Understanding Family Violence. Reforming the Criminal Law relating to Homicide [2016] NZLCR 139; Understanding Family Violence. Reforming the Criminal Law relating to Homicide [2016] NZLCR 139; NZLC R139

	Home \| Databases \| WorldLII \| Search \| Feedback New Zealand Law Commission Reports

The Law Commission is an independent, publicly funded, central advisory body established by statute to undertake the systematic review, reform and development of the law of New Zealand. Its purpose is to help achieve law that is just, principled and accessible and that reflects the heritage and aspirations of the peoples of New Zealand.

The office of the Law Commission is at Level 19, 171 Featherston Street, Wellington

A catalogue record for this title is available from the National Library of New Zealand. ISBN: 978-1-877569-76-0 (Print)

This title is also available on the Internet at the Law Commission’s website: www.lawcom.govt.nz

NZLC R139 – UNDERSTANDING FAMILY VIOLENCE: REFORMING THE CRIMINAL LAW RELATING TO HOMICIDE

I am pleased to submit to you the above Report under section 16 of the Law Commission Act 1985. Yours sincerely

Understanding family violence: Reforming the criminal law relating to homicide iii

There is widespread acceptance that the incidence and level of family violence in New Zealand is a national disgrace, and that we presently need distinct and effective action on a number of aspects of this problem. This is not to decry further research, but there are important things which we can and should safely do now.

The Law Commission was asked to undertake three projects on family violence and sexual violence. The first Report (R136), which has already been tabled in Parliament, was as to alternative models for prosecuting and trying criminal cases. It is still under active consideration by the administration.

The second (R138) addresses the case for a new offence in our criminal code, of non-fatal strangulation. It too has been delivered to Parliament and is under consideration.

This third Report (R139) addresses whether the law in respect of the position of a victim of family violence who kills their abuser can be improved. It addresses the long troublesome issues of self- defence, partial defences and sentencing principles. These are all areas which raise difficult criminal law issues.

The three Reports the Commission has lodged are not the only possibilities in this unhappy area of criminology and will form part of a broader package being considered by Parliament. But they have the potential to greatly improve law relating to the position of the victims of family violence.

We are heartened by the public response to the Reports which are already in the public domain. As many commentators have already noted – our proposals “will take courage”, but we are firmly of the view that real change is needed in this unhappy area of human interaction.

We are grateful to all the people and organisations that provided input during the preparation of this Report. We would particularly like to thank all members of our expert panel. Their advice, contributions and cumulative years of experience were crucial in highlighting key issues relevant to this reference. Members of the expert panel were Dr Ang Jury, Annabel Markham, Judge Caren Fox, Associate Professor Elisabeth McDonald, Helen Cull QC, John Billington QC, Associate Professor Julia Tolmie, former High Court Judge the Hon Ronald Young and Superintendent Tusha Penny. A list of those who provided submissions on the Issues Paper can be found at Appendix B of this Report.

The lead Commissioner responsible for this reference was the Hon Dr Wayne Mapp. The Hon Douglas White QC and Helen McQueen held warrants as Law Commissioners from 9 February 2016. The legal and policy advisers were Bridget Fenton, Nichola Lambie and Jacob Meagher.

Understanding family violence: Reforming the criminal law relating to homicide v

Foreword .............................................................................................................................. iv Acknowledgements ............................................................................................................... v Summary ............................................................................................................................... 4

Introduction .......................................................................................................................................... 4

A better understanding of family violence .......................................................................................... 5

A change to the law of self-defence ..................................................................................................... 7

Greater recognition of reduced culpability .......................................................................................... 9

No partial defence to murder ............................................................................................................	10
Changes to sentencing for homicide .................................................................................................	11
Conclusion ..........................................................................................................................................	13
_{PART 1 Setting the scene .....................................................................................................}	₁₅
Chapter 1 Scope, context and approach .............................................................................. 16
Introduction .......................................................................................................................................	16
Context of this review .......................................................................................................................	17
Scope of this project ...........................................................................................................................	19
Our approach .....................................................................................................................................	21
Chapter 2 Understanding family violence ..........................................................................	23
Introduction .......................................................................................................................................	23
The different forms of family violence .............................................................................................	23
Evolving understanding of intimate partner violence .....................................................................	26
Contemporary understanding of family violence ............................................................................	28
Improving understanding of family violence ...................................................................................	36
Chapter 3 The legal context ................................................................................................	39
Introduction .......................................................................................................................................	39
Homicide law .....................................................................................................................................	39
Summary of our case review .............................................................................................................	45
Conclusion ..........................................................................................................................................	50
Chapter 4 The law reform context .....................................................................................	51
Introduction .......................................................................................................................................	51
Previous New Zealand Law Commission Reports ...........................................................................	51
Reform in other jurisdictions ............................................................................................................	56
PART 2 Self-defence .............................................................................................................	65

Chapter 5 Self-defence in New Zealand .............................................................................. 66

Understanding family violence: Reforming the criminal law relating to homicide 1

Introduction .......................................................................................................................................	66
The theory of self-defence .................................................................................................................	66
History of self-defence in New Zealand ...........................................................................................	67
Operation of self-defence in New Zealand .......................................................................................	68
Chapter 6 Self-defence and family violence – is there a problem? .....................................	71
Introduction .......................................................................................................................................	71
The role of gender in the development of self-defence ....................................................................	72
Persisting misconceptions .................................................................................................................	75
Imminence of the threat and lack of alternatives ............................................................................	75
Proportionality ...................................................................................................................................	82
Self-defence in other jurisdictions ....................................................................................................	85
The role of battered woman syndrome evidence .............................................................................	86
Submitters’ views on whether there is a problem ............................................................................	89
The Commission’s view .....................................................................................................................	90
Chapter 7 Proposals to reform self-defence ........................................................................	93

Introduction ....................................................................................................................................... 93

Proposals for substantive reform ...................................................................................................... 93

Proposals for procedural reform ..................................................................................................... 104

Other procedural matters ................................................................................................................ 110

PART 3 Recognising reduced culpability ........................................................................... 115

_{Chapter 8 The conceptual framework for reduced culpability
........................................ 116}

Introduction .....................................................................................................................................	116
Setting the scene: Two illustrative cases ........................................................................................	117
Options for recognising reduced culpability ..................................................................................	118
Conclusion ........................................................................................................................................	121
Chapter 9 Observations from the cases ............................................................................	123
Introduction .....................................................................................................................................	123
Conviction outcomes .......................................................................................................................	123
Sentencing outcomes .......................................................................................................................	134
Conclusion ........................................................................................................................................	136
Chapter 10 Is a partial defence justified? .........................................................................	137
Introduction .....................................................................................................................................	137
overview of the submissions ...........................................................................................................	137
Arguments for a partial defence .....................................................................................................	138
Arguments against a partial defence ...............................................................................................	143
What would a partial defence look like? ........................................................................................	145
A specific homicide offence .............................................................................................................	157
Conclusions and recommendation ..................................................................................................	159

_{Chapter 11 Sentencing for homicide
.................................................................................
161}

Introduction .....................................................................................................................................	161
The approach of the Sentencing Act 2002 .....................................................................................	161
Overview of the submissions ..........................................................................................................	162
The purposes of sentencing .............................................................................................................	162
Mitigating factors under the Sentencing Act .................................................................................	163
Specific issues in sentencing for murder ........................................................................................	171
Sentencing for manslaughter ..........................................................................................................	175
The three strikes provisions ............................................................................................................	175
APPENDICES .......................................................................................................................	181
Appendix A Terms of reference ........................................................................................	182
Appendix B List of submitters on the Issues Paper .........................................................	184
Appendix C Recommendations .........................................................................................	185
Appendix D Glossary of terms ..........................................................................................	187

Understanding family violence: Reforming the criminal law relating to homicide 3

³When a victim of family violence kills their abuser, they will usually be charged with murder.

In only some of these cases, however, will the defendant be able to successfully rely on self- defence under the Crimes Act 1961. In reality these defendants are more likely to be convicted of the lesser crime of manslaughter, than be convicted of murder or acquitted on the basis of self-defence.

The Law Commission will re-consider whether the law in respect of a victim of family violence who commits homicide can be improved. As part of this review the Law Commission shall consider:

Should the test for self-defence, in section 48 of the Crimes Act, be modified so that it is more readily accessible to defendants charged with murder who are victims of family violence; and

Whether a partial defence for victims of family violence who are charged with murder is justified and if so in what particular circumstances; and

Whether current sentencing principles properly reflect the circumstances of victims of family violence who are convicted of murder?

1 Law Commission Victims of Family Violence Who Commit Homicide (NZLC IP39, 2015).

3 Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001).

5 Ministry of Justice Strengthening New Zealand’s legislative response to family violence: A public discussion document (Wellington, August 2015) at

Understanding family violence: Reforming the criminal law relating to homicide 5

intimate partner homicides, it is the primary victim of the past abuse who kills the predominant aggressor.

6 The Family Violence Death Review Committee reported that between 2009 and 2014 there were 85 intimate partner homicides, 71 of which had a known abuse history. Of those 71 homicides, 13 (18 per cent) of offenders were primary/suspected primary victims, and 58 (82 per cent) were predominant/suspected predominant aggressors. Family Violence Death Review Committee submission at 9. All 13 primary/suspected primary victim homicide offenders were women.

7 Or less than 10 per cent of all family violence homicides. This is based on the Family Violence Death Review Committee’s identification, in the period 2009–2012, of nine cases where a primary victim killed a predominant aggressor and one further suspected case, out of a total of 126 family violence homicides and 297 total homicides. Family Violence Death Review Committee Fourth Annual Report: January 2013 to December

10 Family Violence Death Review Committee Fifth Report: January 2014 to December 2015 (Health Quality & Safety Commission, February 2016)

11 State of Victoria Royal Commission into Family Violence: Summary and recommendations (Parl Paper No 132, March 2016) at 27.

therefore recommends (continued) education to improve understanding within the criminal justice system of the dynamics of family violence.

^R1Judges should continue to receive education, including through the Institute of Judicial Legal

^R2Regular and ongoing education courses on the dynamics of family violence should be made available to all criminal lawyers, including Crown prosecutors and defence counsel.

_·^{reflect contemporary social science understanding of
family violence and victims’}

_·^{explain the primary victim/predominant aggressor
analysis in intimate partner violence;}

_·identify common misconceptions of family violence that persist today and their implications in the criminal justice system.

²¹It should be noted that section 48 applies to all occasions when recourse is made to self-defence;

Every one is justified in using, in defence of himself or herself or another, such force as, in the circumstances as he or she believes them to be, it is reasonable to use.

Understanding family violence: Reforming the criminal law relating to homicide 7

48 may apply even if that person is responding to a threat that is not imminent, provided that the defendant believed their actions to be necessary, and the response was otherwise reasonable. This would apply where a victim of intimate partner violence kills their abusive partner, as well as in other family violence contexts, such as where a person kills a parent or partner of their parent in response to family violence. To avoid the unintended consequence of violent offenders being able to take advantage of the change of law, this amendment is only made available to victims of family violence. It applies in all circumstances where self-defence is relevant and would not be limited to charges of homicide.

14 Afamasaga v R [2015] NZCA 615. See also the observations of the Supreme Court in declining leave to appeal in Vincent v R [2016] NZSC 15.

15 Stella Tarrant “Something is Pushing Them to the Side of Their Own Lives: A Feminist Critique of Law and Laws” (1990) 20 UWAL Rev 573 at 598.

gain a proper understanding of all of the circumstances of the alleged offending. We therefore make a recommendation to identify in legislation the full range of evidence of prior family violence and expert evidence that may be relevant to a jury’s assessment of whether a defendant who is a victim of family violence was impelled to act in the way they did. Expert evidence is likely to assist by explaining the social context of the homicide rather than by focusing on the outdated and much criticised battered woman syndrome.

^R5A new provision should be inserted into the Crimes Act 1961 to ensure that, where a person is responding to family violence, section 48 may apply even if that person is responding to a threat that is not imminent.

^R6The Ministry of Justice should consider whether the term “family violence” should be consistent with the definition of domestic violence in the Domestic Violence Act 1995, incorporating any amendments that may be made following the Ministry of Justice’s current review of domestic violence legislation, or whether an inclusive definition of family violence is preferred, including, but not limited to, the definition of domestic violence in the Domestic Violence Act 1995.

^R7The Evidence Act 2006 should be amended to include provisions based on sections 322J and

322M(2) of the Crimes Act 1958 (Vic) to provide for a broad range of family violence evidence to be admitted in support of claims of self-defence and to make it clear that such evidence may be relevant to both the subjective and objective elements in section 48 of the Crimes Act 1961.

16 The concern of the principle of “fair labelling”, Ashworth and Horder explain, is “to see that widely felt distinctions between kinds of offences and degrees of wrongdoing are respected and signalled by the law, and that offences are subdivided and labelled so as to represent fairly the nature and magnitude of the law-breaking”: Andrew Ashworth and Jeremy Horder Principles of Criminal Law (7th ed, Oxford University Press, Oxford, 2013) at 77.

Understanding family violence: Reforming the criminal law relating to homicide 9

^R8The Solicitor-General should, when next reviewing the Solicitor-General’s Prosecution

Guidelines, consider whether they should include express reference to the potential relevance of a defendant’s history as a victim of family violence.

^._{“Defence-based” partial defences, which normally
require an honest belief that the}

_{defendant’s actions were necessary to defend or preserve him or
herself or another.}

^._{“Provocation-based” partial defences, which
require a loss of control, triggered by}

provocation sufficient to deprive a person of ordinary tolerance, in a similar position, of the power of self-control.

^._{“Diminished capacity-based” partial defences,
which require an abnormality of mental}

function that impaired the defendant’s capacity to understand events and to judge whether their actions are right or wrong, or to exercise self-control.

R9 _{No new partial defence or separate homicide offence should be
introduced in New Zealand.}

17 We adopt the same conclusions as the Law Commission did in Review of Part 8 of the Crimes Act 1961: Crimes Against the Person (NZLC R111,

2009) at 30–31. See also Law Commission Strangulation: The Case for a New Offence (NZLC R138, 2015) at [1.17].

18 See Simon France (ed) Adams on Criminal Law – Sentencing (online looseleaf ed, Thomson Reuters) at [SA9.01]–[SA9.28].

Understanding family violence: Reforming the criminal law relating to homicide 11

relevant, however, the offender’s diminished intellectual capacity or understanding of the situation will also be taken into account and the Court of Appeal has accepted that the psychological effects of sustained abuse may justify a sentencing discount.¹⁹

102); second, finite sentences for murder; and third, the so-called three strikes regime (provided for in sections 86A–86I).

22 The only case in which a shorter sentence was imposed is R v Law (2002) 19 CRNZ 500 (HC), a euthanasia case.

strike offences include sexual offences, most violent offences and offences against property where a weapon is used. These offences range in seriousness.

_·amending section 9(2)(c) to clarify that “conduct of the victim” includes prior family violence against the offender; and

_·amending section 9(2)(e) to clarify that “diminished intellectual capacity or understanding” includes any impairment resulting from being subject to family violence.

The Ministry of Justice should undertake further policy work to address the issues noted in this Report in relation to sections 86D(4) and 86E of the Sentencing Act 2002 as they apply to homicide offenders in exceptional circumstances and specifically:

_·consider the position of victims of family violence who kill their abusers in situations where the three strikes regime would mandate a life sentence; and

_·consider how to amend the legislation to allow judges to impose a finite sentence in deserving cases.

– whether or not the defendants were acting in self-defence – in circumstances most of us do not fully understand and will never experience. This Report recommends changes to the law

24 Those cases are R v Brown HC Napier CRI-2008-020-3130, 24 November 2009; R v Stone HC Wellington CRI-2005-078-1802, 9 December 2005;

25 We understand that the Court of Appeal is scheduled to hear appeals by the Solicitor-General against the sentences imposed in R v Harrison [2014] NZHC 2705 (in which the offender’s first strike offence was an indecent assault) and R v Turner [2015] NZHC 189 (in which the offender’s first strike offence was wounding with intent), on 9 and 10 June 2016.

Understanding family violence: Reforming the criminal law relating to homicide 13

and measures to improve understandings of family violence, appreciating that legal reform is meaningful only when accompanied by shifts in thinking.

If there is a lesson to be learned from the repeal of the provocation defence it may be that legal rules, and therefore law reform, make less of a difference than we might expect in the resolution of criminal cases – or certainly criminal cases that involve the “wicked” problem of family violence ... [I]f the same unexamined assumptions that are typically made in these kinds of cases continue to be made (that women’s use of violence in intimate partnership simply mirrors what we know about men’s use of violence; that help is readily available to those victims who are facing dangerous and potentially lethal violence who seek it; and that leaving a violent relationship is always an option and is an effective means of addressing the violence) and if we fail to understand the manner in which coercive control, social marginalisation and structural exclusion entrap victims of family violence, then it is unlikely that reforms to the legal requirements of self-defence will effect much substantive change either.

26 Julia Tolmie “Defending Battered Defendants on Homicide Charges in New Zealand: The Impact of Abolishing the Partial Defences to Murder” [2015] NZ L Rev 649 at 681 (footnote omitted).

Family violence destroys lives and takes a significant toll on New Zealand society. New Zealand has the highest reported rate of family violence in the developed world,²⁷and nearly half of all homicides are related to family violence.²⁸Disproportionately, family violence affects the lives of women and children,²⁹and women are overwhelmingly more likely to be killed by an intimate partner than commit homicide themselves.³⁰The consequences of family violence can be devastating both for the victims and their families. Discussing intimate partner violence, Jane Maslow Cohen writes:³¹

Terrible and tragic things happen within the contexts of battering relationships, even beyond the violence and resultant injury itself. These tragedies include the death of the battered victim; the physical and psychological abuse of others, especially children, within the household; the destruction of employment situations and opportunities; the withering away of basic trust, particularly trust in intimacy; and, often, the waste of what might, and should, have been rewarding and productive lives.

Unsurprisingly, most family violence homicides are committed by those who have a history of perpetrating violence, usually against the deceased.³²Most homicide offenders are men.³³
In a very small number of cases, accounting for less than five per cent of all homicides in New Zealand, a victim of family violence kills their abuser.³⁴In intimate partner killings, these offenders are almost always women.³⁵Whatever their gender or relationship to an aggressor, however, victims of family violence who kill their abusers have typically suffered years of physical, sexual and/or psychological abuse that can be severe and extreme.

In this Report, we consider how the law responds to victims of family violence who commit homicide and whether the law can be improved.

27 Ministry of Justice Strengthening New Zealand’s legislative response to family violence: A public discussion document (Wellington, August 2015).

28 There were 139 family violence-related homicides in the period 2009–2012 out of a total 297 homicides. See Family Violence Death Review Committee Fourth Annual Report: January 2013 to December 2013 (Health Quality & Safety Commission, June 2014) at 32 and 35; and Ministry of Justice, above n 27.

29 State of Victoria Royal Commission into Family Violence: Summary and recommendations (Parl Paper No 132, March 2016) at 18.

30 Between 2009 and 2012 three-quarters of intimate partner homicide offenders were men, and almost three-quarters of victims were women.

31 Jane Maslow Cohen “Regimes of Private Tyranny: What Do They Mean to Morality and for the Criminal Law?” (1995) 57 U Pitt L Rev 757 at

32 The Family Violence Death Review Committee reported that, for the period 2009–2014, the homicide offender was the predominant/suspected predominant aggressor in 82 per cent of intimate partner homicides where the abuse history was known: Family Violence Death Review Committee submission at 9.

33 In the past 10 years in New Zealand, 84 per cent of homicide offenders were male: Statistics New Zealand “Adults convicted in court by sentence type – most serious offence fiscal year” <nzdotstat.stats.govt.nz>.

34 Or less than 10 per cent of family violence homicides. This is based on the Family Violence Death Review Committee’s identification, in the period 2009–2012, of nine cases where a female primary victim killed a male predominant aggressor and one further suspected case, out of a total of 126 family violence homicides and 297 total homicides: Family Violence Death Review Committee, above n 28, at 34 and 75. An earlier study undertaken by the Ministry of Social Development identified two cases in the five years from 2002–2006 of a female against male homicide where there was documented evidence of the male’s violence towards the female in the past and in the context of the event. This accounted for 1.5 per cent of family violence deaths and 0.7 per cent of total homicides for that same period: Jennifer Martin and Rhonda Pritchard Learning from Tragedy: Homicide within Families in New Zealand 2002-2006 (Ministry of Social Development, April 2010) at 38.

35 The Family Violence Death Review Committee identified 13 intimate partner homicides between 2009 and 2014 in which a victim of intimate partner violence killed their abuser, all of whom were women: Family Violence Death Review Committee submission at 9.

The impetus for this Report is a recommendation from the Family Violence Death Review

Committee (FVDRC) that the Government:³⁶
^._{considers modifying the test for self-defence set out in section 48
of the Crimes Act 1961 so}

_{that it is more readily accessible to victims of family violence;
and}
^._{considers the introduction of a partial defence
to murder that can be utilised by primary}

victims of family violence who are not acting in self-defence at the time they retaliate in response to the abuse they have experienced.

These recommendations were made in the FVDRC’s Fourth Annual Report, published in June

2014, which reported that, over the period 2009–2012, there were 126 family violence deaths. The FVDRC identified that 10 family violence deaths involved a killing by a victim of family violence of their abusive partner.³⁷The defendants in all those cases were women. A further three deaths involved killings by children who had been abused by fathers or step-fathers and had witnessed their mothers being abused.³⁸

The FVDRC observed that victims of family violence who kill their abusers are usually charged with murder and, if convicted, can face long terms of imprisonment rather than having their long-term history of victimisation and, at times, extreme abuse recognised in the criminal justice response to their crimes.³⁹The FVDRC concluded, as a result, that victims of family violence who kill their abusers are not well served by the New Zealand justice system. The FVDRC observed that, compared with similar international jurisdictions:⁴⁰

... Aotearoa New Zealand is out of step in how the criminal justice system responds to [intimate partner violence] primary victims when they face homicide charges for killing their abusive partners

... Firstly, it can be attributed to the fact that the defence of self-defence has been interpreted in a restrictive manner in Aotearoa New Zealand, making it difficult to apply in cases involving primary victims. Secondly, by abolishing provocation New Zealand now has no partial defences to murder for those primary victims whose circumstances do not fit within the full defence of self-defence. These defendants will now be convicted of murder rather than manslaughter. And thirdly, Aotearoa New Zealand retains a presumption of life imprisonment for murder, which is difficult to overturn even in such cases and, when it is overturned, still results in long sentences of imprisonment. As such the violent circumstances (that offenders who were [primary victims] were entrapped in and responding to) do not appear to be reflected in local verdicts to the same degree as they are in comparable international

Also in 2014, the Minister of Justice and the Minister for Social Development announced that family violence and sexual violence are amongst their top priorities and undertook a package of

37 At 42. This included one suspected case. In its submission on our Issues Paper the Family Violence Death Review Committee reported that it had identified a further three intimate partner homicides committed by a victim of family violence between 2012 and 2014: Family Violence Death Review Committee submission at 9.

38 At 65. The circumstances of the three intrafamilial deaths reviewed by the Family Violence Death Review Committee are not clear. In particular, it is unclear whether the deceased was the abusive father/step-father, or another family member. If they were not, those cases are outside our terms of reference. As we discuss in Chapter 2, however, two of the 24 cases we reviewed involved children killing parents.

Understanding family violence: Reforming the criminal law relating to homicide 17

initiatives.⁴¹One of the key initiatives was publication in August 2015 of a discussion document about strengthening the legal response to family violence. Within the scope of the review are the Domestic Violence Act 1995, the Care of Children Act 2004, the Crimes Act 1961, the Bail Act 2000 and the Sentencing Act 2002.

Submissions on that document closed in September 2015, and a report is likely in mid-2016.

This review is one of three family violence or sexual violence projects referred to the

Commission.⁴²The other two were:
^._{Alternative models for prosecuting and trying criminal
cases.}⁴³_{We were asked to}

identify best practice for improving the court experience for complainants, with a particular focus on sexual offence cases. We recommended:

^._{establishing an alternative process outside of court for
resolving sexual violence cases}

(such as in situations where the victim wants to maintain a relationship with the perpetrator);

^._{improving the court experience for victims of sexual
offending by mechanisms such}

as statutory time limits to ensure speedier case resolution, introducing less traumatic methods for complainants to give evidence at trial, and piloting a specialist sexual violence court with expert judges and lawyers who are trained and accredited in dealing with sexual violence cases; and

^._{establishing a sexual violence commission to coordinate and
oversee wraparound care}

and services for victims of sexual violence.

This reference was tabled in Parliament in December 2015.
^._{Creation of a separate
crime of non-fatal strangulation.}⁴⁴_{In that Report,
we examined}

the case for a new offence of strangulation. The Commission found that the risks following strangulation are not well understood by Police, judges and others who assist victims of family violence. In at least half of all cases, strangulation does not result in an obvious external injury, which makes it difficult to lay serious charges against the perpetrator and means that perpetrators of strangulation are often getting a much lower sentence than they deserve. The Commission makes seven recommendations to improve the way that the criminal justice system responds to strangulation in family violence circumstances, including a recommendation that a specific offence of strangulation should be enacted.

This reference was tabled in Parliament in March 2016.

41 Cabinet Paper “Progress on the Work Programme of the Ministerial Group on Family Violence and Sexual Violence” (22 July 2015) SOC 15/

42 The reports are available to download from the Law Commission’s website at <www.lawcom.govt.nz>.

43 Law Commission The Justice Response to Victims of Sexual Violence: Criminal Trials and Alternative Processes (NZLC R136, 2015).

44 Law Commission Strangulation: The Case for a New Offence (NZLC R138, 2016).

The terms of reference for this project, which are in Appendix A, require us to consider whether the law in respect of victims of family violence who commit homicide can be improved, including whether:
^._{the
test for self-defence, in section 48 of the Crimes Act 1961, should be modified
so that it}

_{is more readily accessible for victims of family
violence;}
^._{a partial defence to murder is justified
for victims of family violence and, if so, in what}

_{circumstances; and}
^._{current sentencing
principles properly reflect the circumstances of victims of family
violence}

In this Report, we use the general term “victim of family violence” to refer to all persons who have suffered abuse, regardless of whether that person is an adult or child or is being abused by a partner, parent or some other family member. Similarly, our use of “abuser” refers to all perpetrators of family violence, regardless of their relationship with the victim.

We use the term “family violence” in a broad sense. Consistent with the FVDRC, we adopt the definition of the Taskforce on Violence within Families, which defines family violence as:⁴⁵

... a broad range of controlling behaviours, commonly of a physical, sexual and/or psychological nature, which typically involve fear, intimidation and emotional deprivation. It occurs within a variety of close interpersonal relationships, such as between partners, parents and children, siblings and in other relationship where significant others are not part of the physical household but are part of the family and/or are fulfilling the function of family.

At times in this Report, it is necessary to refer to the different forms of family violence. We use the categorisation adopted by the FVDRC, which separates family violence into three forms of abuse:⁴⁶
^._{Intimate partner violence: any
behaviour within an intimate relationship (including}

current and/or past live-in relationships or dating relationships) that causes physical, psychological or sexual harm to those in the relationship.
^._{Child abuse and neglect: all forms of
physical and emotional ill-treatment, sexual abuse,}

neglect and exploitation that results in actual or potential harm to the child’s health, development or dignity.
^._{Intrafamilial violence: all forms of abuse between family
members other than intimate}

partner violence and child abuse and neglect, including abuse/neglect of older people aged approximately 65 years and over by a person with whom they have a relationship of trust, violence perpetrated by a child against their parent, violence perpetrated by a parent on their adult child, and violence among siblings.

45 Family Violence Death Review Committee, above n 28, at 13, citing Ministry of Social Development "Background to family violence indicators"

Understanding family violence: Reforming the criminal law relating to homicide 19

Throughout this Report, we also use the terms “primary victim” and “predominant aggressor” to describe the dynamics of intimate partner violence.⁴⁷We adopt these terms from the FVDRC, which applies a “primary victim/predominant aggressor analysis” in its review of intimate partner homicides.⁴⁸This analysis requires looking at patterns of abuse and harm across the entirety of an intimate partner relationship. The “predominant aggressor” is the person “who is the most significant or principal aggressor ... and who has a pattern of using violence to exercise coercive control”.⁴⁹The “primary victim” is the person “who (in the abuse history of the relationship) is experiencing ongoing coercive and controlling behaviours from their intimate partner”.⁵⁰The primary victim/predominant aggressor analysis is discussed further in Chapter 2.

These key terms and others are also set out in a glossary in Appendix D of this Report.

While our terms of reference refer to victims of family violence who “commit homicide”, in this Report, we focus on victims of family violence who kill perpetrators of family violence (abusers). We do not consider the law in relation to victims of family violence who kill third parties, although we recognise in some cases that offending may be caused or explained in part by the defendant’s prior history of abuse.⁵¹We have taken this approach for several reasons. First, the FVDRC’s Fourth Annual Report, which led to this reference, raised concerns in relation to primary victims who kill predominant aggressors, rather than primary victims who kill generally.⁵²Second, the problems identified in this Report in relation to self-defence and reduced culpability arise by virtue of the dynamics between the homicide defendant as a victim of family violence and the deceased as a perpetrator of family violence. In most of the cases we have considered, the homicide defendants were primary victims of intimate partner violence by the deceased. Sometimes, however, a person other than a primary victim may kill an abuser.⁵³
Accordingly, in this Report, we also consider victims of family violence who kill in response to other forms of family violence, namely child abuse and neglect, and intrafamilial violence. What will be relevant in any case is the link between the homicide and the prior abusive conduct of the deceased.⁵⁴

This Report is not a first principles review of the law of homicide or defences to homicide. We are limited in our consideration of the law to how it applies to victims of family violence. We do not, therefore, consider the law in respect of other defendants who may also be less blameworthy in a comparative sense. We further note that this reference is limited to the offence of homicide. It does not cover other offences that may be responses to family violence,

47 Most of the cases we discuss involve intimate partner violence, but two, R v Erstich [2002] NZCA 122; (2002) 19 CRNZ 419 (CA) and R v Raivaru HC Rotorua CRI-2004-077-1667, 5 August 2005, involve children who killed violent parents. We are also aware of a third case, before the High Court at the time of publication of this Report, involving the killing by a 19 year old, Daryl Kirk, of her mother’s partner. At the time of publication of this Report Ms Kirk had been found guilty of manslaughter but had not yet been sentenced. The term “primary victim” is less obviously appropriate for these cases, in part because the offender appears more likely to be reacting to the abuse of another (for example, a mother) in addition to themselves (see, for example, Raivaru at [6]–[7] and [19]). Thus, where appropriate, we may describe defendants who kill abusers as simply “victims” of family violence.

51 For example, in R v Rongonui [2000] NZCA 273; [2000] 2 NZLR 385 (CA), the defendant, Janine Rongonui, was ultimately convicted of manslaughter in relation to the killing of her neighbour, who she stabbed 10 times after the neighbour declined to babysit her children. Ms Rongonui had been under extreme social and financial pressure, was suffering from a major depressive episode triggered by recent and historical violence, and was brain damaged as a result of long-term physical and chemical abuse.

53 For example, R v Raivaru, above n 47, in which the primary victim was apparently the defendant’s mother. (See further the explanation above at n 47).

54 See further the discussion of mitigating features for sentencing in Chapter 11.

such as assault or failure to protect a child. We are conscious, however, of the potential for unintended consequences that necessarily arises whenever the law is of specific, rather than general, application.

In this Report, we focus our consideration on:
^._{the law
of self-defence, and how it applies when a person commits homicide in response
to}

_{family violence (Part 2 of this Report); and}
^._{how the law can and should recognise the reduced culpability of
victims of family violence}

who kill their abusers other than in self-defence (including when a plea of self-defence is unsuccessful) (Part 3 of this Report).

Part 3 considers in detail the interrelated issues of partial defences to murder and sentencing reform, both of which can be utilised to recognise the reduced culpability victims of family violence.

We rely on the data collected and reported by the FVDRC to supplement our own analysis of reported cases. The FVDRC is a statutory “mortality review committee” established by the Health Quality & Safety Commission.⁵⁵It was established to review and report on family violence deaths in New Zealand.

We have also drawn extensively on law reform work in other jurisdictions, particularly the Australian states of Victoria, Western Australia, New South Wales, Queensland and Tasmania, as well as England and Wales, Canada and Ireland.

We explained in the Issues Paper that we have framed our examination of the problems in this area of the law, and our analysis of the options for reform, by using the following guiding principles:⁵⁶
^._{The
law of homicide should reflect community values. The sanctity of and right
to life}

_{is inherent in our legal system, and the law of self-defence must
reflect this core principle.}
^._{The law should
achieve substantive equality for all defendants. The law should
apply}

equally irrespective of whether a person is a victim of family violence and should strive to be free from any form of gender or other bias. In this context, we do not mean neutrality of the law on its face but rather substantive equality in the manner in which the law operates. For substantive gender equality to be achieved, the law must not have a discriminatory effect in practice.⁵⁷
^._{The law of homicide should
reflect the context in which homicides typically
occur.}⁵⁸

Reform grounded in abstract philosophical principles or historical legal categories is not useful or realistic. Any reform must be driven by an understanding, based on contemporary social science, of the actual context in which victims of family violence commit homicide.

56 Law Commission Victims of Family Violence Who Commit Homicide (NZLC IP39, 2015) [Issues Paper] at [1.28].

58 A similar approach was also adopted by the Law Reform Commissions in Victoria and Western Australia. See Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 14–15; Jenny Morgan Who Kills Whom and Why: Looking Beyond Legal Categories (Victorian Law Reform Commission, Melbourne, 2002) at 1–2; and Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007) at 9.

Understanding family violence: Reforming the criminal law relating to homicide 21

These principles represent the underlying objectives of our recommendations for reform and are consistent with New Zealand’s international obligations and the New Zealand Bill of Rights Act 1990.

The Commission convened an expert panel to advise it on this reference. The panel was made up of academics, Crown and defence counsel, current and retired judges, victim advocates and Police.

Our Issues Paper was published in November 2015,⁵⁹with submissions open until late December 2015. Due to the relatively condensed timeframe for submissions, we accepted late submissions where requested.

We also held consultation meetings with key stakeholders, including (but not limited to) victim support organisations, the Ministry for Women, the FVDRC, Crown solicitors, the New Zealand Law Society, the Criminal Bar Association, the Public Defence Service, legal academics and experts in family violence.

This Report is divided into three parts.

Part 1 (Chapters 1 to 4) sets the scene of this review. Chapter 1 sets out the scope of the review and our approach to it, and Chapter 2 provides an overview of the present state of knowledge of the dynamics of family violence and how this has evolved over time. In that chapter, we draw on the FVDRC’s Fifth Report, which calls for a shift in how we think about intimate partner violence and victims’ responses. Chapter 3 explains the law of homicide and the criminal trial process and makes some observations as to how the law currently responds to victims of family violence who kill their abusers. Chapter 4 sets out the law reform context to this Report, which includes the history of reform in this area in New Zealand and in other comparable jurisdictions.

Part 2 (Chapters 5 to 7) focuses on the law of self-defence. These chapters set out the current law and explore the problems that arise in applying the law to victims of family violence, before going on to analyse the options for reform.

Part 3 (Chapters 8 to 11) considers how the criminal justice system takes into account the reduced culpability of defendants who kill their abusers in response to family violence, other than in self-defence. Chapter 8 introduces the topic and sets out the options for addressing reduced culpability. Chapter 9 makes a number of observations around how the reduced culpability of a victim of family violence is currently taken into account. Chapter 10 considers and concludes on the question of whether a partial defence is justified, and Chapter 11 goes on to explore the options for sentencing reform.

Understanding why victims of family violence kill their abusers requires a broader understanding of the general nature and dynamics of family violence. Otherwise, the individual circumstances and abuse history in any case can be misinterpreted or minimised, which can affect case outcomes. How New Zealanders think about family violence also drives broader public policy decisions, including decisions around family violence intervention, prevention and punishment for family violence-related offending.

As we explain in this chapter, understandings of family violence, within the community and among practitioners, are evolving. Recognising that previous understandings have propagated erroneous assumptions and misunderstandings, the Family Violence Death Review Committee (FVDRC) has called for a change to the way New Zealand understands intimate partner violence and victims’ responses.⁶⁰

We agree with the FVDRC that understanding of family violence needs to change. Underlying the legal issues discussed in Parts 2 and 3 of this Report are misconceptions that are outdated and unhelpful. The result is that victims of family violence may not be recognised as such (particularly if they have, on occasion, been violent too), they may be seen as being to blame for not having left the relationship, or the threat posed, and the overall impact of the violence on the victim may not be fully appreciated.

In this chapter, we:
^._{identify the
different forms of family violence occurring in different relationship
types;}
^._{discuss evolving understandings of intimate
partner violence and victims’ responses to}

_{violence and some misconceptions that these understandings have
perpetuated;}
^._{explain contemporary social science
understanding of family violence, including the}

_{significance of factors such as gender, race, economic deprivation and
gang association; and}
^._{make recommendations targeted
at improving the understanding of family violence by those}

Family violence comes in many forms, including intimate partner violence, child abuse and neglect, elder abuse, violence by children towards parents, and sibling violence. Family violence homicides are, however, most common in intimate partner relationships.⁶¹Intimate partner violence is defined as “[a]ny behaviour within an intimate relationship (including current and/

60 Family Violence Death Review Committee Fifth Report: January 2014 to December 2015 (Health Quality & Safety Commission, February 2016)

61 Among the 126 deaths the Family Violence Death Review Committee reviewed for its Fourth Annual Report, 63 (50 per cent) were intimate partner deaths, 37 (29 per cent) were child abuse and neglect deaths, and 26 (13 per cent) were intrafamilial deaths. See Family Violence Death Review Committee Fourth Annual Report: January 2013 to December 2013 (Health Quality & Safety Commission, June 2014) at 32.

Understanding family violence: Reforming the criminal law relating to homicide 23

or past live-in relationships or dating relationships) that causes physical, psychological or sexual harm to those in the relationship”.⁶²

Historically, consideration of family violence has focused on male aggression towards women and “battered woman syndrome” developed alongside the organised women’s movement. Family violence in intimate partner relationships is a gendered phenomenon. Perpetrators of violence are usually men, and victims are usually women and children.⁶³In this context, it needs to be recognised that men and women kill their intimate partners for different reasons and in different ways.⁶⁴

Family violence does, however, occur within other intimate relationships. The FVDRC identified one case of intimate partner homicide occurring in a same-sex relationship but reports that same-sex family violence deaths are likely to be undercounted.⁶⁵Intimate partner violence in lesbian, gay, bisexual, transgender, queer and intersex relationships has received less attention in the literature,⁶⁶but there is some evidence that indicates it may be as prevalent as heterosexual violence.⁶⁷Some contend the dynamics of same-sex intimate partner violence are similar to those in heterosexual relationships,⁶⁸while others suggest they may be different in material ways.⁶⁹In any event, it is widely acknowledged that further research is required.

Concepts and models that have traditionally been applied to women in the context of intimate partner relationships might be applied to other victims. Most obviously, battered woman syndrome has in some cases been reframed as “battered person syndrome”⁷⁰and/or applied to victims of all genders. The nature and effects of family violence may perhaps more helpfully be conceptualised in terms of behaviours rather than participant characteristics.

It does not seem to us to be problematic to extend our consideration to the positon of victims or aggressors who are not or who are only minimally represented in the available data. We agree with the Victorian Law Reform Commission that the same legal issues arise for all victims of

62 At 14. The Family Violence Death Review Committee defines intimate partner violence as including acts of physical aggression, psychological abuse, forced intercourse and other forms of sexual coercion, and various controlling behaviours. See also Etienne G Krug and others (eds) World report on violence and health (World Health Organisation, 2002).

63 At 41. In a small number of cases, however, men may be victims of intimate partner violence. One of the 55 intimate partner deaths considered by the Family Violence Death Review Committee in its Fourth Report about which relationship history information was available involved a male primary victim and a female predominant aggressor. In its recent discussion document on New Zealand’s legislative response to family violence, the Ministry of Justice noted men’s experience of domestic violence is often different to that of women. Intimate partner violence perpetrated by women against men is much less severe, and men are more likely to experience other forms of family violence, like sibling violence. See Ministry of Justice Strengthening New Zealand’s legislative response to family violence: A public discussion document (Wellington, August 2015) at 14.

64 The different ways in which men and women commit homicide in the context of intimate relationships is discussed in detail in Chapter 6.

66 For a helpful and recent discussion from the Australian Institute of Criminology, see Alexandra Gannoni and Tracy Cussen “Same-sex intimate partner homicide in Australia” (2014) 469 Trends & Issues in Crime and Criminal Justice 1. See also, for example, “Breaking the taboo of domestic violence in LGBTI relationships” (30 May 2015) Stuff.co.nz <www.stuff.co.nz>; Ally Fogg “LGBT victims of domestic abuse are rarely catered for – or acknowledged” The Guardian (online ed, London, 14 March 2014); Maya Shwayder “A Same-Sex Domestic Violence Epidemic Is Silent” The Atlantic (online ed, Washington DC, 5 November 2013); and Joanna Jolly “Is violence more common in same-sex relationships?” (18 November 2014) BBC News <www.bbc.com>.

67 See for example Social Policy Evaluation and Research Unit (Superu) Reducing the impact of alcohol on family violence (April 2015) at 2. See also United States Centers for Disease Control and Prevention “CDC releases data on interpersonal and sexual violence by sexual orientation” (press release, 25 January 2013).

68 Leonard D Pertnoy “Same Violence, Same Sex, Different Standard: An Examination of Same-Sex Domestic Violence and the Use of Expert Testimony on Battered Woman’s Syndrome in Same-Sex Domestic Violence Cases” (2012) 24 St Thomas L Rev 544 at 545. See also Social Policy Evaluation and Research Unit (Superu), above n 67.

69 Gannoni and Cussen, above n 66. See also Evan Stark Coercive Control: How Men Entrap Women in Personal Life (Oxford University Press, Oxford, 2007) at 397.

family violence who kill their abusers, whatever their gender and whatever their relationship with the abuser.⁷¹

While intimate partner relationships are the most common context in which primary victims kill predominant aggressors, victims of family violence also kill abusers within other close interpersonal relationships. Of the 24 New Zealand cases we have reviewed in which victims of family violence killed abusers,⁷²two involved killings of male parents by male children.⁷³A further case that was before the High Court at the time of publication of this Report involved a female defendant, Daryl Kirk, who, at the age of 19, shot and killed her mother’s partner during a violent confrontation.⁷⁴

In R v Erstich, the defendant had been subjected by his father to abuse that the Crown accepted amounted to “not much short of a reign of terror”.⁷⁵When he was 14 years old, after a decade of being subjected to physical and psychological abuse and witnessing violence towards his mother and brothers, the defendant killed his father by shooting him at close range. The killing was premeditated. Although he was charged with murder, he was convicted of manslaughter. At trial, he claimed the killing was provoked.⁷⁶He ultimately received a suspended sentence of two years’ imprisonment.⁷⁷

In R v Raivaru, the defendant was 15 years old when he stabbed his step-father to death with a carving knife in circumstances the sentencing Judge considered amounted to “serious provocation”.⁷⁸Before the killing, the stepfather had assaulted and verbally abused the defendant and his mother, and the Judge accepted the homicide arose from the defendant’s desire to protect his mother, which “regrettably, resulted in disproportionate use of force with a weapon”.⁷⁹The defendant pleaded guilty to manslaughter and was sentenced to four years’ imprisonment.

Erstich and Raivaru are cases of homicide by children, not intimate partners, but both involved violence by the deceased against other family members, including the defendants’ mothers. The FVDRC notes that intimate partner violence and child abuse and neglect are “entangled” forms of abuse and that:⁸⁰

It is well known that exposure to [intimate partner violence] is a form of child abuse and that there is a high rate of co-occurrence between intimate partner violence and the physical abuse of children. Many children affected by family violence are living with what Edleson et al have described as the “double whammy” – the co-occurrence of being exposed to family violence in relation to other family members and being a direct victim of child maltreatment. Children are also injured in the “crossfire” of a violent assault or attack against the adult primary victim and can be used as “weapons” by abusive (ex-) partners in the context of [intimate partner violence].

71 Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 61.

73 R v Erstich [2002] NZCA 122; (2002) 19 CRNZ 419 (CA); and R v Raivaru HC Rotorua CRI-2004-077-1667, 5 August 2005. Both cases involved the partial defence of provocation, before its repeal in 2009. Provocation, and its repeal, is discussed in detail in Chapter 10 of this Report.

74 Daryl Kirk was charged with murder and claimed self-defence. Shortly before publication of this Report the jury returned a verdict of manslaughter. She has yet to be sentenced.

76 The Court of Appeal recorded at [3] the verdict “may have reflected acceptance of lack of intent to murder, but was more likely on the facts of the case to have entailed the jury’s acceptance of the partial defence of provocation”.

77 The sentence of imprisonment was imposed on appeal. The defendant was, in the first instance, sentenced to two years’ supervision.

80 Family Violence Death Review Committee, above n 61, at 76 (footnote omitted).

Understanding family violence: Reforming the criminal law relating to homicide 25

The FVDRC notes, in addition, that intimate partner violence and child abuse and neglect are “not necessarily separate co-existing forms of violence” and that their co-occurrence may “only [make] sense if you understand family violence as a pattern of coercive control and that actions directed at one individual are not necessarily designed to impact only on that individual”.⁸¹
Other forms of family violence (for example, elder abuse or violence among siblings) can be similarly “entangled”.⁸²

Visibility and understanding of family violence, most particularly intimate partner violence, is relatively recent and still evolving.⁸³In its most recent Report, the FVDRC calls for a change to how we understand intimate partner violence and victims’ responses to such violence.

Historically, intimate partner violence was thought of as “marital conflict” and less serious than stranger violence. Disputes between couples were seen as a private matter and a relationship issue for which both parties were responsible.⁸⁴The FVDRC observes that some still view family violence as “just a domestic”, which minimises the serious impact of the abuse by relegating it to “household affairs”.⁸⁵

Not only does this misunderstand family violence and its impact on victims, it also affects family and whānau members’ perceptions of the seriousness of family violence they may witness or be involved in, and the need for intervention.⁸⁶Evidence suggests that misconceptions held by family, whānau, friends and wider society about violence and victimisation make it harder for victims to seek help and leave violent relationships.⁸⁷

To date, understandings of family violence have, according to the FVDRC, tended to be violent incident focused, that is, a series of violent incidents between which it is assumed the victim is not being abused and, in the case of adult victims, there are opportunities to leave or address the violence.⁸⁸

Incident-focused conceptions may seem useful in court proceedings because “incidents can be asserted and often proven”,⁸⁹but the FVDRC argues that an emphasis on discrete events may obscure the broader dynamics of family violence.⁹⁰Intimate partner violence in particular usually involves a combination of physical, psychological, emotional, social and financial abuse,⁹¹and focusing on discrete episodes can minimise other harmful aspects of the violence,

83 As to which, see Stark, above n 69, at 142–145. See also Martha Mahoney “Legal Images of Battered Women: Redefining the Issue of Separation” (1991) 90 Mich L Rev 96 at 27.

84 Family Violence Death Review Committee, above n 60, at 34; and Stella Tarrant “Something is Pushing Them to the Side of Their Own Lives: A Feminist Critique of Law and Laws” (1990) 20 UWAL Rev 573 at 579–581.

85 Family Violence Death Review Committee, above n 61, at 77; and Family Violence Death Review Committee, above n 60, at 35.

87 Fleur McLaren Attitudes, Values and Beliefs about Violence within Families: 2008 Survey Findings (Ministry of Social Development, March 2010)

such as coercive control (discussed below).⁹²It can also mean that some practitioners⁹³and members of the public are not attuned to the danger posed by possessive and controlling partners.

The concept of “learned helplessness” was developed in an attempt to ensure victims’ experiences and responses to family violence were properly understood. Central to the concept of learned helplessness is Dr Lenore Walker’s work on battered woman syndrome, which applied the cycle of violence and learned helplessness theories to battered women.⁹⁴Unifying Dr Walker’s theory is the proposition that “women stay with abusive men because they are rendered helpless and dependent by violence”.⁹⁵Like other conceptions of battering, it is incident focused, emphasising the type and number of assaults (or other coercive acts).⁹⁶

However, the theory of battered woman syndrome is criticised for a number of reasons:
^._{It promotes a rigid,
limited view of battered women’s experiences and behaviour
that}

overemphasises their psychological reactions and defines women by reference to victimisation.⁹⁷The word “syndrome” is considered misleading as it “medicalises” a person’s response to violence and implies that battered women suffer from a condition or mental disability.⁹⁸
^._{The theory does not take into
account cultural diversity.}⁹⁹_{This is particularly
important}

given that Māori women are overrepresented as victims of family violence who kill in New

_Zealand.¹⁰⁰
^._{The theory risks
creating a stereotype of the battered woman to the detriment of victims
of}

family violence who do not fit that stereotype.¹⁰¹

The FVDRC submits that its regional death reviews demonstrate that victims of intimate partner violence are, in fact, neither passive nor helpless. To the contrary, they are proactive help seekers, and those victims experiencing the lowest levels of informal support from friends, family and whānau are more active in seeking help from agencies.¹⁰²

More recently, the FVDRC has identified that a dialogue of “empowerment” has arisen, with an aim of supporting victims in addressing the abuse they have suffered. This is the approach

93 Including all practitioners who work within agencies to provide services that are accessed by people experiencing, perpetrating and exposed to violence and abuse. This may include dedicated family violence services, as well as mainstream services (such as health, housing and income support). It includes people who operate in the criminal justice system, as well as teachers, psychologists, and those delivering parenting programmes. For further information see the Family Violence Death Review Committee’s discussion of the terms “multi-agency family violence system” and “family violence workforce” at 13–14.

The Battered Woman (Harper & Row, New York, 1979) and The Battered Woman Syndrome (Springer, New York, 1984).

97 Elizabeth Sheehy, Julie Stubbs and Julia Tolmie “Securing Fair Outcomes for Battered Women Charged with Homicide: Analysing Defence

98 Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007) at 286; Victorian Law

Reform Commission, above n 71, at 170; and Elisabeth McDonald “Battered Woman Syndrome” [1997] NZLJ 436 at 437.

101 Law Reform Commission of Western Australia, above n 98, at 286; Victorian Law Reform Commission, above n 71, at 170; Danielle Tyson and others “The Effects of the 2005 Reforms on Legal Responses to Women Who Kill Intimate Partners” in Kate Fitz-Gibbon and Arie Freiberg (eds) Homicide Law Reform in Victoria: Retrospect and Prospects (The Federation Press, Leichhardt, 2015) 79 at 83; and Heather Douglas “A consideration of the merits of specialised homicide offences and defences for battered women” (2012) 45 Australian and New Zealand Journal of Criminology 367 at 377.

Understanding family violence: Reforming the criminal law relating to homicide 27

used by many family violence services currently.¹⁰³The FVDRC criticises the empowerment approach, as it places the burden on the victim rather than the wider family violence response system.¹⁰⁴It states:¹⁰⁵

It is important to put the concept of empowerment within victims’ complex and sometimes chaotic lives, as structural inequities constrain and shape the lives of victims, albeit in different ways. The concept of “empowerment” is problematic when working with victims facing lethal violence, who also frequently face severe structural disadvantages. This is because it may appear as though an individual’s inability to keep themselves or their children safe is a result of their decisions and choices. It renders invisible the systemic barriers that impede those choices (such as lack of stable housing and access to money, poverty, racism, sexism and the legacy left behind by colonisation).

The FVDRC calls for a change to the way New Zealand understands:
^._{intimate partner violence as a pattern of harm, rather than
as a series of incidents; and}
^._{victims’
responses to that violence, which requires understanding family violence as a
form}

of entrapment rather than trying to explain victims’ responses by reference to learned helplessness and battered woman syndrome.

Central to this Report is how we understand and interpret the action of a victim of family violence in killing their abuser. In particular, as we go on to explain in Chapter 6, a victim’s use of fatal force against her abuser must be regarded as “reasonable” in the circumstances for a claim of self-defence to succeed.

Understanding victim responses to family violence can be difficult. People who have not experienced violence themselves can struggle to understand why a victim responded in the way that they did. The question commonly asked is “why didn’t she just leave?” The answer to questions of this kind can be complex and counter-intuitive. The FVDRC argues that victims’ responses can be properly understood only by viewing intimate partner violence as a form of “social entrapment”.¹⁰⁶

The FVDRC suggests that, rather than a series of discrete events, family violence is best understood as a pattern of harmful behaviour.¹⁰⁷This behaviour, it is argued, belongs to the primary aggressor (not the relationship) and is bigger than both the constituent incidents of physical violence and the current relationship.

In its submission on the Issues Paper,¹⁰⁸the FVDRC suggested that thinking of intimate partner violence as a pattern of harm enables a helpful shift in thinking. Episodes of violence can better be understood in the wider context of the primary aggressor’s behaviour (including physical violence as well as controlling and coercive behaviour) in both the current relationship and previous relationships.¹⁰⁹

105 Family Violence Death Review Committee, above n 61, at 83 (footnote omitted).

108 Law Commission Victims of Family Violence Who Commit Homicide (NZLC IP39, 2015).

109 Family Violence Death Review Committee submission at 11. See also Family Violence Death Review Committee, above n 60, at 36.

Viewing intimate partner violence as a pattern of harm also reframes single violent episodes as components of an “escalating spiral of violence” rather than one-off events.¹¹⁰This recognises the cumulative effect of such violence, which is more than the sum of individual acts of violence. The FVDRC’s regional reviews, for example, showed that:¹¹¹

The cumulative and compounding effect of the abuse also frequently resulted in a raft of secondary issues. These included physical and mental health issues, histories of self-medicating with drugs and alcohol, suicide attempts and the inability to hold down employment. [Victims of intimate partner violence] often had difficulty in parenting their children, which – in some cases – resulted in them terminating pregnancies because they could not face bringing another child into “a nightmare situation” or their children being physically removed from them because they were unable to keep them safe.

In a similar vein, American researcher Evan Stark notes it is difficult to reconcile long-term abuse and cumulative harm with the criminal law’s traditional focus on discrete incidents of violence. He says:¹¹²

Sheer repetition is not the issue. Even though pickpockets, muggers or car thieves typically commit dozens of similar offenses, because each harm is inflicted on a different person, the law is compelled to treat each act as discrete. But the single most important characteristic of woman battering is that the weight of multiple harms is borne by the same person, giving abuse a cumulative effect that is far greater than the mere sum of its parts. As British sociologist Liz Kelly has pointed out in her work on sexual predators, a victim’s level of fear derives as much from her perception of what could happen based on past experience as from the immediate threat by the perpetrator.

When family violence is understood as a pattern of harm, misconceptions such as those discussed below can be addressed.

Family violence is more than just physical assaults. It usually involves a combination of physical, psychological, emotional, social and financial abuse.¹¹³Physical violence is generally only one method of maintaining power and control in the relationship.

Failing to understand family violence as being more than physical assaults overlooks the broader dynamics often involved in family violence, including the use of coercive control.¹¹⁴Coercive control operates through the use of a range of abusive strategies tailored to the unique psychology of the target, designed to control the victim even when she is not in the presence of her abuser.¹¹⁵Threats of physical violence, for example, are often as powerful in maintaining control over a victim as actual incidents of violence, as once a perpetrator has shown they are capable of carrying out the threats made, there is no need to resort to physical assaults.¹¹⁶The often unpredictable nature of abusive outbursts leaves some women in a state of constant fear for their lives.¹¹⁷

113 Family Violence Death Review Committee, above n 61, at 71; and Victorian Law Reform Commission, above n 71, at 161.

Understanding family violence: Reforming the criminal law relating to homicide 29

International reviews have consistently found that the number one risk factor for intimate partner homicide is prior family violence.¹¹⁸This poses a considerable challenge in objectively assessing the risk of intimate partner homicide, as only a “tiny proportion” of men who have been violent eventually commit homicide, and there is clear empirical evidence to suggest that, qualitatively, men who kill their spouses do not differ greatly from those who use non-lethal violence.¹¹⁹

Instead, the research tells us that a woman’s own appreciation of risk is the most reliable predictor of her partner’s future violence towards her.¹²⁰This is because the victim will have become hypervigilant and attuned to signals of impending violence.¹²¹Primary victims who misjudge the likelihood of future violence tend to underestimate, rather than overestimate, the risk of violence.¹²²

The FVDRC considers that two assumptions underpin current understanding of victims’

responses:¹²³
^._{With help, victims can
effectively deal with family violence.}
^._{Help is
readily available to those prepared to seek it.}

However, these assumptions are challenged when intimate partner violence is understood as a form of social entrapment with three dimensions:¹²⁴
^._{Social isolation, fear and coercion that the abusive partner creates
through violence.}
^._{The institutional responses to the
victim’s suffering and their attempts to seek
help.}¹²⁵
^._{The ways in which coercive
control (and indifferent institutional responses to victims}

seeking help) can be aggravated by structural inequities of gender, class and racism. The FVDRC argues:¹²⁶

An entrapment approach requires an investigation in each case of the manner in which a particular

victim’s choices have been constrained by the violence they have experienced. This includes considering past responses to their help-seeking and the larger structural constraints of their lives, including the

118 Jennifer Martin and Rhonda Pritchard Learning from Tragedy: Homicide within Families in New Zealand 2002-2006 (Ministry of Social

119 At 39, discussing M Aldridge and K Browne “Perpetrators of spousal homicide: a review” (2003) 4 Trauma, Violence and Abuse 265. According to a Report commissioned by the United States National Institute of Justice (USNIJ) in 2005, several studies have examined “escalation” within intimate partner violence and found that, while patterns vary across different types of relationship and different types of violence, increases in the frequency and intensity of domestic violence were common and unpredictable: Alex R Piquero and others Assessing the Offending Activity of Criminal Domestic Violence Suspects: Offense Specialization, Escalation, and De-Escalation Evidence from the Spouse Assault Replication Program (United States National Institute of Justice, NSJ 212298, December 2005). The findings of another study undertaken for the USNIJ similarly “contradicted overgeneralisations about high-risk batterers” who are not “easily ‘typed’ or predicted”: D Alex Heckert and Edward W Gondolf Predicting Abuse and Reassault Among Batterer Program Participants (United States National Institute of Justice, NCJ 199730, 2004) at 50. Findings of these Reports are discussed in Kellie Toole “Self-Defence and the Reasonable Woman: Equality before the New Victorian Law” [2012] MelbULawRw 7; (2012) 36 MULR 250 at 276–277.

123 Family Violence Death Review Committee, above n 60, at 37; and Family Violence Death Review Committee submission at 11.

125 The Family Violence Death Review Committee refers to this, at 39–42, as the “indifference of powerful institutions to the victim’s suffering”.

structural constraints of their families, whanau and communities. It involves interpreting their behaviour in that context.

This does not involve assuming all victims’ experiences of or responses to abuse are the same, nor does it assume victims automatically possess autonomy and choice or are deprived of autonomy and choice as a result of being abused.¹²⁷

By understanding family violence as a form of entrapment, the misconceptions identified below and others such as “if a victim stays in a relationship or returns to a relationship, the violence can’t have been that bad or the victim must have been partly to blame”¹²⁸and “a person’s cultural background or language is no barrier to accessing help”¹²⁹can be addressed.

A common assumption is that a primary victim can avoid further violence by simply leaving the relationship.¹³⁰However, the evidence demonstrates that it is in fact very difficult for primary victims to safely leave abusive relationships and that women are most likely to be killed by an abusive partner in the context of an attempted separation.¹³¹The FVDRC states that “of the 85 [intimate partner] deaths from 2009–2014, Police records suggested that almost half took place in a separation context. Nine of these were planning separations, while 29 had already separated from the predominant aggressor.”¹³²This trend is consistent with overseas patterns.¹³³Indeed, some studies conclude that, following a break-up of a relationship involving intimate partner violence, the predominant aggressor stalks, harasses, attacks and sometimes kills their ex-partner in 75 per cent or more of all cases.¹³⁴

129 Family Violence Death Review Committee, above n 60, at 37; and Victorian Law Reform Commission, above n 71, at 165.

130 A 2008 New Zealand survey on individual attitudes, values and beliefs about family violence found that 67 per cent of respondents believed that a woman who is beaten by her partner just needs to leave the relationship to be safe. See McLaren, above n 87, at 14. Australian research in 2013 found nearly eight in 10 people found it hard to understand why a woman would stay in a violent relationship, with more than half agreeing women would leave a violent relationship “if they really wanted to”: VicHealth Australians’ attitudes to violence against women: Findings from the 2013 National Community Attitudes towards Violence Against Women Survey (NCAS) (P-MW-147, 2014) at 12. United States research found two-thirds of people believed that women could exit abusive relationships: Alissa Pollitz Worden and Bonnie E Carlson “Attitudes and Beliefs About Domestic Violence: Results of a Public Opinion Survey: II Beliefs About Causes” (2005) 20 Journal of Interpersonal Violence 1219 at

131 Walter S DeKeseredy, McKenzie Rogness and Martin D Schwartz “Separation/divorce sexual assault: The current state of social scientific knowledge” (2004) 9 Aggression and Violent Behavior 675 at 677.

133 In the United States and Canada, compared to married women, separated women are 25 times more likely to be assaulted by ex-partners and five times more likely to be murdered: Deborah K Anderson and Daniel G Saunders “Leaving an Abusive Partner: An Empirical Review of Predictors, the Process of Leaving, and Psychological Well-Being” (2003) 4 Trauma, Violence, & Abuse 163 at 179. In Australia, a quarter of intimate partner homicides (1989–2002) occurred between separated, former or divorced couples, with 84 per cent of those victims being women: Jenny Mouzos and Catherine Rushforth “Family Homicide in Australia” (2003) 255 Trends & Issues in Crime and Criminal Justice 1 at 2.

134 Elisabeth McDonald “Defending Abused Women: Beginning A Critique of New Zealand Criminal Law” (1997) 27 VUWLR 673 at 680–681.

Understanding family violence: Reforming the criminal law relating to homicide 31

Some women retaliate and resist coercive control by using violence themselves, sometimes in an attempt to try and establish a semblance of parity in the relationship, other times in violent self- defence, violent retaliation and violent resistance.¹³⁵Primary victims may also use violence when they sense another attack from the predominant aggressor is about to occur.¹³⁶

However, while a predominant aggressor may not be the first party to initiate violence on any particular occasion, he or she will use violence more – and differently – across the relationship as a whole. The primary victim/predominant aggressor dynamic is important in assessing culpability for relationship violence because it takes account of the whole of the relationship, not just discrete events.

Coercive control is considered by many to be central to contemporary understandings of family violence and particularly intimate partner violence.¹³⁷Stark explains coercive control by separating the different tactics of “coercion” and “control”.¹³⁸

Coercion tactics involve the use of violence or threats to intimidate or hurt victims and instil fear. Examples include severe beatings, sexual violence and acts like strangulation, which has been described as “the domestic violence equivalent of water boarding”,¹³⁹as well as threats, stalking, destruction of property and violence against children or pets.¹⁴⁰Coercive behaviour is apparent in the cases we have reviewed. One expert, cited in the Victorian Law Reform Commission’s Report Defences to Homicide, explained the power of this kind of abuse:¹⁴¹

A commonly reported pattern of abuse is the limited use of physical assaults, with daily threats of physical abuse and verbal abuse. The threats of physical violence are often as powerful in maintaining control over a victim as the actual incidents of violence. Once the perpetrator has shown they are capable of carrying out the threats made, there is no need to resort to physical assaults. The often unpredictable nature of abusive outbursts leaves some women in a constant state of fear for their lives.

Control tactics, on the other hand, are designed to isolate and foster dependence on the abuser and their lifestyle. They include deprivation, exploitation and micro-regulation of everyday life, for example, limiting access to money and food, or controlling how women dress.¹⁴²Together, coercive and control tactics undermine a victim’s ability for independent decision making and inhibit resistance and escape.¹⁴³

Not all family violence can be explained this way, but coercive and controlling behaviours are “prototypical” in intimate partner violence and help explain why victims stay in abusive

137 See for example the Ministry of Justice’s recent discussion document: Ministry of Justice, above n 63. See also United Kingdom Home Office

Strengthening the Law on Domestic Abuse Consultation – Summary of Responses (December 2014).

139 Susan B Sorenson, Manisha Joshi and Elizabeth Sivitz “A Systematic Review of the Epidemiology of Nonfatal Strangulation, a Human Rights and Health Concern” (2014) 104(11) Am J Public Health 54 at 57. Strangulation is the subject of another Law Commission Report: Strangulation: The Case for a New Offence (NZLC 138, 2016).

relationships.¹⁴⁴Further, while coercive control has been applied and discussed principally in connection with intimate partner violence, non-intimate family relationships may also involve behaviours of coercion and control.¹⁴⁵

Stark has said that women with whom he has worked “[insist] that ‘violence isn’t the worst part’ of the abuse they experience”.¹⁴⁶In essence, family violence is more than physical assaults and isolates and entraps victims. The FVDRC states that:¹⁴⁷

[The] nature of coercive control makes it almost impossible for many victims to remove themselves and their families from an abusive partner ... The violence is directed at isolating the victim from potential support and undermining her self-determination.

Examples of coercive controlling behaviour by the primary aggressor recorded by the FVDRC include smashing multiple phones so their partner is uncontactable and cannot contact others, keeping at least one child at home every time a partner leaves so the partner has to return, and controlling access to friends and relatives.¹⁴⁸

These behaviour types receive little attention in the cases we have reviewed. They are less obvious than physical assaults and may, of course, not be reported by primary victims. The FVDRC cautions that, within a coercive and controlling environment:¹⁴⁹

Many women are hypervigilant in order to manage their and their children’s safety. Thus, apparent rejections of help or a lack of response to service enquiries may be an attempt to maintain their personal safety and that of their children.

Just as it is important to consider the use of violence by predominant aggressors within the context of a pattern of harm, the FVDRC argues it is also important to think of violence used by primary victims as resistance to that pattern of harm rather than as an isolated incident of aggression.

Despite the dangers separation poses for primary victims and the coercive and controlling tactics that may inhibit escape, the FVDRC records that many victims go to considerable lengths to try and protect themselves and their children. They may relocate to refuges, take out protection orders, or contact Police, other agencies or family, whānau and friends for help.¹⁵⁰

The FVDRC found in its death reviews that many victims of family violence were unable to access proper support to achieve safety, despite their attempts and the well-meaning efforts of many individuals working within the family violence system. It notes that the “reality is that

144 Stark, above n 69, at 12. See also Evan Stark “Re-presenting Battered Women: Coercive Control and the Defense of Liberty” (paper presented to Violence Against Women: Complex Realities and New Issues in a Changing World, Quebec, Canada, 2012) at 7: “The primary outcome of coercive control is a condition of entrapment that can be hostage-like in the harms it inflicts on dignity, liberty, autonomy and personhood as well as to physical and psychological integrity.”

145 See for example United Kingdom Home Office, above n 137. In a consultation on the proposed enactment of a new offence of domestic abuse to criminalise patterns of coercive and controlling behaviour, the Home Office received submissions on “the importance of any new offence capturing inter-familial abuse as well as intimate partner abuse” and noted, at 9, that, while some respondents considered coercive control is limited to intimate partner relationships, others had submitted it was not and that such behaviours may affect other victims, including the elderly. As enacted, the offence of “controlling or coercive behaviour in an intimate or family relationship” applies both to people in intimate personal relationships and family members who live together: Serious Crime Act 2015 (UK), s 76.

Understanding family violence: Reforming the criminal law relating to homicide 33

real help within our current family violence system is sporadic, unpredictable and frequently not available for victims”.¹⁵¹

The FVDRC cautions that, beyond the dynamics of individual relationships, wider structures contribute to and support entrapment.¹⁵²It observed that a number of primary victims identified in the regional death reviews had unaddressed histories of childhood abuse and trauma and compounding experiences of victimisation throughout their adult life, which left them extremely vulnerable. Primary victims were often grappling with co-occurring issues such as addiction and mental health, and many were in positions of extreme economic disadvantage.¹⁵³
It states:¹⁵⁴

Gender inequity, racism, poverty, social exclusion, disability, heterosexism and the legacy of colonisation shape people’s experiences of abuse. Victims who are in the most dangerous social positions may face higher levels of violence and have less support and resources to manage. These victims may well have extended families and communities that are experiencing intergenerational trauma as the historical legacy of colonisation. They are also more likely to be confronted with discriminatory attitudes when seeking help from services charged with protecting and/or providing support to them and their children.

Māori are disproportionately represented in family violence deaths as both offenders and victims.¹⁵⁵The FVDRC identified that:¹⁵⁶

Māori women are likely to have lower levels of education, be poorer, live in areas with poor quality housing and have their children younger. Māori women are more likely to experience racist attitudes and indifference when seeking help from agencies and services. They are also almost six times more likely to be hospitalised because of assault and attempted homicide, and 1.6 times more likely to die of assault and homicide. When their children are harmed, Māori women tend to be socially demonised, evident in the media’s “mother blaming”, with little consideration of the horrific ongoing abuse and violence the women themselves live with.

The FVDRC considers this is a matter of significant concern and suggests that “patterns of normalisation of violence” revealed by the regional reviews may be “a legacy of colonisation and institutional racism”.¹⁵⁷It states:¹⁵⁸

Violence within Māori whānau (immediate and wider family) cannot be addressed without considering the impacts of colonisation on Māori whānau. The colonising agenda was assimilation of Māori and the dispossession of their land, language and cultural practices. The loss of land, along with the urbanisation of many Māori, disconnected them from their tūrangawaewae (place connected with whakapapa to stand) and their cultural connections. With this disconnection came the loss of the protective supports that are inherent in the traditional functioning of whānau, and also the important cultural beliefs that saw women and children as valued and protected members of Māori society.

151 Family Violence Death Review Committee, above n 60, at 42; and Family Violence Death Review Committee submission at 13.

153 Alcohol and drug abuse is strongly correlated with family violence in New Zealand, and research demonstrates that alcohol escalates aggressive incidents. See Social Policy Evaluation and Research Unit, above n 67, at 4. Alcohol was identified as a factor in 14 of the 24 cases we identified in which victims of family violence killed their abusers.

154 Family Violence Death Review Committee, above n 60, at 45 (footnote omitted).

156 Family Violence Death Review Committee, above n 60, at 42–43 (footnote omitted).

158 Family Violence Death Review Committee, above n 60, at 43 (footnotes omitted).

These losses, combined with the imperative that Māori conform to dominant (ie, colonial) cultural traditions, meant the collective responsibility and obligation to protect and nurture women and children within whānau and hapū disappeared. In addition to structural changes to many whānau, gender roles that were traditionally complementary and involved men having an active role in caring for tamariki were changed. Instead, whānau became the private domain of men, and male dominance became a feature in Māori society. Māori women no longer held equal positions and could not rely on the protective korowai (cloak) of the wider whānau. In today’s society, many Māori men are exposed to, and influenced by, dominant non-Māori forms of masculinity.

The Ministry of Justice, in Strengthening New Zealand’s Legislative Response to Family Violence, also suggests:¹⁵⁹

... compounded disadvantage rather than individual risk factors may underlie the risks of wāhine and tamariki Māori being victims of family violence and tāne Māori being apprehended and convicted of a family violence offence.

The Ministry of Justice also identifies Pacific people and ethnic migrant communities as experiencing higher rates of intimate partner violence than the general population.¹⁶⁰These groups can face distinct socio-economic, cultural and practical barriers that may make it more difficult to seek help. Other groups of people identified as being particularly vulnerable to family violence include older people, who may be at risk of intimate partner violence or financial abuse by other family members, and disabled people, who may rely on others for day-to-day care, increasing the risk of family violence.¹⁶¹

The FVDRC also highlights that gangs are frequently environments where the members compound and exacerbate traditional assumptions about women’s roles and justifications for violence against women.¹⁶²Women who live with gangs are at greater risk of more frequent and severe violence.¹⁶³The FVDRC posits that violence against women and children in gang cultures is often more frequent and extreme than in other contexts, and victims’ fears of retaliation if they leave abusive relationships may be greater. In its submission on the Issues Paper, the FVDRC stated that resistance to a partner’s violence is likely to be impeded and made more dangerous “due to the collective tactics of entrapment and coercion that gang membership can enable”.¹⁶⁴Women may be abused by other gang members as well as their partner, and fear of gang retaliatory violence and intimidation are “very real barriers” to seeking help or leaving a violent relationship.¹⁶⁵

Seven of the 12 cases the FVDRC identified between 2009 and 2014 in which female primary victims of intimate partner violence killed abusive male partners had a gang element,¹⁶⁶yet in our review of cases since 2001, a gang element was identified in only one sentencing decision,¹⁶⁷in which that aspect of the defendant’s background was addressed in expert evidence filed on

Understanding family violence: Reforming the criminal law relating to homicide 35

appeal. Some degree of gang association appears also to have been a feature of in three further cases¹⁶⁸although our source of information in those cases is media reports.

A proper understanding of the dynamics of family violence in the criminal justice system, based on contemporary social science, is important to ensure that victims of family violence who commit homicide are treated equitably before the law. Put simply, if the judge or the jury does not understand the social context of the homicide and the realities of the defendant’s situation, the actions of the defendant cannot be accurately assessed.

How family violence is understood will be relevant to:
^._{the
prosecutor’s decision to lay the charge;}
^._{how
the prosecutor and defence counsel approach pre-trial plea
discussions;}
^._{if the case goes to trial, how the
prosecutor and defence counsel choose to run their cases,}

_{including the evidence they seek to
introduce;}¹⁶⁹
^._{the issues and questions on
which the trial judge directs the jury;}
^._{how juries
assess the credibility of the defendant, their state of mind at the time and, in
the}

case of self-defence, the nature of the threat the defendant faced and whether the defendant’s actions were reasonable in the circumstances; and
^._{if the defendant is convicted, how the judge determines
the appropriate sentence.}

In this chapter, we have identified several ways in which traditional understandings of family violence have contributed to misconceptions. These include misconceptions that:
^._{family violence is comprised of a series of discrete incidents of
physical violence;}
^._{a victim’s fear of future
violence is irrational or unreasonable;}
^._{a victim can
avoid future violence by simply leaving the relationship; and}
^._{if a victim was violent as well, their fear was not
real.}

A proper understanding of the nature of intimate partner violence as a pattern of behaviour with a cumulative effect and as a form of social entrapment, is necessary to counter these misunderstandings. As the FVDRC states:¹⁷⁰

Attempting to reform the current system while we continue to think about family violence in exactly the same way will not produce the kinds of systemic changes we all want.

168 In relation to R v Brown see “Jail time led to p addiction, a life of crime – and a violent death” The New Zealand Herald (online ed, Auckland,

12 January 2010). In relation to R v Keefe see “Jessica Keefe not guilty of murder” (19 September 2013) Stuff.co.nz <www.stuff.co.nz>; and “Murder charge unwarranted – lawyer” (21 September 2013) Radio New Zealand < www.radionz.co.nz>. In relation to R v Tagatauli (also known as Amanda Taitapanui) see “Fatal stabbing in the leg leads to rare sentence of home detention” (21 April 2016) Stuff.co.nz

169 We note, as a particularly grave example, the Australian case of R v Kina [1993] QCA 480, discussed in Sheehy, Stubbs and Tolmie, above n 97, at 704–705 and Rebecca Bradfield “Is Near Enough Good Enough? Why Isn’t Self-defence Appropriate for the Battered Woman?” (1998) 5 PPL

71 at 72. Robyn Kina, an Aboriginal woman, stabbed her de facto partner and was convicted of murder after a trial lasting one day. On advice

from her lawyer, she did not give evidence at trial, and her defence was that she had not intended to kill the deceased. It was not until five years later, after spending years speaking to a particular social worker, that her story emerged, one of extreme and sadistic violence at the hands of the deceased. Her conviction was quashed on the grounds that there had been a miscarriage of justice based on the “exceptional difficulties of communication between her legal representatives and the appellant”, and the prosecution exercised its discretion not to proceed with a retrial.

As we observe throughout this Report, our review of the cases in which victims of family violence killed their abusers strongly suggests the need for improved understanding of family violence. Discussion of prior abuse is largely focused on previous incidents of physical violence. Abusive relationships are often described as “volatile”,¹⁷¹or defendants at times are described as just as violent. Sometimes the significance of tactics of coercion and control, beyond physical violence, appear to be missed or minimised.

The Victorian Royal Commission into Family Violence recently observed, in relation to introducing new family violence-specific offences and sentencing provisions:¹⁷²

Introducing new offences and sentencing provisions often has only a symbolic effect and does not result in changes in practice. Whatever laws we have will be only as effective as those who enforce, prosecute and apply them. Improving these practices—through education, training and embedding best practice and family violence expertise in the courts—is likely to be more effective than simply creating new offences.

To ensure the recommendations made in this Report have more than symbolic effect they must be supported by improvements in the understanding of family violence within the criminal justice system. The importance of supporting legislative reform with measures to improve understanding has been emphasised in recent years by law reform bodies in other jurisdictions, particularly Australia. In their joint Report, the Australian and New South Wales Law Reform Commissions recommended legislative guidance and judicial and legal professional education and training focused on improving the application and effectiveness of existing legal defences in the family violence context.¹⁷³The Victorian Law Reform Commission made similar recommendations to promote a better understanding by judges, jurors and legal representatives of the circumstances of family violence and the range of ways people might react to it.¹⁷⁴The Law Reform Commission of Western Australia recommended including a section on the nature and dynamics of family violence in the judiciary’s “bench book” to assist judges confronted with social and cultural issues around family violence.¹⁷⁵

Reviews of reforms targeting the legal response to victims of family violence who kill abusers also emphasise the need for comprehensive, consistent and ongoing training within the criminal justice system.¹⁷⁶Without that training, the potential of statutory law reform may not be realised.¹⁷⁷

In New Zealand, the Institute of Judicial Studies (IJS) has been active in educating the judiciary on both the family violence context and how to spot tell-tale “red flags” of abuse.¹⁷⁸However, in specific cases, judges can be limited by the evidence put before them. It is therefore important

171 For example R v Stone HC Wellington CRI-2005-078-1802, 9 December 2005; R v Tamati HC Tauranga CRI-2009-087-0868, 27 October 2009;

R v Wihongi, above n 167; and R v Woods HC Gisborne CRI-2011-016-000048, 10 June 2011.

172 State of Victoria Royal Commission into Family Violence: Summary and recommendations (Parl Paper No 132, March 2016) at 27.

173 Australian Law Reform Commission and New South Wales Law Reform Commission Family Violence — A National Legal Response (ALRC R114 and NSWLRC R128, October 2010) at 649–650.

176 Tyson and others, above n 101, at 92; Mandy McKenzie, Debbie Kirkwood and Danielle Tyson “‘Unreasonable’ self-defence?” (2013) 2 DVRCV Advocate 12 at 16; and State of Victoria, above n 172, at 57.

178 As part of the Higher Courts Update Programme a full-day seminar was held on these two specific issues in 2015. Also in 2015, the three-day District Court Judges’ Triennial Conference was devoted to family and sexual violence, the first time in the history of the District Court Bench that the entire conference was devoted to a single subject. Judges received education from psychologists, community workers, academics and Police.

Understanding family violence: Reforming the criminal law relating to homicide 37

that prosecutors and defence counsel are also alert to the circumstances in which family violence will be relevant to criminal offending.

As we discuss in Chapter 7, expert evidence can educate jurors on the dynamics of family violence and dispel misconceptions, but this depends on defence counsel identifying the need for and significance of such evidence. It also depends on the availability of family violence experts and of funding for such expert evidence. Education of prosecutors is equally important to ensure consistency in the exercise of prosecutorial discretion.

The role of Police in homicide trials is primarily investigatory, and the extent to which they identify relevant family violence evidence that may be material to the offending is crucial.

In our view, the FVDRC’s Fifth Report makes a compelling case for changing how New Zealand understands family violence, and we recommend that education be available to all key participants in the criminal justice system. Such education should focus on the contemporary social science of family violence and the importance of identifying and dispelling misconceptions that have arisen previously and remain evident today.

These recommendations are intended to support and promote a change in how we think about family violence. We hope this will remove barriers that result from persisting misconceptions and contribute to the achievement of substantive equality before the law for victims of family violence who commit homicide.

^R1Judges should continue to receive education, including through the Institute of Judicial Legal

^R2Regular and ongoing education courses on the dynamics of family violence should be made available to all criminal lawyers, including Crown prosecutors and defence counsel.

_·^{reflect contemporary social science understanding of
family violence and victims’}

_·^{explain the primary victim/predominant aggressor
analysis in intimate partner violence;}

_·identify common misconceptions of family violence that persist today, and their implications in the criminal justice system.

Homicide is the killing of a human being by another. Culpable homicide, which is the unlawful killing of a human being by another, is regarded as the most serious kind of criminal conduct. Andrew Ashworth writes that this is because:¹⁷⁹

The harm caused by homicide is absolutely irremediable, whereas the harm caused by many other crimes is remediable to a degree. Even in crimes of violence which leave some permanent physical disfigurement or psychological effects, the victim retains his or her life and, therefore, the possibility of further pleasures and achievements, whereas death is final. This finality makes it proper to regard death as the most serious harm that may be inflicted on another person, and to regard the culpable causing of death without justification or excuse as the highest wrong.

There are different forms of culpable homicide, and the law of homicide is complex. This is in part a reflection of the seriousness of the crimes and in part a legacy of the sentencing rules for murder. We discuss these features of homicide throughout this Report and particularly in Part 3.

In this chapter, we introduce the law of homicide and the justice process to which people accused of homicide are subject. We also introduce the homicide cases we have reviewed where victims of family violence have killed abusers. We have identified 24 New Zealand cases in this category over a 15-year period. This is a small sample but some trends are identifiable. We note those trends in the second part of this chapter and consider them in greater depth in Chapter 9.

To be an offence, homicide must be “culpable”, that is, blameworthy according to law. Different legal systems criminalise homicide in different ways. Most jurisdictions comparable to New Zealand have a two-tier division of murder and manslaughter,¹⁸⁰and beyond this basic division, there are some lesser related offences that we discuss below.

179 Andrew Ashworth and Jeremy Horder Principles of Criminal Law (7th ed, Oxford University Press, Oxford, 2013) at 237.

180 Or an offence like culpable homicide, which is the equivalent of manslaughter in some common law jurisdictions, including Scotland.

Understanding family violence: Reforming the criminal law relating to homicide 39

In New Zealand, culpable homicide is defined by section 160 of the Crimes Act 1961:

Homicide may be either culpable or not culpable.

_{Homicide is
culpable when it consists in the killing of any person—}

by an unlawful act; or

by an omission without lawful excuse to perform or observe any legal duty; or by both combined; or
by causing that person by threats or fear of violence, or by deception, to do an act which causes his or her death; or

_{by wilfully frightening a child under the age of 16
years or a sick person.}

Except as provided in section 178, culpable homicide is either murder or manslaughter.

_{Homicide that is not culpable is not an
offence.}

Under the ancient common law of England, only homicides caused in the enforcement of justice (such as execution of a death sentence) were lawful. All others were punishable by death unless the offender received a royal pardon because the homicide was “without felony” and, therefore, excusable. While “murder” came to describe more serious homicides, that had little practical importance because every felonious killing attracted the same penalty.¹⁸¹A sentence of death could, however, be avoided if the offender received the “benefit of clergy”, which transferred the case to the ecclesiastical court, which never imposed the death penalty. This was a device to alleviate the harshness of the uniform penalty for homicide.¹⁸²

Over time, the benefit of clergy was expanded, and in the 15th and early 16th centuries, a series of statutes excluded certain particularly serious killings (referred to as murder with “malice aforethought”) from its scope. The result was a division of culpable homicide into two categories. Killings with malice aforethought were called “murder” and punished by death, while killings without malice aforethought were punished by a brand and imprisonment for not more than a year. “Manslaughter” was the word adopted to describe this lesser form of culpable homicide.¹⁸³

The distinction between murder and manslaughter endures. Murder remains the most serious form of culpable homicide, although in few jurisdictions is it still punishable by death. In New Zealand, murder is defined by sections 167 and 168 of the Crimes Act 1961. Section 167 provides as follows:¹⁸⁴

181 Rollin M Perkins “A Re-Examination of Malice Aforethought” (1934) 43 Yale LJ 537 at 539–541. See also New South Wales Law Reform

Commission Partial Defences to Murder: Diminished Responsibility (Report 82, 1997) at n 1.

184 Section 168 of the Crimes Act 1961, which provides a further definition of murder, derives from the common law “felony murder” rule. It provides that, in certain situations where an offender does not mean to cause death or know that death is likely to follow but does one of a number of specified acts for the purposes of facilitating one of a number of specified criminal offences and death results, it will also amount to murder. This provision is not relevant for the purposes of our case review.

if the offender means to cause the death of the person killed:

if the offender means to cause to the person killed any bodily injury that is known to the offender to be likely to cause death, and is reckless whether death ensues or not:

if the offender means to cause death, or, being so reckless as aforesaid, means to cause such bodily injury as aforesaid to one person, and by accident or mistake kills another person, though he or she does not mean to hurt the person killed:

if the offender for any unlawful object does an act that he or she knows to be likely to cause death, and thereby kills any person, though he or she may have desired that his or her object should be effected without hurting any one.

Manslaughter is culpable homicide that is not murder or some other specific form of homicide. In New Zealand, manslaughter is culpable homicide that does not amount to murder or infanticide.¹⁸⁵This includes all instances when death is caused by an unlawful act but there is no murderous intent.

In jurisdictions with “partial defences”, manslaughter also includes homicides that meet the requirements of the defence. Partial defences, which we discuss in detail in Part 3, avert murder convictions by partially excusing, and categorising as manslaughter, intentional killings committed in certain mitigating circumstances. This form of manslaughter is often referred to as “voluntary manslaughter” because it involves killing with the intent for murder and is categorised as a different form of homicide only because a partial defence applies.¹⁸⁶

As well as murder and manslaughter, the law includes some lesser and overlapping specific homicide offences. In New Zealand, the third form of culpable homicide is infanticide.¹⁸⁷Other specific offences (which fall outside the definition of “culpable homicide” but overlap with the conduct captured by manslaughter) include driving offences causing death.¹⁸⁸

Among the issues we consider in this Report is how prosecutors decide what charges to lay when victims of family violence kill their abusers and, where charges are laid, how these defendants are convicted. We consider these issues in Part 3, but note some key points here.

Charging is governed by the Solicitor-General’s Prosecution Guidelines and a person may be charged if the “evidential” and “public interest” tests for prosecution are met. The evidential test requires credible evidence upon which a jury could “reasonably be expected to be satisfied

186 Law Commission of England and Wales Murder, Manslaughter and Infanticide (Law Com No 304, 2006) at 4.

187 Infanticide is provided for by s 178 of the Crimes Act 1961, subs (1) of which states: “Where a woman causes the death of any child of hers under the age of 10 years in a manner that amounts to culpable homicide, and where at the time of the offence the balance of her mind was disturbed, by reason of her not having fully recovered from the effect of giving birth to that or any other child, or by reason of the effect of lactation, or by reason of any disorder consequent upon childbirth or lactation, to such an extent that she should not be held fully responsible, she is guilty of infanticide, and not of murder or manslaughter, and is liable to imprisonment for a term not exceeding 3 years.” Infanticide is a form of diminished responsibility and has a dual role as an offence and a defence. The role performed in any given case affects the burden of proof. If a person is charged with infanticide, the prosecution must prove all elements beyond reasonable doubt, but if a person is charged with murder or manslaughter, she will be entitled to a finding of infanticide if there is a sufficient evidential foundation for it that leaves the jury with a reasonable doubt. See AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington, 2012) at

Understanding family violence: Reforming the criminal law relating to homicide 41

beyond reasonable doubt” that the defendant committed the offence.¹⁸⁹The public interest test invites consideration of whether prosecution is required in the public interest. This test is likely to be met in most homicide cases, given the value placed on human life. The Solicitor-General’s Prosecution Guidelines include the overarching principle that the nature and number of charges should adequately reflect the criminality of the defendant’s conduct.¹⁹⁰

Murder and manslaughter may be charged in the alternative, and manslaughter is usually included as a lesser offence when a person is charged with murder. This means that, if there is an evidential basis for manslaughter, a jury may convict of that offence, even if the defendant is charged only with murder.¹⁹¹This is a feature of homicide law to which we return in Part 3.

In any criminal case, a defendant may discuss with the prosecution the possibility of the charge being amended. The Solicitor-General’s Prosecution Guidelines permit principled plea discussions and arrangements in which a defendant may agree to plead guilty to a lesser charge instead of defending the more serious charge. Such arrangements can reduce the burden on victims, save time and resources and provide a forum for a defendant to accept responsibility for criminal conduct.¹⁹²The Solicitor-General must approve all plea arrangements in relation to murder charges.¹⁹³

If a defendant is charged with a homicide offence and goes to trial, the case will almost always be heard by a jury in the High Court.¹⁹⁴The jury is the fact-finder, but a judge will preside and sit through the entire trial, deal with questions of law and hear the same evidence as the jury. The judge will summarise the case at the end of the trial and direct the jury on the law. The prosecution must prove each element of the offence and the absence of any relevant defence beyond reasonable doubt.¹⁹⁵If the elements of an offence are not proved or any defence not excluded to this standard, the defendant must be acquitted of the offence charged, although in homicide cases where manslaughter is an alternative or included offence, a jury may acquit of murder but convict of manslaughter.

In addition to the evidence of offending, a prosecutor considering whether to pursue a criminal charge must consider the prospective defendant’s capacity for criminal conduct. The most relevant example of capacity is “infancy”, which means a person is under the age of criminal responsibility.¹⁹⁶In New Zealand, no person under the age of 10 years may be convicted of an offence.¹⁹⁷The law also restricts the situations in which children under the age of 14 may be charged with criminal offending. One situation where a charge may be laid is “where the child is of or over the age of 10 years, and the offence is murder or manslaughter”.¹⁹⁸We have

194 Criminal Procedure Act 2011, ss 74 and 103. Some criminal charges may be tried before a judge without a jury, but in New Zealand murder and manslaughter must be tried before a jury other than in cases of juror intimidation.

195 Some crimes, however, have inbuilt defences that put an onus of proof on the defendant. See for example Crimes Act 1961, s 134A (defence to charge of sexual conduct with young person under 16); s 124(2) (defence to charge of distribution or exhibition of indecent matter); s 124A(3) (defence to indecent communication with a young person under 16); and s 125(2) (defence to indecent act in a public place). See also Simester and Brookbanks, above n 187, at 39.

196 Another potentially relevant instance of capacity is insanity. Insanity is commonly characterised as a defence but may be better conceptualised as an instance of legal incapacity. See Simester and Brookbanks, above n 187, at 316. If a judge or jury finds a person accused of a criminal offence insane, the defendant will be acquitted on that basis and then subject to the range of detention orders provided for by the relevant legislation, the Criminal Procedure (Mentally Impaired Persons) Act 2003.

reviewed no homicides by victims of family violence in this age range or below the age of capacity, although we have identified two cases of homicide by young people.¹⁹⁹

Assuming a charge is not precluded on grounds of capacity, a prosecutor considering whether to pursue a charge must also anticipate and evaluate likely defences.²⁰⁰Defences are provided for in statutes like the Crimes Act and in “common law” principles that have developed through the courts.²⁰¹A defendant who contests a charge at trial²⁰²may also rely on a relevant defence, although in some cases, a judge may withhold a defence from the jury. We discuss this issue in connection with self-defence in Part 2.

Some defences, including self-defence, may be described as “justifications”, while others, like compulsion, are described as “excuses”.²⁰³Justifications effectively claim a defendant acted acceptably and so did not commit a crime. For this reason, they are sometimes also called “permissions”.²⁰⁴Because a person who can rely on a justification is not guilty of a crime, these defences “involve judgments about the situation which apply to all participants”.²⁰⁵To say a person was justified in acting in self-defence means they were right to use the force they did and others would have been wrong to intervene. Excusatory defences are more restrictive. They acknowledge that, in certain circumstances, a person should not be subject to the full extent of the penalties of the criminal law because their actions were understandable. Partial defences to murder are generally considered to be partial excuses, although they may have justificatory elements.²⁰⁶

HLA Hart explained the difference between justificatory and excusatory defences in this way:²⁰⁷

In the case of “justification” what is done is considered as something that the law does not condemn or even welcomes. But when the killing ... is excused, the criminal responsibility is excluded on a different footing. What has been done is something, which is deplored, but the psychological state of the agent when he did it exemplifies one or more of a variety of conditions, which are held to rule out the public condemnation and punishment of individuals. This is a requirement of fairness or of justice to individuals.

For most crime, sentencing is an opportunity for the court to consider a person’s particular culpability. Matters judges must generally consider include “the gravity of the offending in the particular case, including the degree of culpability of the offender” and “the seriousness of the type of offence in comparison with other types of offences”.²⁰⁸The principles of sentencing provide that the maximum penalty for an offence is for the most serious cases,²⁰⁹and the court

201 Section 20 of the Crimes Act 1961 expressly preserves as part of New Zealand law all “rules and principles of the common law which render any circumstances a justification or excuse for any act or omission, or a defence to any charge”.

202 Not all charges are contested. A person charged with an offence may enter a plea – of guilty or not guilty – before trial: Criminal Procedure Act

204 See for example Jeremy Horder Homicide and the Politics of Law Reform (Oxford University Press, Oxford, 2012) at 200.

206 On the justificatory elements of the partial defence of provocation, see Law Commission The Partial Defence of Provocation (NZLC R98, 2007).

207 HLA Hart Punishment and Responsibility (Clarendon Press, Oxford, 1968) at 13–14.

Understanding family violence: Reforming the criminal law relating to homicide 43

should impose the least restrictive outcome that is appropriate in the circumstances.²¹⁰Thus, sentencing permits fine-tuned offender-specific assessments of culpability.²¹¹

Historically, homicide has been exceptional among criminal offences because murder has attracted a mandatory sentence. While the maximum penalty for murder is now generally life imprisonment rather than death, in many countries that sentence is compulsory and judges have no discretion to depart from it, even where there are extenuating circumstances. This has led to the development of unique mechanisms for recognising reduced culpability for homicide, including partial defences. As we discuss in Chapter 10, however, in jurisdictions where judges have discretion in murder sentencing, there is a less compelling case for such mechanisms.²¹²

Mandatory murder sentencing was abolished in New Zealand in 2002. Other than in cases captured by the “three strikes” regime,²¹³life imprisonment is now the maximum, but not a compulsory, penalty. There is, however, a strong presumption in favour of life imprisonment, rebuttable only if the court considers that sentence would be “manifestly unjust”.²¹⁴For manslaughter, there is no presumptive sentence. As the least blameworthy form of culpable homicide, infanticide has a maximum penalty of three years’ imprisonment. These differences in penalty signal that some cases are more serious than others, and particular categories of unlawful killing warrant significant leniency.²¹⁵

Sentencing for homicide, as for other offences tried before a jury, is usually conducted by the trial judge who heard the evidence and legal arguments – unless the defendant pleaded guilty before trial.²¹⁶Whether sentencing follows a conviction after trial or a guilty plea, the prosecution and the defendant may make submissions and can seek to present further evidence. The judge will take into account an offender’s prospects for rehabilitation and risk of reoffending, usually on the basis of a pre-sentence report and any reports by professionals such as psychologists or psychiatrists.²¹⁷The judge will also consider victim impact statements, which, in homicide cases, will be from immediate family members of the victim.²¹⁸We discuss some sentencing issues in Chapter 11.

211 Nicola Lacey “Partial Defences to Homicide: Questions of Power and Principle in Imperfect and Less Imperfect Worlds ...” in Andrew Ashworth and Barry Mitchell (eds) Rethinking English Homicide Law (Oxford University Press, Oxford, 2000) 107 at 108. See also Paul Robinson Structure and Function in Criminal Law (Clarendon Press, Oxford, 1997) at ch 6.

212 Indeed, it is notable that, among comparable jurisdictions, only three that have discretionary sentencing for murder (New South Wales, Western Australia and the Australian Capital Territory) also have partial defences. Other jurisdictions that retain partial defences (England and Wales, Canada and Ireland, as well as Queensland, Northern Territory and South Australia) also have mandatory murder sentencing.

213 The “three strikes” regime, which is provided for by ss 86A–86I of the Sentencing Act 2002, significantly curbs sentencing discretion for violent offences, including murder and manslaughter. We discuss this regime, and its implications for victims of family violence who commit homicide, in Chapter 11.

216 A defendant who is considering whether to plead guilty may seek a sentence indication before deciding whether to do so. A sentence indication is a statement by the court that, if the defendant pleads guilty, the court would or would not impose a sentence of a particular type and/or length: Criminal Procedure Act 2011, s 61. Sentence indications give defendants clarity and certainty about the jeopardy they face if they plead guilty: Taylor v R [2013] NZCA 55 at [17].

218 Victims’ Rights Act 2002, s 17AA. Section 4 of that Act defines “victim” to include “a member of the immediate family of a person who, as a result of an offence committed by another person, dies or is incapable, unless that member is charged with the commission of, or convicted or found guilty of, or pleads guilty to, the offence concerned”.

In Chapter 9, we review how the criminal justice system has responded, on a case-by-case basis, to victims of family violence who have killed their abusers. We focus on cases in which the homicide was not accepted by the court or the jury to have been committed in self-defence.

In the remaining part of this chapter, we highlight some aspects of this case review, which includes acquittals, convictions for manslaughter and convictions for murder. Although we consider the cases in detail later, we summarise key features here because they provide important background to why our package of recommended reforms is focused on education about family violence, self-defence and sentencing.

Although our case review captured a 15-year period, we have identified only 24 cases. This makes it difficult to draw definitive conclusions about the existence or extent of issues or problems in practice.²¹⁹Individual cases can also have a disproportionate effect on the impression of overall trends.²²⁰

Our review has, however, not been an attempt to scientifically discern trends. We have largely relied on reported court decisions and, in some cases, media reports. We have identified cases through news databases for reports on trial outcomes and legal databases for sentencing notes, trial rulings and appeal decisions. Given this project’s time constraints, we have not done a full audit of all relevant homicides, nor have we reviewed Crown prosecutor or defence files, or spoken with the defendants, counsel or judges involved in all cases. We therefore cannot be certain we have identified all relevant cases. We may have missed unreported decisions and homicides for which no charges were laid.

Since 2001,²²¹at least 24 victims of family violence have been prosecuted for killing an abuser or suspected abuser.²²²With one exception,²²³the cases were summarised in the Issues Paper.²²⁴

As set out in the table below, across the cases we have reviewed, four defendants were acquitted and 20 were convicted of murder or manslaughter. Three of the four defendants who were acquitted relied on self-defence.

219 The limitations involved in analysing a case sample on the basis of public records, and where case outcomes are multi-factorial, was also noted by a leading New Zealand commentator who recently considered the effects of the abolition of partial defence of provocation on battered defendants: Julia Tolmie “Defending Battered Defendants on Homicide Charges in New Zealand: The Impact of Abolishing the Partial Defences to Murder” [2015] NZ L Rev 649 at n 4.

220 A number of variables are also unknown. These include the grounds on which the Crown decided in any given case to lay or amend particular charges, the content of plea discussions, the reasons for guilty pleas, the reasons for verdicts and the primary evidence on which juries and judges reached verdicts and sentencing decisions.

221 The year the Law Commission published its Report Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001).

222 We are aware of two further relevant cases that we have not addressed in this Report. In one, the defendant was found unfit to stand trial and so the prosecution was not seen through to completion (whether by a guilty plea or finding of guilt after trial). We understand the details of that case may also be subject to ongoing suppression. In the other, the defendant, Daryl Kirk, was charged with the murder of her mother’s partner. She claimed self-defence, but was found guilty of manslaughter shortly before the publication of this Report. She has yet to be sentenced. In these circumstances, we have excluded these two cases from our review.

223 The 24th case, of which we became aware after the Issues Paper was published, is R v Tagatauli. In 2015, Ms Tagatauli (also known as Amanda Taitapanui) was charged with the murder of her partner, Mura Tagatauli, after she stabbed him in the leg, severing his femoral artery. The charge was, however, subsequently amended to manslaughter, to which she pleaded guilty. She was sentenced on 21 April 2016 to 12 months’ home detention. See “Fatal stabbing in the leg leads to rare sentence of home detention” (21 April 2016) Stuff.co.nz <www.stuff.co.nz>.

224 Law Commission Victims of Family Violence Who Commit Homicide (NZLC IP39 2015) at 124–131.

Understanding family violence: Reforming the criminal law relating to homicide 45

Original charge Guilty plea to murder Guilty plea to manslaughter Charges defended at trial

Among those convicted, sentences ranged from life imprisonment with 10-year minimum periods of imprisonment for murder²²⁶to a suspended sentence of imprisonment in one of the manslaughter cases.²²⁷Among the four murder convictions, in only the most recent two – R v Wihongi²²⁸and R v Rihia²²⁹– was the presumption of life imprisonment displaced. For this reason, those cases are particularly significant, and while they are discussed throughout this Report, it is convenient to note their facts and disposition here:
^._{Wihongi: after a trial, Ms Wihongi was convicted of the murder
of a man with whom she}

had been in a relationship. For years before the homicide, she had suffered physical and sexual violence from the deceased and others.²³⁰She was also cognitively impaired after an overdose at age 13, and her ability to make judgements and reason and plan was in the low to borderline range. She displayed “residual, complex features of post-traumatic stress disorder, and anxiety and depression dating from the rapes and home invasion”.²³¹Initially, she was sentenced to a finite term of eight years’ imprisonment, the trial Judge having found it would be manifestly unjust to impose life imprisonment.²³²On appeal by the Solicitor-General and in receipt of fresh evidence about Ms Wihongi’s future risk of violent reoffending, the Court of Appeal concluded a longer finite term of 12 years’ imprisonment was required.²³³Having considered the legislative history of section 102 of the Sentencing Act and Ms Wihongi’s status as a “battered defendant”, however, the Court of Appeal did not depart from the High

225 In R v Fairburn, the defendant was initially convicted of murder, but on appeal, the conviction was quashed and a new trial ordered: R v Fairburn HC New Plymouth CRI-2008-043-931, 3 March 2009 [Fairburn (HC)]; Fairburn v R [2010] NZCA 44 [Fairburn (CA)]; and Fairburn v R [2010] NZSC 159, [2011] 2 NZLR 63. After the second trial, the defendant was convicted of manslaughter: R v Fairburn [2012] NZHC 28. For the purposes of this analysis, we treat this case as a manslaughter conviction.

226 R v Reti HC Whangarei CRI-2007-027-2103, 9 December 2008 [Reti (HC)] and R v Reti [2009] NZCA 271; and R v Neale HC Auckland

228 R v Wihongi HC Napier CRI-2009-041-2096, 30 August 2010 [Wihongi (HC)]; R v Wihongi [2011] NZCA 592, [2012] 1 NZLR 775 [Wihongi

230 At the age of 14 Ms Wihongi was sexually abused by a drug and alcohol counsellor, and around the same time, she had been prostituted for drugs and money by the older brother of the deceased. The violence she suffered over the years included gang rape and a home invasion that saw her scarred from an assault with a bottle: Wihongi (HC), above n 228, at [19]–[20].

233 This was in part for reasons of community protection and Ms Wihongi’s future risk (about which the Court of Appeal had received fresh evidence) and in part to meet the sentencing purposes of denunciation and deterrence: Wihongi (CA), above n 228, at [98].

Court Judge’s assessment that life imprisonment would be manifestly unjust.²³⁴Ms Wihongi was declined leave to appeal to the Supreme Court.²³⁵

^._{Rihia: some 11 months after the Court of
Appeal’s decision in Wihongi, Ms Rihia was}

sentenced for the murder of her estranged husband, having pleaded guilty to that charge.

²³⁶As in Wihongi, and with reference to “the similarities” of that case, the High Court Judge found it would be manifestly unjust to impose a life sentence and that the appropriate response was a finite term of 10 years’ imprisonment. ²³⁷Among other matters, the Judge was satisfied that Ms Rihia “would not have stabbed Mr Rihia to death had it not been for the significant mental impairment [she] suffer[ed] through years of alcohol abuse and physical abuse most recently, until [she] parted with him, at the hands of Mr Rihia himself”. ²³⁸There was no appeal against sentence.

The outcomes of the cases we have reviewed can be further broken down into offending committed before and after repeal of the partial defence of provocation in December 2009. In
14 of the 24 cases, the offending took place pre-repeal and so provocation was available. In the remaining 10 cases, the offending post-dated repeal. Provocation may, of course, have been unavailable on the facts of any given case.

Among the 14 cases in the pre-repeal group, 10 defendants were tried for murder, and provocation was relied on by five. In three of the five cases, the defendants were convicted of manslaughter, and in the other two, the verdicts were murder. Defendants in the pre-repeal group pleaded guilty to manslaughter in four cases, but the original charge was murder in only one of those cases. In none of the pre-repeal cases was a defendant tried for manslaughter.

Among the 10 post-repeal cases, four defendants were tried for murder, but none was convicted of murder.²³⁹The only murder conviction, R v Rihia, resulted from a guilty plea. Two defendants were charged with and tried for manslaughter, and another three pleaded guilty to manslaughter. In one of the cases that went to trial for manslaughter, the defendant was ultimately acquitted.²⁴⁰In Chapters 9 and 11, we consider whether repeal of provocation has made a difference to case outcomes for victims of family violence who commit homicide but, as we note at the end of this chapter, there is no clear empirical basis to conclude it has.²⁴¹

In R v Woods,²⁴²one of the cases we reviewed that involved an initial charge of murder, that was resolved by a guilty plea to manslaughter, the sentencing Judge, Potter J, remarked upon

237 At [25]–[33]. The 10-year term was calculated from a 12-year starting point (by reference to Wihongi), with a two-year discount to recognise

239 Of the four cases that went to trial for murder, two were convicted of manslaughter and the other two were acquitted.

240 R v Gerbes. See “Hung jury in manslaughter trial” (13 February 2014) Stuff.co.nz <www.stuff.co.nz>; and Tracey Chatterton “Accused ‘just needed to call out’” (12 February 2014) Stuff.co.nz <www.stuff.co.nz>.

241 At least in terms of conviction outcomes, this is consistent with Julia Tolmie’s conclusion that the repeal of provocation “has not shifted the fact that the majority of battered defendants are still convicted of manslaughter”: Tolmie, above n 219, at 663.

Understanding family violence: Reforming the criminal law relating to homicide 47

the “striking similarity” of a number of previous manslaughter cases that counsel had cited to the Court for Ms Woods’ sentencing.²⁴³In each case:

... the offender is female, had been in a volatile relationship with the deceased, had been involved in a domestic dispute immediately preceding the stabbing, was under the influence of alcohol, had been the subject of abuse from the deceased immediately prior to the stabbing (except perhaps in the case of Brown), and had used a kitchen knife to stab the deceased with death being an unintended result. In all these cases the women involved had difficult childhoods, had children, and in most cases had only minor previous convictions.

There are, as Potter J observed, common features among the cases we have reviewed. Most defendants were female, and had killed their abusive male partners. Most of the fatal weapons were kitchen knives or other readily available instruments. Most of the homicides occurred in the context of a confrontation, and most of the deceased abusers had recently been violent towards the defendant or intimated they would be.

A number of defendants endured years of violence and abuse before the homicides, often from other people as well as the deceased. In R v Reti, for example, the Court acknowledged the defendant’s “difficult and miserable life” and “appalling history of childhood physical and sexual abuse”.²⁴⁴Terrible personal histories beyond the abusive relationship were also recounted in R v Fairburn,²⁴⁵R v Wihongi,²⁴⁶R v Hu,²⁴⁷R v Stone,²⁴⁸R v Brown,²⁴⁹R v Woods²⁵⁰and R v Rihia.²⁵¹As well as complex abuse histories, many defendants had mental health and alcohol problems, ranging from a lack of “normal coping mechanisms”,²⁵²to depression,²⁵³post-traumatic stress disorder,²⁵⁴self-harming and borderline personality traits,²⁵⁵genetic pre- disposition to alcoholism,²⁵⁶and cognitive impairment.²⁵⁷These cases bear out observations by the Family Violence Death Review Committee (FVDRC) on intergenerational family violence and “compounding trauma”.²⁵⁸

243 Each of the cases to which Potter J was referred is included in our review: R v Brown HC Napier CRI-2008-020-3130, 24 November 2009; R v Mahari HC Rotorua CRI-2006-070-8179, 14 November 2007; R v Stone HC Wellington CRI-2005-078-1802, 9 December 2005; and R v Tamati HC Tauranga CRI-2009-087-0868, 27 October 2009.

245 Fairburn (HC), above n 225, at [10]. The Judge recorded that Ms Fairburn had “a long history as a victim of both sexual and physical abuse dating back to [her] childhood” and that the deceased was abusive towards her.

246 Wihongi (HC), above n 228, at [19]–[26]. As noted above at [3.31(a)], Ms Wihongi was sexually abused at age 14 and prostituted for drugs and money by the deceased’s older brother from the age of 14 or 15. She had been gang raped and subjected to home invasion in the course of which she was injured and scarred. She had cognitive deficits from a drug overdose at age 13 and “residual, complex features of post-traumatic stress disorder, and anxiety and depression dating from the rapes and home invasion”: Wihongi (CA), above n 228, at [18]–[22].

247 R v Hu [2012] NZHC 54 at [7]. Ms Hu’s pre-sentence and psychiatrist’s reports painted a “bleak picture of [her] life so far”, Ms Hu having

“described sexual and physical abuse as a child and young woman” in China, drug use and chronic depressive disorder.

248 R v Stone, above n 243, at [5]. Ms Stone described her upbringing as “been shocking and [she] recollect[ed] violence from [her] mother and apparently an alcoholic step-father. There were constant changes of school. On the one estimate 11 primary schools, in part suggested so that the abuse of [her] could not be detected.”

249 R v Brown, above 243, at [12]. Ms Brown, the Court noted, had a “disturbed and disrupted upbringing”.

250 R v Woods, above n 242, at [11]. Ms Woods grew up with “frequent domestic violence by her father against her mother”.

251 R v Rihia, above n 229, at [18]–[28]. The sentencing Judge recorded Ms Rihia had a “longstanding history of violence”, and “[t]here was much violence and alcohol abuse in [her] past before [she] met Mr Rihia. [Her] parents were both alcoholics and it was [her] father’s violence towards [her] which required [her] removal as a child to the care of [her] grandparents.” Ms Rihia had “abused” alcohol since secondary school and her first marriage (not to the deceased) had also been violent. She had “significant mental impairment ... through years of alcohol abuse and physical abuse most recently, until [she] parted with him, at the hands of Mr Rihia himself”.

254 Reti (HC), above n 226; Wihongi (HC), above n 228; and R v Rakete [2013] NZHC 1230.

258 Family Violence Death Review Committee Fifth Report: January 2014 to December 2015 (Health Quality & Safety Commission, February 2016) at 51 and 55. The Committee records, at 55, findings from the Adverse Childhood Experiences study that “exposure to family violence during childhood heightens the risk of intergenerational violence, with girls more likely to become victims and boys more likely to perpetrate [intimate partner violence] as adults” (footnotes omitted).

Not all cases, however, mention the presence of significant trauma outside the abusive relationship that resulted in the homicide. The defendant in R v Wickham, for example, suffered from multiple sclerosis and had a violent and controlling husband, but the Court did not record other sources of abuse, and Ms Wickham was acknowledged to have a “widespread network of people who care[d] greatly for [her]”.²⁵⁹

The emotions or psychological states of the defendants at the time of the homicides seem also to vary (as far as we can tell from the record). In some cases, the evidence suggests the defendant was acting primarily defensively,²⁶⁰while in others, the courts record feelings of fear, anger or other distress – or combinations of a number of emotions.²⁶¹In all cases we reviewed, however, the defendant’s lethal force was in some way a response to or explained by abuse the defendant had suffered at the hands of the deceased.

It may be problematic to attempt bright-line distinctions between defensive and non-defensive reactions in cases of this kind. In a submission to the Law Commission of England and Wales on reformulation of the partial defence of provocation to accommodate both provoked and fear- based reactions, the Royal College of Psychiatrists cautioned against the assumption that “the two emotions of anger and fear are distinct”, when:²⁶²

... in medical reality they are not. Physiologically anger and fear are virtually identical, whilst many mental states that accompany killing also incorporate psychologically both anger and fear. Hence, the abused woman who kills in response even to an immediate severe threat will also be driven at least partly by anger ... Again, the woman who waits until the man is ‘helpless’ to kill him, is likely not merely to be angry but also fearful that eventually he will kill her, and/or her children ... Any legal solution to the current perceived problems with partial defences to murder which rested upon the assumption that fear and anger can (even usually) be reliably distinguished must, from a medical perspective, therefore fail.

The caution sounded by the Royal College of Psychiatrists in connection with the physiology and psychology of homicidal reactions to abuse may be relevant more generally. The cases we have reviewed often share important elements, but no two are the same. The nature and effects of the defendants’ histories as victims of family violence and the impact of often-entwined factors, like mental health and alcohol or drug issues, differ. So, too, do defendants’ risk profiles and rehabilitative needs and prospects. The circumstances in which victims of family violence come to kill abusers and the circumstances in which they progress through the criminal justice system are not amenable to a straightforward or one-size-fits-all analysis.

259 R v Wickham HC Auckland CRI-2009-090-010723, 20 December 2010 at 39 per Ellis J.

260 See R v Wickham, above n 259; R v Mahari, above n 243; R v Tamati, above n 243; and R v Paton [2013] NZHC 21.

261 R v Brown, above 243, at [5]; R v Woods, above n 242, at [7]; and R v Rihia, above n 229, at [14] and [21].

262 Law Commission of England and Wales Partial Defences to Murder (Law Com No 290, August 2004) at 53. The same point has been made elsewhere. See for example Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 90; and Brenda M Baker “Provocation as a Defence for Abused Women Who Kill” (1998) 11 Can J L and Jurisprudence 193 at 196. Baker argues that “it is hard to credit the idea that women are always driven solely be fear or terror when they kill or risk such killing, although this is a dominant emotion in many homicides and one that is almost always operative to some degree and so almost always has some explanatory relevance”.

Understanding family violence: Reforming the criminal law relating to homicide 49

From our case review, we can make some preliminary general observations and identify some trends:
^._{Whether the defendants were acquitted
or convicted of murder or manslaughter, tragic}

surrounding circumstances are a constant in these cases, but they take a range of forms and engender different reactions.
^._{It is more
common for these defendants to proceed to trial than to enter a guilty plea, and
in}

_{most cases that go to trial, the charge is murder.}
^._{These defendants are, however, rarely convicted of murder. Usually,
they are convicted of}

_{manslaughter or acquitted.}
^._{It is not clear
whether the repeal of provocation has worked against victims of family}

violence who kill their abusers. In terms of charging, there were proportionately more murder trials before repeal of the defence than after, and the proportion of cases where defendants pleaded guilty to manslaughter was broadly similar pre-repeal and post-repeal. In terms of convictions, three of the four murder verdicts we have identified since 2001 were returned in cases where provocation was or could have been run. In the fourth case, R v Rihia, the defendant pleaded guilty. We do not know what would have been the outcome if her case had gone to trial.
^._{While, since R v
Wihongi, victims of family violence convicted of murder appear to have
a}

good case for displacement of the presumption of life imprisonment, there has been only one subsequent murder conviction (R v Rihia), and to date, murder sentences have still been considerably longer than sentences for manslaughter in this area.

We explore these observations and trends, and their implications, further in Part 3.

In recent decades, the legal response to victims of family violence who commit homicide has been a particular focus of law reform both in New Zealand and around the world. The New Zealand Law Commission has examined the position of victims of family violence in two previous reports, Some Criminal Defences with Particular Reference to Battered Defendants (the
2001 Report)²⁶³and The Partial Defence of Provocation (the 2007 Report).²⁶⁴Internationally, the position of victims of family violence who commit homicide has led to law reform in Australia, England and Wales, and Canada.

In this chapter, we summarise the previous work of the New Zealand Law Commission and the range of reforms pursued in other comparable jurisdictions. We further explore the detail of these reforms where relevant throughout this Report.

In 1999, in response to criticism that existing criminal defences were failing to protect those who commit criminal offences as a reaction to domestic violence, the Law Commission was asked to review these defences. It published its Report in 2001.

The 2001 Report was primarily focused on defences to homicide, but the scope of the Commission’s review also covered defences to other crimes, including the defences of compulsion and duress. Various reforms were recommended.

In relation to self-defence, the Commission noted criticisms of the way self-defence was applied in the context of offending by victims of family violence, particularly in respect of reasonableness of force, imminence and proportionality.²⁶⁵The significance of imminence and lack of alternatives had been discussed and confirmed in the 1989 Court of Appeal decision R v Wang, which remains authoritative in New Zealand.²⁶⁶

As the Commission recognised and as we discuss in detail in Part 2 of this Report, the concepts of imminence and lack of alternatives are a poor fit for the realities of family violence and the circumstances in which victims of family violence (who are most often women) kill abusers. The Commission said:²⁶⁷

The Commission considers that self-defence should not be excluded where the defendant is using force against a danger that it not imminent but is inevitable. In many, perhaps most, situations, the use of

263 Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001).

Understanding family violence: Reforming the criminal law relating to homicide 51

force will be reasonable only if the danger is imminent because the defendant will have an opportunity to avoid the danger or seek effective help. However, this is not invariably the case. In particular, it may not be the case where the defendant has been subject to ongoing physical abuse within a coercive intimate relationship and knows that further assaults are inevitable, even if help is sought and the

We agree that the terms of section 48 do not require the courts to exclude self-defence where danger is inevitable but not imminent. However, we think it preferable to make this explicit by legislative reform, rather than to leave the law to be developed case by case. Relying on the courts to develop the law may require a person to be convicted and then to appeal successfully before the legal position is clarified. While the Court of Appeal would be free to change its earlier approach. A trial judge may feel he or she is required to follow the approach in Wang. Until the Court of Appeal has dealt with the matter, the correct interpretation of section 48 would remain unclear, although some trial judges may approach section 48 in terms of inevitability.

Thus, the Commission recommended reform to make it clear there can be situations where the use of force is reasonable even though the danger is not imminent but “inevitable”.²⁶⁸It was also recommended that, whenever there is evidence capable of establishing a reasonable possibility a defendant intended to act defensively, the question of whether the force used was reasonable should always be for the jury.²⁶⁹

In relation to recognising the reduced culpability of victims of family violence who kill other than in self-defence, the Commission considered a range of partial defences that might be relied on by victims of family violence, including provocation, excessive self-defence, diminished responsibility and a special defence for battered defendants. Of all these options the Commission preferred a partial defence of excessive self-defence²⁷⁰but ultimately concluded that reduced culpability should be taken into account during sentencing rather than via a partial defence:²⁷¹

The Commission does not support the retention or creation of partial defences once a sentencing discretion is available for murder. It does not seem fair to make a distinction between those intentional killers who are able to bring themselves within one of the partial defences and those who cannot, but who are nevertheless sentenced to a finite term because of mitigating circumstances.

We agree with the submission that partial defences have proved to be difficult in practice and that it would be easier to take account of the mitigating circumstances they represent in sentencing.

We do not agree with the submission in paragraph 158, that it is preferable for matters of moral accountability to be set out in partial defences rather than to be sentencing considerations. Matters of moral accountability, such as motive and characteristics of the offender, are typically taken into account at sentencing. A judge exercising a discretion must do so within established principles, in open court and must state reasons. If either the offender or the Crown think the discretion was misapplied, and the sentence excessive or inadequate, they can appeal.

Accordingly, the Commission recommended that provocation be repealed²⁷²and that no new partial defence, whether of general application or specific to battered defendants, be introduced

in New Zealand.²⁷³It also recommended that the mandatory life sentence for murder be abolished and replaced with a sentencing discretion.²⁷⁴

The Government considered legislative change to self-defence was unnecessary because, while the Court of Appeal in R v Wang had glossed the “reasonable force” test with a requirement for imminence or immediacy of life-threatening violence, that case did not appear to have been strictly followed in some subsequent cases.²⁷⁵In R v Oakes, a claim of self-defence had been put to the jury even where it did not appear the defendant had acted on an immediate threat.²⁷⁶In R v Zhou, the defendant was charged with the attempted murder of her husband after she drugged and tied him up and chopped him repeatedly with a meat cleaver when he began to struggle. The husband had, hours earlier, raped the defendant, beaten her with the cleaver and made lethal threats.²⁷⁷The defendant successfully claimed self-defence even though she “presumably had other options open to her” when she tied her husband up.²⁷⁸However, as we discuss in Chapter
6, the courts have, since then, confirmed that Wang remains an authoritative statement of the test for self-defence.

The Ministry of Justice also advised that amendment to self-defence was otherwise undesirable. It noted the fact that some defendants fail on self-defence does not mean juries are applying an imminence or immediacy test. The Ministry also considered there was merit in the conclusion of the Criminal Law Reform Committee in its 1979 Report on Self Defence that a self-defence provision should be framed in general terms, with the “infinite variety” of factual scenarios that may constitute reasonable force left for case-by-case assessment.²⁷⁹The Commission’s recommendations on self-defence were, therefore, not adopted.

In relation to the Commission’s recommendation for recognising reduced culpability for murder through sentencing rather than a partial defence, the Government’s initial focus was on statutory enactment of the sentencing recommendations. The Sentencing Act 2002 abolished the mandatory life sentence for murder and replaced it with a rebuttable presumption in favour of that sentence. Section 102 permits a judge to impose a lesser sentence if “given the circumstances of the offence and the offender, a sentence of imprisonment for life would be manifestly unjust”.

Before deciding whether to repeal provocation, the Government in 2004 asked the Commission to consider further the implications of repeal for mentally ill or impaired offenders. There also appeared to be residual doubt about the impact repeal would have on battered defendants.²⁸⁰The Commission undertook further research and consultation before publishing its Report in 2007.

In the 2007 Report, the Commission reiterated its recommendation that the partial defence of provocation should be repealed. Aside from legal, conceptual and practical difficulties with the defence, the Commission identified a more fundamental issue:²⁸¹

275 Ministry of Justice Criminal Defences Discussion Paper: Provocation and Other Partial Defences, Self Defence, and Defences of Duress (2003) at [32].

276 At [32.1], referring to R v Oakes [1995] 2 NZLR 673 (CA). Mrs Oakes had suffered abuse over an 11-year period and ultimately killed her husband by dissolving a large number of tablets in a cup of coffee. This case is discussed in detail in Chapter 6.

277 R v Zhou HC Auckland T 7/93, 8 October 1993; and Ministry of Justice, above n 275, at [32.4].

279 Criminal Law Reform Committee Report on Self Defence (Report 15, November 1979).

Understanding family violence: Reforming the criminal law relating to homicide 53

Section 169 excuses a homicidal loss of self-control, in the face of a provocation of such gravity that it would have prompted a person with ordinary self-control to do likewise. The defence is thus open- ended about the precise emotions that might be driving the defendant; in other words, on its face, provocation is not necessarily confined to an angry loss of self-control, as opposed to one prompted by fear or sympathy. However, anger is the context in which it is commonly understood to operate, and is most frequently used. We would thus argue that the defence puts a premium on anger – and not merely anger, but homicidally violent anger. This, to our minds, is or should be a central issue in considering whether reform is required: out of the range of possible responses to adversity, why is this the sole response that we choose to partially excuse?

The Commission considered whether provocation should be reformed or replaced to retain jury involvement in decisions about culpability for homicide. The options it considered included:²⁸²
^._{a redrafted partial defence of provocation;}
^._{a smorgasbord of partial defences, to capture the broad range of
cases where culpability is}

_mitigated;
^._{a generic partial
defence;}
^._{degrees of murder; and}
^._{replacing the offences of murder and manslaughter with a single
offence of culpable}

homicide.

Overall, the Commission concluded none of these alternatives were viable.²⁸³A “smorgasbord” of partial defences would be arbitrary, since “[t]here is no way of articulating the distinction between what is properly to be regarded as a partial defence, and what is “merely” a mitigating circumstance”.²⁸⁴A generic partial defence would be simpler, but the Commission doubted juries have any particular capacity to arbitrate who is and who is not properly labelled a “murderer”:²⁸⁵

The reality probably is that, in the absence of any legal guidance, the only delineation will be the extent to which a jury sympathises with various defendants and their predicaments. This has the potential to reduce homicide to a lottery: it is an invitation to jurors to dress up their prejudices as law, and substantially increases the risk that more weight will be placed on jury composition and the advocacy skills of defence counsel than on the legal merits of the case.

Accordingly, the Commission recommended that defendants who would have relied on provocation should be convicted of murder, with any evidence of provocation in their case to be weighed with other aggravating and mitigating factors at sentencing.²⁸⁶

The Commission acknowledged that repealing provocation would limit the available options for battered defendants but considered that, for most such defendants:²⁸⁷

[S]elf defence will tactically offer a preferable alternative to provocation, because it results in an acquittal. We adhere to the Law Commission’s previous view that provocation is not benefiting battered defendants sufficiently to warrant its retention, and our review of case law confirms this... While provocation may in the past have offered one option for some battered defendants in New Zealand, it has also arguably been something of a mixed blessing. Although we were not able to confirm it in our own review of recent New Zealand homicide cases, there is a compelling case in the literature to suggest that provocation is a defence typically working against, rather than for, battered defendants –

by the same violent and controlling jealous spouses that have been the subject of much of the feminist critique of this defence.

The issue of self-defence was not substantively revisited in the 2007 Report. Instead, the

In the light of the further work undertaken by the Ministry of Justice on this matter in the development of the government response to Some Criminal Defences with Particular Reference to Battered Defendants, we are content at this stage to concur with the Ministry’s conclusions.

In the 2007 Report, the Commission also recommended sentencing guidelines to address the “manifestly unjust” test in the Sentencing Act, which it anticipated would help guide the length of finite sentences in particular categories of case, including homicides committed by battered defendants.²⁸⁹This recommendation was made on the basis that the Commission’s recommendations in its previous 2006 Report, Sentencing Guidelines and Parole Reform, had been accepted by the Government. In that Report, the Commission recommended the establishment of a Sentencing Council with a mandate to draft sentencing guidelines.²⁹⁰While legislation establishing the Sentencing Council was passed, the incoming Government declined to implement the Council, taking the view that sentencing matters should either be provided for in statute or left to the discretion of the courts.²⁹¹As a result, sentencing guidelines for murder have not eventuated.

In 2001, the Commission identified problems with the way in which self-defence operates in the context of a battered defendant who kills their abuser. It considered that, while the law was capable of accommodating these defendants, legislative reform was preferable to leaving the law to develop case by case. The Government, while recognising these concerns, did not adopt that recommendation, instead preferring to leave the law to be developed through the courts, noting that the approach in recent cases suggested that the law was developing appropriately to accommodate victims of family violence. This issue was not substantively revisited in the 2007
Report. However, as we discuss in Chapter 6, positive development in the case law has stalled, and the concepts regarded as problematic for victims of family violence have been confirmed in the higher courts.

In both the 2001 and 2007 Reports, the Commission preferred that the reduced culpability of victims of family violence be addressed at sentencing and emphasised the merits of sentencing reform. In neither Report did the Commission shrink from concluding that New Zealand’s provocation defence was irredeemably problematic as a matter of theory and practice. We look at this in Part 3 of this Report to understand whether, in light of developments in the law, there is a case for reform around partial defences to properly recognise the reduced culpability of victims of family violence who kill their abusers.

290 Law Commission Sentencing Guidelines and Parole Reform (NZLC R94, 2006) at 13.

291 Hon Simon Power indicated in early 2008 that the Government did not intend to pursue the Sentencing Council, preferring to put the money set aside for it into funding services for victims. Reasons for this appear to be the concern that a Sentencing Council would undermine judicial independence and that the development of detailed sentencing policy would usurp the role of Parliament. See “National to scrap sentencing council” (2 August 2008) Stuff.co.nz <www.stuff.co.nz>; Warren Young and Andrea King “The Origins and Evolution of Sentencing Guidelines – A Comparison of England and Wales and New Zealand” in Andrew Ashworth and Julian Roberts (eds) Sentencing Guidelines: Exploring the English Model (Oxford University Press, Oxford, 2013) 202 at 203. The Sentencing Council Act 2007 is now listed for full repeal in sch 1 to the proposed Statutes Repeal Bill, an exposure draft on which the Parliamentary Counsel Office sought feedback in 2015. See Statutes Repeal Bill (Consultation Draft) at 13.

Understanding family violence: Reforming the criminal law relating to homicide 55

The legal response to victims of family violence who commit homicide has been a focus of law reform activity in a number of jurisdictions since the New Zealand Law Commission published its 2001 Report. Reviews in most states and territories of Australia,²⁹²as well as in England and Wales,²⁹³Ireland²⁹⁴and Canada²⁹⁵have all considered the availability of homicide defences to victims of family violence, either in the context of general reviews of homicide law or reviews of particular homicide defences.

However, despite a concentration of law reform activity, a range of divergent approaches has emerged to address a common concern – namely, the gendered operation of homicide law and, in particular, the difficulties in accommodating the experiences of women who kill in response to intimate partner violence.²⁹⁶While some jurisdictions reformed self-defence to ensure it accommodates the circumstances of victims of family violence, others reformed existing partial defences or introduced new partial defences with victims of family violence in mind.

The range of diverging reforms can be explained in part by the different contexts in which each jurisdiction has pursued reform. While some reforms followed a broad review of homicide, others stemmed from narrower reviews of particular defences, such as provocation, which were constrained by specific terms of reference. Often, therefore, the range of reform options available for consideration was limited.

It is also important to take into account the legal context in which different jurisdictions have pursued reform. The criminal justice systems in each jurisdiction differ in subtle but important ways. There is, for example, a range of approaches to murder sentencing and partial defences, and the requirements for self-defence can vary from jurisdiction to jurisdiction. These factors would likely have influenced approaches to reform. Simple comparisons of law reform pursued in different jurisdictions are not, therefore, always helpful. This said, a unifying feature of all overseas reforms has been better understanding of the circumstances in which victims of family violence – typically women – commit homicide, the features and effects of family violence and changes in social attitudes towards people who suffer abuse in intimate and family settings.

Below, we summarise recent law reform activity in similar jurisdictions in relation to self- defence and partial defences.

292 In Victoria, the Victorian Law Reform Commission reviewed defences to homicide between 2001 and 2004. A similarly comprehensive review of the law of homicide was undertaken by the Law Reform Commission of Western Australia between 2005 and 2007. Provocation was reviewed by the Queensland Law Reform Commission in 2008 and by parliamentary committees in New South Wales in 2013 and in South Australia in 2014. At the federal level, in 2010 the Australian Law Reform Commission with the New South Wales Law Reform Commission reported on a comprehensive review of the legal frameworks at federal, state and territory levels that relate to family violence, which included a review of the availability of legal defences to family violence victims. More recently, in 2015, the Tasmania Law Reform Institute reviewed self-defence. See Victorian Law Reform Commission Defences to Homicide: Final Report (2004); Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007); Queensland Law Reform Commission A review of the excuse of accident and the defence of provocation (Report No 64, September 2008); Select Committee on the Partial Defence of Provocation The partial defence of provocation (New South Wales Legislative Council 2013); Legislative Review Committee Report of the Legislative Review Committee into the Partial Defence of Provocation (Parliament of South Australia 2014); Australian Law Reform Commission and New South Wales Law Reform Commission Family Violence — A National Legal Response (ALRC R114 and NSWLRC R128, October 2010); and Tasmania Law Reform Institute Review of the Law Relating to Self-defence (Final Report No 20, 2015).

293 The Law Commission of England and Wales considered the position of victims of family violence who commit homicide in two related reviews: Law Commission of England and Wales Partial Defences to Murder (Law Com No 290, August 2004); and Law Commission of England and Wales Murder, Manslaughter and Infanticide (Law Com No 304, 2006).

294 Law Reform Commission of Ireland Defences in Criminal Law (LRC 95, 2009). The recommendations made in that Report appear to have been implemented only in part, in relation to legitimate defence in dwelling, by the Criminal Justice (Defence and the Dwelling) Act 2011 (Ireland).

295 The self-defence provisions in the Canadian Criminal Code RSC 1985 c C-46 were reformed by the Citizen’s Arrest and Self-defence Act SC

2012 c 9. In July 2015, the provocation defence in the Criminal Code was also amended by the Zero Tolerance for Barbaric Cultural Practices

296 Kate Fitz-Gibbon and Arie Freiberg “Introduction: Homicide Law Reform in Victoria – Retrospect and Prospects” in Kate Fitz-Gibbon and Arie Freiberg (eds) Homicide Law Reform in Victoria: Retrospect and Prospects (The Federation Press, Leichhardt, 2015) 1 at 1. We discuss concerns relating to the gendered operation of self-defence in detail in Chapter 6. We also note concerns with the gendered operation of the partial defence of provocation in Chapter 10.

Reforms to the law of self-defence to accommodate the circumstances of victims of family violence have been pursued in Victoria, Western Australia and Canada and have been recently recommended in Tasmania.

Reviews in other jurisdictions also identified problems with the operation of self-defence but did not recommend reform given the limited nature of those reviews. For example, the Law Commission of England and Wales noted the problems encountered by victims of family violence when claiming self-defence²⁹⁷but considered self-defence was outside the scope of its review and so did not explore options for its reform.²⁹⁸Similarly, the Queensland Law Reform Commission noted that the battered person “receives no assistance from the law of self-defence, which [in Queensland] requires an assault”,²⁹⁹but because its terms of reference were limited to a review of the partial defence of provocation, it “did not research domestic violence or review the law of homicide as it applies to battered women generally”.³⁰⁰In New South Wales, the Parliamentary Select Committee established to inquire into the partial defence of provocation also noted the significant concerns regarding the adequacy of self-defence and the need to “strengthen” the defence,³⁰¹but as it was not provided with “strong arguments ... on what methods could effectively be used to do so”, it was unable to make a firm recommendation on the issue.³⁰²

Significant changes to homicide laws were introduced in Victoria in 2005, including reforms to self-defence and partial defences (discussed at paragraph [4.43] below).³⁰³These reforms followed a comprehensive review of homicide defences by the Victorian Law Reform Commission (VLRC). In its 2004 Report Defences to Homicide the VLRC made a number of recommendations for reform, recognising that the existing laws of self-defence and provocation failed to accommodate the circumstances in which women commit homicide in response to family violence.³⁰⁴The resultant changes have been held up as a “trendsetting” example of reforms to “remediate gender imbalances in legal responses”.³⁰⁵

The VLRC noted the New Zealand Law Commission’s recommendations in the 2001 Report relating to self-defence (discussed at paragraph [4.7] above) and agreed that the legislation should specify that self-defence may be available “when the person believes that the harm to which he or she responds is inevitable, whether or not it is immediate”.³⁰⁶The VLRC also recommended the legislation specify that self-defence may be available “even though the force used by that person exceeds the force used against him or her”.³⁰⁷In addition to recommending changes to the substantive law of self-defence, the VLRC also made recommendations to support

297 Law Commission of England and Wales Partial Defences to Murder, above n 293, at 79.

298 The Law Commission of England and Wales took the view that, as self-defence was a general defence, it therefore needed to be looked at as a part of a review of the general law rather than specifically in a homicide context. See Law Commission of England and Wales Murder, Manslaughter and Infanticide, above n 293, at 2.

301 Select Committee on the Partial Defence of Provocation, above n 292, at [1.7]. The Select Committee’s reference required consideration of the adequacy of self-defence for victims of prolonged domestic and sexual violence. However, the Committee considered that was limited to the adequacy of self-defence “in the event that the partial defence of provocation were to be abolished”.

305 Debbie Kirkwood, Mandy McKenzie and Danielle Tyson Justice or Judgement? The impact of Victorian homicide law reforms on responses to women who kill intimate partners (Domestic Violence Resource Centre Victoria, Discussion Paper No 9, 2013) at 3.

Understanding family violence: Reforming the criminal law relating to homicide 57

the improved operation of self-defence, including for specific evidence provisions to provide better guidance to judges and lawyers about the sort of evidence that may assist a jury to assess whether the defendant acted in self-defence where there is a history of prior violence, and a recommendation for improved family violence education and training for Police, lawyers and judges.³⁰⁸

The VLRC’s recommendations were largely adopted and, with some modifications, introduced into law by the Crimes (Homicide) Act 2005 (Vic). That Act provides that self-defence is not excluded even if a person is responding to a harm “that is not immediate” or if the response involves the use of force “in excess of the force involved in the harm or threatened harm”.³⁰⁹
However, this applies only where “self-defence in the context of family violence is in issue”,³¹⁰

despite the VLRC’s recommendation that the proposed changes should be generally applicable.³¹¹

An evidence provision setting out the kind of family violence evidence that might be relevant where self-defence is in issue was also introduced.

In 2014, further amendments to self-defence, consistent with the VLRC’s recommendations, were introduced.³¹²At the same time, a provision was introduced that enabled a defendant to request that the trial judge give a jury direction on family violence where self-defence is claimed.³¹³The jury direction is aimed at countering misunderstandings about how the dynamics of family violence may affect the behaviour of family violence victims, such as why victims of family violence remain in abusive relationships.³¹⁴

In 2007, the Law Reform Commission of Western Australia completed its review of the law of homicide and took a similar view to the VLRC in relation to the availability of self-defence to victims of family violence. In particular, it concluded that, “in order to remove any gender- bias associated with the law of self-defence in Western Australia ... it should be made clear that imminence and proportionality are not decisive factors for self-defence”.³¹⁵However, rather than addressing these issues in the substantive provisions on self-defence, it considered the best way to achieve this was by providing that a trial judge must direct a jury about these factors, recommending:³¹⁶

That a new section be inserted into the Evidence Act 1906 (WA) to provide that when the defence of self-defence is raised under s 248 of the Criminal Code (WA) the judge shall inform the jury that:

an act may be carried out in self-defence even though there was no immediate threat of harm, provided that the threat of harm was inevitable; and

that a response may be a reasonable response for the purpose of self-defence ... even though it is not a proportionate response.

312 Crimes Amendment (Abolition of Defensive Homicide) Act 2014 (Vic), s 4. This amendment changed the objective test for self-defence from whether the defendant’s belief was reasonable to whether their conduct was reasonable in the circumstances as perceived by the defendant. This brought the objective test in Victoria in line with the objective test in s 48 of the New Zealand Crimes Act 1961 and with self-defence in other Australian jurisdictions. The Victorian Government considered this new formulation, which limited the objective assessment to the circumstances “as perceived by the accused” was particularly important in the context where a person kills in response to family violence. See Victorian Law Reform Commission, above n 292, at 89–90; and Victoria Department of Justice Defensive Homicide: Proposals for Legislative Reform – Consultation Paper (September 2013) at 41.

313 Crimes Amendment (Abolition of Defensive Homicide) Act 2014 (Vic), s 4. See also the Jury Directions Act 2015 (Vic), ss 58–60.

While the Government appeared to accept the Law Reform Commission’s conclusions, it departed from its recommendation for a jury direction, preferring instead to amend the substantive self-defence provisions to make it clear that a person may be acting in self-defence if the person “believes the act is necessary to defend the person or another person from a harmful act, including a harmful act that is not imminent”.³¹⁷The Explanatory Memorandum to the Bill introducing the reforms explained that:³¹⁸

The reference to whether or not the threatened harmful act is imminent allows this defence to apply to the battered spouse scenario so long as the response is reasonable in the circumstances as the person believed them on reasonable grounds to be. The requirement that the response be reasonable would preclude pre-emptive attacks where it would instead be reasonable for police to be called.

As this amendment was made to the substantive provision on self-defence, it has general effect, unlike in Victoria, where the 2005 self-defence reforms apply only where family violence is in issue. Western Australia also departed from the approach in Victoria and the recommendations of the Law Reform Commission of Western Australia by not adopting the recommendation regarding “proportionate response”. Other recommendations for guidance on the relevance of family violence evidence in the context of self-defence have also not yet been adopted.³¹⁹

The reforms to self-defence in Victoria, and in particular the family violence evidence provisions, have been endorsed elsewhere in Australia. In 2010, the Australian Law Reform Commission and the New South Wales Law Reform Commission published a joint Report: Family Violence – A National Legal Response. That Report considered whether the current defences to homicide available to victims of family violence were adequate across the different Australian jurisdictions.³²⁰Without making specific recommendations as to what defences should be available, they recommended that states and territories adopt evidential provisions along the lines of the Victorian provisions:³²¹

The Commissions maintain their view expressed in the Consultation Paper that state and territory criminal legislation should provide express guidance about the potential relevance of family-violence related evidence in the context of homicide defences, in similar terms to s 9AH of the Crimes Act 1958 (Vic) ...

The Commissions consider that there is considerable merit in focusing attention on the potential relevance of such evidence in homicide defences, given its importance in these circumstances. The Commissions endorse the views of the VLRC that such a provision would assist in avoiding ‘unnecessary arguments concerning ... relevance and ensure the range of factors which may be necessary to represent the reality of the accused’s situation are readily identified’.

In New South Wales, while the Parliamentary Select Committee considering provocation did not make recommendations to reform self-defence as noted above, it did endorse the family violence evidence provisions in Victoria and recommended:³²²

That the NSW Government introduce an amendment similar to section 9AH of the Victorian Crimes Act

_{1958, to explicitly provide that evidence of family violence may
be adduced in homicide matters.}

317 Criminal Law Amendment (Homicide) Act 2008 (WA), s 8, amending the Criminal Code 1913 (WA), s 248(4)(a).

318 Criminal Law Amendment (Homicide) Bill (WA) 2008 (explanatory memorandum), cl 4.

319 Law Reform Commission of Western Australia, above n 292, at 293; and Stella Tarrant “Self Defence In The Western Australian Criminal

320 Australian Law Reform Commission and New South Wales Law Reform Commission, above n 292, at ch 14.

322 Select Committee on the Partial Defence of Provocation, above n 292, at xii.

Understanding family violence: Reforming the criminal law relating to homicide 59

In 2015 the Tasmania Law Reform Institute reviewed the law of self-defence in that State and made recommendations for reform broadly consistent with the self-defence reforms in Victoria, including clarification of the substantive law, as well as family violence evidence provisions and jury directions.³²³

In Canada, reforms were introduced in 2011 to simplify the law of self-defence.³²⁴These reforms clarify that the jury must engage in a contextual analysis of the reasonableness of self-defence claims.³²⁵They also codified an earlier ruling of the Supreme Court of Canada in R v Lavallee that imminence was only a factor to be considered when a defendant is responding to family violence as opposed to being a requirement.³²⁶Other relevant factors listed in the legislation include “the nature and proportionality of the person’s response to the use or threat of force” and the size, age, gender and physical capabilities of the parties in the incident.³²⁷

The Law Commission of Ireland, when considering the codification of self-defence, took a different approach, recommending that the requirements for imminence and proportionality remain but that, to address the “difficult cases” such as family violence killings, the circumstances as the accused reasonably believed them to be should be taken into account.³²⁸
The Commission’s recommendation to codify self-defence has not yet been implemented.

In jurisdictions with partial defences, there has been considerable focus on improving their availability to victims of family violence. Some jurisdictions have reformed existing partial defences, while others have introduced new defences designed with victims of family violence in mind, but no jurisdiction has created a new partial defence for victims of family violence where no general partial defence already existed. The various partial defences operating in other jurisdictions are discussed in Chapter 10. Below, we briefly summarise key reform activity in Australia, England and Wales, and Canada.

In addition to recommending reforms to self-defence, the VLRC also recommended the abolition of provocation, taking the view that degrees of culpability should generally be dealt with through the sentencing process rather than through the continued existence of partial defences.³²⁹However, an exception to this approach was made in relation to excessive self- defence, which it recommended be reinstated, noting that it would provide a partial defence to victims who kill in response to family violence but who cannot rely on self-defence.³³⁰The Crimes (Homicide) Act 2005 (Vic) repealed provocation, but rather than introducing a new partial defence of excessive self-defence, a new offence of “defensive homicide” was introduced, which operated as an alternative verdict to murder where a defendant acted defensively but did not have reasonable grounds for that belief (and therefore could not claim self-defence).³³¹

323 Tasmania Law Reform Institute, above n 292, at viii. The test for self-defence in Tasmania is very similar to the test for self-defence in New

324 Department of Justice, Canada Bill C-26 (SC 2012 c 9) Reforms to Self-Defence and Defence of Property: Technical Guide for Practitioners (March

325 Vanessa MacDonnell “The New Self-Defence Law: Progressive Development or Status Quo?” (2013) 92 Can Bar Rev 301 at 317.

326 R v Lavallee [1990] 1 SCR 852 at 852; and Department of Justice, Canada, above n 324, at 24.

Similar recommendations were made by the Law Reform Commission of Western Australia, but the recommendation to repeal provocation was on the basis that the mandatory penalty of life imprisonment for murder also be abolished in favour of a presumptive life sentence.³³²As in Victoria, a key rationale for Western Australia’s introduction of excessive self-defence was accommodation of victims of family violence who committed homicide but could not rely on self-defence. There was a concern that, without a partial defence “some women may be unjustly convicted of murder if the extremity of their circumstances was not recognised in a trial”.³³³The Criminal Law Amendment (Homicide) Act 2008 (WA) repealed provocation and enacted the partial defence of excessive self-defence.

While excessive self-defence remains a partial defence in Western Australia, defensive homicide proved problematic in practice and was repealed in Victoria in 2014.³³⁴The Victorian Department of Justice identified several problems with the operation of defensive homicide. First, there were inherent problems in trying to distinguish between conduct that amounted to self-defence and conduct that amounted to defensive homicide (or excessive self-defence).³³⁵
Second, while defensive homicide was intended as a safety net for victims of family violence, the Department considered that:³³⁶

On balance ... it is difficult to conclude that this offence clearly works to the benefit of women who kill in response to family violence. Accordingly, it is not clear that it achieves its intended objective. Further, defensive homicide may work to the detriment of women who kill in response to family violence and its existence may inhibit attempts to drive further cultural change in consideration the situation of women who kill in response of family violence.

The Department was, in particular, concerned that the existence of defensive homicide risked suggesting that a woman who kills in response to family violence “is not acting reasonably, or will often not be acting reasonably, and therefore, it is better to plead guilty to defensive homicide than raise self-defence at trial”.³³⁷Finally, the Department identified “clear evidence that defensive homicide inappropriately provides a partial excuse for men who kill”.³³⁸

In a review of the defence of provocation in 2008 the Queensland Law Reform Commission concluded “there can be no doubt that the law of provocation, as it presently works in Queensland does not satisfy the test of substantive gender equality”³³⁹but recommended retention of the defence on the basis that the mandatory life sentence for murder was to remain.³⁴⁰The Commission recommended reform of provocation to ensure it was not available to those who kill out of sexual possessiveness or jealousy,³⁴¹but to accommodate victims of family violence who kill abusers, it preferred the creation of a separate defence for battered persons rather than distorting the defence of provocation.³⁴²

335 Victoria Department of Justice, above n 312, at 21. We discuss these problems with reference to the defence-based partial defences in Chapter

338 At viii. In its seven years of operation, men comprised 25 of the 28 people convicted of defensive homicide. In many of those cases men had killed in circumstances that were very similar to those where provocation previously applied.

Understanding family violence: Reforming the criminal law relating to homicide 61

A new partial defence of “killing for preservation in an abusive domestic relationship” was subsequently introduced in 2010. If successfully argued, this defence would reduce murder to manslaughter where the victim had committed acts of serious domestic violence against the defendant in the course of an abusive domestic relationship and where the defendant believed, on reasonable grounds “having regard to the abusive domestic relationship and all the circumstances of the case”, that it was necessary for their preservation from death or grievous bodily harm to do the act or make the omission causing the death.³⁴³

The introduction of a specific partial defence was (and remains) a novel approach to unique circumstances. As noted above, the Queensland Law Reform Commission was operating under terms of reference limited to a review of provocation and with a clear direction of “the Government’s intention not to change the [mandatory life sentence for murder]”.³⁴⁴The test for self-defence in Queensland is also substantially narrower than in other Australian jurisdictions.³⁴⁵The Queensland approach has attracted criticism from commentators and law reform bodies on the grounds it may jeopardise self-defence claims by victims of family violence³⁴⁶and, as a partial defence, is less helpful than sentencing discretion to recognise mitigating factors and is problematic to the extent it is confined to a single category of defendant.³⁴⁷

The Parliamentary Select Committee established in New South Wales to review provocation recommended its retention and reform in 2013.³⁴⁸One factor in the Committee’s decision to retain provocation was the concern that self-defence did not properly accommodate the circumstances of victims of family violence who commit homicide.³⁴⁹The Committee recommended that provocation be reformed, including reforms to ensure that it:
^._{was more accessible to victims of family violence by
removing the requirement for a loss of}

_{self-control;}³⁵⁰_and
^._{was not available except in cases of “gross provocation”
where the defendant had a justifiable}

sense of being seriously wronged to ensure it did not provide a partial defence to undeserving offenders.³⁵¹

343 Criminal Code 1899 (Qld), s 304B. Few cases have considered this provision. In the first to do so the defendant was in fact acquitted as having acted in self-defence rather than convicted of manslaughter on the basis of the new partial defence: R v Falls Supreme Court of Queensland, 3

345 In Queensland, self-defence requires a reasonable apprehension of death or grievous bodily harm caused by an unlawful and unprovoked assault.

It is therefore extremely difficult, if not impossible, to apply self-defence to a victim of family violence who kills in the absence of an imminent threat. See Criminal Code 1899 (QLD), ss 271-272; and Queensland Law Reform Commission, above n 292, at 313.

346 Thomas Crofts and Danielle Tyson “Homicide Law Reform in Australia: Improving Access to Defences for Women Who Kill their Abusers” [2013] MonashULawRw 28; (2013) 39 Mon LR 864 at 890. See also Anthony Hopkins and Patricia Easteal “Walking in Her Shoes: Battered women who kill in Victoria, Western Australia and Queensland” (2010) 35 Alt LJ 132; Michelle Edgely and Elena Marchetti “Women Who Kill Their Abusers: How Queensland’s New Abusive Domestic Relationships Defence Continues to Ignore Reality” (2011) 13 Flinders LJ 125; Elizabeth Sheehy, Julie Stubbs and Julia Tolmie “Defences to Homicide for Battered Women: A Comparative Analysis of Laws in Australia, Canada and New Zealand” [2012] SydLawRw 22; (2012) 34 Syd L Rev 467; and Jeremy Horder Homicide and the Politics of Law Reform (Oxford University Press, Oxford, 2012) at 208. The Law Commission of England and Wales made a similar observation in its 2006 Report when it noted that a defendant may not wish to run a partial defence (in that case, provocation) if they do “not want the jury to be side-tracked by a partial defence if his or her main claim is for a complete acquittal”: Law Commission of England and Wales Murder, Manslaughter and Infanticide, above n 293, at [5.15].

347 Australian Law Reform Commission and New South Wales Law Reform Commission, above n 292, at 638–642.

New South Wales replaced provocation with “extreme provocation” in 2014, which requires that the conduct of the deceased constitute a “serious indictable offence”.³⁵²The requirement for a loss of control remains, but the legislation provides that it need not occur immediately before the act causing death.³⁵³

In England and Wales, provocation was replaced with the partial defence of “loss of control”³⁵⁴

following a 2004 review of partial defences by the Law Commission of England and Wales.

The partial defence of loss of control is restricted to cases where the defendant has lost self- control in response to a “qualifying trigger”. A loss of self-control will have a “qualifying trigger” if it is attributable to either “circumstances of an extremely grave character”, which “caused [the defendant] to have a justifiable sense of being seriously wronged”, or a “fear of serious violence from [the deceased] against [the defendant] or another identified person”.³⁵⁵In this way, the partial defence has been modernised and is:³⁵⁶

... designed to make a formal statement of symbolic value, in this instance by turning on its head the law’s former implicit endorsement of male violence against unfaithful wives in the way that it shaped the categories of admissible provocation (“qualifying triggers”).

The inclusion of “fear of serious violence” as a qualifying trigger was drafted with victims of family violence in mind.³⁵⁷It is novel, extending the partial defence of provocation to cases where a defendant overreacted to a fear of serious violence with unreasonable force.³⁵⁸This was intended to capture scenarios traditionally caught by defence-based partial defences like excessive self-defence. The defence also excludes anything “said or done constitut[ing] sexual infidelity” as a qualifying trigger.³⁵⁹

Some commentators have criticised the English approach on the basis it adds to the increasingly complex role of the jury in homicide trials.³⁶⁰The VLRC considered but rejected the English model because:³⁶¹

... the provision proposed by the Law Commission [of England and Wales] does not overcome the very real concerns we have about provocation providing a proper basis for a defence. In particular

... it remains an overly subjective assessment of what constitutes sufficient provocation, and involves speculation about how a person might have reacted in the circumstances. While recognising anger as a possible motivator, the provision explicitly excludes actions carried out “in premeditated desire for

353 Crimes Act 1900 (NSW), s 23(2) and (4). The requirement for a loss of self-control is discussed further, in relation to provocation-based defences, in Chapter 10.

357 Catherine Elliot “A Comparative Analysis of English and French Defences to Demonstrate the Limitations of the Concept of Loss of Control” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 231 at 231; and Susan SM Edwards “Loss of Self-Control: When His Anger is Worth More than Her Fear” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 79 at 90.

358 Law Commission of England and Wales Murder, Manslaughter and Infanticide, above n 293, at 87.

360 Kate Fitz-Gibbon Homicide Law Reform, Gender and the Provocation Defence: A Comparative Perspective (Palgrave Macmillan, Hampshire, 2014)

Understanding family violence: Reforming the criminal law relating to homicide 63

In Canada, provocation was restricted in 2015 to serious indictable offences punishable by five years or more in prison to ensure a less serious act or insult would no longer be sufficient in claiming provocation.³⁶²Provocation in Canada continues to require a sudden loss of self- control.³⁶³In contrast, the Law Commission of Ireland recommended that provocation should not be limited to unlawful conduct of the deceased, and that insulting words and gestures that are unacceptable by ordinary community standards should be capable of amounting to provocation, but that provocation should not require the provocation to have occurred immediately before the act causing death.³⁶⁴

Some jurisdictions in Australia have also reformed provocation to exclude non-violent sexual advances.³⁶⁵This issue was considered by the Parliament of South Australia in 2013, and a 2014
Report of the Legislative Review Committee recognised this issue, as well as wider concerns with the operation of provocation, but was unable to settle on options for broader reform.³⁶⁶
Following a high profile court case, the issue is now being reconsidered by that Committee.³⁶⁷

In 2015, the Tasmania Law Reform Institute considered, but did not recommend a partial defence for Tasmania. In doing so it observed the experience with defensive homicide in Victoria³⁶⁸and noted that, unlike Queensland, Tasmania has a broad and flexible self-defence test and discretionary sentencing for murder.³⁶⁹

It is apparent from this brief introduction to overseas law reform activity that a number of comparable jurisdictions have examined the criminal law’s response to victims of family violence who commit homicide. These jurisdictions have recognised, to varying degrees, the difficulties these defendants face in relying on self-defence and establishing reduced culpability where self-defence is not available. It is, however, also apparent there is a range of approaches to reform, but no best practice exists. As we explore in Parts 2 and 3 of this Report, an understanding of the different legal contexts and prevailing policy choices (most obviously, around the mandatory sentence for murder) is crucial to any evaluation of the merits and disadvantages of particular options for reform.

365 See Crimes Act 1990 (ACT), s 13 (amended by the Sexual Discrimination Legislation Amendment Act 2004 (ACT)); and Criminal Code Act

367 Legislative Review Committee Interim Report of the Review of the Report of the Legislative Review Committee into the Partial Defence of Provocation (Parliament of South Australia, PP 244, 2016). That Interim Report resolved that it would not be prudent to make findings or recommendations until the completion of the high profile retrial of Mr Michael Joseph Lindsay on a charge of the murder of Mr Andrew Negre.

Self-defence recognises that a person is justified in using reasonable force in the defence of themselves or another. It is contained in section 48 of the Crimes Act 1961. Self-defence is a complete defence, resulting in an acquittal if successful. It is also a general defence, which means it can be engaged as a defence to justify the use of force by any person against almost any form of attack or threat to that person or any other and is not limited to defence against unlawful assault.³⁷⁰

This chapter discusses the theory and history of self-defence in New Zealand and sets out how it applies generally. The subsequent chapter then discusses the problems that arise in the operation of self-defence where a victim of family violence kills their abuser. An analysis of the options for reform follows in Chapter 7.

Self-defence represents a balance between the needs of an ordered society (in which people are generally not permitted to use force and “take the law into their own hands”) and the right of individuals to ensure their own protection where the state cannot.³⁷¹

While the exact formulations of self-defence vary in different jurisdictions, the theoretical underpinnings of the defence are similar. There is a fundamental requirement of “necessity” to exert the level of defensive force used.³⁷²As one commentator puts it:³⁷³

Society holds life to be sacred, abhors the killing of human beings, and considers premeditated killing among the most offensive of crimes. However, killing may be justified if it is necessary to prevent an act that is as or more offensive than the killing. The law of self-defence therefore justifies killings that are necessary to defend oneself from death or serious bodily injury.

As noted in Chapter 3, self-defence is what is called a “justification-based” defence. It means that any person who successfully claims self-defence was morally justified, and blameless, for acting as he or she did. It is an exception to the criminality of the act.

370 AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington, 2012) at 507; and Simon France (ed)

371 Law Reform Commission of Ireland Defences in Criminal Law (LRC 95, 2009) at 26. See also Leason v Attorney-General [2013] NZCA 509, [2014] 2 NZLR 224 at [64].

372 R v Wang [1989] NZCA 290; [1990] 2 NZLR 529 (CA) at 539; and Boaz Sangero Self-Defence in Criminal Law (Hart Publishing, Portland, 2006) at 3.

373 Nan Seuffert “Battered Women and Self-Defence” (1997) 17 NZULR 292 at 298.

Self-defence has a long history in the common law. The Royal Commission first appointed to consider codification of the criminal law in England said in 1879:³⁷⁴

We take one great principle of the common law to be, that though it sanctions the defence of a man’s person, liberty and property against illegal violence, and permits the use of force to prevent crimes, and to preserve the public peace, and to bring offenders to justice, yet all this is subject to the restriction that the force used is necessary; that is, that the mischief sought to be prevented could not be prevented by less violent means; and that the mischief done by, or which might reasonably be anticipated from the force used is not disproportioned to the injury or mischief which it is intended to prevent.

Self-defence was first codified in New Zealand in the Criminal Code Act 1893, which enacted the provisions drafted by the English Royal Commission. Self-defence was divided into three provisions applying to self-defence against provoked assaults, unprovoked assaults and defence of a person under protection. These provisions were re-enacted in the Crimes Act 1908 and again, largely unmodified, in the Crimes Act 1961.

In 1979, 100 years after the self-defence provisions were first drafted by the Royal Commission, a review of the law of self-defence was carried out in New Zealand by the Criminal Law Reform Committee with a view to rationalising and simplifying the defence.³⁷⁵Problems with the existing provisions included uncertainty as to whether they involved an objective or subjective test and difficulties in determining who started the particular incident (as different tests applied for provoked and unprovoked self-defence).³⁷⁶The Committee reported to the Government in November 1979 and recommended replacing the existing provisions with one simple provision that applied in all cases. The recommended provision was enacted without alteration, becoming section 48 of the Crimes Act 1961.

The Criminal Law Reform Committee favoured a simple comprehensive self-defence provision for the following reasons:³⁷⁷

Briefly, such a provision will require no abstruse legal thought and no set words or formula to explain it; and only common sense is needed for its understanding. The jury will decide the question of reasonableness in the light of the Judge’s summing up of the evidence. In summing up, the Judge will no longer be faced with varying statutory tests and distinctions that are extremely difficult, if not impossible, to explain simply to a jury.

The Committee considered and rejected the option of providing a list of evidentiary guidelines for the court, stating:³⁷⁸

The Judge will in any case sum up to the jury on the evidence relating to such matters as the degree and mode of force used or threatened by the original aggressor or used in his own defence by the accused, the danger apprehended by the accused, and his opportunity (if any) to avoid the original assault or prevent it by other means. But to list such things in the legislation as matters to which the Court must have regard is in our view unwise and unhelpful in relation to self-defence, where the question is one

374 Criminal Code Bill Commission Report of the Royal Commission Appointed to Consider the Law Relating to Indictable Offences: With an Appendix Containing a Draft Code Embodying the Suggestions of the Commissioners (C2345, Eyre & Spottiswoode for HMSO, London, 1879) at 11, referred to in Criminal Law Reform Committee Report on Self Defence (Report 15, November 1979) at [5]. The Royal Commission included Sir James Fitzjames Stephen, who first drafted a bill to state the law relating to murder, which became known as the “Stephen Code”. See Jeremy Finn “Codification of the Criminal Law: the Australasian parliamentary experience” (paper presented to Comparative Histories of Crime Conference, Christchurch, September 2003).

377 At [20]. The Committee stated these reasons were those expressed by Lord Morris, in stating the common law principles of self-defence in

Understanding family violence: Reforming the criminal law relating to homicide 67

of fact to be decided in the light of an infinite variety of circumstances in different cases. It might well introduce into the law complexities of interpretation, resulting in a further body of case law and the risk of elevating evidentiary principles into rules of law.

Everyone is justified in using, in the defence of himself or herself or another such force as, in the circumstances as he or she believes them to be, it is reasonable to use.

It is for the prosecution to prove beyond reasonable doubt that the defendant was not acting in self-defence, and a defendant may be discharged if no jury could properly exclude self-defence. However, before self-defence goes to the jury, there must be evidence of a credible or plausible narrative that might lead the jury to entertain the reasonable possibility of self-defence.³⁷⁹It is for the judge to determine, on the view of the evidence most favourable to the defendant, whether there is sufficient evidence to leave the defence to the jury, and self-defence must be left unless the judge is satisfied that it would be impossible for the jury to entertain a reasonable doubt.³⁸⁰

While section 48 states the law of self-defence in “deceptively simple terms”,³⁸¹it is well established that section 48 involves three inquiries:³⁸²
^._{What were the circumstances
as the defendant believed them to be at the time?}
^._{In
those circumstances, was the defendant acting to defend himself or herself or
another?}
^._{Given that belief, was the force used
reasonable?}

The first two inquiries are subjective. The situation is assessed by looking at it as the defendant did. The third inquiry is objective, although it is applied to the defendant’s subjective view of the circumstances. The Criminal Law Reform Committee, when recommending the current wording of section 48, explained as follows:³⁸³

To restate our proposal in legal terms, we think that a subjective rather than an objective test should be applied in determining the accused’s belief as to the facts, but that an objective test should be used in assessing the accused’s response to the facts as he believes them to be.

_{The subjective inquiries – the “circumstances as he or
she believes them to be”}

As two commentators note, “a cardinal principle of criminal responsibility is that moral obligation is dependent not merely upon the actual facts but also upon the actor’s perception of them”.³⁸⁴Accordingly, where self-defence is raised, a defendant’s conduct is to be assessed according to the circumstances as he or she believed them to be at the time force was used.

This means that, where the defendant is acting under an honest but mistaken or objectively unreasonable belief, the force used must be assessed against that belief as if it were accurate. As

379 R v Wang, above n 372, at 534; R v Tavete [1988] 1 NZLR 428 (CA) at 430; R v Kerr, above n 377, at 340; and France, above n 370, at [CA48.17].

380 R v Wharerau [2014] NZHC 1857 at [8]; R v Wang, above n 372, at 534; R v Tavete, above n 379, at 431; and R v Kerr, above n 377, at 340.

382 R v Li [2000] CA140/100, CA141/100, 28 June 2000 at 6; Fairburn v R [2010] NZCA 44 at [34]; R v Wharerau, above n 380, at [4]; and R v

383 Criminal Law Reform Committee, above n 374, at [23]. See also Simpson v R [2010] NZCA 140 at [68], where the Court of Appeal said that, if a defendant's subjective perception is clouded by alcohol, this is a matter that is able to be taken into account under the subjective part of the s 48 inquiry.

the Criminal Law Reform Committee explained, when recommending the current wording of section 48:³⁸⁵

For example, if the jury determines that the accused believed he was being attacked – when in fact he was not – the jury should nonetheless find that the use of force in repelling the attack was justified unless it is satisfied that the force used was more than was necessary for overcoming the danger the accused thought he faced.

This was recently confirmed by the Court of Appeal in Fairburn v R, which stated:³⁸⁶

A further point of potential significance in this instance is: what if an accused’s beliefs are unreasonable? The accused may have had a misapprehension, and one which might objectively be said to be unreasonable. The present principle in New Zealand is that an unreasonable belief that force was necessary may still support a defence provided that belief is honestly held. As this Court said in R v Savage: “When the knife was used, the accused must have seen himself as under a real threat of danger, and not merely think there may be some future danger to him.” Or, to put it another way, there must be an honest belief of a threat of the requisite danger. Thus, to take an extreme example, even an insane delusion might require the defence to be put to the jury. As Wright has correctly observed, “[t]he cases that really concern the judges seem to be those where the defendant’s view of the circumstances is wholly unreasonable”. But on the present state of the law, so long as the belief is honestly held it does not matter that it was unreasonable.

However, while even an “insane delusion” might need to be put to the jury (as noted in Fairburn), in practice, the more unreasonable the belief, the more the defendant’s credibility may be doubted, and a jury may simply not believe that an unreasonable belief was genuinely held.

The third inquiry in section 48 is an objective one, but it also has a subjective element. The fact- finder must assess whether the defendant’s actions were reasonable in the circumstances as the defendant believed them to be. This requires the fact-finder to “look at the situation ... through the eyes of the accused to determine whether the force may have been reasonable having regard to that person’s perception of the circumstances”.³⁸⁷

The Court of Appeal recently endorsed the judicial direction given by the trial Judge – Woolford

Whether the force used was reasonable, will require consideration of the perceived imminence and seriousness of the attack or anticipated attack, whether the defensive reaction was reasonably proportionate to the perceived danger and whether there were alternative courses of action of which [the defendant] was aware. In the context of lethal or deadly force, reasonableness requires that the force be absolutely necessary.

385 Criminal Law Reform Committee, above n 374, at [23]. See also Simpson v R, above n 383, at [68].

Understanding family violence: Reforming the criminal law relating to homicide 69

Accordingly, whether the force used by the defendant was a reasonable response to the facts as he or she believed them to be depends on:³⁹⁰
^._{the perceived imminence and seriousness
of the attack or threatened attack;}
^._{whether there
were alternative courses of action reasonably available of which the
defendant}

_{was aware; and}
^._{whether the defensive reaction
was reasonably proportionate to the perceived danger.}

How these three concepts operate where a victim of family violence kills their abuser in response to long-term abuse lies at the heart of the problems with the current law of self- defence. The history and problems with these three concepts are discussed in detail in the following chapter.

390 France, above n 370, at [CA48.08]; and Afamasaga v R, above n 381, at [43]–[47].

Self-defence is often claimed by victims of family violence who kill their abusers, but is not usually successful. Our review of cases since 2001 identified that self-defence was claimed in
10 out of 16 cases that went to trial, but only three defendants were successful and were acquitted.³⁹¹Six were convicted of manslaughter³⁹²and one of murder.³⁹³In all 10 cases where self-defence was claimed, the defendant was a woman who had killed an abusive male partner.

Given the weight of academic literature and law reform activity on this topic, it has, as Julia Tolmie puts it, become “trite” to point out that defences to murder do not equitably accommodate the circumstances in which victims of family violence, typically women, tend to kill their abusers.³⁹⁴This inequity is said to arise because the law of self-defence, which developed primarily in the context of male violence and male standards of reasonableness, fails to recognise the different ways in which women typically use defensive force. Moreover, because women are most likely to use defensive force in response to family violence, this bias is deepened by persistent misunderstandings about family violence and victims’ responses to it.

In this chapter, we consider the role of gender in the development of self-defence and how misunderstandings around family violence can disadvantage victims who claim self-defence. We then examine the interrelated concepts of imminence, lack of alternatives and proportionality which the courts have developed to assess whether the use of defensive force was reasonable.

391 The cases of Honor Stephens, Natalie Ford and Jessica Keefe. In relation to Honor Stephens, see Bridget Carter “Jury accepts battered- wife defence in murder trial” The New Zealand Herald (online ed, Auckland, 24 April 2002); Natalie Ford, see R v Ford HC Auckland CRI-2010-044-000132, 22 July 2011, and Edward Gay “Ford found not guilty of murder” The New Zealand Herald (online ed, Auckland, 11

August 2011); and Jessica Keefe, see “Jessica Keefe not guilty of murder” (19 September 2013) Stuff.co.nz <www.stuff.co.nz>, and “Murder charge unwarranted – lawyer” (21 September 2013) Radio New Zealand <www.radionz.co.nz>. We identified one further case in which the defendant, Juliette Gerbes, was acquitted of manslaughter after stabbing her partner. However on the basis of the information available it appears the defendant’s claim was that, while she picked up a knife to scare or stop the deceased during an argument, the deceased grabbed her and pulled the blade into himself. Therefore it was not claimed that the stabbing was an act of self-defence. See “Hung jury in manslaughter trial” (13 February 2014) Stuff.co.nz <www.stuff.co.nz>.

392 R v Mahari HC Rotorua CRI-2006-070-8179, 14 November 2007; Fairburn v R [2010] NZSC 159, [2011] 2 NZLR 63; R v Wickham HC Auckland

CRI-2009-090-010723, 20 December 2010 at 39; R v Rakete [2013] NZHC 1230; R v Paton [2009] NZCA 155; and R v Wharerau [2014] NZHC

1857. In Fairburn, the claim of self-defence was withheld from the jury. That case is discussed further in the following chapter. We also note the case of Daryl Kirk, before the High Court at the time of publication of this Report. Ms Kirk was charged with the murder of her mother’s partner, and claimed self-defence. On 20 April 2016 the jury returned a verdict of manslaughter. Ms Kirk has yet to be sentenced.

394 Julia Tolmie “Battered Defendants and the Criminal Defences to Murder – Lessons from Overseas” [2002] WkoLawRw 5; (2002) 10 Wai L Rev 91 at 91. See Chapter

4 for a summary of recent law reform activity on this issue. Examples from the literature include Susie Kim “Looking at the Invisible: When Battered Women are Acquitted by Successfully Raising Self-Defence” (University of New South Wales, 13–04 UNSWLJ Student Series, 2013); Mandy McKenzie, Debbie Kirkwood and Danielle Tyson “‘Unreasonable’ self-defence?” (2013) 2 DVRCV Advocate 12; Thomas Crofts and Danielle Tyson “Homicide Law Reform in Australia: Improving Access to Defences for Women Who Kill their Abusers” (2013) 39 Mon LR

864; Kellie Toole “Self-Defence and the Reasonable Woman: Equality before the New Victorian Law” [2012] MelbULawRw 7; (2012) 36 MULR 250; Heather Douglas “A consideration of the merits of specialised homicide offences and defences for battered women” (2012) 45 Australian and New Zealand Journal of Criminology 367; Anthony Hopkins and Patricia Easteal “Walking in Her Shoes: Battered women who kill in Victoria, Western Australia and Queensland” (2010) 35 Alt LJ 132; Oliver Quick and Celia Wells “Getting tough with defences” [2006] Crim LR 514; Elisabeth McDonald “Defending Abused Women: Beginning A Critique of New Zealand Criminal Law” (1997) 27 VUWLR 673; Rebecca Bradfield “Is Near Enough Good Enough? Why Isn’t Self-defence Appropriate for the Battered Woman?” (1998) 5 PPL 71; Stanley Yeo “Resolving Gender Bias in Criminal Defences” [1993] MonashULawRw 5; (1993) 19 Mon LR 104; and Stella Tarrant “Something is Pushing Them to the Side of Their Own Lives: A Feminist Critique of Law and Laws” (1990) 20 UWAL Rev 573.

Understanding family violence: Reforming the criminal law relating to homicide 71

We conclude that, while section 48 is, on its face, capable of accommodating the experience of victims of family violence who kill their abusers, these concepts can operate as a barrier to a claim of self-defence in those circumstances. The concern is that defensive action by a victim of family violence can be dismissed as not being an action of self-defence simply because it does not accord with perceptions of what self-defence “really” is.³⁹⁵While evidence of battered woman syndrome has gone some way to addressing these issues, its use is now criticised for a number of reasons, discussed below, and is no substitute for substantive equality in terms of the law itself.

In New Zealand, the overwhelming majority of violent offenders are men.³⁹⁶This is typical across the world.³⁹⁷As a result, the law of self-defence has developed primarily in response to male violence and in the context of male standards of reasonableness. However, there is a substantial body of literature and empirical evidence that points to differences between who, why and how men and women kill and use defensive force.

Men are more likely to kill strangers or acquaintances, while women are most likely to kill those with whom they have an intimate relationship.³⁹⁸However, because female homicide offenders are proportionately small in number, intimate partner homicide is still overwhelmingly characterised by a male killing a female partner (or ex-partner). In New Zealand, the Family Violence Death Review Committee (FVDRC) reports that three-quarters of intimate partner homicide offenders are men and almost three-quarters of homicide victims are women.³⁹⁹These figures mirror recent Australian data.⁴⁰⁰

Research shows that when men kill in the context of intimate relationships, they tend to do so out of jealously or a desire for control.⁴⁰¹Men who kill in this context tend to also have histories of aggression. The FVDRC, in its submission on our Issues Paper,⁴⁰²reported that 79 per cent of intimate partner homicides between 2009 and 2014 with a known abuse history were committed by a male predominant/suspected predominant aggressor.⁴⁰³Women, by contrast,

395 Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 77.

396 Eighty-four per cent of offenders convicted of homicide and related offences in the past 10 years were male: Statistics NZ “Adults convicted in court by sentence type - most serious offence fiscal year” <nzdotstat.stats.govt.nz>. Statistics New Zealand data also shows that for the 12 months ending October 2015, men committed 82 per cent of homicide and related offences, 74 per cent of acts intended to cause injury, 96 per cent of sexual assaults and related offences, and 79 per cent of abductions, harassment and other related offences against a person: Statistics NZ “Recorded crime offenders statistics – unique offenders” <nzdotstat.stats.govt.nz>.

397 Victorian Law Reform Commission Defences to Homicide: Options Paper (2003) at xiv; Toole, above n 394, at 255. However, some suggest that women account for as little as 10 per cent of homicide offending across Australia, Canada, England and the United States: Kathryn Whitely “Women as Victims and Offenders: Incarcerated for Murder in the Australian Criminal Justice System” (Thesis, Queensland University of Technology, 2012) at 12.

398 Whitely, above n 397; Quick and Wells, above n 394, at 524; and Vickie Jensen Why Women Kill (Lynne Rienner Publishers, Boulder, 2001).

399 For the period between 2009 and 2012: Family Violence Death Review Committee Fourth Annual Report: January 2013 to December 2013

400 Tracy Cussen and Willow Bryant “Domestic/family homicide in Australia” (2015) 38 Research in Practice 1 at 2.

402 Law Commission Victims of Family Violence Who Commit Homicide (NZLC IP39, 2015).

403 Family Violence Death Review Committee submission at 9. The terms “predominant aggressor” and “primary victim” are discussed in Chapter

tend to kill intimate partners in response to long-term family violence.⁴⁰⁴The FVDRC reported that 13 of the 15 women who killed an intimate partner between 2009 and 2014 with a known abuse history were primary/suspected primary victims of family violence.⁴⁰⁵Two commentators explain the differences in why men and women kill as follows:⁴⁰⁶

Male killing is about power and control. Women killing abusers is about avoiding power and control ... Women do not often kill from anger, while anger fuels many male killings.

When men kill an intimate partner, it is most commonly by “overkill”, which the FVDRC describes as “the use of violence far beyond what would be necessary to cause death and encompasses multiple stabbings and/or multiple forms of violence”.⁴⁰⁷The way in which a woman kills is often dictated by her physical strength relative to the deceased.⁴⁰⁸A woman may not be able to effectively defend herself with her bare hands in a direct confrontation with a bigger, stronger male partner, and so when women kill intimate partners, a weapon is almost always used, and it is most often a kitchen knife used to inflict one or sometimes two stab wounds.⁴⁰⁹

Because of the differences in who men and women tend to kill, self-defence is claimed in different scenarios. When men kill in self-defence, they normally do so in the context of a spontaneous encounter with a male stranger or acquaintance of relatively equal strength.⁴¹⁰
Women are more likely to claim self-defence in the context of an intimate partner relationship.

Most often, they will be defending themselves against a violent assault, but sometimes they may kill in a non-confrontational situation, while the abuser’s guard is down, rather than waiting to match their strength against their abuser in a direct confrontation.⁴¹¹As was observed by then Victorian Attorney-General Rob Hulls:⁴¹²

[W]omen who kill in response to protracted campaigns of violence against their partners know only too well that any attempt to defend themselves when facing an immediate threat frequently leads to an escalation of the violence against them. Women in this situation are simply not capable of defending themselves then and there, whether because they lack the physical strength, because they are attempting to diffuse the assault or because they are trying to protect children who are, tragically, so often present when this kind of violence occurs.

404 Family Violence Death Review Committee, above n 399, at 40–41; Danielle Tyson and others “The Effects of the 2005 Reforms on Legal Responses to Women Who Kill Intimate Partners” in Kate Fitz-Gibbon and Arie Freiberg (eds) Homicide Law Reform in Victoria: Retrospect and Prospects (The Federation Press, Leichhardt, 2015) 79; Crofts and Tyson, above n 394, at 879; Douglas, above n 394, at 367; Quick and Wells, above n 394, at 524; Victorian Law Reform Commission, above n 395, at 61; Victorian Law Reform Commission, above n 397, at xiv; and Bradfield, above n 394, at 71.

407 Family Violence Death Review Committee, above n 399, at 47. In its submission on the Issues Paper the Family Violence Death Review Committee reported that overkill was the method of killing in 36 of the 85 identified intimate partner homicides for the period 2009–2014. All offenders who used overkill were male: Family Violence Death Review Committee submission at 19.

409 In its submission on the Issues Paper the Family Violence Death Review Committee reported that in 11 out of 12 cases where a female primary victim killed a male predominant aggressor, a kitchen knife or implement was used: Family Violence Death Review Committee submission at

19. All of those homicides took place at the family home, half of them in the kitchen. In our case review, we identified one or two stab wounds as the cause of death in 17 out of 24 cases since 2001. See also Family Violence Death Review Committee, above n 399, at 47; McKenzie, Kirkwood and Tyson, above n 394; and Toole, above n 394, at 256–257.

410 Toole, above n 394, at 256; and Victorian Law Reform Commission, above n 395, at 61.

411 Tasmania Law Reform Institute Review of the Law Relating to Self-defence (Final Report No 20, 2015) at 65; Kim, above n 394, at 6; McKenzie, Kirkwood and Tyson, above n 394; Toole, above n 394, at 256; Victorian Law Reform Commission, above n 395, at 62; and Law Commission Battered Defendants (NZLC PP41, 2000) at 12–13.

412 At the launch of the Crimes (Homicide) Act 2005 (Vic) in October 2005. Victorian Department of Justice Defensive homicide: Review of the offence of defensive homicide: Discussion paper (August 2010) at 30.

Understanding family violence: Reforming the criminal law relating to homicide 73

Not all women who respond to abuse with fatal force will be acting defensively,⁴¹³but the explanation for their conduct is nonetheless likely to be the psychological stress and trauma of intimate partner violence.⁴¹⁴We also note that sometimes women may be motivated by both fear and anger, and it is not always appropriate to try to separate out the two emotions.⁴¹⁵

This is borne out in empirical studies.⁴¹⁶The Victorian Law Reform Commission (VLRC) undertook a study of homicide prosecutions in that State between 1997 and 2001.⁴¹⁷They identified that self-defence was far more commonly claimed by men. It was most often raised in the context of a spontaneous encounter (such as a pub brawl), and it was most likely to be successful when raised in that context. In contrast, while women were most likely to kill in the context of sexual intimacy and in response to alleged violence perpetrated by the deceased, only two women were able to raise self-defence, and neither was successful, both being convicted of murder.⁴¹⁸The VLRC considered that this demonstrated the exclusion of women from the use of self-defence, in that “self-defence is seen to involve a single, isolated attack between two men of approximately equal strength, who are either strangers or acquaintances”.⁴¹⁹

The gendered differences in who, why and how men and women kill mean that the actions of women may not “conform to established patterns of male violence”.⁴²⁰The failure to equally accommodate the ways women use lethal force to defend themselves or another constitutes a gender bias in the operation of the law. As one English commentator notes:⁴²¹

The relative scarcity of female killers has resulted in a paradigmatically male ideal model and this, together with the incompatibility of aggressive force with stereotypical femininity, means that the apparently gender-neutral concept of reasonableness is actually weighted against the female defendant.

Because the overwhelming majority of victims of family violence who kill their abusers are women, the operation of any gender bias in the law is of central importance in this review. We recognise, however, that men can also be victims of family violence and can kill abusive partners or parents.⁴²²We also recognise that family violence can be perpetrated in same-sex relationships. We agree with the VLRC that issues that arise for female primary victims who kill their abusive partners are also likely to arise for other victims of family violence who kill

414 Whitely, above n 394; Fiona Leverick Killing in Self-defence (Oxford University Press, Oxford, 2006); Jensen, above n 398; and Melanie Randall “Domestic Violence and the Construction of ‘Ideal Victims’: Assaulted Women’s ‘Image Problems’ in Law” (2004) 23 St Louis U Pub L Rev 107 at 119.

415 See also paragraph [3.40] above. Law Commission of England and Wales Partial Defences to Murder (Law Com No 290, August 2004) at 53. The same point has been made elsewhere. See for example Victorian Law Reform Commission, above n 395, at 90; and Brenda M Baker “Provocation as a Defence for Abused Women Who Kill” (1998) 11 Can J L and Jurisprudence 193 at 196. Baker argues that “it is hard to credit the idea that women are always driven solely be fear or terror when they kill or risk such killing, although this is a dominant emotion in many homicides and one that is almost always operative to some degree and so almost always has some explanatory relevance”.

416 A number of empirical studies have been undertaken in Australia in particular. See discussion in Tyson and others, above n 404; Toole, above n 394; Victorian Law Reform Commission, above n 397; Bradfield, above n 394; and Tarrant, above n 394.

421 Aileen McColgan “In Defence of Battered Women Who Kill” (1993) 13 Oxford Journal of Legal Studies 508 at 515, discussed in Catherine Elliot “A Comparative Analysis of English and French Defences to Demonstrate the Limitations of the Concept of Loss of Control” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham,

422 We are aware of at least four cases in New Zealand in which the defendant was a son or stepson of the deceased and killed in response to a long history of abuse from the deceased, although none involved a claim of self-defence: R v Raivaru HC Rotorua CRI-2004-077-1667, 5 August

abusers, particularly when their relationship is marked by the characteristics of coercion and control discussed in Chapter 2.⁴²³

In the context of self-defence, a jury’s understanding of family violence affects how they assess a claim of self-defence by a victim of family violence. Misconceptions can undermine their assessment of the defendant’s credibility or the reasonableness of his or her actions.

We identified in Chapter 2 persisting misconceptions, including the belief that a primary victim of family violence can avoid further violence by leaving an abusive relationship; that fear of future violence is irrational or unreasonable; and, if the primary victim used violence in the past, her fear was not real. We discussed the need to understand family violence as a pattern of harmful behaviour with a cumulative effect and a form of entrapment. Victims’ responses must be considered in the context of:⁴²⁴
^._{the manner in which their choices have been constrained by the
violence they have}

_experienced;
^._{what the past responses to their
help seeking have been; and}
^._{the wider structural
constraints of their lives, including the structural constraints of
their}

families, whānau and communities.

This shift in thinking is necessary to counter misconceptions that can affect a jury’s assessment of self-defence claims by victims of family violence. In a similar vein, the concepts of imminence, lack of alternatives and proportionality are difficult to reconcile with contemporary understanding of family violence, as we discuss below.

In judging the reasonableness of the force used by a defendant, courts have traditionally required there to be an immediacy of life-threatening violence to justify the killing of another in self-defence.⁴²⁵This is known as the requirement for “imminence”. Closely related to the concept of imminence is whether the defendant had an alternative to the use of force, such as leaving or seeking help from Police.

_{Problems with imminence and lack of alternatives in the context of
family violence}

Imminence and lack of alternatives developed in the context of the stereotypical danger envisaged by self-defence, which emanates from an immediate, violent one-off confrontation with a stranger or acquaintance.⁴²⁶Many argue that these concepts are difficult to reconcile with the contemporary understandings of family violence.⁴²⁷As we explained in Chapter 2, intimate partner violence is best understood as a cumulative pattern of harmful behaviour that

424 Family Violence Death Review Committee Fifth Report: January 2014 to December 2015 (Health Quality & Safety Commission, February 2016)

427 See for example Tasmania Law Reform Institute, above n 411, at 65; Crofts and Tyson, above n 394, at 877–880; Angelica Guz and Marilyn

McMahon “Is Imminence Still Necessary? Current Approaches to Imminence in the Laws Governing Self-Defence in Australia” (2011) 13

Flinders LJ 79 at 118–120; Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September

2007) at 274; Victorian Law Reform Commission, above n 395, at 78; Bradfield, above n 394, at 76–77; Nan Seuffert “Battered Women and

Self-Defence” (1997) 17 NZULR 292 at 300–301; Yeo, above n 394, at 115–116; and Tarrant, above n 394, at 597–599.

Understanding family violence: Reforming the criminal law relating to homicide 75

is bigger than the incidents of physical violence on any one occasion.⁴²⁸In some cases, coercive and controlling tactics by a predominant aggressor can mean a primary victim faces a constant and ongoing threat. Imminence, however, focuses on danger that is close at hand. As the Law Commission observed in 2001, this limits the inquiry to the discrete incident of violence or threat immediately preceding the defendant’s use of force.⁴²⁹As one commentator explains:⁴³⁰

The model of self-defence which focuses on an isolated ensuing or imminent assault envisages a killing in response to an extraordinary eruption in normal existence. It is the interruption of normal existence which allows for the deviation from a simple application of the laws against killing. Such a model may be appropriate in the context of an isolated contest between strangers. However, women who kill in retaliation to systemic abuse are killing in response to an aspect of their ordinary existence and the law at its most primary level does not contemplate the possibility of the need to defend oneself against normality. The danger faced by women in violent intimate relationships is embodied not in an isolated attack, nor even in a series of attacks, but in an on-going life of being abused and the fear which accompanies that life. The nature of these two kinds of “assaults” are of a fundamentally different nature.

Because victims of family violence are not facing the stereotypical scenario of a one-off confrontation, the way in which they respond is different. They may respond where an immediate assault appears relatively minor, or they may respond at a time when the violence has ceased. They may arm themselves in anticipation of an attack, or they may act to protect themselves by a pre-emptive strike.⁴³¹

Focusing on the immediate event means that juries may be less likely to hear evidence on the relationship history.⁴³²Even if the jury does hear such evidence, it may only be for the limited purpose of understanding the circumstances of the immediate event in determining whether the defendant honestly apprehended death in light of the latest incident.⁴³³One commentator notes several problems with introducing evidence for this purpose. First, evidence of past abuse tends to be elicited as a “list” of events, with little connection made between it and the killing, and without eliciting the experience and effects of living a life of being abused.⁴³⁴Second, the weight of evidence of past abuse diminishes over time, and abuse that occurred several years before will be of less importance, whereas for an abused woman, its significance may increase if it indicates the fear with which she was living.⁴³⁵Third, the introduction of evidence of past abuse may in fact work to undermine the reasonableness of the defendant’s actions, as if she had survived all past abuse the danger may be seen to be less serious, or if she had “warning” of the danger then she could have left.⁴³⁶Such evidence, it is argued, is not sufficient to enable a jury to gain a proper understanding of the circumstances of the killing.⁴³⁷

A further problem with the concept of imminence is it assumes that, where a threat cannot be carried out immediately, there will be options other than the use of force to avoid the threatened harm, such as fleeing or calling the Police. However, such options may simply delay the threatened harm. As the Law Commission said in 2001:⁴³⁸

429 Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001) at 10.

431 Bradfield, above n 394, at 76; Robertson, above n 425, at 279; and Tarrant, above n 394, at 598.

Threats of future violence are typically used to keep a battered woman from leaving a relationship ... Going to the police, for example, is a reasonable alternative to the use of force if thereby the defendant obtains effective protection, but not if it will ensure the defendant’s safety only for as long as she is in the presence of police officers.

The Supreme Court also recently observed, when discussing the defence of compulsion,⁴³⁹that the criteria of immediacy will be “highly indicative” of coercive circumstances leaving no practical alternative, but that it may be under-inclusive, because “threats which do not meet the immediacy ... criteria may none the less be very coercive in the sense of leaving no reasonable alternative”.⁴⁴⁰

For these reasons, it is widely accepted that “the traditional understanding of what amounts to imminent peril is unsuited to self-defence in the context of [family] violence”.⁴⁴¹It can impede victims of family violence from relying on self-defence even where they genuinely believe there is no option to escape death or serious injury other than to kill their abusers.⁴⁴²

We consider below how the courts currently apply these concepts in New Zealand.

The leading case on imminence and lack of alternatives in the context of a victim of family violence who kills their abuser is R v Wang.⁴⁴³Xiao Jing Wang was charged with the murder of her husband Jing Wah Li after stabbing him while he slept. Ms Wang was an immigrant from China. She was married to Mr Li for approximately four years before killing him. There was evidence that earlier in the evening of the fatal stabbing the deceased had threatened to kill Ms Wang and her sister if her family in China did not send him money. The trial Judge also noted that, if Ms Wang’s evidence was believed, there was “ample material” about past threats by the deceased.⁴⁴⁴At trial a psychiatrist gave evidence that Ms Wang was suffering from a major depressive illness and, in her mental state, she would have believed that the threats of her husband would be carried through, and “the only course she could think of was to kill her husband”.⁴⁴⁵

The trial Judge withheld the issue of self-defence from the jury. This was on the basis that “a reasonable person in the accused’s position had a number of alternative courses open to her” and “the only view of the evidence open is that the accused was in no immediate danger”.⁴⁴⁶On the question of imminence, the Judge said:⁴⁴⁷

Here there is no suggestion that the victim had a weapon, nor had made any move to suggest the intended use of any object as such. The contention on behalf of the defence has to go to the length of asserting that a jury could reasonably find that an accused under no immediate threat or danger,

439 Akulue v R [2013] NZSC 88, [2014] 1 NZLR 17. In that case the appellant was Nigerian and faced charges of importing and conspiring to supply methamphetamine. He claimed that he was coerced into the offending by a cousin living in Nigeria who threatened to kidnap and kill members of his family also living in Nigeria unless he assisted in the importation of drugs into New Zealand. The Supreme Court was asked to consider whether he could rely on the defence of compulsion or necessity. The defence of compulsion is provided for in s 24 of the Crimes Act 1961 and is narrowly constructed to only apply where the defendant is under compulsion “by threats of immediate death or grievous bodily harm from a person who is present when the offence is committed”. The Supreme Court, at [23]–[24], while recognising that s 24 could be seen as underinclusive, concluded that it was not drafted with a view to allowing a defence of compulsion based on the belief of the defendant that assistance from the authorities would not be forthcoming. This can be contrasted to s 48, which is clearly drafted to require the fact-finder to determine whether the use of force was reasonable “in the circumstances as [the defendant] believes them to be”.

444 R v Wang HC Christchurch T 40/88, 27 February–6 March 1989 (trial rulings) at 10.

Understanding family violence: Reforming the criminal law relating to homicide 77

however elastic an interpretation is given to that concept, who had alternative courses open none of which she had tried or seemingly considered, was or at least might reasonably be justified in deliberately killing the other party with a knife. To accede to that proposition in these circumstances would I think be close to a return to the law of the jungle.

The Court of Appeal upheld the decision of the trial Judge, stating:⁴⁴⁸

... what is reasonable force to use to protect oneself or another when faced with a threat of physical force must depend on the imminence and seriousness of the threat and the opportunity to seek protection without recourse to the use of force. There may well be a number of alternative courses of action open, other than the use of force, to a person subjected to a threat which cannot be carried out immediately. If so, it will not be reasonable to make a pre-emptive strike.

The Court considered that:⁴⁴⁹

In our view what is reasonable under the second limb of s 48 and having regard to society’s concern for the sanctity of human life requires, where there has not been an assault but a threatened assault, that there must be immediacy of life-threatening violence to justify killing in self-defence or the defence of another.

The Court of Appeal concluded that “it would be impossible for the jury to entertain a reasonable doubt on the point” as there was no imminent danger to be averted by instant reaction, and Ms Wang was not held hostage and was free to seek protection in other ways.⁴⁵⁰
The decision in Wang is, therefore, authority for need for immediacy of life-threatening

Wang was decided over 25 years ago, and has been considered and followed in the higher courts on a number of occasions.⁴⁵¹Cases decided since Wang have done two things. First, they have confirmed the requirement for imminence. Second, they have clarified the approach to considering alternatives to the use of force.

Because cases involving victims of family violence claiming self-defence are uncommon the higher courts have not considered the operation of imminence and lack of alternatives in this specific context since Wang. Accordingly, subsequent consideration of Wang has typically been in markedly different factual circumstances involving a male defendant and a male homicide victim.⁴⁵²They establish how the law of self-defence is to be applied in a general sense rather than in the specific context of family violence.

Early indications following Wang were that the requirement for imminence was not being strictly enforced by the courts. In the early 1990s cases of R v Zhou⁴⁵³and R v Oakes,⁴⁵⁴both involving the use of force by a victim of family violence against their abuser, the question of
self-defence was left to the jury even though in neither case was there an immediate threat of

harm present on the facts.⁴⁵⁵Neither case cited Wang. The Ministry of Justice quite reasonably inferred from these cases that Wang was not being strictly followed and thus victims of family violence were not being improperly excluded from relying on self-defence.⁴⁵⁶As a result, the Government did not implement the Law Commission’s recommendations on self-defence in the
2001 Report,⁴⁵⁷preferring instead to allow the courts to continue to develop the law on a case-

by-case basis.

However, more recent Court of Appeal decisions, albeit not in cases in the family violence context, confirm that imminence remains central to the law of self-defence. In R v Richardson,⁴⁵⁸the appellant appealed his convictions for firearms possession. At trial he had claimed self- defence, arguing that he was under a “constant” threat of attack from an acquaintance as a result of his involvement as a prosecution witness in a murder trial. Dismissing his appeal against conviction, the Court of Appeal, while not citing Wang, concluded that:⁴⁵⁹

[T]here was no suggestion that there was any imminent danger ... the case seems to have been predicated on the quite unsustainable premise that if a person subjectively believed that they were under constant threat, they would be able to carry loaded weapons. That is not the law and the ground is unmeritorious.

The FVDRC considers that this decision has the potential to present a legal barrier to raising self-defence for a victim of family violence who is responding to an “omnipresent threat which has the potential to crystallise at any point in time and which the police are unable to defuse”.⁴⁶⁰

More recently, two Court of Appeal decisions have considered and approved Wang, and confirmed the role of imminence and lack of alternatives in any claim of self-defence.

In Vincent v R, the appellant was a prison inmate who stabbed another inmate four times in the neck. The attack followed an incident in the exercise yard four days earlier, where it was alleged the victim deliberately kicked a basketball towards Mr Vincent. At trial, the appellant claimed he was acting pre-emptively in self-defence in response to a threat of future violence from the victim. He appealed his conviction for wounding with intent to cause grievous bodily harm on the ground that the trial Judge erred in withdrawing self-defence from the jury.

The Court of Appeal referred to Wang and observed that, in certain circumstances, self-defence may be available where a defendant takes pre-emptive action to defend himself or herself (or another) from a perceived threat.⁴⁶¹The Court said:⁴⁶²

While the imminence of the threat is not treated as a distinct or separate requirement, the authorities have emphasised that the imminence or immediacy of the threat is a factor that is to be weighed in assessing whether the defence is available. This is a question of fact and degree. Amongst other things, the opportunities available to the defendant to seek protection or adopt some other alternative course of action are to be considered. The defendant must have seen himself or herself as under a real threat of danger and not merely believe there may be some future danger.

455 In Zhou the defendant was charged with attempted murder of her husband after she drugged and tied him up and chopped him repeatedly with a meat cleaver when he began to struggle. The husband had, hours earlier, raped the defendant, beaten her with the cleaver, and made lethal threats. In Oakes¸ the defendant and her children were physically, sexually and emotionally abused by her partner. One night after being verbally abused and threatened she spiked his coffee with more than 30 sleeping pills, causing him to die. Oakes is discussed in further detail at paragraph [6.57] below.

456 Ministry of Justice Criminal Defences Discussion Paper: Provocation and Other Partial Defences, Self Defence, and Defences of Duress (2003) at 8.

Understanding family violence: Reforming the criminal law relating to homicide 79

The Court of Appeal went on to consider Mr Vincent’s claim of self-defence against the concepts of imminence, lack of alternatives and proportionality and upheld the Judge’s decision to withhold self-defence from the jury, finding “there was no realistic possibility that the jury could entertain a reasonable doubt” that the defendant was acting defensively in terms of section 48.⁴⁶³Mr Vincent sought leave to appeal to the Supreme Court, but that was declined. The Supreme Court observed:⁴⁶⁴

Although the proposed appeal raises questions as to when self-defence is available in the case of a pre- emptive strike, the absence of immediacy in relation to the alleged threat and the alternatives available to the applicant were material considerations. In dealing with this part of the case, the Court of Appeal applied settled law.

The concepts of imminence and lack of alternatives were considered again by the Court of Appeal in the case of Afamasaga v R. The prosecution’s case was that, with five others, Mr Afamasaga, a gang prospect, planned to shoot the leader of a rival group. The plan involved provoking the victim to turn up at an address where he knew he would find Mr Afamasaga and his associates. As the victim got out of the car and walked towards the house the appellant, who was in a darkened bedroom about 10 to 12 metres away, shot him. Mr Afamasaga claimed self- defence, arguing that he feared for his life and that he thought he saw a pistol in the deceased’s hands. The trial Judge’s summing up to the jury included the following passage:⁴⁶⁵

Whether the force used was reasonable, will require consideration of the perceived imminence of the seriousness of the attack or anticipated attack, whether the defensive reaction was reasonably proportionate to the perceived danger and whether there were alternative courses of action of which Mr Afamasaga was aware.

Mr Afamasaga was convicted and appealed, arguing that the trial Judge failed to direct the jury adequately on self-defence. Citing Wang, the Court of Appeal said: “As the Judge properly emphasised, a threat has to be ‘imminent or immediate [with] no alternative available’ for a pre- emptive strike to amount to self-defence”.⁴⁶⁶

The Court of Appeal concluded that the Judge directed the jury correctly on the elements of self-defence.⁴⁶⁷

While the cases of Richardson, Vincent and Afamasaga did not involve victims of family violence who had used force against their abusers, as we read them these cases comprise general statements about the law of self-defence that could apply equally to a defendant in that context. Afamasaga, in particular, clearly confirms that imminence and lack of alternatives are necessary elements of any claim of self-defence. It would be open to the Court of Appeal or Supreme Court to take a different approach in a specific context. However, given that cases involving victims of family violence who kill their abusers are uncommon and given how recently Vincent and Afamasaga were decided, it seems unlikely that the roles of imminence and lack of alternatives will be revisited in the near future.

465 R v Afamasaga [2014] NZHC 2142 at [68], cited in Afamasaga v R, above n 451, at [43].

In Wang the Courts determined that self-defence was not available because the defendant “was not held hostage and was free to seek protection in other ways”.⁴⁶⁸However, as one commentator notes:⁴⁶⁹

Wang herself may have seen the circumstances differently. She had been physically, sexually and emotionally abused throughout the marriage. She was an immigrant to New Zealand and spoke little English. The Court of Appeal admitted that she “was not conversant with social opportunities or avenues for help” and described this as a case of “a weakening of the accused’s ability to reason leading to a situation with an apparent absence of alternatives”.

Several commentators noted with concern that the approach in Wang meant that reasonableness was considered in that case on an almost purely objective basis, ignoring the defendant’s subjective beliefs about the options available to her.⁴⁷⁰Ms Wang had testified that “I had to kill him, there was no other way”.⁴⁷¹If this was taken as being, in the words of section
48, the “circumstances as she believed them to be”, then “it is hard to see how the question of

the reasonableness of the force used should not have been left to the jury”.⁴⁷²It was argued the proper interpretation of section 48 is that any belief a defendant has about an available option is a “circumstance” he or she believes to exist for the purposes of section 48 and must, therefore, be treated as if it was correct.⁴⁷³

It appears that this issue has been clarified by the Court of Appeal in the subsequent cases of Fairburn v R,⁴⁷⁴McNaughton v R⁴⁷⁵and Afamasaga v R.⁴⁷⁶In Fairburn, the Court of Appeal said: “The present principle in New Zealand is that an unreasonable belief that force was necessary may still support a defence provided that belief is honestly held”.⁴⁷⁷

As one commentator notes, if self-defence in Wang was withheld because the defendant’s belief that “she had to kill him, there was no other way” was unreasonable, although honestly held, that seems contrary to the approach in Fairburn.⁴⁷⁸

In the second case, McNaughton, the Court of Appeal considered the finer issue of the appellant’s belief about the options available to avoid the use of force. The appellant had shot the deceased during a pre-arranged fight between rival groups and claimed self-defence. A core component of the prosecution’s case was that he had a number of alternative options when he fired the gun. He could have fired into the ground or the air or fled.⁴⁷⁹The trial Judge, in a written memorandum for the jury, identified the issue as whether there were “other options that he knew he might take in the time available, such as getting help or fleeing”.⁴⁸⁰Responding to a question from the jury during deliberations, Miller J gave the following direction:⁴⁸¹

469 Fran Wright “The Circumstances as She Believed Them to Be: A Reappraisal of Section 48 of the Crimes Act 1961” (1998) 7 Wai L Rev 109 at

470 At 123–125; McDonald, above n 394, at 681; Seuffert, above n 427, at 312; Jeremy Finn “R v Xiao Jing Wang (Case and Comment)” (1990) 14

Crim L Rev 200 at 202–203; and Kurt Van Wirkervoort Crommelin “Case Note: R v Wang” (1990) 3 Auckland U L Rev 456 at 458–459.

Understanding family violence: Reforming the criminal law relating to homicide 81

The belief that he had other options is a dimension of reasonable response. In other words, you may conclude that the force used was unreasonable if the Crown have proved to beyond reasonable doubt,

– and that is a matter for you – that he knew he had other options available to him and those options were reasonably available to him in the time he had to react. That is for you to assess.

Mr McNaughton appealed his conviction for murder, arguing the prosecutor failed in cross- examination to directly challenge his claim that he did not have time to think of other options. The Court of Appeal, considering the sufficiency of Miller J’s direction, said:⁴⁸²

As noted, the prosecutor had not questioned [the defendant] on the options said to be available – principally, firing the gun into the air or the ground or fleeing. The first or subjective stage of the self- defence inquiry had not been tested by the Crown. The situation required that the Judge direct the jury to consider whether the Crown had proved by an appropriate evidential foundation both that, to use the Judge’s words, Mr McNaughton (a) “knew he had other options available to him” and (b) “those options were reasonably available to him in the time he had to react”. An explicit direction of this nature was required but was not given.

Accordingly, the Court of Appeal concluded that the trial Judge’s direction was insufficient and thus in error. For this and other reasons, the Court allowed the appeal, quashing the appellant’s conviction and ordering a retrial.⁴⁸³

The third and most recent case, Afamasaga, is discussed in detail at paragraph [6.38] above. In that case, the appellant challenged his conviction for murder on the basis that there were deficiencies in the trial Judge’s summing up, including on self-defence. The Court of Appeal was satisfied that the trial Judge directed correctly on the elements of self-defence, which included the direction that whether the force used was reasonable requires consideration of “whether there were alternative courses of action of which Mr Afamasaga was aware”.⁴⁸⁴

In our view, the decisions in Fairburn, McNaughton and Afamasaga clarify that the reasonableness of a defendant’s use of force should be assessed objectively but on the basis of any beliefs the defendant held about the “circumstances”. This includes beliefs both as to the nature and seriousness of the threat and the options available to avoid the use of force. Where a victim of family violence kills their abuser and claims self-defence, this appears to allow the jury to go further than the courts’ consideration of alternatives in Wang. It allows them to consider, for example, how any language and cultural barriers could have influenced the defendant’s assessment of the situation and their ability to leave and whether they could have sought effective protection from going to the Police. In some cases, it may be reasonable for the jury to conclude that leaving was not a viable alternative.

When assessing the reasonableness of defensive force, the decision maker must consider, alongside the imminence of the threatened harm and alternatives to the use of force, whether the defensive reaction “was reasonably proportionate to the perceived danger”.⁴⁸⁵

As discussed above, due to differences in size and strength, women are likely to use a weapon, typically a kitchen knife, to defend themselves against an abusive partner armed only with fists. Further, a victim of family violence is likely to be responding to a threat that emanates from a cumulative pattern of harm rather than a single threat or event. For these reasons, many argue

that it can be difficult to apply a simplistic measuring process based on the proportionality of the force used against the threatened harm.⁴⁸⁶A defendant’s actions may not “look” proportionate, particularly if all the circumstances are not understood.⁴⁸⁷When a victim of family violence kills their abuser in a non-confrontational situation, proportionality is even more difficult to measure, because at that particular moment, the deceased was incapable of using force.

The term “proportionality” can create confusion. Sometimes it is used and understood simply as a synonym for reasonableness,⁴⁸⁸but this can be misleading because a disproportionate response can still be reasonable. As the Court of Appeal noted in R v Howard, reasonable force:⁴⁸⁹

... may include force which is not in reasonable balance with the believed threat, if for instance the accused has no real choice of means, other than a means which might be seen in the normal course as way out of balance with the threat.

The recent Court of Appeal decision of Afamasaga approved the statement of the law of self- defence by Woolford J in the High Court, set out at paragraph [6.38] above, which identifies “reasonable proportionality” as a factor in – but not a replacement for – the assessment of reasonable force. In our view, this is the right approach because it acknowledges that a disproportionate response may still be reasonable.

Some argue that the law gives conflicting signals about the degree of force that may permissibly be used in self-defence.⁴⁹⁰On the one hand, the law does not require strict proportionality; a person defending himself or herself “cannot weigh to a nicety the exact measure of his [or her] necessary defensive action”.⁴⁹¹On the other hand, section 62 of the Crimes Act provides that everyone authorised to use force is criminally liable for any excess. In this discussion, we are concerned with force that may be disproportionate but still reasonable – rather than force that is excessive and unreasonable.

The concepts of proportionality and imminence are interconnected. In Wang the Court of

... prior to his falling asleep on his bed in a drunken state, her husband had not armed himself in any way to carry out his threat to kill ... [therefore] one could not reasonably have considered that those threats might be carried out by him, “at any moment”, in his then state.

Proportionality was also in issue in the 1995 case of R v Oakes, discussed at paragraph [6.31]

above.⁴⁹⁴In that case, the defendant and her children had been physically, sexually and

486 Kim, above n 394, at 7; McKenzie, Kirkwood and Tyson, above n 394; Toole, above n 394, at 257; Mark Campbell “Pre-Emptive Self-Defence: When and Why” (2011) 11 OUCLJ 79 at 96–99; Bradfield, above n 394, at 77–78; Seuffert, above n 427, at 301–302; Yeo, above n 394, at 108; Law Reform Commission of Western Australia, above n 427, at 165–166; Law Reform Commission of Ireland Defences in Criminal Law (LRC

95, 2009) at [22.04]; Victorian Law Reform Commission, above n 395, at 83; and Law Commission of England and Wales, above n 415, at [4.20].

488 For example, see R v McGrath [2010] EWCA Crim 2514, [2010] All ER (D) 185 at [5], in which the Court said the “critical question for the jury is: was his response reasonable, or proportionate (which means the same thing)”. Similar difficulties arise with the term “excess force”, which has a special meaning in s 62 of the Crimes Act 1961 and is also associated with the partial defence of “excessive self-defence”, which is not recognised in New Zealand. Section 62 of the Crimes Act provides that everyone authorised by law to use force is criminally responsible for any excess, according to the nature and quality of the act that constitutes the excess.

490 AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington, 2012) at 528; and Warren Brookbanks “Partial Defences to Murder in New Zealand” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 271 at 286.

Understanding family violence: Reforming the criminal law relating to homicide 83

emotionally abused by her partner for 11 years. Ms Oakes had previously complained to Police, obtained protective orders against the deceased and had fled several times to a women’s refuge.⁴⁹⁵One night, after being verbally abused and threatened and fearing a serious beating, Ms Oakes spiked the deceased’s coffee with more than 30 sleeping pills. Ms Oakes claimed that her actions were in self-defence, and that she was acting in a state of fear and panic. The prosecution argued that this was a case of premeditated murder, committed in order to rid the defendant of a man who would not leave her alone and to rid her daughter of her abuser.⁴⁹⁶
There was evidence that Ms Oakes had spiked the deceased’s coffee the previous day “to obtain some respite from him”,⁴⁹⁷and a witness for the prosecution, P, gave evidence that Ms Oakes told her that she wanted to kill him and that, once the deceased “was out of it” from the drugged coffee, she proceeded to eye drop other drugs into his throat.⁴⁹⁸Ms Oakes denied this, but her claim of self-defence was rejected by the jury, and she was convicted of murder.

Ms Oakes appealed her conviction, arguing that the Judge’s summing up, both on the relevance of battered woman syndrome (which we discuss at paragraphs [6.71]–[6.74] below) and on the requirements of self-defence, was inadequate.⁴⁹⁹The trial Judge had explained the requirements of self-defence by comparing “shooting a person about to slap one in the face on the one hand, and shooting an assailant wielding an axe on the other”.⁵⁰⁰The appellant argued this example was “male-oriented and not helpful in the context of this case”.⁵⁰¹

The Court of Appeal noted that the reasonableness of the defendant’s response is to be judged in the light of her perception of the threat and, accordingly that “the use of more drastic means than might otherwise be thought appropriate” might be justified.⁵⁰²The Court then said:⁵⁰³

It is important too in the assessment of the nature of the woman’s response to remember her physical limitations. The point was well put by Utter J delivering the majority judgment of the Supreme Court of Washington in State of Washington v Wanrow Wash 559 P 2d 548 (1977) at p 558:

“In our society women suffer from a conspicuous lack of access to training in and the means of developing those skills necessary to effectively repel a male assailant without resorting to the use of deadly weapons.”

He went on to caution against leaving the jury with the impression that the objective standard to be applied is that applicable to an altercation between two men.

However, in dismissing the appeal the Court concluded that, in relation to the trial Judge’s explanation of self-defence:⁵⁰⁴

The examples the Judge gave of justified and unjustified self-defence may not have been entirely apposite, but nonetheless they did demonstrate the need for proportionality of response, which was certainly relevant in this case, bearing in mind the quantity of drugs used in comparison with the drugged coffee the morning before and, if it was accepted, [P’s] evidence about the eye-dropper.

Despite the Court of Appeal’s comments on matters of legal principle in Oakes, confirming the need to take into account the circumstances as the defendant believes them to be, including

495 At 683. For a discussion of the facts in Oakes, see also Seuffert, above n 427, at 319.

physical disparities and the history of abuse in the relationship when assessing whether the force used was reasonably proportionate, the application of those principles on the facts appears to have been grounded in what is now an outdated understanding of family violence, focused on the psychology of the defendant and battered woman syndrome. As we discuss below, the Court considered that the poisoning could not be seen as a reasonable or appropriate response “without the emotional fragility and the altered perception that are features of [battered woman] syndrome”.⁵⁰⁵The passage quoted at paragraph [6.60] above seems to focus on the imminent threat posed by the deceased, and what response – or how many pills – was in balance with that imminent threat, rather than viewing the threat faced as emanating from a cumulative pattern of harm.

With a proper understanding of all the circumstances, however, based on an understanding of family violence that reflects contemporary social science, we consider the requirement for reasonable proportionality is capable of accommodating victims of family violence.

The FVDRC argues that self-defence has been interpreted in a more restrictive manner in New

Zealand compared with similar jurisdictions.⁵⁰⁶

In Australia, imminence, lack of alternatives and proportionality are simply factors in the jury’s assessment of self-defence at common law.⁵⁰⁷Queensland is alone in requiring by statute that defensive action be taken in response to an assault.⁵⁰⁸However, even in that State, the courts have been sensitive to look past the question of imminence.⁵⁰⁹The Queensland Supreme Court in the case of R v Falls said:⁵¹⁰

[I]t doesn’t matter that at the moment [the defendant] shot Mr Falls in the head he didn’t at that moment offer or pose any threat to her. He had assaulted her. There was the threat that there would be another one and another one and another one after that until one day something terrible happened. It might have been the next day, it might have been the next week, but the risk of death or serious injury to her was ever present.

The only Australian jurisdiction that retains a direct reference to proportionality in its statute is South Australia. However, this requirement is qualified, consistent with the approach at common law, by the statement that:⁵¹¹

A requirement ... that the defendant's conduct be (objectively) reasonably proportionate to the threat that the defendant genuinely believed to exist does not imply that the force used by the defendant cannot exceed the force used against him or her.

The relatively flexible approach in the Australian common law has not prevented several states from pursuing law reform to clarify that these concepts do not operate as a barrier to successful claims by victims of family violence. As discussed in Chapter 4, the VLRC considered that legislative reform was desirable to encourage a more careful analysis by jurors of circumstances

507 Zecevic v DPP [1987] HCA 26; (1987) 162 CLR 645; Hopkins and Easteal, above n 394, at 132–133; Law Reform Commission of Western Australia, above n

508 Criminal Code 1899 (QLD), ss 271–272, discussed in Hopkins and Easteal, above n 394, at 135–136.

509 Elizabeth Sheehy, Julie Stubbs and Julia Tolmie “Defences to Homicide for Battered Women: A Comparative Analysis of Laws in Australia, Canada and New Zealand” [2012] SydLawRw 22; (2012) 34 Syd L Rev 467 at 471.

510 R v Falls Supreme Court of Queensland, 2–3 June 2010, cited in Sheehy, Stubbs and Tolmie, above n 509, at 471.

511 Criminal Law Consolidation Act 1935 (SA), s 15B. During the second reading of the 2003 Bill introducing this qualifying statement, the Hon P Holloway explained the amendment was not a new principle but rather a statement of the current principle in the codified law on self-defence. See (26 May 2003) SAPD LC 2393.

Understanding family violence: Reforming the criminal law relating to homicide 85

in which a person may reasonably believe his or her life is in danger, even where that person is not under immediate attack or at risk of immediate harm.⁵¹²It also recommended reform to discourage juries from placing undue emphasis on the proportionality of the response to the force used or threatened against the defendant in determining whether their actions were reasonable.⁵¹³Similar findings were made in Western Australia and Tasmania. We consider the detail of relevant reforms pursued in these jurisdictions in the following chapter.

In Canada, in the 1990 case of R v Lavallee, the Supreme Court similarly relaxed the requirement for imminent danger in respect of victims of family violence who kill their abusers.⁵¹⁴To require a victim of family violence to wait until the physical assault is “underway” before her apprehensions can be validated in law would, that Court considered, be tantamount to sentencing the defendant to “murder by instalment”.⁵¹⁵Given the context in which family violence occurs, the Court also considered the “mental state of an accused at the critical moment she pulls the trigger cannot be understood except in terms of the cumulative effect of months or years of brutality”.⁵¹⁶Subsequent statutory reforms to the law of self-defence in Canada, discussed in Chapter 4, codified the decision in Lavallee to make it clear that the jury must engage in a contextual analysis of the reasonableness of self-defence claims, of which imminence, lack of alternatives and proportionality are but relevant factors for consideration.⁵¹⁷

In England and Wales, self-defence requires that the defendant was acting in response to what he or she perceived as an “actual or imminent unlawful assault” and that the defendant’s actions were reasonable in the circumstances as he or she believed them to be.⁵¹⁸The degree of force is not to be regarded as having been reasonable in the circumstances if it was “disproportionate in those circumstances”, but the jury must take into consideration that a person cannot “weigh to a nicety the exact measure of his necessary defensive action”, and evidence of a person having only done what they honestly and instinctively thought was necessary constitutes strong evidence that only reasonable action was taken.⁵¹⁹The requirement for proportionality is therefore interpreted flexibly, consistent with the approach in New Zealand. In particular, the use of a weapon by a woman against an unarmed but violent man does not necessarily make their act unlawful, as the nature of the threat must be considered against the background circumstances.⁵²⁰It is recognised, however, that the continued requirement for imminence may exclude cases in which the defendant acted in genuine fear for their life.⁵²¹

We have discussed above the problems with applying the concepts of imminence, lack of alternatives and proportionality to victims of family violence. These problems are not new. In an effort to overcome these problems, courts in New Zealand and overseas have accepted evidence of battered woman syndrome.⁵²²This provides a framework within which the jury

514 R v Lavallee [1990] 1 SCR 852 at 876–877. See Campbell, above n 486, at 82–84. See also R v Petel [1994] 1 SCR 3 at [22].

517 Citizen’s Arrest and Self-defence Act SC 2012 c 9, s 34. See also Vanessa MacDonnell “The New Self-Defence Law: Progressive Development or Status Quo?” (2013) 92 Can Bar Rev 301 at 317.

519 Criminal Justice and Immigration Act 2008 (UK), s 76, codifying the Privy Council decision in Palmer v R, above n 491, at 832. See Law

520 The Law Commission of England and Wales, endorsing the submission by HHJ Goddard QC. See Law Commission of England and Wales, above n 415, at 76.

can understand the circumstances of women who have suffered long-term violence and enables appropriate assessment of the reasonableness of these defendants’ actions.⁵²³The FVDRC notes that, without such evidence, the fact a victim did not leave an abusive partner or seek help from Police every time they were assaulted or threatened is taken as evidence that the abuse was not as bad as claimed or that the victim chose to stay in the situation and was, therefore, partially responsible for it.⁵²⁴The role and relevance of expert evidence is discussed in detail in the next chapter. Below we identify the problems with battered woman syndrome evidence in the context of self-defence claims.

As we explained in Chapter 2, battered woman syndrome is based on the theory that ongoing intimate partner violence results in women experiencing learned helplessness – the belief that, whatever they do, victims cannot change their situation, so they make no attempts to break out of the cycle of violence.⁵²⁵

The admissibility of battered woman syndrome evidence in criminal trials is “beyond controversy” in New Zealand.⁵²⁶Its relevance to claims of self-defence is primarily to correct juror misconceptions they may have about family violence and its effect on victims.⁵²⁷In R v Zhou, an attempted murder case, the trial Judge explained battered woman syndrome was relevant to “rebut a suggestion that she could not have been battered because she would have stopped seeing him”.⁵²⁸In R v Oakes, discussed at paragraphs [6.57]–[6.61] above, the Court of Appeal confirmed its relevance to the defendant’s belief as to the nature and seriousness of the threat they faced, noting that a woman suffering from the syndrome “may genuinely perceive danger earlier than others would, and a threat of more serious harm than others might see”.⁵²⁹

The importance of expert evidence to explain the effects of battered woman syndrome was also underlined in Oakes, when the Court noted that “[experts] may provide an answer to questions which would naturally occur to the average juror”.⁵³⁰The Court cited the Canadian case of R v Lavallee, which also confirmed the relevance of expert evidence on battered woman syndrome to claims of self-defence. In that case Wilson J said:⁵³¹

The average member of the public (or of the jury) can be forgiven for asking: Why would a woman put up with this kind of treatment? Why should she continue to live with such a man? How could she love a partner who beat her to the point of requiring hospitalization? We would expect the woman to pack her bags and go. Where is her self-respect? Why does she not cut loose and make a new life for herself? Such is the reaction of the average person confronted with the so-called ‘battered wife syndrome’. We need help to understand it and help is available from trained professionals.

However, despite the Court of Appeal’s clear recognition of the relevance of expert evidence on battered woman syndrome in Oakes, the Court went on to find that there was no misdirection by the trial Judge in that case, even though there was only one brief reference to battered woman syndrome in the Judge’s summing up and only in relation to the partial defence of provocation.⁵³²The impact of this has been noted:⁵³³

531 R v Lavallee, above n 514, at 871–872; and R v Oakes, above n 454, at 679.

Understanding family violence: Reforming the criminal law relating to homicide 87

The crucial relevance of the entire context of abuse to Oakes's perception of the circumstances and the reasonableness of a pre-emptive strike in self-defence was not clarified for the jury by the Judge. It appears therefore that the jury may not have had any information on the reasonableness of a pre- emptive strike in the circumstances as Oakes believed them to be and the relevance of battered woman syndrome to those circumstances. Therefore it seems unlikely that the jury could have grasped the relevance of Oakes's and the expert's testimony as to Oakes's fear on the night of the killing and her possible perception, due to an ability to predict the abusive acts of Gardner, that he would kill her that night.

The Court of Appeal noted that, if the trial Judge had said more about battered woman’s syndrome and its bearing on this case, he could have found it necessary to underline that the syndrome could explain why the accused had deliberately killed the deceased, and thus “provide a motive for murder”.⁵³⁴This “remarkable statement” has been criticised as being apparently contrary to all the evidence given in the case, and in other cases, about the nature and effects of battered woman syndrome.⁵³⁵Commentators also argue that battered woman syndrome is often interpreted as explaining the defendant’s subjective state of mind, but not the state of mind of a reasonable person in her position.⁵³⁶In other words, evidence could be interpreted to explain that the defendant had an unreasonable but understandable reaction, rather than a normal or reasonable response in the circumstances. In Oakes, for example, an expert witness for the defence gave evidence that the defendant’s “ability to reason rationally would have been impaired” as a result of suffering from battered woman syndrome,⁵³⁷and the Court of Appeal said:⁵³⁸

Whether or not Mrs Oakes was suffering from the battered woman’s syndrome at the time ... was critical to self-defence ... for without the emotional fragility and the altered perception that are features of the syndrome the poisoning could not be seen as a reasonable or appropriate response to such a threat as the deceased actually posed.

For these reasons, along with the general criticisms of battered woman syndrome discussed in Chapter 2, many commentators argue that such evidence does not adequately inform judges and jurors of the relevant effects of abuse and can in fact undermine a defendant’s claim to self-defence.⁵³⁹Others, however consider that despite its limitations, battered woman syndrome evidence has been useful for educating juries about the perceptions that women in situations of prolonged family violence may develop.⁵⁴⁰

In its 2001 Report the Law Commission recommended that the term “battered woman syndrome”, or any use of the term “syndrome” in this context be dropped and that reference be made instead to the nature and dynamics of battering relationships and the effects of battering.⁵⁴¹
In 2004 the VLRC recommended a move away from battered woman syndrome evidence and towards the use of expert evidence that focuses on the social context of family violence.⁵⁴²

535 McDonald, above n 394, at 683; Seuffert, above n 427, at 322; and Jeremy Finn “Oakes – Case and Comment” (1995) 19 Crim LJ 291 at 293.

539 Crofts and Tyson, above n 394, at 881–882; Law Reform Commission of Western Australia, above n 427, at 287; McDonald, above n 394, at

683–684; Seuffert, above n 427, at 327; and Tarrant, above n 394, at 603–604.

540 Tasmania Law Reform Institute, above n 411, at 39; and Robertson, above n 425.

Almost without exception, submitters considered that a victim of family violence should be entitled to rely on self-defence even if the harm sought to be avoided was not imminent or the fatal force was not strictly proportionate to the threatened harm.

The FVDRC emphasised the need to assess the actions of victims of family violence in the context of the threat they were experiencing because of their ongoing relationship with a violent perpetrator, not just because of what was happening on any one particular occasion. That assessment must be informed by an understanding of family violence as a form of entrapment. The FVDRC considers legislative reform is necessary to address the decision in Wang. It also notes the decision in Richardson (discussed at paragraph [6.32] above), which it says has the potential to present a further legal barrier for a victim of family violence who is responding to an omnipresent threat that has the potential to crystallise at any point in time and that the Police are unable to defuse.

The Auckland Coalition for the Safety of Women and Children said that, all too often, the assessment of what has occurred when the defendant is responding to family violence is “heavily influenced by male gendered normative notions of the options available to women victims of domestic violence”. They note a number of factors that must be considered relevant to any consideration of self-defence in this context, such as whether the defendant had access to a telephone, whether she could speak English, whether she knew what agencies to contact for help and what help she received in the past if she did make contact, the size of the defendant relative to the deceased, whether the deceased had invaded the home of the defendant and whether the defendant had been raped by the deceased. The idea that the use of a weapon means the use of force was disproportionate is, the Coalition says, highly gendered when a woman is unable to defend herself effectively in any other way.

Members of the relevant committees of the New Zealand Law Society⁵⁴³were divided on this issue. Some noted a major concern with the concepts of imminence and proportionality and considered that the approach in Wang needs to be changed. They observed that imminence can act as a “roadblock” for defendants because it requires a defendant who has suffered cumulative abuse over a lengthy period of time to artificially nominate a single point of confrontation and wrongly assumes threats of violence are always avoidable and that a delayed threat will always allow for non-violent intervention. One member also noted that the concept of imminence can be particularly problematic for youth offenders and must be balanced against their ability to foresee consequences when they have themselves experienced abuse throughout their lives. Other members, however, were equally concerned that these conclusions were based on unconfirmed assumptions about jury behaviour and that, while self-defence is historical, and applies mostly to male violence, this did not necessarily lead to the conclusion that the existing law is inadequate for victims of family violence. We note, however, that our previous discussion focuses not on jury verdicts for which reasons are unknown but on authoritative statements of the law by the higher courts.

The Criminal Bar Association and the Public Defence Service also considered that section 48 should be amended to remove any requirement for imminence or proportionality where self- defence in the context of family violence is in issue. The Public Defence Service noted, however, that proportionality should not be quite as significant an impediment as imminence, because proportionality is assessed in the circumstances as the defendant believes them to be, which allows the defendant’s account of a history of physical disadvantage vis-à-vis the deceased to

Understanding family violence: Reforming the criminal law relating to homicide 89

be introduced. They also noted that proportionality ties in with the need to recognise that imminent harm is not a requirement for use of force to be reasonable, in light of the fact that, due to disparity in size and strength, a pre-emptive strike may be justified. An academic also noted that proportionality should not be interpreted to permit disproportionate force in every case but rather that the concept of proportionality ought to be interpreted in light of the context that force is used, including not only on the threat of harm but also the respective sizes of the parties and the history of abuse between the parties.

Some submitters noted that, while they agreed that imminence and proportionality should not be requirements, it should still be open to the court to reject a claim of self-defence because the harm was not imminent or proportionate where those questions are central to the assessment of reasonableness. The few submitters that took a contrary view expressed concern that the sanctity of life may be diminished or that undeserving defendants might exploit any change in the law.

The problems in applying the concepts of imminence, lack of alternatives and proportionality to an assessment of the reasonableness of force used by a victim of family violence are well known and widely accepted.

These concepts in the context of family violence are problematic because they assume continued association of self-defence with a one-off confrontation rather than an ongoing threat of harm. This association developed in the context of male violence and male standards of reasonableness. It jars with contemporary understandings of the nature and dynamics of family violence and, as one English commentator notes, means that self-defence is “skewed to the detriment of women since a defendant’s action is only considered ‘reasonable’ when killing is a proportionate response to an immediate threat of deadly force”.⁵⁴⁴

The theory that underpins self-defence, discussed in the previous chapter, is that the use of force is justified only if it is “necessary”. The jury assesses necessity by asking whether, in the circumstances as the defendant believed them to be, the force used was reasonable. The concern is that the concepts of imminence and proportionality are not always indicative of necessity in the context of ongoing family violence. They can focus on the immediate circumstances to the exclusion of the wider context, including the cumulative and compounding nature of family violence and the history of the relationship between the defendant and the deceased. In practice, therefore, these concepts can operate as a barrier or unattainable threshold even where a defendant genuinely believes himself or herself to be at risk of death or serious injury, and that the use of force is necessary to avert the risk.

This is not to say the concepts should be abandoned or that every victim of family violence who kills an abuser should be entitled to an acquittal. As the Law Commission said in 2001:⁵⁴⁵

In many, perhaps most, situations, the use of force will be reasonable only if the danger is imminent because the defendant will have an opportunity to avoid the danger or seek effective help. However, this is not invariably the case. In particular, it may not be the case where the defendant has been subject to ongoing physical abuse within a coercive intimate relationship and knows that further assaults are

544 Susan Edwards “Abolishing Provocation and Reframing Self Defence – The Law Commission’s Options for Reform” [2004] Crim L Rev 181 at

We therefore consider, consistent with the Commission’s view in 2001, that imminence and lack of alternatives should not be strictly applied where a victim of family violence claims self- defence and should not overshadow or substitute the jury’s assessment of whether the use of force was, in the circumstances as the defendant believed them to be, “reasonable”. We also consider that, when a victim of family violence kills an abuser, self-defence should not be excluded simply because the force used by the defendant is not strictly proportionate to the force used against the defendant in the immediate circumstances. The reasonableness of the force used must be assessed in light of all the circumstances, including physical disparities and the history of abuse between the parties.

An important principle of law reform is that non-legislative alternatives to achieving a policy objective are considered and unnecessary legislation avoided.⁵⁴⁶Legislative reform to clarify the correct interpretation of the law may, however, be preferred where it serves an important declaratory or educative function. The VLRC, for example, observed that there was no requirement for imminence or strict proportionality in Australian common law but nonetheless recommended that the substantive test for self-defence be clarified in statute in order to ensure jury directions dealt with the issues adequately and to encourage a more careful analysis by jurors of self-defence claims, particularly by victims of family violence.⁵⁴⁷

In New Zealand, in 2001, the Commission recognised that section 48 did not require the courts to exclude self-defence where a danger was inevitable but not imminent. However, the Commission considered it was preferable to make this explicit through legislative reform, rather than leave the law to be developed case by case, because:⁵⁴⁸

Relying on the courts to develop the law may require a person to be convicted and then to appeal successfully before the legal position is clarified. While the Court of Appeal would be free to change its earlier approach, a trial judge may feel he or she is required to follow the approach in Wang. Until the Court of Appeal had dealt with the matter, the correct interpretation of section 48 would remain unclear, although some trial judges may approach section 48 in terms of inevitability.

Since then, as outlined above, the Court of Appeal has not changed its approach to the requirement for imminence. Instead, it has continued to follow Wang, which the Supreme Court recently recognised as part of the settled law of self-defence.⁵⁴⁹In the absence of any Court of Appeal or Supreme Court decision to the contrary High Court judges would, therefore, be bound to follow Wang. Since 2001, however, an extensive body of literature has examined the operation of self-defence and maintains that the concept of imminence is problematic in the context of victims of family violence who kill, which has led to targeted reform in other jurisdictions.

For these reasons, we conclude the case for reform to address the problems with the concept of imminence when self-defence is claimed by victims of family violence is made out. We discuss the options for reform in the next chapter.

The requirement for “reasonable proportionality” between the force used and the perceived danger is, we consider, capable of being interpreted in a manner that takes into account the full circumstances of victims of family violence who use defensive force. The statements in Oakes

546 See the Legislation Advisory Committee Guidelines: Guidelines on Process and Content of Legislation (2014). This is also a requirement of the

Understanding family violence: Reforming the criminal law relating to homicide 91

and Afamasaga confirm that proportionality is assessed in the circumstances as the defendant believes them to be. We do not, therefore, consider there is a need to clarify in legislation the operation of this aspect of the test for self-defence. However, for this concept to operate fairly for victims of family violence, the “circumstances” considered must include not only the physical disparities between the defendant and their abuser but also the history of abuse between the parties. These circumstances may be explained through direct evidence from the defendant as well as expert evidence on the dynamics of family violence. We consider that the education of judges, lawyers and Police on the dynamics of family violence, recommended in Chapter 2, is also critical to ensuring the full circumstances are put before the jury.

In the preceding chapter, we concluded that legislative reform was necessary to address problems with the concept of imminence when self-defence is claimed by victims of family violence. We also observed the need for the jury, in considering a claim of self-defence by a victim of family violence, to have a full understanding of the dynamics of the violent relationship including the history of violence, the defendant’s prior responses to that violence and the effects of the violence on the defendant. Without a full appreciation of all the circumstances, a jury may find it difficult to believe the defendant’s account or to understand the defendant’s use of force as reasonable.

In this chapter, we consider the options for reform and make recommendations to:
^._{reform the substantive
law of self-defence to address the problems encountered with the}

_{concept of imminence by victims of family violence; and}
^._{support the proper application of self-defence where a defendant is
responding to family}

violence, in particular, by ensuring that relevant evidence is provided to the jury and the relevance of that evidence to the legal requirements of the defence is understood.

In making these proposals for reform, we have drawn on the experiences of comparable jurisdictions and on the substantial body of literature on this issue. We also refer to the Law Commission’s previous recommendations in the 2001 Report.

In this section, we consider the different options for substantive reform of self-defence and whether reform should be limited to victims of family violence and/or homicide offences.

In the Issues Paper, we set out three different options for substantive reform of self-defence:
^._{Option 1
– introduce a legislative provision clarifying that, under section 48,
the force used}

by the defendant may be reasonable even though the defendant is responding to harm that is not immediate or uses force in excess of that involved in the harm or threatened harm.
^._{Option 2 – amend section
48 to replace by statute the Wang concept of imminence with
that}

_{of “inevitability”.}
^._{Option 3
– introduce a new complete defence to extend the concept of
self-defence to apply}

to the specific circumstances in which victims of family violence kill their abusers out of necessity.

The first option we put forward in the Issues Paper was to clarify the operation of section 48 in a new provision in the Crimes Act 1961. The intended effect of such a provision would be to reverse the presumption in R v Wang that a victim of family violence who kills their

Understanding family violence: Reforming the criminal law relating to homicide 93

abusive partner is not acting in self-defence unless the threat is capable of being carried out immediately⁵⁵⁰and to avoid undue emphasis being placed on the proportionality of the level of force used by the defendant against the force used or threatened by the deceased. It would not alter the substantive requirements of self-defence in section 48, only clarify how those requirements should be applied. In particular, the concepts of imminence and proportionality would remain relevant in assessing whether the force used was reasonable but would no longer act as a barrier or threshold for victims of family violence to successfully relying on self-defence.

Similar reforms have been pursued in Victoria and Western Australia, and the operation of those reforms is discussed below.

In response to recommendations of the Victorian Law Reform Commission (VLRC) in its

2004 Report Defences to Homicide, the Victorian Crimes Act 1958 was amended to include the following provision, now found in section 322M(1) of that Act:⁵⁵¹

Without limiting section 322K [self-defence], for the purposes of an offence in circumstances where self-defence in the context of family violence is in issue, a person may believe that the person’s conduct is necessary in self-defence, and the conduct may be a reasonable response in the circumstances as the person perceives them, even if—

the person is responding to a harm that is not immediate; or

the response involves the use of force in excess of the force involved in the harm or threatened harm.

Bench Notes issued by the Judicial College of Victoria explain that this provision “clarified existing law”, namely that:⁵⁵²

... a person is not required to wait until an attack is in progress or immediately threatened before using defensive force. S/he is entitled to take steps to forestall a threatened attack before it has begun (Osland v R [1998] HCA 75; (1998) 197 CLR 316). Similarly, the force used is not required to be precisely proportionate, as long as the accused believed it was necessary (Zecevic v Director of Public Prosecutions [1987] HCA 26; (1987) 162 CLR 645) and the conduct was a reasonable response in the circumstances.

The operation of the self-defence reforms were reviewed by the Victorian Department of Justice in 2010⁵⁵³and again in 2013.⁵⁵⁴In 2010, the Department identified that the self-defence reforms had been relevant to two decisions to discontinue proceedings where there was clear evidence that the defendant had killed a family member in response to ongoing family violence.⁵⁵⁵It concluded, therefore, that the reforms had “introduced significant improvements to the criminal

551 This provision was originally enacted in 2005 as s 9AH of the Crimes Act 1958 (Vic).

552 Judicial College of Victoria Bench Notes: Statutory Self-Defence (Judicial College of Victoria, June 2015) at [75].

553 Victorian Department of Justice Defensive homicide: Review of the offence of defensive homicide: Discussion paper (August 2010).

554 Victoria Department of Justice Defensive Homicide: Proposals for Legislative Reform: Consultation Paper (September 2013) at vii–viii.

555 The first case involved a teenage girl (SB) who killed her stepfather after four years of “relentless sexual abuse”. She was threatened with a shotgun and forced to perform a sexual act. In fear for her life, she followed his demands, and immediately after, she picked up the shotgun and shot her stepfather in the back of the head. Rather than proceed to trial, the Director of Public Prosecutions decided to discontinue the prosecution on the basis that no jury would find the defendant guilty given the circumstances and the volume of evidence supporting her testimony, and also taking into account the new self-defence provisions. The second case involved a 57-year-old woman (Freda Dimitrovski) who killed her husband in response to an immediate, violent attack. Dimitrovski had been physically and psychologically abused by her husband for over 30 years. During the attack, he hit her in the face and knocked her to the ground in the presence of her daughter and four-year-old grandson. When he attempted to attack Dimitrovski’s daughter, she stabbed him with a pocket knife. Following a three-day committal hearing the presiding Magistrate concluded that the evidence “overwhelmingly” supported a history of abuse and discharged Ms Dimitrovski on the basis that she was not satisfied there was sufficient evidence for a jury to convict. See Victorian Department of Justice, above n 553, at 30–32.

justice system in dealing with situations in which a woman kills in response to long-term violence”.⁵⁵⁶

Some commentators, however, caution that celebration of the success of the Victorian self- defence reforms might be premature.⁵⁵⁷Both cases in which proceedings had been discontinued involved “traditional notions of self-defence” and, in particular, had concerned responses to an immediate threat.⁵⁵⁸In 2013, the Department of Justice noted that the self-defence reforms had still not been tested in the context of a non-immediate threat. How the reforms might operate in that context, the Department considered, was “exceptionally difficult to answer in the abstract”,⁵⁵⁹but that:⁵⁶⁰

What is essential for these purposes is that it is well within the operation of the laws of self-defence that a woman could be acquitted on the basis of self-defence in this situation. Whether that is the case as a matter of fact would need to be determined by a jury on a case by case basis.

Since then, it appears self-defence has been successfully relied on in at least one case where a victim of family violence (in that case, the male de facto partner of the deceased) killed an abusive partner in the context of a non-immediate threat.⁵⁶¹

We note that the full impact of the self-defence reforms in Victoria may still be unknown, given the operation of defensive homicide in that jurisdiction until 2014. Defensive homicide, discussed in Chapter 4, was noted by the Department of Justice as having the unintended effect of “shift[ing] the focus of debate from the adequacy of complete self-defence to defensive homicide”.⁵⁶²

Self-defence in Western Australia was codified following a review of homicide by the Law Reform Commission of Western Australia.⁵⁶³Adopting the Commission’s recommendations to clarify the role of imminence, the statutory provision on self-defence in that jurisdiction now requires that:⁵⁶⁴

the person believes the act is necessary to defend the person or another person from a harmful act, including a harmful act that is not imminent; and

the person’s harmful act is a reasonable response by the person in the circumstances as the person believes them to be; and

_{there
are reasonable grounds for those beliefs.}

557 Kellie Toole “Self-Defence and the Reasonable Woman: Equality before the New Victorian Law” [2012] MelbULawRw 7; (2012) 36 MULR 250 at 270. See also Danielle Tyson and others “The Effects of the 2005 Reforms on Legal Responses to Women Who Kill Intimate Partners” in Kate Fitz-Gibbon and Arie Freiberg (eds) Homicide Law Reform in Victoria: Retrospect and Prospects (The Federation Press, Leichhardt, 2015) 79.

561 DPP v Bracken [2014] VSC 94. The Hon Justice Marcia Neave, who was Chairperson of the Victorian Law Reform Commission at the time Defences to Homicide was published (Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 77), observes that the jury “must have accepted [Mr Bracken] was the victim of family violence”. She also observes that the expert evidence given in that case, given by a forensic psychiatrist, was likely to have been very important. See Marcia Neave “The More Things Change the More They Stay the Same: Homicide Law Reform in Victoria” in Kate Fitz-Gibbon and Arie Freiberg (eds) Homicide Law Reform in Victoria: Retrospect and Prospects (The Federation Press, Leichhardt, 2015) 9 at 11.

563 Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007).

Understanding family violence: Reforming the criminal law relating to homicide 95

The general operation of this provision was considered by the Supreme Court of Western

Australia in the case of Goodwyn v State.⁵⁶⁵Buss JA explained the law as follows:⁵⁶⁶

So, it is apparent that s 248(4) enumerates four elements. First, the accused (subjectively) believes the harmful act is necessary to defend the accused or another person from a harmful act, including a harmful act that is not imminent (s 248(4)(a)). Secondly, the accused's harmful act is a reasonable (objective) response by the accused in the circumstances as the accused (subjectively) believes them to be (s 248(4)(b)). Thirdly, there are reasonable (objective) grounds for the accused's (subjective) belief that the harmful act is necessary to defend the accused or another person from a harmful act, including a harmful act that is not imminent (s 248(4)(a) read with s 248(4)(c)). Fourthly, there are reasonable (objective) grounds for the accused's (subjective) belief as to the circumstances (s 248(4)(b) read with s 248(4)(c)).

The first and second elements are similar to the subjective and objective requirements in section

48 of the New Zealand Crimes Act, but the third and fourth elements (derived from subsection

(c) of the Western Australian self-defence provision) are not applicable in New Zealand.⁵⁶⁷

In relation to imminence, Mazza JA in Goodwyn noted:⁵⁶⁸

The law as it previously stood assumed that self-defence would only arise in a case of immediate threat which was required to be met with an immediate response. The current provisions contemplate a person acting in respect of a harmful act which is not imminent.

The jury direction given by the trial Judge in Goodwyn on the relevance of imminence to the first element (not in issue on appeal) was as follows:⁵⁶⁹

The law says that people are not required to wait until they are injured or killed, or other persons are injured or killed, before defending themselves or that person. They are entitled to use force before that situation arises. As I've said, the threat need not be imminent. It is sufficient if the person believes that his act is necessary to defend himself from such a threatened assault.

Accordingly, as one commentator has observed, the Western Australian self-defence reforms “[put] it beyond doubt that response[s] to on-going domestic violence can come within the defence whether or not the accused responded at the time of a physical attack”.⁵⁷⁰

Unlike the reforms pursued in Victoria, the Western Australian self-defence provision is not limited to where family violence is in issue⁵⁷¹and does not clarify the requirement for proportionality.⁵⁷²

A further issue is whether Option 1, if recommended, should be limited to where a defendant is responding to family violence. Some argue that any clarification of self-defence should have general effect, as in Western Australia, because it may be equally relevant in non-family violence situations. It may be relevant, for example, where a defendant is held hostage and

570 Stella Tarrant “Self Defence In The Western Australian Criminal Code: Two Proposals for Reform” (2015) 38 UWAL Review 1 at 7.

571 The extended definition of self-defence has been relied on outside the context of family violence in Western Australia. See for example Goodwyn v The State of Western Australia, above n 565.

572 Despite a recommendation from the Law Reform Commission of Western Australia to clarify that a response may be reasonable even though it is not proportionate. See Law Reform Commission of Western Australia, above n 563, at 172.

uses pre-emptive force.⁵⁷³We recently published our Report Strangulation: The case for a new offence, which recommends a separate offence of strangulation to address the problems with prosecuting strangulation offenders in the family violence context. We concluded, however, that the new offence should be of general application, as the problems identified are likely to apply equally to other contexts, and the requirement to prove the presence of family violence circumstances may impede prosecution.⁵⁷⁴We considered that limiting a new offence to family violence circumstances would risk anomalous and inconsistent treatment of different classes of offenders whose conduct may be equally culpable.⁵⁷⁵

In this context, however, any perceived widening of self-defence may lead to greater use of the defence by those prone to violence.⁵⁷⁶This was the experience in Victoria in relation to defensive homicide, which, as we explained in Chapter 4, was used primarily by violent men rather than the victims of family violence for whom it was intended. We note, however, that this risk cannot be entirely avoided even if Option 1 were limited to family violence. For example, a predominant aggressor could potentially claim they killed a primary victim in self-defence and seek to rely on Option 1, particularly if there is evidence of a history of violence on both sides.⁵⁷⁷

Further, if Option 1 has general application, there is a risk that achievement of the policy objective of the reform – to accommodate the experiences of victims of family violence who kill in self-defence – could be undermined, as it would not be apparent on the face of the provision.

The second option is to amend section 48 itself to expressly replace the Wang concept of imminence with a new concept of “inevitability”.

This option arguably sits closer to the Law Commission’s recommendation in its 2001 Report, which was to “make it clear that there can be fact situations in which the use of force is reasonable where the danger is not imminent but is inevitable”.⁵⁷⁸That view was endorsed by the VLRC and the draft provision it put forward in its Report, Defences to Homicide, provided as follows:⁵⁷⁹

a person may believe that the conduct is necessary; and

_{the
person’s response may be reasonable}

when the person believes that the harm to which he or she responds is inevitable, whether or not it is immediate.

This approach, however, was not adopted in Victoria and has been criticised for several reasons.⁵⁸⁰First, it is argued that the inclusion of a legal requirement of “inevitability” would extend the current judicially determined parameters of self-defence, which might have the

573 Victorian Law Reform Commission, above n 561, at 80; Nicola Wake “‘His home is his castle And mine is a cage’: a new partial defence for primary victims who kill” (2015) 66 NILQ 149 at 158; and Elisabeth McDonald “Defending Abused Women: Beginning A Critique of New Zealand Criminal Law” (1997) 27 VUWLR 673 at 690.

574 Law Commission Strangulation: The Case for a New Offence (NZLC R138, 2016) at 39–40.

576 Angelica Guz and Marilyn McMahon “Is Imminence Still Necessary? Current Approaches to Imminence in the Laws Governing Self-Defence in Australia” (2011) 13 Flinders LJ 79 at 101.

577 This issue is discussed further in Chapter 10 in relation to partial defences.

578 Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001) at 12.

580 AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington, 2012) at 528; Guz and McMahon, above n 576, at 103–104; Mark Campbell “Pre-Emptive Self-Defence: When and Why” (2011) 11 OUCLJ 79 at 93–95; and Kevin Dawkins and Margaret Briggs “Battered Defendants and Criminal Defences” [2001] NZ L Rev 345.

Understanding family violence: Reforming the criminal law relating to homicide 97

unintended consequence of allowing cases of uncertain merit to succeed.⁵⁸¹For example, considering the dangerous nature of prison environments or gang culture (both being contexts in which self-defence has been recently claimed – but rejected – in New Zealand),⁵⁸²changing the test from imminence to inevitability might allow prisoners and gang members to justify pre-emptive strikes on the basis of inevitable future harm from other prisoners or rival gang members.⁵⁸³

Second, it is argued that it would be more difficult for a jury to assess whether a danger is “inevitable” compared with “imminent”.⁵⁸⁴Imminence is a measure of proximity, whereas inevitability is a measure of “moral certainty” and is almost impossible to establish (because the future is unknowable), so exactly how “inevitable” must a threat be to justify self-defence?⁵⁸⁵
An inevitability standard may necessarily involve speculation and dangerously raise the level of error when predicting inevitable violence at some future but unspecified point in time.⁵⁸⁶A judge may need to give detailed directions to a jury to ensure they are aware of the difference between imminence and inevitability, and these issues may result in a further body of appellate case law to clarify how “inevitable” the threat must be.⁵⁸⁷These issues were considered by the VLRC when it recommended the provision set out at paragraph [7.25] above.⁵⁸⁸It concluded, however:⁵⁸⁹

The question of inevitability under this formulation focuses on the subjective belief of the accused. There is no requirement that inevitability be established in an objective sense, nor is it implied that by establishing that the accused believes the threat was inevitable the objective test of reasonableness will be satisfied. Rather, the provision aims to ensure that the actions of an accused person are not automatically excluded from the scope of the defence on the basis that the harm threatened is not immediate. Ultimately the question of whether the accused acted reasonably in self-defence will remain one for the jury.

The third option we explored in the Issues Paper is a new, complete defence that applies only where a defendant is responding to family violence and focuses on the underlying principle of self-defence – necessity. A separate provision would exist alongside the “traditional” self- defence provision in section 48 and would be tailored to take account of the nature of the danger that victims of family violence face.

The Law Commission in 2001 considered but rejected a separate defence for victims of family violence, considering it was preferable that the general requirement of reasonableness in section
48 be interpreted so that it can incorporate the use of defensive force against violence that may not be imminent.⁵⁹⁰

581 Dawkins and Briggs, above n 580, at 348–350; Guz and McMahon, above n 576, at 104; and Campbell, above n 580, at 95.

584 Dawkins and Briggs, above n 580, at 348–349; Simester and Brookbanks, above n 580, at 528; and Guz and McMahon, above n 576, at 103.

587 Dawkins and Briggs, above n 580, at 350; and Guz and McMahon, above n 576, at 103. Dawkins and Briggs also argue, at 348, that inevitability potentially blurs the justificatory underpinnings of self-defence and “edges closer to excusing the accused’s behaviour in acting out of a sense of fear or terror as an understandable human frailty”. However, as self-defence can apply where the defendant is operating under a genuine but mistaken belief, it is arguable that this distinction between justification and excuse in section 48 is already “blurred”. See Fairburn v R [2010] NZCA 44 at [39].

We are not aware of any comparative jurisdictions adopting a separate, complete defence for victims of family violence. Queensland is the only jurisdiction to have a specific defence for victims of family violence, but that only operates as a partial defence to murder.⁵⁹¹A new complete defence was proposed, but not adopted, in Western Australia by the Taskforce on Gender Violence, established by the then Chief Justice the Hon David Malcolm AC QC.⁵⁹²The Taskforce proposed a new complete defence formulated as follows:⁵⁹³

Conduct is carried out by a person in self-defence if the person is responding to a history of personal violence against herself or himself or another person and the person believes that the conduct was necessary to defend himself or herself or that other person against the violence.

Law reform bodies that have addressed this issue have generally considered it preferable if existing general defences are made capable of accommodating the experience of family violence victims rather than introducing separate defences for this specific category of defendant.⁵⁹⁴The risks of creating a separate defence were identified by the Australian and New South Wales Law Reform Commissions in their joint report on family violence in 2010:⁵⁹⁵

The Commissions consider that criminal defences should not recognise the circumstances of family violence victims in an ‘atypical context’, or typecast the reactions of family violence victims who kill as the product of ‘extraordinary psychology.’ There is substantial force in stakeholders’ arguments that separate, family-violence specific defences may result in the differential treatment of persons who have killed in response to family violence, compared with those who have killed in response to non- familial violence. To this end, it is preferable for family-violence related circumstances to be integrated into existing defences of general application. In the Commissions’ view, existing defences–in particular self-defence–are doctrinally capable of accommodating the diverse situational and psychological circumstances of family violence victims.

The Commissions concluded that any problems associated with the practical application of general defences should be addressed by improving the existing defences in the family violence context, through legislative clarification and guidance where necessary, and professional education and training for lawyers and the judiciary.⁵⁹⁶

Most submitters preferred Option 1 over Options 2 or 3. Reasons included that it was the least risky option for reform and that it struck the right balance in addressing any risk of undue emphasis being placed on the concepts of imminence and proportionality, whilst leaving them as relevant factors to be considered by the courts.

The Family Violence Death Review Committee (FVDRC), the Criminal Bar Association, the Public Defence Service and most members of the relevant committees of the New Zealand Law Society⁵⁹⁷favoured Option 1. The FVDRC submitted that Option 1 should also include a requirement that the threat be assessed in the context of the victim’s ongoing relationship with a violent perpetrator, not just what was happening on any one occasion. Another submitter noted that section 48 is a clear, principled statement of the law and thus any reform should be carefully drafted as a clarification rather than an extension.

592 Taskforce on Gender Bias Report of the Chief Justice’s Taskforce on Gender Bias (1994).

595 Australian Law Reform Commission and New South Wales Law Reform Commission Family Violence — A National Legal Response (ALRC R114 and NSWLRC R128, October 2010) at 649–650.

Understanding family violence: Reforming the criminal law relating to homicide 99

Submitters were divided on whether Option 1 should apply generally or only where a defendant is responding to family violence. Reasons for limiting Option 1 to family violence included the distinctive features of intimate partner violence compared with other forms of interpersonal violence and the risk of unmeritorious arguments in respect of other forms of interpersonal violence. The New Zealand Law Society also noted that applying the proposed reform more widely would diminish the weight that should be given to victims of family violence in trying to change perceptions and understandings of the dynamics of family violence. However, committee members of the Law Society differed as to how Option 1 should be limited. Some considered it should be limited to victims of family violence, while others believed that fairness requires its availability to any defendant who is subjected to duress to a similar degree or of the same nature outside a familial relationship, for example, non-familial elder abuse.

Those who submitted that Option 1 should have general effect included the Auckland District Law Society and the Criminal Bar Association, who cited the need to protect other equally vulnerable defendants. The Criminal Bar Association considered that the risk of it being used by defendants who, in the overall interests of justice, ought not be able to could be addressed by a requirement that the judge make a finding, on the basis of expert reports, that imminence and proportionality could be dispensed with. They said that, in effect, this would require the judge to determine that the defendant had been affected by a history of family violence. Several other submitters, including the Public Defence Service and the Auckland Crown Solicitor’s Office, noted the need for further research and consultation if Option 1 were to have general effect.

Some submitters preferred Option 2. However, the FVDRC noted that the concept of inevitability in the hands of those who do not understand entrapment could actually raise the standards that primary victims are expected to meet if they are to successfully establish self- defence. One submitter was concerned that Option 2 could narrow the current statutory test of reasonableness and restrict its development, while another submitter was concerned it could be interpreted too widely. The Public Defence Service was also concerned that Option 2 carries the potential to blur the justificatory underpinnings of self-defence, noting that any amendment that sees section 48 becoming “excuse based” should be avoided.

Submitters in favour of Option 3 included the Auckland Crown Solicitor’s Office, which considered Option 3 to be less nuanced and potentially problematic than the other options. The Auckland Coalition for the Safety of Women and Children also preferred Option 3, noting that different treatment is needed in this context because the circumstances in which victims of family violence kill are quite different from homicide committed in other contexts. The Public Defence Service considered, however, that self-defence can sufficiently accommodate the situational and psychological circumstances of family violence. One submitter preferred a separate defence that applies more broadly to victims of family violence who commit other offences when acting under coercion. Another submitter proposed a new defence drawing on elements from duress, necessity and self-defence.

The Commission considers that Option 1, clarifying the operation of section 48 in a new provision in the Crimes Act, is the best mechanism for achieving the policy objective of reform, which is ensuring that self-defence properly accommodates the experiences of victims of family violence who commit homicide.

Because Option 1 does not expressly introduce any new concepts, it has the advantage over Options 2 and 3 of minimising the risks of unintended consequences and of introducing complexities of interpretation that could result in a further body of case law. In relation to Option 2, while we agree that self-defence should not be excluded in circumstances where the

danger is inevitable but not imminent, we consider the objective of this option, and of the Commission’s recommendation in the 2001 Report, can be achieved through Option 1 with a lower risk of adverse consequences. Given the problem arises not with the wording of section
48 but with the way in which the section has been applied, in our view, it is undesirable to amend section 48 itself.

We also prefer Option 1 to Option 3 because, while Option 3 has the advantage of enabling the law to be drafted with the particular circumstances of victims of family violence in mind, we are concerned about the risk of perpetuating different, unequal treatment of this group of defendants. We agree with the Law Commissions of Australia, New South Wales, Victoria and Western Australia that, where possible, the application of existing homicide defences should be improved to ensure substantive equality in the law, rather than introducing new defences.⁵⁹⁸We consider that self-defence can and should accommodate the diverse situational and psychological circumstances of family violence victims.

In Chapter 6, we concluded that Wang remains authority for the need for an “imminent or immediate threat [with] no alternative available”.⁵⁹⁹In relation to the concept of proportionality, however, we concluded that the law of self-defence is capable of accommodating the experiences of victims of family violence who use a level of force during an immediate confrontation that may not be strictly proportionate to the force used or threatened against them, especially when expert evidence is adduced.⁶⁰⁰We are concerned, therefore, that clarifying the operation of section 48 in respect of proportionality would do little to change the substantive approach. We are also aware of the risk that such a clarification could invite confusion, and could risk a suggestion that unreasonable, “excessive” force could be considered legitimate self-defence.

Where the issues of imminence and proportionality are interrelated (as they were in Wang), we consider that removing a requirement for imminence will necessarily require the jury to look beyond the immediate circumstances and undertake a broader contextual analysis of whether the use of force was reasonably proportionate, by reference not to the single preceding event but in anticipation of a future of repeated violence.

We therefore recommend that the Crimes Act be amended only to overrule the requirement in Wang for an imminent threat. This would bring the interpretation and application of self-defence in the context of family violence in line with the comparable jurisdictions of Australia (at common law⁶⁰¹and in the statutory self-defence provisions of Victoria and Western Australia) and Canada.⁶⁰²Under such a provision, imminence would remain a relevant consideration in the jury’s assessment of self-defence claims, as it remains so in Victoria and Western Australia.⁶⁰³Self-defence would, however, in the words of the Supreme Court of Western Australia, “contemplate a person acting in respect of a harmful act which is not imminent”.⁶⁰⁴

We have considered the view of the Criminal Law Reform Committee, responsible for drafting what is now section 48, that the substantive provision on self-defence should not include a list of evidentiary guidelines for the courts, as that would be “unwise and unhelpful in relation to

598 See Victorian Law Reform Commission, above n 561, at 68; and Law Reform Commission of Western Australia, above n 563, at 289.

602 See Chapter 6 for a discussion of the operation of self-defence in those jurisdictions.

Understanding family violence: Reforming the criminal law relating to homicide 101

self-defence, where the question is one of fact to be decided in the light of an infinite variety of circumstances in different cases”.⁶⁰⁵While we agree with that approach in principle, we are satisfied that the problems identified with the concept of imminence in the previous chapter, and the recent statements of approval of Wang by the Supreme Court in Vincent⁶⁰⁶and the Court of Appeal in Afamasaga⁶⁰⁷justify an exception to that approach to overrule Wang.

We therefore recommend that the Crimes Act be amended to include a provision clarifying that section 48 may apply even where a person is responding to a threat that is not imminent. We do not make recommendations as to the precise wording of such a provision. This would be an appropriate task for the Parliamentary Counsel Office in conjunction with the Ministry of Justice. Such a provision must, however, be carefully drafted so that the threat need not be imminent in order for either the subjective or objective elements of self-defence to be satisfied, that is, a defendant may believe that the use of force was necessary and the use of force may be reasonable where a threat is not imminent. The provisions in Victoria and Western Australia may be useful examples, but we note that the tests for self-defence in those jurisdictions differ from section 48.

While the concept of imminence may also be problematic for defendants other than victims of family violence, given the limitations of our terms of reference and the time constraints on this project, we have not been able to consider properly the position of other defendants. Without this additional work, we are concerned about the risk of unintended consequences of recommending reform with general application. These risks are significantly higher than the risks identified in recommending an offence of strangulation with general application⁶⁰⁸and, in our view, require a narrower approach. We are also persuaded of the merit in explicitly identifying the context of family violence in promoting changes in perceptions and improvements in understandings of the dynamics of family violence.

For these reasons, we recommend that the proposed reform be limited to the context of family violence. We acknowledge that this cannot entirely exclude the risk of predominant aggressors or other “unmeritorious” defendants claiming that they acted in self-defence when they used force against an intimate partner or other family member. We think, however, the risk of such defendants establishing a probative evidentiary foundation for self-defence in these circumstances is low, given the continued requirement for objective reasonableness.⁶⁰⁹We also note that, in such circumstances, the trial judge would be able to exercise his or her discretion and either:
^._{withhold the question of self-defence from the
jury, if satisfied that no jury could entertain a}

_{reasonable doubt on the issue (discussed below at paragraphs
[7.100]–[7.103]; or}
^._{rule any evidence proposed
to be adduced as irrelevant to the question of self-defence and}

inadmissible, pursuant to the general rules of admissibility in the Evidence Act 2006.

Because the proposed reform is intended only to clarify the application of section 48 in the specific context of family violence, we do not consider there is a need for any further threshold or procedure for defendants seeking to rely on the proposed reform.

605 Criminal Law Reform Committee Report on Self Defence (Report 15, November 1979) at 8.

609 A similar finding was made by the Law Reform Institute of Tasmania when endorsing the Victorian model provision in s 322M of the Crimes

Act 1958 (Vic). See Tasmania Law Reform Institute Review of the Law Relating to Self-defence (Final Report No 20, 2015) at 67.

A further issue that is raised when limiting the proposed reform to family violence is what we mean by “family violence”. We discussed our use of this term in Chapter 2. The Crimes Act does not include a definition of family violence. The Domestic Violence Act 1995, however, defines domestic violence as “violence against that person by any other person with whom that person is, or has been, in a domestic relationship”.⁶¹⁰“Violence” encompasses physical, sexual and psychological abuse, including intimidation, harassment, threats and financial or economic abuse.⁶¹¹A “domestic relationship” includes relationships between spouses, partners, family members, others who are ordinarily members of the same household and close personal friends.⁶¹²That definition is also used in the Evidence Act 2006.

In Victoria, section 322J of the Crimes Act 1958 (Vic) defines family violence in relation to a person to include “violence against that person by a family member”. The definitions of “violence” and “family member” are similar to the definitions in the New Zealand Domestic Violence Act, although the definition of family member is inclusive rather than exhaustive.

The FVDRC considers that the existing definition of domestic violence is problematic because it does not accord with contemporary understanding of family violence as a cumulative pattern of harm.⁶¹³It argues the existing definition supports an incident-based response to domestic violence rather than a consideration of each person’s role in the intimate relationship abuse history.⁶¹⁴The FVDRC also identifies that an unintended consequence of the current definition is that defensive behaviour by primary victims can be misconstrued as acts of perpetration.⁶¹⁵

The Ministry of Justice is currently reviewing family violence legislation. Included within this review is a consideration of the legal definition of domestic violence. The Ministry has sought views on whether it is up to date and whether principles could more clearly guide how the law is implemented.⁶¹⁶

The Commission considers it is appropriate that any definition of family violence for the purposes of the proposed reform be consistent with the definition of domestic violence in the Domestic Violence Act, incorporating any amendments that may be recommended by the Ministry of Justice as part of its current review. We note, however, that there may be merit in defining family violence in an inclusive rather than exhaustive fashion for the purposes of the proposed reform (by stating that family violence includes rather than has the same meaning as domestic violence in the Domestic Violence Act).

Finally, we have considered whether the proposed reform should be limited to homicide. Our reference is limited to victims of family violence who kill. However, self-defence is a general defence that can also be claimed in relation to offences such as attempted murder and assault.

Limiting the proposed reform to murder and manslaughter would give rise to anomalies. A

clear example would be a victim of family violence who is charged with attempted murder. This

611 Domestic Violence Act 1995, s 3(2). Section 3(5) also confirms that a single violent act may amount to abuse and that a number of acts that form a pattern of behaviour may amount to abuse for that purpose, even though some or all of those acts, when viewed in isolation, may appear to be minor or trivial

616 Ministry of Justice Strengthening New Zealand’s legislative response to family violence: A public discussion document (Wellington, August 2015) at

Understanding family violence: Reforming the criminal law relating to homicide 103

issue was identified by the VLRC when it recommended similar reforms limited to homicide in Victoria. The VLRC recommended serious consideration be given to their application to offences more generally.⁶¹⁷In 2013, the Victorian Department of Justice considered this issue and determined that the application of the VLRC’s recommendations should not depend simply upon whether a person dies or not as a result of the defendant’s conduct.⁶¹⁸

On balance, we consider that the proposed reform should not be limited to murder or manslaughter charges. Unlike the issue of whether the proposed reform should be limited to family violence, where there clearly may be unintended consequences of general application, here, the risk of unintended consequences arises if we do not have general application in terms of the offence charged. Limiting the proposed reform to homicide would create anomalies in the law depending on whether the defendant’s use of force was lethal. We therefore recommend the proposed reform not be limited to homicide offences.

If the requirement for imminence is removed from self-defence, we consider the defence is, in substance, capable of accommodating the experiences of victims of family violence, typically women, who kill their abusers. However, problems will remain in the application of the law if the jury does not adequately understand the general dynamics of family violence and the defendant’s experiences specifically.⁶¹⁹As the Australian and New South Wales Law Reform Commissions noted in 2010, “a focus on the doctrinal content of defences is insufficient to ensure that the experiences of family violence victims who kill are accommodated in practice”.⁶²⁰

In Chapter 2, we explained that misunderstandings and misconceptions of family violence persist today and recommended further education on the dynamics of family violence for those working within the criminal justice system. In this section, we consider proposals to ensure that a jury, considering a claim of self-defence by a victim of family violence, understands the full circumstances of the defendant’s actions, including the dynamics of their relationship with the deceased, the history of violence in that relationship and its effects on the defendant.

619 Julia Tolmie “Defending Battered Defendants on Homicide Charges in New Zealand: The Impact of Abolishing the Partial Defences to Murder” [2015] NZ L Rev 649 at 681; Toole, above n 557, at 252; and Anthony Hopkins and Patricia Easteal “Walking in Her Shoes: Battered women who kill in Victoria, Western Australia and Queensland” (2010) 35 Alt LJ 132 at 134.

620 Australian Law Reform Commission and New South Wales Law Reform Commission, above n 595, at 651; and Kate Fitz-Gibbon and Julie Stubbs “Divergent directions in reforming legal responses to lethal violence” (2012) 45 Australian & New Zealand Journal of Criminology 318 at 331.

In a criminal trial, the context and circumstances of the alleged offending are presented to the jury through evidence introduced by the prosecution and the defence. This includes evidence of facts and, in some cases, expert opinion evidence. In self-defence claims, evidence will assist the jury to assess the nature of the threat faced, the defendant’s state of mind and whether his or her actions were reasonable in the circumstances.

Where a victim of family violence claims self-defence, evidence of prior family violence will be critical to the jury’s understanding of the circumstances. The VLRC explained the relevance of this evidence to the elements of self-defence as follows:⁶²¹

For jurors, the application of force or use of a weapon by those who are subjected to abuse, particularly in non-confrontational circumstances, may raise issues about the reasonableness of the accused’s belief in the need to use fatal force. Jurors may believe there were other options available to the accused to escape the violence, or that the use of a weapon was out of proportion to the nature of the threat. Jurors may also have questions about the honesty of the accused’s belief in the need to use force to defend himself or herself because of the apparently planned nature of her actions. The broader context of prior violence will often be critical to the jury’s evaluation of whether the accused acted in self- defence.

Even where evidence of prior family violence is admitted, however, persisting myths and misunderstandings may mean that the significance of prior family violence may be missed or that jurors, relying on their own limited knowledge, may make unjustified behaviour assumptions or rely on illegitimate reasoning when assessing that evidence. For this reason, many acknowledge the importance of evidence on the general dynamics of family violence given by experts on this topic. As some commentators note:⁶²²

Juries will commonly need assistance, for example, in understanding why leaving may not resolve domestic abuse, why failed attempts to get help might make future help-seeking by the victim more difficult or dangerous, why there may be few independent witnesses to corroborate the accused’s account, why the accused may have made allegations of violence in the past and recanted, and why her story may take some time to fully emerge. In fact it may only emerge as a truthful and complete account after considerable time has passed and perhaps even then only within a relationship of trust. Expert testimony will be particularly significant in understanding past retaliatory violence by the accused.

Evidence of fact is generally admissible under the Evidence Act 2006 where it is relevant to a fact in issue.⁶²³Expert opinion evidence is admissible if the fact-finder “is likely to obtain substantial help from the opinion in understanding other evidence in the proceeding or in ascertaining any fact that is of consequence to the determination of the proceeding”.⁶²⁴

An expert is someone “who has specialised knowledge or skill based on training, study, or experience”.⁶²⁵The judge must determine whether the expert witness is properly qualified to

622 Elizabeth Sheehy, Julie Stubbs and Julia Tolmie “Securing Fair Outcomes for Battered Women Charged with Homicide: Analysing Defence

623 Evidence Act 2006, s 7. Subsection (3) states that evidence is relevant “if it has a tendency to prove or disprove anything that is of consequence to the determination of the proceeding”. See also Mathew Downs (ed) Cross on Evidence (online looseleaf ed, LexisNexis) at [EVA7.1]; and Richard Mahoney and others The Evidence Act 2006: Act and Analysis (2nd ed, Thomson Reuters, Wellington, 2010) at 39–47. Evidence may be excluded, however, if its probative value is outweighed by the risk that the evidence will have an unfairly prejudicial effect on the proceeding, or needlessly prolong the proceeding: Evidence Act 2006, s 8. Certain exceptions also apply to hearsay evidence (Evidence Act 2006, ss 16-22) and prior consistent statements (Evidence Act 2006, s 35). We are not aware of any problems with the operation of these exceptions in this context.

624 Evidence Act 2006, s 25. See also Downs, above n 623, at [EVA25.1]–[EVA25.13]; and Mahoney and others, above n 623, at 95–103.

Understanding family violence: Reforming the criminal law relating to homicide 105

testify. Parties may also agree on a statement of expert opinion rather than calling witnesses to give evidence.

Evidence of battered woman syndrome has been accepted in New Zealand courts to “disabuse jurors of misconceptions they may have about the nature of domestic violence and its effect on women exposed to it”.⁶²⁶Battered woman syndrome evidence is regarded as “syndrome evidence”, usually given by qualified forensic psychologists and forensic psychiatrists.⁶²⁷

Cases involving victims of family violence who kill their abusers are rare, and we have not identified any legal difficulties with admitting evidence of prior family violence to support claims of self-defence under the Evidence Act. However, in the previous chapter, we identified the broader concern that the focus on the immediate circumstances of the alleged offending, consistent with the stereotypical self-defence scenario of a one-off violent confrontation with a stranger, may mean juries are less likely to hear evidence of prior family violence (as it is not seen as relevant to the claim of self-defence). Furthermore, where such evidence is introduced, it may be for the limited purpose of understanding the circumstances of the immediate event and may not be sufficient to enable a jury to gain a proper understanding of all the circumstances of the killing.⁶²⁸

We also discussed in the previous chapter the problems with battered woman syndrome, in particular, that it promotes a rigid, limited view of women’s experiences and behaviour that overemphasises their psychological reactions. In light of these problems, commentators and law reform bodies have recommended redefining the scope of expert evidence on family violence to place a greater emphasis on the broader social context of a defendant’s situation and to reflect the current state of knowledge about the nature and dynamics of family violence and its effects.⁶²⁹It is argued that expert evidence on family violence in self-defence claims can perform a wide range of useful functions, including:⁶³⁰
^._{educating jurors on the phenomenon
of family violence;}
^._{addressing common
misconceptions;}
^._{bridging the gap between the legal
requirements of self-defence and the defendant’s evidence,}

for example, by explaining why the defendant’s actions could be objectively reasonable, given their perception of danger or their belief that no lesser measure could have protected him or her; and
^._{supporting the defendant’s credibility, by explaining his or
her state of mind and normalising}

aspects of their behaviour or demeanour that may be troubling for the jury.⁶³¹

We also note that funding constraints and availability of suitably qualified experts may impact detrimentally on the use of expert evidence in these cases. Most homicide defendants rely on legal aid funding. Under the legal aid framework, lawyers must apply for prior approval of funding for an expert witness. Approval will only be granted if the lawyer can demonstrate why the expert’s attendance is required, how it will contribute to a successful outcome for their client

628 Stella Tarrant “Something is Pushing Them to the Side of Their Own Lives: A Feminist Critique of Law and Laws” (1990) 20 UWAL Rev 573 at 599.

629 Victorian Law Reform Commission, above n 561, at 172–173; and Tarrant, above n 570, at 19–21.

631 For example, assisting the jury in determining whether the defendant did in fact fear for her life by explaining the heightened sensitivity of a primary victim to the predominant aggressor’s acts: R v Lavallee [1990] 1 SCR 852 at 882.

and confirm that any potentially less expensive sources of evidence have been considered.⁶³²We also understand that there is a small pool of expertise on family violence in New Zealand, with perhaps as few as six or eight people properly qualified as experts in this field.

Finally, we note the observations of the Hon Justice Stevens of the Court of Appeal in 2015 about inconsistent reliance on expert evidence generally (albeit by prosecutors, not defence counsel) across New Zealand:⁶³³

The practice of Crown solicitors and prosecutors, in determining whether to seek to adduce such evidence, is not uniform across New Zealand. This is a matter of discretion for individual prosecutors and regional variations have emerged. My enquiries suggest in some regions, such as Rotorua, it is considered that juries are well-equipped to deal with complex family dynamics and the concomitant issues that arise in these sexual assault cases. In Tauranga, however ... counter-intuitive evidence is seen by the Crown as serving a useful educative function.

A broader understanding by jurors of what it must be like for a victim of abuse to live in a situation of ongoing and serious violence is crucial to the further development of self-defence. Without a proper appreciation of the circumstances of the accused, including the nature of the threat he or she faced, and other personal circumstances, juries are unlikely to be able to make an informed assessment of whether the accused acted in self-defence.

While the VLRC recognised family violence evidence was generally accepted by the courts, it recommended providing specific guidance in legislation:⁶³⁵

... the importance of this evidence in supporting a plea of self-defence has persuaded us that its status should be clarified in legislation. This will avoid any unnecessary arguments concerning its relevance and ensure the range of factors which may be necessary to represent the reality of the accused’s situation are readily identified.

The Crimes Act 1958 (Vic) was subsequently amended to include a provision that, in circumstances where self-defence in the context of family violence is in issue, “evidence of family violence” may be relevant to determine whether the defendant carried out conduct while believing it to be necessary in self-defence, or whether their conduct was a reasonable response in the circumstances as the defendant believed them to be.⁶³⁶

632 Legal Services Commissioner Legal Aid Disbursement Policy: Criminal (Ministry of Justice, 31 March 2014) at 13.

633 Justice Stevens “Counter-Intuitive Evidence in Sexual Assault Cases: Current Issues and Future Solutions” (paper presented to the Triennial

Understanding family violence: Reforming the criminal law relating to homicide 107

The legislation includes an extensive definition of what is meant by “evidence of family violence”, which provides:⁶³⁷

_{Evidence of family violence, in relation to a person, includes evidence
of any of the following—}

the history of the relationship between the person and a family member, including violence by the family member towards the person or by the person towards the family member or by the family member or the person in relation to any other family member;

the cumulative effect, including psychological effect, on the person or a family member of that violence;

social, cultural or economic factors that impact on the person or a family member who has been affected by family violence;

the general nature and dynamics of relationships affected by family violence, including the possible consequences of separation from the abuser;

the psychological effect of violence on people who are or have been in a relationship affected by family violence;

social or economic factors that impact on people who are or have been in a relationship affected by family violence.

Subsections (a) to (c) relate to evidence of prior family violence. Subsections (d) to (f) recognise the relevance of expert social context evidence. The VLRC endorsed the view that the people best qualified to give expert evidence on family violence are likely to include those with direct experience of working with victims of family violence and with knowledge of current research in the field.⁶³⁸

An independent review of the operation of the Victorian provisions in 2015 identified an increased awareness by legal professionals of the relevance of family violence in homicide cases and noted that the provisions made it more likely that family violence will be considered and linked to the defendant’s actions and the available defences.⁶³⁹

The Victorian provisions were endorsed by the Australian and New South Wales Law Reform Commissions in 2010, recommending that the criminal legislation in each state and territory of Australia provide guidance about the potential relevance of family violence-related evidence in the context of a defence to homicide, “given its importance in these circumstances”.⁶⁴⁰The Law Reform Commission of Western Australia, the New South Wales Select Committee on Provocation and the Tasmania Law Reform Institute also endorse specific family violence evidence provisions, often citing the important educative function it would serve for the legal profession and the broader community.⁶⁴¹In 2010 Queensland introduced a provision to clarify that evidence of the history of the domestic relationship between the defendant and the person against whom the offence is committed is admissible in relation to certain offences.⁶⁴²

637 Crimes Act 1958 (Vic), s 322J(1). Subsection (2) defines various terms, and subsection (3) confirms that a single act of violence, and a number of acts forming part of a pattern of behaviour, may amount to abuse for the purposes of the definition in subsection (2).

639 Mandy McKenzie, Debbie Kirkwood and Danielle Tyson “‘Unreasonable’ self-defence?” (2013) 2 DVRCV Advocate 12 at 15; and Tyson and others, above n 557, at 92. The authors concluded, however, that there was limited use of the Victorian evidence provision and that the potential of the reforms to challenge gender-based stereotypes around self-defence has not yet been fully realised. They call for improvement through comprehensive, consistent and ongoing training for prosecution and defence counsel, judges, expert witnesses and other legal professionals.

640 Australian Law Reform Commission and New South Wales Law Reform Commission, above n 595, at 652–654.

641 Law Reform Commission of Western Australia, above n 563, at 293; Government of New South Wales Government response to the report of the Legislative Council Select Committee on the Partial Defence of Provocation (2013) at 185–186; and Tasmania Law Reform Institute, above n 609, at 63–64.

In the Issues Paper, we asked submitters whether they considered there was a problem with admitting family violence evidence under the Evidence Act 2006 and whether a specific evidence provision similar to the Victorian provision should be introduced in New Zealand.

Submitters generally agreed that evidence of prior family violence should be admissible in cases involving claims of self-defence. However, they did not identify any particular problems with the current operation of the Evidence Act. The Law Society considered there may be problems around the “relevance” requirement in section 7, but the current rules were generally flexible enough to deal effectively with family violence evidence. One academic noted, however, that the test of “substantial helpfulness” for expert evidence is a high threshold.

Around half of submitters were in favour of introducing a family violence evidence provision. Some emphasised the need to move on from an approach focused on the defendant’s psychology and battered woman syndrome. Others, however, including the New Zealand Law Society and the Public Defence Service, noted that education and guidance around the myths and misconceptions of family violence could mitigate the need for such a provision.

Some submitters also noted individuals with field experience and an understanding of the contemporary social science of family violence should be considered “qualified” to give expert opinion evidence, even if they do not possess formal qualifications as a psychologist or psychiatrist. One expert we spoke with, however, noted that those with field experience may not be best placed to give this evidence if they appear to be advocates and that medical professionals may be preferred for their perceived neutrality.

We also asked submitters whether they favoured a family violence evidence provision being included in the Crimes Act or the Evidence Act. Most submitters considered the Evidence Act was the logical place for an evidential provision. The minority of submitters who preferred an amendment to the Crimes Act argued it would have an important educative function, and it would be more likely lawyers would be appraised of the provisions in the context in which they were relevant.

... is particularly difficult to convey in the criminal justice context because it spans a period of time (often lengthy), has a cumulative impact on those who survive it that affects how they see and respond to the world, is a pattern of behaviour rather than an event or events, is hidden, has been culturally minimised, and is more complex than an account of the physical incidents of violence that have taken place would suggest.

Rules of evidence do not determine the elements of self-defence, but they affect how a case is prepared and a defendant’s story is heard at trial.⁶⁴⁴In this context, where victims of family violence usually rely on self-defence in the “non-stereotypical” scenario of an ongoing threat posed by an intimate partner and where myths and misconceptions around family violence persist, we consider there is substantial value in a family violence evidence provision to encourage a proper assessment of self-defence claims at trial.

A statutory amendment based on the Victorian evidence provision would draw attention to the evidence that would normally be considered relevant in order to fully represent the defendant’s

Understanding family violence: Reforming the criminal law relating to homicide 109

circumstances to the jury. The trial judge would still have a role, under the Evidence Act, to determine the admissibility of such evidence in individual cases. We also expect that clarifying the relevance of expert evidence would lead to greater reliance on experts at trial. Such a provision would operate alongside Recommendation 5 above, which recommends a new provision clarifying that section 48 may apply even if a person is responding to a threat that is not imminent, to provide a comprehensive package of reform focused on improving the way in which the experiences of victims of family violence are accommodated in the application of self-defence.

We have considered whether such a provision should be included in the Crimes Act or the Evidence Act. We recognise that there might be some merit in including it in the Crimes Act, if, as some submitters suggested, it may have greater visibility in that Act and thus better serve its educative purpose. It is not unusual for evidential issues specific to certain offences or defences to be addressed in that Act.⁶⁴⁵However, evidential issues should normally be addressed in the Evidence Act, and on balance, our preference is that this provision be included in that Act.

We endorse the view that people with experience in the field, as well as those with formal qualifications such as psychiatrists and psychologists, might qualify as “experts”. We consider the Evidence Act is capable of providing such recognition,⁶⁴⁶and therefore do not make any further specific recommendations in this regard.

Consistent with our views on amendments to self-defence at paragraph [7.57] above, we do not recommend that the family violence evidence provision should be limited to homicide offences. We also refer to our discussion at paragraphs [7.50]–[7.54] above as to the appropriate definition of “family violence”.

^R7The Evidence Act 2006 should be amended to include provisions based on sections 322J and

In addition to the introduction of expert evidence, another way to correct any misconceptions and assumptions about family violence is through the judge’s directions to the jury. In jury trials, it is the role of the judge to direct the jury on the relevant law and facts in issue. Jury directions should be tailored to the particular facts of the case and questions in issue.⁶⁴⁷Trial judges are guided by higher court decisions on appropriate directions and by the guidance set out in the Criminal Jury Trial Bench Book, maintained by the Institute of Judicial Studies.⁶⁴⁸

645 See, for example, s 75 of the Crimes Act 1961 that relates to evidence of treason, which is an offence under s 73 of that Act. See also s 112 relating to evidence of perjury, false oath or false statement.

646 The provisions in the Evidence Act 2006 dealing with expert evidence in criminal proceedings are sections 4 (definition of expert) and 25 (admissibility of expert opinion evidence). See also Downs, above n 623, at [EVA25.1]–[EVA25.13]; and Mahoney and others, above n 623, at

647 Simpson v R [2010] NZCA 140 at [100], citing R v Tukaki CA360/05, 14 June 2006 at [12].

648 The Bench Book is a reference guide, based on Court of Appeal decisions, for High Court judges issuing directions to the jury. Unlike the Crown Court Bench Book in the United Kingdom, it is not publicly available. See Christopher Pitchford Crown Court Bench Book - Directing the Jury (Judicial Studies Board of England and Wales, 2010).

Some jury directions are prescribed by legislation. The Evidence Act includes jury directions about unreliable evidence, offering evidence in different ways, warnings about giving undue weight to evidence that a defendant lied and warnings of special caution in relying on identification evidence.⁶⁴⁹

Legislated “standard” jury directions, intended to correct erroneous beliefs juries may hold, are rare. There is only one example in the Evidence Act in the context of sexual offending. Section
127 provides:

Subsection (2) applies if, in a sexual case tried before a jury, evidence is given or a question is asked or a comment is made that tends to suggest that the person against whom the offence is alleged to have been committed either delayed making or failed to make a complaint in respect of the offence.

If this subsection applies, the Judge may tell the jury that there can be good reasons for the victim of an offence of that kind to delay making or fail to make a complaint in respect of the offence.

The advantage of legislating for a specific jury direction to address juror misconceptions is that it can fill a “gap” where expert evidence is not introduced. However, if it is too removed from the facts of the case, it may be unlikely to have any positive effect. Research suggests, in the context of sexual offending, jury directions aimed at addressing stereotypical and biased expectations of witness behaviour may be more effective in preventing preconceptions if given at the beginning of the trial.⁶⁵⁰Directions given later, during the judge’s summing up of the case, may have little or no effect on verdicts.⁶⁵¹

The Supreme Court recently considered the different methods for addressing juror misconceptions in sexual offending cases in DH v R.⁶⁵²It observed:⁶⁵³

We do not think it is appropriate to be prescriptive about how erroneous beliefs or assumptions are best to be countered in criminal trials. Judicial directions, s 9 statements and expert evidence are all possibilities. We do, however, consider that a cautious approach needs to be taken to the ambit of expert evidence given at trials of this kind to ensure that such evidence is confined to what would be substantially helpful, there is focus on live issues and that the evidence is not unduly lengthy or repetitive and is expressed in terms that address assumptions and intuitive beliefs that may be held by jurors and may arise in the context of the trial.

Jury directions of the type used in England and Wales on topics where there is a general acceptance of the topic are a worthwhile alternative to expert evidence. If all the areas that would otherwise be covered by expert evidence are amenable to jury direction, this would obviate the need for the evidence and it would no longer be substantially helpful. If not, the jury directions could reduce the scope of the evidence to topics not covered in the directions.

Accordingly, jury directions may be appropriate to counter juror misconceptions where there is a general acceptance of the topic and little controversy. If a topic is amenable to judicial direction, then that may reduce the need for expert evidence.

In Victoria, VLRC in its Report Defences to Homicide recognised that while those with expertise in family violence are best placed to address misconceptions, the trial judge has an important

650 Elisabeth McDonald and Yvette Tinsley “Evidence Issues” in Elisabeth McDonald and Yvette Tinsley (eds) From “Real Rape” to Real Justice: Prosecuting Rape in New Zealand (Victoria University Press, Wellington, 2011) 371 at 372.

Understanding family violence: Reforming the criminal law relating to homicide 111

role in assisting the jury to recognise the significance of prior violence and to make the connections between expert evidence and the issues at trial.⁶⁵⁴Where expert evidence is not led, the judge’s directions to the jury take on even greater significance. However, the VLRC did not favour legislating to require a set jury direction to be delivered when a history of family violence is raised:⁶⁵⁵

The Commission accepts that a ‘one size fits all’ approach to jury directions will not allow sufficiently flexibility. Moreover, we think that a standard charge suffers from the fundamental difficulty of the trial judge intruding into territory which belongs exclusively to the jury. But it is in many cases vital, if the trial is to be fair, that relevant matters be brought to the jury’s attention. In our view, this should be the role of social framework evidence, and of the experts who are appropriately qualified to give it. The trial judge will play an important role in highlighting the relevance of a history of abuse, and of the social framework evidence, to the particular facts in issue in the case.

Subsequently, however, in 2014, the Victorian Government introduced a standard jury direction aimed at juror misconceptions around family violence, observing that many members of the community do not fully understand the dynamics of family violence and that jury directions can play an important role in addressing juror misconceptions.⁶⁵⁶

The Jury Directions Act 2015 (Vic) provides that, where requested by defence counsel or the defendant, the trial judge must give a direction on certain matters unless there are good reasons for not doing so.⁶⁵⁷The direction may be given at any time during the trial.⁶⁵⁸Where relevant, the judge must inform the jury that self-defence is in issue and that evidence of family violence may be relevant to determining whether the defendant acted in self-defence.⁶⁵⁹The judge may also include any of the following matters in the direction:

is not limited to physical abuse and may include sexual abuse and psychological abuse;

may involve intimidation, harassment and threats of abuse;

may consist of a single act;

may consist of separate acts that form part of a pattern of behaviour which can amount to abuse even though some or all of those acts may, when viewed in isolation, appear to be minor or trivial;

people may react differently to family violence and there is no typical, proper or normal response to family violence;

_{it is not uncommon for
a person who has been subjected to family violence—}

to stay with an abusive partner after the onset of family violence, or to leave and then return to the partner;

_{not to report family violence
to police or seek assistance to stop family
violence;}

decisions made by a person subjected to family violence about how to address, respond to or avoid family violence may be influenced by—

that, as a matter of law, evidence that the accused assaulted the victim on a previous occasion does not mean that the accused could not have been acting in self-defence or under duress (as the case requires) in relation to the offence charged.

In the Issues Paper, we sought views on the desirability of a statutory jury direction and asked what matters it should address. Those submitters in favour included the Public Defence Service, who thought there would be little utility in adopting Recommendation 5 above without an accompanying jury direction, and the FVDRC, who noted the importance of jury directions to address the misunderstandings surrounding the cumulative and compounding effect of family violence and victim entrapment. Members of the relevant committees of the New Zealand Law Society⁶⁶⁰noted, however, a lack of evidence that acquittal or conviction rates are a result of misconceptions and were concerned that a “one size fits all approach” would be misguided. Experts in jury research we spoke with stressed the importance of any jury directions being tailored and flexible and noted that judicial education could lead to appropriate jury directions in any event. One submitter recommended that any supporting legislative reforms be deferred until the operation of any changes to section 48 can be reviewed.

In the Commission’s view, it is preferable that juror misconceptions be addressed by expert evidence rather than through the judge’s direction to the jury. While we note there may be an issue with expert availability, given the relative infrequency of these cases, we think that the issue is best addressed by encouraging greater use of expert evidence. This is the intended effect of Recommendation 7 above which provides a clear statement of the type of evidence that may be relevant to claims of self-defence.

In relation to jury directions on the operation of self-defence if Recommendation 5 is adopted, we are of the view that appropriate jury directions will naturally follow. An express jury direction is, therefore, unnecessary.

A final issue in relation to self-defence is the role of a trial judge to withhold claims of self- defence from the jury. In R v Wang, the Court of Appeal confirmed that a trial judge may withhold self-defence from the jury if he or she considers no jury could properly regard the defendant’s use of force as reasonable.⁶⁶¹This has been confirmed in subsequent cases.⁶⁶²

The Law Commission in its 2001 Report considered that the determination of what is reasonable force in self-defence calls for the application of community values, and contrary to Wang, that issue should always be left to the jury.⁶⁶³The Commission observed that a judicial filter would remain because a trial judge would still have to decide whether, on the evidence, there was a reasonable possibility that the defendant intended to act defensively.⁶⁶⁴

661 R v Wang, above n 550. This case is discussed in detail in Chapter 6 at paragraphs [6.24]–[6.28].

Understanding family violence: Reforming the criminal law relating to homicide 113

This issue was considered by the Court of Appeal in R v Bridger.⁶⁶⁵In that case, the appellant argued, consistent with the Commission’s recommendation in 2001, that the question of the reasonableness of the force used in the perceived circumstances should always be left as a matter for the jury.⁶⁶⁶However, the Court considered that a distinction between ruling upon whether the accused might possibly have been acting defensively and ruling upon whether, in so acting, the accused used no more than reasonable force was:⁶⁶⁷

... more than a little artificial since the question of whether the accused acted purely defensively in the perceived circumstances is surely to be assessed in part by reference to what the accused did in response to those circumstances ... Although the reasonableness of the response is a separate question, it is so closely connected with whether the response was defensively motivated that we can see no sensible reason for allowing the Judge to make an assessment of the first two matters but not the third in deciding whether self-defence should be left to the jury.⁶⁶⁸

An assessment of reasonableness involves the application of the community’s sense of justice. Community views on what is reasonable can, and do, change over time. We consider that only in exceptional cases should the issue of self-defence be withheld from the jury if there is evidence of a reasonable possibility that the defendant intended to act in self-defence. Vincent may be regarded as such a case.⁶⁶⁹However, we do not consider there is a need to recommend a change to the general operation of section 48 in this regard. We note the Court of Appeal’s comments in Bridger, as quoted in paragraph [7.102] above. Further, as we apprehend it, the problem identified in the Commission’s 2001 Report was the Court’s decision to withhold self- defence in the case of Wang on the basis that an imminent threat was not made out on the evidence. While we agree self-defence should have gone to the jury in that case, the effect of Wang is most appropriately mitigated by legislative reform to remove the requirement for imminence rather than by limiting judicial discretion to withhold claims of self-defence from the jury. We therefore make no recommendations in this regard.

668 The three matters referred to in this quote are the three elements of self-defence discussed in Chapter 5: (a) What were the circumstances as the accused honestly believed them to be? (b) In those circumstances, was the accused acting in the defence of himself or another? (c) Was the force used reasonable against the circumstances as the accused believed them to be?

669 In this case, the Court of Appeal upheld the trial Judge’s decision to withhold self-defence from the jury, and the Supreme Court dismissed the appellant’s leave to appeal. The defendant was a prison inmate who stabbed another inmate four times in the neck. The attack followed an incident on the exercise yard four days earlier where it was alleged the victim deliberately kicked a basketball towards Mr Vincent. At trial, the appellant claimed he was acting pre-emptively in self-defence in response to a threat of future violence from the victim. The decisions of the trial Judge, Court of Appeal and Supreme Court are discussed in Chapter 6. See Vincent v R, above n 582; and Vincent v R [2015] NZCA 201.

In Chapter 3, we introduced some aspects of the law of homicide and the key features of the 24

New Zealand cases we have identified in which victims of family violence have been prosecuted for killing their abusers since 2001. In Part 2 we examined how the law of self-defence applies to these defendants. Self-defence is a complete defence to a charge of murder or manslaughter. A person who successfully defends a homicide charge on grounds of self-defence will be acquitted.

In this part, we put self-defence to one side and turn to consider cases where a homicide cannot be justified but the defendant can point to circumstances that reduce their culpability. Building on our introductory discussion of homicide law in Chapter 3, we consider how the law deals with people who commit homicide in extenuating or mitigating circumstances:
^._{In this chapter, we consider the range of ways the law can provide
for recognition of reduced}

_{culpability for homicide.}
^._{In Chapter 9, we
build on the summary of our case review in Chapter 3 and look in more}

depth at recent New Zealand cases of homicide by victims of family violence. We consider whether the cases reveal any problems or gaps in this area of the law.
^._{In Chapters 10 and 11, we explore whether a partial
defence for victims of family violence is}

justified and whether New Zealand’s sentencing law is fit for purpose for these defendants.

There are two principal means by which reduced culpability may be recognised in homicide cases:
^._{The charge: instead of a charge
of murder, a person may be charged with a lesser offence}

such as manslaughter or infanticide. These are less serious types of homicide. Even if a person is charged with and tried for murder, a jury may find them guilty of manslaughter or, in appropriate cases, infanticide. This could be the outcome for a number of reasons, as we discuss in Chapter 9.
^._{The sentence: in some countries, murder attracts a mandatory
life sentence. In others, a}

person convicted of an unlawful killing, including murder, may be sentenced in a way that reflects their individual level of culpability within the offence category, whether murder, manslaughter or some lesser offence. New Zealand’s murder sentencing rules fall somewhere between mandatory and discretionary. The mandatory life sentence was abolished in 2002, but there remains a strong presumption in favour of life imprisonment and prescriptive rules for minimum periods of imprisonment where a life sentence is imposed. Furthermore, in

2010, the so-called three strikes legislation introduced outright mandatory sentencing for certain repeat violent offenders, including murder and manslaughter.⁶⁷⁰

As we noted in the Issues Paper, over the past 15 years there has been significant law reform activity in New Zealand and overseas that has addressed victims of family violence who commit homicide and the role of charging, partial defences and sentencing in recognising reduced culpability. It is apparent from this body of work that there is a range of ways reduced culpability can be taken into account. A great deal turns on the circumstances of and prevailing attitudes in individual jurisdictions and their wider homicide law.⁶⁷¹

To illustrate how charging and sentencing decisions may enable recognition of reduced culpability for homicide, it is useful to consider two New Zealand cases – those of Tikiahi Erstich and Jacqueline Wihongi.⁶⁷²

Erstich and Wihongi are not the only cases we could have selected to contrast the different means by which the law may recognise reduced culpability for homicide, but they stand out. Having committed homicide when he was just 14 years old, Tikiahi Erstich is the youngest defendant in our case review. Ms Wihongi’s case, as noted in Chapter 3 and discussed in the following chapters, is significant in large part because the Court of Appeal considered the circumstances in which the presumption of life imprisonment for murder may be displaced for offenders who kill after severe and prolonged abuse.

We noted the case of 14 year old Tikiahi Erstich in Chapter 2.⁶⁷³In 2001, after he had been subjected by his father to a 10-year “reign of terror” that included beatings with pipes and sticks, having his head battered against hard surfaces and being thrown against walls, Tikiahi Erstich shot and killed his tormentor at point-blank range. He was charged with murder and, at trial, relied on the partial defence of provocation,⁶⁷⁴which the jury apparently accepted because it returned a verdict of manslaughter. He was ultimately sentenced to a two-year suspended term of imprisonment, imposed by the Court of Appeal following a Crown appeal against sentence.⁶⁷⁵

We summarised the facts and disposition of R v Wihongi in Chapter 3,⁶⁷⁶and so we reiterate only immediately salient points here.

In 2009, Jacqueline Wihongi fatally stabbed her partner. She was in her early 30s and had suffered years of sexual and physical violence, including gang rape, at the hands of the deceased and others. She was charged with murder and pleaded not guilty. She did not rely on self- defence or the partial defence of provocation, and the jury convicted her of murder. She was

670 Sentencing and Parole Reform Act 2010. We consider the implications of the three strikes law for victims of family violence who commit homicide in Chapter 11.

674 The partial defence of provocation was available until December 2009 when it was repealed by the Crimes (Provocation Repeal) Amendment

675 Tikiahi Erstich was initially sentenced in the High Court to a sentence of two years’ supervision with special conditions: R v Erstich, above n

Understanding family violence: Reforming the criminal law relating to homicide 117

ultimately sentenced by the Court of Appeal to a 12-year term of imprisonment. She was spared life imprisonment because the Court of Appeal found that sentence would be manifestly unjust in the circumstances.

These cases share two important features. The first is that it is probable both defendants were found to have killed with murderous intent.⁶⁷⁷The second is that, in both cases, it was recognised that the defendants’ culpability was, in differing ways, reduced because of the violence they had suffered before the killings.

The outcomes in these cases demonstrate how partial defences and sentencing can reflect reduced culpability for homicide. In Erstich, the partial defence of provocation likely meant the defendant was convicted of manslaughter, rather than murder, and sentenced on that basis, while Ms Wihongi, who did not rely on a partial defence, was convicted of murder. Despite this difference, both offenders received sentences that took account of their personal circumstances. Tikiahi Erstich’s sentence reflected the violence he had suffered, his youth and rehabilitative prospects and his low risk of future offending.⁶⁷⁸Ms Wihongi was sentenced as a murderer but also in recognition of her personal circumstances. In combination with her cognitive deficits, Ms Wihongi’s history of abuse justified displacement of the presumption of life imprisonment,⁶⁷⁹although her level of risk together with the need to deter and denounce the taking of a life were found to require a 12-year finite term.⁶⁸⁰

New Zealand law no longer includes the partial defence of provocation or any other relevant partial defence,⁶⁸¹and so if Tikiahi Erstich was tried today, he could not defend the murder charge on that basis. A key question in this part is whether there is a need for a partial defence for cases of this kind or whether reduced culpability can – and should – be recognised principally through sentencing, as it was in Ms Wihongi’s case.

There is a range of options for recognising reduced culpability during the criminal justice process in homicide cases.

Some argue a murder-manslaughter bifurcation is, in itself, problematic, and it would be preferable to divide murder by degrees to better distinguish grades of blameworthiness for homicide.⁶⁸²In 2006, the Law Commission of England and Wales concluded that the offences of murder and manslaughter in that jurisdiction had been strained to accommodate “changing and deepening understandings of the nature and degree of criminal fault and the emergence of new

677 For Ms Wihongi, this follows from the conviction for murder, while in Tikiahi Erstich’s case, the sentencing Judge acknowledged the jury’s verdict “may have reflected acceptance of lack of intent to murder” but observed it “was more likely on the facts of the case to have entailed the jury’s acceptance of the partial defence of provocation”: R v Erstich, above n 672, at [2].

681 While infanticide and killing pursuant to a suicide pact (Crimes Act 1961, ss 178 and 180) are available as pathways to lesser convictions in certain narrow groups of cases that would otherwise meet the criteria for murder, there are no general partial defences in New Zealand. For the purposes of this Report, we largely disregard infanticide and killing pursuant to a suicide pact. As Warren Brookbanks has noted, they are closely circumscribed (infanticide) and rarely engaged (killing pursuant to a suicide pact): Warren Brookbanks “Partial Defences to Murder in New Zealand” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 271 at 271. They are also not relevant in cases where victims of family violence kill abusers. We refer to infanticide insofar as it may be some precedent for a separate homicide offence for victims of family violence but not otherwise.

682 See for example Brenda Midson “Degrees of blameworthiness in culpable homicide” [2015] NZLJ 220 at 234. See also Jeremy Horder Homicide and the Politics of Law Reform (Oxford University Press, Oxford, 2012) at 92.

partial defences” and recommended that murder be split into first and second degree offences.⁶⁸³

The Commission considered that questions about who should be labelled a “murderer” had become identified with questions about punishment.⁶⁸⁴This recommendation was not taken up, but degrees of murder are used in Canada⁶⁸⁵and in American law⁶⁸⁶and have previously been proposed in New Zealand.⁶⁸⁷

The ambit of New Zealand’s homicide offences and the merit of degrees of murder is beyond the scope of this reference. By their nature, these issues go wider than victims of family violence who commit homicide. We therefore do not explore them in this Report, although we touch on associated questions as we review options for distinguishing degrees of culpability, through partial defences and sentencing, in Chapters 10 and 11.

A partial defence may arise when a person is charged with murder. If successfully relied on, a partial defence results in a lesser conviction of manslaughter, even where the homicide otherwise fits the criteria for murder. A partial defence cannot be invoked if a person is charged with manslaughter (or any other criminal offence). This serves to illustrate that these defences were originally a way to circumvent the capital punishment for murder.⁶⁸⁸Creation of sentencing discretion is not the only contemporary explanation for partial defences, however. Other frequently cited rationales are fair labelling, the desirability of sharing assessments of culpability between the judge and the jury, ensuring homicides are not “overcharged” as murder and minimising perverse conviction outcomes.

New Zealand repealed its only general partial defence – provocation – in December 2009⁶⁸⁹after two Law Commission reports recommending abolition in 2001 and 2007.⁶⁹⁰On both occasions when it recommended repeal, the Commission concluded it was irrational to single out provocation (or anything else) as capable of reducing an intentional killing to manslaughter and emphasised that the defence was anachronistic.⁶⁹¹

We discuss the arguments for and against partial defences in Chapter 10.

It is a feature of partial defences that, if accepted, they always lead to a conviction for manslaughter. Because of the breadth of offending manslaughter captures, a manslaughter conviction in itself conveys little information.⁶⁹²The Law Commission noted this in 2007.⁶⁹³

683 Law Commission of England and Wales Murder, Manslaughter and Infanticide (Law Com No 304, 2006) at 4.

687 Degrees of Murder Bill 1996 (157). See also Midson, above n 682, at 234; and Law Commission Battered Defendants (NZLC PP41, 2000) at 46.

688 See discussion in Nicola Lacey “Partial Defences to Homicide: Questions of Power and Principle in Imperfect and Less Imperfect Worlds ...” in Andrew Ashworth and Barry Mitchell (eds) Rethinking English Homicide Law (Oxford University Press, Oxford, 2000) 107 at 111. See also Law Commission of England and Wales, above n 683, at 48.

690 Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001); and Law Commission The Partial

691 Law Commission Some Criminal Defences with Particular Reference to Battered Defendants, above n 690, at 41–42; and Law Commission The

692 The range of killings manslaughter captures is more or less wide depending on how a particular jurisdiction’s homicide law is structured. It includes cases of “involuntary” manslaughter, which are unlawful killings that are not murder because of the absence of murderous intent and, where partial defences exist, cases of “voluntary” manslaughter, which are unlawful killings that are not murder even though the defendant may have killed with that intention. See AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington,

2012) at 560. See also Law Commission The Partial Defence of Provocation, above n 690, at 53. We consider the implications of this potentially wide capture in connection with the “fair labelling” argument for partial defences, in Chapter 10.

Understanding family violence: Reforming the criminal law relating to homicide 119

Other than in provocation cases, the Commission observed, cases of manslaughter in New Zealand “are all unintentional killings, perhaps arising from careless or dangerous driving, medical misadventure, or a misjudged assault”.⁶⁹⁴

An alternative and perhaps preferable way to recognise reduced culpability is a separate homicide offence. Separate homicide offences have the advantage of not “lumping together”, under a common manslaughter label, people who kill intentionally in mitigated circumstances with people who may have had no intention to kill or even injure; they may avoid straining the scope of manslaughter and achieve clearer labelling.⁶⁹⁵

Examples of specific offences include infanticide,⁶⁹⁶which exists across common law jurisdictions, and, of more recent vintage, Victoria’s now-repealed “defensive homicide”. Defensive homicide was in substance akin to a partial defence of excessive self-defence, but the Victorian Government preferred to enact a separate offence partly on the basis it would confer more accurate information for sentencing.⁶⁹⁷

The merits of a new homicide offence are considered alongside our discussion of partial defences in Chapter 10.

Sometimes a decision to charge the lower-level offence of manslaughter or to substitute a charge of murder with manslaughter is in itself a way to recognise a person’s reduced culpability for homicide.⁶⁹⁸

An apparent example in the context of family violence is R v Woods,⁶⁹⁹in which the defendant was charged with the murder of her partner after stabbing him twice in the upper chest. Just before trial, a charge of manslaughter was substituted when Ms Woods offered to plead guilty to that charge. The sentencing Judge accepted “unreservedly the sincerity and integrity of [the defendant’s] remorse”.⁷⁰⁰She also noted that Ms Woods had responded to her family’s view she should have claimed self-defence with the explanation that “she would ‘never be able to live with [herself]’ if she were acquitted”.⁷⁰¹We understand from the prosecutor in Woods that, while there was an evidential basis to put murder to the jury, there were both evidential and public interest reasons that supported amending the charge to manslaughter. This suggests that, within New Zealand’s current homicide offence structure, there may be scope for recognition of reduced culpability at the charging stage.

Another circumstance in which a person may be convicted of manslaughter, despite a charge of murder, is if the jury finds the person guilty of the lesser charge. This may occur because the prosecution fails to prove the elements of murder beyond reasonable doubt. A jury might also return such a verdict if it sympathises with the defendant.

695 Horder, above n 682, at 92; and Law Commission of England and Wales, above n 683, at 9.

696 As noted above, infanticide is a hybrid provision; it is a separate substantive homicide offence, but it can also be pleaded as a form of defence to a charge of murder or manslaughter. It is a partial defence to the extent it results in a conviction of less gravity but not an acquittal. See paragraph [3.11] and the footnotes contained therein.

697 See the discussion in Kate Fitz-Gibbon Homicide Law Reform, Gender and the Provocation Defence: A Comparative Perspective (Palgrave

698 For a general discussion (in the English context) of opportunities for discretion in cases of homicide by family violence victims, see Martin

Wasik “Cumulative Provocation and Domestic Killing” [1982] Crim L Rev 29 at 32–34.

We look at New Zealand charging practice in cases of homicide by victims of family violence and the role played by juries in our case review in Chapter 9.

Finally, an offender’s reduced culpability may be recognised at sentencing. Where mandatory murder sentencing has been abolished, judges have discretion to make case-specific decisions about culpability for murder, although that discretion may still be fettered.

Jurisdictions that have abolished mandatory sentencing include New Zealand (outside of the three strikes regime), Victoria, Western Australia, New South Wales, the Australian Capital Territory and Tasmania.⁷⁰²Life imprisonment remains mandatory in England and Wales, Ireland, Canada, South Australia, the Northern Territory and Queensland.

Whether or not the mandatory sentence has been abolished, murder sentencing in some jurisdictions is still constrained by a presumption in favour of life imprisonment⁷⁰³and, in New Zealand, prescriptive rules about mandatory minimum periods of imprisonment where a life sentence is imposed. This is not the case for manslaughter for which sentences cover a wide range.⁷⁰⁴

Because of the discretion it generally affords judges to make decisions tailored to particular cases and offenders, the current sentencing process is well suited to enquiring into and accommodating “hard cases” that might fall through the cracks of the trial structure or encourage strained and artificial arguments.⁷⁰⁵Sentencing also permits a full range of responses rather than the much more limited and often binary possibilities available through verdict. We discuss in Chapter 11 whether New Zealand’s current sentencing laws are fit for purpose for victims of family violence who kill their abusers.

Historically, a conviction for murder resulted in a mandatory sentence: first, the death penalty and, later, life imprisonment. This limited judges’ capacity to recognise and reflect mitigating circumstances in sentencing. It was in this context that partial defences, which reduce murder to manslaughter in certain circumstances, developed. Outside of the three strikes regime, New Zealand no longer has mandatory murder sentencing, and at present, New Zealand law includes no general partial defences.

The question in this part of the Report is whether New Zealand’s current approach to the recognition of reduced culpability is sufficiently flexible to accommodate cases in which victims of family violence kill their abusers.

In this chapter, we have discussed means by which reduced culpability may be recognised in the course of a homicide prosecution. These include degrees of murder, tailored homicide offences, partial defences and sentencing rules. They also include “softer” mechanisms not governed by statute, namely charging practices and jury decision making.

We discuss the “softer” mechanisms in our case review in Chapter 9, where we recommend that the Solicitor-General consider whether the Solicitor General’s Prosecution Guidelines should

702 Life imprisonment is, however, mandatory in New South Wales if the person killed is a police officer in certain circumstances: Crimes Act 1900 (NSW), s 19B.

703 There is a presumption of life imprisonment for murder in New Zealand and Western Australia and, formerly, in New South Wales. For discussion, see Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007).

705 See Law Commission The Partial Defence of Provocation, above n 690, at 48–49.

Understanding family violence: Reforming the criminal law relating to homicide 121

refer to the potentially mitigating effect of a defendant’s history as a victim of family violence. However, while prosecutorial discretion and jury decision making may sometimes be favourable to defendants, due to the way the law is currently structured they lack transparency, cannot be consistently relied on and are not as readily amenable to reform.⁷⁰⁶Thus, our first line of focus in this part is the suite of statutory mechanisms for recognising reduced culpability for homicide.

More specifically, we are focused on statutory mechanisms that could be introduced or reformed with some degree of specificity to improve the legal response to victims of family violence who kill their abusers. We have not substantively considered options that are inherently wider than our terms of reference, like degrees of murder. Reform options that would necessarily affect the whole of the law of homicide would need to form part of a broader review.

706 Prosecutorial discretion is largely outside the control of defendants (even plea discussions are negotiations and not overseen by the courts) and jury deliberations are secret in all but exceptional circumstances which promotes “the finality of verdicts and uninhibited discussion during jury deliberations”. See Evidence Act 2006, s 76; and Neale v R [2010] NZCA 167 at [9]–[13].

If section 48 of the Crimes Act 1961 is clarified as we have recommended in Chapter 7, we expect that the defence of self-defence will be more accessible to primary victims of family violence. However, that defence will provide no relief for victims who kill their abusers other than in self-defence or defence of another, including those whose claims to self-defence are rejected by the jury. This is significant because the Family Violence Death Review Committee (FVDRC) considers that, although self-defence will be the more appropriate defence in the majority of cases in which victims of family violence kill their abusers:⁷⁰⁷

... not every victim of severe [intimate partner violence] uses retaliatory physical violence from a position of self-protection, as opposed to reacting with anger to what has been done to her. This means that it cannot be assumed that even if self-defence is appropriately reformed and sensitively applied in such cases, that it will necessarily always be available on the facts. In one of our regional reviews, the female offender had a very strong case for provocation but was unable to argue it because it had been abolished prior to the killing. Self-defence was also not available on the facts.

The question is how the law should deal with such homicides, which, although not justified, often involve significant mitigating circumstances. Currently, the approach in New Zealand, for all homicides where judges have discretion, is to recognise reduced culpability mainly through sentencing.⁷⁰⁸

In this chapter, we look at how the criminal justice system has responded on a case by case basis to victims of family violence who have killed abusers, particularly where the defendant did not rely successfully on self-defence. We look at how defendants were charged, whether they pleaded guilty or went to trial and how they were convicted and sentenced. We build on the summary of our case review in Chapter 3.

A homicide defendant may be convicted of a crime by one of two routes: by pleading guilty or by being found guilty after trial. The range of possible conviction outcomes in a given case will be affected by the ultimate charge (which may, for example, be amended from murder to manslaughter), and, if the case proceeds to trial, the jury’s decision.

Under the Solicitor-General’s Prosecution Guidelines, which set core standards for the conduct of public prosecutions, criminal charges are required to adequately reflect the criminality of a

707 Family Violence Death Review Committee Fourth Annual Report: January 2013 to December 2013 (Health Quality & Safety Commission, June

2014) at 119. As well as certain intimate partner homicides, the Family Violence Death Review Committee considers children who kill abusive caregivers may in particular fall outside the scope of self-defence, as the facts of such cases can tend to “look like retaliation rather than defensive force”: Family Violence Death Review Committee submission at 21 (footnotes omitted).

708 Exceptions are cases of infanticide and killing pursuant to a suicide pact (Crimes Act 1961, ss 178 and 180) and, where relevant, cases in which a defendant is insane. Otherwise, reduced culpability for homicide may be recognised outside the black letter of the law, through the use of prosecutorial discretion to reduce the charge and/or accept a plea to a lesser charge, or jury nullification.

Understanding family violence: Reforming the criminal law relating to homicide 123

defendant’s conduct but should not be inflated to increase the likelihood of an offer to plead guilty to a lesser charge (overcharging).⁷⁰⁹The overarching consideration in any plea discussions is the interests of justice.⁷¹⁰

We have not been able, in the time available to prepare this Report, to ascertain methodically how decisions to lay or amend charges were made or how plea negotiations unfolded in all the cases we have reviewed. Charging and plea discussions are neither public nor formally overseen by the courts,⁷¹¹and we have been almost entirely limited to the public record. In some cases, we can infer the charge was a straightforward reflection of the evidence and the Solicitor General’s Prosecution Guidelines. In others, the basis of the original charge or amendment of a charge from murder to manslaughter is less obvious.

Most victims of family violence who kill their abusers are charged with murder, but few are found guilty of murder at trial. Most are convicted of manslaughter. There have also been proportionately fewer guilty pleas to manslaughter in New Zealand to date than in the comparable jurisdictions of Australia and Canada.⁷¹²Some have argued these features of the cases suggest a comparatively more punitive criminal justice response to victims of family violence in New Zealand and a lack of understanding of contemporary social science around family violence.⁷¹³We explore this possibility below.

Cases we have reviewed in which defendants were charged with manslaughter usually involved force that did not suggest murderous intent,⁷¹⁴and in some cases where the charge was murder, the defendant’s use of force was such that intent could be readily inferred.⁷¹⁵

In other cases, however, the basis for the charge is less clear. Examples are R v Hu and R v Gerbes. In Hu, the defendant stabbed the deceased in the neck while he was sitting at a desk with his back turned. She said she thought she was aiming for his shoulder.⁷¹⁶In R v Gerbes, the defendant allegedly stabbed her boyfriend multiple times in the stomach during a violent confrontation. She said the deceased grabbed the blade of the knife and stabbed himself.⁷¹⁷Ms Hu and Ms Gerbes were both charged with manslaughter. Ms Hu pleaded guilty to that offence. Ms Gerbes was ultimately acquitted at trial.⁷¹⁸

The prosecutorial decisions to lay manslaughter charges in Hu and Gerbes may be contrasted with the murder charges in a number of cases that involved single stab wounds inflicted in

711 The Victorian Law Reform Commission noted that charging and plea practices had been generally criticised for their lack of transparency and accountability and that there is arguably a public interest in records of charging and plea decisions being publicly available: Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 108.

712 Elizabeth Sheehy, Julie Stubbs and Julia Tolmie “Battered Women Charged with Homicide in Australia, Canada and New Zealand: How Do

714 Including a stab in the leg (R v Stone HC Wellington CRI-2005-078-1802, 9 December 2005; and R v Tamati HC Tauranga CRI-2009-087-0868,

27 October 2009) and a hit over the head with a pepper grinder (R v Rakete [2013] NZHC 1230).

715 For example, multiple stab wounds to the upper body (R v Neale HC Auckland CRI-2007-004-3059, 12 June 2009; and Neale v R [2010] NZCA

167); and the use of a shotgun (R v Wickham HC Auckland CRI-2009-090-010723, 20 December 2010).

717 “Hung jury in manslaughter trial” (13 February 2014) Stuff.co.nz <www.stuff.co.nz>; and Tracey Chatterton “Accused ‘just needed to call out’” (12 February 2014) Stuff.co.nz <www.stuff.co.nz>.

718 Ms Gerbes was acquitted at her second trial; the jury in her first trial was hung: “Hung jury in manslaughter trial”, above n 717.

response to violent assaults.⁷¹⁹In R v Mahari,⁷²⁰for example, the defendant stabbed her partner in the shoulder area behind his neck with a kitchen knife as he broke into the cabin she shared with the deceased. In R v Wharerau, the defendant killed her partner with a single stab wound to the chest. The couple had been arguing, and the deceased slapped her in the face and broke her phone. He followed her into the kitchen, and Ms Wharerau grabbed a knife off the bench and swung it in his direction to scare him away. After stabbing him, the defendant sought help from a neighbour and returned home to support the victim until the ambulance arrived.⁷²¹
Despite being charged with, and tried for, murder, Ms Mahari and Ms Wharerau were both

convicted of manslaughter.

Looking only at the inflictions of fatal injuries in these cases, there is little in isolation from the defendants’ accounts to distinguish them in terms of murderous intent. The manslaughter charges in Hu and Gerbes are apparently most readily explicable on the basis that, in those cases, weight was accorded to the defendants’ version of events.

In four cases, including R v Tagatauli,⁷²²the defendants avoided a trial for murder by pleading guilty to manslaughter.

One such case pre-dated the repeal of provocation, and the decision to accept the plea seems to be explained by a provocation narrative.⁷²³In the other three, R v Brown,⁷²⁴R v Woods⁷²⁵and R v Tagatauli,⁷²⁶the defendants stabbed their abusive partners in the context of confrontations. In R v Brown, the sentencing Judge considered that manslaughter was the right outcome on the basis that, while the defendant had “intended to stab [the victim] and in a general way to hurt him, she had no intention to kill and had no realisation that she might kill him or even that she had done so”. Thus, she had committed a “deliberate but unthinking act”, which, in law, “add[ed] up to the offence of manslaughter”.⁷²⁷In R v Woods, the defendant said she did not mean to hurt the deceased. She acknowledged she had been angry at the time of the stabbing but said she believed he was going to continue his assault.⁷²⁸

One commentator has recently observed that a significant number of cases since 2009 in which victims of family violence have been alleged to have killed their abusers appear to have been resolved by guilty pleas to manslaughter.⁷²⁹If so, and if such charging decisions are in part

719 R v Raivaru HC Rotorua CRI-2004-077-1667, 5 August 2005; R v Mahari HC Rotorua CRI-2006-070-8179, 14 November 2007; R v Brown HC Napier CRI-2008-020-3130, 24 November 2009; R v Ford HC Auckland CRI-2010-044-132, 22 July 2011; R v Paton [2013] NZHC 21 [Paton (HC)]; R v Paton [2009] NZCA 155 [Paton (CA)]; R v Wharerau [2014] NZHC 1857 [Wharerau (HC)]; and Wharerau v R [2015] NZCA 299 [Wharerau (CA)]. See also in relation to Honor Stephens, Bridget Carter “Jury accepts battered-wife defence in murder trial” The New Zealand Herald (online ed, Auckland, 24 April 2002); and in relation to Jessica Keefe, “Jessica Keefe not guilty of murder” (19

September 2013) Stuff.co.nz <www.stuff.co.nz>, and “Murder charge unwarranted – lawyer” (21 September 2013) Radio New Zealand

722 Aaliyah Tagatauli is also known as Amanda Taitapanui. See “Woman admits killing partner” (23 March 2016) Stuff.co.nz

726 See “Fatal stabbing in the leg leads to rare sentence of home detention” (21 April 2016) Stuff.co.nz <www.stuff.co.nz>.

729 Julia Tolmie “Defending Battered Defendants on Homicide Charges in New Zealand: The Impact of Abolishing the Partial Defences to Murder” [2015] NZ L Rev 649 at 663.

Understanding family violence: Reforming the criminal law relating to homicide 125

recognition of the defendants’ reduced culpability, it is debatable whether, in practice, there is any problem with charging.

On the other hand, whatever the outcome for individual defendants, reliance on prosecutorial discretion may be problematic if there is inconsistency in approach across the cases. Some also argue that in any event discretion is no substitute for transparent legal rules about when homicide will and will not amount to murder.⁷³⁰

On the basis of the information we have identified, it appears that, in some cases (notably Mahari and Wharerau), it would have been tenable to charge manslaughter rather than murder. However, we do not have all the information to make that assessment, and even if the lesser charge would have been tenable, that does not mean the decisions to charge murder were wrong. Murderous intent is a question of fact and therefore legitimately put to the jury where it is in issue and the evidential test for prosecution is met. A decision to charge murder is not impugned just because the charge is subsequently amended or the jury convicts of manslaughter. That is particularly so when the jury has scope to convict the defendant of manslaughter or acquit. The small number of cases in which defendants pleaded guilty to manslaughter after first being charged with murder also suggests it is unlikely there is “overcharging”.

We do not consider the low number of guilty pleas overall in our case review is in itself evidence of a problem. It means most defendants are testing the prosecution case at trial. In jurisdictions where most victims of family violence who commit homicide plead guilty, concerns have been raised that the limits of the law are not being tested, risking inconsistent outcomes, slow development in the law and limited guidance for judges and lawyers as to how such cases should be treated.⁷³¹

We have therefore identified no evidence of a practical problem in this area. It appears that the fact a defendant may have been a victim of family violence can be taken into account in charging for homicide, and the Crown Law Office has told us there have been recent cases in which that fact was very much to the fore of the prosecutor’s mind in respect of requests for consideration of amending charges of murder to charges of manslaughter. We do not, however, know the extent to which this is reflective of standard practice or whether approaches are consistent among prosecutors.⁷³²The Solicitor-General will be in the best position to make that judgement.

We understand the Solicitor-General’s Prosecution Guidelines are periodically reviewed. In these circumstances, the Solicitor-General may wish to give consideration to whether the Solicitor- General’s Prosecution Guidelines should include express reference to the potential relevance of a defendant’s history as a victim of family violence.

730 See James Chalmers “Partial Defences to Murder in Scotland: An Unlikely Tranquility” in Alan Reed and Michael Bohlander (eds) Loss of

Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 167. Chalmers argues, at

180–181, that, in Scotland, “even if the current law is operating satisfactorily in terms of the outcomes arrived at in particular cases (something which is difficult to assess in the absence of empirical research), the law as it stands is overly reliant on the benign exercise of prosecutorial discretion” and that it is necessary that “the letter of the law creates a clear, transparent and just framework for determining when a verdict of culpable homicide is to be preferred to one of murder”.

731 Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007) at 284.

732 The Attorney-General’s Introduction to the Solicitor-General’s Prosecution Guidelines notes that, unlike most similar jurisdictions, “New Zealand has no centralised decision-making agency in relation to prosecution decisions”, and that “[t]he absence of a central decision-making process underscores the importance of comprehensive guidelines, and the acceptance of core prosecution values”: Crown Law, above n 709, at 1.

^R8The Solicitor-General should, when next reviewing the Solicitor-General’s Prosecution

Guidelines, consider whether they should include express reference to the potential relevance of a defendant’s history as a victim of family violence.

While, in most cases we reviewed the defendants were charged with murder, few were convicted of murder. Setting to one side cases resolved by guilty pleas, out of the 14 victims of family violence who defended a murder charge, three were convicted of murder. Eight were found guilty of manslaughter, and three were acquitted.

In the absence of a partial defence, on a charge of murder, the only legal route to a manslaughter conviction is a finding that the prosecution has not proved that the defendant had murderous intent.⁷³³As we identified in the Issues Paper, however, it may be difficult to explain manslaughter verdicts for some cases on this basis.

An alternative explanation for such verdicts is that the jury departed from a “strict” application of intent. The jury may have sympathised with the defendant if, for example, they believed the defendant was acting defensively, but that self-defence did not apply, or some verdicts may reflect jurors’ differing assessments of the facts.

In this section we examine trial outcomes. Because jury decision making is secret and juries do not give reasons, we have been able only to examine whether the manslaughter and murder verdicts are explicable in terms of the law and the facts disclosed by the reported material (usually sentencing decisions).

To be convicted of murder, a defendant must have either intended to kill or intended to cause bodily injury known to them to be likely to cause death and been reckless whether death ensued (reckless intent).⁷³⁴Intent is a question of fact that can be inferred from the defendant’s conduct and the circumstances he or she must have known.⁷³⁵If, on a person’s trial for murder, the jury finds the defendant committed a culpable homicide but is not satisfied they did so with murderous intent, it should find the defendant guilty of manslaughter.

In light of the criminal standard of proof and the seriousness of a murder conviction, it is to be expected juries will scrutinise evidence of murderous intent. Given New Zealand has no general partial defences, it is also unsurprising that intent may be the “battleground” in these trials.⁷³⁶

Murderous intent (particularly reckless intent) is often complicated, particularly when the defendant was responding to an attack or confrontation. Determining the factual position involves fine judgements. Proof of intent will turn on evidence about the defendant’s state of mind, why they used force and the circumstances in which the injury was inflicted. Many of the cases we have reviewed involved frenetic confrontations marked by fear and other emotions.

733 For recent discussion of the issue of murderous intent in cases of homicide by victims of family violence, see Tolmie, above n 729, at 664.

734 Crimes Act 1961, s 167. A person may also be guilty of murder under the “felony murder” rule, noted in Chapter 3, but that is not relevant in the present context.

735 AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington, 2012) at 554.

736 This was observed by the Irish Law Reform Commission when considering the implications of the repealing provocation: Law Reform Commission of Ireland Consultation Paper on Homicide: the Plea of Provocation (LRC CP 27-2003, 2003) at 127–128. See also Law Commission The Partial Defence of Provocation (NZLC R98, 2007) at n 93.

Understanding family violence: Reforming the criminal law relating to homicide 127

We would suggest they illustrate the complexities involved in murderous intent in this context. Some examples provide helpful illustrations.

In R v Wickham,⁷³⁷the defendant, who suffered from multiple sclerosis, called the Police before retrieving a gun and shooting her husband. She said afterwards she could not recall pulling the trigger. She claimed self-defence, saying that her husband had grabbed her by the throat, knocked her over, thrown a full bottle of spirits at her and threatened to drown her in the swimming pool and “gut [her] like a fish”.⁷³⁸The jury rejected her claim of self-defence but convicted her of manslaughter. The sentencing Judge had “no doubt” the defendant had been scared at the time of the killing, considered it “likely that the jury considered that the force that [she] used in ... presenting a loaded shotgun at [the deceased] was disproportionate to the threat [she was] facing” and “agree[d] with the submission made by [her] counsel and the Crown that in finding [her] guilty of manslaughter the jury must have concluded that [she] pulled the trigger by accident”.⁷³⁹

R v Wharerau and R v Mahari, where the defendants were convicted of manslaughter, also involved claims of self-defence. In Wharerau, the defendant claimed that she grabbed a knife off the kitchen bench and swung it at the deceased to scare him away but did not mean to stab him. In Mahari, the defendant stabbed the deceased in the shoulder as he broke into the caravan in which she had barricaded herself.

That an unsuccessful claim of self-defence may see a defendant convicted of manslaughter for reasons relating to intent has also been noted overseas. In its report on partial defences to murder, the Law Commission of England and Wales observed that, “notwithstanding the complete nature of the defence [of self-defence], the facts which fall short of substantiating self-defence may, nonetheless, form the basis of a conviction for manslaughter”,⁷⁴⁰citing the following passage from the English Court of Appeal case of McInnes:⁷⁴¹

[I]f a plea of self-defence fails for the reason stated, it affords the accused no protection at all. But it is important to stress that the facts upon which the plea of self-defence is unsuccessfully sought to be based may nevertheless serve the accused in good stead. They may, for example, go to show that he may have acted under provocation or that, although acting unlawfully, he may have lacked the intent to kill or cause serious bodily harm, and in that way render the proper verdict one of manslaughter.

As well as factual complexity where a fatal injury is inflicted during a confrontation, it is possible juries are taking into account histories of family violence when analysing murderous intent. Rebecca Bradfield has researched Australian cases where women killed male partners and concluded that, in some, lack of intent was “being used as a defacto defence of domestic violence”.⁷⁴²The violence the defendants had suffered before the homicides, Bradfield suggested, was:⁷⁴³

... appropriated to explain the emotional state of the accused at the time of the killing. It has been accepted that emotional turmoil and anger are factors relevant to the issue of whether the accused had the requisite intention for murder (Cutter [(1997) 94 A Crim R 168] at 156 per Brennan CJ

_{and Dawson J). The impact of the history of violence on the
accused’s psychological/emotional state,}

737 R v Wickham, above n 715. See also Brenda Midson “Degrees of blameworthiness in culpable homicide” [2015] NZLJ 220; and Victoria

Robinson “Home detention for killing husband” (20 December 2010) Stuff.co.nz <www.stuff.co.nz>.

740 Law Commission of England and Wales Partial Defences to Murder (Law Com No 290, August 2004) at 75.

742 Rebecca Bradfield “Women Who Kill: Lack of Intent and Diminished Responsibility as the Other ‘Defences’ to Spousal Homicide” (2001) 13

together with the accused’s fear or anger before the killing, are used to explain the finding of lack of intent.

This approach may have arisen in cases in our sample. In R v Paton,⁷⁴⁴the defendant’s partner followed her into the kitchen after attacking her in the hallway, and she stabbed him in the neck with a kitchen knife. The jury rejected her claim of self-defence but convicted her of manslaughter. At sentencing, the Judge noted this must reflect a finding of lack of intent but said he viewed the homicide as very close to murder:⁷⁴⁵

The jury rejected your claim that you acted in self-defence, but found that you did not have murderous intent. Now murderous intent can be present where you do not actually intend to kill, but intend to cause injuries of a type that are likely to kill, and where you are reckless as to whether death does occur. A stab wound to the neck with a large kitchen knife is likely to kill the victim. You may recall the evidence from the pathologist that the injury was effectively “unsurvivable”. I treat the jury analysis as recognising that the view as to the risk of death from a stab wound of this kind, by a woman in your position, would not be analysed as it would be by most of us. The prolonged history of beatings conditioned you to downplay the risks and consequences of violent attacks, so that a woman in your position would not appreciate the risk of causing death when others, who had not experienced the sad domestic history you had, could reasonably be expected to recognise that risk.

I accept the jury’s analysis but I see the level of violence used in a stab to the neck with a large knife as being at the most serious end of any scale of attacks that might not involve murderous intent. That means, Ms Paton, that a stabbing of this type, in these circumstances, is only a very short way from murder.

_{The manslaughter verdicts: was the verdict a result of sympathy or
jury nullification?}

Explaining manslaughter verdicts in terms of lack of intent, however construed, puts them in a legal framework, but there is no way to be certain such an analysis actually reflects the juries’ reasoning. It is also possible juries are reaching manslaughter verdicts on grounds of sympathy rather than a strict application of the law. Wickham and Paton may be examples.⁷⁴⁶Some verdicts may, alternatively, reflect compromises between jurors with different views of the evidence. Yet another possibility is that some verdicts reflect mingled doubt about defendants’ moral blameworthiness and legal culpability. One person we spoke with who is experienced in jury research suggested that, even if asked, some juries may not readily be able to articulate the reasons for their decisions. A jury might both have doubt about the evidence and consider the defendant’s culpability to be less than that reflected by the offence charged. It may be unsurprising for legal and moral considerations to be mingled in juries’ decision making.⁷⁴⁷

“Jury nullification” describes verdicts that result when juries deliberately do not apply the law because doing so would run counter to their view of justice in a particular case. It has been observed to be a legitimate feature of the jury system and consistent with the jury’s function to act as the “conscience of the community” and “safeguard against arbitrary and oppressive government”.⁷⁴⁸This is an important function, even if there are few cases in which the jury

744 Paton (HC), above n 719; and Paton (CA), above n 719. See also the discussion of the approach to intent in Paton in Tolmie, above n 729, at

746 We also note the case of Daryl Kirk, before the High Court at the time of publication of this Report. Ms Kirk was charged with the murder of her mother’s partner, and claimed self-defence. On 20 April 2016 the jury returned a verdict of manslaughter. Ms Kirk has yet to be sentenced.

747 Martin Wasik “Partial Excuses in the Criminal Law” (1982) 45 MLR 516 at 517–518.

748 Law Commission Juries in Criminal Trials Part One: A Discussion Paper (NZLC PP32, 1998) at 15–18.

Understanding family violence: Reforming the criminal law relating to homicide 129

actually acts as such a “check”.⁷⁴⁹The special role of jury nullification in murder trials has been noted in Australia by White J in R v Marshall:⁷⁵⁰

In a murder case, community values are reflected in a special way on such subject matters as provocation, self-defence, intention and manslaughter; in the latter case, the jury has a “constitutional right” to bring in a merciful verdict of manslaughter even where the elements of murder are proved. That merciful verdict belongs to the jury ...

If juries were consistently returning manslaughter verdicts even where the elements of murder were made out, however, there would be an argument the law should be brought into line with this collective view of culpability. It might be thought wrong that a jury could feel forced either to make a decision at odds with its finding on a defendant’s moral culpability or to convict of manslaughter or acquit “perversely”.⁷⁵¹

Brenda Midson develops this argument by reference to three apparently intentional New Zealand homicides – one of which was Wickham, discussed above – where the defendants were acquitted or convicted of manslaughter.⁷⁵²She argues that the outcomes in these cases were “undoubtedly fair if moral blameworthiness is the basis of criminalisation” but that:⁷⁵³

The fundamental problem with these decisions is that they signal that the outcome for a defendant depends upon the whims of the jury rather than the application of legal principles. If these decisions continue unchecked, inconsistent outcomes will result. In other words, if there are degrees of culpability then it would be as well to be upfront about them.

The point about fairness, as Midson notes and some submitters raised,⁷⁵⁴is that decision making that depends on sympathy or prejudice is inconsistent. Different juries have different sympathies, and some defendants evoke more sympathy than others. It seems unfair for decisions about culpability to turn on these “whims”. Cases of jury nullification might also be troubling if they make it harder for judges to determine facts for sentencing,⁷⁵⁵although, as we discuss in the next chapter, a legal mechanism such as a partial defence would not necessarily provide greater clarity.

If the cases demonstrate that juries are already empowered to reach verdicts that reflect their views of moral culpability, thus averting substantive unfairness, it is difficult to argue that introduction of a partial defence would lead to different results.⁷⁵⁶Because our review is limited to victims of family violence, we have not considered manslaughter conviction rates for other murder trials and cannot say to what extent the trends we have observed are particular to these defendants. As a general point, though, there will always be cases in which the elements of

749 Law Reform Commission of Canada The jury in criminal trials: Working Paper 27 (1980) at 10–12.

750 R v Marshall (1986) 43 SASR 448 at 449. Discussed in Law Commission, above n 748, at 26.

752 The other two cases were R v Apatu HC Wellington CRI-2009-041-3156, 14 October 2010; and R v Scollay [2014] NZHC 465. Mr Apatu was acquitted of the murder of his neighbour, whom he shot. He admitted going to the neighbour’s house but said he only intended to scare him. There was evidence at trial about Mr Apatu’s mental state at the time of the killing. Midson notes that, in sentencing Mr Apatu on a minor offence to which he had pleaded guilty, Miller J, at [6], observed that the acquittal must be explained by a jury finding that the Crown had not negated beyond reasonable doubt that the killing was accidental. See Midson, above n 737, at 231. Ms Scollay was found guilty of the manslaughter of her husband, whom she stabbed. She said she did not intend to kill him but wished to shake him out of his deeply depressed state: at [7]. See Midson, above n 737, at 231.

754 The Family Violence Death Review Committee, for example, noted it was important to ensure that manslaughter convictions are placed on a solid legal foundation rather than relying on the jury to “do the right thing regardless of the legal position”. The Auckland District Law Society and Jeremy Hammington, Barrister, expressed similar views. The Law Commission of England and Wales took the same view in respect of prosecutorial discretion in its 2004 Report on partial defences: Law Commission of England and Wales, above n 740, at 54.

755 See, for example, the Courts’ observations as to the likely grounds for the verdicts in R v Erstich [2002] NZCA 122; (2002) 19 CRNZ 419 (CA) at [2]; and R v King

756 James Chalmers suggests, for example, that a possible explanation for the lack of impetus for reform of provocation in Scotland is that

“substantive injustice” in the operation of the defence has been averted by prosecutorial discretion and jury nullification: Chalmers, above n

an offence seem to be proved but the jury appears to sympathise with the defendant. The law cannot necessarily anticipate all exceptional or “hard” cases.⁷⁵⁷

Any decision making that depends on value judgements by the jury presents a risk of inconsistency. Determining whether conduct was “reasonable” for the purposes of self-defence, for example, requires application of community values and reference to individual experiences. These will inevitably vary, depending on the composition of the jury.

We identified three cases where defendants were found guilty of murder after trial, and a fourth case where the defendant pleaded guilty to murder:
^._{In R v Reti,}⁷⁵⁸_{a 2008 case, the
defendant relied unsuccessfully on provocation. The day of}

the murder, after an argument, Ms Reti stabbed the victim in the leg. She called for help, and the victim was treated. Later that day, they argued again and Ms Reti got a knife and stabbed the victim “with so much force that the whole blade was buried in his chest”. Ms Reti said that, after she retrieved the knife, she asked the victim if he wanted her to “poke” him, and he “kept saying bring it on bitch bring it on”, spat in her face and kicked her in the stomach. Ms Reti said she “just retaliated” by attacking him.⁷⁵⁹
^._{In R v
Neale,}⁷⁶⁰_{a 2009 case, the defendant and victim had
been in an on-again, off-again}

relationship. Ms Neale went to the victim’s apartment with a knife in her handbag and stabbed him in the chest while he was in the shower. She inflicted eight more stab wounds (to the back, throat, shoulder, arm and abdomen).⁷⁶¹
^._{In R v
Wihongi,}⁷⁶²_{a 2010 case the circumstances of which
we discuss elsewhere in this}

Report, the defendant followed the victim out of the house and stabbed him in the chest when they were both outside the house. When the victim drove away, Ms Wihongi followed and, when she reached his car, punched at him through the window.
^._{In R v Rihia, a 2012 case and the one case
in our review involving a guilty plea to murder,}

the defendant stabbed the victim in the chest as he lay on the couch. He had been assisted to lie down there after Ms Rihia threw a stereo speaker at his head.

Of the murder cases, R v Reti may be the most finely balanced in terms of intent, as Ms Reti’s evidence that she intended only to injure the victim was supported by the fact she sought assistance when he began to bleed. However the fatal wound was inflicted only hours after Ms Reti had stabbed the deceased in the leg (for which he had received paramedic treatment). The sentencing Judge observed that the murder verdict suggested that the jury:⁷⁶³

... would have found that when you stabbed him in the chest you did intend to cause him grievous bodily harm. You knew there were risks but you were simply reckless whether death ensued or not. You showed that by stabbing him in in the area of the chest where you did and, as I said, embedding the knife up to the handle. Also, while accepting you may have suffered from post-traumatic stress disorder as a result of what you have suffered as a young child, the jury in my assessment would have

758 R v Reti HC Whangarei CRI-2007-027-2103, 9 December 2008 [Reti (HC)]; and R v Reti [2009] NZCA 271.

761 We note our information source for Neale includes media reports, given the bare nature of the court decisions. See Kim Ruscoe “Accused was ex’s prostitute” (20 April 2009) Stuff.co.nz <www.stuff.co.nz>.

762 R v Wihongi HC Napier CRI-2009-041-2096, 30 August 2010 [Wihongi (HC)]; R v Wihongi [2011] NZCA 592, [2012] 1 NZLR 775 [Wihongi

Understanding family violence: Reforming the criminal law relating to homicide 131

found that even if you were provoked you did not display the self-control expected of a person in those circumstances.

The four murder cases share a number of features:
^._{The
level of violence in the fatal attack was strong evidence of intent either to
kill or to inflict}

_{very serious injury with recklessness as to death
resulting.}
^._{While the relationship between the
defendant and the deceased was violent and abusive,}

the court decisions we have reviewed do not record that there was evidence the defendant believed her life was threatened either on the day of the offending, or more generally in the future. There may be an argument this suggests the presentation or interpretation of the evidence in those cases was affected by persisting misconceptions about family violence, although we do not have the information to identify whether or not that is so.
^._{The fatal attacks occurred in the context of a confrontation or
argument, but the deceased}

had earlier been physically impeded or was not advancing at the time.

Cases in which manslaughter verdicts were returned (except where provocation was relied on and apparently accepted), by contrast, tend to feature:
^._{evidence the defendant intended to injure the
deceased but not that they were reckless as to}

the risk death might result (for example, where the defendant stabbed the victim in the leg);

_and/or
^._{evidence the defendant’s use of
force was defensive and in the face of a violent assault or}

threat of violence by the deceased.

Typically, in the cases that resulted in a manslaughter conviction despite a murder charge, the defendants were alleged to have acted with reckless murderous intent. As we have noted, fine judgements are likely to be involved in determining the factual position where a defendant is responding to an assault or confrontation. In all the manslaughter cases (other than where provocation was apparently accepted), there was evidence suggesting either that, given the nature of the force, the defendant did not act with murderous intent or the defendant was trying only to ward off an assault or threat. It does not appear, in other words, that these manslaughter verdicts were at odds with the evidence.

At first glance, the low rate of murder convictions could suggest a reluctance to convict victims of family violence of murder even when the elements of murder have been proved. However, this is not necessarily borne out by close analysis. Murder convictions are rare, but they appear generally to be returned in cases where there is evidence of higher-level offending. It appears juries are carefully construing the requirements of murderous intent and returning manslaughter verdicts where there is reasonable doubt about the defendant’s state of mind when the fatal injury was inflicted.

We do not think the cases disclose arbitrary or inexplicable conviction outcomes. The possibility of jury nullification cannot be discounted, however, because jury decision making is conducted behind closed doors. For this reason, we do not rely solely on what we have gauged from the cases. In the next chapter, we examine the in-principle case for partial defences to assess whether there is a gap in the law.

The partial defence of provocation was, at least in law, available in 10 cases we reviewed but (apparently) successfully relied on in only three (R v Suluape, R v Erstich and R v King). It was relied on but rejected in two others (R v Neale and R v Reti).

The FVDRC suggests that a comparison of cases before and after the repeal of provocation demonstrates that repeal has adversely affected the position of victims of family violence who kill abusers.⁷⁶⁴The FVDRC compares sentences in cases in which the verdict was manslaughter by reason of provocation – R v Suluape (in which the defendant killed her husband by striking him a number of times on the head with an axe while his back was turned) and R v King (where the defendant killed her husband by putting ground up sleeping pills in his food) – with those in R v Wihongi and R v Rihia, in which the verdicts were murder.

R v Wihongi pre-dates repeal of provocation, but Ms Wihongi did not rely on the defence, and so we are reluctant to draw conclusions from her case.⁷⁶⁵R v Rihia is one post-repeal case in which provocation might have availed the defendant if it had still been in force.⁷⁶⁶In that case, on the day of the homicide Ms Rihia’s youngest child had been removed to Child, Youth and Family custody. The sentencing Judge said:⁷⁶⁷

I am satisfied in your case also that you would not have stabbed Mr Rihia to death had it not been for the significant mental impairment you suffer through years of alcohol abuse and physical abuse most recently, until you parted with him, at the hands of Mr Rihia himself. It was that background of abuse which led to all three of your children of that relationship being taken from you. I am satisfied that the immediate cause of your offending was the further and final repetition of those distressing events and that they led you to ‘just snapping’, as you said.

On its face, this account of events might have supported a finding of provocation. That may have enhanced Ms Rihia’s ability to negotiate a manslaughter charge or encouraged her to go to trial rather than pleading guilty, thereby increasing the likelihood of a manslaughter verdict. The difficulty is that we do not know why Ms Rihia pleaded guilty, and we do not know what would have been the outcome if she went to trial – with or without the aid of the partial defence.⁷⁶⁸

Further, while Ms Rihia was convicted of murder, the sentencing Judge took into account her “extreme reaction” to her despair at losing her child and concluded it would be manifestly unjust to impose life imprisonment.⁷⁶⁹The circumstances that led to and surrounded Ms Rihia’s offending were, therefore, recognised in a concrete way, albeit through sentencing discretion, not a partial defence.

Only six years have passed since provocation was repealed, and homicides by victims of family violence are not common. Before repeal, provocation was relied on, and not always successfully, in only a small number of cases. It may be too early to tell whether repeal of the partial defence

765 Although Julia Tolmie has suggested Ms Wihongi may have had an arguable case for manslaughter by reason of provocation: Tolmie, above n

768 It is worth noting that, in R v Reti, which bears some factual similarity to R v Rihia, the defendant relied unsuccessfully on provocation. At sentencing, the Judge said: “Ms Reti at trial your defence was two-fold, first a lack of intent to kill and secondly, provocation. I heard the same evidence the jury heard. In finding you guilty of murder the jury rejected those defences, I have to say rightly in my view ... while accepting you may have suffered from post-traumatic stress disorder as a result of what you have suffered as a young child, the jury in my assessment would have found that even if you were provoked you did not display the self-control expected of a person in those circumstances”: Reti (HC), above n 758, at [10].

Understanding family violence: Reforming the criminal law relating to homicide 133

has disadvantaged this group of defendants.⁷⁷⁰At present, we consider there is insufficient evidence to conclude that it has.

Unlike jury verdicts, sentencing is conducted in public, and decisions are almost always transcribed. It is therefore relatively straightforward to review and compare reasons for decisions, but we are still restricted to the public record, and we have not seen or heard the primary evidence (although sentencing notes would usually record that evidence).

There is variation among sentences, particularly between the murder and manslaughter cases we have reviewed. Beyond the sentences themselves, two features of the courts’ approach to the sentencing process warrant mention at the outset:
^._{There is some variation in how the courts have characterised and
approached defendants’}

_{histories of violence and abuse in these cases.}
^._{Where relevant, the courts have consistently adverted to
defendants’ future risks of}

reoffending and prospects of rehabilitation in determining the appropriate sentence.

We discuss these features in detail in Chapter 11.

Manslaughter captures a wide range of conduct, and the range of penalties the courts are willing to entertain is correspondingly wide. Sentences in the manslaughter cases we have reviewed (of which there were 15) ranged from a suspended term of imprisonment (R v Erstich) and
12 months’ home detention (R v Wickham and R v Tagatauli) to five years and six months’
imprisonment (R v Brown). We have not identified a problem with sentences for manslaughter in this area.⁷⁷¹

Sentencing for murder is relatively complex. Section 102 of the Sentencing Act 2002 prescribes a presumption in favour of life imprisonment that is rebuttable only if, “given the circumstances of the offence and the offender, a sentence of imprisonment for life would be manifestly unjust”. Where a life sentence is imposed, sections 103 and 104 prescribe mandatory minimum terms of imprisonment. The appropriate minimum term will depend on the circumstances of the offending.⁷⁷²

The “manifestly unjust” test is strict. Since it was enacted in 2002, the presumption of life imprisonment has been displaced in only six cases, as far as we have ascertained.⁷⁷³The

770 This was a point made during Victoria’s 2010 consultation on abolition or retention of the offence of defensive homicide, which had been introduced in 2005. Some submitters believed it was too early to know whether the new offence was a necessary “safety net” for women who kill abusive partners, and others recommended retention on the basis that operation of the offence would be reviewed further in three or five years’ time: Victoria Department of Justice Defensive Homicide: Proposals for Legislative Reform – Consultation Paper (September 2013) at 12–13. We note, however, that Julia Tolmie considers enough time has passed since the repeal of provocation to be able to assess the effects of repeal: Tolmie, above n 729, at 650.

771 Although Julia Tolmie has recently suggested that subsisting misconceptions about family violence, in conjunction with difficulties in accommodating the nature and effects of family violence within existing sentencing doctrine and structure, may mean sentences imposed on victims of family violence convicted of manslaughter are disproportionately long when compared to sentences imposed on predominant aggressors convicted of manslaughter: Tolmie, above n 729, at 680.

773 According to Statistics New Zealand records, 371 life sentences have been imposed since 2001. Finite sentences for murder are therefore highly exceptional – representing less than two per cent of the total.

presumption was not discussed in the two earlier murder cases in our review, R v Neale and R v Reti, in both of which the offenders were sentenced to life imprisonment with minimum periods of imprisonment of 10 years.⁷⁷⁴The presumption was, however, reviewed and found to be rebutted in R v Wihongi and R v Rihia. The Court of Appeal in Wihongi considered section

Overall, we see this as a case of a battered defendant who has reacted in an extreme way to her abuser in circumstances where both the history of abuse and the offender’s cognitive deficits have played a significant role in that extreme reaction arising. We see this as a case falling within the class of cases that Parliament contemplated would justify the displacement of the presumption [of life imprisonment].

In Rihia, the Judge discussed Wihongi and concluded the two cases were sufficiently similar that they fell into the same category for the purposes of the “manifestly unjust” test.⁷⁷⁶There was no Crown appeal against the sentence in that case.

The Courts in Wihongi and Rihia imposed finite terms of 12 and 10 years respectively. In Wihongi, after considering fresh evidence as to Ms Wihongi’s risk of future offending and the need to take account of the sentencing purposes of denunciation and deterrence, the Court of Appeal substituted the 12-year term for the eight-year term originally imposed by the High Court. In Rihia, the Judge relied on Wihongi in setting the finite term and assessing the starting point.⁷⁷⁷

Some commentators and the FVDRC (in its Fourth Annual Report and its submission on the Issues Paper) have suggested that, notwithstanding displacement of the presumption of life imprisonment, the finite sentences imposed in Wihongi and Rihia may be too high.⁷⁷⁸To some extent, this depends on a comparison of murder and manslaughter cases, which assumes juries’ decisions whether to return manslaughter or murder verdicts are somewhat arbitrary. The cases do not suggest this is valid. It is also plausible that the murder cases involved more serious offending and/or stronger evidence of murderous intent.

It is also not self-evident that finite terms of 10 and 12 years are “too” high. Attitudes towards punishment vary. Some members of the community will think that, despite mitigating circumstances, the sentences in Wihongi and Rihia do not adequately reflect the seriousness of the offending. Those sentences are, moreover, among the lowest ever imposed for murder in New Zealand,⁷⁷⁹and the Supreme Court declined Ms Wihongi leave to appeal against the Court of Appeal’s decision.⁷⁸⁰We explore these issues in Chapter 11.

Whatever view is taken of individual sentences, it must be borne in mind that sentencing decisions are made by judges who have heard the evidence, considered relevant reports and other material (including victim impact statements) and heard legal submissions. Without being

774 At sentencing in R v Neale, above n 715, the High Court Judge said, at [1], “Ms Neale ... As you know, the maximum, and indeed the mandatory, sentence on a charge of murder is one of life imprisonment”, notwithstanding that Ms Neale’s case post-dated enactment of the Sentencing Act

2002. There appears to have been no suggestion that the s 102 presumption might be displaced in Ms Neale’s case, however.

777 Ms Rihia was sentenced on the basis of a 12-year starting point (following Wihongi) with a two-year discount to reflect her guilty plea.

778 Tolmie, above n 729, at 667; Family Violence Death Review Committee, above n 707, at 121; and Family Violence Death Review Committee submission at 36.

779 The only case we have identified with a lower sentence for murder is R v Law (2002) 19 CRNZ 500 (HC), a euthanasia case, in which the High Court imposed a finite term of 18 months’ imprisonment. The Court held, at [51]–[53], that the circumstances of the offence were such that a sentence of life imprisonment would be manifestly unjust and accepted the defence submission that the full range of sentencing options under the Sentencing Act was available. Mitigating factors were the offender’s age, health, motives, guilty plea, acceptance of responsibility, remorse and previous good character.

Understanding family violence: Reforming the criminal law relating to homicide 135

fully appraised of the facts, it is difficult for us or others not involved in the cases to say with confidence whether a given sentence properly reflects the Sentencing Act and relevant case law.

Overall, we do not think the charging, trial or sentencing outcomes in these cases suggest that the structure of New Zealand’s homicide law – with two main categories of culpable homicide, no partial defences and generally discretionary sentencing – is giving rise to problems for victims of family violence who kill abusers. As we discuss in Chapter 11, we think there may be scope to improve the consistency with which histories of family violence are approached at sentencing, but we do not consider sentence methodology or outcomes are otherwise troubling.

Our review has, however, taken in a fairly small number of cases, and there are methodological limitations in our analysis. In these circumstances and given our terms of reference require us to examine whether a partial defence is justified, we consider it is important also to review the case for partial defences in principle.⁷⁸¹We do so in the following chapter.

781 The Victorian Department of Justice took a similar approach in its 2013 review of defensive homicide. The Department considered that its small sample of defensive homicide cases and the low rate of such cases (such that it would take too long to obtain a “potentially significant” sample) meant it was “important to consider defensive homicide both in policy terms as well as how it applies when men kill”: Victoria Department of Justice, above n 770, at viii. To the extent our case review also precludes definite conclusions – because of both its size and the uncertainties and limitations we have identified – we think such a two-pronged approach is helpful.

Partial defences are a controversial aspect of the criminal law. Even in countries that no longer have mandatory murder sentencing, questions remain over the role they should play in determining culpability for homicide. The purpose of this chapter is to consider both in- principle arguments for and against partial defences, and the merits of some particular formulations, with reference to victims of family violence.

However, we do not have a remit or the time to look at the broader question of whether a partial defence is warranted for other defendants. This presents a dilemma. As a number of submitters identified, considering partial defences in terms of a single group has the potential to lead to unintended consequences and unfairly differential treatment.⁷⁸²To the extent required to assess possible consequences and potential unfairness, this chapter includes some comment on the application of partial defences to other defendants. Such comments are, however, necessarily brief. Proper consideration of these wider issues would best be undertaken as part of a first- principles review of the law relating to homicide, which would be a significant undertaking.

Submitters were fairly evenly split on the merits of a partial defence, but most qualified their support or opposition by noting the difficulties associated with formulating such a defence. Perhaps unsurprisingly, the division of views among members of our expert panel largely mirrored those of submitters; some considered a partial defence had merit, while others were concerned such a defence could raise problems and lead to unintended consequences.

The Family Violence Death Review Committee, the Criminal Bar Association, the Auckland District Law Society and two academics said that, with no partial defence to guide plea negotiations or the presentation of evidence at trial, defendants are subject to prosecutorial discretion and face considerable jeopardy if they defend a murder charge. The possibility of acquittal goes hand in hand with the risk of conviction for murder, with no middle ground unless intent is not proved by the prosecution. This has implications for sentencing. Juries, too, face binary decisions, and while they may “informally” recognise reduced culpability with manslaughter verdicts, such verdicts should have a foundation in law.

One submitter argued that a partial defence is more likely than discretionary elements of the justice system, such as prosecutorial decision making, to result in transparent and even-handed treatment of defendants. Another thought a partial defence may reduce the number of murder charges and thereby neutralise some of the jeopardy defendants currently face.

782 This point has been made in other jurisdictions, too. See, for example, Law Reform Commission of Western Australia Review of the Law of Homicide: Final Report (Project 97, September 2007) at 289; and Victoria Department of Justice Defensive Homicide: Proposals for Legislative Reform – Consultation Paper (September 2013) at 11.

Understanding family violence: Reforming the criminal law relating to homicide 137

The “messaging” potential of a partial defence was identified as a practical benefit. When a jury accepts a partial defence and convicts of manslaughter, it is argued, that sends a signal about reduced culpability, which helps the judge set the appropriate sentence and helps the public understand and accept the verdict and any leniency in sentencing.

Two academic submitters argued that, as a matter of legal theory, sentencing determines punishment, but culpability is determined through verdict. For some intentional homicides, a murder verdict will not accurately reflect an offender’s culpability.

One person we spoke with thought the empirical case for a partial defence for victims of family violence is not made out and that this was critical to whether there is a case for reform. This consultee thought education may enhance sentencing practice, such that a new defence would be a “sledgehammer to crack a nut” and noted that most jurisdictions that have partial defences also have mandatory sentencing for murder.

On a practical level (the New Zealand Law Society, Aviva and Women’s Refuge), several organisations noted there is a risk the “wrong” defendants will rely on any partial defence. Others were concerned a partial defence might work against victims of family violence by detracting from self-defence and promoting compromise pleas or verdicts or, if specific to victims of family violence, “normalising” these homicides of abusers as manslaughter.

A number of submitters and people we spoke with, including the New Zealand Law Society, thought the proper way to recognise reduced culpability is through sentencing. One noted that the rest of the criminal law functions without partial defences, and it is anomalous for murder to be treated differently.

The arguments in favour of partial defences, while inter-related, generally fall into one of the following categories:
^._{They promote
fair labelling.}
^._{They enhance the role of the jury in
making value judgements.}
^._{They have a positive impact
on charging practices.}
^._{They have a positive impact on
sentencing.}

We assess these arguments, in turn, below.

In respect of homicide, this argument rests, at its most fundamental, on the premise that, for some intentional killings, “murder” is not the right ascription and the law should permit such killings to be categorised as manslaughter.⁷⁸³This argument does not focus on the practical effects of a murder conviction but rather on how certain killings should be regarded and categorised.

James Chalmers and Fiona Leverick have argued that labels applied by the law may be important both to describe and to differentiate criminal conduct.⁷⁸⁴Some differentiations, like the

783 On fair labelling generally, see Andrew Ashworth and Jeremy Horder Principles of Criminal Law (7th ed, Oxford University Press, Oxford,

784 James Chalmers and Fiona Leverick “Fair Labelling in Criminal Law” (2008) 71 MLR 217 at 220–222.

distinction between murder and manslaughter, are themselves descriptive, but ideally descriptive labels should be “intelligible on a freestanding basis”.⁷⁸⁵Labelling may be important in the interests of fairness to offenders and to communicate with offenders, victims, the public and agencies within and outside the criminal justice system.⁷⁸⁶

Labelling is relevant across the criminal law, but it is often invoked in connection with homicide, perhaps because of the particular stigma and condemnation of a “murder” conviction. Andrew Ashworth has said that “the distinction between murder and manslaughter is significant in terms of its declaratory meaning and the stigma for the offender” and that partial defences may be warranted because juries and the public expect significant forms of mitigation to reflect that distinction.⁷⁸⁷“Murder”, on this view, should be reserved for and applied to only the “most heinous” killings.⁷⁸⁸

While labelling is important and fair labelling is in itself unobjectionable,⁷⁸⁹what amounts to a fair label is complex. Any system of offence categorisation will involve competing considerations and some compromise,⁷⁹⁰and the significance of labels may depend on the audience.⁷⁹¹In the context of homicide, it should not be presumed that the public necessarily looks more benignly on people convicted of manslaughter. The Law Commission said in 2007 that:⁷⁹²

... it is in fact arguable that the stigma attached to any given homicide varies depending on the circumstances of the case as much as the name of the crime. For example, a drunk driver who crashes and kills the occupants of another car will be convicted of manslaughter, which reflects lack of criminal intent, but not the public abhorrence of this kind of crime; whereas an elderly spouse who kills his failing partner, by consent or believing that it is in her best interests, is dubbed a murderer. We consider that the argument might plausibly be made that some murders (e.g. mercy killing) may be more sympathetically regarded by society than some instances of manslaughter that are widely regarded as particularly abhorrent.

The way the public understands a sentence in a given case is likely to be more complex than the visceral connotations of the offence label viewed in isolation. We are not aware of any particular public outcry at the sentence imposed by the Court of Appeal in R v Wihongi, for example, even if the case was remarked on.⁷⁹³

The complexity involved in analysing what is a fair label is readily apparent in the case of murder and manslaughter when the question of partial defences arises. Generally, murder describes intentional killings, and manslaughter describes unintentional killings.⁷⁹⁴Partial defences carve out exceptions to that distinction based on mitigating circumstances and confer

785 A description of a homicide as manslaughter, Chalmers and Leverick note, at 222, “draws explanatory value from the fact that it has in some way been differentiated from murder”.

787 Andrew Ashworth “Reforming the Law of Murder” [1990] Crim L Rev 75 at 83.

789 Glanville Williams “Convictions and Fair Labelling” (1983) 42 CLJ 85 at 86; and Chalmers and Leverick, above n 784, at 246.

790 Chalmers and Leverick note, for example, that a balance must be struck between particularism and breadth, and that some categorisations of criminal offending may be driven by pragmatic considerations. An example is the greater likelihood of securing convictions for particularised homicide offences (like dangerous driving causing death and corporate killing) where juries are reluctant to convict of manslaughter for the same conduct): Chalmers and Leverick, above n 784, at 239–240.

791 Chalmers and Leverick suggest, at 237–239, that for people with no particular legal knowledge, the descriptive aspect of a label is important, whereas for people within the justice system (like prosecutors and judges), it is the differentiation of offences and offenders that matters, and pure descriptions are secondary.

792 Law Commission The Partial Defence of Provocation (NZLC R98, 2007) at 53 (footnotes omitted).

793 The Commission made a similar observation in the 2007 Report, at 52–53, about the sentence imposed in the euthanasia case of R v Law (2002)

794 Law Commission, above n 792, at 53. See also Victorian Law Reform Commission Defences to Homicide: Final Report (2004) at 9; and Law

Understanding family violence: Reforming the criminal law relating to homicide 139

a generic manslaughter label despite the presence of murderous intent. Partial defences are therefore a blunt tool for labelling purposes.⁷⁹⁵As the Law Reform Commission of Canada (since disestablished) said:⁷⁹⁶

[E]ven if ‘murder’ seems an inappropriate term for killing under provocation, ‘manslaughter’ is surely (with all due respect to the common law) as singularly inappropriate a term for killing with intent (which killing under provocation is).

The Law Commission of England and Wales noted that deployment of fair labelling as a rationale for partial defences may conflate questions of offence categorisation and punishment⁷⁹⁷and that, when it comes to partial defences, fair labelling is secondary to the “sentence mitigation principle”.⁷⁹⁸That Commission has said:⁷⁹⁹

The secondary importance of labelling considerations explains, in part, why partial defences do not reduce other crimes that carry discretionary maximum sentences, such as attempted murder, to lesser included offences, such as, in appropriate cases, wounding with intent to do grievous bodily harm. We found it significant that those who thought labelling considerations ought to have played more of a role in structuring our proposals for partial defences to homicide, did not extend the logic of their arguments to those homicide-related crimes that have lesser included offences.

While labels may be important, therefore, it does not follow that fair labelling is well served by partial defences. It may be better served by a separate offence, tailored to a subset of homicides, but even this will not get around the problem of arbitrariness. If there are many circumstances that may mitigate culpability for homicide, should they not all be able to be taken into account in a meaningful way? This is where, as we discuss below, sentencing seems to have a procedural advantage.

It has been argued that, when a jury returns a verdict, it is not deciding an abstract question but rather how a person ought to be held responsible. This involves both legal and moral considerations so that, in some cases:⁸⁰⁰

... it may be that the greater offence is made out in law, but the jury feels that it is morally inappropriate to convict. Here, the jury will be tempted to convict of a lesser offence or, perhaps more likely, to acquit “perversely.” This tendency will be seen by some as an example of the jury’s usurping the function of the judge, and by others as the jury’s taking steps to nullify oppressive law.

To the extent jury nullification may sometimes be a reality, some consider the law should squarely accommodate these kinds of moral assessments rather than force juries to make binary choices between murder and acquittal.⁸⁰¹Martin Wasik says there is “something to be said for allowing the jury the option of a lesser offence in some cases rather than forcing them to

795 Law Commission, above n 792, at 54. Writing in the English context, Nicola Lacey suggests that, if the mandatory penalty were abolished in that jurisdiction, “one would have to ask hard questions about the real importance of labelling homicides as ‘murder’ and ‘manslaughter’, and about the relationship between labelling and grading ... If labelling is really important, one might argue that we should then reconstruct voluntary manslaughter in relation to a discrete set of labels: provoked killing; killing under duress; killing in panic; mercy killing and so on.”: Nicola Lacey “Partial Defences to Homicide: Questions of Power and Principle in Imperfect and Less Imperfect Worlds ...” in Andrew Ashworth and Barry Mitchell (eds) Rethinking English Homicide Law (Oxford University Press, Oxford, 2000) 107 at 130.

796 Law Reform Commission of Canada Homicide (Working Paper No 33, 1984) at 74. See also Law Reform Commission of Western Australia, above n 782, at 218.

797 See Law Commission of England and Wales Murder, Manslaughter and Infanticide (Law Com No 304, 2006) at 9. Where the mandatory penalty subsists, the Commission noted, at 48, “[m]atters of verdict and sentence are effectively fused in murder cases. Partial defences affect the verdict of murder, and only that verdict, because a verdict of murder is the only one that carries in its wake a mandatory sentence of such gravity (life imprisonment).”

800 Martin Wasik “Partial Excuses in the Criminal Law” (1982) 45 MLR 516 at 518–520 (footnotes omitted).

801 At 520. See also Ashworth and Horder, above n 783, at 251. Also see the discussion above at paragraphs [9.32]–[9.38].

choose between [conviction for the offence charged and acquittal] when neither reflects their true finding”, which might “[reduce] the temptation for the jury to acquit perversely in cases where the strict application of the law would lead to a conviction for a morally inappropriate offence”.⁸⁰²On this view, partial defences may be desirable not just for individual defendants but in the overall interests of justice. Lord Bingham has said:⁸⁰³

The interests of justice are not served if a defendant who has committed a lesser offence is either convicted of a greater offence, exposing him to greater punishment than his crime deserves, or acquitted altogether, enabling him to escape the measure of punishment which his crime deserves. The objective must be that defendants are neither over-convicted nor under-convicted, nor acquitted when they have committed a lesser offence of the type charged.

There is a risk of a “perverse” outcome in any case if a jury acquits on grounds of sympathy rather than law, and a middle-ground offence could avoid perversity in many cases. The risk of perverse outcomes is, in other words, not unique to homicide. Even for homicide, partial defences can address this risk only to the extent they apply. There will not be a partial defence for every potentially mitigating circumstance.

Partial defences may also lead to a perversity that disadvantages defendants in the form of compromise verdicts. In a trial situation, if a jury cannot agree whether a defendant should be acquitted on the basis of self-defence, for example, a partial defence may provide an attractive pathway to a middle-ground manslaughter verdict.⁸⁰⁴On the face of it, this risk is greatest with excessive self-defence, as we discuss below, but it could attach to any partial defence, and compromise outcomes may also arise in plea negotiations.⁸⁰⁵In the present context, where self-defence may be the appropriate defence in many cases,⁸⁰⁶an increased risk of compromise verdicts and pleas would be a particular concern.

A further element of this argument is that significant questions of mitigation should be left to juries. It is argued this can legitimise lenient sentences since there will be community endorsement of the relevant mitigating circumstances.⁸⁰⁷A jury decision on mitigation may also assist the sentencing judge by sending a clear message on the jury’s view of culpability, thus promoting sentencing consistency and fairness.⁸⁰⁸

One difficulty with arguments about the acceptability of consequences that follow convictions (like sentencing) is that they involve speculation. Somewhat like fair labelling, there is intuitive force in the idea a jury verdict may render a sentence more comprehensible or legitimate, but matters are probably much more complex.

Further, although in jury trial cases factual questions are mainly for the jury, judges may still make findings of fact for sentencing. A guilty verdict tells the judge the jury found the elements of the offence proved but not how it reached that conclusion. Thus, while a judge “must accept

804 Victorian Law Reform Commission, above n 794, at 102–103; Michelle Edgely and Elena Marchetti “Women Who Kill Their Abusers: How Queensland’s New Abusive Domestic Relationships Defence Continues to Ignore Reality” (2011) 13 Flinders LJ 125 at 141; Select Committee on the Partial Defence of Provocation The partial defence of provocation (New South Wales Legislative Council 2013) at 61; and Law Commission of England and Wales Partial Defences to Murder (Law Com No 290, August 2004) at 79.

805 Elizabeth Sheehy, Julie Stubbs and Julia Tolmie “Defences to Homicide for Battered Women: A Comparative Analysis of Laws in Australia, Canada and New Zealand” [2012] SydLawRw 22; (2012) 34 Syd L Rev 467 at 492. See also Victoria Department of Justice, above n 782, at 27 and 30. The Department of Justice considered defensive homicide may have been “operating to distort the role of self-defence” and detracting from a proper focus on self-defence, which, the Department considered, “should be the primary focus” in the context of defensively motivated homicides by victims of family violence.

806 Family Violence Death Review Committee Fourth Annual Report: January 2013 to December 2013 (Health Quality & Safety Commission, June

Understanding family violence: Reforming the criminal law relating to homicide 141

as proved all facts, express or implied, that are essential to a plea of guilty or a finding of guilt”,⁸⁰⁹there may be other matters still to make findings on. An example in the cases we have reviewed is the circumstances surrounding the homicide. Was the offender threatened before he or she inflicted the fatal injury? Were they frightened? Angry?

Even if a manslaughter verdict is returned after a defendant has relied on a partial defence, the judge will have to infer whether the verdict reflected acceptance of the defence or a finding of lack of intent,⁸¹⁰and as the Commission noted in 2007, if a partial defence is run and rejected, “little if anything is articulated about the way in which the issues have been dealt with”⁸¹¹– to the public or the sentencing judge.

Some argue that partial defences can encourage better charging practice. It is argued that, if the prosecution is guided by a partial defence when making decisions about when a charge of manslaughter should be preferred to murder, there would be less risk of overcharging. Defendants would have more leverage in plea negotiations and may feel safer defending a murder charge knowing their defence need not be “all or nothing”.⁸¹²There may also be positive public resource implications if less money were expended on murder trials.⁸¹³

There is a degree of circularity here in the assumption of a problem with charging practice and the implication that defendants should have scope to negotiate a lesser charge even when the evidence is that they killed with murderous intent. In any event, as we have discussed, we do not think the evidence from our case review discloses an “overcharging” problem.

Further, just as a jury may take a middle ground verdict option if one is presented, a defendant who has the opportunity to plead guilty to a lesser offence may take it rather than proceed to trial on a murder charge despite the possibility of a better outcome (acquittal) at trial.⁸¹⁴In other words, while a middle ground option may relieve some pressure, it does not follow that the defendant will necessarily ultimately benefit.

FVDRC suggests that a partial defence is required to achieve fair sentencing outcomes for victims of family violence who are currently convicted or at risk of being convicted of murder. Even if the presumption of life imprisonment is rebutted, a sentence for murder is likely to be higher than for manslaughter – whatever the facts of the offending.⁸¹⁵The FVDRC compares sentences in cases where the conviction was for manslaughter (particularly manslaughter cases where provocation was successfully run) with those where the conviction was for murder.

There is no doubt the sentences in the two most recent murder cases (R v Wihongi and R v Rihia), where provocation was not run or not available,⁸¹⁶are longer than those imposed in the manslaughter cases, but it does not follow that this points to a need for a partial defence.

816 As we noted in Chapter 9, provocation was in law available, but not run, in R v Wihongi, but the offending in R v Rihia post-dated repeal and so the defendant in that case could have sought to rely on the partial defence. See paragraph [9.48]. See also R v Wihongi HC Napier CRI-2009-041-2096, 30 August 2010 [Wihongi (HC)]; R v Wihongi [2011] NZCA 592, [2012] 1 NZLR 775 [Wihongi (CA)]; R v Wihongi [2012] NZSC 12 [Wihongi (SC)]; and R v Rihia [2012] NZHC 2720.

As we discussed in Chapter 9, it cannot be assumed there is no material difference in the culpability or risk between offenders convicted of murder and those convicted of manslaughter or that the murder convictions are in themselves evidence something went wrong. Nor is it valid to argue that the murder sentences were too high because they were longer than sentences in manslaughter cases where provocation was successfully run. This presumes, rather than illuminates, the merit of a partial defence and does not engage with the policy behind the repeal of provocation. In its 2007 Report, the Law Commission said,⁸¹⁷in a passage cited with approval in R v Hamidzadeh,⁸¹⁸that:

... if provocation is repealed on the policy basis that the defendants who rely upon it are not inherently more deserving of favourable treatment than many others who are presently convicted of murder, then it would make no sense to endorse and take steps to ensure an ongoing lower tariff simply for provocation. It may be that a more flexible approach to sentencing for murder ought to be taken to allow better recognition of the wide range of mitigating factors (including provocation) that can be present in cases of intentional killing, but that is a different issue.

Even if there were an unwarranted discrepancy between sentences for murder and sentences for manslaughter in this area, it would not follow that the structure of the law was the problem.⁸¹⁹If the issue is sentencing practice, it should be considered by looking at whether the law is structured, and the criminal justice system has the resources, for sentencing to be approached in a way that is flexible, responsive and fair. We discuss issues around sentencing in Chapter 11.

In assessing the arguments in favour of partial defences we have, necessarily, addressed many of the counter-arguments already, but there are four stand-alone arguments against partial defences that warrant mention.

As highlighted throughout this part, with few exceptions, partial defences go hand in hand with mandatory sentencing for murder. Law reform bodies here and overseas have noted the mandatory penalty’s historical co-extensiveness with partial defences, and jurisdictions where mandatory sentencing has been abolished have, in the main, found that to point against partial defences.⁸²⁰Two law reform bodies to have recommended against abolition of partial defences, the Law Commission of England and Wales and the Queensland Law Reform Commission, were

820 The co-extensiveness of partial defences and the mandatory penalty for murder was noted in Law Commission of England and Wales, above n

797, at 48–49; Law Reform Commission of Ireland Defences in Criminal Law (LRC 95, 2009) at 112; Select Committee on the Partial Defence of Provocation, above n 804, at ch 2; Tasmania Law Reform Institute Review of the Law Relating to Self-defence (Final Report No 20, 2015), at 53 and 71; Queensland Law Reform Commission A review of the excuse of accident and the defence of provocation (Report No 64, September

2008), at 500; Victorian Law Reform Commission, above n 794, at xx, 11, and 241; Victoria Department of Justice, above n 782, at 17; and Law Reform Commission of Western Australia, above n 782, at 5–6. Among jurisdictions to have abolished mandatory sentencing but retained (and in some cases reformed) or reintroduced partial defences (New South Wales, Western Australia and the Australian Capital Territory), New South Wales and Western Australia have identified particular reasons for doing so. The New South Wales Select Committee on the Partial Defence of Provocation accepted there was a case for abolition but, unable to reach a consensus and without sufficient information to enable a clear recommendation on how self-defence could be reformed to accommodate victims of family violence, did not recommend abolition but did recommend a review of homicide and defences after five years and that recommendation has been accepted (Select Committee on the Partial Defence of Provocation, above n 804, at 87 and 209). In Western Australia, the Law Reform Commission noted that mandatory murder sentencing “has provided one of the strongest justifications for the partial defence of provocation”. The Commission considered no partial defences should be retained or introduced “unless the circumstances giving rise to the defence always demonstrate reduced culpability” but found that excessive self-defence met this test (Law Reform Commission of Western Australia, above n 782, at ch 4.)

Understanding family violence: Reforming the criminal law relating to homicide 143

working on the basis that the mandatory sentence for murder was not open for reform, and that was highly relevant to their recommendations.⁸²¹

For other crimes, if the elements of an offence are proved, that is sufficient to convict, and differences in the seriousness of particular cases are addressed at sentencing. Culpable homicide is the only kind of crime where, despite proof of the elements of murder, a person may still be convicted of another less serious offence (manslaughter or, in relevant cases, infanticide) if a partial defence applies. These are cases where the defendant intentionally caused death but extenuating circumstances are seen to justify a description of culpability or sentencing consequences other than those that follow a murder conviction.⁸²²Thus, the Law Commission of England and Wales has described the idea of a partial defence as “something of a misnomer”, better understood as “the way that the law has created space for discretion in sentencing in murder cases”.⁸²³Despite this, some suggest a partial defence model could be applied beyond homicide,⁸²⁴while others resist the idea partial defences are anomalous just because they are unique to the law of homicide.⁸²⁵

_{Partial defences are not well suited to taking account of mitigating
circumstances}

A compelling argument, which has considerable practical significance and was reviewed at length by this Commission in 2007,⁸²⁶is that partial defences are a second-best way to take account of mitigating circumstances.

One reason for this is that a wide range of circumstances may mitigate a defendant’s culpability, and unless it is suggested there should be a partial defence for every such circumstance, there will inevitably be arbitrariness in which mitigating factors are singled out for recognition via a partial defence.⁸²⁷

Partial defences also reduce complex issues to binary inquiries. They require a jury to decide whether a defendant fits the criteria for mitigation as a “yes or no” question. This may mean mitigating circumstances are excluded from consideration at sentencing, if a jury considers the defendant does not quite come within the partial defence,⁸²⁸or that “undesirable constraints” are placed on the presentation of evidence at trial,⁸²⁹partly because trials are adversarial and set up to “generate a ‘winner’ and a ‘loser’”.⁸³⁰Issues that turn on matters of degree are therefore not well served,⁸³¹and there is a strong argument that it is preferable for evidence of mitigating

821 Law Commission of England and Wales, above n 797, at 43–44; and Queensland Law Reform Commission, above n 820, at 9–10.

824 Jeremy Horder Excusing Crime (Oxford University Press, Oxford, 2004) at 143–146; and Lacey, above n 795, at 120. But compare Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001) at 41: “The argument for jury participation in determining levels of culpability should logically extend to all crimes and not be confined to murder. For good reason this has never been suggested. Instead, the task of crafting penalty to fit blameworthiness has long been the daily diet of judges.” The possibility of an expanded partial defence model is obviously beyond our remit, although we would note it is relevant, in this connection, that homicides are a small number of all violent offences, including those committed by victims of family violence responding to abuse.

825 See for example Victor Tadros “The Homicide Ladder” (2006) 69 MLR 601 at 615.

829 Law Commission of England and Wales, above n 797, at 49. See also, discussing the partial defence of diminished responsibility, Jeremy Horder

Homicide and the Politics of Law Reform (Oxford University Press, Oxford, 2012) at 233.

831 Horder, above n 829, at 233–234; Law Commission of England and Wales, above n 797, at 49; and Victorian Law Reform Commission, above n

circumstances to be dealt with through the more inquisitorial and holistic sentencing process.⁸³²

By making evidence of diminished responsibility relevant to verdict, when (as in cases of second degree murder) it could simply be made relevant to sentence, one would needlessly force experts to distort the relevance of their evidence. Likewise, by making evidence of provocation relevant to verdict, one forces the jury to work with only a partial picture of the context in which the provocation was alleged to have been given, a partial picture largely provided (usually uncontested) by D. These drawbacks may be something that we must live with when the verdict would otherwise entail the passing of an inappropriate mandatory sentence but they should not be tolerated outside that context.

In addition to the risk of artificially rigid presentation of evidence, partial defences can have other undesirable effects. One, which we have noted, is the risk of compromise pleas and verdicts. Another is complexity for juries, which, in homicide cases, must in any event grapple with difficult concepts and evidence.⁸³⁴

In the present context, two additional potential perversities should be noted. First, partial defences can undermine the operation of self-defence. This was noted by the Victorian Department of Justice when it reviewed the operation, and recommended repeal, of defensive homicide in that jurisdiction. The risk has also been acknowledged by other law reform bodies.⁸³⁵Second, partial defences can lead to unintended consequences, particularly the excusing of conduct by “undeserving” defendants. Victoria’s defensive homicide is a clear example of this in practice – it was primarily used by violent men.⁸³⁶

Bearing in mind the in-principle arguments for and against partial defences, we have considered what a partial defence aimed at victims of family violence would look like. This practical enquiry informs our view on whether a partial defence is justified.

We have drawn on the Commission’s previous work, and in particular Some Criminal Defences with Particular Reference to Battered Defendants,⁸³⁷and overseas law reform work, but there is no standard approach in this area. Other jurisdictions’ approaches vary depending on policy choices about what conduct is partially excusable and local context, including sentencing law and procedure, and whether and what partial defences already exist.

Experience in other jurisdictions suggests that the essential focus of partial defences can be broken down into three categories. Each emphasises different elements, although there can be significant overlap:
^._{Defence-based partial defences
– normally requiring an honest belief that the
defendant’s}

_{actions were necessary to defend or preserve him or herself or
another.}
^._{Provocation-based partial defences
– requiring a loss of control by the defendant,}

triggered by provocation sufficient to deprive a person of ordinary tolerance, in similar circumstances, of the power of self-control.

835 Law Reform Commission of Western Australia, above n 782, at 294; and Tasmania Law Reform Institute, above n 820, at 71.

Understanding family violence: Reforming the criminal law relating to homicide 145

^.Diminished capacity-based partial defences – requiring an abnormality of mental functioning that impaired the defendant’s capacity to understand events, to judge whether their actions are right or wrong, or to exercise self-control.

We have also considered a fourth option of a trauma-based partial defence, which would focus on the trauma suffered by a defendant and their response to it.

Below, we discuss each of the four types of partial defence, before commenting briefly on the difficulties presented by the question of whether any such defence should be general or limited to victims of family violence.

Defence-based partial defences operate alongside self-defence to provide a middle tier of culpability where the defendant satisfies the subjective, but not the objective, element of self- defence, which is usually a requirement that the use of force must be reasonable.

Among the submitters supportive of a defence-based partial defence were the Public Defence Service, the Auckland Coalition for the Safety of Women and Children and an academic. Others considered it preferable to reform self-defence to better accommodate family violence victims.

Below, we discuss the merits of two examples of defence-based partial defences: excessive self-defence and killing for preservation in an abusive domestic relationship (killing for preservation). The specific offence of defensive homicide in Victoria (now repealed) has a similar focus to these partial defences. Given that it is a stand-alone offence, however, we have chosen to analyse the merits of this option separately at paragraph [10.98] below.

In its 2001 Report, the Law Commission considered the merits of excessive self-defence, and while it favoured this partial defence over others, ultimately, it preferred a sentencing discretion for murder.⁸³⁸The Commission considered a partial defence of excessive self-defence might be worded as follows:⁸³⁹

It is a partial defence to a charge of murder (reducing the offence to manslaughter) if, in the defence of himself [or herself] or another, a person uses more force than it is reasonable to use in the circumstances as he [or she] believes them to be.

As the Commission recognised, this defence would apply where a defendant intended to do something “lawful within limits”, unlike provocation and diminished responsibility, which apply where a defendant does something unlawful.⁸⁴⁰A defendant who honestly believes they need to defend themselves but is mistaken about the level of force required to counter the threat is, arguably, morally less culpable than people who kill intentionally and should not be labelled a murderer.⁸⁴¹This convinced both the Victorian Law Reform Commission (VLRC) and the Law Reform Commission of Western Australia to recommend versions of this partial defence.⁸⁴²

In respect of victims of family violence, excessive self-defence may be more appropriate than a provocation or diminished capacity-type defence, as there is no need for a loss of control and the defendant’s actions are not treated as if they arise from a mental condition.⁸⁴³Not all victims

842 Victorian Law Reform Commission, above n 794, at 101; and Law Reform Commission of Western Australia, above n 782, at 181–182.

843 Victorian Law Reform Commission, above n 794, at 102; and Law Commission, above n 824, at 25.

of family violence who kill their abusers will be acting defensively, however; some will be retaliating out of anger for what has been done to them. Excessive self-defence would therefore not apply in all cases. In fact, it is not clear that any of the four murder convictions identified in our case review would have benefited from a defence-based partial defence.⁸⁴⁴

There are problems with excessive self-defence, which have led courts in New Zealand,⁸⁴⁵

England,⁸⁴⁶Canada⁸⁴⁷and Australia⁸⁴⁸to disavow the partial defence at common law. In New

Zealand, the Court of Appeal recently considered excessive self-defence in McNaughton v R.⁸⁴⁹

Concluding that no such partial defence exists at common law, the Court noted there was “little statutory support” for excessive self-defence outside New South Wales,⁸⁵⁰South Australia,⁸⁵¹
Western Australia⁸⁵²and a few American states (North Carolina, Massachusetts). The Court quoted the Supreme Court of Canada, which has stated, in R v Faid, that excessive self- defence:⁸⁵³

... lacks any recognizable basis in principle, would require prolix and complicated jury charges and would encourage juries to reach compromise verdicts to the prejudice of either the accused or the Crown.

In Australia, the Model Criminal Code Officers Committee rejected the reintroduction of excessive self-defence on the ground that it is inherently vague and would bring unnecessary complexity to the law,⁸⁵⁴although the New Zealand Law Commission was not convinced excessive self-defence would involve the same complexities in this jurisdiction. Because excessive self-defence would sit alongside the complete defence of self-defence, it would not involve new concepts and would fit naturally into jury directions on self-defence.⁸⁵⁵

While the wording of the suggested provision in paragraph [10.51] above seems straightforward, there is an issue around how the jury decides what is “reasonable force” when determining whether a defendant should be acquitted or convicted of manslaughter (assuming the subjective limbs are met). Self-defence does not turn on fine distinctions. As discussed in Chapter 5, the courts do not apply a strict proportionality test, recognising that, when responding to a threat, “a person defending himself cannot weigh to a nicety the exact measure of his necessary defensive action”.⁸⁵⁶If our recommendations in Part 2 are accepted and it is clarified that self- defence is not necessarily excluded where the degree of force used in defending a threat exceeds the degree of threatened force, it would be difficult for the jury to distinguish between self-

844 In R v Neale [2010] NZCA 167, the defendant claimed self-defence, but the facts do not appear to have supported it. The defendant stabbed the deceased nine times as he got out of the shower. In R v Reti HC Whangarei CRI-2007-027-2103, 9 December 2008 [Reti (HC)]; R v Reti [2009] NZCA 271, excessive self-defence could have led to a different outcome on the facts, as the defendant stabbed the deceased after he kicked her in the stomach. However, she had previously stabbed him several hours earlier, and the jury rejected her claim of provocation. In R v Wihongi, above n 816, the defendant stabbed the deceased after an argument and while he was leaving the property. In R v Rihia [2012] NZHC 2720, the defendant stabbed the deceased during an argument and while the deceased was lying on the couch, intoxicated and after the deceased had thrown a speaker at his head.

853 McNaughton v R, above n 845, at [64], quoting R v Faid, above n 847, at 271.

854 Model Criminal Code Officers Committee Model Criminal Code: Fatal Offences Against the Person (1998) at 107–113.

Understanding family violence: Reforming the criminal law relating to homicide 147

defence and excessive self-defence. This could lead to compromise verdicts, with juries opting for the “middle ground” of manslaughter.⁸⁵⁷

Queensland’s partial defence of killing for preservation applies only to victims of family violence where the defendant believes their actions are necessary for their “preservation from death or grievous bodily harm”.⁸⁵⁸The defence also requires reasonable grounds for the defendant’s subjective belief. In that respect, it is narrower than traditional formulations of excessive self- defence.

The Queensland defence has been criticised on the basis that cases captured by the defence would, in other jurisdictions, result in complete acquittals on the basis of self-defence.⁸⁵⁹In Queensland, self-defence is available only where a person is facing an immediate threat of harm or is being assaulted.⁸⁶⁰Queensland also retains a mandatory sentence for murder, and the defence was introduced in response to concerns raised by the Queensland Law Reform Commission that, given the unsuitability of the partial defence of provocation, the circumstances of victims of family violence were unable to be considered in mitigation of the mandatory life penalty for murder.⁸⁶¹Rather than repeal provocation and the mandatory sentence (as in New Zealand), a separate partial defence was introduced to fill the gap where a defendant is responding to a threat of harm that is not immediate. Queensland’s legal context is, therefore, unique and not a wholly helpful comparator.

Provocation-based partial defences do not require a defendant to be motivated by a need to protect him or herself or another from a threat. They apply where a defendant has lost control in the face of provocation.

Among submitters, the Criminal Bar Association and the FVDRC favoured a partial defence based on loss of control or something akin to New South Wales’s “extreme provocation”, which we discuss below. The FVDRC recognised this would capture circumstances where the victim is retaliating but the behaviour is not defensive (although the FVDRC had a number of concerns about the detail of the New South Wales provision). Other submitters, including the New Zealand Law Society and the Public Defence Service, opposed any reintroduction of a provocation-based defence.

857 S Kadish The Model Penal Code’s Provocation Proposal and its Reception in the State Legislatures and Courts of the United States of America, with Comments Relating to the Partial Defenses of Diminished Responsibility and Imperfect Self-Defence (February 2004). See Appendix F of Law Commission of England and Wales, above n 804.

860 See discussion in Tasmania Law Reform Institute, above n 820, at 70; and Victorian Department of Justice Defensive homicide: Review of the offence of defensive homicide: Discussion paper (August 2010) at 50. See also Queensland Law Reform Commission, above n 820, at 313.

861 Queensland Law Reform Commission, above n 820, at 489. See also discussion at paragraph [10.36] above, and the footnotes therein.

Provocation-based partial defences vary from jurisdiction to jurisdiction, but a traditional formulation normally requires that:

the defendant lost control;

the loss of control was caused by provocation; and

the provocation would have been sufficient to deprive a person with the same characteristics as the defendant, but a normal degree of tolerance or the ordinary person’s power of self-control, of the power of self-control.

In New Zealand, provocation was relied on in five out of 11 cases involving victims of family violence between 2001 and 2009. Three of those cases resulted in a manslaughter verdict, which could be attributed to an accepted claim of provocation.⁸⁶²In two cases, the defendant was convicted of murder. Since provocation was repealed, there has been one conviction for murder, which followed a guilty plea,⁸⁶³and the facts of that case may have fit within a provocation- based defence.⁸⁶⁴Accordingly, while the number of cases we have reviewed is too small to draw conclusions, the circumstances in which victims of family violence kill their abusers may, in some circumstances, come within the elements of a provocation-based defence.

Provocation has historically been problematic, however, and was repealed in New Zealand in

2009.⁸⁶⁵We do not attempt to explore the problems with the defence, which was considered in detail by the Law Commission in 2007.⁸⁶⁶We note, however, that in the context of family violence, provocation is regarded as problematic for two reasons. First, the traditional version of the defence requires a sudden loss of control, but it is argued this does not reflect the reality of the experiences of victims of family violence, who may respond to prolonged abuse some time after the provocation has been endured.⁸⁶⁷Second, provocation can operate to excuse perpetrators of family violence who kill their victims in unexceptional circumstances, such as relationship break-downs.⁸⁶⁸

While several jurisdictions have abolished provocation, others have reformed it in an attempt to address these problems. Reforms have focused on guarding against inappropriate reliance on provocation by excluding certain conduct from the scope of the defence⁸⁶⁹and making the

862 The jury could, however, also have found that the defendant lacked the necessary intent for murder, which would also render a manslaughter verdict.

864 See the discussion of Ms Rihia’s case and the partial defence of provocation in Chapter 9 at paragraphs 9.49–9.51.

865 As discussed in Chapter 4. In Australia, provocation was also abolished in Tasmania in 2003, in Victoria in 2005 and in Western Australia in

2008. See Criminal Code Amendment (Abolition of Defence of Provocation) Act 2003 (Tas); Crimes (Homicide) Act 2005 (Vic); and Criminal

867 Victorian Law Reform Commission, above n 794, at 58; Queensland Law Reform Commission, above n 820, at 298; Law Commission Battered

Defendants: Victims of Family Violence Who Offend (NZLC PP41, 2000) at 28–29; and Law Commission of England and Wales, above n 797, at

81. See also Barry Mitchell “Loss of Self-Control under the Coroners and Justice Act 2009: Oh No!” in Alan Reed and Michael Bohlander (eds)

Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 39 at 42.

869 Reform of provocation in England and Wales expressly excludes sexual infidelity and claims where the provocative conduct has been incited by the defendant as an excuse to use violence (Coroners and Justice Act 2009 (UK), s 55(6)). In New South Wales provocation was reformulated to “extreme provocation” and applies only when the deceased’s conduct constitutes a serious indictable offence and excludes conduct incited by the deceased in order to provide an excuse to use violence (Crimes Act 1900 (NSW), s 23(2) and (3)). In Queensland provocation was restricted to exclude conduct of the deceased to bring about the end of a relationship, except in circumstances of a “most extreme and exceptional character” (Criminal Code 1899 (QLD), s 304(3)). In Canada, amendments were introduced in 2015 to restrict provocation to circumstances where the deceased’s conduct constituted an indictable offence punishable by five or more years of imprisonment (Criminal Code RSC 1985 c C-46, s 232(2)).

Understanding family violence: Reforming the criminal law relating to homicide 149

defence more accessible to victims of family violence.⁸⁷⁰The most notable reformulation is

England and Wales’s “loss of control”, and New South Wales’s “extreme provocation”.

The partial defence of “loss of control” replaced provocation in England and Wales, following two reports of the Law Commission in that jurisdiction.⁸⁷¹Loss of control applies where:⁸⁷²

the defendant’s conduct resulted from a loss of self-control;

the loss of self-control had a qualifying trigger; and

a person of the defendant’s sex and age, with a normal degree of tolerance and self-restraint and in the circumstances of the defendant, might have reacted in the same or in a similar way.

A loss of self-control will have a “qualifying trigger” if it is attributable to:⁸⁷³

the defendant’s fear of serious violence from the deceased against the defendant or another identified person;

a thing or things done or said (or both) that:

constituted circumstances of an extremely grave character; and

caused the defendant to have a justifiable sense of being seriously wronged; or

Some commentators consider the extension to fear of serious violence is “close to redundant” for victims of family violence because of constraints on the operation of the defence and this trigger.⁸⁷⁴The most significant and controversial constraint is the requirement to prove loss of self-control, which remains despite the Law Commission of England and Wales’s recommendation that it be removed. The Commission considered the loss of control requirement was unnecessary and undesirable because the wrongful words or conduct, and not the loss of control, were the justification for the defence. The Commission also noted the criticism that a loss of control requirement “privilege[d] men’s typical reactions to provocation over women’s typical reactions” and disadvantaged women responding to family violence:⁸⁷⁵

It was clear to us that when a battered woman uses excessive force against her abusive partner only because she fears for her safety in any direct confrontation, it would be wrong to rule out her plea simply because there was no evidence of a loss of self-control.

The requirement for a loss of control may, in addition, be illogical when the defence is invoked in connection with a rational fear of serious violence.⁸⁷⁶

870 The requirement that the loss of control be “sudden” or for the provocation to occur immediately before the defendant’s use of force was removed by statute in England and Wales (Coroners and Justice Act 2009 (UK), s 54(3)); New South Wales (Crimes Act 1900 (NSW), s 23(4)); and the Australian Capital Territory (Crimes Act 1900 (ACT), s 13(2) and (4)(b)).

871 Law Commission of England and Wales, above n 804; and Law Commission of England and Wales, above n 797.

874 Catherine Elliot “A Comparative Analysis of English and French Defences to Demonstrate the Limitations of the Concept of Loss of Control” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 231 at 231.

The recommendations of the Law Commission of England and Wales were endorsed by the Select Committee established to review the partial defence of provocation in New South Wales.⁸⁷⁷The Committee agreed that the requirement for a loss of control should be abandoned, and the focus shifted to the nature of the provocative conduct:⁸⁷⁸

The Committee is concerned that the practical effect of the partial defence requiring a loss of self- control inappropriately lends itself to killings in which extreme violence is used to reduce a defendant’s culpability and, in relation to intimate partner homicides, this tends to favour male defendants who kill women, further contributing to concerns about gender bias. Conversely, the requirement to show a loss of self-control tends to disadvantage those who kill, usually women, who kill in ‘slow burn’ cases.

However, in neither jurisdiction was the recommendation to discard the requirement for a loss of control adopted. Ultimately, the concern was that removing that requirement would risk opening up the partial defence to cold-blooded killing.⁸⁷⁹The Ministry of Justice in England considered that:⁸⁸⁰

... there is ... a fundamental problem about providing a partial defence in situations where a defendant has killed while basically in full possession of his or her senses, even if he or she is frightened, other than in a situation which is complete self-defence.

Rather than remove the requirement for a loss of control, the legislation in both England and

Wales and New South Wales clarified that the loss of control does not need to be sudden.⁸⁸¹

This may not have a major impact, however, as the period of time between the provocation and the defendant’s response will remain relevant.⁸⁸²The explanatory notes to the loss of control provision state, for example, that delay could be evidence as to whether the defendant actually lost control, and the greater the delay the more likely the defendant acted out of calculated revenge.⁸⁸³It has been argued that this will continue to be problematic for defendants who use force where there is no immediate threat.⁸⁸⁴

In 2007, the New Zealand Law Commission considered the English proposal to include fear- based responses in a reformulated provocation defence. Overall, it was unconvinced this offered any advance on New Zealand’s formulation and noted labelling and conceptual issues with combining provocation and excessive self-defence.⁸⁸⁵Since then, the experience in England and Wales and New South Wales demonstrates that, while the most troublesome element of provocation-based partial defences for victims of family violence – the requirement for a sudden loss of control – is not inseparable from the defence, legislators have been unwilling to part with it given the risk the defence will then accommodate undeserving conduct.⁸⁸⁶

Partial defences premised on diminished capacity look “into the actor’s mind to see whether he [or she] should be judged by a lesser standard than that applicable to ordinary men [or

879 Department of Attorney General and Justice Reform of the Partial Defence of Provocation: Call for Submissions on the exposure draft Crimes

Amendment (Provocation) Bill 2013 (New South Wales, 2013) at 8; and Mitchell, above n 867, at 44–45.

880 Ministry of Justice Murder, Manslaughter and Infanticide (MoJ CP No 19, London 2008) at [36]. Discussed in Mitchell, above n 867, at 44.

883 Coroners and Justice Act 2009 (UK) (explanatory notes) at [337]. Discussed in Elliot, above n 874, at 232.

886 A defence that partially excuses some provoked intentional killings is also fundamentally problematic for the reason recorded in the Law Commission’s 2007 Report; it “assumes that the ordinary person, faced with a severely grave provocation, will in consequence resort to homicidal violence, when it is in fact arguable that only the most extraordinary person does this”: Law Commission, above n 792, at 42. The English reformulation does not resolve this conceptual problem.

Understanding family violence: Reforming the criminal law relating to homicide 151

women]”.⁸⁸⁷These defences do not share the objective requirement of provocation-based defences (that the provocation have a similar effect on a similar person).⁸⁸⁸They are generally intended to capture conduct falling short of insanity, where the defendant suffers from an abnormality of mental functioning that impairs their capacity to understand events, to judge whether their actions are right or wrong or to exercise self-control.⁸⁸⁹Diminished responsibility was considered, but rejected, by the VLRC and the Law Reform Commission of Western Australia.⁸⁹⁰The VLRC noted that it had been argued that introducing diminished responsibility would serve to entrench misleading stereotypes of women by attributing the homicide to a psychological disturbance rather than a defensive reaction to ongoing and severe family violence.⁸⁹¹

The few submitters who supported a partial defence of diminished responsibility, including the Auckland District Law Society, did so on the basis that it should apply generally rather than only to victims of family violence. Others questioned the appropriateness of a capacity-based defence for victims of family violence, noting it may imply some kind of mental disturbance and perpetuate gender stereotypes. The FVDRC considered this kind of defence would inappropriately focus on the defendant’s mental state rather than the social context within which they were responding. The Criminal Bar Association was not comfortable with the implication that victims of family violence have lost their ability to make reasonable judgements about their situation. The Public Defence Service was divided on the issue, noting concerns with describing defensive actions as a manifestation of some psychological disturbance but cautioning against rejecting the defence out of hand on that ground if it might avert a conviction for murder. They considered additional research may be needed to further investigate the potential of this partial defence for victims of family violence.

Diminished capacity-based partial defences are recognised in several Commonwealth jurisdictions including in England and Wales,⁸⁹²New South Wales,⁸⁹³Queensland,⁸⁹⁴the Northern Territory⁸⁹⁵and the Australian Capital Territory.⁸⁹⁶

An example is the English provision, which was reformed following two reports of the English

A person (“D”) who kills or is a party to the killing of another is not to be convicted of murder if D

887 Model Penal Code (US) § 210(3) (1985) (commentary) at 71–72, noted in Paul H Robinson “Abnormal Mental State Mitigations of Murder: The US Perspective” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 291 at 307.

888 For example, loss of control applies where “a person of [the defendant’s] sex and age, with a normal degree of tolerance and self-restraint and in the circumstances of [the defendant], might have reacted in the same or in a similar way”. See Coroners and Justice Act 2009 (UK), s 54(1)(c).

890 Victorian Law Reform Commission, above n 794, at 239; and Law Reform Commission of Western Australia, above n 782, at 259.

893 New South Wales reformed its version – “substantially impaired capacity” – in 1997. See Crimes Act 1900 (NSW), s 23A.

897 Law Commission of England and Wales, above n 804; and Law Commission of England and Wales, above n 797.

898 Coroners and Justice Act 2009 (UK), s 52 (amending Homicide Act 1957 (UK), s 2).

substantially impaired D's ability to do one or more of the things mentioned in subsection

(1A), and

_{provides an explanation for D's acts and omissions in
doing or being a party to the killing.}

to understand the nature of D's conduct;

to form a rational judgment;

_{to exercise self-control.}

...

Diminished responsibility is considered an anomaly in English law, driven by retention of the mandatory sentence for murder.⁸⁹⁹It reflects the continuum of mental illness (with insanity at one end and full competence at the other) and offers an alternative verdict for those not willing or able to raise the complete insanity defence.⁹⁰⁰It is also supported on the basis murder is not the appropriate label for someone who is mentally ill, although as we have noted, the force of the “labelling” argument was rejected by the Law Commission of England and Wales.⁹⁰¹

Diminished capacity is a difficult concept to define. This makes it problematic to apply and could lead to inconsistency.⁹⁰²Some commentators note that there is no reliable way to gauge degrees of responsibility and question the utility of psychiatric evidence on this issue.⁹⁰³For these reasons, diminished responsibility has never been part of New Zealand law. It was considered by the Crimes Consultative Committee established in 1989, but the Committee’s preference was to deal with matters of impaired responsibility at sentencing.⁹⁰⁴The Committee considered that difficulties with establishing a discrete provision would be exacerbated by complexities involved in achieving sufficiently precise wording.⁹⁰⁵

Despite conceptual problems, the partial defence has, however, been successfully relied on by victims of family violence in other jurisdictions,⁹⁰⁶and New Zealand courts have recognised that any capacity-based partial defence might be relied on by defendants who have been victims of family violence and suffered from battered woman syndrome.⁹⁰⁷

However, a partial defence of diminished responsibility is arguably problematic in this context for two reasons. First, it can entrench misleading stereotypes of women by attributing homicides of abusers to psychological disturbance or mental abnormality rather than defensive reactions or acts of desperation in response to ongoing and severe violence.⁹⁰⁸Second, it may be used by predominant aggressors in the context of intimate partner violence (for example,

899 Alan Reed and Nicola Wake “Anglo-American Perspectives on Partial Defences: Something Old, Something, Borrowed, and Something New” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 183 at 183–184.

902 Law Commission, above n 824, at 47; and Victorian Law Reform Commission, above n 794, at 239.

904 Warren Brookbanks “Partial Defences to Murder in New Zealand” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished

Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 271 at 283.

907 R v Gordon (1993) 10 CRNZ 430 (CA). A capacity-based partial defence may also be relatively well suited to young offenders. The Family Violence Death Review Committee said in its submission that self-defence is often problematic on the facts of cases where children kill abusive parents or caregivers and noted children have different capacity to evaluate risk and consider outcomes” (citing Nelson v R [2012] NZHC 3570). See also Brookbanks, above n 904, at 282–284.

908 Victorian Law Reform Commission, above n 794, at 239; and Law Commission, above n 824, at 47.

Understanding family violence: Reforming the criminal law relating to homicide 153

by affording a defence to “depressed husbands who kill their partners when they end the relationship”).⁹⁰⁹

For these reasons, the Law Commission recommended against introduction of the defence in New Zealand in 2001.⁹¹⁰When it rejected the partial defence, the VLRC considered degrees of criminal responsibility should be assessed at sentencing and that introducing diminished responsibility would conflict with the recommendation to abolish provocation.⁹¹¹The Law Reform Commission of Western Australia took a similar view, noting substantial impairment by mental abnormality does not always accurately reflect culpability or render such homicides equivalent to unintentional killings.⁹¹²

We observe, finally, that in New Zealand the Sentencing Act 2002 provides that, if “the offender has, or had at the time the offence was committed, diminished intellectual capacity or understanding”, this may be a mitigating factor.⁹¹³This applies to all offences, not just homicide, and we discuss it in Chapter 11.

In addition to the existing types of partial defence, we have considered the option of a new partial defence based on trauma that would be tailored towards victims of family violence who kill their abusers.

As we apprehend it, the rationale for a partial defence in this context is to recognise reduced culpability of victims of family violence who have been traumatised by a history of abuse. This includes, but is not limited to, defendants who overreact to a threat and defendants who act out of anger after reaching a “breaking point” as a consequence of abuse. Neither a defence-based nor a provocation-based partial defence will capture both kinds of case. Further, we agree with the FVDRC and other submitters that a diminished capacity-based defence is inappropriate, as it would assume a victim of family violence was suffering from an abnormality of mental functioning caused by a recognised medical condition. This may not always be the case. A tailored partial defence based on the trauma of family violence could better recognise and fairly capture the unique circumstances of family violence victims.

A helpful analogue is the partial defence of extreme mental or emotional disturbance in the

A homicide which would otherwise be murder [is manslaughter when it] is committed under the influence of extreme mental or emotional disturbance for which there is a reasonable explanation or excuse. The reasonableness of such explanation or excuse shall be determined from the viewpoint of a person in the actor’s situation under the circumstances as he believes them to be.

One of the features of this provision is that it requires a “reasonable explanation or excuse” only for the extreme mental or emotional disturbance under which the killing was committed and not for the killing itself. Some commentators have noted that this distinction avoids some of the difficulties with provocation’s reasonableness requirement.⁹¹⁵Because this provision does not require a loss of control or reaction to an external threat, it also avoids the problem of proving a temporal connection between an external circumstance and the defendant’s conduct. However,

910 Law Commission, above n 824, at 47–48. See also Brookbanks, above n 904, at 284.

914 Model Penal Code (US) § 210(3)(1)(b) (1985), cited and discussed in Mitchell, above n 867, at 49.

915 See, for example, Kadish, above n 857, at 273–274. See also the discussion in Robinson, above n 887, at 292–296.

this feature has also been criticised on the basis it has “permitted juries to return a manslaughter verdict in cases where the defendant claims passion because the victim left, moved the furniture out, planned a divorce, or sought a protective order”,⁹¹⁶and while juries rejected the partial defence in those cases, the point may still have force, since:⁹¹⁷

... remitting issues to the bare sympathies of the juries invites the illicit and the prejudiced, particularly troublesome when the prejudice tends to be part of the culture, as it is to some degree with sex roles and behavior.

Another distinctive feature of the MPC provision is that the objective aspect of the test (an extreme mental or emotional disturbance for which there is a reasonable explanation or excuse) is determined from the viewpoint of a person in the actor’s situation under the circumstances as he or she believes them to be. This was considered “potentially radical”⁹¹⁸(although it bears some similarity to the mixed subjective-objective test for self-defence and defence of another in section 48 of the Crimes Act 1961, which requires the defendant’s use of force to be reasonable “in the circumstances as he or she believes them to be”). It may be criticised on the basis it leaves the jury with no guidance for determining which abnormalities should and should not count and “individualises” the reasonableness standard.⁹¹⁹

The Law Commission of England and Wales closely considered the MPC’s “extreme mental or emotional disturbance” as an alternative to the concept of “loss of control”.⁹²⁰The concept received support from a significant number of submitters in that jurisdiction, including representative bodies of the legal profession and women’s groups, but was opposed by a majority of judges and academics on the basis that it was too vague.⁹²¹That Commission rejected the concept in 2004 (although it was noted as an alternative to the preferred option in the 2006
Report) on the basis it was “unduly vague and indiscriminate⁹²²and could lead to “intense legal scrutiny and, no doubt, to a number of cases in the appeal courts to determine its meaning and scope”.⁹²³The phrase has, however, formed the basis of a provocation defence in at least some states in the United States.⁹²⁴

Drawing on the MPC provision, the elements of a trauma-based defence could include that:
^._{the defendant had
been subjected to repeated serious violence;}
^._{the
defendant had reacted in a state of extreme mental or emotional disturbance
caused by}

_{the violence she or he had experienced; and}
^._{there is a reasonable explanation or excuse for the extreme mental or
emotional disturbance}

in the circumstances as the defendant believed them to be.

Such a provision would, however, raise complex issues and involve a number of policy choices, including whether the relevant violence was limited to that perpetrated by the deceased. Proper

919 Thereby departing from the intention of the partial defence of provocation that the provocation must have been so grave as to have affected most reasonable persons in the same fashion: Kadish, above n 857, at 276–277.

920 Law Commission of England and Wales, above n 797, at 82; and Law Commission of England and Wales, above n 804, at 41–44.

922 Law Commission of England and Wales, above n 797, at 82; Law Commission of England and Wales, above n 804, at 41–44.

924 Law Commission of England and Wales, above n 797, at 82; and Kadish, above n 857, at 272. When the Law Commission of England and Wales considered this in 2004, none of the 34 United States jurisdictions that had revised their criminal codes had adopted the Model Penal Code proposal wholesale, but five (Arizona, Arkansas, Connecticut, Kentucky and New York) had adopted the proposal with omission of only the term “mental” from the phrase “extreme mental or emotional disturbance”.

Understanding family violence: Reforming the criminal law relating to homicide 155

application of any objective requirement would also necessitate a sound understanding of family violence or risk the same failings as loss of control in England and Wales.

A further issue is whether any partial defence should be limited in terms to apply only in the context of family violence. Given that we do not recommend the adoption of any partial defence, we only address this issue in brief to further highlight the difficulties with such an approach.⁹²⁵

The introduction of a partial defence of general application would have significant implications for the criminal law and would carry a high risk of unintended consequences. Given the current absence of general partial defences in New Zealand, it would be natural to expect the limits of any new partial defence to be tested vigorously. As victims of family violence who kill their abusers account for only a small fraction of homicide offenders, other defendants are far more likely to attempt to engage the partial defence.

On the other hand, we see several issues with limiting a partial defence to victims of family violence. As a general principle, the law should, wherever possible, apply equally to people who have the same state of mind. A partial defence limited to specific circumstances or defendant characteristics would likely result in anomalous and inconsistent treatment of other classes of defendant whose conduct may be comparably culpable.

Limiting a partial defence also heightens the risk that a woman who kills in response to family violence is not seen to be acting reasonably or will often not be acting reasonably.⁹²⁶This may discourage victims of family violence from going to trial and relying on self-defence and may distort or detract from self-defence reforms.⁹²⁷

Finally, if a partial defence were limited to victims of family violence, definitional issues, and therefore policy choices about scope, would inevitably arise. Queensland is the only jurisdiction to have enacted a defendant-specific partial defence of “killing for preservation”. This defence applies only where “the deceased has committed acts of serious domestic violence against the person in the course of an abusive domestic relationship” and “the person believes that it is necessary for the person’s preservation from death or grievous bodily harm to do the act or make the omission that causes the death”.⁹²⁸This would not capture a defendant who killed to protect another (for example, a woman who killed an abusive partner to protect her children from ongoing sexual and physical abuse), which has been subject to some academic comment.⁹²⁹
However, as we suggest in Chapter 11 in connection with sentencing, it is not difficult to see why Queensland might have decided to enact a narrowly framed defence.⁹³⁰

925 Submitters were divided on the question of whether a partial defence should be limited to victims of family violence. Those who favoured limiting a partial defence were concerned about the risk of unintended consequences and of undeserving defendants using the partial defence to their benefit. Others, including the Criminal Bar Association and Auckland District Law Society, could not see the justification for limiting a partial defence. They recommended providing additional criteria or a threshold test if the concern was to prevent its use by “undeserving” defendants. Some submitters emphasised the difficulties in identifying a victim of family violence, noting that many aggressors were primary victims in previous situations (as children). Historical and intergenerational violence could therefore be problematic to deal with under a partial defence. Another submitter said criminal defences should be principles based, not ad hoc.

At present in New Zealand, culpable homicide is either murder or manslaughter⁹³¹unless it is infanticide, which is New Zealand’s single specific homicide offence.⁹³²A more contemporary example of a specific homicide offence is Victoria’s now-repealed defensive homicide.

Creation of a new specific homicide offence may be an alternative way for the law to recognise reduced culpability for homicide. Such an offence could include any of the elements intended for a partial defence and reflect reduced culpability by setting a lower maximum penalty than applies to murder.⁹³³Like manslaughter, a specific offence could also operate as an alternative verdict so that a defendant charged with murder might instead be convicted of the specific offence; as might a defendant charged with manslaughter, if the maximum penalty was lower. As partial defences and specific offences both operate to recognise reduced culpability, the merits and drawbacks canvassed above would largely apply also to a separate homicide offence.

The offence of defensive homicide was introduced in Victoria in 2005 in response to the VLRC’s recommendation for a partial defence of excessive self-defence. To have found a person guilty of defensive homicide a jury must have been satisfied beyond reasonable doubt that:⁹³⁴
^._{the defendant killed in circumstances that would otherwise constitute
murder; but}
^._{the defendant was not guilty of murder
because he or she believed his or her conduct to be}

necessary to defend himself or herself against the infliction of death or really serious injury;

_and
^._{the defendant did not have reasonable
grounds for that belief.}

The offence carried the same maximum penalty as manslaughter in Victoria: 20 years’

^10.100Victoria decided to introduce defensive homicide rather than a partial defence to murder, because a separate offence would make it clear to the sentencing judge on what basis a verdict was reached, thus enabling imposition of a sentence that accurately reflected the crime.⁹³⁵If a partial defence were introduced, the basis for the jury’s verdict would not be clear, and it would be up to the judge to decide the basis on which to sentence.⁹³⁶

932 The law of infanticide in New Zealand has not been reviewed recently and infanticide is a controversial offence in the western world. See PJ Dean “Child Homicide and Infanticide in New Zealand” (2004) 27(4) Int J Law Psychiatry 339; and Eric Vallillee “Deconstructing Infanticide” (2015) 5(4) Western Journal of Legal Studies 1.

933 We consider the impact of the three strikes law on sentencing for murder and manslaughter convictions in the following chapter. We note that if the three strikes law remains unchanged, a separate homicide offence with a maximum penalty less than life imprisonment could potentially lead to a significantly different result if a defendant is convicted of that offence rather than the offences of murder or manslaughter. However we consider it preferable to focus directly on problematic aspects of the three strikes law rather than to propose a specific homicide offence as a “work-around”.

935 See speech during second reading by Attorney-General Hulls: (6 October 2005) VicPD LA 1351. See also Kate Fitz-Gibbon Homicide Law

Reform, Gender and the Provocation Defence: A Comparative Perspective (Palgrave Macmillan, Hampshire, 2014) at 123.

Understanding family violence: Reforming the criminal law relating to homicide 157

^10.101The other advantages of a specific offence identified in Victoria were that:

^._{a separate homicide offence provides juries and sentencing
judges with more options, in a}

_{self-defence case than the “all or nothing” choice between
murder and acquittal;}⁹³⁷

^._{a new offence would involve fewer complexities than a
partial defence and be easier for a}

^._{a separate offence means that data on excessive
self-defence is easily identifiable, thereby}

^10.102Defensive homicide proved problematic in practice, however, and for the reasons outlined in

^10.103As a general principle, offences should be general rather than context or victim-specific. New Zealand’s Crimes Act largely reflects this approach. In a review of crimes against the person, the Law Commission noted the need to guard against the “risk of ad hoc specific offences being randomly inserted on to the statute book, every time an issue arises that causes political or public concern”.⁹⁴¹Specific offences should only be recommended if there is a compelling rationale.⁹⁴²While context-specific offences can be helpful from a labelling perspective, the Commission has been sceptical of specific offences that capture conduct that could be adequately covered by generic offences, because:⁹⁴³

^._{they broaden prosecutorial discretion and can result in
inconsistent charging practice;}

^._{they single out one aggravating or mitigating factor among
the many possible factors that may}

be present in any given case, giving rise to arbitrary disparities across different offending;

^._{the overuse of specific offences could result in a
patchwork of offences without logical or}

^10.104The Legislation Advisory Committee Guidelines provide that a new offence can only be justified if it can be shown that:⁹⁴⁴

^10.105Accordingly, the Commission’s approach is to recommend a specific offence only where the case is sufficiently strong to overcome these issues, such as where there is a clear gap in the law.⁹⁴⁵

937 Victorian Department of Justice, above n 860, at 24; and Fitz-Gibbon, above n 935, at 122–123. Although, as we have noted, where a defendant is charged with murder, a jury already has the option of convicting of manslaughter even absent a partial defence or separate offence, and evidence that supports a claim of self-defence may also tend to support a claim of lack of intent, which may result in a conviction for manslaughter.

938 Victorian Department of Justice, above n 860, at 26. Some commentators, however, argue, to the contrary, that creation of a separate offence would unduly complicate the law and trials by adding to the number of matters a jury must consider: Wasik, above n 800, at 526.

941 Law Commission Review of Part 8 of the Crimes Act 1961 : Crimes Against the Person (NZLC R111, 2009).

942 Law Commission Strangulation: The Case for a New Offence (NZLC R138, 2016) at [1.17].

944 Legislation Advisory Committee Guidelines: Guidelines on Process and Content of Legislation (2014) at ch 1.

945 For further discussion see Law Commission, above n 941, at 30–31; and Law Commission, above n 942, at [8.32]–[8.36].

^10.106Few submitters supported introduction of a specific offence. The small number who did noted a separate offence might achieve better labelling and enable more tailored penalties but would in practice function much like a partial defence and, therefore, achieve similar ends and pose similar problems.⁹⁴⁶The Public Defence Service observed that the question of whether a separate offence is preferable is largely philosophical.

^10.107While we can identify some practical advantages of a specific offence over a partial defence, we are unconvinced they are sufficiently compelling to recommend a specific homicide offence targeted at victims of family violence who kill abusers. We agree with the concern expressed by some that, where the policy objective is to recognise reduced culpability of offenders in certain circumstances, creation of a new offence may send a confusing message.⁹⁴⁷

^10.108In this chapter we have considered the in-principle case for and against a partial defence or a separate homicide offence and the form any such defence or offence might take. We have sought to review the arguments and the options fairly and comprehensively, but for the following six reasons, we do not recommend that a partial defence or a specific offence be introduced in New Zealand:

^._{Although a number of arguments may be marshalled in
support of partial defences, a}

mandatory sentence is their driving and most compelling rationale. In a jurisdiction, such as New Zealand, that has abolished mandatory murder sentencing, that rationale falls away. To the extent there are vestiges of mandatory sentencing (such as the three strikes law) or other problems with sentencing, it is preferable that they be addressed directly in that context.

^._{Partial defences are conceptually problematic. They are
anomalous, as they apply only to}

homicide, and they introduce arbitrariness and rigidity into the recognition of mitigating circumstances. They are not the best way to recognise reduced culpability.

^._{The experience of partial defences in New Zealand and
elsewhere bears out these conceptual}

problems. In whatever form and with whatever focus, partial defences have given rise to complexity and difficulties.

^._{It is well recognised that partial defences, and
particularly excessive self-defence, can work}

to the detriment of defendants who seek to rely on self-defence. The recommendations we make in this Report are intended to work as a package, and we are concerned to avert any reform that might undermine the effectiveness of our recommended clarification of section

^._{As in previous Law Commission reports and in common with a
number of overseas law}

reform bodies, commentators and submitters, we consider mitigating circumstances are best addressed at sentencing. We would only recommend a partial defence or new offence if there was evidence of a clear and strong need. We have not identified such evidence. Most victims of family violence who commit homicide are charged with murder but are convicted of manslaughter. As we discuss in the next chapter, in the rare case where a victim of family violence is convicted of murder, we consider the essential structure of New Zealand’s

946 A specific offence might more reliably ensure sentences reflect reduced culpability. A manslaughter conviction leaves a person at least liable to a life sentence, even if that is unlikely to be imposed in sympathetic cases. This was noted in the 1991 Report of the Crimes Consultative Committee, which observed that abolition of infanticide “would significantly increase the potential penalty for this class of offender, although it is plain enough that a degree of leniency would continue to be extended in practice”: Crimes Consultative Committee Crimes Bill 1989: Report of the Crimes Consultative Committee (Crimes Consultative Committee, Wellington, 1991) at 54. See AP Simester and Warren Brookbanks Principles of Criminal Law (4th ed, Thomson Reuters, Wellington, 2012) at 593.

Understanding family violence: Reforming the criminal law relating to homicide 159

sentencing framework is capable of taking into account the mitigating circumstances of victims of family violence who commit homicide.

^._{Finally, our review is limited to victims of family
violence and we have not had the}

opportunity (or a remit) to consider the position of homicide defendants generally. Even a carefully tailored partial defence, based on a criterion like trauma, would raise complexity, and there would be significant risks with limiting any defence to a narrow group of defendants.

^10.109The centrepiece of our recommended legislative reforms is to make self-defence more accessible to victims of family violence who kill their abusers. Amendments to the Crimes Act will clarify that, in the context of family violence, self-defence may apply in the absence of an imminent threat. Whether the force used was reasonable will remain a question of fact for the jury.

^10.110Cases where an offender has a history of being abused by the deceased but there is no viable claim of self-defence will involve some culpability, but it might be mitigated because of the history of abuse. In such cases, our view is that sentencing provides a better avenue than partial defences for addressing variation in homicide offending. This presumes the sentencing process is able to properly take into account the various factors that might mitigate culpability, however. In the next chapter, we consider the adequacy of New Zealand’s sentencing law in this context.

R9 _{No new partial defence or separate homicide offence should be
introduced in New
Zealand.}

In this chapter we briefly set out the law of sentencing as it relates to victims of family violence who kill their abusers and analyse how this has been applied in practice and whether there is any case for reform. We consider the general structure of the Sentencing Act 2002 and the purposes of sentencing. We then canvass in greater detail the mitigating factors specified in the Sentencing Act and how these are taken into account in family violence cases, before turning to issues specific to homicide sentencing, including the three strikes provisions. Discussing sentencing, we refer to “offenders”, although we have used the term “defendants” elsewhere in this Report. That is because, by the time a person is sentenced, the charge against them has been proved and they have been convicted.

The question for this chapter is how to ensure the reduced culpability of victims of family violence who kill their abusers can be adequately addressed in sentencing.

A key policy issue in sentencing is how best to facilitate decisions that respond to the particular facts of the offending and the offender’s personal circumstances while also promoting consistency with cases of a similar nature. Fairness requires that both be achieved so far as possible. However, different people within the community and different judges will have different views on how serious a particular offence is and the degree to which an offender’s personal circumstances reduce their culpability or otherwise justify a more lenient sentence. These questions are inevitably contentious.

Under the traditional common law model, the judiciary has considerable discretion to determine the appropriate sentence in an individual case, subject to the applicable maximum penalty.⁹⁴⁸An alternative approach is to provide more guidance to judges either through statute or sentencing guidelines.⁹⁴⁹At the most directive end of the scale, mandatory sentences might be imposed through legislation – often reflecting a community view about the gravity of the offence.⁹⁵⁰

In New Zealand, the approach to sentencing sits somewhere in the middle of the scale. The Sentencing Act sets out principles that must be taken into account when deciding a sentence, including the gravity of offending, the desirability of consistency and the need to impose the least restrictive outcome appropriate in the circumstances.⁹⁵¹It also sets out a list of purposes the judge may consider, such as holding the offender to account, deterrence, denunciation,

948 Warren Young and Andrea King “The Origins and Evolution of Sentencing Guidelines – A Comparison of England and Wales and New

Zealand” in Andrew Ashworth and Julian Roberts (eds) Sentencing Guidelines: Exploring the English Model (Oxford University Press, Oxford,

950 See, for example, the observations on the United States’ “grid” sentencing model in Andrew Ashworth and Julian Roberts “The Origins and Nature of the Sentencing Guidelines in England and Wales" in Andrew Ashworth and Julian Roberts (eds) Sentencing Guidelines: Exploring the English Model (Oxford University Press, Oxford, 2013) at 2.

Understanding family violence: Reforming the criminal law relating to homicide 161

rehabilitation and community protection.⁹⁵²The Sentencing Act also lists a number of aggravating and mitigating factors that must be taken into account in determining a sentence.⁹⁵³
These factors are not exhaustive.⁹⁵⁴

The Act therefore creates a framework for the exercise of judicial discretion, providing guidance on both the purposes of sentencing and factors to be considered in an individual case. The Act is not prescriptive, and the factors do not dictate a particular outcome.

For murder, there is a strong presumption in favour of life imprisonment. The presumption may be displaced if a sentence of life imprisonment would be “manifestly unjust”,⁹⁵⁵but the threshold for displacement is a high one. It is apparent from case law that it will be met only in exceptional cases.

The Law Commission has previously reviewed aspects of sentencing in its Report Sentencing Guidelines and Parole Reform,⁹⁵⁶which considered overall issues of fairness and consistency within our sentencing system. In this review, we are considering the narrower question of whether the existing law allows courts sufficient flexibility to impose sentences that take account of an offender’s experience as a victim of family violence where such offenders kill their abusers.

All submitters agreed that it is important for judges to have sufficient flexibility to impose reduced sentences when victims of family violence kill their abusers and are found guilty of manslaughter or murder. The Issues Paper invited comment on whether additional mitigating factors could be included in the Sentencing Act or whether the threshold for displacing the presumption of life imprisonment needed to be reformed.

The majority of submitters believed that there should be changes to sentencing law. Many considered that there would be benefit in elaborating on family violence as a mitigating factor. Some noted that sentencing reform could be a viable alternative to a partial defence. Other submitters were, however, more equivocal, stating that, although changes to sentencing may highlight societal concern, the list of mitigating factors is already non-exhaustive and the presumption of life imprisonment has previously been displaced in family violence homicide cases.

Sentencing is not meant simply to punish. It is also intended to promote community protection and assist offenders to rehabilitate and reintegrate.⁹⁵⁷To this end, pre-sentence reports include information on the offender’s rehabilitative needs and future risk of offending.

954 Sentencing Act 2002, s 9(4). The court is not prevented from taking into account any other aggravating or mitigating factor that the court thinks fit, and a factor referred to in s 9 need not be given greater weight than any other factor that the court might take into account.

955 The exception to the discretionary approach is found in the three strikes law, which is discussed below beginning at paragraph [11.72].

The courts in the cases we have reviewed have consistently addressed the defendants’ prospects of rehabilitation and risk of reoffending. In R v Erstich,⁹⁵⁸R v Raivaru⁹⁵⁹and R v Wharerau,⁹⁶⁰the defendants’ youth and prospects of rehabilitation and reintegration were found to warrant sentence discounts. As for risk of reoffending, there are examples in both directions. In some cases, low risk has mandated a less restrictive penalty,⁹⁶¹and in one case the Court considered a lengthy sentence might actually increase the risk of reoffending, because in an adult prison the defendant would be exposed to “older criminals who may lead you further astray”.⁹⁶²

Among cases in which the risk of reoffending was considered to be higher are R v Wihongi and R v Rihia – the two most recent murder cases – and in both, risk was relevant to the length of the finite term of imprisonment. In Wihongi, fresh evidence about the appellant’s risk contributed to the Court of Appeal’s conclusion that the finite term should be increased,⁹⁶³and in Rihia, the Court noted the need for “community protection” in determining sentence.⁹⁶⁴

The Sentencing Act prescribes several mitigating factors to be taken into account by the sentencing judge, including the following that might be relevant where a victim of family violence kills their abuser:
^._{the conduct of the victim of the offending, which, in the context of
homicide, refers to the}

_{conduct of the deceased;}⁹⁶⁵_and
^._{the offender’s diminished intellectual capacity or
understanding.}⁹⁶⁶

Depending on the circumstances, other mitigating factors such as the offender’s age may also be relevant. There may also be aggravating factors. In all cases of the type we are concerned with, however, the conduct of the deceased and diminished capacity of the offender are potentially prominent issues and worthy of particular consideration.

Mitigating factors can relate to the nature of the offending or the personal circumstances of the offender, which affect assessment of their culpability relative to others convicted of similar offending.⁹⁶⁷While the conduct of the deceased is relevant to the nature of the offending, the offender’s mental state is a personal mitigating factor. Mitigating factors can also arise in combination. This is particularly likely in cases where the offending can be attributed in part to the deceased’s conduct and in part to the offender’s psychological state. Homicides by victims of family violence of their abusers will often be cases of this sort.

961 For example, R v Wickham, in which the sentencing Judge saw no need to impose a sentence that protected the community from the defendant, as it was “extremely unlikely that you will ever reoffend”: R v Wickham HC Auckland CRI-2009-090-010723, 20 December 2010 at [31].

963 The Court concluded, in the particular circumstances of Ms Wihongi’s case, that “the sentencing principle of community protection is better met by a longer finite sentence, providing a longer period of imprisonment (during which treatment can be provided to Ms Wihongi with a view to reducing future risk). It would also provide a longer period during which some supervision with the possibility of recall is available.”: R v Wihongi [2011] NZCA 592, [2012] 1 NZLR 775 at [98].

The modern approach to sentencing uses as a reference point a starting point taking into account aggravating and mitigating features of the offending, but excluding aggravating and mitigating features relating to the offender. Put another way, a starting point “is the sentence considered appropriate for the particular offending (the combination of features) for an adult offender after a defended trial.”

Understanding family violence: Reforming the criminal law relating to homicide 163

Unlike a defence or partial defence, which is either accepted or rejected, a mitigating factor can be given a different weighting depending on the circumstances of the case. For example, in a case where abuse was particularly severe but self-defence is not made out, the judge would still have scope to recognise that mitigating circumstance through a sentencing discount.

In cases in which victims of family violence kill their abusers, the conduct of the deceased is likely to be the most obvious mitigating factor.

The Court of Appeal has made it clear that the behaviour of the victim of the offending includes behaviour that provokes a violent reaction or that justifies a degree of force, albeit less than that used.⁹⁶⁸Below, we consider recent statements by the Court of Appeal setting out how prior violence by the victim of the offending should be considered as a matter of law. We then discuss how this issue has been treated in the cases we have reviewed.

The leading case relevant to this review is Tuau v R.⁹⁶⁹The offender was an adult man convicted of wounding with intent to cause grievous bodily harm after he stabbed his father through the eye with a knife. He had been abused by his father as a child and was living with him temporarily following the death of his mother. On the day of the offending, the father had accused the offender of being a child molester. At issue in the appeal was the extent of sentencing discount appropriate for each of three features: the provocation on the day of the offending and in the immediate prior period, the history of abuse by the father towards his family including the offender and the way the offender’s schizophrenia and paranoia made him perceive the situation. The Court of Appeal held that the history of abuse by the father was a mitigating factor that should be taken into account when considering the offender’s response to the provocation that precipitated the offending. The Court said:⁹⁷⁰

[The pattern of prior violence] explains how [the offender] decided he had no alternative but to attack his father to avoid being attacked; and no alternative but to render his father unconscious and stab him out of fear that, unless he went that distance, his father would retaliate with even greater force.

While noting the three factors identified converged and it was important not to over-count, the Court of Appeal considered the trial Judge “should have allowed Mr Tuau a significantly greater discount than he did to recognise why it was that he offended and as seriously as he did”.⁹⁷¹
This is a clear statement that the historical conduct of the victim of the offending is relevant to how the immediate conduct is viewed. It strongly supports an approach of considering the immediately threatening conduct in the wider context of family violence. In Tuau, the offender’s sentence was reduced by a year to four and a half years.

As for sentencing methodology, the standard approach, as explained by the Court of Appeal in R v Taueki, is to fix a starting point by reference to the aggravating and mitigagting features of the offending, of which “conduct of the victim” may be one.⁹⁷²This may, however, be done in different ways. For example, in R v Paton,⁹⁷³the Judge fixed the starting point by reference to previous manslaughter cases where victims of family violence had killed their abusers.⁹⁷⁴

In another manslaughter case, R v Rakete,⁹⁷⁵the Judge explained that, but for any mitigating factors, he would set a starting point of between three years six months and four years. He then adverted expressly to the conduct of the deceased as a mitigating factor, saying:⁹⁷⁶

The victim was the primary aggressor, a much larger person than you, intoxicated and threatening toward you. In reaching this view I have had regard to evidence of his aggression earlier in the day and at the time of the incident (though as I have said I do not accept that there was a verbal threat to kill), and I have had regard to the evidence of his prior acts of violence towards you in a domestic context and when he was jealous of another man.

Given the prior acts of violence by the deceased, as well two other mitigating factors relevant to the circumstances of the offending (that Ms Rakete was fearful of an assault and that the use of weapon was largely impulsive and motivated primarily by fear), the Judge adopted a starting point of three years.⁹⁷⁷

Adoption of a starting point based on analogous cases, as in Paton, may promote sentencing consistency, a goal affirmed by section 8(e) of the Sentencing Act. However, this way of achieving sentencing consistency may tend to reinforce the approach of previous cases, which will not always be helpful.⁹⁷⁸Where a victim of family violence is convicted of killing an abuser, a focus on previous cases might, for example, inhibit consideration of offending in light of up- to-date research about family violence – notwithstanding the useful precedent set in Tuau. We expect our recommendations for continued education will, to some extent, address this concern.

When a victim of family violence kills their abuser, the conduct of the deceased will often be the most relevant mitigating factor. It enables the sentencing judge to understand, consistent with contemporary social science, family violence as a pattern of harmful behaviour that belongs to the abuser, rather than the relationship or the family violence victim.⁹⁷⁹However, consideration of that factor alone may not always be sufficient to capture the full range of mitigating features in the present case. As discussed in Chapter 2, the cumulative and compounding effect of family violence can result in a raft of secondary issues such as mental health issues. For example, in many of the cases we reviewed the offender was said to be suffering from post-traumatic stress disorder at the time of the offending.⁹⁸⁰Further, many offenders have a long history of being abused not just by the deceased but by others. These matters can be taken into account in sentencing under the mitigating factor of diminished capacity or understanding.

In Chapter 10, we considered the possibility of a partial defence based in diminished responsibility. A persuasive argument against such a partial defence is that the offender’s psychological characteristics are better considered in sentencing, where a more flexible approach is available. In sentencing, the offender does not need to demonstrate that their

977 R v Rakete, above n 975, at [36]. The ultimate sentence, with discounts for mitigating factors to the defendant personally, was two years’

978 Julia Tolmie has suggested that Paton demonstrates deficiencies in the “standardised” approach to sentencing that requires determination of a starting point, by reference to certain aggravating features of the offending, and then adjustment to take account of the offender’s personal circumstances. Tolmie considers it is arguable this process “supports and replicates many of the misconceptions about family violence ... It places primary focus on one incident of violence, decontextualized from the history of violence, the larger architecture of coercive control in the relationship and the entrapment of the victim.” The author acknowledges that a history of family violence victimisation is sometimes factored into the sentence starting point in these cases, but that is “not a given”: Julia Tolmie “Defending Battered Defendants on Homicide Charges in New Zealand: The Impact of Abolishing the Partial Defences to Murder” [2015] NZ L Rev 649 at 675–677.

979 Family Violence Death Review Committee Fifth Report: January 2014 to December 2015 (Health Quality & Safety Commission, February 2016)

980 For example, Rakete, above n 975, at [40]; Wihongi, above n 963, at [34]; Rihia, above n 964, at [20]; R v Reti [2009] NZCA 271 at [5]; and R v

Understanding family violence: Reforming the criminal law relating to homicide 165

mental functioning falls below a prescribed level of criminal responsibility. Instead, any level of impaired decision making is potentially relevant if it contributes to the offending, and the weight given to this factor can be tailored to the degree of impairment.

The Court of Appeal case of R v Whiu⁹⁸¹is significant for the category of offenders within the scope of this review. In Whiu, the offender drove dangerously while intoxicated and caused the death of another driver. She was charged with manslaughter. She had earlier been driving the car with her abusive partner who had ordered her to drive him, punched through the glass of one of the windows and continued punching her while she drove. After he instructed her to drop him off, she drove back towards their home and crashed into another car on her street. The other driver later died, and the passenger’s injuries were serious.

The Court of Appeal considered the extent to which psychological effects resulting from a history of intimate partner violence were mitigating. As the deceased in this case was an innocent bystander, there was no question of the contribution of the deceased’s conduct to the offending. The Court noted that “the appellant’s characteristics or symptoms must be related to what occurred if they are to have any significance in terms of her sentencing”,⁹⁸²and made the following observations:⁹⁸³

... we agree that it is not necessary for there to be a formal diagnosis of battered women’s syndrome before prolonged abuse suffered by a woman at the hands of a partner or family member can be taken into account on sentencing. The critical point is that, whatever label is used, there must be evidence which supports the view that prolonged abuse suffered by an offender materially contributed to her offending. Typically a psychiatrist or psychologist would give such evidence, and where it exists, it should be taken into account like any other relevant factor.

If an offender wishes to argue that she has suffered prolonged abuse at the hands of a partner or family member and that this has contributed materially to her offending and so is relevant to sentence, she will have to point to an evidential basis for the submission. The evidence will need to address the underlying facts of the abuse, its impact on the offender and the way in which it is said to have made a material contribution to the offending.

The Court accepted that the psychological effects of sustained abuse were relevant as a mitigating factor and justified a discount in the order of 20–25 per cent from the starting point.⁹⁸⁴

The cases in our review demonstrate a mixed approach to the offender’s psychological state as a mitigating factor. Some include extensive discussion, while others touch on the issue briefly or not at all.⁹⁸⁵This factor is sometimes identified to explain a verdict of manslaughter rather than murder – for example, that the offender’s prolonged history as a victim of extreme violence means they cannot be taken to have appreciated, in the same way as a person not exposed to high levels of violence, that the wound they inflicted was likely to cause death.⁹⁸⁶

984 The Court rejected the submission that, in order to properly reflect the experiences of offenders who have suffered abuse, this mitigating factor should be taken into account in fixing the starting point and concluded there was “no reason of principle to treat post-traumatic stress disorder arising from battered women’s syndrome differently” to other psychological and behavioural conditions relevant to sentencing (at [33]). See Tolmie, above n 978, at n 152.

985 Cases in which no reference is made to psychological reports include R v Paton, above n 973; R v Neale HC Auckland CRI-2007-004-3059, 12

June 2009; R v Tamati HC Tauranga CRI-2009-087-0868, 27 October 2009; and Wharerau (HC), above n 960.

In some cases, the cumulative effect of abuse from different people including the deceased will be relevant to the offender’s psychological state. In R v Fairburn, the Judge took account of the offender’s history as a victim of family violence and sexual abuse, including the way that history informed the offender’s belief that the deceased was a threat to both the offender and her daughter.⁹⁸⁷

While the judgment in Whiu expressly provided that psychological effects are relevant, it imposed an evidential requirement. It is not enough for the defendant to assert they have suffered abuse that has caused psychological disturbance relevant to the offending. There must be an evidential basis for that claim.

A defendant’s lawyer may seek to present relevant evidence, and under the Criminal Procedure (Mentally Impaired Persons) Act 2003, in some circumstances, a court may on its own initiative order a health assessor such as a psychologist or psychiatrist to prepare a report even if counsel does not. Section 38(1) provides that such a report may be ordered to assist the court to determine whether a person is unfit to stand trial or is insane, the type and length of sentence that might be imposed and the nature of a requirement the court may impose as part of, or a condition of, a sentence or order. A report might also be ordered by the court even if another report has been separately arranged by counsel for the defendant.⁹⁸⁸Section 48, however, is not a general power to require relevant expert evidence for proceedings beyond the purposes specified in section 38(1). It does not provide for the provision of expert evidence at trial generally, for example, to explain the dynamics of family violence relevant to a claim of self-defence. Such expert evidence was discussed in Chapter 7 above.

Therefore, at least in cases where the court has sufficient information to identify a need for evidence from a health assessor for sentencing purposes and section 38 applies,⁹⁸⁹this provision could avoid prejudice if evidence about an offender’s psychological state is not sought or made available for sentencing by counsel. There may, however, still be variation in access to expert evidence and depending on whether the court invokes section 38 in a given case, the costs of a defendant obtaining it independently may be a barrier, especially for offenders whose legal representation is being paid for through legal aid.⁹⁹⁰

If psychological reports (whether court-ordered or otherwise) are prepared and taken into account, this can lead to a significantly reduced sentence. In Rakete, for example, the Judge, Whata J, recorded:⁹⁹¹

While your PTSD does not excuse what you have done; your action, or more correctly, your over reaction, occurred against a backdrop of [...] emotional fragility and dysfunction caused by it. Indeed I am satisfied, applying the various clinical factors also identified by Dr Evans, that all of this affected your capacity to act in a proportionate way when confronting Mr Simeon. Secondly, a lengthy term of imprisonment is likely to have a disproportionate effect on you given your history of social phobia, agoraphobia and the panic attacks. I refer also to the report by the Ashburn Clinic which raises concerns about your ability to cope if imprisoned. Third, the likelihood of reoffending is low as your current offending is very context specific.

989 As to which, see Togia v Police, above n 988, in which the Court of Appeal held that s 38 may apply to persons on bail if they appear in and are therefore under the supervision of the court at the time the report is ordered (at [31]).

990 See also the discussion of expert evidence in connection with self-defence, in Chapter 7.

991 R v Rakete, above n 975, at [41]–[42] (ellipses in square brackets from original).

Understanding family violence: Reforming the criminal law relating to homicide 167

These factors taken together support a discount of 20%. I note for completeness that I did not include your psychiatric condition [...] in formulating the starting point for the offending – so there is no double counting. Rather, I preferred to apply a discount after fixing the starting point for offending, in a transparent way, to acknowledge the impact of your emotional and mental health on your offending and for the purpose of understanding the full impact of a sentence of imprisonment on you.

Within our sample, there are also cases in which psychological effects of intimate partner violence are not explicitly considered as a mitigating factor, despite there being sufficient comment about the offender’s history that this would seem to be relevant. In Rakete, quoted above, Whata J went on to say “while your condition does not form part of the observable facts of the offending, I consider that it is a relevant factor when assessing your overall culpability but only with the benefit of expert evidence”.⁹⁹²Given the extent of the sentencing discount that may be provided, there is a risk of injustice if potentially relevant evidence is not before the court, although, as we have noted, section 38 of the Criminal Procedure (Mentally Impaired Persons) Act 2003 provides some guard against this.

We have considered whether some reform is required to ensure that family violence is properly considered as a mitigating factor in cases where victims of family violence kill their abusers. Our review suggests that while the reduced culpability of family violence victims can be taken into account in sentencing, there may be inconsistencies in approach. Whether and to what extent that is so is difficult to be sure of given, as we have noted, that a key mitigating factor – the conduct of the victim – may routinely be factored into the sentence starting point.

In these circumstances, arguments for and against reform are finely balanced. We are mindful of the need to avoid unintended consequences and to ensure that the reform is targeted to the identified problems. In our view, the most significant problems are as follows:
^._{While the Court of Appeal’s decision in
Tuau provides a clear statement that the conduct}

of the victim includes historical conduct in the context of family violence, this has not been cited in any of the subsequent cases in our homicide sample, perhaps because Tuau was not a homicide case. In any event, we cannot be sure from the sentencing notes whether judges are always giving due weight to the broader context of family violence over the course of the relationship leading up to the incident that constituted the offending, such as, in particular, non-physical abuse including tactics of coercion and control. See paragraph [2.58] above.
^._{There is a possibility
psychological evidence may not always be prepared despite the}

circumstances of the offending being suggestive of diminished intellectual understanding or capacity resulting from the trauma of family violence. This concern is mitigated by the provisions of the Criminal Procedure (Mentally Impaired Persons) Act 2003, but if evidence is not presented by counsel or sought by the court under the Act in a given case, the judge will not have the information required to properly consider this mitigating factor.

^._{amending section 9(2)(c) of the Sentencing Act to clarify
that “conduct of the victim”}

^.amending section 9(2)(e) of the Sentencing Act to clarify that “diminished intellectual capacity or understanding” includes any impairment resulting from being subject to family violence.

This would help ensure that family violence is brought to the attention of judges and would prompt defence lawyers to consider the need for submissions targeted at these areas and supported by expert evidence. Inclusion of these factors in the statute may also make it more straightforward for lawyers to apply for legal aid funding for expert evidence in the event evidence is not otherwise before the Court.

We recognise that, in some cases, relevant “conduct of the victim” (that is, the deceased, in homicide cases) may include prior family violence against people other than the offender (for example, a child or a parent, as in R v Raivaru). Despite this, our recommended amendment to section 9(2)(c) is limited to violence against the offender. There are two reasons for this:
^._{First, in almost
all cases we have reviewed, the offenders have been primary victims
of}

intimate partner violence perpetrated by the deceased. Thus, it appears that, most frequently, the relevant “conduct of the victim” will be violence against the offender. Further, in all cases we reviewed, the conduct of the deceased included abuse of the offender, even if other people were also victims. As a matter of reality, a primary victim may not be the only person subject to abuse by a predominant aggressor because intimate partner violence and child abuse and neglect are “entangled” forms of abuse and often co-occur.⁹⁹³Violence against another family member may also amount to violence against the offender because family violence encompasses psychological abuse.⁹⁹⁴This would appear capable of including witnessing or fearing abuse of a child, for example.⁹⁹⁵Under the Domestic Violence Act, psychological abuse of a child specifically includes exposure to abuse of people with whom the child has a domestic relationship.⁹⁹⁶In these circumstances, it does not seem necessary to frame the amendment any more widely than we have recommended.
^._{Second, we are concerned it would be difficult to frame an amendment
more widely without}

risking unintended consequences or going beyond the focus of this Report, which is the nature of the dynamics between victims of family violence (who kill the perpetrators) and the perpetrators who are killed.⁹⁹⁷Our recommended amendment will, of course, not limit the existing scope of section 9(2)(c).

These recommendations complement our approach to self-defence. Even with an expanded definition of self-defence, there will be cases of homicide against an abuser that fail the test because the force used is not a reasonable response to the threat. There will also be cases, like Rihia, in which the fatal attack occurred in the extenuating context of prior violence but was not primarily related to the perception of a future threat. In such cases, there is a need for a flexible approach to recognise both the harm caused by the offending and the mitigating factors.

993 Family Violence Death Review Committee Fourth Annual Report: January 2013 to December 2013 (Health Quality & Safety Commission, June

994 At 13. See also Domestic Violence Act 1995, s 3. In Queensland, this was part of the rationale for recommending that any new partial defence for victims of serious abuse not extend to victims who act to protect third parties, such as children: “The conduct of the abuser towards the third party could be encompassed within our definition of violence towards the victim of abuse, specifically as psychological abuse. It would form part of the history of serious violence in the relationship”: Geraldine Mackenzie and Eric Colvin Homicide in Abusive Relationships: A Report on Defences (2009) at 46.

996 Family Violence Death Review Committee, above n 993, at 13, citing the Domestic Violence Act 1995, s 3.

997 Again, Queensland’s experience of enacting a defendant-specific partial defence is helpful. Mackenzie and Colvin, above n 994, observed, at 45, that “[a]ny broadening of the defence for victims of abuse is fraught with risks” and, despite recommending that any new partial defence extend to “family members” who act to protect victims of serious abuse, even that limited extension to third parties was not taken up. See also Michelle Edgely and Elena Marchetti “Women Who Kill Their Abusers: How Queensland’s New Abusive Domestic Relationships Defence Continues to Ignore Reality” (2011) 13 Flinders LJ 125 at 152.

Understanding family violence: Reforming the criminal law relating to homicide 169

As with self-defence, we do not suggest restricting the reforms to homicide. Similar issues will arise for other offences by a primary victim against a predominant aggressor, such as assault, and it is important that these too can be recognised.

The list of factors in the Sentencing Act “demonstrates Parliament’s view that the specification [of factors] is primarily a legislative rather than a judicial responsibility”.⁹⁹⁸While judges can take account of any relevant factor, it is open to Parliament to guide the courts to take particular account of certain features relevant to the offending. If there is a case that judges may give too little or inconsistent weight to family violence, this suggests there is an argument for reform.

The counter argument is that reform is not needed because the experience of family violence victims can be considered within the ambit of existing mitigating factors. The law already allows judges to impose reduced sentences for offenders who kill primary abusers.

Fair sentencing practice can, however, be promoted and ensured only if those involved in sentencing – the lawyers for both sides, experts preparing reports and the judge – are aware of the nature of family violence and able to consider the offender’s circumstances based on evidence rather than misconceptions. We are not satisfied this is currently achieved in all cases. We have considered whether the problems identified can be adequately addressed through education and have reached the view that education is likely to be helpful but not sufficient. Education does not carry the force of legislation.

Our assessment is that there is a case for a declaratory provision in the statute to draw attention to family violence as a feature within the scope of existing mitigating factors. In the absence of a facility such as sentencing guidelines (which we understand are unlikely to be introduced in the near future), greater detail in the statute is warranted.⁹⁹⁹In particular, we are concerned that the highly relevant feature of historical conduct of the deceased may not always be given sufficient weight and that it is possible evidence on psychological effects of sustained family violence may not always be sought or available. Statutory recognition of family violence – as it relates to existing grounds of mitigation – would help in this regard.

At the same time, we acknowledge that judicial flexibility needs to be protected. It is generally accepted that the question of relative weighting of mitigating and aggravating factors should be left to the sentencing judge, who is best placed to consider all relevant features of the offending. For this reason, we do not suggest anything more prescriptive than the amendments outlined above. As these amendments are consistent with statements in recent Court of Appeal cases, we perceive a minimal risk of unintentional consequences.

We have considered the alternative of adding a new mitigating factor, such as “the offender’s actions are the direct result of being a victim of family violence”. An option along these lines was suggested in the Issues Paper. We have reached the view that this raises several complicated issues of application, but for completeness, we set out the arguments for and against this option.

999 The Commission’s earlier report, Sentencing Guidelines and Parole Reform, above n 956, recommended establishment of a Sentencing Council which would have the role of preparing sentencing guidelines. This was recommended because guidelines can promote consistency while protecting a judge’s ability to respond to particular facts. Without a mechanism of this sort, there is a potentially wide gap between the abstract principles and factors expressed in the Sentencing Act and the facts presented by a particular case, leaving a lot of room for judicial discretion. This can be partially addressed by guideline judgments of the higher courts. However, the Court of Appeal has declined to issue guideline judgments for murder and manslaughter on the basis that there is too much variation in offending. The gap therefore remains and is a potential issue for this category of cases.

The chief argument in favour of this approach, as compared with the more conservative option above, is that it would provide a stronger signal on how a history of family violence should be considered in sentencing. However, as a more significant reform, we are concerned with how it might be applied.

We are mindful that violent offenders – both men and women – are overwhelmingly more likely than the general population to have themselves been victims of violence, including sustained violence throughout childhood.¹⁰⁰⁰Whether such previous abuse is directly causative of offending will be difficult to assess, and judges may interpret a new mitigating factor based on experiences of family violence more broadly than we anticipate. This would raise policy issues that are well outside the scope of this project and that we do not have time to consider. On one view, a history of family violence can be seen as mitigating moral culpability because the offender’s background is one in which violence is normalised. Conversely, from the perspective of community protection, this group of offenders may have complex rehabilitative requirements and pose a higher risk of reoffending – therefore potentially justifying longer sentences. These are fairly significant issues of policy that affect a wide range of offending types and offenders. They would require an examination of rehabilitation services as well as sentencing law. Such matters are not within our terms of reference.

_·amending section 9(2)(c) to clarify that “conduct of the victim” includes prior family violence against the offender; and

_·amending section 9(2)(e) to clarify that “diminished intellectual capacity or understanding” includes any impairment resulting from being subject to family violence.

Section 102 of the Sentencing Act provides a presumption in favour of life imprisonment for those convicted of murder. This was introduced in 2002, when the mandatory life sentence for murder was abolished in New Zealand. The presumption is rebuttable but only if a life sentence would be “manifestly unjust”.

As might be expected, given the inherent gravity of a murder conviction, the “manifestly unjust” threshold for departing from a life sentence with a 10-year minimum period of imprisonment is very high. Since the Sentencing Act was enacted, there have only been six cases in which a finite sentence has been substituted for a sentence of life imprisonment. This small group includes two cases in our review, R v Wihongi¹⁰⁰¹and R v Rihia.¹⁰⁰²The other cases that have met the “manifestly unjust” threshold are:
^._{A “mercy killing” where a husband killed his wife who had
advanced Alzheimer’s
disease.}¹⁰⁰³

Understanding family violence: Reforming the criminal law relating to homicide 171

^._{A case in which the offender was convicted as a party to
the murder but was not directly}

^._{A case in which the offender was suffering from
schizophrenia and experiencing paranoid}

delusions. The defence of insanity was rejected by the jury, but the sentencing Judge noted that the attack was entirely out of character and was motivated only by the mental illness.¹⁰⁰⁶

Other than in exceptional cases where the presumption of life imprisonment is displaced, the Sentencing Act also includes a hierarchy of minimum periods of imprisonment if a sentence of life imprisonment is imposed. Section 103 prescribes a 10-year minimum period of imprisonment, and section 104 requires a court to impose a minimum period of 17 years for murder committed with one of several aggravating features, such as a “high level of brutality”. As with section 102, the 17-year minimum period can be departed from in qualifying cases only if the judge considers that a minimum term of that length would be “manifestly unjust”.¹⁰⁰⁷

The range of circumstances in which the courts have been willing to depart from the presumptive 17-year minimum period is broader than that in which the section 102 presumption in favour of life imprisonment has been displaced, notwithstanding that the language of “manifestly unjust” is used in both sections. The more severe sentence of a 17-year minimum period and the range of qualifying features that lead to that sentence mean that more offenders are likely to fall within the class for whom the punishment is clearly unduly severe, meeting the “manifestly unjust” threshold.¹⁰⁰⁸

Thus, while murder sentencing in New Zealand (outside of the three strikes regime) is now discretionary, it is still tightly constrained, and the threshold for departing from the presumption of a life sentence will be met only in truly exceptional cases. In these circumstances, it is especially striking that, in two cases, a history of abuse has satisfied the test. Wihongi comes close to identifying a class of defendants who will ordinarily receive reduced sentences.¹⁰⁰⁹As a Court of Appeal case, Wihongi is binding on lower courts and has been followed on this point in Rihia. Wihongi includes commentary on relevant principles and parliamentary debates that indicate that the legislature envisaged that victims of “severe and prolonged abuse” would be a class of offender for whom the presumption may be set aside.

The section 102 presumption has, however, not always been displaced in cases of this kind. It was not addressed in R v Neale or R v Reti,¹⁰¹⁰the other two murder convictions in our case review. These cases preceded Wihongi and Rihia, but given they comprise half of the murder convictions in our case review, it is necessary to consider whether they affect the precedential value of the later cases in terms of section 102. It appears that Neale and Reti are distinguishable on the basis they are not so readily amenable to a primary victim/predominant aggressor analysis. In Reti, for example, the relationship involved violence, but there was also

1005 R v McNaughton [2012] NZHC 815; upheld in R v Cunnard [2014] NZCA 138. There is a second case with similar facts that also resulted in a finite term at sentencing (R v Innes [2014] NZHC 2780), but the conviction was overturned on appeal: R v Baker [2015] NZCA 306.

1007 For discussion of the principles and approach applicable to the provisions governing murder sentencing, see R v Williams [2004] NZCA 328; [2005] 2 NZLR 506 (CA) at [30]–[34] and [45]–[68]; and R v Gottermeyer [2014] NZCA 205 at [71]–[86] (leave to appeal to the Supreme Court declined: Gottermeyer v R [2014] NZSC 115).

1010 As we noted in Chapter 9, notwithstanding that Ms Neale’s case post-dated enactment of the Sentencing Act, the sentencing Judge said that life imprisonment was “the maximum, and indeed the mandatory” sentence for murder. However there appears to have been no suggestion that the s 102 presumption might have been displaced in Ms Neale’s case. See R v Neale, above n 985.

evidence the defendant was generally the aggressor.¹⁰¹¹Thus, it appears tenable to conclude that, in light of Wihongi, there is at least an emerging trend towards displacement of the section 102 presumption in cases where a primary victim of family violence kills their abuser. The Family Violence Death Review Committee (FVDRC) made a similar observation in its submission, albeit with the rider that, in its view, the starting point for the finite sentences imposed in these cases is still “very high”.

The Court of Appeal in Wihongi, relying in large part on fresh evidence about Ms Wihongi’s risk of reoffending, increased the sentence from eight to 12 years. In Rihia, the Judge adopted a 12-year starting point, based on similarities to Wihongi. After a two-year guilty plea discount, Ms Rihia was sentenced to a 10-year finite term of imprisonment.

As we have already said, some commentators have expressed concern about the length of the sentences imposed in these cases and argued that, given the history of abuse the defendants suffered before the homicides, even finite sentences of 12 and 10 years are too high.

This concern is premised partially on a comparison with sentences imposed in manslaughter cases. As noted in Chapter 9, our review of the cases suggests the murder convictions tended to follow evidence of more serious offending than in the manslaughter cases,¹⁰¹²with stronger evidential bases for intention and less clearly self-defensive elements.¹⁰¹³We are therefore sceptical of comparing sentences in manslaughter and murder cases, although we acknowledge there may not always be a significant culpability gap between murder and manslaughter. In Paton, for example, the Judge considered the offending was “only a very short way from murder”¹⁰¹⁴but sentenced Ms Paton to five years and three months’ imprisonment.

Bearing in mind that murder involves the intentional taking of life other than in self-defence, a reasonably lengthy finite sentence is likely to be appropriate even in highly unusual mitigating circumstances. Further, the sentences in Wihongi and Rihia are among the lowest ever imposed for murder in New Zealand. The only case we have identified in which a person has received a sentence of less than 10 years’ imprisonment for murder is R v Law, the euthanasia case.¹⁰¹⁵

Each case also turns on its own facts. In Wihongi, the Court of Appeal had considerable fresh evidence on reoffending risk, and the sentence increase was based on that evidence and the Court’s concern the finite term should address the sentencing purposes of denunciation and deterrence. This is apparent from the Court of Appeal’s judgment:¹⁰¹⁶

In our view, in the particular circumstances of this case, the sentencing principle of community protection is better met by a longer finite sentence, providing a longer period of imprisonment (during

1012 Putting aside the manslaughter cases that appeared to follow from a successful claim of provocation.

1015 The sentence imposed in R v Law, above n 1003, was 18 months' imprisonment. The Court held that the circumstances of the offence were such that life imprisonment was manifestly unjust (at [51]), and accepted the defence submission that the full range of sentencing options under the Act were available (at [52]). Mitigating factors were the offender’s age, health, motives, guilty plea, acceptance of responsibility, remorse, and previous good character (at [53]). The sentencing Judge went on to say, at [62]:

Plainly, you do not represent any risk to the community but that is not the only matter to be considered. While there are substantial factors which argue for a compassionate approach, those important considerations must be tempered by the high value which the Courts and the community rightly attach to the sanctity of human life. The taking of a human life, even for the highest and best motives, is not permitted under our law and, for good reason, murder is ordinarily regarded as the most serious crime in our statute books. According to the evidence, there are many persons in our community suffering from Alzheimer’s disease and other forms of dementia. The Court would be sending the wrong message to the community if it were prepared to allow the deliberate killing of someone suffering from such a disease or other affliction to go unpenalised, even in the tragic circumstances of a case like this.

Understanding family violence: Reforming the criminal law relating to homicide 173

which treatment can be provided to Ms Wihongi with a view to reducing future risk). It would also provide a longer period during which some supervision with the possibility of recall is available. A longer finite term also assuages the concern that a sentence of eight years’ imprisonment does not adequately meet the sentencing purposes of denunciation of a crime involving the taking of a life and deterrence.

When it declined Ms Wihongi leave to appeal further the Supreme Court emphasised the case’s particularity, noting that the issues the Court of Appeal addressed and on which its decision was based were “very fact-specific”.¹⁰¹⁷Thus, it seems plausible that a shorter finite term could be imposed in a case with a minimal risk of reoffending (such as Wickham, if that case had resulted in a murder verdict), although Wihongi was a Solicitor-General’s appeal and so the range of the appropriate finite term may have extended beyond the 12 years actually imposed.¹⁰¹⁸

It is also helpful to consider penalties for other crimes against the person. For example, the maximum sentences for “wounding with intent” or “aggravated wounding or injury” are 14 years, and for “injuring with intent”, the maximum penalty is 10 years. As for manslaughter, in Whiu, the conviction resulted in a sentence of seven and a half years, even with a discount for psychological disturbance from family violence. This suggests that a finite term of 10 or 12 years, or even slightly more,¹⁰¹⁹is broadly in the range that could be expected for murder. To the extent there remains a difference between sentences for murder and sentences for manslaughter (with Paton perhaps the clearest example in our review), the seriousness of a murder conviction should not be discounted. Further, where the presumption of life imprisonment (which is not engaged for manslaughter) is displaced, that will, in itself, be a material recognition of reduced culpability even before the length of the finite term arises.¹⁰²⁰

Finally, returning to Wihongi, it is relevant the Court of Appeal declined to impose a minimum period of imprisonment. The 12-year finite term equates to a non-parole period of four years under the Parole Act 2002.¹⁰²¹The Court of Appeal considered whether a longer minimum term was required and decided not to impose one. Ms Wihongi has now appeared before the Parole Board three times, first in June 2014, and has not yet been released. We have had the benefit of reviewing the notes of the Parole Board for each of the hearings. The Parole Board has reached the view that further reintegration planning is required before Ms Wihongi is released. At none of the hearings has her legal representative sought release.

For these reasons, we are satisfied that a close review of Wihongi and Rihia does not demonstrate there is a problem with the sentences in those cases. What it does demonstrate is that, in sentencing, judges are required to balance a range of matters. Mitigating factors must, in particular, be weighed against community protection and deterrence and denunciation of serious crime.

In Chapter 9, we considered whether the repeal of provocation has adversely affected the position of victims of family violence who commit homicide in terms of trial outcomes. We concluded there is at present insufficient evidence to say that it has.

1018 Per the principle, that “[u]pon a successful appeal by the Solicitor-General a sentence is adjusted by no more than the minimum extent necessary to remove the element of manifest inadequacy”: Sipa v R [2006] NZSC 52, (2006) 22 CRNZ 978 at [9] per Blanchard J. See also R v N [1998] 2

1019 As noted above, the “range” disclosed by Wihongi, above n 963, and Rihia, above n 964, may extend beyond 12 years if the Court of Appeal’s decision on the length of the finite sentence in Wihongi reflected the principles usually applied to sentence appeals brought by the Solicitor- General.

1020 Tolmie notes this point in connection with the sentences imposed in Wihongi, above n 963, and Rihia, above n 964, albeit in support of an argument that “a sentencing discretion for murder does not make murder a neutral alternative to a manslaughter conviction based on a provocation defence”: Tolmie, above n 978, at 667.

When it comes to sentencing, we have already observed that it is problematic to compare pre- repeal cases (like R v Suluape and R v King) with post-repeal cases (like R v Rihia). We do not repeat that discussion here but record the following statement of the Court of Appeal in R v Hamidzadeh,¹⁰²²made by reference to the section of the Law Commission’s 2007 Report that we referred to in Chapter 10:¹⁰²³

We do not discern any Parliamentary intention to diminish the high threshold necessary to establish manifest injustice under s 102. Indeed, the Law Commission specifically rejected the adoption of any lower standard in its report which led to the abolition of the provocation defence. We see no reason to depart from the general approach adopted in Rapira, recently endorsed in Wihongi, that manifest injustice under s 102 is likely to be established only in exceptional circumstances. It necessarily follows from the abolition of the defence that a conviction for murder is no longer to be treated as manslaughter where provocation is established. Sentencing for murder must therefore be approached on the footing that the killing was intentional or, where applicable, that the offender intended to inflict

_{injury known to be likely to cause death but was reckless as to whether
death ensued or not.}

The maximum sentence for manslaughter is life imprisonment. There is no minimum sentence, and sentencing is highly fact dependent.

Where manslaughter results from the deliberate infliction of injury, an established approach is for the judge to begin with the bands for grievous bodily harm established in R v Taueki.¹⁰²⁴This reflects the fact that manslaughter is inadvertent killing and, as such, the focus should be on the action that caused the death and not only the result. For example, in R v Leuta the Court of Appeal stated:¹⁰²⁵

In sentencing in cases of violent offending the element of deterrence must be directed towards the aspect of the conduct of the offender which was intentional and which created the risk of serious harm or death. In the case of manslaughter the likely deterrent effect of sterner sentences must be measured against that aspect, not against the unintended consequence of death.

This ensures a level of consistency in how manslaughter offending is treated compared with the position if the same action had not resulted in death. There have been exceptional cases where following a guilty plea, an offender has been discharged without conviction.¹⁰²⁶At the other end of the spectrum, cases of severe child abuse leading to death commonly result in terms of over
10 years’ imprisonment and, in one case, 16 years’ imprisonment.¹⁰²⁷

The Sentencing Act was amended in 2010 to introduce provisions commonly known as the three strikes law. These are contained in sections 86A–86I of the Act. These amendments significantly curb sentencing discretion for violent offences. Under the three strikes law, people convicted of a first strike offence must be given a warning about the effects of a second strike

1023 At [58], discussing Law Commission The Partial Defence of Provocation (NZLC R98, 2007) (footnotes omitted).

1024 R v Taueki, above n 967. For comments on this approach, see R v Tae [2010] NZCA 598 at [11]–[23]; and Ioata v R [2013] NZCA 235 at

1026 For example R v X [2015] NZHC 1244, in which a mother inadvertently caused the death of her 16-month-old son after leaving him unattended in a car on a hot day. She had been working late several days in a row and forgot he was in the car, as her husband usually dropped him at childcare in the morning.

Understanding family violence: Reforming the criminal law relating to homicide 175

offence. If they are convicted of a second strike offence, they will be required to serve the full term of imprisonment rather than being eligible for parole. On a third strike, they must serve the maximum penalty available, also without parole, unless that would be manifestly unjust. Strike offences include a range of sexual and violent offending. There are specific provisions for murder committed as a second or third strike offence and for manslaughter committed as a third strike offence, which we discuss below.

Under the second and third strike murder provisions in section 86E, the offender will be required to serve a life sentence without parole, although the prohibition on parole can be departed from if the court is satisfied that life without parole would be “manifestly unjust”.

If the “manifestly unjust” threshold is met, the court must order a minimum period of imprisonment of at least 20 years for a third strike offence and at least 10 years for a second strike offence. The 20-year minimum period of imprisonment for a third strike offence can be departed from if the court considers it would be manifestly unjust. In all circumstances, the minimum period of imprisonment is 10 years. There is no scope to impose a finite sentence. This applies to both second and third strike murder convictions.

To understand the effect of the three strikes law for murder, it is necessary to return to the discussion of section 102 of the Sentencing Act. Section 102 provides that those who are convicted of murder will be required to serve a sentence of life imprisonment with a minimum period of imprisonment of 10 years unless this is manifestly unjust. Under the three strikes provisions, offenders who would otherwise have met the “manifestly unjust” threshold (including victims of family violence who kill their abusers) will now be required to serve a life sentence with a 10-year minimum period of imprisonment. Section 102 was amended when the three strikes law passed to provide that it is subject to new section 86E(2),¹⁰²⁸so there is no room to read down the three strikes provisions that do not permit discretion in cases of manifest injustice.

If a second strike offence is manslaughter, the offender will be required to serve the full term imposed without parole, but the court retains discretion as to the duration of the term imposed (that is, it need not be life).¹⁰²⁹This is a significant difference between second strike manslaughter and second strike murder.

However, if manslaughter is a third strike offence, a similar approach applies as for murder,¹⁰³⁰though with a more lenient starting point. A life sentence must be imposed with a minimum period of imprisonment of no less than 20 years. This can be reduced to a minimum period of no less than 10 years if the court finds that a minimum period of 20 years would be manifestly unjust.¹⁰³¹As with a second or third strike murder conviction, there is no ability to depart from the life sentence even in circumstances of manifest injustice.

1028 Section 86E(2) of the Sentencing Act 2002 provides that, if s 86E applies, which it will if the offender is convicted of murder as a second or third strike offence, the court must (a) sentence the offender to imprisonment for life for that murder; and (b) order that the offender serve that sentence of imprisonment for life without parole unless the court is satisfied that, given the circumstances of the offence and the offender, it would be manifestly unjust to do so.

1030 See the discussion in Warren Brookbanks “Partial Defences to Murder in New Zealand” in Alan Reed and Michael Bohlander (eds) Loss of Control and Diminished Responsibility: Domestic, Comparative and International Perspectives (Ashgate, Farnham, 2011) 271 at 289–290. The author notes that “[i]n terms of the minimum penalty that must be imposed upon a second or third strike conviction for either murder or manslaughter, both offences are now in an undifferentiated category” (at 289).

As mentioned above at [11.68], in sentencing decisions in manslaughter cases that involve deliberate injury, the courts have followed the guideline for grievous bodily harm sentencing in Taueki. The three strikes legislation cuts across this approach. It requires courts to focus solely on the fact of the manslaughter conviction and not on the seriousness of the action that caused death.

A minimum period of imprisonment has a very different effect when the offender is serving a life sentence rather than a finite sentence. In the case of a finite sentence, the minimum period of imprisonment is the minimum portion of the sentence that must be served, and on release, the offender will be subject to parole conditions for the remaining duration of the finite term. When a person is sentenced to life imprisonment, they are subject to parole conditions until they die and can be recalled to prison if they breach their parole conditions at any stage in the future. There is also no guarantee that the offender will ever be released on parole – the assessment is undertaken with regard to community safety and the offender’s rehabilitation progress and reintegration prospects.

At the time of writing, there have been four homicide convictions under the three strikes law – all for murders committed as second strike offences.¹⁰³²In each of the four cases, the judge has taken the view that it would be manifestly unjust to require the defendants to serve sentences of life imprisonment without parole. The earlier two cases, R v Harrison and R v Turner, are awaiting hearing by the Court of Appeal, the Solicitor-General having appealed against the sentences on the basis of the “manifestly unjust” question.¹⁰³³Given these two appeals, we make no specific comment on the cases. However, we make some general observations about the three strikes law and the exceptions provided through the “manifestly unjust” proviso.

The term “manifestly unjust” is used in the Sentencing Act as follows:
^._{Section 102, allowing the judge to depart from the default sentence
of life imprisonment with}

a 10-year minimum period for murder. If this threshold is met, the judge may impose a finite sentence with a minimum period of one third of the finite sentence.
^._{Section 104, allowing the judge to depart from
the default 17-year minimum period of}

imprisonment where the murder has certain aggravating features. If this threshold is met, the judge may impose a minimum period of between 10 and 17 years.
^._{Section 86E, allowing the judge to depart from the
default sentence of life imprisonment}

without parole for a second strike murder. If this threshold is met, the judge may impose a minimum period of no less than 10 years, bringing it in line with the default sentence for murder under section 102.
^._{Section 86E, allowing the judge to depart from the default sentence
of imprisonment for}

life without parole for third strike murder. If this threshold is met, the judge may impose a minimum period of no less than 20 years. However, if the 20-year minimum period would also be manifestly unjust, the judge may impose a minimum period of no less than 10 years, bringing it in line with the default sentence for murder under section 102.

1032 These cases are R v Harrison [2014] NZHC 2705 (in which the offender’s first strike offence was an indecent assault); R v Turner [2015] NZHC

189 (in which the offender’s first strike offence was wounding with intent); R v Kingi [2016] NZHC 139 (in which the offender’s first strike offence was robbery); and R v Herkt [2016] NZHC 284 (in which the offender’s first strike offending was also robbery).

1033 We understand that the Court of Appeal is scheduled to hear appeals by the Solicitor-General against the sentences imposed on Mr Harrison and Mr Turner, with an appeal by Mr Harrison’s co-offender to be heard at the same time, on 9 and 10 June 2016.

Understanding family violence: Reforming the criminal law relating to homicide 177

^.Section 86D, allowing the judge to depart from the default sentence of imprisonment for life with a minimum period of 20 years for third strike manslaughter. If this threshold is met, the judge can impose a minimum period of no less than 10 years, bringing it in line with the default sentence for murder under section 102.

^._{Section 86D, allowing the judge to depart from the
requirement to impose the maximum}

sentence with no parole for third strike offences other than murder and manslaughter. If this threshold is met, the judge may allow parole but cannot impose a lesser sentence.

For each section, the judge must consider whether the default sentence under the statute would be manifestly unjust. Under sections 102 and 104, if the threshold is met, the judge will then determine the sentence in accordance with the ordinary principles of the Sentencing Act. Under the three strikes provisions, the judge is able to depart from the presumptive sentence of life imprisonment without parole for murder or the presumptive sentence of life imprisonment with a 20 year minimum period for manslaughter but cannot impose a finite sentence and cannot impose less than a 10-year minimum period.

This means it is possible that some offenders who would have received a finite sentence prior to the three strikes law will now receive a sentence of life imprisonment with a minimum period of imprisonment of 10 years. However, somewhat anomalously, an offender who previously would have received a life sentence for murder with a minimum period of 10 years may not receive any uplift under the three strikes law.

This is particularly problematic for third strike manslaughter convictions and, in the section below, we consider how this might apply to some of the cases we have reviewed. It is apparent that there could be a substantial disparity between manslaughter sentences imposed where the offender has two prior qualifying convictions compared with a similar offending not caught by the three strikes provisions.

The effect of the three strikes provisions on victims of family violence who kill their abusers will depend on the number of previous strikes and whether the conviction is for murder or manslaughter. The requirement to impose a life sentence could arise if:
^._{an offender is convicted of murder
after having already committed one or two strike}

_{offences; or}
^._{an offender is convicted of
manslaughter after having committed two strike offences.}

In each case, the most lenient result possible is a life sentence with a minimum period of imprisonment of 10 years. This would put any such case at the very highest end of the cases identified in our case review.¹⁰³⁴We consider below the potential sentencing uplifts that could apply. It is significant that this could easily arise for a manslaughter conviction as well as in cases, less common in our review, involving murder convictions.

A range of offences could qualify as a first or second strike. These are specified in section

86A. Most sexual offences are included together with most violent offences other than low-level assault. Offences against property where a weapon is involved are also included. The list of qualifying offences includes some offences that span a considerable range of seriousness, such as “wounding with intent to injure” and “aggravated burglary”.

1034 Sentences of life imprisonment with minimum non parole periods of 10 years were imposed in both R v Reti, above n 980, and R v Neale, above n 985, although both pre-date Wihongi.

Within our case review of 20 convictions, we have identified three cases in which prior offences might have been caught by the three strikes law had they occurred after it was enacted. Two of these cases involved manslaughter verdicts and one case a murder verdict. All involved guilty pleas. There may be other cases in which the offender had a prior violent conviction that was not mentioned in the sentencing notes.

In R v Brown,¹⁰³⁵the offender killed her abusive partner during an argument and pleaded guilty to manslaughter. The sentencing Judge noted a prior conviction for “injuring with intent”. Other prior convictions were alluded to but not identified. If any of these had been qualifying offences and the three strikes law had applied, Ms Brown would have been sentenced to life imprisonment with a minimum period of imprisonment of 10 years as opposed to a finite sentence of five years and six months with eligibility for parole in the normal course. This is a stark difference in outcome.

In R v Stone,¹⁰³⁶the offender killed her partner by a stab to the leg during an argument and pleaded guilty to manslaughter. The Judge noted “a past of violent convictions, two for assault and others for wilful trespass and driving with an excess breath alcohol level”.¹⁰³⁷It is not clear whether the assault charges referred to are common assault or more serious offending that would be caught by the three strikes law. The sentence in Stone was three years’ imprisonment, which reflected the strange circumstances of a wound to the leg resulting in death. Despite these particular circumstances, if her case had been caught by the three strikes legislation, Ms Stone would have been required to serve a life sentence with a minimum period of imprisonment of
10 years.

The judge in Rihia noted a previous conviction for assault by Ms Rihia against Mr Rihia. It was not stated whether this was common assault or a more serious offence. If this conviction resulted from the incident described in the judgment in which Ms Rihia “attacked Mr Rihia repeatedly over the head with a table leg”,¹⁰³⁸it could have been charged as “wounding with intent to injure” and thereby qualify as a first strike offence. Similarly, in Wihongi, while prior convictions were not mentioned, the Judge referred to a previous episode of violence in which Ms Wihongi injured her partner.¹⁰³⁹Both Ms Rihia and Ms Wihongi received finite terms. If, in the future, analogous offending is caught by the three strikes law, the possibility of finite terms of imprisonment and the normal parole entitlements will be precluded.

There is a strong case that the three strikes law could cause injustice to the category of offender we are concerned with in this review. It is only a matter of time before a case comes before the court where an offender is convicted of manslaughter or murder and would have received a relatively short finite sentence but for the three strikes law.¹⁰⁴⁰

There are further issues of potential relevance to victims of family violence who commit offences against their abusers that are worth noting. For example, there is no provision for a judge to decline to issue a first or second strike warning if, in the circumstances of the offence, categorising the offending as a strike would be manifestly unjust. Nor is there provision to

1040 We note that the first case to present problematic issues of mandatory sentencing for culpable homicide under the three strikes law may or may not involve a victim of family violence. It could be a murder conviction for one of the other categories of defendant for whom life imprisonment has been held to be manifestly unjust, or it could be a manslaughter case that would otherwise be treated with lenience. It might be a case that falls just outside the boundaries of our terms of reference – such as in Whiu, above n 981, the motor manslaughter case, where the offender was a victim of family violence but the homicide victim was a bystander. As the courts have repeatedly pointed out, manslaughter is an offence that captures a wide range of culpability and for which sentencing flexibility is, therefore, particularly important.

Understanding family violence: Reforming the criminal law relating to homicide 179

depart from the mandatory maximum sentence for third strike offences except for murder or manslaughter. The three strikes law may bear heavily on victims of family violence who injure their abusers in analogous situations to those in our case sample but without causing death.

Under the three strikes law the legislature made a deliberate policy decision to impose mandatory life sentences for murder and manslaughter offenders who may otherwise have received finite sentences.¹⁰⁴¹This review has identified one group of offenders for whom this is likely to be unjust, particularly if the prior qualifying offence (or offences) is (or are) similarly connected to or explained by a history of abuse. We have not considered the position of other offenders who may also be disproportionately disadvantaged by the three strikes law. Doing so would take us well outside the scope of this review.

We are also mindful there are underlying questions around the application of the three strikes law to homicide offences, given that, in all four murder cases to date, the sentencing judges have reached the conclusion that a sentence of life imprisonment without parole would be manifestly unjust. As noted above, appeals in two of these cases are due to be heard shortly by the Court of Appeal.¹⁰⁴²In the face of this uncertainty around interpretation and the broader policy issues, we have not developed recommendations for amendments to the Sentencing Act. Instead, we suggest that the Ministry of Justice considers the issues raised by the three strikes legislation for homicide offenders in exceptional circumstances. In particular, we recommend that the Ministry of Justice consider how the mandatory life sentence could be dispensed with for victims of family violence who kill their abusers in circumstances where the three strikes regime would otherwise mandate a life sentence.

The Ministry of Justice should undertake further policy work to address the issues noted in this Report in relation to sections 86D(4) and 86E of the Sentencing Act 2002 as they apply to homicide offenders in exceptional circumstances, and specifically:

_·consider the position of victims of family violence who kill their abusers in situations where the three strikes regime would mandate a life sentence; and

_·consider how to amend the legislation to allow judges to impose a finite sentence in deserving cases.

1041 The mandatory life sentence for manslaughter was criticised at the time. See Warren Brookbanks and Richard Ekins “The Case against the Three Strikes Sentencing Regime” [2010] NZ L Rev 689 at 705. It was also subject to debate at the Committee stage. An amendment to exclude certain types of manslaughter was voted on and rejected: (18 May 2010) 663 NZPD 10901.

In 2001 the Law Commission published a report examining the legal defences available to protect those who commit criminal offences as a reaction to domestic violence: “Some Criminal defences with Particular Reference to Battered Defendants” LCR73. Of particular note the Report recommended repeal of the partial defence to murder of provocation, an amendment to the defence of self-defence and abolition of the mandatory sentence of life imprisonment for murder.

In 2002 Parliament introduced discretionary sentencing in murder cases, subject to a presumption in favour of life imprisonment.

In 2007 the Law Commission published a second Report: “The Partial Defence of Provocation” LCR98. The Report again recommended repeal of this partial defence. The Commission concluded that its major deficiency was that the partial defence of provocation had been primarily used by violent offenders in respect of unwelcome advances or slights against their honour. It was seldom available to victims of family violence. Given this conclusion, the Commission re-examined whether the defence of self-defence should be amended to ensure that it is available to victims of family violence in appropriate cases. In answering this question the Commission noted the work undertaken as part of the Government Response to the Commission’s 2001 Report. That work concluded that amendment to section 48 of the Crimes Act 1961 (self-defence and defence of another) was not required to meet the needs of battered defendants, and might be undesirable in light of the fact that the section is generally regarded as working well. The Ministry reviewed recent case law, which tended to suggest that problems previously encountered were being ironed out in the courts; it thus concluded that the real problem previously was one of social awareness, rather than of law. The Ministry found that overwhelmingly stakeholders were comfortable with letting matters take their course. The Commission stated: “we are content at this stage to concur with the Ministry’s conclusions”.

In 2009 Parliament repealed section 169 of the Crimes Act 1961, which had provided for the partial defence to murder of provocation.

Since the 2009 repeal, the Family Violence Death Review Committee has been gathering data on all family violence homicides in New Zealand. In its Fourth Annual Report published in 2014, the Committee concluded that New Zealand is out of step in how the criminal justice system responds to victims of family violence when they face criminal charges for killing their abusive partners. To address this, the Committee recommended that the Government re-examine the options for amending the defence of self-defence and introducing a targeted partial defence to murder.

The Government has asked the Law Commission to conduct the re-examination recommended by the

The Law Commission will re-consider whether the law in respect of a victim of family violence who commits homicide can be improved. As part of this review the Law Commission shall consider:

Should the test for self-defence, in section 48 of the Crimes Act 1961, be modified so that it is more readily assessable to defendants charged with murder who are victims of family violence; and

Whether a partial defence for victims of family violence who are charged with murder is justified and if so in what particular circumstances; and

Whether current sentencing principles properly reflect the circumstances of victims of family violence who are convicted of murder?

The reference forms part of a broader range of initiatives relating to family violence being undertaken by the Ministry of Justice. It also forms part of two other projects being undertaken by the Law Commission, being alternative trial processes with particular focus on sexual offence cases and whether a separate offence of non-fatal strangulation is desirable.

^._{Conducting preliminary research including a review of
recent New Zealand cases and an}

^._{Consulting with targeted agencies within New Zealand
including the Police, Judiciary, Ministry}

of Justice, Family Violence Death Review Committee, New Zealand Law Society and other knowledgeable agencies;

^._{Engaging with an expert panel, made up of both public and
non-public sector advisers, during the}

Understanding family violence: Reforming the criminal law relating to homicide 183

^._{New Zealand Law Society (Criminal Law Committee, Youth
Justice Committee and Family Law}

^R1Judges should continue to receive education, including through the Institute of Judicial Legal

^R2Regular and ongoing education courses on the dynamics of family violence should be made available to all criminal lawyers, including Crown prosecutors and defence counsel.

_·^{reflect contemporary social science understanding of
family violence and victims’}

_·^{explain the primary victim/predominant aggressor
analysis in intimate partner violence;}

_·identify common misconceptions of family violence that persist today and their implications in the criminal justice system.

^R7The Evidence Act 2006 should be amended to include provisions based on sections 322J and

Understanding family violence: Reforming the criminal law relating to homicide 185

^R8The Solicitor-General should, when next reviewing the Solicitor-General’s Prosecution

Guidelines, consider whether they should include express reference to the potential relevance of a defendant’s history as a victim of family violence.

R9 _{No new partial defence or separate homicide offence should be
introduced in New Zealand.}

_·amending section 9(2)(c) to clarify that “conduct of the victim” includes prior family violence against the offender; and

_·consider the position of victims of family violence who kill their abusers in situations where the three strikes regime would mandate a life sentence; and

_·consider how to amend the legislation to allow judges to impose a finite sentence in deserving cases.

Terms and abbreviations commonly used in this Report have the meanings set out below. Where appropriate we have adopted and summarised the definitions of key terms used by the Family Violence Death Review Committee in their Fourth Annual Report.

Child abuse Includes all forms of physical and emotional ill-treatment, sexual abuse, neglect and exploitation that results in actual or potential harm to the child’s health, development or dignity. Five subtypes can be distinguished: physical abuse, sexual abuse, neglect and negligent treatment, emotional abuse and exploitation. May include exposure to intimate partner violence.

A broad range of controlling behaviours, commonly of a physical, sexual and/or psychological nature, which typically involve fear, intimidation and emotional deprivation. It occurs within a variety of close interpersonal relationships, such as between partners, parents and children, siblings and in other relationships where significant others are not part of the physical household but are part of the family and/or are fulfilling the function of family.

_{Definition from the Taskforce on Violence Within
Families} <www.msd.govt.nz>.

Any behaviour within an intimate relationship (including current and/or past live-in relationships or dating relationships) that causes physical, psychological or sexual harm to those in the relationship.

All forms of abuse between family members other than intimate partners or parents of their children.

Issues Paper Law Commission Victims of family violence who commit homicide (NZLC IP39, 2015). Available to download from the Law

The person who is the most significant or principal aggressor in a violent intimate partner relationship, and who has a pattern of using violence to exercise coercive control.

The person who (in the abuse history of the relationship) is experiencing ongoing coercive and controlling behaviour from their intimate partner.

2001 Report Law Commission Some Criminal Defences with Particular Reference to Battered Defendants (NZLC R73, 2001). Available to download from the Law Commission website <www.lawcom.govt.nz>.

2007 Report Law Commission The Partial Defence of Provocation (NZLC R98, 2007). Available to download from the Law Commission website <www.lawcom.govt.nz>.

Understanding family violence: Reforming the criminal law relating to homicide 187

New Zealand Law Commission Reports

Understanding Family Violence. Reforming the Criminal Law relating to Homicide [2016] NZLCR 139; Understanding Family Violence. Reforming the Criminal Law relating to Homicide [2016] NZLCR 139; NZLC R139

New Zealand Law Commission Reports

Understanding Family Violence. Reforming the Criminal Law relating to Homicide [2016] NZLCR 139 (12 May 2016); Understanding Family Violence. Reforming the Criminal Law relating to Homicide [2016] NZLCR 139; NZLC R139