Joint Physical Custody in Europe: A Comparative Exploration

Abstract

Joint physical custody (JPC)—where children alternate between parental households after a separation—has been found to be on a rise in a diversity of European countries. However, variations in sampling, data and operationalization consistently complicate the comparative mapping of JPC prevalence and its correlates. In this respect, the 2021 EU-SILC ad hoc module on Living arrangements and conditions of children in separated and blended families provides a unique opportunity to study JPC prevalence across Europe. The current study aims to validate and expand on existing research by employing the module’s second release, concerning children in households across 21 countries. Our descriptive overview of shared versus sole and main residence supports and adds to the previously noted diversity of JPC in Europe. We further note a similar father- to mother-residence ratio in countries with high and low incidence of JPC, warranting consideration of how social and legal norms regulate the granting of custody to mothers versus fathers in various country settings. Subsequently, we analyse the association of child-, parent-/household- and country-level characteristics with JPC using a three-level linear mixed model. The results underline the importance of a multi-level approach to understanding the correlates of JPC and prompt the elaboration of country comparisons using the EU-SILC module.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10680-025-09732-y.

Introduction

Following a parental separation, a decision must be made concerning the residence of the child(ren). Either a child can live mainly or exclusively with one parent, which is generally defined as sole physical custody, or they can alternate between both parent’s households. This sharing of childcare, or joint physical custody (JPC), has gained increasing (academic) interest over the past decades, mainly in the light of its supposed psychological, social, and physical benefits for both parents and children (Steinbach et al., 2021; Turunen, 2017). While there remains debate whether these outcomes are due to more advantageous (e.g. socio-economical) characteristics of separated parents who practice JPC, sharing childcare has irrefutably become more widespread across Europe (Braver & Votruba, 2021; Hakovirta et al., 2023; Nielsen, 2018). Insofar as the diffusion of JPC entails a growing diversity in families’ characteristics, this could imply a diminishing socio-economic vantage of those who practice JPC, and thus more broadly experienced benefits by parents and children (Bergström et al., 2014).

However, a diversification of families’ characteristics is not of sole importance in understanding the increased prevalence of JPC. Insofar as uptake is moderated through a country’s legal system and reigning social norms, the specific country-setting is also vital to understanding patterns and differences underlying JPC prevalence and outcomes across Europe. Currently, documentation on JPC uptake remains scarce, and the small-sample focus of existing studies greatly limits cross-country inferences (Nielsen, 2018; Zilincikova, 2021). Notable headway has recently been made in this respect by studies that employ the first release of the 2021 EU-SILC ad hoc module on Living arrangements and conditions of children within separated or blended families to comparatively assess JPC prevalence across Europe. Using this module, Hakovirta et al. (2023) noted a doubling of JPC in less than 20 years, while Salin et al. (2024) subsequently identified key socio-economic and socio-demographic characteristics at the individual, family and country level that correlate to JPC presence in 17 countries.

We add to these findings in three ways. First, we employ the second release of the EU-SILC module, allowing for an assessment of JPC prevalence across a larger sample of 21 countries. Second, we provide a descriptively more detailed account of custody arrangements across these countries than has currently been done, by differentiating between equal JPC, main and sole father residence, and main and sole mother residence. We further expand upon socio-economic correlates of JPC uptake by also considering psychosocial aspects, such as the perceived ability to make ends meet, life satisfaction and trust in others. Third, building on the finding of Salin et al. (2024) that the introduction of legal joint custody (JLC) is associated with JPC presence, we consider the validity of clustering countries by the timing of its implementation. In doing so, we not only solidify earlier findings using the EU-SILC module, but also expand our knowledge of the importance of the country context in explaining how JPC uptake takes shape across Europe.

Background

Existing research suggests that the correlates of JPC uptake are located at different levels. Firstly, and mainly sourced by single-country studies, individual characteristics of parents and children appear to be salient factors in predicting custody arrangements. Secondly, parent and household characteristics are of importance, as they create (practical) vantages or restraints for specific custody choices. Finally, increased attention is being given, mainly by comparative studies, to the importance of the country setting, in terms of, e.g. the legal context, as well as the socially normative setting surrounding separation and the sharing of childcare. In this section, we discuss these correlates and their importance in the light of JPC prevalence across Europe.

Child-Level Characteristics

Concerning characteristics of the child, there appears to be a curvilinear relationship between JPC and age, with very young children and older adolescents being more likely to reside in single-parent than in shared custody (Juby et al., 2005; Steinbach, 2019). In terms of gender, results vary by country, with some studies indicating that boys are more likely to be in JPC, while others note no gender difference (Bauserman, 2012; Poortman & Van Gaalen, 2017). Further, as stated previously, JPC is generally associated with higher well-being among children, related to both social, mental and academic outcomes (Nielsen, 2018). However, using cross-sectional data, selectivity or reverse causality may imply that such aspects are (also) determinants, and not merely outcomes of JPC, as parents may be more likely to expose children to frequent household transitions when children are in a more stable “state” (Augustijn, 2022; Fransson et al., 2018). As such, this study considers children’s age, gender and well-being as potential correlates of JPC.

Parent- and Household-Level Characteristics

Studies repeatedly note that separated parents practicing JPC tend to differ from those with other custody arrangements in terms of their socio-economic and socio-demographic profiles. First, sharing care of children across households appears to be more likely when both parents are working, have a higher educational degree and earn a greater (joint) income (Bergström et al., 2019; Poortman & Van Gaalen, 2017; Vanassche et al., 2017; Walper et al., 2021). This is partly due to practical requirements: JPC has long been identified as a resource-intensive arrangement, requiring dual housing for children and the duplicate purchase of basic amenities (Melli & Brown, 1994). Again, this does not necessarily imply causality. While opting for JPC requires initial investments, it can facilitate (especially mothers’) employment (Bonnet et al., 2022), thus increasing a parent’s financial resources as a result of (and as a way to maintain) the sharing of childcare. Meanwhile, objective income measures may fall short without consideration for the larger (familial) context, e.g. in terms of the number of children and the presence of a new partner. While results are mixed, some studies indicate that JPC is more difficult to implement for more children—both due to the dual (initial) costs for each child, as the logistics of moving and managing multiple children on a frequent basis (Claessens & Mortelmans, 2021; Kalmijn & De Graaf, 2000). Concurrently, having a (new) partner may burden or contribute to the household’s finances, thus (indirectly) affecting the maintainability of JPC. Furthermore, for women, sharing childcare after separation has been found to be positively related to the presence of a (new) partner, due to lessened time restraints (Bakker & Mulder, 2013; Cancian et al., 2014). On the parent level, we therefore include educational degree, work status, poverty risk, the number of children and the presence of a partner in the household, along with an objective resource measure. Furthermore, considering the importance of the perceived (as well as the actual) ability to make ends meet, we also include a subjective resource measure as a psychosocial aspect potentially affecting JPC take-up.

As JPC arrangements adapt over time according to individuals’ and families’ needs (Claessens & Mortelmans, 2021; Poortman & Van Gaalen, 2017), we further argue the importance of considering the psychosocial dynamics that potentially underlie the tenability of JPC. We firstly take into account life satisfaction, which, while often mediated through other relational and contextual factors, has been found to be positively associated with sharing childcare for mothers and (to a lesser extent) fathers (Augustijn, 2023a, 2023b; Steinbach, 2019). We secondly consider trust in others. Sharing care has been found to improve cooperation between parents (Spruijt & Duindam, 2009), but simultaneously (and reversely) requires trust in both the other parent (who takes care of the child for a significant portion of time) as in the larger familial and support networks in which the child resides (Ver Steegh & Gould‐Saltman, 2014; Volgy & Everett, 2014). Moreover, Sodermans et al. (2015) suggest that trust goes hand in hand with social interaction and cooperation in social networks, which are both essential to functional JPC arrangements. As such, we consider both parental life satisfaction and general level of trust in others.

Country-Level Characteristics

Through the rise in available data for cross-country comparative JPC research, the importance of the macro-context in assessing and understanding JPC prevalence and its diversification is becoming increasingly clear. First, we consider the normalization of relationship dissolution as a catalyst to both social and legal trends underlying JPC uptake. The overall increase in divorce and separation across Europe can be perceived as an expression of both greater (financial) independence—especially of women— and of changing expectations of relationships, towards, e.g. a greater valuation of personal autonomy (Boertien, 2020)—which are both linked to a rise in JPC uptake. The proportion of non-intact families in a country therefore provides an indication of the likeliness of JPC uptake—which requires certain social institutionalization in order to become a generally accepted practice (Fučík, 2020).

Concurrently, increased involvement of fathers in childrearing has been noted across a diversity of country settings (Haux & Platt, 2021; Yavorsky et al., 2015), motivated by social support for father involvement in children’s daily lives and the benefits in terms of relational and psychological well-being (Cano et al., 2019; Oren & Hadomi, 2020; Popov & Ilesanmi, 2015). Paternal involvement pre-separation, in turn, is linked to post-separation childcare (Poortman, 2018; Westphal et al., 2014), suggesting it to be a driver of increased JPC uptake. At the foundation of increased paternal involvement also lay (implicit) assumptions concerning gender roles embedded in, among others, labour market and family policies (Dilli et al., 2019). A useful indicator in this respect is the female labour market participation in a country, as, especially for women, work creates time restraints and motivates the sharing of childcare (Claessens & Mortelmans, 2021). As the division of care tends to perpetuate after separation, a greater proportion of working women in a country thus potentially coincides with the normalization of JPC (André et al., 2013).

Second, generally prevailing gender norms are notably reflected in the gender equality index (GEI), a composite measure capturing country-level gender equality in six core domains (work, money, knowledge, time, power and health) across Europe. As noted by Salin et al. (2024), countries with a higher GEI score are more likely to have systems in place that support dual-worker households and, simultaneously, motivate responsibility for more gender equal earning and childcare, thus lowering the threshold to entering a JPC arrangement.

Finally, legal frameworks surrounding custody of children after separation function both as a catalyst and consequence of evolving social norms. Over the past decades, most European countries revised their legal approaches to child custody after separation, generally aiming to stress the importance of continued parental involvement. This minimally refers to legal joint custody (JLC), where both parents (are motivated to) share legal parental responsibilities, irrespective of their relationship or living situation, in the light of the best interests of the child. In this respect, Steinbach (2019) suggests that a legal motivation to ensure children with regular and meaningful contact with both parents through JLC can be reflected in more accepting attitudes towards and social norms surrounding JPC, thus motivating the sharing of childcare.

As social and legal norms concerning relationship dissolution and the sharing of childcare simultaneously affect the uptake of JPC (Fučík, 2020; Salin et al., 2024), we take into account the broader country context when assessing JPC prevalence. More precisely, we consider the proportion of non-intact families, GEI, female labour market participation, and introduction of JLC legislation for each country under analysis. Considering the (practical) feasibility and restraints vis-à-vis JPC uptake created especially by the legal context (and its reported significance as a single indicator (Salin et al., 2024)), we group countries by the introduction period of JLC as a first assessment of the validity of group-wise comparisons. Furthermore, as we know from the Second Demographic Transition theory, accounting for distinct period-effects is highly relevant when considering normative change, such as surrounding separation and childcare across Europe (Lesthaeghe & Surkyn, 1988).

Data and Methods

Data

The European Union Statistics on Income and Living Conditions (EU-SILC) collects comparable cross-sectional and longitudinal data on income, poverty, social exclusion and living conditions for nationally representative samples of persons aged 16 or older in 32 European countries. As a four-year rotating panel, one-quarter of the longitudinal sample is refreshed yearly with a refresher sample. Consequentially, the data contain four-wave longitudinal information and are cross-sectionally representative for each country on a yearly basis (Eurostat, 2022b). Longitudinal data are collected individually, whereas the cross-sectional data concern both the individual and household level. On the individual level, all household members aged 16 and above are invited to provide demographic and income data. Additionally, one respondent answers the household questionnaire for each household.

This study uses the 2021 ad hoc module ‘Living arrangements and conditions of children in separated and blended families’, collected from the household respondent. This respondent is first asked for each child (aged 17 or younger) in their household whether they have a parent who is part of a different household.1 The module then provides questions on children whom this regards, thus collecting detailed information on physical custody arrangements in particular households. The analyses in this study concern the second (spring 2023) release of the 2021 EU-SILC data, for which we consider the ad hoc module for 21 European countries (see Table 1).2 The total dataset consists of 185,013 households, of which 51,788 intact households with children and 6939 households for which at least one child has a parent living outside the household. At the child level, we analyse data on 10,721 children. After listwise deletion for missing variables,3 the multivariate analysis pertains to 6945 children.

Table 1.

Proportion of non-intact families, GEI, introduction of JLC, and female labour market participation (aged 20–64) by country

Proportion of non-intact familiesa		Gender equality index 2022b		Introduction of joint legal custodyc		Female labour market participationd
Country	%	Country	GEI	Country	Year	Country	%
Belgium	6.40	Sweden	83.9	Sweden	1976	Switzerland	82.4
France	6.11	Denmark	77.8	Finland	1983	Sweden	82.0
Malta	5.17	Switzerland	77.8	Denmark	1986	Estonia	81.9
Czech Rep	4.81	Finland	75.4	Cyprus	1990	Czech Rep	81.3
Sweden	4.68	France	75.1	Estonia	1995	Hungary	80.2
Finland	4.50	Spain	74.6	Hungary	1995	Denmark	80.1
Spain	4.26	Belgium	74.2	Belgium	1995	Malta	80.1
Denmark	4.20	Austria	68.8	Czech Rep	1998	Lithuania	79.0
Italy	3.66	Slovenia	67.5	Lithuania	2000	Finland	78.4
Hungary	3.47	Malta	65.6	Austria	2001	Cyprus	77.9
Switzerland	3.41	Italy	65.0	France	2002	Slovenia	77.9
Cyprus	3.16	Estonia	61.0	Croatia	2003	Austria	77.3
Estonia	3.14	Croatia	60.7	Romania	2004	Poland	76.7
Austria	3.13	Lithuania	60.6	Slovenia	2004	France	74.0
Slovenia	2.88	Serbia	58.0	Spain	2005	Belgium	71.9
Serbia	2.52	Poland	57.7	Serbia	2005	Croatia	70.2
Lithuania	2.45	Cyprus	57.3	Italy	2006	Spain	69.3
Poland	2.18	Czech Rep	57.2	Poland	2009	Serbia	68.8
Romania	1.97	Hungary	54.2	Malta	2011	Romania	68.5
Greece	1.45	Romania	53.7	Switzerland	2014	Greece	66.3
Croatia	0.92	Greece	53.4	Greece	2021	Italy	64.8

^aAggregated from 2021 EU-SILC module on Living arrangements and conditions of children in separated and blended families (second release)

^bEuropean Institute for Gender Equality (2022), data mostly from 2020; for Switzerland: authors’ estimation, inferred from the Gender Inequality Index 2020 (UNDP, 2020)

^cAustria: Böheim et al., 2012; Belgium: Vanassche et al., 2017; Switzerland: Schwenzer & Keller, 2014; Cyprus: Nicolaou, 1996; Czech Republic: Fučík, 2020; Denmark: Graversen, 1986; Estonia: Kullerkupp, 2001; Greece: Stylianidis, 2021; Spain: Solsona & Spijker, 2016; Finland: Kurki-Suonio, 2000; France: Fulchiron, 2005; Croatia: Graovac, 2022; Hungary: Szeibert, 2013; Italy: de Blasio & Vuri, 2013; Lithuania: Maslauskaitė & Tereškinas, 2017; Malta: Sammut, 2018 (set as 2011 due to legalization of divorce and standard practice of favouring joint legal custody of parents); Poland: Sokołowski, 2021; Romania: Slabu, 2019; Serbia: Kitanovic, 2014; Sweden: Singer, 2008; Slovenia: Kraljic, 2008

^dAged 20–64, Eurostat (2022a). Employment and activity by sex and age. https://doi.org/10.2908/lfsi_emp_a

Measures

The dependent variable was derived from the question ‘How many nights does [child’s name] stay in your household during a typical month?’, recoded to a dichotomous indicator measuring the presence of JPC. In line with the consensus among JPC researchers concerning the required time spent with each parent for the custody arrangement to be considered as JPC, the boundaries for overnight stays were set at a minimum of 10 and maximum of 20 nights per month (taken as 31 days). This corresponds with living 33% to 66% of monthly time with the household respondent, of which the complement is spent with the other parent. For the descriptive results shown in Table 2, JPC is distinguished from main residence (living 67%-99% of the time with one parent) and sole residence (100% of the time with one parent) with either the mother or the father. These latter categories are merged into the reference group ‘no JPC’ for the descriptive statistics (Table 3) and multivariate analysis (Table 4). It is important to note that as physical custody arrangements can differ for multiple children in one household, the presence of JPC is a characteristic on the child level. In our sample, 9.54% of households have multiple children who do not all have the same physical custody arrangement.

Table 2.

Physical custody arrangements by country

	N	Sole physical custody—mother (100%)	Main residence—mother (67–99%)	JPC (33%–66%)	Main residence—father (67–99%)	Sole physical custody—father (100%)	Ratio father-to-mother custody
Austria	337	50.52	18.01	9.51	9.23	12.73	0.32
Belgium	847	41.33	12.54	30.15	10.43	5.54	0.30
Switzerland	388	19.86	30.98	21.41	18.20	9.54	0.55
Cyprus	249	47.18	26.80	10.32	4.79	10.91	0.21
Czech Rep	552	42.47	27.65	13.66	7.49	8.73	0.23
Denmark	441	21.08	20.11	39.89	11.68	7.24	0.46
Estonia	370	56.18	10.08	18.69	5.19	9.85	0.23
Greece	325	45.50	28.58	1.66	9.32	14.94	0.33
Spain	1495	36.28	18.28	21.88	10.42	13.14	0.43
Finland	665	24.71	22.82	31.76	11.00	9.71	0.44
France	1576	31.11	23.27	29.25	10.48	5.89	0.30
Croatia	85	46.16	9.72	5.29	13.28	25.55	0.69
Hungary	396	65.54	14.68	5.40	3.34	11.04	0.18
Italy	766	56.59	16.35	5.35	8.63	13.09	0.30
Lithuania	220	83.17	5.00	0.59	3.71	7.54	0.13
Malta	204	51.94	21.30	5.86	4.99	15.90	0.29
Poland	592	65.75	4.34	2.89	7.93	19.09	0.39
Romania	103	58.12	6.53	4.90	4.78	25.66	0.47
Serbia	161	51.27	11.36	5.36	8.69	23.31	0.51
Sweden	667	27.13	6.95	53.72	3.64	8.55	0.36
Slovenia	282	44.54	21.32	22.13	5.18	6.83	0.18

N = 10,721; 2021 EU-SILC module on Living arrangements and conditions of children in separated and blended families (second release), weighted with the EU-SILC personal cross-sectional weight factors

Table 3.

Descriptive statistics of the dependent and independent variables by JPC presence

	JPC: Yesa			JPC: No
	%	M	SD	%	M	SD
Outcome variable at the child-level
Joint physical custody	21.66			78.34
Child-level characteristics
Age		10.60	1.15		10.51	1.25
Gender
(ref. Female)
Male	54.05			51.93
Health
(ref. Fair & Bad)
Very good	66.20			64.47
Parental/household-level characteristics
Gender
(ref. Male)
Female	58.12			80.29
Educational level
(ref. Middle (ISCED 3–4))
Low (ISCED 0–2)	10.81			22.23
High (ISCED 5–8)	52.79			30.47
Self-defined economic status
(ref. Other status)
Full-time	88.50			72.85
Part-time	6.76			15.27
Unemployed	1.72			1.61
Make ends meet
(ref. Not difficult/not easy)
Difficult	17.55			33.93
Easy	29.71			14.57
At risk of poverty
(ref. No)
Yes	17.83			29.35
Household size		3.39	3.27		3.63	3.90
Partnered
(ref. No)
Yes	22.66			26.11
Life satisfaction
(ref. Not low/not high)
Low	1.40			2.65
High	52.22			43.88
Trust in others
(ref. Not low/not high)
Low	9.50			15.64
High	28.50			21.19

N = 10,721; ^aJPC = Joint physical custody (33–66% of nights per month with each parent); 2021 EU-SILC module on Living arrangements and conditions of children in separated and blended families (second release), weighted with the EU-SILC personal cross-sectional weight factors

Table 4.

Hierarchical logistic regression predicting the presence of JPC^a (Three Levels)

	Model 1			Model 2			Model 3
	b (OR)	se		b (OR)	se		b (OR)	se
Intercept	− 6.60 (0.00)	1.36	***	− 4.79 (0.01)	1.33	**	− 24.60 (0.00)	4.49	***
Child-level characteristics
Age	1.13 (3.10)	0.27	***	1.12 (3.06)	0.27	***	1.14 (3.13)	0.27	***
Age2	− 0.06 (0.94)	0.01	***	− 0.06 (0.94)	0.01	***	− 0.06 (0.94)	0.01	***
Gender
(ref. Male)
Female	− 1.24 (0.29)	0.30	***	− 1.23 (0.29)	0.30	***	− 1.29 (0.28)	0.29	***
Health
(ref. Fair & Bad)
Very good	− 2.34 (0.10)	1.12	***	− 2.36 (0.09)	1.11	*	− 2.39 (0.09)	1.11	*
Parent-/household-level characteristics
Gender
(ref. Male)				− 3.28 (0.04)	0.27	***	− 3.49 (0.03)	0.25	***
Educational level
(ref. Middle (ISCED 3–4))
Low (ISCED 0–2)				− 1.06 (0.35)	0.36	**	− 0.93 (0.39)	0.37	*
High (ISCED 5–8)				1.66 (5.26)	0.36	***	1.78 (5.93)	0.38	***
Self-defined work status
(ref. Other status)
Full-time				1.41 (4.10)	0.49	**	1.58 (4.85)	0.56	**
Part-time				− 0.05 (0.95)	0.48		− 0.06 (0.94)	0.54
Unemployed				0.98 (2.66)	1.13		0.55 (1.73)	1.16
Make ends meet
(ref. Not difficult/not easy)
Difficult				− 0.76 (0.47)	0.29	*	− 0.56 (0.57)	0.30
Easy				0.51 (1.67)	0.38		0.38 (1.46)	0.43
At risk of poverty (ref. No)				0.11 (1.12)	0.29		− 0.03 (0.97)	0.25
Household Size				− 0.21 (0.81)	0.12		− 0.22 (0.80)	0.12
Partnered (ref. No)				− 0.95 (0.39)	0.35	**	− 0.98 (0.38)	0.36	**
Life satisfaction
(ref. Not low/not high)
Low				0.06 (1.06)	0.64		− 0.31 (0.73)	0.65
High				0.13 (1.14)	0.24		0.15 (1.16)	0.26
Trust in others
(ref. Not low/not high)
Low				− 0.74 (0.48)	0.31	*	− 0.58 (0.56)	0.34
High				0.41 (1.51)	0.31		0.46 (1.58)	0.34
Country-level characteristics
Prop. non-intact families							1.93 (6.89)	1.29
Introduction JLC
(ref. 1990’s)
1970’s-1980’s							− 1.36 (0.26)	0.73
2000’s and later							− 1.74 (0.18)	0.57	*
Gender equality index							0.23 (1.26)	0.03	***
Female labour market participation							0.06 (1.06)	0.05
N	6945			6945			6945
Deviance (-2LL)	290,955.4			292,969.1			283,120.5
Unexplained variance (household)	3.81	***		3.25	***		3.42	***
Unexplained variance (country)	0.18	***		0.14	***		0.01	***

^aJPC = Joint physical custody (33%-66% of nights per month with each parent), measured at the child-level; b = Unstandardized beta, OR = odds ratio, se = standard error; * p < .05 **, p < .01 ***, p < .001; 2021 EU-SILC module on Living arrangements and conditions of children in separated and blended families (second release), weighted with the EU-SILC personal cross-sectional weight factors

In line with the theoretical framework presented above, indicators at the child, parent/household and country level were included as independent variables. As the EU-SILC study’s primary aim is collecting detailed and reliable data on income and social exclusion, only a limited number of characteristics of the children in separated or blended families are measured. Therefore, we can only include the child’s age (and its quadratic form, as previous research has noted a curvilinear association between children’s age and JPC prevalence (Bergström et al., 2014)), gender and a health measure, as a proxy for well-being. In this respect, the responding parent answered the question ‘How would you describe [child’s name]’s health in general?’ with answers ranging on a scale from 1 (very good) to 5 (very bad). As the variable was highly skewed, we included it as a dummy variable contrasting the category ‘very good’ with the four lower ranked categories.

The parent-level indicators contain measures both at the household and individual level, concerning the parent answering the module for the child. The responding parent’s educational level was harmonized using the international ISCED classification. We regrouped the categories as follows: ISCED 0–2: ‘low education’ (less than high school), ISCED 3–4: ‘middle education’ (high school) (ref.) and ISCED 5–8: ‘high education’ (tertiary education). Their self-defined work situation was recoded to ‘full-time work’, ‘part-time work’, and ‘unemployed’, consisting of those who are retired, in education or not actively working, with ‘other status’ (fulfilling domestic tasks, compulsory military or civilian service or other) as the reference category. Next, two financial resource indicators were included. First, we included the harmonized risk of poverty (AROP) measure of EU-SILC. The indicator is defined as having an equivalized disposable income below the risk-of-poverty threshold, set at 60% of the national median equivalized disposable income (after social transfers) (Eurostat, 2022b). The reference category in the analysis was ‘not at risk’, i.e. above the 60% threshold. Second, we included a measure of the degree to which the responding parent perceives their household as able to make ends meet, defined as ‘The ability to pay for its usual necessary expenses’, which serves as a psychosocial aspect potentially affecting JPC uptake. We contrasted three categories: ‘easy to make ends meet’, ‘not difficult or not easy’ (ref.) and ‘difficult to make ends meet’. Further, two additional psychosocial indicators at the parent level were life satisfaction and trust in others. Life satisfaction of the responding parent was measured through a single item (‘Overall, how satisfied are you with your life these days?’), measured on an 11-point scale. This was recoded to ‘low satisfaction’ (scale scores 0–2), ‘high satisfaction’ (scale scores 8–10) and ‘medium satisfaction’ (scale scores 3–7) as the reference category. We further included a trust indicator, ‘To what extent do you trust other people?’, which was recoded similarly. Finally, two controls were included: household size of the responding parent (as a continuous measure) and a dichotomous indicator of whether the responding parent is partnered.

To account for the country context, four country-level indicators were included. First, we considered the proportion of non-intact families in a country, internally aggregated from the EU-SILC data. The second indicator concerned the 2022 Gender Equality Index, which contains data mostly from 2020 and comprises 6 subdimensions: work, knowledge, time, money, health and power (European Institute for Gender Equality, 2022). Third, building on the work of Salin et al. (2024), we aimed to further unravel the previously found significant association between the introduction of JLC legislation with JPC, by testing its validity as grouping unit. As such, we included dummies to account for the introduction of JLC legislation between 1970 and 1989 (Sweden, Finland, Denmark), from 1990 to 1999 (Cyprus, Belgium, Estonia, Hungary, Czech Republic), and from 2000 onwards (Lithuania, Austria, France, Croatia, Romania, Slovenia, Spain, Serbia, Italy, Malta, Switzerland, Poland and Greece). Finally, we included the 2022 female labour market participation in a country, expressed as a percentage of actively working women aged 20–64, derived from Eurostat (2022a).

Analytical Strategy

For the multivariate analysis, we firstly estimated weighted descriptive statistics for the prevalence of JPC in the considered countries, as well as for the dependent and independent variables. The personal cross-sectional weight factors calculated by EU-SILC were used to correct for survey design effects and to produce nationally representative (and comparable) results. Next, we used linear mixed modelling with a logit link function to predict the probability of a child being in JPC, applying Full Maximum Likelihood estimation. As children are clustered in households which are clustered in countries, a three-level model was applied with the same weights as in the descriptive tables. The first level considered only the child characteristics, with the second and third model adding the parent/household- and country-level variables, respectively.

Results

Prevalence of JPC in Europe

As shown in Table 2, there was considerable variation in the prevalence of JPC across Europe. The Northern European countries showed the highest proportions of JPC, with Sweden (53.7%), Denmark (39.9%) and Finland (31.8%) leading the ranks. Next came mainly Western European countries such as Belgium (30.2%) and France (29.3%), with some notable exceptions (e.g. Slovenia (22.1%) and Spain (21.9%)). Confirming previous studies (Hakovirta et al., 2023; Salin et al., 2024), we found the lowest prevalence of JPC in Eastern Europe, which, due to our use of the second release of the EU-SILC module, included countries such as Malta (5.6%), Serbia (5.4%) and Poland (2.3%).

In all countries, sole and main mother residence remained dominant over father residence. Notably, the ratio of father versus mother custody (as a sum of both sole and main residence) appeared higher in both countries with the highest and the lowest incidence of JPC. For example, whereas the highest-ranking JPC countries show a ratio of father- to mother-residence upwards of 0.35 (Sweden and Finland) and even 0.45 (Denmark), similar results occurred for several countries with the lowest incidence of JPC, such as Greece (0.33), Poland (0.39), Romania (0.47), Serbia (0.51) and Croatia (0.69).

Descriptive Statistics

Table 3 presents the descriptive statistics of all variables, according to the presence of JPC. Roughly one out of five children across all considered countries resided in a JPC arrangement (21.7%). Minimal differences appeared on the child level concerning the age, gender and health composition of the groups with and without JPC. More notable were the differences in the parent-level characteristics, starting with a higher prevalence of reporting mothers (than fathers) among the group without JPC. Conversely, higher education levels, being in full time work, easily making ends meet and a lower poverty risk were more prevalent among parents with (than without) JPC, along with higher scores on life satisfaction and trust in others.

Multivariate Analysis

Table 4 presents the three-level logistic regression analysis. First, the child characteristics were related to being in JPC across all models. For age of children we found a consistent quadratic relation, with the likelihood of residing with both parents first increasing, and then plateauing or slightly decreasing at later ages. Boys appeared less likely than girls to reside in both parental households across all models, and being in very good health was also associated with a lower likelihood of JPC. While the addition of the parent-/household-level characteristics and subsequent country-level characteristics in, respectively, Models 2 and 3 decreased the significance of the association between the child’s health and JPC (from p < 0.001 to p < 0.05), the negative relationship was retained across models (Model 1: OR = 0.10; Model 2 and 3: OR = 0.09).

On the parent/household level, being more highly educated was strongly and positively linked to the prevalence of JPC both in Model 2 (OR = 5.26, p < 0.001) and Model 3 (OR = 5.93, p < 0.001), while being lower educated revealed a slightly weaker negative association in each model (Model 2: OR = 0.35, p < 0.01; Model 3: OR = 0.39, p < 0.05). Being in full-time employment was positively related to JPC in both models (Model 2: OR = 4.10, p < 0.01; Model 3: OR = 4.86, p < 0.01), whereas no differences were found in poverty risks among households with and without JPC. Experiencing difficulties in making ends meet and a low reported trust in others were weakly and negatively associated with physically sharing care in Model 2 (resp. OR = 0.47, p < 0.05 and OR = 0.48, p < 0.05), but both associations disappeared with the inclusion of the country-level characteristics in Model 3. Finally, having a partner showed a consistently negative association with JPC across Model 2 (OR = 0.39, p < 0.01) and Model 3 (OR = 0.38, p < 0.01).

On the country level, the Gender Equality Index was strongly and positively related to JPC (OR = 1.26, p < 0.001), while later introduction of JLC (from 2000 onwards, compared to the 1990’s) showed a negative association (OR = 0.18, p < 0.05). For the proportion of non-intact families and female labour market participation, no association with JPC was found.

Conclusion

This study maps the prevalence of JPC across Europe, taking into account child-, parent-/household- and country-level indicators. By employing the second release of the 2021 EU-SILC module on Living arrangements and conditions of children in separated and blended families, we build on existing studies in three ways. First, we provide an expanded descriptive overview of JPC prevalence in terms of custody arrangements and number of countries. Second, we assess the relevance of psychosocial indicators along with previously identified socio-economic factors related to JPC, and third, we broaden the consideration of the country level by including more controls while assessing the validity of grouping countries by their introduction of JLC legislation.

Our descriptive results both confirmed and expanded on previous studies using the EU-SILC module. We found JPC prevalence (compared to main and sole custody) to be highest in Northern European countries, but closely followed by several others (e.g. France and Belgium). While the incidence of mother residence remained higher than father residence in all countries, we noted a similarity in the father- to mother-residence ratio for countries with the highest and lowest JPC incidences. This was especially prominent due to the inclusion of several countries with lower JPC levels not previously considered using the first release of the EU-SILC module. It could be expected that, in the light of country-level differences concerning legal and social care norms across Europe (in terms, e.g. of labour market and family policies: Dilli et al., 2019; Saraceno, 2022), mother-oriented care would remain most prominent in contexts with less institutionalization of father-involved and gender equal care-giving. This warrants further investigation, preferably building on the EU-SILC module, of how and why dissimilarities in JPC prevalence align with the father- to mother-residence ratio in potentially very different policy contexts across Europe.

Next, the multivariate analysis yielded several expected results concerning the individual- and household-level indicators. Confirming previous findings, age and gender of the child, as well as parental employment and education, were significantly associated with JPC across all models (Juby et al., 2005; Poortman & Van Gaalen, 2017; Zilincikova, 2021). Meanwhile, the child’s health, which we included as a proxy for well-being, appeared to be negatively rather than positively associated with JPC. While living with both parents is generally considered beneficial for a child’s well-being (even when controlling for parents' socio-economic profile: Nielsen, 2018), their health can be simultaneously, and reversely, associated with JPC. Health issues may be burdensome for parents with sole or main custody, thus motivating the sharing of childcare. We must also consider the possibility of reversed causality, as frequent movements across households may negatively impact the child’s (mental) health. While in countries with a high incidence of JPC less health issues have been found among children in shared care (compared to those in other custody arrangements) (Bergström et al., 2015), how this association takes form in settings with less legal and social normalization of and support for JPC (of which several countries are present in the EU-SILC dataset), remains unclear. Furthermore, Augustijn (2023b) found that the link between JPC and children’s health can be partly explained by the quality of the mother–child relationship, indicating the importance of assessing the broader (social) context in which a child resides. As such, although the limited number of child-level indicators prohibits further investigation using the EU-SILC module, our results prompt future comparative research into child’s health across countries with varying JPC prevalence.

Neither our objective (poverty risk) nor subjective (perceived ability to make ends meet) measures of financial resources were related to JPC. The lack of statistical significance is in any case not due to a markedly high correlation between these two variables representing financial resources (C = 0.034, p < 0.05). As such, while potentially a positive indication concerning legal implementations that financially support the physical sharing of care (e.g. the automatic sharing of benefits: Hakovirta & Rantalaiho, 2011; Merla et al., 2021), we must consider that poverty risk may simply be too crude a measure. Indeed, using income quintiles, Salin et al. (2024) found a significantly lower likelihood of JPC for the four lower, compared to the highest, income group(s). In the light of the supposed diversification of JPC families, financial feasibility of JPC may be becoming more nuanced across income groups and thus warrants more fine-tuned consideration. The same is true for household size, which in its considered form also captured the presence of, e.g. step-children, and may thus not suffice in reflecting practical restraints to JPC due to a higher number of shared children (Claessens & Mortelmans, 2021).

Notable was the negative association between being partnered and having a JPC arrangement. In the current analysis, the rationale behind this result cannot be entirely unravelled. A potential explanatory factor could be the inclusion of a large number of countries with low JPC prevalence (10 countries under 10% JPC), which negatively skew its association with being partnered. It may be that in countries where separation and JPC have become more normatively accepted, the benefits in terms of finding and spending time with a (new) partner have become more closely intertwined with sharing childcare than in countries with a continuing normative disapproval of separation and low incidence of JPC (Kalmijn & Uunk, 2007). This line of reasoning could also be extended to the lack of association between JPC, life satisfaction and trust in others. While in line with recent studies that indicate a disappearance or even reversal of the positive relation between sharing care of children and parental life satisfaction (Augustijn, 2023b; Riser et al., 2023), additional analyses for individual countries (not reported) suggested a positive association for specific countries. This could lead to a ‘neutralization’ of results and thus again urges more elaborate country-comparisons using the EU-SILC module. This is further supported by our finding that having low trust in others was negatively related to JPC in the model excluding country-level characteristics; without accounting for the larger (legal and social) country context, individual experiences and convictions may thus be more influential in opting to share childcare with an ex-partner.

The country-level indicators showed that, as expected, higher levels of gender equality in a country are associated with higher JPC prevalence, while later introduction of JLC legislation linked to lower JPC levels—even when controlling for the proportion of non-intact families and female labour market participation in a country. As stated above, countries with higher GEI scores are more likely to have systems in place that support equal work and care division in households, thus lowering the threshold and increasing the motivation for JPC take-up (DiFonzo, 2014; Westphal et al., 2014). Meanwhile, countries that introduced JLC from 2000 onwards have lower JPC levels than those with earlier JLC legislation, suggesting that later (legal) normalization of joint parental responsibilities of children coincides with lower levels of JPC. As such, and through our inclusion of two country-level controls and four additional countries, our results support and validate the findings of Salin et al. (2024). Furthermore, grouping countries by their distinct period of JLC introduction proved to be useful for the identification of country-level differences in JPC and thus serves as a step-up for future comparisons using the EU-SILC module. For example, we currently lack insight into whether and how GEI and the introduction of JLC legislation function differently in various country contexts or regions. Meanwhile, the proportion of intact families and female labour participation, which showed no relation to JPC in our analysis, potentially serve as catalysts of JPC in earlier stages of its normalization, and less when JPC has become the norm. More detailed country-level comparisons would aid to better understand the explanatory power of these contextual factors vis-à-vis JPC prevalence.

In the light of the proposed future research perspectives, some limitations must be acknowledged. Firstly, the scope of the analysis was bounded by the limited number of indicators provided by EU-SILC on the child level, restricting elaboration on, e.g. the negative association between child health and JPC. Other variables of importance for the tenability of JPC could therefore also not be accounted for, such as distance between parental households, satisfaction with the custody arrangement, pre-separation childcare responsibilities and the quality of interpersonal relationships within the family system. Second, we lack information on the other parent (living outside of the household). As more equal financial resources, work time and care responsibilities pre-separation tend to increase the likelihood of JPC (Claessens & Mortelmans, 2021), the relative ratio of these characteristics would be useful for understanding whether and how the (implicit) power division between ex-partners affects JPC prevalence (potentially differently) across countries. Third, the EU-SILC module’s cross-sectional design restricts causal inferences and limits us to associative statements. As such, it is important to keep in mind that a JPC arrangement in itself can affect employment (and resources), health, life satisfaction and trust in others, and increased prevalence on the country-level can both be a catalyst as consequence of changing social and legal norms.

Despite these restrictions, and especially amidst the conceptual, methodological and contextual heterogeneity of existing JPC studies, the EU-SILC module provides a currently unique opportunity for comprehensively mapping and comparing the prevalence and correlates of sharing care across Europe. We draw three distinct conclusions from the current study. First, along with diversity in JPC, sole and main residence across Europe, we noted a similar father- to mother-residence ratio in countries with high and low incidence of JPC. This warrants consideration of how social and legal norms currently regulate the granting of custody to mothers versus fathers in diverse country contexts. Second, our results concur with existing research that highlights the importance of understanding JPC prevalence from a multi-level perspective and underline the significance of the country context in understanding JPC uptake. Finally, while there appears to be a commonality in the association of certain individual socio-economic and socio-demographic characteristics with JPC, the potentially different nature of other (e.g. psychosocial) aspects across countries remains unclear. We therefore underline the need for (and strongly encourage) studies that more elaborately assess country differences, especially in the light of varying social and legal norms surrounding JPC. Exploring different configurations of country groupings could not only aid in better understanding which country-level factors guide the uptake and normalization of JPC, but also allow for more in-depth considerations of how individual-level factors come into play in different country settings. This is a vital next step in unravelling the true nature—and thus potential outcomes of—the diversification of JPC families across Europe.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

EC was mainly responsible for the elaboration of the introduction and background, whereas DM prepared the data and conducted the analyses. Both authors contributed to the interpretation of the results and the dissemination of the results, discussion and conclusion and approved the final manuscript.

Funding

Funded by the European Union (ERC, Project 101054654—SINGLETON). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency (ERCEA). Neither the European Union nor ERCEA can be held responsible for them.

Data Availability

The dataset analysed during the current study is available via https://ec.europa.eu/eurostat/web/income-and-living-conditions/database

Declarations

Conflict of interests

There are no competing interests related to this publication.

Ethical Approval

Not applicable.

Consent for Publication

Not applicable.