Abstract
Significant strides have been made in the law's recognition of harms arising from domestic abuse. In England and Wales, the Serious Crimes Act 2015, and in Scotland, the Domestic Abuse (Scotland) Act 2018, have supported a more holistic understanding of the dynamics of abuse and the means by which coercion and control are deployed to cement and supplant perpetrators’ violence. In this article, we explore what the introduction of these offences means in other situations where questions regarding the impact of abuse upon victims’ agency arise: specifically, where victims commit an offence that might have been compelled by abusive behaviour or take their own lives in contexts that might indicate perpetrators’ liability for suicide. In particular, drawing on interviews with professionals across both jurisdictions, we highlight the precarity of recognition of the effects of coercive control and the need to engage in more complicated discussions about when and why context matters.
Introduction
In the past decade, significant strides have been made, across many jurisdictions, in the criminal law's recognition of, and capacity to respond to, harms arising in the context of domestic abuse that do not fall within pre-existing frameworks for criminalising physical or sexual violence. This has been advocated by many domestic abuse specialists for some time, based on well-established understandings of the complex operation of abusive behaviour, the dynamics of control and violence through which it functions, and the insidious nature of its negative effects upon victims (Scott, 2020). Legislative innovation has now created the potential for this more holistic understanding to receive increased legal recognition – in particular, by criminalising patterns of coercion and control which manifest through forms of emotional abuse, financial control, or orchestrated surveillance or social isolation, and operate to cement, support and even supplant the use by perpetrators of physical violence.
In England and Wales, this framework was initiated via section 76 of the Serious Crimes Act 2015 (hereafter ‘SCA’), which established an offence where a person (i) repeatedly or continuously engages in behaviour towards another to whom they are personally connected that is controlling or coercive, (ii) that behaviour had a serious effect on the other person such that it caused them serious alarm or distress having a substantial adverse effect on their daily activities or causing them to fear on at least two occasions that violence would be used against them, and (iii) the perpetrator knew, or ought to have known, that the behaviour would have this serious effect on the person to whom it was directed. Though investigative and prosecutorial backlogs have impacted the justice journeys of domestic abuse survivors substantially during the Covid-19 pandemic and its aftermath (Nott, 2022), there is evidence of growing utilisation of this offence since its implementation. In 2019/20, for example, police in England and Wales recorded 24,856 section 76 offences, which compares to a total of 4246 in 2016/17, with an increase of 18% in the rate of cases proceeding to prosecution in 2019 compared to the previous year (ONS, 2022). Equally, however, concerns remain regarding its marginal usage in the wider context of the overall volume of domestic abuse-related incidents reported to police in England and Wales (Barlow et al., 2020; Myhill et al., 2022).
Robinson et al. (2018) have documented difficulties with the appropriateness of existing risk assessment tools as well as with the attitudes of frontline police officers who utilise those tools, which they suggest are likely to undermine the potential of the section 76 offence. More recently, Barlow et al. (2020) have noted that further training and resourcing are needed in England and Wales to ensure the effective identification of these offences by police and the development of a gender-sensitive understanding regarding the harms and behaviours involved. Equally, other commentators have raised concerns about the potential challenges encountered at the prosecution stage, given that the model of repeated and continuous controlling or coercive behaviour envisaged within this legislation is often an unfamiliar one for those who argue, judge and decide these cases (Bettinson and Robson, 2020). Evidence in support of all these concerns was uncovered in the present study (see further, Bettinson et al., forthcoming). At the same time, though, it is important to bear in mind that report-to-charge ratios can often tell a partial story since attrition due to victim withdrawal is also a significant factor. Though there are legitimate reasons to interrogate what agencies might have done differently in the face of a victim's hesitancy to support a prosecution, in some cases it will be clear that proceeding against their wishes would be apt to increase their vulnerability (for discussion in the English and Welsh context, in particular, see Porter, 2020). Here, attrition may reflect concern about the effects of criminalisation more than its appropriateness.
Meanwhile, in Scotland, the Domestic Abuse Scotland Act 2018 (hereafter ‘DASA’) adopted a somewhat different approach to the framing of an offence designed to address these complex forms of psychological and emotional abuse. While section 76 was designed with a view to capturing behaviour that was not already covered by existing criminal laws, the DASA provisions took a more holistic approach, aiming to create a bespoke offence that targeted the wrongdoing at the heart of domestic abuse, including where necessary behaviours that were already criminalised under existing offences (Burman and Brooks-Hay, 2018). The DASA offence is designed to place emphasis on the behaviour of the accused rather than on the severity of its effects on the victim, and it targets only partner or ex-partner relationships rather than wider family relationships. Thus, section 1 DASA creates an offence where a person (i) engages in a course of behaviour which is abusive to a partner or ex-partner, (ii) they intend or are reckless as to whether that behaviour will cause physical or psychological harm, and (iii) a reasonable person would consider the behaviour likely to cause such harm to the recipient. It defines abusive behaviour as that which is violent, threatening or intimidating, or which has or would reasonably be understood to have as one of its purposes making the recipient dependent on or subordinate to the perpetrator, isolating the recipient from sources of support, controlling or regulating their daily activities, depriving or restricting their freedom of action, or otherwise frightening, degrading, humiliating or punishing them.
Even more so than its counterpart in England and Wales, the conjunction between the implementation of DASA and the Covid-19 pandemic has complicated the ability to evaluate its uptake and outcomes. That said, recent analysis indicates that, in 2021/2, there were 33,425 charges reported by police to prosecutors with a domestic abuse identifier, reflecting a 9% increase on the previous year (HMICS, 2023; Scottish Government, 2023). The vast majority of these offences remained for Breach of the Peace, Common Assault or Crimes against Public Justice, however. Though reflecting an increase of 13% on the previous year's total (n = 1581), in 2021/22, only 1790 charges were reported under DASA, accounting for 5.5% of all domestic abuse charges. As with England and Wales, then, concerns remain in Scotland regarding the consistency of response across police officers, and their skills in recognising and applying the DASA legislation, particularly where they are not part of specialist investigative units (HMICS, 2023). There is evidence of poor communication between justice agencies and victims, failures to adequately risk assess and safety plan for victims, and experiences of re-traumatisation during the trial process (Scottish Government, 2023). At the same time, the existence of a ‘clear strategic commitment to tackling domestic abuse at a senior level’ with ‘good governance structures’ and ‘strategic scrutiny of the local response’ has also been identified as ‘evidenced and embedded’ in Police Scotland (HMICS, 2023: 12). Importantly, this sits in a context in which there was a substantial investment in professional training and public education to accompany the implementation of DASA in Scotland (Scott, 2020), shown by independent research to have the potential to improve outcomes (Brennan et al., 2021).
Though our primary focus is on criminal justice responses, it is worth noting that, in both jurisdictions, recent years have also seen the evolution of more expansive regimes of civil protection. In England and Wales, the police have a role in enforcing breaches of non-molestation orders and the Domestic Abuse Act 2021 has extended the power of police officers to issue a Domestic Abuse Protection Notice (‘DAPN’) against a suspected perpetrator, which can be turned – through a successful application to the magistrates’ court – into a longer-term Domestic Abuse Protection Order. Once piloted, these new orders will replace the existing regime of Domestic Violence Protection Notices and Orders (‘DVPNs’ and ‘DVPOs’) first introduced in 2014. Although, in theory, the police can use them, for example, to prohibit contact and to regulate access to shared property, in practice there is variable use across forces in England and Wales, and it seems that the orders are under-utilised, with the enforcement of both non-molestation orders and DVPOs being weak (Bates and Hester, 2020). The Domestic Abuse (Protection) (Scotland) Act 2021 establishes a similar structure of emergency barring orders which are capable of being actioned by police officers. In both jurisdictions, barriers to using these civil/criminal hybrid orders include resourcing and training, and concerns have been raised about how they operate alongside the prosecution of substantive criminal offences, given that they are designed to supplement not replace effective prosecutions, including prosecutions for coercive control (Burton, 2022).
Aims and Scope of the Current Study
Against this backdrop, the authors undertook fieldwork during 2021–2022, which aimed to explore ‘on the ground’ experience of implementing these new criminal offences, and where appropriate to identify lessons to be learned across jurisdictions in terms of the framing of offences, investigative and trial process, and wider training/education around the complexities and harms of domestic abuse. To do so, we conducted a series of semi-structured interviews, which were situated alongside an extensive literature review of existing policy and academic materials regarding the design and implementation of coercive control laws. All interviews – which typically lasted approximately 60 minutes – were conducted and recorded on MS Teams, professionally transcribed and then coded for subsequent qualitative analysis using Nvivo. In total, we spoke with 20 professional stakeholders in Scotland (comprised of five police officers, six Crown Office prosecutors, six sheriffs, and three third-sector specialists) and 25 stakeholders in England and Wales (comprised of 12 police officers, seven magistrates, and six third-sector specialists). Police officers were recruited from a range of locations in Scotland and across two police forces in England, in order to provide some representation across urban and rural communities and diverse demographic contexts. Though the structure of specialist units varies, officers were typically located in departments where they would have more concentrated exposure to the investigation of domestic abuse, as distinct from being frontline officers only called to the incident initially, and prosecutors typically had wide-ranging experience across a large number of domestic abuse cases. In England and Wales, we focused on magistrates as they are a neglected part of the legal response to domestic abuse, apart from in the context of specialist domestic abuse courts, last studied in depth more than two decades ago. As coercive control is a triable either-way offence, it is likely to be dealt with more often by magistrates than Crown Court judges, given the mode of trial allocation guidelines which encourage their retention of cases. Across the interviews, we deliberately recruited magistrates who sat in specialist courts, as well as those who did not. In Scotland, the majority of criminal and civil cases are heard in the sheriff’s court, and the sheriffs that we spoke with had extensive experience presiding over criminal trials, including those that involved domestic abuse and coercive or controlling behaviour. Though the third-sector specialists that we spoke with worked in a variety of capacities and roles, they were also all highly experienced colleagues employed by or leading within organisations that have long provided advocacy and/or service support to victims of domestic abuse.
Those third-sector specialists were recruited through a snow-balling technique based on existing networks, with selection to ensure a mix of perspectives across legal advocacy and support provision, and enabling insight into the experiences of marginalised groups (including migrant or minority communities and victims with offending histories or addiction issues). Permission to recruit police participants was secured in the first instance from senior colleagues in the relevant force areas, in accordance with their research access protocols. Once permission was granted, the authors liaised directly with a designated point of contact, who suggested a list of relevantly experienced colleagues of varying seniority for us to approach. A similar approach was also taken to the recruitment of Procurator Fiscals in Scotland where, after securing permission from the Crown Office, we adopted a mixed approach of issuing an open call and targeted approaches to specific prosecutors identified to us by others. In relation to sheriffs and magistrates, again, a formal process for securing permission was completed after which we were assisted in recruitment by the Scottish Courts and Tribunals Service and Magistrates Association respectively. In the former case, we were given contact details for the head clerk in each sheriffdom with whom we liaised as to the willingness of individual sheriffs to be interviewed. In the latter case, we liaised with the Regional Leadership Magistrates who cascaded a recruitment email to members or approached people they thought particularly well-positioned to participate in the research. Access was also requested to interview Crown Prosecution Service colleagues in England and Wales, but the delay on obtaining approvals meant that it was not possible. Though the interview protocol was modified slightly to accommodate the different constituencies, in line with its semi-structured nature, it was designed to address the same broad themes across discussions. The authors coded the interview data collaboratively, first establishing a coding frame on a grounded basis using a sub-sample of transcripts: when synergies had been identified, structures imposed, and points of inconsistency addressed, we continued to code using this coding frame, with the flexibility to flag and add emergent themes as needed on an iterative basis.
In other work, we have explored in detail our findings regarding the challenges and opportunities encountered in identifying and responding to the scale and complexity of harms associated with domestic abuse within these legislative frameworks, as well as in successfully evidencing and prosecuting coercive control offences in and across both jurisdictions (Bettinson et al., forthcoming). In this article, we focus attention on another key aim of the research which was to explore the potential precarity of the contextual understanding of coercive control that those legislative reforms purported to bolster. More specifically, we wanted to explore potential tensions between justice professionals’ experiences of implementing these offences and responding to other situations in which the fact of being subjected to coercive control might imprint upon victims’ choices – in particular, where the controlling circumstances prompted victims to commit offences themselves or led to a situation in which victims took their own lives.
Our starting point for analysis here is that the creation of coercive control offences is well-intentioned and important. It ushers into the criminal law a recognition of patterns of abuse that can isolate victims from networks of support and radically reduce their capacity or confidence to exercise agency. Though coercive and controlling behaviour may often be accompanied by physical violence, perpetrators do not need to rely on it, and it is not required by statute in the two jurisdictions under consideration. To recognise that coercive controlling behaviour has the potential to make victims feel hopeless, such that resisting the perpetrators’ demands or taking steps to leave the relationship seem futile, is not to position them as passive or pathologised. Rather, it is to highlight the complex and interconnecting ways in which relational, social and structural factors frame victims’ experiences of living within and navigating abusive relationships. In this respect, recent work that has infused understandings of the dynamics and effects of coercive control with a ‘social entrapment’ lens is instructive. Such an approach attends to: first, the social isolation, fear and coercion placed on the victim which is created by the perpetrator's use of control; second, the institutional responses to victims’ suffering that are formally available, and factors grounded in prior engagement or circumstances that impact on victims’ assessment of their feasibility and efficacy; and third, the exacerbation of individual experiences of control and coercion arising from factors such as gender, class, race, age and disability (Ptacek, 1999: 10). As Tolmie et al., have argued, this analysis can be crucial in focussing explanations on ‘the objective realities of the circumstances in which the primary victims are located’ (2018: 207), in a context in which the isolating and disempowering effects of abusers’ behaviour are often compounded by layers of other situational or structural disadvantages, and victims’ co-existing vulnerabilities and precarities are manipulated and capitalised upon by perpetrators.
In this context, then, there may be ramifications of the recognition of the effects of coercive control that extend well beyond the immediate confines of the offences outlined above. By exploring participants’ understandings of how coercion might operate in a wider range of situations, and, specifically, how it might impact assessments of the agency and responsibility with which crimes are committed by victims or deaths by suicide occur, we sought to probe the potential precarity of existing understandings. While each individual and circumstance is distinctive, experiencing domestic abuse can – and often does – cause substantial harm to victims, reducing their self-confidence and sense of agency, isolating them from support, and creating conditions in which ceding to the demands or pressure of perpetrators or taking their own lives can, in some cases, feel like victims’ best option. At one level, this remains a choice made by the victim, but in other important respects, it is also a consequence of the actions of perpetrators. Moreover, the dynamics that construct and constrain that choice are also multi-layered, with causative factors shifting in and out of view over time and context in complex ways, and intersectional vulnerabilities exhibited by the victim that may variously predate, arise from or have been amplified by the effects of domestic abuse. Victims’ prior engagement with services also frequently cements their perception of a lack of import, bolstering conclusions about the futility of evasive alternatives, and often in ways that perpetrators will directly rely upon and manipulate to extend control.
In this article, then, we reflect on the extent to which such complexity is capable of being accommodated within existing criminal justice responses, in Scotland and England and Wales. Informed throughout by a feminist analysis that takes seriously the situated perspectives and lived experiences of victims of abuse, we interrogate how relevant evidential and legal thresholds were understood and applied by participants, and the extent to which they are, or could be, capable of recognising appropriately the impact of inter-personal power dynamics, contextual constraints and structural conditions on domestic abuse victims’ scope for agency. There may be sound reasons – grounded, for example, in moral responsibility or public policy – why the effects of coercive control ought to be responded to differently by criminal justice systems when deployed either to excuse victims’ criminality or to contextualise and attribute responsibility to perpetrators for their suicide. But, in this analysis, we seek to probe the extent of the inconsistency, reflect on its implications in practice, and assess the adequacy of any accounts provided by justice personnel to justify and explain this divergence. Though, as outlined above, the dataset we draw upon is relatively modest in scale, particularly when segmented across jurisdictions and participant roles, the findings, whilst not claiming to be representative of all, are robust due to the commonality across interviews. And in this fast-evolving landscape, the opportunity to reflect across contexts in which coercive control is a prominent concern – contexts that might, at first sight, seem diverse and are unlikely in justice processes to be thought about together by professional stakeholders – provides an important opportunity to generate new insight by looking more holistically and subjecting the assumptions that inform one context to critical scrutiny from the vantage point of another.
In the first part of the article, then, we highlight the partial and tentative nature of professionals’ understanding of the role that coercive control can play as a cause of victims’ criminality, and reflect on the barriers to justice that this continues to pose. In a context in which others have already highlighted difficulties presented in relation to violent offending against an abusive partner, we address issues of mitigation for homicide and self-defence more briefly, before turning to victims of abuse who are coerced or compelled to non-violent forms of criminality. We suggest that, while this accounts for a significant proportion of victims’ offending, the relevance of coercive control has not been adequately considered nor responded to here, either in theory or practice. In the second part of the article, we turn to a different set of circumstances in which, nonetheless, the role and relevance of coercion in influencing and informing victim behaviour are also pertinent, namely where victims of domestic abuse take their own lives. By interrogating professionals’ hesitancy to consider the criminal liability of abusers in this context, we emphasise shifting understandings of cause, coercion and control in the criminal justice process and the extent to which acknowledging victims’ diminished agency emerges as increasingly complex. In the concluding section, we reflect on these scenarios, juxtaposing professionals’ responses to the rhetoric that animates coercive control as an offence. We reflect on the precarity of that discourse and the extent to which taking coercive control seriously may require justice professionals to engage in more complicated discussions about when, why, and to what extent, such context matters.
Coercive Control as a Cause of Criminality
The frequent involvement of domestic abuse in women's pathways to criminality has been widely documented. In 2017, the Prison Reform Trust (hereafter ‘PRT’) reported, for example, that 57% of women in prison in England and Wales had been the victims of domestic abuse: a figure that was, moreover, ‘likely to be an underestimate’ given well-established barriers to disclosure (Gelsthorpe et al., 2007). Though there may not always be a direct causal link, this abuse was identified as a key driver to offending for the ‘majority of women in prison’ (PRT, 2017: 9). More recently, these findings have been reinforced by the Centre for Women's Justice (hereafter ‘CWJ’) (CWJ, 2022), which has highlighted key failings across the criminal process, particularly tied to a lack of understanding and recognition of coercive or controlling behaviour. These concerns have also been echoed by the All Party Parliamentary Group on Women in the Penal System, which concluded that, notwithstanding the U.K. Government's Female Offender Strategy (2018), which commits to a reduction in women's incarceration particularly in respect of serving short sentences, ‘far too many women are drawn into the criminal justice system’ and come ‘into conflict with the police when they need support’, including in respect of domestic abuse victimisation (APPG, 2020: 1). Similar concerns have been expressed in Scotland where recent research has revealed that almost four in five women prisoners had a significant history of head injury that, in most cases, occurred in the context of domestic abuse, to which they were often subjected for many years (McMillan et al., 2021).
Thus, even where domestic abuse takes more corporeal forms – manifest through physical or sexual violence – it is not clear that the effects of experience that abuse on creating, or at least facilitating, pathways to offending have been adequately recognised or responded to within the criminal justice system. Moreover, current defences and diversionary alternatives may be particularly apt to fail women who have experienced forms of non-physical domestic abuse, whether alongside or in lieu of physical violence (PRT, 2017). This is partly because such abuse may not generate the corroborative trail of hospital admissions or agency interventions that, though still absent in many cases, may be more likely where there has been physical violence, but it may also be because the severely oppressive effects of such abuse continue to be less readily recognised by justice professionals. The consequence of this, as Loveless cautions, is that ‘non-culpable abused women’ are ‘left defenceless’ (2010: 108).
Professional participants in the present study were largely cognisant of this empirical reality in which many female offenders have been subjected (often repeatedly and over prolonged periods) to domestic abuse. However, when asked about the potential connections between being a victim of coercive control and committing criminal offences, they often provided responses that reflected this lack of recognition, as well as a lack of confidence about steps to be taken where victimisation was recognised. Participants observed, for example, that ‘if they’re committing crime because they’re being forced to do so, you’d like to think that would be massively mitigating’ (Police 2, emphasis added) and assured that ‘it's certainly something that we’d identify if it screams out to us’ (Police 5, emphasis added). Such responses did not, however, yield any systematic approach to identifying and responding to such abuse.
Violent Resistance, Mitigation and Self-Defence
Without further probing, most participants engaged with the issue of female offending in the context of domestic abuse only in relation to so-called ‘dual perpetrator’ cases, where the form of criminality involved arose as a consequence of reciprocal violence (Hester, 2013). Such cases are sometimes (incorrectly) labelled mutually abusive, and this has proved problematic across the legal system, not just in the criminal justice sphere. In our study, participants emphasised that ‘you should never be arresting both parties … you should always be trying to establish … what has actually happened’ (Police 2). Indeed, although the extent to which such guidance would assist in contexts of coercive control where the main behaviour involves, for example, verbal rather than physical abuse, was unclear, several interviewees in Scotland highlighted the existence of joint police and prosecutor guidance addressing this issue, which was intended to ensure ‘primary aggressors’ were appropriately identified (Brooks and Kyle, 2015; Police Scotland and COFPS, 2019). As one officer put it, ‘it's looking at that big picture, why did the victim have to do that, why did they have to go to that extent? You’ve now got a serious assault or an attempted murder … it isn’t just, oh, the female stabbed him or she had an ashtray and fractured his skull, you know, look at the whole circumstances’ (Police Scotland 1). Meanwhile, interviewees in England and Wales often referenced the high-profile case of Challen [2019] EWCA Crim 916, in which the Court of Appeal overturned a conviction for murder after accepting psychiatric evidence outlining the impact on the victim of having been subjected to coercive and controlling behaviour across a 40-year relationship. While Challen's defence at her first trial relied solely on a diagnosis of depression to support her plea of diminished responsibility, this failed to put before the jury information about the impact of coercive control on her mental health, which it was recognised could give rise to mitigation to homicide (Howes et al., 2021). Broadly, our participants welcomed such developments, which they felt enabled them to better address cases of ‘violent resistance’.
Even within this terrain, however, the use of doctrinal avenues for limiting liability has been widely criticised as inadequate (CWJ, 2022; Howes et al., 2021; McPherson, 2022). Pleas of loss of control or diminished responsibility are not only limited to those whose violent resistance resulted in the death of their abuser but impose a number of restraining categorisations on victims in order to obtain that mitigation. Indeed, loss of control, like its predecessor, provocation, is gendered. Historically more amenable to the ‘red mist’ killings of angry men than the ‘slow burn’ cumulative impact of domestic abuse (Howe, 2012), Edwards (2021) has argued that the revised requirement for fear of serious violence in the Coroners and Justice Act 2009 in England and Wales continues to undermine much understanding of how an abused, coerced or controlled woman might react (see also Williams, 2020). Meanwhile, in Scotland, critics have underscored the profound injustice of retaining a ground for provocation based on sexual infidelity (Drury v HMA 2001 SCCR 583), whilst failing to acknowledge the plight of victims of domestic abuse (McDiarmid, 2010, 2019). Further, though Challen has opened space for reliance on diminished responsibility, it has also continued to rely on pathologising explanations of what might be argued to be normal reactions to sustained and systemic abuse (Bettinson, 2019; Burton, 2022; Wake, 2013).
Moreover, where the victim's violent resistance does not result in the death of her abuser, there is no scope for such partial mitigation under homicide rules. Unlike in some other common law jurisdictions, self-defence is not a partial defence to either homicide or assault in England and Wales, nor in Scotland, and the prospect of those who have survived a prolonged course of abusive behaviour using it successfully to avoid responsibility is remote (Bettinson and Wake, forthcoming). The Scottish Law Commission (2021) has observed that its focus on imminent danger ignores the reality that victims are often required to use stealth or surprise to increase the prospects of successful defensive action; and that it fails to recognise the extent to which victims’ assessment of the feasibility of escape is impacted both by their experiences of abuse, as well as by empirically defensible concerns, particularly within some communities, about the reliability and availability of safe alternatives (Douglas et al., 2021; Tolmie et al., 2018). This has been further supported by McPherson's recent analysis (2022), which found that the defence is rarely successful when relied upon by victims of domestic abuse. McPherson has argued that this is because it privileges (male) experiences of (public) violence over (female) experiences of domestic abuse. Meanwhile, in England and Wales, similar barriers to reliance on self-defence have long been acknowledged (McColgan, 2000). Howes et al. (2021) have highlighted the need for changes in attitude, practice and legal doctrine alike to improve prospects for justice: in particular, they highlight the disparity in current treatment between abused women whose use of apparently disproportionate force (e.g., use of weapons) precludes reliance on the defence and householders who are now legally permitted to use disproportionate force when faced with an intruder, with this assessment based on the circumstances as they believed them to be (see also Wake, 2013).
Our research suggests that these doctrinal shortcomings might be particularly pronounced in contexts of coercive or controlling behaviour, remembering that – as a general defence – self-defence ought to be available for a variety of non-fatal offences and not solely where lethal force is used. The abuse experienced, though undoubtedly deeply harmful to victims, may not involve the sort of direct application of force thought to legitimately prompt defensive violence. Equally, the corrosive effects of coercive control and the inequities of institutional responses mean that victims are often isolated from support, and may be especially sceptical of the prospects of alternative means of (lasting) escape (Bettinson and Wake, forthcoming). This was reflected upon by third-sector stakeholders. As one put it, for example, though the advent of coercive control legislation offers ‘an opportunity to make transparent the mechanisms of women's inequality and domestic abuse’ and ‘reconceptualise this notion about cause and consequence’, to work effectively the complexities need to be ‘unpicked a lot more’: ‘the idea of being coerced to kill somebody who has never physically assaulted you is easy to understand if you work in the sector, completely easy to understand … but really, really under operationalised in the ways that we talk and think about women who are accused’ (Third Sector 7). Whilst some justice professionals that we spoke to also acknowledged these difficulties with applying existing defences to violent offending precipitated by coercive control, they were considerably more sanguine about the prospects for ‘unpicking’ and ‘reconceptualising’. As Prosecutor 2 put it, ‘there is discussion to be had about whether we widen out what realistically was the exit, when somebody has been told for years that they have nowhere else to go’ but under existing frameworks, ‘it's hard to see what you would do with that … it will always come back down to there was another option available to you and you could have left’. Thus, though the ‘why didn’t she leave’ trope that reduces empathy for survivors of abuse and diminishes their prospects for securing justice and protection from perpetrators may be a less potent feature of contemporary responses to victimisation, participants indicated it remains a feature of defence doctrine in circumstances where the abuse suffered compelled victims towards criminality (Douglas et al., 2021; Tolmie et al., 2018).
Coercion, Duress and Non-Violent Offending
Participants’ lack of ambition in recognising coercive control in the context of defensive violence mapped onto their responses in regard to other types of criminality with which victims of domestic abuse might be charged, and the alternative defences that such charges might bring to the fore. More specifically, in a context in which the majority of female offending does not involve violence, it has been suggested that victims of domestic abuse may nonetheless be compelled to criminality in order to ensure their wider self-preservation (Barlow, 2016). This may involve ‘status’ offences in which victims are placed in a precarious position by an abusive partner, for example, in respect of benefits entitlements; ‘consequential’ offences where victims are put into criminality under pressure from partners, for example, drug dealing or prostitution; or ‘liberation’ offences where the crime is intended to improve or remove the victim from their abusive situation, for example, shoplifting to redress a lack of access to independent finances. In such contexts, a victim's capacity for choice may be eroded, and failure to comply with partners’ demands may provoke a punishment worse than legal consequences (Bettinson et al., 2023; Schloenhardt and Markey-Towler, 2016). Third-sector interviewees often underscored the scale of this problem. One commented, for example, that ‘we hear it so often, how women end up committing offences just because of the pressure that they’re under in the relationship’ (Third Sector 1), whilst another observed that ‘we have lots of clients who are coerced by the perpetrator to commit benefit fraud’ (Third Sector 5). Further, another confirmed ‘we do see that … things like, you know, limiting her access to money so she can’t buy nappies or food for her kids and that causes either shoplifting or other kind of behaviours’ (Third Sector 3). But such participants also highlighted the difficulties of recognising and responding to this within existing duress defences and suggested that there's an ‘unwillingness to engage on such minor offending’ (Third Sector 1), both at the policy level and in the mechanics of professional criminal practice.
That suggestion certainly seemed to be borne out in the contributions made by justice professionals in this study. As one officer observed, in the context of those charged with purchasing, or stealing in order to purchase, drugs for a partner, ‘taking into account coercive control for the offending they’ve done, as bad as this may sound, I don’t think it's ever featured on my radar’ (Police 10). Meanwhile, others insisted that this was simply not something they had encountered – as one prosecutor put it, for instance, ‘examples like that, where complainers have been sort of compelled into doing certain things … [are] quite rare’ (Prosecutor 6). At the same time, other contributors were willing to acknowledge that such scenarios do arise, but reflected on the challenges they raised, both substantively and procedurally. As one prosecutor put it, ‘you do often get these cases coming in and it's hard to know what to do with them’ (Prosecutor 2), whilst a police officer remarked ‘we’re much more sympathetic to victims who perhaps assault a partner because they’ve just had years and years of coercive control and they’ve just snapped …. Where I think we struggle to know what to do is, for example, where you might get someone arrested for possession with intent to supply and the defence is, well, I didn’t want to do it but my abusive partner made me do it … that's a tricky one’ (Police 6). Often these participants also went on to point out that there are a variety of reasons why a person might commit an offence – ‘being in a controlling relationship might be one factor, but not the only factor’ (Prosecutor 4) – and that it was important to acknowledge a level of choice remaining with the victim to commit the criminal act – ‘it's the same as if someone is an alcoholic, which isn’t their fault and it is a recognised illness, but you can’t blame every offence you commit on being drunk’ (Police Scotland 2). Indeed, as one sheriff put it, ‘that's a messy can of worms to start picking apart’ (Sheriff 1).
While duress would be the obvious defence for victims to seek to rely upon in such scenarios, significant doctrinal barriers to its effective deployment remain, both in England and Wales and Scotland. In particular, duress has been understood to be limited to threats or circumstances that present the risk of death or serious injury, failing to recognise psychological or financial forms of abuse. It also looks for evidence that the person threatened had no reasonable alternatives for escaping the harm other than by committing the offence, based on a modified version of the reasonable person standard that fails to recognise the deleterious impacts of abuse and disappointing prior agency engagement upon victims’ assessment of those prospects (Bettinson et al., 2023; Loveless, 2010; Prison Reform Trust (PRT), 2017). Existing case law – across both jurisdictions – also highlights the demanding thresholds that continue to be imposed, whereby courts require, in effect, a complete erasure of agency in order to avoid liability.
In R v A [2012] EWCA Crim 434, for example, a rape victim who was prosecuted for perverting the course of justice following her retraction of a complaint against her partner, was denied the defence. The court emphasised that ‘the circumstances in which different individuals are subjected to pressure, or perceive that they are under pressure are virtually infinite’ but ‘duress should not and cannot be confused with pressure’ (para 63). Thus, whilst threats of rape would constitute a sufficiently serious injury to be relevant, they could only ground a defence of duress where they operated to overbear the victim's will. In this case, notwithstanding the background of domestic abuse, it was ‘inconceivable’ to the court that the victim would not have told authorities at the time such threats were made against her, had they been genuine. Therefore, it was deemed that she had reasonable alternatives to retracting the complaint in order to avoid the threat of harm (para 70). Meanwhile, in R v GAC [2013] EWCA Crim 1472, while the court accepted in principle that Battered Women's Syndrome may be a relevant factor to be taken into account when considering whether an individual is acting under duress, it emphasised that ‘not every woman who suffers domestic violence goes on to suffer from BWS’ and ‘not every women who suffers BWS can claim the defence of duress’ (Para 49). The court concluded that ‘an accused would have to be suffering from BWS in a severe form to be in a position to claim their will was overborne’ (para 51). Notwithstanding expert evidence to the contrary, the court maintained the appellant did ‘not come close to establishing she may have been subjected to serious physical violence so bad that she had lost her free will’ (para 51). This is perhaps surprising given the dangerousness of GAC's ex-partner who had been convicted of murder. Thus, GAC established a narrow test that focussed on ‘chronic’ manifestations of ‘severe physical violence’ where victims were subject to ‘total domination’ (para 55). That GAC was described as ‘feisty’ underscored that she did not present as sufficiently ‘helpless’ to make use of the defence. Thus, rather than engaging in a more contextualised analysis of the sort that would emerge through a social entrapment approach, the court in GAC failed to appreciate the significance of the underwhelming institutional responses to the victim's previous help-seeking and the ways in which this, alongside the coercive behaviour of the perpetrator, progressively closed down her space for action, including in respect of resisting offending.
Moreover, the advent of the coercive control offence in England and Wales appears to have done little to adjust this doctrinal landscape. In Johnson v R [2022] EWCA Crim 832, the court likewise indicated that a victim of domestic abuse would need to establish that she ‘effectively had no will’ in order to avail herself of the defence of duress, having made false statements to the police decades prior during investigations into murders for which her partner was subsequently convicted. As in GAC, the court underscored that, in its view, she had the opportunity to escape from or avoid any threat posed by her partner or his family, by going to the police after the threats were made or prior to giving her incriminating testimony. Indeed, the court arguably held the victim to a higher standard as a consequence of her previous – albeit failed – attempts to engage support services, observing that since she had made prior requests for assistance ‘she did not consider that she had no choice – she would have welcomed and accepted police protection – yet she took no steps to secure that help’.
In Scotland, though the case law is less developed, the leading case of Ruxton v Lang deploys a similar approach. The accused had driven under the influence of alcohol in order to escape an abusive ex-partner's armed attack on her and her male companion but was unable to avail herself of the defence of necessity. The sheriff opined that at some point in the less than two-mile journey before she was apprehended, the threat compelling her would no longer have been ‘immediate’; and that even if it were, the accused had opportunities that she could reasonably be expected to have taken to avoid continuing the offence. As part of the feminist judgments project, Cowan and Munro (2019) critique this decision. In their feminist re-imagining, they argue that ‘the law must acknowledge that … the precise parameters of the risk of serious violence posed by any one incident depend on a background pattern of behaviour and cannot be evaluated in isolation or with exacting precision’ (2019: 92). Thus, they conclude ‘the accused remained in immediate danger … until such time as she secured protection’ (2019: 93). Moreover, Cowan and Munro challenge the suggestion that the accused had prudent evasive alternatives available: the sheriff's conclusion that she ought to have pulled over the car to avoid further offending and continued her journey by foot ‘takes for granted the erroneous assumption that public spaces are by default safe spaces for women’, notwithstanding daily court and common experience indicating otherwise (2019: 96). The failure of the sheriff in Ruxton v Lang to consider these factors established a line of highly restrictive authority. This was, however, further supported in D v Donnelly [2009] HCJAC 37, where the appellant was again denied the defence of necessity to a charge of drink-driving when – after having a couple of drinks at a social function – she was sexually assaulted by four men in a car parked outside of the club. Though she was partially dressed and in a state of distress, the sheriff concluded that, since the assailants had left and returned to the function, just a short distance from where the car was parked, the appellant was not facing any immediate threat. Moreover, as the appellant had her phone with her, there was a reasonable alternative to driving in that she could have remained in the car and called for assistance. This invokes a very narrow interpretation of immediacy in respect of the threat, fails to acknowledge the impact of trauma, and simplistically presumes reporting to the police to be a viable and effective solution for ensuring the protection of victims of such abuse.
Also informing these restrictive doctrinal thresholds, of course, is a public policy concern around not allowing undue latitude for law-breaking. This was reflected in the Westminster Government's rejection of calls to create a bespoke duress defence for victims of domestic abuse, akin to that currently in operation in respect of victims of modern slavery, on the basis that it was necessary to balance ‘recognition of the abuse that has been suffered and the impact that it has had on a victim against the need to ensure that people, wherever possible, do not revert to criminal behaviour’ (Lord Wolfson, DA Bill Debate, HoL, 2021). Many of our participants, when probed about the feasibility of recognising the effects of coercive control to reduce or absolve criminal liability were also keen to caution that it would be open to abuse. Indeed, several invoked the language of ‘a get out of jail free card’ (e.g., Police 8). One officer noted, ‘I could go on an offending spree and then claim I’m being controlled, and people will do this, we know that's human nature’ (Police 1), whilst another observed ‘it would be easy for victims to say they’re a victim of controlling and coercive behaviour … we’d have to be careful with throwing it around willy-nilly, as opposed to using it in genuine cases’ (Police 3). There are good reasons to be circumspect about the potential for unintended consequences arising from innovations designed to achieve progressive purposes. And in the context of domestic abuse, there are particularly apposite examples from other jurisdictions, including the ill-fated defence of defensive homicide in Victoria, Australia, which was introduced in large part to mitigate the liability of women who used lethal force against abusive partners but was abolished quickly thereafter on evidence of its unanticipated use primarily by male offenders to justify excessive use of force (Fitz-Gibbon and Pickering, 2012; Ulbrick et al., 2016). Equally, however, the suggestion by a number of justice professionals in this study that it would be ‘easy’ for offenders to claim that they had been coercively controlled is arguably at odds with extensive literature documenting barriers to disclosure and escalating risks tied to victims’ reporting. Moreover, to the extent that fabrications may be made by those charged with offences, the implication of such professional responses is that such claims, once made, would automatically be afforded credence. The reality, however, is that it would be necessary for such claims to be investigated appropriately before being relied upon. Nonetheless, this disquiet about the purported ease of potential misuse of contextual protections to victims illustrates the extent to which traditional reactions of scepticism and trivialisation towards allegations of domestic abuse may continue to linger.
For other interviewees, the concern was less about widespread abuse and more about the appropriate stage in the criminal justice process for any such context to be taken into account. As one police officer put it ‘at the end of the day, it's up to the jury to decide …. I just investigate stuff’ (Police 8); similarly others observed ‘that's ultimately for the sheriff to decide … as police officers, yes, I use my discretion when I need to … but if crimes are being committed, we’re duty bound to investigate and report the circumstances … to then be brought before court’ (Police Scotland 3) and ‘I know it's potentially passing the buck … but you can’t really discriminate where the evidence is there, you can’t cherry pick who you’re charging and who you’re not charging with an offence’ (Police Scotland 2). Meanwhile, prosecutors noted ‘our role is as a prosecutor, we see an offence, we will prosecute it’ (Prosecutor 3) and reflected that ‘we would probably be quite risk averse in that situation … we would probably mark it for court and say we’ll see how it comes out in the trial’ (Prosecutor 5). Though this may reflect the different institutional logics of the limbs of the justice process, such reluctance amongst professionals at earlier stages to actively turn their minds to considering the potential existence and weight of a defensive claim risks disproportionate exposure to prosecution, and arguably falls foul of the commitment to consider public interest, which lies at the foundation of prosecutorial codes in both jurisdictions. Moreover, as Third Sector interviewees often observed, even if mitigation is ultimately successful at trial, the accused had to go through the distress of the trial process. The prospect of this in itself can render victims more reluctant to come forward, even where doing so may protect themselves and the public from greater harm: for example, one interviewee recounted a case in which a perpetrator had recorded a video of the victim taking cocaine which he threatened to show to the police if she reported his abuse towards her. The victim specifically sought reassurance from the charity that she would not be prosecuted for that offence if she reported to the police but ‘we can’t guarantee that’ (Third Sector 2).
Moreover, while some participants highlighted the benefits of adversarialism, insisting ‘if their defence lawyer is doing their job properly, they can explain why it is that a person isn’t guilty or guilty but less so because of these various reasons’ (Magistrate 2), the speed and resource with which low-level offences are tried raises questions about whether an evidenced account of coercion will be put forward or even whether a victim will be able to disclose. Indeed, one sheriff observed that, by the trial stage, ‘we’d be perfectly content to simply avoid it because that's a very awkward situation to start picking at … do you simply accept the suggestion of coercive control, do you make any investigations into that’. Though other judicial participants did indicate that they would be more sympathetic to arguments heard at trial about the impact of coercive control in precipitating criminality, for many – absent further authority – this was seen to be unlikely to absolve responsibility: ‘I think there is sympathy for victims of abuse who offend but there is also a feeling that it's sad but nonetheless the offence has been committed … I suspect it would be, we’re very sorry, we understand why you did it, but you did it, and therefore you’re guilty’ (Magistrate 3). Meanwhile, another observed that ‘ultimately, they must know that they’re doing wrong’ (Magistrate 5), whilst another emphasised the need to establish, prior to any mitigation, that ‘there is no option for you to get away, go to the police, do something different’ and ‘if you’re taking the view that continuing a relationship with [the partner] is more important’ that won’t suffice (Sheriff 1).
For others, the relevance of domestic abuse in this context appeared to be contingent on the perceived severity of the threat that perpetrators posed to victim-offenders. As one judge put it, ‘I think we’d probably have less sympathy for [coercive control] than we would the more serious side of things’ (Magistrate 5), which he suggested would extend to threats of harm to a child, dog or criminal damage. Meanwhile, for others, it seemed the relevance of abuse would hinge significantly on the respectability of the accused: one magistrate recounted a case involving ‘a lovely, lovely woman, who had been a fairly prolific shoplifter’ after leaving her home to escape domestic abuse. Noting that there was something ‘almost elegant about her’, he reflected ‘I suddenly realised this was a middle-class woman who had just hit really bad times … she was a criminal because she committed offences, but she wasn’t a criminal: that wasn’t her nature, but she’d been forced into it by her circumstances’ (Magistrate 2). Though this recognition of coercive circumstance is important, it was far from clear that the same sympathy would have been afforded by that judge to the majority of female offenders who, alongside histories of abuse, navigate chaotic lives and social exclusion. Indeed, as one officer put it, in such cases, there is a much greater likelihood that the impact of coercive control ‘could get lost in the noise of everything that's gone on’ (Police Scotland 1).
Overall, then, the findings of our study support the body of existing literature which suggests that domestic abuse is hard to utilise in the context of defences, be that partial defences to homicide or generic defences to homicide or other crimes. Amongst criminal justice professionals that we spoke with, there was often a reluctance to acknowledge its relevance in the context of victims’ offending behaviour. And though there was an openness amongst some decision-makers to see it as potential mitigation in the ‘right’ circumstances, what constituted the right circumstances were, it would seem, often informed by assumptions about ‘serious’ domestic abuse and ‘ideal’ victims, of a sort that it might have been hoped would have been left behind with the creation and implementation of coercive control offences themselves. The dominant response to this amongst commentators has been to argue for legal reform that will render the substantive tests embedded within key defences more clearly applicable to the domestic abuse context and/or to create a statutory exemption from prosecution in a way that mirrors protections available to other cohorts. Though such reform may be the best long-term solution, such initiatives have met with limited success to date; and it is important to underscore that, with greater professional curiosity, critical reflection and contextualised application, there is a good deal more that could be done within existing legal parameters to recognise the effects of coercive control in reducing victim-offenders’ space for action and address the consequences of that in terms of their culpability. That endeavour will require far greater openness and awareness amongst professionals across all stages of the justice process than our data suggests is currently in place, however.
Coercive Control as a Cause of Suicidality
The challenges that many justice professionals encountered in recognising the potential impact of domestic abuse in precipitating criminality were also mirrored in their reflections regarding its role as a cause of suicidality. In contrast to the substantial evidence base that indicates a co-occurrence, if not a connection, between experiencing coercive control and female offending, the strength of the relationship between domestic abuse and suicide is less clear in the United Kingdom. In one early report, based on data within one locality, it was estimated that as many as half of all women in Asian communities, who had attempted suicide or self-harmed, may have suffered domestic abuse (Chantler et al., 2001). Meanwhile, in 2011, Southall South Black Sisters reported that – across a sample of 409 domestically abused women that the organisation had worked with – some 44% had contemplated suicide or self-harm and a further 18% had made attempts to do so. In addition, during the 8-year period reviewed in this research, a further eight women had ended their lives by suicide (Siddiqui and Patel, 2011).
Subsequent to this, in 2018, Aitken and Munro analysed data across a cohort of 3519 clients who had interacted with REFUGE between April 2015 and March 2017. Some 24% responded positively to one or more measures of suicidality: 18.9% reported feeling suicidal currently or recently, and 18.3% confirmed having made plans to end their lives, with 3.1% declaring that they had made at least one failed attempt to do so. This suicidality was present across clients with a diverse range of types of abuse experienced, but the correlation was heightened for those who had suffered coercive and controlling behaviours over a long period of time, or perpetrated by more than one person. In addition, the researchers reported that clients who expressed suicidality scored significantly higher than peers in the CORE-10 psychological distress questionnaire, with measures tied to feeling despairing and hopeless, or depressed and isolated, being prominent; and often intersecting in complicated ways with an increased prevalence of issues in relation to the misuse of drugs or alcohol (Munro and Aitken, 2020).
More recently, McManus et al., based on an analysis of the 2014 Adult Psychiatric Morbidity Survey, which involved over 7500 participants, reported that in past year suicide attempts were almost three times more common in victims of intimate partner abuse, and almost four times more common amongst those victimised in the previous year (2022: 6). Indeed, amongst those who reported having attempted suicide in the past year, 58.4% of the women had experienced intimate partner violence over their lifetime, as had 39.6% of the men (2022: 6). In line with the REFUGE research, McManus et al. also indicated the existence of a ‘dose-response’ relationship between domestic abuse and suicidality, with those who experienced two or more types of abuse having higher odds of suicidality within their sample (2022: 7).
It is as a consequence of this evolving knowledge base around proclivity and prevalence that the U.K. Government's most recent Domestic Abuse plan has expressed ‘concern’ about the effects of domestic abuse on suicides (Home Office, 2022: 7), noting that ‘in too many cases, these harms [inflicted by domestic abuse perpetrators] can result in a victim taking their own life’ (Home Office, 2022: 11). That Plan is right to acknowledge the centrality of feelings of hopelessness and isolation to victims’ suicidality (Munro and Aitken, 2020; O’Connor, 2003) in its observation that ‘it is devastating to know that those trapped by domestic abuse can feel so hopeless that they believe the only way out is suicide’ (2022: 60). But it is equally important to underscore that this is not an inevitability, and there is much that can be done through improved training, risk assessment and support provision tailored to this context.
Bates et al.'s analysis of 39 domestic abuse suicides in the year to March 2021, which were identified by police as part of a wider Domestic Homicide Project in England and Wales, is instructive here. It reveals continuities with previous research regarding the profiles of those involved in fatal domestic violence. Ninety percent of the victims were female and in almost all cases where the relationship was known, the domestic abuse was perpetrated by an intimate partner or ex-partner (94%) (2021: 50). At the same time, levels of agency contact were noticeably higher in suicide cases than amongst other domestic homicide victims, with known involvement in 74% of cases; and the perpetrators of abuse in suicide cases were found to be three times more likely to have engaged in coercive and controlling behaviour (2021: 58).
Those findings are also supported by recent studies that have explored the learning arising from domestic homicide reviews (DHRs) conducted to date in England and Wales in suicide cases where the circumstances of death give rise to a concern about a background of domestic abuse. These studies provide a powerful profile of deceased victims (the majority of whom are women) with complicated needs, including in relation to drug or alcohol misuse, who had interacted – often repeatedly – with agencies (in particular, police and mental health services). Many had long histories of victimisation involving physical, sexual, psychological and financial abuse, and a large proportion had disclosed prior suicidality, often tied to feelings of entrapment and hopelessness (Dangar et al., 2023; Monckton-Smith et al., 2022).
Cumulatively, then, this has raised increasingly urgent questions regarding, amongst other things, the ability of police officers and other agency professionals to identify risks that victims of domestic abuse may pose to themselves as a result of their victimisation, and to recognise those risks as attributable to the conduct of the perpetrator rather than to any ‘instability’ or ‘precarity’ of the victim. The appropriateness of holding perpetrators criminally liable when victims take their own lives has also become a focus of debate. In R v Dhaliwal ([2006] EWCA Crim 1139), there was a failed attempt by the Crown Prosecution Service in England and Wales to bring manslaughter charges against an abusive husband who had subjected his wife to sustained psychological abuse prior to her suicide. At the time, the case fell at the first hurdle with bodily harm being defined to exclude psychological forms of injury, such that there could be no foundational criminal act upon which to ground constructive liability. The court did observe, however, that if the CPS had relied instead on the husband's less frequent but nonetheless extant acts of physical violence, this could have provided such a foundation; and indicated that, had they done so, the court would then have been open to hearing arguments seeking to establish the necessary causal link between the husband's behaviour and the victim's death. While some commentators have argued the court in Dhaliwal took an unduly restrictive approach in their interpretation of bodily harm (Munro and Shah, 2010), what is clear is that, in the intervening period, the creation and implementation of coercive control legislation – as well as the extension of complex PTSD diagnostic criteria – has created a situation in which the doctrinal terrain is now quite different, with the obstacles posed by that first hurdle in Dhaliwal being automatically overcome upon proof of coercive control. In addition, in the subsequent case of Allen, a conviction for manslaughter was secured in respect of a victim-suicide case involving domestic abuse, albeit via return of a guilty plea. 1
Though the doctrinal landscape is thus still evolving, this opens up the possibility for perpetrators of domestic abuse to be held responsible where a victim takes their own life in the context of that abuse. Further momentum towards formally recognising the role of abuse in causing victims’ suicides has also been created by recent coronial outcomes in England and Wales, with greater emphasis being placed upon professionals who engage with victims, either during the lifetime or in the aftermath of death, to display professional curiosity, undertake effective risk assessments, and explore preventative or punitive interventions. 2
In the absence of a specific legislative provision to create this criminal liability for suicide, however, it is dependent upon the ability to establish – to appropriate standards of proof – that the perpetrator's behaviour caused the death of the victim. At the heart of this is the question of whether the victim's act of taking their own life can be understood as a voluntary one, such that it breaks the causal chain. This, in turn, engages our understanding of coercive control and its effects. After all, if we take seriously the rationale for criminalising coercive control – namely, that its effects can be insidious and significant, often depriving its victims of freedom and diminishing their sense of self, while isolating them from and encouraging their distrust of providers of support – then the extent to which suicide is an agentic act can also be contested. Though the outcome in Dhaliwal has been described by some as ‘an affront to justice’ (Stannard, 2010: 534), as noted above, the court was clear that if the initial unlawful act hurdle had been overcome (as it now would be in any case grounded upon coercive control), it would have been open to considering whether such behaviour caused the victim's death in a sufficiently proximate way to ground the construction of manslaughter liability. Such consideration commences from the position, articulated by the House of Lords in Kennedy (No 2), that ‘causation is not a single unvarying concept to be mechanically applied without regard to the context in which the question arises’. 3 Thus, it requires engaging with the dynamics and effects of coercive control in the circumstances (Munro and Aitken, 2018).
In our interviews, we probed participants’ understanding of this connection between domestic abuse and suicide, and the extent to which they felt that there ought to be, or could currently be, criminal liability. One thing that was immediately clear here was that, despite often having worked on domestic abuse for many years, justice professionals had rarely considered the existence and relevance of connections to suicide until it was raised with them. Police referred to it, for example, as a ‘really interesting’ proposition, but indicated that they would be unsure of whether or how to ‘take that forward’ (Police 6). Or, as another put it, ‘it's an interesting one and my answer is, why not, sort of thing’ (Police 7). This sense of the novelty of such considerations in the criminal justice arena was echoed by a Third Sector expert who noted that, although in her experience the links between domestic abuse and suicide were clear, ‘I cannot even know how we would start the conversation’ about liability for perpetrators because ‘it would be completely a new avenue’ (Third Sector 5).
When asked to consider it, however, several interviewees were supportive in principle of the idea that there ought to be a liability for perpetrators, given what they saw as the feasibility of the fact a victim might feel compelled to take her own life as a result of coercive control. As one police officer put it, for example, ‘I’d be supportive of that because … the coercive control that they’re under, that must feel like the only way out on some occasions’ (Police 4). Meanwhile, another observed that ‘if they’re the reason somebody's decided to take their own life, I mean 100% there should be something, they should be held accountable for that’ (Police 11), whilst a third commented that ‘if the coercive control in that context has pushed somebody to take their own life, then that perpetrator is a very dangerous person … and if we don’t hold them accountable for that … we are leaving other people at risk’ (Police 11). But, many also noted that, in the words of one police officer, ‘it would be a bloody nightmare to put together a case for that though wouldn’t it?’ (Police 11). One sheriff described it as a ‘hornet's nest’ (Sheriff 4), for example, while a prosecutor observed that ‘if somebody is driven to take their own lives, there's not necessarily going to be one factor which has led to it’ and expressed concerns about the potential – in the process of seeking criminal liability – to tarnish the deceased's character, with defence arguments focused on establishing ‘they were unstable, on all these drugs, they were constantly depressed’, and as such the death wasn’t the fault of the perpetrator (Prosecutor 2). The challenges of this when conviction requires convincing a jury who may be apt to ‘over-simplify it, because they’ve not had those lived experiences’ was also raised – ‘in principle and in theory, it sounds like something we should be considering, but in practice, it probably would be really, really difficult … for a prosecutor to say, for example, this woman committed suicide because she suffered abuse and that abuse was he didn’t let her go to the shops when she wanted, and the jury will say, well, hold on, I can’t go to the shop when I want to, but I don’t kill myself’ (Prosecutor 6).
Participants may be right to highlight these difficulties, albeit that the reference to perpetrators not allowing victims to go to the shops is unhelpfully flippant. What is striking, though, is the very high threshold for causation invoked. One officer asked, ‘how could we show that coercive control was solely the principal cause of somebody taking their life?’ (Police Scotland 1). Another observed that ‘you need that sort of line between the perpetrator and the deceased to say coercive control has caused it but as soon as you’ve got other potential strands hitting into that you’re causing some sort of doubt as to whether it was, exclusively, the coercive control’ (Police 7). This suggests a need for domestic abuse to be the cause of suicide when the legal test is less demanding, requiring that the perpetrator's behaviour be a substantial and operating cause, potentially among others; and although a voluntary act of the victim could break such a causal chain, it will not do so where it was a reasonably foreseeable consequence of the perpetrator's behaviour (Wallace [2018] EWCA Crim 690). As noted above, there is now ample evidence demonstrating that domestic abuse can reduce the actual and perceived agency of its victims, isolate them from support networks and generate a sense of hopelessness that is, in turn, pivotal to proclivities to suicidal ideation and behaviour. Indeed, the links to victim suicidality appear particularly pronounced in contexts that involve coercive and controlling behaviour, with disappointing prior engagement with services, situational and personal vulnerabilities, and the manipulation thereof by perpetrators intersecting in complex but compounding ways (Munro, 2023). In this context, as Munro and Shah argued in their feminist reimagining of Dhaliwal, ‘the abuser does not pull the trigger or provide the rope. The victim may even see the act of suicide as a form of liberation or a final expression of rebellion or subversion against a partner's control. But this does not mean that the actions of the abuser are not a significant cause of death, and nor does it mean that the act of taking one's life is a reflection of voluntary agency’ (2010: 270).
Even in cases where such a demanding threshold for causation could potentially be met, moreover, participants also indicated that they doubted the feasibility of securing any prosecutions, since without a live victim to provide contemporaneous testimony, there would be little possibility of curating a sufficiently compelling evidence base. As one officer put it, ‘it's hard because they are not there to tell you their side of the story, so again you’re relying on physical stuff, text messages, things like that … the suicide note’ (Police 8), whilst another noted ‘you would almost need that person to say he's done this and I’m going to commit suicide because of it for that direct link to be made’ (Police Scotland 2). This sits uncomfortably, however, alongside the insistences that participants also gave in our discussions about having developed the skills to gather information effectively from a variety of sources to appropriately establish patterns of coercive control (Bettinson et al., forthcoming). One officer remarked that ‘I just don’t see that we would ever be able to prove that … with such a personal crime …you never know what goes on behind closed doors’ (Police 12). But it is unclear why this would necessarily be more of a problem in cases where the victim is now deceased but has, for example, made disclosures of abuse during her lifetime than it would be in cases where the victim is alive but withdrawn cooperation from the investigation. As noted above, though there may be evidential and ethical difficulties with prosecution in that latter context, it was not presented by officers as posing insurmountable obstacles in the way it was in respect of cases that ended in victim suicide. Instead, the assumption driving responses here appeared to be that suicidal victims would have struggled in silence, without a record of disclosure or agency engagement in respect of the domestic abuse or its effects.
That will indeed have been the case for some victims, particularly given the widely acknowledged under-reporting of domestic abuse overall, which can be amplified in certain communities. Equally, the existing evidence regarding victims who take their own lives in the context of domestic abuse indicates markedly high levels of service interaction, with many ‘struggling in plain sight’ of a range of professionals (Dangar et al., 2023: 5). Indeed, a recent study of Domestic Homicide Reviews in suicide cases in England and Wales found that over half of the deceased had engaged with domestic abuse services and almost two-thirds with mental health or counselling services. Three-quarters were in regular contact with their GPs, over 90% had a history of police contact, and more than half had prior or ongoing contact with housing services often due to being in a precarious situation that was making it more difficult to leave an abusive partner, and in over one-third of cases, there was social service involvement in relation to children (Dangar et al., forthcoming, 2023). There were clear shortcomings, however, in many of these cases, in terms of the confidence and curiosity of professionals in those interactions to ask questions about domestic abuse, suicide or the connection between the two. There were also wider errors in the use of risk assessment tools, siloed agency responses, inadequate resourcing and poor protocols around multi-agency working; and in several cases, it was clear that disappointing or distressing service interactions (especially with criminal justice or social services) were a contributory factor to suicidality. This raises important questions about lessons that can be learned, through DHRs, inquests and beyond, to improve such responses, as well as about state responsibilities in this context (Bates et al., 2021; Dangar et al., 2023; Herring, 2022). But, mirroring the discussion above regarding victims’ prospects for evasive action in respect of duress, this analysis is also important in underscoring that the relevance of a history of interaction often lies less in indicating any particular confidence or trust in those services, and more in documenting a risk that may have been navigated and negotiated by victims for years prior to their suicide. Though such documentation may exist in many suicide cases, it requires to be sought out, so that the connection to abuse can indeed be evidenced: thus, any assumption by professionals that, in the absence of a victim to provide testimony, the evidential trail falters may be self-fulfilling.
Interestingly, some of the prosecutors that we spoke with were more engaged in the possibility and more optimistic about the prospects, of establishing an evidential baseline, but noted that, at the heart of doing so, would be ‘good police work’ to bring sources of information together in order to provide the narrative of ‘someone else's behaviour leading you to the point where you take your own life’ (Prosecutor 5). As we have noted elsewhere, there are questions regarding the capacity and resourcing of officers to undertake that work (Bettinson et al., forthcoming), but participants’ responses also indicated uncertainty regarding what inferences different types of evidence might appropriately carry in this context. In particular, there was a contradictory approach taken in relation to a record of reported victim mental health difficulties. While one police officer identified this as ‘the best case scenario’ (Police 5) in terms of corroborating the impact of domestic abuse during the victim's lifetime that might have precipitated suicidality, others saw it as a barrier to establishing a credible causal link to the perpetrator: ‘how do we know that that person didn’t just have underlying mental health difficulties or psychological problems? If they did, were those because of the domestic abuse or was it something else, or a combination of both?’ (Prosecutor 6). To the extent that existing knowledge regarding domestic abuse suicide does indicate the presence of mental health diagnoses amongst many victims, this is an important consideration (Dangar et al., 2023); but wrestling with it must involve the recognition that mental ill-health is often aggravated by, if not a consequence of, that domestic abuse. So too it must avoid the psychiatric pathologising of victims’ reactions to abuse when manifesting in suicidality, understanding this as a reflection of hopelessness rather than, as Magistrate 7 put it, ‘mental disorder’.
Subject to such obstacles being capable of being overcome, it was clear that the majority of participants were supportive of, or at least interested in, the possibility of pursuing perpetrators’ liability for suicide in domestic abuse cases. This could be seen to reflect cognisance of the deep impacts of coercive control on victims’ sense of isolation and capacity for agency, albeit with hesitancy in respect of the ability of existing evidential and doctrinal frameworks to recognise that reality. At the same time, however, there was a small number of interviewees who were significantly more dubious. For some, this was because they felt addressing it as an aggravating factor within existing offences would be more appropriate than seeking to establish a direct route for suicide liability. As one officer put it, for example, ‘I don’t think it changes much … he's not actually killed her, she's committed suicide, and the best we can really do is control and coercive’ (Police 3). For others, it was precisely because they felt it would ignore the responsibility that victims continued to hold for their taking their own lives and be open to abuse – one prosecutor maintained that it was important to keep in mind that ‘ultimately suicide involves an element of choice on the part of the victim’ (Prosecutor 4), whilst a sheriff underscored that ‘invariably, it is always the person who commits suicide's decision to do it’ (Sheriff 1). Meanwhile, a police officer who observed ‘the problem you have with domestic abuse is that a lot of victims need to take responsibility as well for some of their actions’ went on to suggest ‘on the flipside of it, if I lose my job tonight and go home and kill myself because my boss has sacked me … does that mean then that my boss is held accountable for killing me?’ (Police Scotland 3). Such comments display a disregard for the complex dynamics and effects of coercive control, which the same participants pointed to when it arose before them as an offence, and may highlight the precarity of victim empathy. This was reflected particularly acutely in the observation from one magistrate that ‘people would find it hard … a police officer would find it hard that anybody would take their own life because they’ve got an awkward husband’ (Magistrate 6).
As was the case in respect of victim offending, discussed in the previous section, it might be argued that the optimal long-term solution for addressing liability for domestic abuse-related suicide within the criminal justice system lies in the creation of a bespoke offence. However, it is important to bear in mind that there is in fact a great deal more that could be done, within the parameters of existing manslaughter provisions, to hold perpetrators to account. Indeed, though it has required commendable persistence from bereaved families and their advocates, recent innovations in the coroners’ court illustrate the extent to which change is possible within existing frames when a more contextualised understanding of coercive control and its effects is applied. In the criminal justice context, this requires, amongst professionals, an appreciation of the scale of the problem and nature of the relationship between abuse and suicide; a curiosity to ask questions, training to assess risk effectively and the capacity to perform trauma-informed preventive interventions during the lifetime of the victim; and, in the aftermath of their death, the motivation to investigate abuse, holistically consider its effects, and build cases creatively. Whether in the context of implementing new legislative provisions or evolution of the interpretation of existing frameworks, our findings in this study suggest that this will require a change in many professionals’ attitudes, anxieties, and assumptions.
Conclusion
Challenges undoubtedly remain, both in Scotland and in England and Wales, regarding effective identification and risk assessment of coercive or controlling behaviour, and robust investigation and prosecution of offences as a consequence. Nonetheless, our data did indicate that – particularly amongst those criminal justice professionals who had undergone specialist training in relation to psychological, emotional and financial forms of domestic abuse – there was a growing appreciation of the need for a holistic understanding of the nature and severity of harms, and complexity of behaviours, involved (Bettinson et al., forthcoming). This is to be welcomed in terms of the prospects for successful investigation and prosecution of coercive control offences. Despite this, it was also clear that, in respect of other contexts in which coercive control might arise as an issue, participants had substantially less nuanced understandings of how to navigate the complexities involved. This often resulted in a tendency to diminish the impact of such abuse on victims’ freedom and agency, whether in regard to their engaging in criminal activity or ending their lives by suicide; and often despite the fact that, in discussions around the offence itself, those same participants had been keen to emphasise the ways in which that abuse had deep and lasting impacts. Without diminishing the complex evidential and doctrinal issues at stake, we have suggested here that this divergence in approach may reflect the precarity of trauma-informed assessments of the impact of coercive control. It also demonstrates the ways in which, when individuals are no longer amenable to simplistic positioning within the mould of powerless victims, empathetic engagement with the profoundly restricted scope of their agentic action, and condemnation of the perpetrator for orchestrating that condition, can be substantially diminished. The capacity of policy, practice and personnel in the criminal justice system to recognise and respond to victims’ experiences, in all their complexity, is limited: and often in ways that undermine the prospects for justice. This was reflected poignantly in the suggestion by one officer, in respect of a hypothetical case where a victim is arrested for shoplifting and discloses her history of domestic abuse, that ‘she could wear two hats – she could wear the thief hat and the domestic abuse victim hat simultaneously, and we’d probably get different officers to deal with them but that's not an insurmountable challenge’ (Police 1). Though signposting to support in respect of the domestic abuse is no doubt welcome in this scenario, this officer's contribution powerfully misses the point that these are not two hats worn simultaneously, but one hat, one experience, one identity that the victim-suspect is being required to navigate. Justice processes might struggle to grasp that complexity, but understanding it is vital to properly reflect the circumstances under which offending and abuse occur.
In this article, we have drawn on key findings from our fieldwork to open up space for more critical reflection, across criminal justice contexts, about how coercive control and its effects are recognised and responded to. There is, no doubt, more to be done to unpack the complexities at stake. Given the distinctive moral, policy and social issues involved, as we have noted above, there may be legitimate reasons why such effects, recognised in one justice process, cannot be similarly recognised, or weighted in the same way, in another. But to the extent that there is inconsistency, we have suggested that it requires to be considered, evaluated and justified. Thus, more robust and transparent interrogation by professionals is needed around when and why context matters, and how coercion or control interrupts agentic choices, across contexts. Without this, our confidence in their assertions that the criminal justice system takes seriously the holistic impacts on victims of coercive control is in jeopardy.
Footnotes
Acknowledgement
The authors are indebted to the University of Warwick for funding, through a Faculty of Social Science Research Development Award, much of the underlying fieldwork. We are grateful to the stakeholders who participated as interviewees in the research and to key contacts in policing, the third sector, the Crown Office, Scottish Courts and Tribunals Service and the Magistrates Association who facilitated our access. The research was conducted in line with ethical approvals provided by the University of Warwick Humanities and Social Science Research Ethics Committee – reference HSSREC 69/20-21.
Author's Note
Vanessa Bettinson is now affiliated at Northumbria University, UK.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
