1. Introduction:

This study aims to provide insight in the extent to which formerly cohabiting parents in the Netherlands comply with the Promotion of Continued Parenting and Proper Divorce Act (Staatsblad 2008, 500). The primary aim of this law, which came into force on 1 March 2009 (Staatsblad 2009, 56), is to encourage equal parenthood after divorce and a ‘careful divorce’, in order to mitigate the adverse consequences of divorce for children (Jeppesen de Boer, 2014). A ‘careful divorce’ means that parents are encouraged to think carefully about post-divorce childrearing in advance. To this end, parents need to draw up a parenting plan (Schonewille, 2009). This is a legal document in which parents write down how they will exercise their parental responsibilities after their divorce.
The construction of a parenting plan is mandatory for both married and cohabiting parents (if they have parental authority) with minor children (Staatsblad 2008, 500). However, unlike married couples or parents who have a ‘registered partnership’, Dutch cohabiters do not have to go through legal proceedings when they separate (Schonewille, 2009). Their falling outside the legal system implies that the requirement of a parenting plan is not enforceable for formerly cohabiting parents. Only when they bring a conflict regarding parental authority to court it is regulated by law that these parents have to present their parenting plan (Tomassen-van der Lans, 2015). Even though the judge is required to demand a parenting plan in the case one is absent, previous research has shown that this rarely happens (Tomassen-van der Lans, 2015). It therefore remains to be seen how many cohabiting parents actually create a parenting plan and when they do not, whether they lay down their agreements in some other document, only make verbal agreements or perhaps even make no arrangements at all.
In this study we examine to what extent formerly cohabiting parents have parenting plans and if not, what other arrangements they make. Additionally, we examine which parents are more likely to draw up a parenting plan. So far, little is known about former cohabiting parents because previous empirical studies about parenting plans focused on former married parents (Tomassen-van der Lans, 2015; Smits, 2015). More knowledge regarding former cohabiting parents is especially relevant when we consider that cohabiters have a higher risk of union dissolution than married couples (De Graaf, 2005; Liefbroer & Dourleijn, 2006; Amato, 2010), which makes it more likely that children born into a cohabiting family will experience the dissolution of their parents’ relationship than children born to married couples (Manning, Smock, & Majumdar, 2004). Moreover, if many cohabiting parents do not lay down their agreements in a parenting plan or other written document – as parenting plans are not enforceable for this group –, these parents have no legal document to fall back on in case of future (legal) disputes. To the extent that parenting plans may mitigate the adverse consequences of parental separation for children – although research examining the effects on children’s wellbeing is to our knowledge virtually absent – , the absence of such plans means that children of cohabiting parents are particularly vulnerable when their parents separate. This may all the more be true if it turns out that many cohabiting parents do not make any agreements at all about post-separation childrearing responsibilities. An examination of whether cohabiting parents create parenting plans, other (written or verbal) or no arrangements, shows whether cohabiting parents and their children are indeed a vulnerable group. Studying which cohabiting parents are more likely to construct a parenting plan, shows which cohabiters, and their children, are particularly at risk. Some cohabiters may be more inclined to create parenting plans or other written arrangements than others, for instance because they have more knowledge about legal issues. In this study, we examine whether the chances of creating a parenting plan and other arrangements depend upon characteristics of children (e.g., their age), parents (e.g., their educational attainment) and the former relationship (e.g., parental conflict). We use the New Families in the Netherlands survey (NFN) (Poortman, Van der Lippe & Boele-Woelki, 2014), a representative large-scale survey of Dutch divorced and separated parents, to answer our research questions. This dataset is unique because it contains a large sample of cohabiting parents who separated after the 2009 divorce law came into effect. Another strength of the NFN is that it includes information about how parents dealt with their agreements about post-separation child rearing.

2. Theoretical background:

We distinguish three types of factors that we assume are important for the likelihood of creating a parenting plan: child characteristics, parent characteristics and relationship characteristics.

Child characteristics

Children’s development requires continuous care (Linker, Stolberg, & Green, 1999). In addition, children may benefit from regularity and continuity (Lamb, Sternberg, & Thompson, 1997). A parenting plan may ensure such continuity as clear agreements can foster stability and structure in the post-separation period and may prevent that parents leave things to chance or ad-hoc decision-making. We expect that parents will be more likely to create a parenting plan when the child’s need for stability and structure is higher. Because children who face psychological and/or physical problems and very young children likely need more structure and stability, we hypothesize that the likelihood of a parenting plan is higher for younger children and children with psychological or physical problems.

Parent characteristics

The chance that parents construct a parenting plan depends upon their legal knowledge. At the very least, parents must have a basic knowledge of the law, i.e. they must be aware of the mandatory nature of parenting plans. Couples may be unaware of existing family laws and policies (Perelli-Harris & Gassen, 2012) or their implications (Boele-Woelki, Curry-Sumner, Jansen, & Schrama, 2007), which will lower the likelihood that these parents create a parenting plan.
We assume, first of all, that the parents’ educational level is related to their knowledge of the law. People have to collect the necessary information and fully understand the implications of the law, and higher educated people probably need less time and effort to collect and understand this information (Giesen, 1999). Research indeed shows that cohabiters who legalize their union (i.e. registered partnership) or draw up a cohabitation contract tend to be higher educated than those who do not (Poortman & Mills, 2012). We therefore expect that higher educated cohabiters are more aware of the legal procedures, and thus more likely to draw up a parenting plan.
Second, we assume that parents’ knowledge and awareness of the law depends upon whether or not they consulted a legal practitioner. Legal practitioners engage with the law and know the law as professionals (Constable, 2014). Because legal practitioners are aware of the mandatory nature of the parenting plan, it is likely that couples who consult a lawyer, notary or mediator more often construct a parenting plan than couples who do not consult a legal practitioner.
Besides legal knowledge, personal problems of parents (e.g., psychological problems, addiction) may also affect the likelihood that a parenting plan is constructed. Such problems may restrict communication between parents, thereby reducing the likelihood that a parenting plan is constructed since its construction requires extensive communication. Alternatively, parents may have a greater need for a parenting plan in case of personal problems. Because such problems can affect the respective parent’s childrearing competencies and decrease the extent to which parents trust each other regarding the way the other partner is raising the child (Maccoby, Buchanan, Mnookin, & Dornbusch, 1993), parents may want to clearly write down who is responsible for what. A parenting plan offers the opportunity to do so and makes it easier to hold the other parent to his/her end of the bargain. We thus expect that personal problems of either parent increase the chance of having a parenting plan.

Relationship characteristics

Parents who opt for residential co-parenting, whereby the child alternates evenly between the parents’ homes (Nikolina, 2015), may be particularly inclined to create a parenting plan. Because these parents have to communicate and inform each other more often than parents in sole-custody arrangements, residential co-parenting makes coordinating the child care a greater challenge (Van der Heijden, Poortman & Van der Lippe, 2016). To reduce the complexities of residential co-parenting, parents may want to make clear agreements in advance by creating a parenting plan.
The geographical distance between the former partners’ homes may also affect the likelihood that a parenting plan is constructed. At least one of the ex-partners has to find a new place to live after the separation (Mulder & Malmberg, 2011). This move can cause communication problems, since communication about child-related decisions is harder for separated parents when they are in different locations (McBroom, 2011). If their homes are relatively far away from each other, communication may be particularly problematic. Longer distances make it harder for parents to visit each other (Cooksey & Craig, 1998), and instead of communicating directly and discussing child-related issues face-to-face, these parents will have to rely on the telephone or e-mail. Long distances may therefore increase the need for clear arrangements. We expect that the farther away ex-partners live from each other, the more likely it is that they will create a parenting plan.
Finally, parental conflict may be another important relationship characteristic affecting the likelihood for a parenting plan. In general, separation is an unpleasant event for both former partners and in most cases there is at least some amount of conflict immediately after separation (Kim, 2011). Parents with high levels of conflict may be more inclined to construct a clear and detailed parenting plan, because clear and written agreements leave less room for own interpretation (Klein Woolthuis, Hillebrand, & Nooteboom, 2005) and ensure that the other parent will keep to the agreements. Severe conflict may, however, make it impossible for parents to create a parenting plan. Parents should at least be able to communicate and negotiate with each other. Because hostility between ex-partners is likely to hamper communication (Linker, Stolberg, & Green, 1999), we assume that it is almost impossible to reach agreement about the post-separation period in high conflict situations. The arguments we pose concerning the role of conflict are contradictory, but it is quite possible that both are true to some extent. If so, we expect that the level of conflict will positively affect the likelihood of a parenting plan being created, up to a certain point. After this point, the level of conflict is too severe for communication and negotiation, reducing the likelihood that a parenting plan will be created.

3. Method

Data

In this study, we analyze data from the New Families in the Netherlands (NFN) survey (Poortman, Van der Lippe & Boele-Woelki, 2014) which were collected in 2012 and 2013. The survey involved formerly married and cohabiting parents from the Netherlands who ended their relationship in 2010, after the new divorce law came into effect in 2009. The sample was randomly drawn by Statistics Netherlands. Both former partners were contacted by letter and invited to complete a web survey covering their legal arrangements and other characteristics before, after and during their separation. In about one third of the couples contacted, both former partners participated in the survey. Concerning the overall response rate, about 39 percent of the parents who were approached participated, with the response rates being higher for formerly married (43 percent) than for formerly cohabiting parents (32 percent). In terms of households, i.e. when at least one parent in the former household participated in the survey, the response rate was higher (58 percent), again with formerly married households being more likely to respond (63 percent) than formerly cohabiting households (48 percent). These response rates are relatively high for a web-based survey in the Netherlands, a country known for its relatively low response rate (De Leeuw & De Heer 2001). Moreover, the response rates are quite high considering that the respondents who were approached were recently separated individuals in the middle of the post-separation period and might therefore have been less willing to participate.
Former cohabiters, men (particularly those with young children), younger persons, people of non- Western descent, people with low incomes and those on welfare were underrepresented, whereas men with children officially registered at their address were overrepresented. The group of former cohabiters was particularly selective: also parents from the most urban areas and men with fewer than two children were underrepresented (Poortman, Van der Lippe & Boele-Woelki, 2014). In addition, non-responders may have been less willing to participate in the survey owing to severe problems and conflicts in the post-separation period. If so, we may overestimate the percentage of people who create a parenting plan and underestimate the effects of problems and conflict. Caution is therefore advised in interpreting the results.
The total sample consisted of 4,481 recently divorced/separated respondents who had children together (for former cohabiters this is determined by legal parenthood). Since the focus of this paper is on former cohabiters, 3,177 formerly married respondents and 179 respondents who had registered their partnership were excluded. Even though the sample was based on respondents who had separated after 2009, we excluded another 111 respondents who reported separating before March 2009 (the month and year the new divorce law became effective). Most variables included in the analyses had some missing values, so another 44 respondents were excluded due to list-wise deletion. Because our dependent variable is at the household level, we randomly selected one respondent from each household when both former partners participated (n = 172) in the survey, to prevent their decisions from being overrepresented in the analyses. Additionally, this also overcomes the problem that not all former partners agree on whether they had a parenting plan or not. However, from these 172 households 84 percent reported the same on having a parenting plan or not (analyses not presented). The total number of cases depends upon the specific analyses we conduct. For the analyses focused on whether or not parents constructed a parenting plan, 798 households are under study. When we analyze which other arrangements (e.g. verbal, no arrangements) are made instead of a parenting plan, the number of households under study is 707. For more details, see below.

Measures

Parenting plan

For our dependent variable we constructed two measures. Respondents were asked ‘Did you draw up a parenting plan in the separation/divorce?’ with ‘yes’ = 1 and ‘no’ = 0. In a note respondents were informed about what a parenting plan is (“A parenting plan is a document in which the agreements about children (for example with whom the children will live or child support) are recorded in writing. The parenting plan is sometimes included in the separation/divorce agreement. If this is the case, select “yes””). With this question we constructed our first measure, that is a dichotomous variable indicating whether a parenting plan is constructed or not. For our second measure, we examined what sort of arrangements were made when a parenting plan was not constructed. In case respondents indicated to have no parenting plan, they were asked whether arrangements concerning the children had been documented in some other way and, if so, in what sort of document. The response alternatives were ‘Don’t know,’ ‘Not applicable: no arrangements concerning the children were made,’ ‘Not applicable: court- issued ruling regarding the children,’ ‘Did make agreements, but they weren’t recorded in writing (verbal arrangements),’ ‘Recorded in divorce agreement,’ ‘Recorded in divorce petition,’ and ‘Recorded in other document.’ Respondents were allowed to give multiple answers, and 16 in fact did. Most were somewhat contradictory, however, so we decided that the most formal arrangement would be the respondent’s sole answer. For instance, if respondents gave the answers ‘court’ and ‘verbal arrangement’ the respondent was marked as reporting only ‘court’. Some of the answers were given by only a few respondents (as will be shown later on – Table 2), which makes it difficult to compare them. We therefore selected only the three largest groups for comparison, resulting in a variable indicating whether respondents have 1 ‘a parenting plan’, 2 ‘verbal agreements’ or 3 ‘no arrangements’.
In sum, we constructed two dependent variables: 1) a dichotomous variable showing whether or not a parenting plan has been created, and 2) a nominal variable showing whether the former couple had a parenting plan, a verbal arrangement, or no arrangement at all. The descriptive statistics of these variables will be discussed in the results section. Note that for the second measure some cases were excluded (i.e., respondents indicating ‘divorce petition’, ‘divorce agreement, ‘court ruling’, ‘don’t know’ or ‘other document’), which results in a lower number of cases when we examine our second measure of the dependent variable compared to when the first measure is examined.

Child problems

Respondents were asked whether or not the following had happened to one of their children before parental separation: (1) severe psychic illnesses (e.g. disability), (2) behavioral problems (e.g. stealing), and (3) social or psychological problems (e.g. being bullied). We constructed a variable counting the number of problems the children faced, ranging from 0 to 3.

Age of youngest child

The question was ‘What is the age of the youngest child that you and your ex-partner had together or adopted?’ Because this variable indicates the age of the youngest child at the time of the interview, we subtracted the number of years between the separation and the interview. As a result, this variable represents the age of the youngest child at the time of the parents’ separation.

Education

Educational level is divided into ten categories, ranging from 1: ‘Did not finish elementary school’ to 10: ‘Post-academic.’ The respondents were asked to indicate their own and their ex- partner’s highest level of education. Based on this information, we selected the highest educational level of the two.

Contacted a legal practitioner

Respondents were asked ‘Did you and/or your partner use a lawyer during the divorce/separation?’ The response alternatives were ‘yes’ = 1 and ‘no’ = 0. We asked the same question about a notary and a mediator, using the same coding. This resulted in three dichotomous variables indicating whether the parents had contacted a lawyer, notary and/or mediator. It was possible for respondents to have consulted all three types of legal practitioners (only 3,63% did this).

Parents’ personal problems

Respondents were asked whether the following situations had happened to them and/or their ex-partner during their relationship: (1) severe physical illness or handicap, (2) severe psychological problems, (3) violence, drug or alcohol addiction, and (4) trouble with the law (traffic violations excluded). We constructed a variable counting the number of problems, ranging from 0 to 4.

Residential co-parenting

The question was ‘What was agreed on during the divorce or decided by the court concerning with whom the child will live most of the time?’ The response alternatives were ‘With me,’ ‘With my ex-partner,’ ‘With both (about) equally (residential co-parenting),’ and ‘Other’. We constructed a variable indicating whether parents had opted for residential co-parenting (coded 1) or not ( all other alternatives = 0).

Geographical distance between partners

The geographical distance between the partners is measured in minutes. Respondents were asked how many minutes it takes them to travel from their house to the house of their ex-partner (one-way trip). We performed the same analyses with several discrete measures with all sorts of different cut-off points, these showed inconsistent but mostly insignificant effects and did not altered the results. Therefore, we decided to use the linear measure of geographical distance in our analyses.

Level of conflict

Respondents were asked to what extent there were conflicts between the ex- partners concerning things that had to be arranged during the divorce/separation. The response alternatives were ‘no conflicts’ = 1, ‘few conflicts’ = 2, ‘somewhat frequent conflicts’ = 3, and ‘very frequent conflicts’ = 4. Since a non-linear relationship was expected, we constructed four dummy variables for each possible answer category.

Control measure

In the analyses we controlled for the duration of the former cohabiters’ relationship. Earlier analyses (not presented here) showed that this variable had a non-linear relationship with the dependent variable. Thus, we measured duration by four dummies (with more or less the same number of respondents per dummy) indicating whether the relationship lasted four years or less, five to eight years, nine to twelve years, and the reference category thirteen years and longer. We checked the correlation between the age of the youngest child and duration. The correlation was moderate (r = .636), suggesting that there are no problems of multicollinearity. The descriptive statistics for all independent and control variables used in the analyses can be seen in Table 1.

Method

To examine to what extent former cohabiters create a parenting plan and what their alternatives are, we used descriptive statistics. To identify the correlates of creating a parenting plan, we estimated a logistic regression model whereby the outcome variable distinguishes between having and not having a parenting plan. We also estimated a multinomial logistic model to capture differences between the determinants of three different post-separation arrangements, namely parenting plan, verbal arrangement and norrangement. For ease of interpretation, we calculated the predicted probabilities of having each type of post-separation arrangement based on our multinomial logistic regression models. These predicted probabilities also more clearly demonstrate the size of the effects for our key variables. As explained before, the number of respondents included in the multinomial logistic model differs from the number included in the descriptive results and the logistic model.

4. Results

Descriptive results

We first examined to what extent former cohabiters create a parenting plan after separation. As can be seen in Table 2, about half of the former cohabiters had drawn up a parenting plan (51,1 percent). We also investigated what sorts of arrangements are made when former partners do not create a parenting plan. The table shows that about a quarter of the former cohabiters (24,3 percent) make verbal arrangements concerning their children. A smaller percentage of parents make no arrangements at all for the post- separation period (13,2 percent). One out of ten former cohabiters chooses from among a wide variety of different alternatives to the parenting plan, with some respondents laying out their child-related arrangements in a document other than one of the options presented (3,9 percent). Although legal proceedings are not mandatory for former cohabiters, the courts nevertheless do rule on some arrangements (3,8 percent). Also remarkable is that some respondents do not know what kind of arrangements have been made (1,6 percent), and some mention options that pertain to married couples only, such as a divorce petition (0,3 percent) and divorce agreement (1,9 percent). This suggests that some people are unaware of the actual agreements they have made or do not know the formal name for the arrangement in question.

Multivariate results

Table 3 (in appendix) displays the results of our logistic regression model concerning whether or not a parenting plan is created. Given the many different alternatives to the parenting plan available to former cohabiters, it is interesting to see whether verbal or no arrangements do indeed offer an alternative to the parenting plan, and for whom. Moreover, certain parents may be more likely to create a parenting plan than make no arrangements but have no preference when asked to choose between a parenting plan and verbal arrangements. We therefore present in Table 4 (in appendix) the multinomial logistic regression, which displays three contrasts: (1) creating a parenting plan versus making verbal arrangements, (2) creating a parenting plan versus making no arrangements, and (3) making verbal arrangements versus making no arrangements. When we compare the results from Table 3 and Table 4, we see that the same correlates are significant, yet Table 4 gives a more nuanced view on who differs from whom.
We start with discussing the effects of the child characteristics. From Table 3 we see that the number of problems a child faces does not significantly affect the likelihood that a parenting plan will be created. When we compare between the different sorts of post-separation arrangements (Table 4), again no effect of the number of problems the child faces is found. From Table 3 we can see that the age of the youngest child affects the likelihood that a parenting plan will be created: the younger the child, the higher the likelihood. The multinomial logistic model (Table 4) shows that while the age of the youngest child is indeed a factor when it comes to creating a parenting plan rather than making no arrangements, it does not influence the decision to create a parenting plan rather than make verbal arrangements. Also, the younger the youngest child, the more likely that parents will make verbal arrangements rather than no arrangements. This indicates that former cohabiters are more likely to make some sort of arrangements (be it written or verbal), rather than no arrangement at all, the younger their child is. These findings are illustrated with predicted probabilities presented in Figure 1. Here we see that parents whose youngest child is four years or younger have a higher probability to construct a parenting plan (67.9%) compared to those with an older child (59.5%). The probabilities that verbal arrangements are made are quite comparable with 22.1% for parents with their youngest child being four years or younger compared to 23.4% for those where their youngest child is older than four year. Parents with an older youngest child do have a higher probability to make no arrangement (17.1%) compared to those with a younger child (9.9%).

When we look at the effects of the parents’ characteristics, we see that the level of education does not help to explain whether a parenting plan is constructed or not. If we look at Table 4, however, we see that higher educated parents are more likely to create a parenting plan than to make no arrangements, and to make verbal arrangements rather than no arrangements. This suggests that higher educated parents are less likely to make no arrangements than other households.
Parents who consulted a legal practitioner are more likely to have a parenting plan as can be seen from Table 3. Table 4 provides us the information that this difference applies to both verbal and no arrangements. Figure 2 illustrates these findings, and clearly shows particularly the probability to construct a parenting plan is much higher when respondents consulted a lawyer (86.7% vis-à-vis 47,9% for those who did not consult a lawyer). Comparable probabilities are shown for those who consulted a notary (83.6%) versus those who did not (59,7%), and respondents who consulted a mediator (85.6% versus 53,1%). Conversely, the probabilities that verbal arrangements are made or no arrangements at all are much higher when respondents did not consult a legal practitioner.

Whether one or both parents face psychological and/or physical problems, does not seem to affect the likelihood that a parenting plan will be created as can be seen from Table 3. Moreover, Table 4 provides the information that this characteristic also does not help to explain why a parenting plan, verbal arrangement or no arrangement is made.
Finally, we discuss the findings regarding the relationship characteristics. Table 3 shows that parents who opted for residential co-parenting are significantly more likely to have a parenting plan. When we look at Table 4 we see that parents with a residential co-parenting arrangement are more likely to create a parenting plan or verbal arrangements as compared to no arrangements. Thus, when parents opt for residential co-parenting, they are more likely to make some form of arrangement (verbal or written in a parenting plan) as opposed to no arrangements. Figure 3 illustrates these findings and shows that parents with a residential co-parenting arrangement have a higher probability to make some sort of arrangement (94%) than those with other residence arrangements (85%).

The geographical distance between the former partners does not affect the likelihood of them creating a parenting plan (see Table 3). We also find no evidence that there is an effect of geographical distance to explain why parents are more likely to make a parenting plan, verbal arrangements or no arrangements, as can be seen from Table 4.
The level of parental conflict affects the probability to construct a parenting plan (see Table 3). Compared to the highest level of conflict during the separation process, all other levels of conflict show a greater likelihood that a parenting plan will be created. Additional analyses (not presented here) show that no statistical differences are visible between the three lower levels of conflict. This suggests that parents are less likely to create a parenting plan only in very high conflict situations. When we compare the different post-separation arrangements (Table 4), we see that the level of conflict affects the likelihood of a parenting plan versus no arrangements: parents are more likely to create a parenting plan than make no arrangements when there is no or little conflict between them, compared to a high level of conflict. However, no difference between creating a parenting plan and making verbal arrangements or between making verbal arrangements or no arrangements is found. Figure 4 clearly illustrates that couples with very frequent conflict are least likely to construct a parenting plan (55% versus e.g. 70% for parents with little conflict) and most likely to make no agreements regarding their children (20% versus 10% for the lowest conflict groups).

With respect to the control variables, we find that parents who had relatively shorter relationships are less likely to draw up a parenting plan (Table 3). A further look shows that people who were together for just a short period are more likely to make no arrangements than to create a parenting plan or make verbal arrangements (Table 4).

Conclusion and discussion

In March 2009 the Promotion of Continued Parenting and Proper Divorce Act came into force, which encourages equal parenting and a ‘careful divorce.’ One of the major differences compared to previous legislation is that both married and cohabiting Dutch parents of minor children are now obliged to create a parenting plan when they separate. However, given the informal nature of cohabitation, it is by no means certain that former cohabiters actually will create a parenting plan. In this study, we examined the extent to which formerly cohabiting parents create parenting plans, and if not, whether and how they settled their agreements about post-separation child rearing responsibilities, for example by making verbal agreements. Furthermore, we investigated which former couples were more likely to create a parenting plan. As far as we are aware, this is the first empirical study to focus on former cohabiting parents – previous studies focused on parenting plans of married parents only (Tomassen-van der Lans, 2015; Smits, 2015). Data from the survey New Families in the Netherlands allowed us to answer our research questions because these data contain a relatively large number of formerly cohabiting parents, who were questioned about their parenting plans and their arrangement otherwise.
We found that half of former cohabiters create a parenting plan. In other words, half of the former cohabiters do not create a parenting plan and therefore do not comply with the legal demands of the 2009 law. We caution that this number may be an underestimation, because it is likely that couples who went through a problematic and conflict-laden separation did not participate in the survey. As these couples will be less likely to have a parenting plan, the percentage of former cohabiters without a parenting plan may be even higher in reality. A likely explanation for the high percentage of cohabiters without a parenting plan is that cohabiters do not have to go through a legal procedure when they want to separate; the legal requirement to construct a parenting plan can therefore often not be enforced. A closer look at those who did not construct a parenting plan furthermore revealed that almost a quarter makes verbal agreements regarding their children and another 13% makes no arrangements at all. The remaining ten percent was quite diverse in their arrangements (e.g. court ruling, other document) or did not precisely know how they settled their agreements about the children. These figures suggest that almost half of formerly cohabiting parents do not lay down their arrangements in writing, which could lead to problems should disputes arise later on. Because they do not have a formal written document to fall back on, any disputes may be more difficult to resolve. Moreover, about an eighth of cohabiting parents report that they made no agreements about their children whatsoever, suggesting that the children from these parents may be a vulnerable group as their parents did not agree upon, for instance, child support payments or visitation schedules.
We also found systematic differences between parents with regard to whether they construct a parenting plan or not, and whether they make verbal agreements or no arrangements at all. Parents with younger children, parents who consulted a legal practitioner, parents with low levels of conflict and those who had been in a long-term relationship more often constructed a parenting plan as compared to any other type of arrangement. Given the diversity in the other types of arrangements, we tried to obtain a more nuanced view by comparing the probability of constructing a parenting plan with that of making verbal arrangements and that of making no arrangements (i.e., the three options that were mentioned most). Results show that involvement of a legal practitioner in the separation process strongly increased the chances that parents constructed a parenting plan as compared to both verbal and no agreements. Legal practitioners are well aware of the 2009 law and the possible repercussions in the longer term when parents do not have a legal document to fall back on. They will therefore likely discourage to rely on verbal agreements or make no arrangements at all. Consultation of a legal practitioner was the only characteristic that increased the chance of a parenting plan in particular, that is, relative to the other types of arrangements. The other characteristics that were found to matter, differentiated between making some kind of arrangement, be it a parenting plan or verbal agreements about the children, and making no agreements at all. More specifically, parents with younger children, with a residential co-parenting arrangement, a higher education, low to medium conflict and those who had been with their ex-partner for more than four years, were more likely to have made agreements about post-separation childrearing responsibilities, be it verbal or laid down in a parenting plan, rather than not making any agreements. Parents with younger children or a residential co-parenting arrangement probably do not want to resort to ad-hoc decision making or to just wait how things will go after separation, because coordination can be complex for co-parents and younger children in particular need structure and continuous care. People with a higher education may also be less inclined to make no arrangements at all, because they probably are more aware of the importance of making at least some arrangements in case disputes arise. Not surprisingly, parents in a conflict-laden separation are more likely to make no agreements regarding their children, as it may simply be impossible to communicate or negotiate with each other. Also the group of cohabiters who have only been together for a very short time are more likely to make no agreements.
One of the main conclusions is that half of former cohabiting parents do not draw up a parenting plan, and that those who are better informed about the regulations of the 2009 law are more likely to draw up a parenting plan. When parents include a third party with legal knowledge in the process, parents will become aware of the obligation to draw up a parenting plan and thus become more likely to comply with the 2009 law. With this finding we strongly agree with Tomassen-van der Lans (2015), who pointed to the necessity of proper information provision to separating parents. Awareness of the mandatory nature of the parenting plan is a significant requirement for parents drawing up a parenting plan. Creating more awareness of the obligation to draw up a parenting plan among separating parents may lead to higher compliance with the 2009 law and thus to higher percentages of parents actually drawing up a parenting plan. Another conclusion is that about one eighth of former cohabiting parents makes no arrangements at all (not even verbal), and that this especially concerns the children born to lower educated cohabiting parents, and those whose parents had high conflict or were together for only a short period of time. Thus, providing the information of a mandatory parenting plan is particularly of interest to these groups. A next step for future research would be to examine whether or not parenting plans and the other type of arrangements can limit the negative effects of parental separation on children’s well-being.