (Bauer and Sousa-Poza Reference Bauer and Sousa-Poza2015; Fast et al. Reference Fast, Keating, Otfinowski and Derksen2004; Haberkern and Szydlik Reference Haberkern and Szydlik2010; Lee and Tang Reference Lee and Tang2015). Previous research has documented gendered differences in the types of care-giving provided, indicating that daughters typically take on more personal and time-intensive tasks, whereas sons more often engage in activities like financial management or home repairs (Evandrou et al. Reference Evandrou, Falkingham, Gomez-Leon and Vlachantoni2018; Matthews and Heidorn Reference Matthews and Heidorn1998). However, these studies primarily utilize cross-sectional approaches, lacking a detailed longitudinal perspective and not fully capturing the complexity of filial support patterns over time (Hortová and Souralová Reference Hortová and Souralová2019; Verbakel Reference Verbakel2018).
Consequently, this study addresses existing gaps by examining filial support types longitudinally, using latent Markov analysis to better capture temporal changes in care-giving practices among adult children. By classifying filial support into latent states ranging from no support to very intense support, this study comprehensively captures variations in care-giving intensity over nearly two decades (2004–2022). Additionally, previous literature often broadly categorizes countries into welfare regimes (Northern/Western versus Southern/Eastern Europe) without capturing finer distinctions in informal care-giving patterns across countries (Albertini Reference Albertini2016; Brandt et al. Reference Brandt, Haberkern and Szydlik2009; Sundström et al. Reference Sundström, Jegermalm, Abellán, Ayala, Pérez, Pujol and Souto2018). Using a multi-level latent class model allows us to address this limitation by systematically clustering countries based on their specific filial support patterns, thereby highlighting nuanced differences that may challenge prevailing typologies.
The current study aims to answer two primary research questions: (1) how are gender and labour force participation associated with filial support provided by adult children across Europe and Israel; and (2) how do filial support patterns vary across different country clusters? By combining a longitudinal latent Markov model with a multi-level latent class approach, this research offers new insights into the dynamics of care-giving, clarifying the nuanced impacts of gender and employment status within distinct cultural and institutional contexts.
Literature review
Care arrangements have become increasingly significant across European countries owing to population ageing, greater female labour force participation and a growing number of older individuals living independently (Da Roit and Le Bihan Reference Da Roit and Le Bihan2019; Sundström et al. Reference Sundström, Jegermalm, Abellán, Ayala, Pérez, Pujol and Souto2018). In many cases, healthy spouses and adult children serve as the primary sources of informal support, with spouses typically providing the most intensive care when available (Hequembourg and Brallier Reference Hequembourg and Brallier2005). However, as public resources for formal care-giving remain constrained, the responsibility placed on informal care-givers – particularly adult children – has expanded considerably (Bauer and Sousa-Poza Reference Bauer and Sousa-Poza2015; Knickman and Snell Reference Knickman and Snell2002). This responsibility is expected to increase further as baby boomers age and family sizes shrink, leaving fewer children available to share care-giving roles (Tolkacheva et al. Reference Tolkacheva, Broese van Groenou and Van Tilburg2010). Adult children’s care-giving decisions are shaped by a complex interplay of factors, including labour market participation, gender and country-specific cultural norms and welfare arrangements (Bolin et al. Reference Bolin, Lindgren and Lundborg2008; Verbakel Reference Verbakel2018).
The relationship between labour force participation and care-giving is theoretically grounded in Becker’s economic theory of time allocation, which proposes that care-giving responsibilities directly compete with time allocated for paid employment, leading care-givers to experience significant opportunity costs (Becker Reference Becker1965). Previous research consistently shows that adult children who are employed tend to provide less intensive care owing to the constraints imposed by their work commitments (Fast et al. Reference Fast, Keating, Otfinowski and Derksen2004; Haberkern and Szydlik Reference Haberkern and Szydlik2010; He and McHenry Reference He and McHenry2016; Van Houtven et al. Reference Van Houtven, Coe and Skira2013). Policies that promote employment participation may inadvertently reduce the availability of family care-givers, thus placing additional pressure on formal care-giving services (He and McHenry Reference He and McHenry2016).
The gender dimension of care-giving is well documented, grounded in social role theory and feminist perspectives, which posit that care-giving responsibilities are socially constructed and disproportionately assigned to women owing to societal norms, expectations and gender roles (Eagly and Wood Reference Eagly, Wood, Van Lange, Kruglanski and Higgins2012; Grigoryeva Reference Grigoryeva2017; Haberkern and Szydlik Reference Haberkern and Szydlik2010). Research consistently demonstrates substantial gender disparities, highlighting that daughters are more involved in care-giving tasks requiring emotional and physical intimacy, such as personal care and household tasks, whereas sons typically perform less demanding and less frequent tasks such as paperwork or financial management (Evandrou et al. Reference Evandrou, Falkingham, Gomez-Leon and Vlachantoni2018; Grigoryeva Reference Grigoryeva2017; Matthews and Heidorn Reference Matthews and Heidorn1998). These gender differences are shaped by gendered employment patterns, labour market inequalities and societal expectations about care-giving roles (Bittman et al. Reference Bittman, England, Sayer, Folbre and Matheson2003; Gerson Reference Gerson1993). For instance, women’s participation in care-giving is not only higher in quantity but also more adaptable and responsive, as women frequently adjust their work schedules or shift to part-time employment to meet care-giving demands, highlighting persistent inequalities in the intersection of care and labour market participation (Ingersoll‐Dayton et al. Reference Ingersoll‐Dayton, Neal, Ha and Hammer2003; Plantenga Reference Plantenga2002).
Cross-national differences in filial support can be explained through Esping–Andersen’s welfare regime theory, which differentiates between familialistic welfare states – where care-giving responsibilities predominantly rest with families – and defamilialized welfare states, characterized by greater state responsibility and more comprehensive formal care provisions (Albertini Reference Albertini2016; Brandt et al. Reference Brandt, Haberkern and Szydlik2009; Esping-Andersen Reference Esping-Andersen1990). Mediterranean countries, identified with strong familialistic values, typically exhibit high-intensity care-giving despite less frequent support exchanges, reflecting substantial familial responsibilities in care-giving (Albertini Reference Albertini2016; Reher Reference Reher1998). Conversely, countries with comprehensive welfare states, such as the Nordic countries, exhibit frequent but typically less intensive care-giving exchanges, indicating a specialization where formal care handles more intensive tasks and informal care-giving remains complementary (Brandt et al. Reference Brandt, Haberkern and Szydlik2009; Verbakel et al. Reference Verbakel, Tamlagsrønning, Winstone, Fjær and Eikemo2017). Pichler and Wallace (Reference Pichler and Wallace2007) emphasized the role of informal and formal social capital, suggesting that in contexts where formal care infrastructures are weak, dense family and community networks become vital sources of support – particularly relevant in Eastern European countries. Recent comparative studies have further shown that care-giving practices differ significantly across European countries based on these cultural norms and welfare policies, underscoring the importance of analysing care-giving in a cross-national context (Barczyk and Kredler Reference Barczyk and Kredler2019; Suanet et al. Reference Suanet, Van Groenou and Van Tilburg2012; Sundström et al. Reference Sundström, Jegermalm, Abellán, Ayala, Pérez, Pujol and Souto2018).
In sum, exploring filial support through the lenses of gender, employment status and cross-national differences provides critical insights into the complexity of care-giving practices. This theoretical framework underscores the necessity of adopting a nuanced, longitudinal and comparative approach to fully comprehend informal care-giving dynamics and to inform targeted and equitable policy responses.
Hypotheses
Intergenerational care and its determinants and long-term care policies’ impact on care-giving behaviour have been extensively studied. We aim to further this research by investigating different levels of filial support and its determinants in both cross-national and longitudinal settings at an individual level. This study seeks to answer two primary research questions:
(1) How are gender and labour force participation associated with filial support provided by adult children across Europe and Israel?
(2) How do filial support patterns vary across different country clusters?
To address these questions, we have formulated the following hypotheses:
H1: Countries with less extensive formal care systems, such as those in Southern and Eastern Europe, show higher involvement of filial support.
H2: Countries with more comprehensive formal care systems, such as those in Northern and Western Europe, display a more moderate to low level of filial support.
H3: Employed adult children are more likely to be involved in a less intense level of filial support compared to those who are not employed.
H4: Daughters are more likely to be involved in an intense level of filial support compared to sons.
Data and method
This study employs longitudinal data from the Survey of Health, Ageing and Retirement in Europe (SHARE), covering Waves 1 through 9, conducted biennially between 2004 and 2022. SHARE is a comprehensive, cross-national panel survey that collects data on health, socio-economic status and social networks among individuals aged 50 and above across European countries and Israel (Börsch-Supan et al. Reference Börsch-Supan, Brandt, Hunkler, Kneip, Korbmacher, Malter, Schaan, Stuck and Zuber2013). Our analysis focuses on respondents aged 65 years or older, resulting in 186,071 person-period observations.
The primary unit of analysis is the dyad of ageing parents and their adult children. Accordingly, each parent may appear multiple times in the dataset if they have more than one adult child. Each adult child is analysed individually, resulting in multiple observations per parent. This structure enables a detailed investigation of adult children’s care-giving behaviours towards their ageing parents. Importantly, unlike many care-giving studies that rely on care-giver-reported data, our analysis uses care recipients’ reports of support received.
We included respondents regardless of their health status or frailty level, recognizing that filial support needs vary significantly. This inclusive approach allows us to capture the full spectrum of support patterns, including cases in which no support is provided, reflecting both need-based and culturally driven care-giving behaviours. Excluding individuals based on health status would limit our ability to observe the broader dynamics of filial support.
The final dataset includes data from 28 countries:Footnote 1 Austria, Germany, Sweden, the Netherlands, Spain, Italy, France, Denmark, Greece, Switzerland, Belgium, Israel, Czech Republic, Poland, Luxembourg, Hungary, Portugal, Slovenia, Estonia, Croatia, Lithuania, Bulgaria, Cyprus, Finland, Latvia, Malta, Romania and Slovakia. Ireland was excluded owing to incomplete longitudinal data availability.
Our methodological approach follows a two-step modelling strategy, conceptually similar to the framework proposed by Bakk and Kuha (Reference Bakk and Kuha2018) for multi-level latent class modelling. In the first step, we apply a latent Markov model that identifies longitudinal transitions in filial support behaviours over time. This model assumes a first-order Markov structure, meaning that the latent support state at a given wave depends on the state in the previous wave. We used three binary indicators of support – paperwork help, household help and personal care – based on SHARE’s original ‘type of support’ variables. These indicators were treated as manifest variables to construct seven latent filial support states, ranging from no support to very intense support. The model captures how combinations of support types co-occur and change over time, offering a dynamic view of care-giving trajectories.
In the second step, we use a multi-level latent class model in which the latent states from the first step are treated as observed classifications. This model accounts for country-level variation and clusters countries based on the prevalence of the identified support states. By doing this, we create country clusters that reflect low, moderate and high levels of filial support involvement. Within this framework, we analyse how adult children’s gender and labour force participation relate to filial support patterns across these clusters, addressing our research questions on demographic and cross-national variation.
This two-step modelling strategy separates the estimation of care-giving patterns from the analysis of their covariates, in line with the approach developed by Bakk et al. (Reference Bakk, Tekle and Vermunt2013) and extended by Bakk and Kuha (Reference Bakk and Kuha2018). This separation helps reduce bias and improves estimation robustness. All analyses were conducted using Latent GOLD 6.0 (Vermunt and Magidson Reference Vermunt and Magidson2021; see also Vermunt and Magidson Reference Vermunt and Magidson2016 for technical details on the software and modelling procedures).
This integrated longitudinal and multi-level approach allows for a nuanced understanding of care-giving dynamics over time and across national contexts. The longitudinal component reveals how filial support evolves as parents age, while the multi-level design incorporates structural and cultural country-level differences, offering valuable insights for both theory and policy.
Results
Descriptive results
The descriptive statistics (compare with Table 1) provide an overview of various variables across five selected years, covering data from 2004 to 2022. The data include the number of parents and adult children surveyed, ranging from 11,781 to 42,486 parents and from 63,235 to 142,830 adult children. The number of person-period observations remains constant at 186,071 for each year.
Table 1. Descriptive statistics
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
The dataset reveals several key trends in intergenerational support, demographics and employment status among the sample across various European countries and Israel over this period. The average age of parents has slightly increased from 73.7 in 2004 to 74.8 years in 2022, while the average age of adult children rose from 35.7 to 41.9 years. The gender distribution among parents shows a modest increase in the percentage of females, from 54.2% in 2004 to 56.7% in 2022, while the gender ratio of adult children remains relatively stable, with the proportion of females at approximately 49%.
The employment rate among adult children shows an upward trend for both genders, particularly for women. The employment rate among female adult children increased from 70.4% in 2004 to 82.3% in 2022, while the employment rate for males rose from 84.4% to 89.9% over the same period.
Support activities such as paperwork assistance, household help and personal care exhibit varying trends over the years. Paperwork assistance increased from 4.5% in 2004 to a peak of 6.9% in 2006 before stabilizing at around 4.4%–5.0% in later years. Household help showed a significant rise from 7.9% in 2004 to 11.4% in 2006, then levelled off between 8.9% and 9.9%. Personal care assistance rose from 1.9% in 2004 to a peak of 3.5% in 2006, before declining to 1.7% in both 2019 and 2022.
Results of measurement model: forming latent states to capture longitudinal filial support types
To accurately capture the complex and dynamic nature of filial support provided by adult children to their ageing parents over an extended period (2004–2022), we applied latent Markov analysis. This methodological approach, following the framework outlined by Bakk and Kuha (Reference Bakk and Kuha2018), is particularly suitable for analysing longitudinal data because it allows us to model how individuals transition between different states of care-giving over time, rather than treating care-giving as static or uniform.
SHARE originally categorized filial support into eight distinct categories, based on the question ‘Which types of help has been provided in the last twelve months?’. The response categories were: (I) no support, (II) paperwork only, (III) household help only, (IV) household and paperwork help, (V) personal care only, (VI) personal care and paperwork, (VII) personal care and household help, and (VIII) personal care, household and paperwork help.
In SHARE, personal care involves activities such as assistance with dressing, bathing, eating, moving around the home and using the toilet. Household help refers to activities like home repairs, transportation, gardening, shopping and daily chores. Paperwork support entails assistance with administrative tasks such as filling forms and managing financial or legal matters.
To streamline the complexity and facilitate clearer interpretation of care-giving patterns, we re-coded the original categorical variable into three binary indicators: paperwork help, household help and personal care assistance (see Table 2). This binary recoding made it easier to identify specific combinations of support types clearly and consistently over multiple waves.
Table 2. Coding example of three binary variables
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
After creating these binary indicators, we estimated latent Markov models with one to eight latent states to identify the optimal number of filial support types. Model fit was assessed using the Bayesian Information Criterion (BIC) and the Akaike Information Criterion (AIC), as presented in Table 3. These indices help identify the optimal number of latent classes by balancing model complexity with sample size (Weller et al. Reference Weller, Bowen and Faubert2020). Model selection was primarily based on BIC, with AIC considered in cases where the choice between models was less clear. The model with seven latent states provided the best fit according to these criteria, offering the optimal balance between complexity and explanatory capability.
Table 3. Fit statistics for the estimated latent Markov models examining type of filial support
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
Table 4 presents the distribution for each of the seven latent filial support states identified through the latent Markov model. These probabilities indicate how likely it is that an adult child provided a specific type of support (paperwork, household help or personal care) in each latent state. For example, a high probability of personal care in a given state suggests that this state represents intensive care-giving. The table helps interpret what each latent state represents in substantive terms. Further, the resulting seven filial support states represent a spectrum of support intensity,Footnote 2 ranging from ‘no support’ to ‘very intense support’. Specifically, the categories were: no support, extremely low, very low, low, moderate, intense and very intense support (see Table 4). These seven states reflect distinct patterns with implicit intensities: The ‘no support’ state comprises individuals who provide no care-giving tasks. The states labelled ‘extremely low’, ‘very low’ and ‘low’ support represent minimal involvement in filial support, predominantly focused on paperwork or minor household assistance. The ‘moderate support’ state captures support primarily involving paperwork or household tasks, with limited engagement in personal care. The ‘intense support’ state denotes substantial involvement in household tasks and paperwork, occasionally supplemented by personal care. Finally, the ‘very intense support’ state involves extensive filial support across all three types – paperwork, household tasks and especially personal care. This classification enables a nuanced examination of support patterns of adult children, enhancing our understanding of care-giving behaviours and their evolution over time.
Table 4. Distribution of filial support types
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
Table 4 further shows that the most prevalent category, accounting for approximately 47% of the sample, is the ‘no support’ group, in which adult children provided no tasks. Among those providing support, the largest proportion fell into the ‘low support’ category (29%), indicating involvement in minimal tasks – primarily low-intensity household tasks (5%) and paperwork assistance (2%). Similarly, ‘very low support’ represented around 11% of adult children, characterized by limited support tasks, typically minor household chores or paperwork assistance. The ‘extremely low support’ group, representing about 6% of the sample, exhibited minimal support, primarily paperwork assistance.
The categories of ‘moderate’, ‘intense’ and ‘very intense’ support indicate progressively higher involvement and more extensive support activities. Although the ‘moderate support’ category includes only about 1% of adult children, it represents significant involvement in paperwork assistance (85%) and moderate household help (18%), with limited personal care (1%). The ‘intense support’ category (5%) is characterized by substantial household help (100%) and notable paperwork involvement (33%), with some engagement in personal care tasks (4%). Finally, the ‘very intense support’ category (2%) reflects adult children extensively involved in all three care-giving types – personal care (81%), household tasks (92%) and paperwork assistance (74%).
In the next analytical step, these seven latent filial support states were used in a multi-level latent class model to examine cross-country variations in filial support, explicitly exploring how gender and employment status relate to these care-giving patterns.
Results of multi-level latent class analysis
In the second analytical step, the latent states of filial support identified through the latent Markov analysis were held fixed and incorporated into a multi-level latent class model. This model allowed us to explore how care-giving patterns varied systematically across countries, explicitly considering gender and employment status differences within these contexts. By estimating models with one to five country-level clusters, we evaluated their fit using the BIC and the AIC. Based on these indicators (see Table 5), the optimal solution was determined to consist of three distinct country clusters.
Table 5. Fit statistics for the estimated multi-level latent class models
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
The filial support states identified in the first analytical stage were used to interpret the care-giving patterns within these three clusters: ‘low involvement’, ‘moderate involvement’ and ‘higher involvement’ in filial support. Table 6 displays the probability of each type of filial support (paperwork, household help and personal care) conditioned on the country-level cluster (for more information, please check Table A1 in appendix.) identified through the multi-level latent class analysis. The three clusters represent different levels of overall filial support involvement: low (Cluster 1), moderate (Cluster 2) and higher involvement (Cluster 3).
Table 6. Filial support type probabilities conditioned on the country-level cluster
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
The first cluster (‘low involvement in support’), comprising approximately 53% of the countries, is characterized by relatively low probabilities across all three care-giving indicators, with limited involvement in paperwork (4%), household work (7%) and personal care (1%). Countries within this cluster include Germany, Sweden, the Netherlands, France, Denmark, Switzerland, Belgium, Israel, Hungary, Slovenia, Croatia, Finland, Latvia, Bulgaria and Slovakia.
The second cluster (‘moderate involvement in support’), covering about 32% of the sample, reflects a slightly higher yet still moderate involvement across all types of support, with similar proportions of paperwork (4%), household help (5%) and personal care (3%). Countries in this cluster include Spain, Italy, Poland, Luxembourg, Portugal, Lithuania, Cyprus, Malta and Romania.
The third cluster (‘higher involvement in support’), accounting for 15% of the countries, shows greater involvement in care-giving activities, especially in household tasks (13%) and paperwork assistance (9%), with relatively limited involvement in personal care (3%). This cluster includes Austria, Greece, the Czech Republic and Estonia.
Table 7 presents the membership probabilities of each filial support state, conditioned on the three country-level clusters: low, moderate and higher involvement in support. The seven latent states represent varying intensities of support, ranging from no support to very intense support, as identified through the latent Markov model. The ‘low involvement’ cluster predominantly features low (63%) and very low (32%) support states, indicating minimal care-giving involvement. In the ‘moderate involvement’ cluster, the dominant support states are ‘very low’ (62%) and ‘no support’ (32%). The ‘higher involvement’ cluster shows a broader distribution, with substantial proportions in ‘low’ (60%) and ‘very low’ (27%) states, and notably higher probabilities for intense (6%) and very intense support (3%).
Table 7. Filial support state membership probabilities conditioned on the country-level cluster
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
Further analyses in Table 8 illustrate how the distribution of different filial support states varies according to gender, employment status and country clusters. Among unemployed sons, the probability of providing intensive support varies considerably by country cluster: around 20% of unemployed sons in both the low- and the higher-involvement clusters provide intensive care, rising to approximately 60% in the moderate-involvement cluster. Employed sons follow a similar, though less pronounced, pattern.
Table 8. Distribution of sons and daughters by the level of filial support and country classes
Source: Survey of Health, Ageing and Retirement in Europe (Waves 1 through 9 release 9.00), calculation by authors.
Note: Observations are based on respondents who are aged 65 or older, N = 186,071.
Gender differences are prominent: daughters, regardless of employment status, provide substantially more intensive care-giving than sons. Among unemployed daughters, those in the moderate-involvement countries show the highest level of intensive support (81%), considerably surpassing their counterparts in low- (50%) and higher-involvement clusters (42%). Employment reduces daughters’ intensive care-giving rates by about 10% across all clusters, yet employed daughters still exhibit higher care-giving involvement compared to sons. This pattern underscores the strong gender dimension in care-giving roles, highlighting persistent inequalities.
These results highlight meaningful variations in care-giving behaviour driven by gender and employment status across different cultural and institutional contexts, as captured by country clusters. The explicit identification of these clusters facilitates targeted policy recommendations and interventions tailored to address care-giving challenges effectively within diverse European contexts.
Discussion
As populations across Europe age, the question of who will provide informal care and support to older adults with declining health has become a critical topic within gerontological research. Informal care-giving remains essential in many European countries, largely owing to limited public resources and demographic pressures. Despite significant social transformations – such as increased competition in labour markets, growing female participation in paid employment and heightened job mobility (Bolin et al. Reference Bolin, Lindgren and Lundborg2008; Kalmijn Reference Kalmijn2013) – adult children continue to play a pivotal role as primary informal care-givers within familial networks (Evandrou et al. Reference Evandrou, Falkingham, Gomez-Leon and Vlachantoni2018).
This study contributes to our understanding by examining how filial support differs across country clusters, gender and the employment status of adult children. Initially, we identified seven distinct filial support states, ranging from no support to very intense support. Subsequently, through multi-level latent class analysis, we categorized 28 European countries into three clusters: low, moderate and higher involvement in filial support.
Our findings demonstrate pronounced gender and employment-status disparities within these country clusters. Unemployed sons providing intensive support accounted for approximately 20% in both low- and higher-involvement clusters, a figure that rises dramatically to around 60% within the moderate-involvement cluster. Several explanations are possible. First, in countries within this cluster, multi-generational living arrangements may be more common, increasing daily care-giving involvement simply owing to co-residence. Second, unemployed men may engage in care-giving as a form of role compensation or social contribution in the absence of paid work. In such contexts, care-giving can serve as a substitute role when formal labour market participation is limited. Third, some men may have left the labour market owing to escalating care-giving demands – suggesting a potential selection effect where unemployment is a consequence, rather than a cause, of care-giving. These explanations reflect a complex interplay between care needs, labour force exclusion and family dynamics. Although employed sons exhibited a similar trend, their involvement was consistently lower. Thus, sons’ filial support appears to vary in relation to employment conditions and prevailing care-giving norms within their country context.
For daughters, the gender disparity was even more striking. About 81% of unemployed daughters in countries with moderate involvement provided intensive care-giving, compared to 50% and 42% in low- and higher-involvement countries, respectively. Employment reduced the likelihood of intensive care-giving by approximately 10% for daughters across all clusters, yet employed daughters still provided more care compared to employed sons. These findings underscore the persistent gendered nature of informal care and support provision, reflecting broader gender inequalities across different national contexts.
Our analysis supports all four hypotheses (H1 to H4). In line with H1 and H2, we observed higher filial support involvement in countries characterized by limited formal care provision, primarily located in Southern and Eastern Europe. Conversely, countries with more developed welfare systems, notably in Northern and Western Europe, exhibited moderate or low filial support involvement, predominantly focused on household and administrative tasks. Consistent with H3 and H4, employed adult children provided less intensive care compared to unemployed counterparts, and daughters were consistently more involved in care-giving than sons.
These results align with previous studies exploring gender disparities in care-giving. Daughters generally provide more intensive and responsive care-giving, often adapting their employment schedules or working part-time to accommodate care-giving demands (Bittman et al. Reference Bittman, England, Sayer, Folbre and Matheson2003; Ingersoll‐Dayton et al. Reference Ingersoll‐Dayton, Neal, Ha and Hammer2003; Plantenga Reference Plantenga2002). Arber and Ginn (Reference Arber and Ginn1995) examined how gender fundamentally shapes the experience of ageing by highlighting the cumulative disadvantages women face across the lifecourse. They argue that social structures, especially the gendered division of labour, produce long-term inequalities in income, health and care-giving responsibilities. The article stressed that older women are disproportionately responsible for providing informal care, often involving intensive and time-consuming tasks, and that this care work is undervalued and insufficiently supported by welfare systems. Further, economic theories highlight that employment opportunities and gender wage gaps play significant roles in shaping care-giving responsibilities. Men, benefiting from better job prospects and higher wages, typically face higher opportunity costs when providing informal care, pushing care-giving responsibilities towards women (Gerson Reference Gerson1993; Sarkisian and Gerstel Reference Sarkisian and Gerstel2004). The European Commission (2022) confirms a gender wage gap, noting that women earn, on average, 13% less per hour than men, further influencing care-giving dynamics.
Additionally, cultural factors significantly affect care-giving practices. Studies suggest that parental preferences often favour daughters over sons in care-giving roles, particularly concerning personal care tasks, owing to perceived gender compatibility and comfort levels (Grigoryeva Reference Grigoryeva2017; Sigurdardottir and Kåreholt Reference Sigurdardottir and Kåreholt2014). Parents commonly anticipate that daughters will better understand and respond to their care-giving needs (Lee et al. Reference Lee, Dwyer and Coward1993; Suitor et al. Reference Suitor, Gilligan, Rurka, Peng, Meyer and Pillemer2019; Suitor and Pillemer Reference Suitor and Pillemer2006). Thus, care-giving preferences within families strongly reflect gendered cultural expectations.
The country clusters identified through our multi-level latent class analysis also warrant detailed consideration. The first cluster (low involvement), characterized by low filial support engagement across all domains (paperwork, household tasks and personal care), includes countries such as Germany, Sweden, the Netherlands, France, Denmark, Switzerland, Belgium, Israel, Hungary, Slovenia, Croatia, Finland, Latvia, Bulgaria and Slovakia. The second cluster (moderate involvement) comprises countries with slightly higher but still moderate support engagement, particularly regarding household and personal care tasks, including Spain, Italy, Poland, Luxembourg, Portugal, Lithuania, Cyprus, Malta and Romania. The third cluster (higher involvement), encompassing Austria, Greece, Czech Republic and Estonia, demonstrated notably higher involvement in tasks across all three domains.
Our country-cluster findings broadly align with previous literature that associates care-giving patterns with national welfare models and cultural norms. Northern and Western European countries, featuring more extensive welfare provisions, typically exhibit lower filial support involvement, whereas Southern and Eastern European countries, characterized by more familialistic care models, show higher involvement. However, several unexpected groupings emerged; for instance, Eastern European countries like Slovakia, Slovenia, Croatia and Bulgaria appeared within the low-involvement cluster, despite having comparatively less comprehensive welfare systems. Similarly, Austria’s placement in the higher-involvement cluster is surprising given its welfare system characteristics. These anomalies indicate more complexity in care-giving patterns across European contexts than initially anticipated, highlighting potential avenues for future research. It is important to acknowledge a limitation related to the size of certain subgroups in our analysis. Although the overall sample is large, the number of individuals categorized as providing intense or very intense support is relatively small across all country clusters. In particular, for smaller countries, this may result in limited absolute numbers within specific subgroups, which can affect the stability and generalizability of some estimates. These findings should therefore be interpreted with caution, especially when drawing conclusions about rare care-giving patterns within smaller national contexts.
This research is important in explaining gender and labour force participation from the perspective of filial support across different country clusters. We have encompassed the majority of tasks that adult children can offer to their ageing parents, spanning a broad spectrum of support levels (from no support to extremely intense support). This study marks the initial step towards a more comprehensive exploration of factors influencing filial support states. Future studies could examine the differences between countries, particularly those that do not align with their country class, such as Luxembourg, Bulgaria, Croatia, Slovenia and Slovakia. Additional demographic characteristics, such as adult children having their own children, marital status and education status, should also be examined.
Appendix
Table A1. Distribution of country clusters
Source: Survey of Health, Ageing and Retirement in Europe (wave 1 through wave 9 release 9.00), calculation by authors.
Observations are based on respondents who are 65 or older, N = 186,071.
Open access
