Observations from the cases
9.4A 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.
Charging and plea discussions
9.5Under 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 defendant’s conduct but should not be inflated to increase the likelihood of an offer to plead guilty to a lesser charge (overcharging). The overarching consideration in any plea discussions is the interests of justice.
9.6We 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, and 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.
9.7Most 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. Some 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. We explore this possibility below.
9.8Cases we have reviewed in which defendants were charged with manslaughter usually involved force that did not suggest murderous intent, and in some cases where the charge was murder, the defendant’s use of force was such that intent could be readily inferred.
9.9In 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. In 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. Ms Hu and Ms Gerbes were both charged with manslaughter. Ms Hu pleaded guilty to that offence. Ms Gerbes was ultimately acquitted at trial.
9.10The 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 response to violent assaults. In R v Mahari, for 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. Despite being charged with, and tried for, murder, Ms Mahari and Ms Wharerau were both convicted of manslaughter.
9.11Looking 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.
Charges amended from murder to manslaughter
9.12In four cases, including R v Tagatauli, the defendants avoided a trial for murder by pleading guilty to manslaughter.
9.13One such case pre-dated the repeal of provocation, and the decision to accept the plea seems to be explained by a provocation narrative. In the other three, R v Brown, R v Woods and R v Tagatauli, the 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”. In 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.
Do the cases reveal a problem with charging practice?
9.14One 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. If so, and if such charging decisions are in part recognition of the defendants’ reduced culpability, it is debatable whether, in practice, there is any problem with charging.
9.15On 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.
9.16On 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”.
9.17We 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.
9.18We 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. The Solicitor-General will be in the best position to make that judgement.
9.19We 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.
- R8 The 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.
9.20While, 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.
9.21In 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. As we identified in the Issues Paper, however, it may be difficult to explain manslaughter verdicts for some cases on this basis.
9.22An 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.
9.23In 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).
The manslaughter verdicts: was murderous intent absent or not proved?
9.24To 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). Intent is a question of fact that can be inferred from the defendant’s conduct and the circumstances he or she must have known. If, 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.
9.25In 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.
9.26Murderous 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. We would suggest they illustrate the complexities involved in murderous intent in this context. Some examples provide helpful illustrations.
9.27In R v Wickham, the 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”. The 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”.
9.28R 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.
9.29That 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”, citing the following passage from the English Court of Appeal case of McInnes:
[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.
9.30As 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”. The violence the defendants had suffered before the homicides, Bradfield suggested, was:
… 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, together with the accused’s fear or anger before the killing, are used to explain the finding of lack of intent.
9.31This approach may have arisen in cases in our sample. In R v Paton, the 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:
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?
9.32Explaining 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. Some 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.
9.33“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”. This is an important function, even if there are few cases in which the jury actually acts as such a “check”. The special role of jury nullification in murder trials has been noted in Australia by White J in R v Marshall:
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 ...
9.34If 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”.
9.35Brenda 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. She argues that the outcomes in these cases were “undoubtedly fair if moral blameworthiness is the basis of criminalisation” but that:
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.
9.36The point about fairness, as Midson notes and some submitters raised, is 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, although, as we discuss in the next chapter, a legal mechanism such as a partial defence would not necessarily provide greater clarity.
9.37If 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. Because 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 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.
9.38Any 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.
The murder verdicts
9.39We 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, 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.
- In R v Neale, 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).
- In R v Wihongi, 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.
9.40Of 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:
… 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 found that even if you were provoked you did not display the self-control expected of a person in those circumstances.
Do the cases reveal a problem with trial outcomes?
9.41The 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.
9.42Cases 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.
9.43Typically, 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.
9.44At 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.
9.45We 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.
Has the repeal of provocation made any difference?
9.46The 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).
9.47The 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. The 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.
9.48R 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. R v Rihia is one post-repeal case in which provocation might have availed the defendant if it had still been in force. In 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:
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.
9.49On 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.
9.50Further, 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. The 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.
9.51Only 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 has disadvantaged this group of defendants. At present, we consider there is insufficient evidence to conclude that it has.