Executive summary of homicide trends, patterns and outcomes in England and Wales, using data from the Home Office Homicide Index.
Executive summary
Preventing homicide is a core feature of the police operational duty to protect life and property. Understanding patterns in homicide is essential to informing robust prevention strategies and to ensuring that responses are targeted and effective.
This paper uses data from the Home Office Homicide Index to provide a comprehensive summary of homicide trends, patterns and outcomes in England and Wales between 2019/20 and 2021/22. It also discusses the implications of this information for the role of police in prevention of homicide.
Homicides are rare in England and Wales, with an annual average of about 11 homicides per million people. Individual homicides are hard to predict but, when aggregated, clusters in homicide characteristics can be identified. This report uses seven homicide sub-types identified by the Home Office based on victim characteristics and incident context to make the examination of patterns and the planning of homicide prevention more manageable.
Trends and frequency
During the three-year period, 2019/20 to 2021/22, rates of homicide were stable. The median number of homicides per week in England and Wales was 12, with that weekly count ranging between 3 and 43. There was no significant change in the weekly frequency of homicide over the three-year period and no observable seasonal pattern. There was no effect of COVID lockdowns on the number of homicides, although the proportion of homicides where victim and suspect lived together increased from 18% to 26%, suggesting a change in the nature of homicide during these periods.
Area differences
The frequency of homicide varied considerably across forces. In most forces, the monthly frequency of homicide followed patterns indicative of random variation. Across forces, the most common number of homicides per month was zero but the average monthly number was 1.18 with a range of 0 to 39 (18 if excluding the manslaughter of 39 people in a lorry in Essex in October 2019).
Adjusted for population, the average rate of homicide in London was three times the rate of homicide in the South East of England (the region with the lowest homicide rate). Homicides are spatially and socio-economically concentrated: 56 of the 317 local authority districts and the 27% most deprived neighbourhoods each accounted for half of all homicides in England and Wales. However, risk was not a uniquely urban phenomenon: three of the ten area local-authority areas with the highest rate of homicide were in London, but three were rural areas.
Predictability
At police-force level, the number of homicides each month tended to follow a pattern similar to random variation, which is common in rare events and suggests that individual homicides are likely to be very difficult to predict. The difference between a homicide and a serious but non-fatal assault is often the difference between “a [weapon] too near and a hospital too far” (Felson and Eckert 2018, 3), making it hard to identify in advance which violent incidents are likely to become homicides. Furthermore, the rarity of these events makes it extremely difficult to identify when, where or to whom homicide will happen.
Victimisation
The analysis revealed some clear patterns in homicide victimisation. Almost three-quarters of victims were male. Male victimisation peaked in late adolescence and early adulthood while female victimisation remained more consistent over the life course. While the majority of homicide victims were white, black people were disproportionately likely to be victims across all age groups. This was particularly acute between the ages of 10 and 29 years, with black
males around 13 times more likely to be a victim of homicide than similarly aged white males.
Suspects
A suspect was charged in 60% of cases recorded in the Homicide Index, although is likely to be an underestimate of the proportion who will eventually be charged. 93% of convicted suspects were male and over three-quarters were aged between 13 and 27 years. Across most age groups, convicted homicide suspects were disproportionately black. This was particularly acute between the ages of 15 to 25 years. Almost 60% of convicted suspects were unemployed; one-quarter had a previous criminal conviction.
Victim-suspect relationship
Suspects were most-commonly acquaintances of victims (40% of known victim–offender relationships), although a quarter were strangers. The victim–offender relationship varied with sex: over one-third of female homicide victims were killed by a current or former partner, compared to only 3% of male victims. Although convicted suspects tended to be slightly younger than their victims, their ages were generally similar.
Homicide methods
Sharp instruments accounted for 42% of homicides. Victims in adolescence and early adulthood were disproportionately more likely to be killed with a sharp object. Strangulation accounted for 8% of homicides (with half of those victims being female).
Homicide setting
Homicides of males were more geographically concentrated than homicides of females, reflecting the number of homicides of male victims that were related to drug markets or the night-time economy.
The location and setting of homicide varied by gender: 78% of female victims and 41% of male victims were killed in a home. Across genders, the proportion of victims who were killed in a street or open space peaked in late adolescence (with 73% of male victims aged 10 to 19 being killed in a public place) and decreased consistently with age (for example over 80% of female victims aged 20 or older were killed at home).
Homicide context
Around one-quarter of homicides met the definition of domestic abuse. Homicides of children fell into two distinct groups: children under 12 killed by family members (44% of homicides of children), and children 12 or older killed by acquaintances or strangers outside the home (38% of homicides of children). The context and method of killing in these two categories varied considerably.
Around 13% of homicides were of victims aged 65 years or older, which were around three times more likely to be committed by a family member than homicides of younger victims. Nearly two-thirds of homicides involved substance use by either the victim, suspect or both, while 17% of homicides were linked to mental health. Around 10% of homicides were recorded as gang-related, but this accounted for 34% of homicides of 10 to 19 year olds and 22% of homicides of 20 to 29 year olds: less than 1% of victims of homicides classed as gang-related were female.
Homicide outcomes
Between 2019/20 and 2020/21, a suspect was charged in around 70% of homicides. Investigations often led to a charge relatively quickly: half of suspects were charged within four days of the homicide occurring, and 81% of charges occurred within four weeks of a homicide occurring. The rate of charge in the dataset for 2021/22 was considerably lower, but this likely reflects the time taken for a case to be closed and updated on the Homicide Index database rather than a real change in charging rates.
The proportion of homicides charged varied by homicide sub-type: a suspect was charged in over 80% of cases involving a female victim of domestic homicide but this proportion was 40% for non-domestic homicides with female victims. Of suspects charged, 82% were convicted of at least one offence, of which almost two-thirds were for murder. There were substantial differences in the proportion of homicides that resulted in a charge: some forces had charge rates that were one-third that of similar-sized forces. Charge rates for homicide were associated with the type of homicide: when an incident took place outside a home, the charge rate was around half of that for domestic incidents. Differences in homicide context may account for some of the force-level variation in outcomes.
Clustering homicide
Clustering homicide by victim characteristics and the incident setting appears to be a manageable way to analyse homicide. It provides a useful starting point to think about the monitoring and prevention of homicide for groups with very different patterns of and vulnerability to homicide. However, the current Home Office sub-types may be failing to accommodate the distinct features of elder abuse, and does not distinguish the different patterns of homicide of victims aged less than 10 compared to those in their early teens. These validity of these sub-types should be tested and monitored regularly.
Child abuse and elder abuse
Patterns of homicide are different for very young victims and older victims, relative to working-age adults. The risk of homicide victimisation was highest for children under one, and young children were most likely to be killed by a family member. Similarly, homicide victimisation increased in females over 65 years of age, and older victims were more likely to be killed by a family member than other adult victims. This suggests a categorical difference in the nature of homicide in these two age groups that will require age-specific intervention recognising the increased risk that perpetrators are family members and recognising higher levels of vulnerability in young children and older adults (Caman et al. 2017).
Strategic insights
The low number of homicides in most police-force areas may mean that specific force-level homicide-prevention strategies are unnecessary or infeasible. Given the association between homicide and non-fatal violence with injury, force-level homicide-prevention activities should be integrated into serious violence prevention, which should be separated by violence occurring in different circumstances.
A purely criminal-justice approach to homicide may not only be ineffective but may hinder longer-term efforts to reduce homicide. The unpredictable nature and upstream causes of homicide mean that it is unlikely police alone can effectively prevent homicide. A long-term preventive approach where the causes of violence are understood as the result of individual, interpersonal, organisational, societal and situational factors, is likely to produce fairer and more effective responses than framing violence and homicide as solely a criminal-justice problem.
The inequitable risk of homicide faced by black boys and men aged between 10 and 29 years of age should be considered a public health problem that requires urgent and sustained attention.
The seven Home Office homicide sub-types have shown that the setting and victim characteristics determine the context and nature of homicide incidents, meaning interventions to prevent different types of homicide require different strategies. A national strategy for homicide and local strategies for serious violence should adopt a sub-type system.
Force-level variations in homicide investigation outcomes are partially explained by patterns in homicide characteristics but the differences appear too great to be explained by this factor alone. These differences require further investigation.
The Homicide Index is an invaluable tool for understanding homicide in England and Wales and its maintenance and sustainability should be well-resourced. However, its value for analysing some questions (for example, understanding emerging trends) may be limited by the length of time between police becoming aware of information and it being recorded in the Homicide Index. For example, police are allowed up to 30 days between charging a suspect and adding suspect details to the Index. The Home Office should consider what effect this has on analysis and whether there is anything further to be done to ensure the Homicide Index data is sufficiently up-to-date to be used to produce timely analysis to support homicide prevention.
This executive summary is from the report 'Patterns and trends in homicide in England and Wales and their implications for policing: a problem profile' by Prof Iain Brennan and Dr Matt Ashby.