Introduction: contact and domestic abuse

ABSTRACT

This article introduces the special issue of the Journal of Social Welfare and Family Law on contact disputes and allegations of domestic abuse. It first describes the aims and findings of the International Symposium on Contact Disputes and Allegations of Domestic Violence – Identifying Best Practices at which the papers in the special issue were originally presented. It then outlines the position in England and Wales regarding allegations of domestic abuse in child arrangements cases, highlighting the difference between the ‘law in the books’ and the ‘law in action’. Thirdly, it discusses the research evidence on another prominent international approach to domestic abuse allegations – legislative presumptions against custody or unsupervised visitation/contact for abusive parents. The experience of presumptions in the USA and New Zealand suggests that a similar gap between ‘law in the books’ and ‘law in action’ exists, together with potential problems of legislative drafting. Finally, the article outlines the contributions of the other papers in the special issue to our understanding of international approaches to ensuring safety for children and resident parents in family proceedings where allegations of domestic abuse are raised.

KEYWORDS:

• Child arrangements
• domestic abuse
• international approaches
• Practice Direction 12J
• presumptions against custody
• presumptions against unsupervised contact

The international symposium on contact disputes and allegations of domestic violence – identifying best practices

Post-separation arrangements for children where domestic abuse¹ is a feature of parental relationships is an area of law and practice that has troubled women’s groups and feminist scholars, including the editors of this special issue, for many years. In England and Wales domestic abuse was virtually ignored in legal and family law professional practice relating to post-separation child contact, jeopardising the safety and well-being of children and victim parents, until the landmark case of Re L, V, M, H (Contact: Domestic Violence) [2001] Fam 260, when the Court of Appeal laid down guidelines for courts in such cases. Since then, we have seen a ‘cycle of failure’ – demonstrable failure to apply the Re L guidelines, new practice provisions designed to strengthen the guidelines, continuing concerns about implementation of the new provisions prompting further reforms, followed by renewed concerns about implementation.

This led the editors of this special issue to ponder how we could escape the cycle of failure and, in particular, how real and sustained cultural change in legal and professional practices and perceptions could be achieved. It was apparent from international research and literature that other jurisdictions experienced similar issues in family law to those encountered in England and Wales. Rather than focusing on our domestic failings, we wondered what we could learn from the way in which those other jurisdictions managed such cases.

The result of these questions was an International Symposium held on 4–5 May 2017 at the Inner Temple in London. Sponsored by The Honourable Society of the Inner Temple, Queen Mary University of London and Brunel University London, the aim of the Symposium was to identify best practices for dealing with allegations of domestic abuse in family law proceedings concerning post-separation arrangements for children. Experts in the field from Australia, New Zealand, Sweden, Spain, the USA, Canada, England and Wales, Scotland and Ireland were invited to present papers about the measures adopted in their jurisdictions to address this issue and to evaluate how those measures were working in practice. We were particularly interested to hear about the success or otherwise of specific approaches including presumptions against unsupervised contact with perpetrators of domestic abuse (New Zealand, the USA), legislation prioritising child safety concerns (Australia, Sweden), the explicit framing of domestic abuse as gender violence/violence against women and of children as direct victims of gender violence (Spain), specialist domestic violence courts (the USA, Canada, England and Wales), judicial education (Canada, the USA) and strategies of listening to children affected by abuse (Scotland, Ireland). Our interest was also determinedly socio-legal. Speakers were asked to address not just what their laws or other measures provided for, but how they operated in practice, and what factors helped or hindered their effective implementation. Representatives of key stakeholder groups in England and Wales, including the judiciary, policy makers, government agencies, women’s groups, and academics, attended and participated in the international dialogue about what works to prioritise safety and minimise harm for children and adult victims of abuse in private family law cases, and how models adopted elsewhere might translate (or not) into the local context. This special issue presents most of the papers from the Symposium, providing a rich compilation of international experience on contact and domestic abuse. Where, for one reason or another, Symposium participants were not able to write up their paper, we have included key information from their jurisdictions and presentations in this Introduction and in the Concluding article.

While we had expected to identify a range of ‘good practice’ models from overseas which might prompt a rethink of the approach taken in England and Wales, it is fair to say that the main message of the Symposium was not of the existence of a wealth of better options, but of widely shared common problems, which persist across a range of different legislative and practice configurations. While this suggests that the unsatisfactory state of affairs in England and Wales is by no means unique – and indeed may not be as internationally lagging as we had expected and can be seen to have some virtues in that context – it does mean that there are not many positive alternatives available to be drawn upon for future practice development. Nevertheless, a handful of approaches discussed at the Symposium stood out as offering at least a potential way forward. The two major obstacles to the effective implementation of any measures designed to ensure safety from domestic abuse which emerged clearly from the presentations and discussions are a pro-contact culture in the family justice system and, related to this, the under-resourcing of services and systems to support children, affected parents and perpetrator parents, and to promote judicial and practitioner expertise in understanding and addressing issues of abuse. If these characteristics are present, other initiatives aiming to foster safety are likely to be undermined. Three approaches which can help to achieve change in the underpinning culture and related resourcing are the adoption of a human rights framework for determining contact disputes involving allegations of domestic abuse, the adoption of trauma-informed practice by family justice systems and actors, and the development of integrated services which seek to address issues of domestic abuse in a holistic way, focusing on the needs of the family as a whole rather than solely on the legal matter raised in an application for contact or a protective injunction. These three approaches are the subject of the final article in this special issue.

The Symposium opened with an account of the current situation concerning domestic abuse allegations in child arrangement proceedings in England and Wales, for the benefit of our international audience. This was followed by an analysis of the evidence from the USA and New Zealand on presumptions against contact. We now turn to a discussion of these two issues, before introducing the subsequent articles in this special issue.

Child arrangements and domestic abuse – the state of play in England and Wales

Despite persistent evidence over the last 20 years that domestic abuse is minimised by family courts and professionals dealing with post-separation arrangements for children, and that (mostly) mothers and children may be exposed to unsafe contact arrangements, little real progress has been made. Drawing on the available research evidence and recent case law, we suggest reasons why the framework for dealing with allegations of domestic abuse in child arrangements proceedings is problematic and why the current approach may not be working effectively.

Background

Family law cases involving litigation between divorcing or separating couples over post-separation arrangements for their children are known as ‘child arrangements cases’ in England and Wales. This replaces the previous statutory language of ‘custody’ and ‘access’, and subsequently ‘residence’ and ‘contact’. ‘Child arrangements orders’ are defined as orders ‘regulating arrangements relating to…with whom [and when] a child is to live, spend time or otherwise have contact…with any person’ (s 8 Children Act 1989). Section 1 of the Children Act 1989 provides that, when a court is considering making an order, the child’s welfare must be the paramount consideration. Courts are further guided by the ‘welfare checklist’ set out in s 1(3) of the Children Act 1989, which requires courts to ‘have regard in particular’ to a range of factors including the child’s wishes and feelings, their characteristics and needs, the capability of the parents in meeting those needs, and any harm which the child has suffered or is at risk of suffering.

Since the late 1970s family policy and legal decision-making and professional practice in parenting proceedings has been shaped by the strong assumption that children need contact with non-resident fathers for their emotional, psychological and developmental health, leading to what has been described as a de facto ‘presumption of contact’ (Bailey-Harris, Barron, & Pearce, 1999; Hunt & Macleod, 2008). The higher courts have repeatedly emphasised that ‘cogent’ or ‘compelling’ reasons are required to refuse contact and that courts should not ‘give up’ on endeavouring to ensure that contact happens.² This construction of the welfare principle, together with the perception that ‘conflict’ and contested court hearings are ‘bad’ for children, has led to a strong imperative by judges and legal and child welfare professionals to encourage agreement for contact (Bailey-Harris, Davis, Barron, & Pearce, 1998; Hunter & Barnett, 2013). Mothers who oppose or seek to restrict contact have been seen as ‘implacably hostile’, i.e. selfish, manipulative, irrational or unreasonable (Harrison, 2008; Kaganas & Day Sclater, 2004) or, more recently, as potentially ‘alienating’ their children against their fathers (Eaton, Jarmain, & Lustigman, 2016).

Despite the strength of the ‘presumption of contact’ and the rarity of refusals of contact orders (less than 0.3% in 2011) (Ministry of Justice, 2012), fathers’ rights groups pressed for a presumption of shared care, asserting that courts were unreasonably excluding fathers from children’s lives. The Family Justice Review in 2011 recommended against any presumption (Family Justice Review, 2011, p. 23). Contrary to this recommendation, in April 2014 the government amended s 1 of the Children Act 1989 to include a presumption that ‘unless the contrary is shown…involvement of [a] parent in the life of the child concerned will further the child’s welfare’ so long as that parent can be involved in the child’s life in a way that does not put the child at risk of suffering harm.

The cycle of failure: child arrangements/contact and domestic violence

UK Government strategies to end violence against women and girls focus on criminal prosecutions, education, and support for victims of violence (HM Government, 2014, 2016). In addition, local authorities take the protection of children from exposure to domestic violence to be a central part of their child protection role (Local Government Association, Cafcass, & Women’s Aid, 2006). However, domestic violence policies in England and Wales rarely address the area of private family law. Marianne Hester has observed that the ‘differences in culture, practice and discourse’ between professionals working in criminal justice, child protection and family law proceedings is so striking that they could be said to occupy three different planets (Hester, 2011).

Yet a large proportion of child arrangements/contact cases in England and Wales take place within a context of domestic abuse (at least 50% – see, e.g. Cafcass & Women’s Aid, 2017; Harding & Newnham, 2015; HMICA, 2005; Hunt & Macleod, 2008; Perry & Rainey, 2007). Despite this prevalence and the well-documented effects of domestic abuse on victims and children, courts and professionals in England and Wales virtually ignored the issue of domestic abuse when considering applications for residence or contact. The pioneering study by Marianne Hester and Lorraine Radford in 1996 revealed that the perceived importance for children of maintaining contact with non-resident fathers led courts and professionals to minimise domestic violence and to focus on persuading mothers to agree to contact rather than on fathers’ behaviour, with the result that very few of the contact arrangements agreed or ordered in cases involving domestic violence were ultimately safe (Hester & Radford, 1996; see also Anderson, 1997; Barnett, 2000; Kaganas & Piper, 1999).

As a result of the research by Hester and Radford and pressure from women’s groups, the Children Act Sub-Committee of the Lord Chancellor’s Advisory Board on Family Law (CASC) published ‘good practice’ guidelines for the judiciary setting out the approach which ought to be taken when domestic violence was put forward as a reason for denying or limiting parental contact (Lord Chancellor’s Advisory Board, 2001). At the same time, the Court of Appeal heard the combined appeals of Re L, V, M, H (Contact: Domestic Violence) [2001] Fam 260. Drawing on a report to the court by expert child psychologists, Drs Sturge and Glaser (2000), the court acknowledged that domestic violence involves a ‘significant failure in parenting’ and laid down guidelines for courts and professionals in contact cases where allegations of domestic violence are made, based on those developed by CASC. These include the requirement to hold a fact-finding hearing on disputed allegations of domestic violence and the imperative to ensure that the risk of harm is minimised and the safety of the child and resident parent is secured before, during and after contact.

Subsequent research indicated that the Re L and CASC guidelines were largely ignored and their application was inconsistent and ‘patchy’ because courts and professionals continued to prioritise contact over children’s and resident parents’ safety. Domestic violence continued to be minimised by courts and professionals and women who raised allegations of domestic violence continued to be viewed with suspicion, disbelieved and treated as obstructive. Fact-finding hearings were inconsistently and rarely held, and even in cases of proven domestic violence, applications for direct contact were very rarely refused; the most common final outcomes were for direct, unsupervised contact (Aris & Harrison, 2007; DCA and DfES, 2004; HM Inspectorate of Court Administration, 2005; Humphreys & Harrison, 2003; Hunt & Macleod, 2008; Perry & Rainey, 2007; Trinder, Connolly, Kellet, & Thoday, 2004; Trinder, Firth, & Jenks, 2009).

Following the publication by Women’s Aid of a report about 29 children who were killed in England and Wales between 1994 and 2004 as a result of contact arrangements (Saunders, 2004), the Family Justice Council (FJC) issued a report which called for a ‘cultural change…with a move away from “contact is always the appropriate way forward” to “contact that is safe and positive for the child is always the appropriate way forward”’ (Craig, 2007, p. 27). The FJC recommended that a Practice Direction should be issued embodying the CASC and Re L guidelines and its own recommendations. In response, a Practice Direction was issued by the President of the Family Division in May 2008 which was revised in 2009 and subsequently incorporated into the Family Procedure Rules 2010 as Practice Direction 12J (PD12J). PD12J now establishes the framework to be followed by courts and professionals in child arrangements cases where allegations of domestic abuse are raised.

The history of PD12J since its introduction bears a striking resemblance to the history leading up to it. Research following its introduction showed that PD12J was not being implemented as intended. Although there were some variations in practices and approaches, judges and professionals tended to see only recent, serious physical violence as being relevant to the issue of contact (Barnett, 2014, 2016; Coy, Perks, Scott, & Tweedale, 2012; Hunter & Barnett, 2013; Thiara & Gill, 2012). Hunter and Barnett (2013) found a marked difference between ‘legalistic’ understandings of domestic abuse, focused on discrete incidents of physical violence (largely held by family lawyers and the judiciary), and social science understandings, which recognise its power and control dynamics. Judges and professionals continued to view women’s complaints about domestic violence with suspicion, perceiving them to be a delaying tactic and/or designed to disrupt the other party’s relationship with the child (Barnett, 2015; Coy et al., 2012; Hunter & Barnett, 2013). Risk was not always adequately assessed because the nature of the abuse was misunderstood (Trinder, Hunt, Macleod, Pearce, & Woodward, 2013), and the risk and welfare assessment provisions of PD12J were not applied properly or at all. Direct contact was invariably agreed or ordered and refusals of contact orders continued to decline (Barnett, 2014, 2016; Coy et al., 2012; Hunter & Barnett, 2013). The most likely type of contact ordered would be supervised or supported, although resort was often had to informal means of supervision by family or friends because of the limited availability of contact services (Hunter & Barnett, 2013). Hunter and Barnett (2013) concluded, therefore, that the ‘cultural shift’ called for by the FJC remained incomplete.

PD12J was revised in April 2014 to bring its provisions and terminology into line with other amendments to child arrangements proceedings that came into effect at that time. The amendments included inserting a new, broader definition of ‘domestic violence’ focusing on coercive control, implementing some of Hunter and Barnett’s recommendations in order to improve protection for children and victim parents, and attempting to make the process more user-friendly for litigants in person (LIPs).

In 2016 Women’s Aid published Nineteen Child Homicides, which updated its earlier report and documented the deaths of a further 19 children killed by their fathers in the context of contact arrangements between 2005 and 2015. At the same time, Women’s Aid launched its Child First campaign ‘to stop avoidable child deaths as a result of unsafe child contact with dangerous perpetrators of domestic violence’ (Women’s Aid, 2016). A Parliamentary Briefing Paper by the All-Party Parliamentary Group on Domestic Violence (2016) and a House of Commons debate on domestic abuse, child contact and the family courts held on 15 September 2016, both highlighted serious concerns about the treatment and experiences of victims of domestic abuse in family courts and made further recommendations for reform. In November 2016 Mr Justice Cobb, having reviewed PD12J at the request of the President of the Family Division, produced a report with proposed revisions to PD12J (Cobb, 2016). The final, revised (and current) version of PD12J came into effect on 2 October 2017, having been somewhat watered down from Mr Justice Cobb’s original draft. PD12J 2017 reinforces the mandatory nature of many of its provisions, clarifies its broad concern with all forms of domestic abuse, and requires courts to record in their orders certain information demonstrating that its provisions have been followed. Whether the latest revisions will overcome the dominant ‘contact at all costs’ approach and the other obstacles to achieving a real cultural shift discussed below, remains to be seen.

In theory: the procedure for addressing allegations of abuse

PD12J sets out in considerable detail a four-stage procedure that courts are supposed to follow when allegations of abuse are raised as an issue in child arrangements proceedings.

First, the issue of abuse needs to be brought to the court’s attention. This is not entirely left to the parties. In all child arrangements proceedings the court’s child welfare service – known as Cafcass (the Children and Family Court Advisory and Support Service) – will undertake initial safeguarding checks with police and local authorities, and separately with the parties, to identify any prior criminal or social services actions or concerns relating to domestic or child abuse. The results of safeguarding checks are supposed to be available to the court at the first hearing, termed the First Hearing Dispute Resolution Appointment (FHDRA). If the results of the safeguarding checks are not available at the FHDRA ‘and no other reliable safeguarding information is available, the court must adjourn the FHDRA until the results of safeguarding checks are available’ and should not generally make any interim child arrangements order in the absence of safeguarding information (PD12J, para. 12).

Secondly, if there are contested allegations of domestic abuse which are relevant to any child arrangements order the court may make, the court must determine as soon as possible whether it is necessary to conduct a fact-finding hearing to adjudicate on the disputed allegations, in order to provide a factual basis for the subsequent welfare determination, including the assessment of risk (para. 16). Where disputed allegations of domestic abuse are undetermined, PD12J prescribes that the court should not make an interim child arrangements order unless it is satisfied that it is in the interests of the child to do so and the order would not expose the child or the non-abusive parent to an ‘unmanageable’ risk of harm (para. 25).

Thirdly, once it is established that domestic abuse has been perpetrated (whether as a result of findings of fact, admissions or otherwise), the court must consider whether to order any expert assessment of any party or the child (including a safety and risk assessment), and whether any party should engage in any intervention (such as participation in a Domestic Violence Perpetrator Programme (DVPP)) as a precondition to any child arrangements order (paras. 33–34). The assessment is often undertaken by a Cafcass officer, but other experts may be instructed.

Finally, a determination should be made of what arrangements will be in the child’s welfare. In making this determination, the court must take into account the abuse found to have occurred, any risk assessment, any harm the child and the resident parent has suffered or is at risk of suffering, a range of factors derived from the original Sturge and Glaser report (Sturge & Glaser, 2000) (including the motivation of the abusive parent in seeking contact and their likely behaviour during contact), the other factors in the welfare checklist, and the need to ensure that future arrangements will be in the best interests of the child, will not expose the child to an ‘unmanageable’ risk of harm, and will not expose the resident parent to further domestic abuse (paras. 35–37).

Additionally, in recognition of the fact that a high proportion of child arrangements orders are made by consent, PD12J requires the court to scrutinise proposed consent orders to ensure they accord with the welfare of the child (para. 6). In this respect, it tries to ensure that courts have sufficient information available to them about (potential) issues of abuse to enable them to determine whether proposed agreements are likely to be safe.

This process, on paper, seems to focus attention appropriately on determining the existence and effects of domestic abuse and to prioritise the safety of children and their non-abusive parents. However, when it comes to implementation, ongoing problems have been encountered at every stage.

In practice: barriers to implementation

The assumption in favour of contact

PD12J operates in the context of the very strong assumption – now reinforced by the legislative presumption – that contact is always in children’s best interests. This includes the notion that contact is always better than no contact, that direct contact is always better than indirect contact, and that unrestricted contact is the ideal to be worked towards. The onus is very firmly on the party seeking to deny or limit contact, and a very high level of risk must be established before contact will be considered potentially harmful. This position is reinforced by case law from the Court of Appeal, which has been at pains to promote contact and so, in many cases, has downgraded and minimised anything other than severe, recent physical violence. In domestic abuse-contact cases reported in 2014–16, the Court of Appeal allowed five out of seven appeals. In all of these cases, a fact-finding exercise had been undertaken at which trial judges made findings of abusive behaviour and refused to order direct contact, but the appeal court overturned their decisions (Barnett, 2017). Mr Justice Cobb’s original amendments to PD12J provided that the statutory presumption of contact contained in s 1(2A) of the Children Act 1989 should not apply ‘[w]here the involvement of a parent in a child’s life would put the child or other parent at risk of suffering harm arising from domestic violence or abuse’ (Cobb, 2016, p. 12). But this was omitted from the final draft which only requires courts to ‘consider carefully whether the statutory presumption applies’ when domestic abuse is alleged or admitted (para. 7) – a missed opportunity to mitigate the strong pro-contact culture.

Pressures to settle

As discussed above, parents in dispute over post-separation arrangements for their children are pressured from every angle to settle rather than pursue their dispute in court. The government expects parents to take responsibility for their own private disputes rather than relying on public resources to resolve them, and reinforced this message in 2013 by withdrawing legal aid for family court proceedings (other than for those victims of abuse who can produce specified evidence of the abuse) while strongly promoting the alternative avenue of mediation (Legal Aid, Sentencing and Punishment of Offenders Act 2012). There is compelling evidence that cases involving domestic abuse are inappropriately accepted into and ‘settled’ in mediation (Barlow, Hunter, Smithson, & Ewing, 2017). And parties to cases that do go to court are still encouraged by the court and lawyers to reach agreement (Barnett, 2016).

While PD12J exhorts courts to scrutinise proposed consent orders to ensure that they do not expose children and non-abusive parents to the risk of harm, its provisions continue to be undermined by the strong pro-settlement culture and lack of resources to adjudicate in most or even a sizeable proportion of cases involving issues of domestic abuse (Hunter & Barnett, 2013). A recent study by Cafcass and Women’s Aid (2017) found that the vast majority of interim and final orders, including in cases where allegations of domestic abuse were raised, were made by consent.

When is abuse considered relevant?

Hunter and Barnett (2013) documented a range of views among judges and lawyers as to when allegations of violence would be considered ‘relevant’ to the question of contact (see also Barnett, 2015; Coy et al., 2012). In particular there remained a widely-held belief that violent men could be ‘good enough’ fathers, that violence against the mother did not necessarily have any impact on the children, that ‘historic’ violence which had occurred earlier in the relationship, when children were too young to remember, or which had been succeeded by apparently ‘successful’ contact could be disregarded, or that only severe physical violence directed at or in the presence of the children was relevant.

As noted above, the 2014 revisions to PD12J broadened the definition of domestic abuse to emphasise coercive and controlling behaviour, and also reinforced the fact that children may suffer psychological or emotional harm from living with violence and abuse and may also be harmed indirectly ‘where the domestic abuse impairs the parenting capacity of either or both of their parents’ (para. 4). Nevertheless, recent case law shows some judges continuing to dismiss ‘historic’ violence and focusing solely on physical violence. For example, in Re V (A Child) (Inadequate Reasons for Findings of Fact) [2015] EWCA Civ 274, McFarlane LJ emphasised the fact that all but one of the allegations took place prior to the child’s birth or when he was still an infant despite an ongoing pattern of coercive and controlling behaviour. For this reason, he said:

most, if not all, of the allegations were not in fact relevant to whether or not young T in 2012/2013, who had been having contact with his father, including staying contact, at whatever regularity, could carry on having contact to this father in the future. [37]

Yet Cafcass and Women’s Aid point out that ‘contact taking place before proceedings and [by] consent may not always equate to an “agreement” about contact and may instead be indicative of a context of coercion or fear’ (Cafcass & Women’s Aid, 2017, p. 3). Similarly, in Re K (Children) [2016] EWCA Civ 99 the trial judge dismissed as irrelevant an incident of physical violence that took place five years earlier on the basis that there was ‘no evidence of recent physical violence between father and mother’ (at [13]), failing to contextualise the earlier violence within his findings of continuing controlling, aggressive behaviour. When allegations considered insufficiently serious or recent are ‘weeded out’ and disregarded, the welfare determination then proceeds as if those allegations did not exist.

Other pressures not to hold a fact-finding hearing

The pre-2014 research suggested that fact-finding hearings were disfavoured because they introduced delays in the proceedings, increased ‘acrimony’ between the parties, used up scarce resources, and were, in any event, to little effect because findings of abuse carried relatively little weight in the welfare determination (Barnett, 2014, 2015; Hunter & Barnett, 2013). Thus, there was a tendency to avoid fact-finding hearings if possible, but scant evidence to suggest that allegations of abuse were dealt with at the final, welfare hearing. These pressures on court time and judicial resources have only intensified since 2014, coupled with a new procedural emphasis on judges managing cases tightly and concluding them expeditiously (Practice Direction 12B – the Child Arrangements Programme). This was seen recently, for example, in the case of Re T [2016] EWCA Civ 1210, in which, after lengthy delays, the trial judge discharged orders for a fact-finding hearing on the basis that proceeding with the hearing was no longer either necessary or proportionate. As the Court of Appeal pointed out, however, rather than moving to a composite hearing at which both the contested allegations of violence and the other welfare issues could be dealt with together, the judge had simply jettisoned the allegations and treated them as no longer relevant to the question of welfare, a decision that was held to be wrong in principle.

While there has been no systematic review of court files since the advent of PD12J to determine how often fact-finding hearings are held, a recent small-scale study by Cafcass and Women’s Aid of 216 files found 62% contained allegations of domestic abuse but a fact-finding hearing was held in only five cases (2017, p. 10). Nor is there any evidence to suggest that contested allegations are routinely dealt with at composite hearings instead. PD12J 2017 provides detailed guidance as to how a court should determine whether it is necessary to engage in the fact-finding exercise, and requires courts to record in the order their reasons for determining that a fact-finding hearing is not necessary. The list of potential reasons, however, remains a long one.

The focus of fact-finding on ‘incidents’ of violence

Separate fact-finding hearings have their origins in child protection proceedings, where they were introduced to facilitate the process of determining whether the ‘threshold criteria’ for the court’s intervention are met – that is, that the child has suffered or is at risk of suffering significant harm due to lack of reasonable parental care (Children Act 1989 s 31). Courts were encouraged to consider ‘whether or not there were questions of fact within a case which needed to be determined at an early stage’ (Wall, 2010, para. 11). Bracewell J in Re S (Care Proceedings: Split Hearing) [1996] 2 FLR 773, explained that cases suitable for such ‘split hearings’ ‘would be likely to be cases in which there is a clear and stark issue [in dispute], such as sexual abuse or physical abuse’ (at 775). From their inception, therefore, fact-finding hearings are a product of and are underpinned by a legalistic, incident-based approach to allegations of abuse.

The practice also developed of requiring the use of so-called Scott Schedules for fact-finding hearings, which are itemised tables setting out the dates and brief descriptions of the specific allegations the alleged victim seeks to prove, together with the alleged perpetrator’s response. These schedules can have the effect of compelling victims to construct and articulate the abuse they have sustained in a disaggregated and decontextualised way. Furthermore, the pre-2014 research found that many judges, in an effort to contain the scope of fact-finding hearings, would limit the number of ‘incidents’ that could be included in the schedule, and anecdotal evidence suggests that this practice continues (Barnett, 2015; Hunter & Barnett, 2013). At best, then, findings would be made on a limited number of discrete incidents of physical violence, which meant that the full extent of the risk posed to the victim and child was minimised or invisible. The 2014 revisions to PD12J attempted to counter the incident-focus of fact-finding by specifically directing the court to consider ‘what evidence is required in order to determine the existence of a pattern of coercive, controlling or threatening behaviour, violence or abuse’ (para. 19).³ Unfortunately, at a late stage of the drafting, an amendment was added so that the court is also required to consider whether ‘the key facts in dispute can be contained in a … Scott Schedule’ (para. 19).

A further problem with fact-finding hearings is that the alleged victim has the burden of proving the abuse. So if the judge cannot decide which party is telling the truth, and there is no ‘independent’ evidence to assist, they can simply decide that the victim has not proved her case. In Re V (Inadequate Reasons for Findings of Fact) [2015] EWCA Civ 274, for example, the trial judge found that the mother had failed to prove that the father had bombarded her with threatening calls and messages. This, in turn, led the Court of Appeal to highlight the ‘historical’ nature of the findings that were made against the father, which was one of the reasons why his appeal was allowed. Victims’ ability to prove the abuse they have sustained may also be impeded by the suspicion and disbelief with which women’s allegations of abuse are met, and the inability of courts and professionals to understand the effects of abuse on women, whose coping strategies can include ‘dissociating themselves from the violence, “forgetting” about abuse, retaining vague and sketchy memories of violent incidents, [and] minimising the seriousness of the violence’ (Hunter, 2006, p. 742). This can be seen in the perception that mothers who are ‘credible’ in their testimony should be able to provide a coherent narrative. In Re V (Inadequate Reasons for Findings of Fact) McFarlane LJ unpicked the trial judge’s findings against the father by comparing what the mother had said in her Scott Schedule against her oral evidence and that of her witnesses and in so doing, found inevitable inconsistencies and contradictions which undermined her credibility.

Skipping or truncating the risk assessment/intervention stage

The research studies and case law on PD12J also suggest that where domestic abuse is found to have occurred, there is then considerable pressure to move the case on – ideally to continue or restore contact – as quickly as possible (Hunter & Barnett, 2013; Trinder et al., 2013). For example, judges may dispense with expert risk assessments and make their own (minimising) assessment of the effect of the abuse and the degree of ongoing risk posed by the father (Women’s Aid, 2016). In Re W [2012] EWCA Civ 528, the judge had made significant findings of sustained domestic abuse by the father at a fact-finding hearing but had then refused an application for a report by Cafcass and a psychological assessment of the father, stating that he was satisfied that the father had been ill but was now much better and that further assessment was therefore unnecessary. His decision that contact should proceed was overturned on appeal, with Black LJ making clear that the assessment of the father’s mental state and the risk of further violence was a matter on which the judge required further professional assistance (see also Re F (Children) [2015] EWCA Civ 1315). Additionally, risk assessments may be obtained from unqualified sources, leading Women’s Aid (2017) to recommend that PD12J be amended to require that in all cases where domestic abuse has occurred, the court should obtain a safety and risk assessment conducted by a specialist domestic abuse practitioner working for an appropriately accredited agency. This proposal was not adopted in the 2017 revision of PD12J, but in any event, any such assessment may fail to capture the entirety of the risk if findings have been made which reflect only a fraction of the whole dynamic of a coercive and controlling relationship. The recent case law indicates that the appellate courts focus almost solely on the risk of physical violence (in some cases only to the child). For example, in Re M (Children) [2013] EWCA Civ 1147 the Court of Appeal allowed the father’s appeal because the trial judge failed ‘to adequately address why the children’s safety and the management of the mother’s anxieties cannot be achieved under any circumstances of supervision’ (per Macur LJ at [19]).

As far as interventions for perpetrators are concerned, Hunter and Barnett (2013) found that courts were disinclined to order fathers to undertake DVPPs because they took ‘too long’ to commence and to complete, a view shared by some family lawyers (Barnett, 2014). Anger management courses or parenting programmes might be preferred because they were more available and quicker, despite their acknowledged inefficacy in addressing the risk of future abuse (Iwi & Newman, 2015). The 2014 revisions to PD12J included specific acknowledgement that ‘acceptance on a DVPP is subject to a suitability assessment by the service provider, and that completion of a DVPP will take time in order to achieve the aim of risk-reduction for the long-term benefit of the child and the parent with whom the child is living’ (para. 34). As well as the low number of fact-finding hearings noted above, however, the Cafcass and Women’s Aid study found only five referrals to DVPPs among their file sample (2017, p. 10).

Decision-making: the welfare stage

As discussed above, where domestic abuse is found to have occurred, PD12J requires courts and risk assessors to address the Sturge and Glaser (2000) factors now set out in paras. 36 and 37. The pre-2014 research found an inconsistent application of these factors by Cafcass officers and the lower courts (Barnett, 2014; Coy et al., 2012; Hunter & Barnett, 2013) and recent case law indicates a continuing failure to refer to or apply these provisions, despite the Court of Appeal repeatedly stressing that the discipline in these paragraphs offers the surest guide to correct decision-making. For example, in Re P [2015] EWCA Civ 466, the Court of Appeal overturned the orders for unsupervised contact made by the Recorder who had failed to obtain a risk assessment or to apply para. 36 to the established history of domestic abuse. In Re F (Children) [2015] EWCA Civ 1315 the Court of Appeal allowed the mother’s appeal in a case where the trial judge completely ignored his own findings concerning the father’s violence, and focused only on the father’s application for a residence order (which he granted), not even considering the mother’s contact application. The Court of Appeal were highly critical of the trial judge for failing to follow PD12J or to apply the welfare checklist.

However, there is considerable reluctance even by the appellate courts to hold fathers to account for their abuse or to require evidence that they have acknowledged its impact on their families and sought to make amends. In Re K [2016] EWCA Civ 99, the Court of Appeal castigated the guardian and the trial judge for requiring the father to apologise to the children and to the mother for his behaviour and demonstrate that he had changed, instead of commending them for what appears to have been a proper application of PD12J. Barnett found that it could be very difficult to persuade courts that a father seeking contact ‘is motivated by anything other than a desire to see the children’ (2014, p. 450). Participants in her study reported judges readily accepting expressions of contrition at face value and expressing sympathy for abusive fathers.

The pre-2014 research indicated that the typical response to findings of domestic abuse was a period of supported or supervised contact, with the expectation that this would eventually move on to unsupervised contact (Barnett, 2014, 2016; Hunter & Barnett, 2013). The more recent, small-scale Cafcass and Women’s Aid (2017) study found that the most common order made when allegations of abuse are raised is for unsupervised direct contact (39%). Supervised or supported contact was ordered in 11% of such cases. Although Cafcass officers recommended either indirect or no contact in nearly a quarter of cases in which domestic abuse was alleged,⁴ no direct contact was ordered in only 2%, with no order being made in 12% (and final disposal being unknown in 31%).

The quality of supervision was also an issue raised in the pre-2014 research, with some courts preferring supervision by family members to the use of child contact services, caused partly but not entirely by the limited availability of supervised contact centres (Hunter & Barnett, 2013). Responding to concerns raised by Women’s Aid about this apparently ongoing problem, PD12J 2017 states that: ‘Where a risk assessment has concluded that a parent poses a risk to a child or to the other parent, contact via a supported contact centre, or contact supported by a parent or relative, is not appropriate’ (para. 38). What effect this provision will have, or whether it will be circumvented by consent orders and the lack of risk assessment, remains to be seen.

Litigants in person

The withdrawal of legal aid for private family law proceedings from April 2013 has resulted in a significant increase in litigants in person in the family courts (Ministry of Justice, 2017). The forensic and adversarial structure of raising allegations, fact-finding, risk assessment and welfare determination is particularly difficult for litigants in person (Trinder et al., 2014). This creates another disincentive to fact-finding, and further opportunities for issues of violence and abuse to be disregarded (Barnett, 2016). The 2014 amendments to PD12J suggest that a fact-finding hearing ‘can be an inquisitorial (or investigative) process’, and that ‘the judge…should be prepared where necessary and appropriate to conduct the questioning of the witnesses on behalf of the parties, focusing on the key issues in the case’ (para. 28). However, many judges remain reluctant to take on this role (Corbett & Summerfield, 2017). Women’s Aid (2017) raised concerns about the prevalence of fathers acting in person and using the opportunity to perpetuate their abuse of mothers by directly cross-examining them in family court proceedings. Abusive cross-examination is barred in criminal trials for sexual assault, but not in family proceedings (House of Commons, 2017). Proposed legislation which would have introduced a very limited ban on direct cross-examination by alleged perpetrators of abuse in family courts was abandoned due to the June 2017 general election. Instead, Practice Direction 3AA to the Family Procedure Rules, which came into effect on 30 November 2017, simply requires courts to consider making directions about the way in which a ‘vulnerable witness’ (which includes victims of domestic abuse) may be cross-examined (para. 5.5).

In summary, while there is always room for improving the detail of PD12J, the problems discussed above suggest that simply attempting to fine-tune it is unlikely to achieve the cultural change called for by the Family Justice Council 10 years ago. The gap that exists between the law on paper and the law in action suggests that the barriers to a real shift in judicial and professional practice and perceptions need to be addressed at a more fundamental level.

Presumptions against contact

The trajectory in England and Wales could have been quite different had the Court of Appeal in Re L accepted Sturge and Glaser's (2000) recommendation that a finding of domestic abuse should have a clear consequence in raising a presumption against direct contact. After reviewing the effects of domestic abuse on children and setting out the risks created for children of direct contact with an abusive parent, Sturge and Glaser concluded:

[T]here should be no assumption that contact to a previously or currently violent parent is in the child’s interests; if anything, the assumption should be in the opposite direction and the case of the non-residential parent one of proving why he can offer something of such benefit not only to the child but also to the child’s situation (i.e. act in a way that is supportive to the child’s situation with his or her resident parent and able to be sensitive to and respond appropriately to the child’s needs) that contact should be considered. We would go as far as to suggest…a position in which a [parent]…who has been found to have been domestically violent to the child’s parent should need to show positive grounds as to why, despite this, contact is in the child’s interests in order for an application to be even considered. There could be a requirement that that parent sets out how he proposes to help the child heal and recover from the damage done. (2000, p. 10)

Regrettably, the Court of Appeal rejected this recommendation, deciding that established domestic violence should not necessarily be a bar to contact and requiring instead that courts balance the evidence of domestic violence and its effects with the other factors in the welfare checklist, with the results outlined above. Other jurisdictions have enacted presumptions against direct contact. This section examines the operation of these presumptions and finds that they ‘have met with very mixed results’ (Lemon, 2001, p. 602).

The USA

Presumptions against the award of custody to perpetrators of domestic violence began to enter the statute books in the USA during the 1980s in response to the calls of domestic violence advocates to protect women and children from abuse. The first US national policy statement supporting such a presumption was House of Representatives Congressional Resolution 172 passed in 1990 which stated that ‘credible evidence of physical abuse of a spouse should create a statutory presumption that it is detrimental to the child to be placed in the custody of the abusive spouse’.

In 1994, the National Council of Juvenile and Family Court Judges published the Model Code on Domestic and Family Violence (Lemon, 2001, p. 606). Section 401 states that where the court makes a determination that ‘domestic or family violence has occurred’, this ‘raises a rebuttable presumption that it is detrimental to the child and not in the best interest of the child to be placed in sole custody, joint legal custody, or joint physical custody with the perpetrator of family violence’. The American Bar Association as well as the American Psychological Association recommended that states adopt statutory presumptions against custody for abusers and also urged that safeguards be put in place in relation to visitation; unsupervised contact should not be awarded unless the abuser has undergone treatment and can show he is no longer a threat (see Lemon, 2001, p. 607).

Garvin’s (2016) review of the field found that over 40 US States have adopted statutory presumptions and/or other provisions to consider the impacts of domestic violence on children when determining custody. Twenty-four have enacted rebuttable presumptions based on the Model Code, most of which shift the burden of proof and/or guide judicial discretion. However the legislation varies. In some jurisdictions, the presumption refers to all types of custody. In others it operates only in disputes concerning joint custody. Definitions of domestic violence and the types of evidence required to trigger the presumption vary from State to State. The evidence required to rebut the presumption also varies. There is little consistency in the way States deal with counter-allegations of violence, and in States which have provisions favouring the ‘friendly parent’ or presumptions in favour of joint custody, the domestic violence provisions are undermined. Courts’ application and interpretation of the presumptions has also limited their effectiveness.

Presumptions in favour of joint custody, friendly parent provisions and other risks

As Davis, Lizdas, Tibbetts, Murphy and Yauch point out, presumptions against custody in cases of domestic violence do not take effect unless the ‘fact that domestic violence has occurred’ is established. However presumptions in favour of joint custody ‘simply start with a conclusion without any foundational showing whatsoever’ (2010, p. 6). Unless the parent opposing joint custody has the wherewithal and the evidence to challenge the joint custody presumption, joint custody will be deemed to be in the best interests of the child (Davis et al., 2010, p. 7). Challenges to joint custody can also ‘backfire’ under friendly parent provisions embedded in State custody laws. ‘Friendly parent’ provisions require the court to consider each parent’s willingness to support and encourage contact between the child and the other parent. A parent who challenges the joint custody presumption because of concerns about violence will be perceived to be hostile to contact and so fail to meet the friendly parent standard. ‘Consequently, the rebuttal to the JPC presumption works worst when a child needs it most’ (Davis et al., 2010, p. 11).

Even in the absence of friendly parent provisions, victims of abuse may not raise the issue for fear of retaliation, for fear of not being believed or for fear of losing their children in child protection proceedings (Davis et al., 2010, p. 17). Indeed a mother may lose custody because, as a victim of domestic violence, she may be considered less able to provide the emotionally and financially stable home favoured by US courts in applying the best interests of the child test (Conner, 2009; p. 197; Lamprecht, 2011; p. 363; Morrill, Dai, Dunn, Sung, & Smi, 2005; p. 1078).

Definitions of domestic violence

In order to activate the presumption against custody, there must be evidence or proof of domestic violence. Garvin observes that a ‘common critique of statutes addressing domestic violence is that they rely on inadequate legal definitions, which tend to focus on violent acts and threats of physical violence’ (2016, para. 11B). They do not cover emotional, sexual or financial abuse (Lemon, 2001, p. 615) or other forms of controlling behaviour or patterns of abuse (Conner, 2009, pp. 206–7). Critics have also pointed out that statutes tend to exclude all but recent incidents of violence and focus on discrete events, ignoring patterns of coercive control. For example, the presumption in the Florida legislation against shared parental responsibility is triggered only if a ‘parent has been convicted of a misdemeanour of the first degree or higher involving domestic violence’ or has been incarcerated (Florida Statutes Annotated s 61.13 4(c)(2)). ‘Domestic violence’ is defined only as physical or sexual assault, stalking, kidnapping or false imprisonment. So the focus is almost exclusively on serious physical violence leading to a conviction.

Evidence

Some States require that violence has occurred more than once. In others there must be a pattern or history of abuse or at least one serious incident. In terms of standards of proof, in some States, the statute merely provides that there must be a ‘finding of domestic violence’ or ‘credible evidence of domestic violence’ (Lemon, 2001, p. 615), while in others, as in Florida, there must be a criminal conviction.

The difficulties of proving domestic abuse – as something that occurs in private and is often unreported – are well known. In many cases, there is no evidence, or the only witness may be a child whom the mother and her lawyer are very reluctant to call into court (Conner, 2009, p. 184). Even reporting may not help, as ‘Law enforcement involvement… does not guarantee proper documentation of injuries, property damage, or other evidence’ (Conner, 2009, p. 184). In this context, requiring proof of multiple acts of abuse is particularly problematic (Conner, 2009, p. 199).

Rebutting the presumption against custody

Some States leave the requirements for rebuttal to the courts’ discretion while others list specific factors that the court must consider in finding that the presumption has been rebutted. In California, for example, the court must consider:

1) whether the perpetrator has shown that it is in the best interest of the child to be in the custody of that parent; 2) successful completion of a batterer’s program; 3) successful completion of a program for alcohol or drug abuse if found appropriate by the court; 4) compliance with court orders and with probation and parole conditions, if applicable; and 5) whether there has been any further violence. (California Family Code s 3044; cited in Lemon, 2001, p. 618)

As Garvin points out, however, completion of a ‘batterer’s’ program does not necessarily render the abuser safe and the word ‘successful’ is open to interpretation (2016, para. 111A). In Popovich v Newton (No B194324, 2008 WL 116261, California Court of Appeal, 2^nd District, 14 January 2008) the mother claimed that the father’s completion of a batterer’s program could not be described as successful since he continued to deny, minimise and persist with his abusive behaviour. This claim was rejected, the presumption was rebutted and the best interests requirement of contact with the non-resident parent prevailed.

Unintended consequences

States have had to adopt dominant aggressor provisions so as to avoid penalising victims who act in self defence, and to deal with false counter-accusations. Lemon refers to observations by an attorney that perpetrators were filing civil protective orders in order to benefit from the presumption against custody, casting the victim as a perpetrator. They were also strongly defending allegations of abuse in order to avoid the presumption being invoked against them (Lemon, 2001, pp. 635, 665).

Supervised contact as a panacea

Statutory presumptions in the US are largely confined to the award of custody. While the majority of States provide in various ways for supervised visitation for the protection of children (Young, 2005, p. 7 and see fn. 17), there are no presumptions against visitation per se; supervised contact, it seems, is assumed to be safe. By contrast, Sturge and Glaser (2000) did not see supervised contact as the panacea in cases of domestic abuse, but considered it was likely to be useful only to maintain safety, manage contact or support the child where change in the short term was likely (2000; pp. 12–13; see also Perry & Rainey, 2007). In particular, supervised contact does not protect children from the emotional harm of having to have contact with a parent who has previously traumatised them (Robertson et al., 2007, p. 98). As discussed below, heavy reliance on supervised contact in conjunction with the New Zealand presumption was one of the factors that made it unworkable.

Judicial interpretations of the legislation

The US literature is critical of judges who, in applying the presumptions, have shown little understanding of the dynamics of domestic violence and patterns of domination and control. For example, the North Dakota Supreme Court confirmed in Tulintseff v Jacobsen (615 NW 2d 129 (2000) at 134) that domestic violence not resulting in serious injury which was committed three or more years before the proceedings was too remote. In the Massachusetts case of Custody of Zia (50 Mass App Ct 237 (2000)) the court held that, despite two abuse protection orders and a pending charge of assault and battery, the mother’s challenge to the judge’s finding that ‘[t]here is no history of domestic violence perpetrated by or upon either party’ was not persuasive. There was no history or pattern of domestic violence that would preclude an award of custody to the father (at 246). In Brown v Brown (867 P2d 477, Oklahoma Court of Appeal, 1993) there was evidence that the father once shoved the mother roughly up against a doorway, threatened her with violence when he suspected her of infidelity and broke the windows in another man’s car because he believed, wrongly, that the wife was having an affair with him. The court said:

we do not see in this behavior any sort of pattern of abuse; much less do we find any evidence which supports [the mother’s] claim that [the father’s] acts constitute ‘ongoing domestic abuse’… We…construe the phrase ‘ongoing domestic abuse’ to mean abuse which is still occurring, or has occurred with sufficient frequency and recency to give rise to some expectation that it will continue or will recur, and thus will constitute a threat to any child of whom the abusive person is granted custody. As such, ‘ongoing domestic abuse’ is not merely one or two isolated instances of proscribed behaviour. (at [7]-[10])

Judges’ failure to understand the dynamics of abuse also leads them to disbelieve victims’ accounts, for example, if they do not recollect details, contradict themselves or appear nervous. By contrast, perpetrators are often manipulative and confident and their demeanour is interpreted as evidence of truthfulness (Conner, 2009, pp. 174, 177). In addition, according to Lamprecht, ‘There is a long-standing judicial belief that women tend to exaggerate the violence that they have endured in order to manipulate the court into awarding them custody’ (2011, p. 364).

Further, judges also appear to have been implicated in what commentators have described as a backlash against these presumptions. Lemon cites a report by an Arizona lawyer to the effect that judges were refusing to hear evidence of domestic violence and were giving strong preference to joint custody (2001, p. 660). Sometimes they would hear the evidence but then decide cases according to their assessment of the child’s best interests, holding that best interests considerations rebutted the presumption (p. 661). Lemon also refers to a survey conducted in California that revealed resistance by some judges to applying the presumption, partly due to lack of time and resources but also because they disliked having their discretion curtailed. They considered joint custody almost always best, even in cases of domestic violence. They would also try to get parties to agree an outcome by repeatedly ordering them to try to mediate. The presumption was also seen to be having an effect on applications for restraining orders, with judges being increasingly reluctant to grant these and demanding corroboration of abuse (p. 664; see also Garvin, 2016).

Effects of a presumption

Despite all these problems, presumptions do appear to have some positive effects. Morrill et al. (2005) found that in States with a presumption, courts were more likely to award sole legal custody to mothers and less likely to award joint legal custody. However in States with competing legislation such as presumptions in favour of joint custody, those provisions prevailed and joint custody was the common outcome. The researchers also found that, although there was no difference in terms of the proportion of visitation applications granted, there was more likely to be structured or conditional visitation in States with a presumption. Nevertheless, the US experience demonstrates both limited understanding of domestic violence on the part of legislators (and hence unduly narrow drafting of presumptions as well as the existence of contradictory provisions), and judicial interpretation and application of the law in a way that makes the effect of presumptions even less protective. Most US academics writing on this subject recommend better training and education for judges and others working in the legal system. However Morrill et al. (2005) came to the conclusion that whether judges had attended domestic violence training made little difference to their decision-making, although the quality of that training could not be ascertained.

The New Zealand presumption against unsupervised contact

In 1994 Alan Bristol, a New Zealand father engaged in a custody battle with his ex-wife, killed himself and his three daughters. The subsequent inquiry by New Zealand’s Chief Justice into the family court proceedings in the case suggested a need for law reform, and this opportunity was grasped by domestic violence campaigners to enact a version of the US Model Code provision in New Zealand (Robertson et al., 2007, p. 102). The Guardianship Act 1968 was amended in 1995 to include a presumption against the award of day to day care or unsupervised contact ‘where a party is shown to have used violence against a child of the family or the other party in custody or access proceedings’, ‘unless the court is satisfied that the child would be safe during such care or contact’ (Guardianship Act 1968 s 60(3)). The amendments also specified a set of risk assessment factors to be taken into account in determining whether the presumption was rebutted (s 61) and what should occur if allegations of violence were not proved but the court was nevertheless satisfied there was a real risk to the child’s safety (s 60(6); later Care of Children Act 2004 ss 58–61; now repealed).

A 2006 review of experience with the presumption found that those working in the family justice system generally considered that it enhanced children’s safety from the risk of family violence (Perry, 2006, 21). However, the presumption encountered a range of by now familiar problems in implementation. The legislative definition of ‘violence’ was confined to physical and sexual abuse, although in practice, judges were generally willing to take psychological abuse into account as well (Perry, 2006, pp. 6–7). The intention of the legislation was undermined by court orders for parties to attend counselling and mediation, resulting in ‘consent’ orders for contact, with women being advised by their lawyers to agree so as not to appear obstructive of contact (Robertson et al., 2007, pp.88–89, 97). Practical difficulties arose in that courts found themselves with limited information when making decisions about children’s safety, since parents’ written statements formed the only evidence in a high proportion of cases and these were often short on detail (Perry, 2006, p. 11). There were concerns about delays between the making of allegations of violence in an application for custody or a protection order, triggering a suspension of contact, and the making of final orders, with fathers being barred from seeing children for many months (Perry, 2006, p. 10). And the government’s insistence that supervised contact centres should be run on a user pays basis resulted in under-resourcing and fewer and fewer providers willing to offer services (Perry, 2006, p.10). Consequently, families were often left to make their own supervision arrangements, and in many cases, supervision was undertaken by the custodial parent – precisely the situation the legislation had sought to avoid (Perry, 2006; p. 11; Robertson et al., 2007; p. 100). Moreover, in the face of judicial and contact parent objections to supervised contact, this came to be seen as a last resort when children’s safety simply could not be otherwise assured (Perry, 2006, p.11), and as the worst case scenario, making it impossible for mothers to seek less than supervised contact (e.g. indirect or no contact) without being seen as hostile (Robertson et al., 2007, pp. 98–99).

The legislation was not well received by family court judges and judicial resistance included applying standards to determine safety which failed to follow the risk assessment checklist in s 61 (Robertson et al., 2007, pp. 110, 124). One respondent to Robertson et al.’s study summarised the judgment in which shared custody had been granted to her abusive ex-husband as follows: ‘You’re not with him now? He’s not going to beat you. And has he ever beaten the child? No. He’s not dangerous to her’ (Robertson et al., 2007, p. 98). The backlash against the presumption was joined by the fathers’ rights lobby, academics and policy-makers, who supported perceptions that the operation of the presumption tended to inflame conflict and make disputes more contentious, and that domestic violence allegations were used as a weapon in highly conflicted parenting disputes (Perry, 2006, pp. 12, 16–19). The presumption was finally repealed in a neoliberal reform of the Care of Children Act 2004 which came into force in 2014 and which, like the Legal Aid, Sentencing and Punishment of Offenders Act 2012 and Children and Families Act 2014 in England and Wales, was justified by concerns about adversarialism and delay and placed major emphasis on agreement out of court (see also Atkin, 2015). The new s 52 of the Care of Children Act 2004 contains a virtual presumption of contact in all cases, which some may argue has been implicit all along in the approach of the New Zealand family courts.

The following articles

The New Zealand experience suggests that even a well-drafted presumption may be subject to the same gap between law in the books and law in action as has been seen in the implementation of PD12J. The following articles explore a range of alternative approaches to allegations of domestic abuse in post-separation children’s matters. The collection starts with Linnéa Bruno’s analysis of how positive reforms in Sweden aimed at recognising women’s bodily integrity, protecting children exposed to domestic abuse and prioritising children’s wishes, have been undercut by political and professional discourses and understandings of domestic abuse manifested in a range of settings including the family courts. Drawing on the empirical research she undertook for her doctoral dissertation, Times of trouble: Fathers’ violence, the state and the separating family (published in Swedish in 2016), Bruno explores the impact in custody proceedings of two social processes she identifies as endemic in Sweden: familialisation (the process through which kinship as a power structure is reproduced as seen, for example, in the promotion of shared custody and the positioning of children as objects) and selective repression (whereby responsibility for domestic violence is selectively distributed along racialised and gendered lines). Bruno’s article explores how these processes result in abuse being actively administered rather than countered, in ways which contribute to its continuation.

A similar contradiction between progressive reforms and their implementation in practice is explored by Teresa Picontó Novales’s study of the pioneering Spanish legislation which explicitly frames domestic abuse as gender violence/violence against women, together with further innovative measures enacted in 2015, which include attributing direct victim status to children living with domestic abuse and restrictions on contact and communication between fathers who have perpetrated gender violence and their children. However, Picontó Novales demonstrates how conservative judicial interpretation of these measures, underpinned by the greater priority given to the child’s perceived ‘need’ for a relationship with the father over his abuse of the mother, means that real change is slow, with a minimal increase in the use of these new measures over the past two years.

The importance of listening to children is addressed by Stephanie Holt (Ireland) and Kirsteen Mackay (Scotland). Stephanie Holt observes that children’s theoretical rights under Article 12 of the United Nations Convention on the Rights of the Child 1989 to participate in legal proceedings that affect them has struggled to achieve translation into practice in Ireland. Rather, universal assumptions of children’s best interests result in children’s competence to participate being questioned and their participation being restricted on grounds of age and maturity. Drawing on the narratives of children and young people involved in family court proceedings where domestic abuse was an issue, Holt concludes that when children’s voices were listened to, contact arrangements were made safer and children were empowered.

Kirsteen Mackay considers the approach taken in Scotland to child contact cases in which there are allegations of domestic abuse. Sheriffs have a statutory duty to consider the need to protect a child from abuse and from the risk of abuse when making an order relating to parental rights and responsibilities (which includes child contact). Drawing on her analysis of court papers from 208 child contact disputes, surveys and interviews with sheriffs, professionals, former litigants and children, Mackay discusses the strengths and limitations of four key features of the case process in terms of protecting women and children – the availability of legal aid; the cautious process of successive child welfare hearings; the use of court appointed child welfare reporters to investigate all the circumstances of the child; and the statutory requirement that the child should be given an opportunity to express their views. Despite these measures, like Holt, Mackay finds that the assumption in Scottish case law that contact with non-resident parents will best promote the welfare of the child can work to undermine positive measures for seeking children’s views, particularly where the impact of domestic abuse and children’s fears are not taken seriously by professionals.

One way in which professional practices concerning domestic abuse may be improved is via enhanced training. The focus of the article by Jaffe et al. is on a particular judicial education initiative, ‘Enhancing Judicial Skills in Domestic Violence Cases’ (EJS), which was developed and implemented over the last 20 years in the United States by the National Council of Juvenile and Family Court Judges and Futures without Violence. Jaffe et al. present their findings of a preliminary evaluation of the programme based on the self-reports of 480 judges from diverse courts and jurisdictions across the US who had participated in a four-day workshop between 2006 and 2010. Overall, judges reported the programme to be engaging and effective, with most identifying specific benefits and behaviour changes in the areas of access to justice, judicial leadership, victim safety, and abuser accountability as a result of participating in the programme. Jaffe et al. observe, however, that ultimately the extent to which the programme is having an impact on the experiences of victims, perpetrators and children who come before the courts is unknown.

A further step from judicial education is the establishment of specialist domestic violence courts (SDVCs) for promoting the safety of children and resident parents in family cases where domestic abuse is an issue. Jennifer Koshan’s article provides a wide-ranging and insightful review of SDVCs in the United States and Canada together with the more recent development (primarily in the United States) of integrated domestic violence courts (IDVCs). While there is much diversity in the jurisdiction of SDVCs, the most common type deal with criminal cases, while IDVCs have jurisdiction to hear criminal, family and civil protection matters in one setting. The article focuses on Koshan’s qualitative study of the IDVC in Manhattan, New York, which she situates within existing evaluations of other IDVCs in the US and Toronto. She considers the advantages of one judge following the family through multiple proceedings, and conversely, disadvantages such as the high settlement rate of cases in IDVCs, in light of existing concerns about parties being pressurised to settle contact cases and about domestic abuse being minimised in such cases. It is clear from Koshan’s article that more research is required on the specific impact of IDVCs on parenting disputes.

Mandy Burton provides critical insights into SDVCs in the criminal justice system in England and Wales, and the implications for such specialisation in the family justice system. She undertakes an in-depth examination of the one attempt in England and Wales at establishing an IDVC (in Croydon), and the reasons for its failure. Looking more generally at the evaluations reviewed by Koshan, Burton questions whether the empirical evidence makes a convincing case for developing IDVCs in England and Wales as a means of addressing child arrangement cases involving domestic abuse. She goes on to explore, as an alternative, the potential advantages of a model of specialisation within each of the criminal and family justice systems.

The concluding chapter summarises the most promising strategies raised at the symposium. Four programmes recently established in Australia take a trauma-informed, whole-of-family approach to meeting the legal and non-legal support needs of families affected by domestic abuse. Two of these operate outside the court setting in dispute resolution services, while two seek to address the needs of parties and their children holistically when they encounter the court system. The final strategy is a human rights approach to child arrangements cases. The European Convention on Human Rights is relevant to such cases not only via Article 8 (right to respect for family life: a qualified right), but also via Article 3 (right to freedom from torture and inhuman and degrading treatment: an unqualified right). A human rights approach would enable the rights of all family members to be taken into account, and would prioritise freedom from inhuman and degrading treatment for both children and their non-abusive parents. While the success of these approaches cannot be guaranteed, they do appear to offer the greatest prospects for building on the good practices and overcoming the problems identified in the foregoing articles.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Rosemary Hunter

Rosemary Hunter, FAcSS, Professor of Law and Socio-Legal Studies, Queen Mary University of London.

Adrienne Barnett

Dr Adrienne Barnett, Lecturer, Brunel Law School, Brunel University London.

Felicity Kaganas

Felicity Kaganas, Professor of Law, Brunel Law School, Brunel University London.

Notes

1. Different jurisdictions use different terminology. The current preferred term in England and Wales is ‘domestic abuse’, but it has previously been ‘domestic violence’. In Australia the term ‘family violence’ is preferred; in the USA it is still common to use the term ‘battering’, while Spanish legislation addresses ‘gender violence’. These terms are used interchangeably in this Introduction and throughout the special issue, unless it is clear at any point that a distinction is being drawn between them. In each case, we and the other authors are referring to patterns of harmful, threatening, intimidating, coercive and controlling behaviours which are, in all of the societies discussed, overwhelmingly perpetrated by men against women and children.

2. See, eg, Re S (Contact: Promoting Relationship with Absent Parent) [2004] EWCA Civ 18; Re M (Children) [2009] EWCA Civ 1216; Re J-M (A Child) [2014] EWCA Civ 434; Re T (A Child: Suspension of Contact: Section 91(14) CA 1989) [2015] EWCA Civ 71.

3. PD12J 2017 retains this provision in slightly different wording: ‘what evidence is required in order to determine the existence of coercive, controlling or threatening behaviour, or of any other form of domestic abuse’ (para. 19).

4. The study did not report on findings of domestic abuse.