Urban crime rankings

Urban crime rankingsClick here to download the report

The publication of information is a key means to improve the performance of public services. For services such as health and education, it allows users to exercise choice between providers, often supported by advisers such as GPs. There has been a dramatic increase in consumer information for health and education in recent years. For services such as policing, where choice does not apply, it allows local communities to compare the performance of different police units and to demand better performance.

The performance of public law and order agencies in England and Wales needs improvement. Despite recent falls, on the latest surveys crime remains amongst the highest in the developed world. The Cabinet Office Strategy Unit has described England and Wales' performance on violent crime as a "weakness" compared to other countries. The latest British Crime Survey results show that crime ceased falling in the last quarter of 2005.

Accurate, relevant and easily intelligible local data on crime would put pressure on police forces to improve. At present, however, such data is not available:

  • Police recorded crime is a measure of all crime reported to the police and subsequently recorded by them as an offence. Police recorded crime at the national and regional level is published quarterly.
  • The Home Office has recently begun to record crime by Crime and Disorder Reduction Partnerships (CDRPs), based on local government boundaries. It does not however publish a ranking of CDRPs. Only a minority of individual police forces present detailed data on recorded crime in their area although all are required by law to compile it.
  • The annual British Crime Survey is a national survey based on 40,000 adults living in private households. While it records many crimes not reported to the police, such as domestic violence and failed burglaries, it does not record many others including crimes against children and shoplifting.
  • Performance assessments for each police force as part of the Police Performance Assessment Framework (PPAF) are published annually by the Home Office. While full of useful data, the results are opaque to the general reader and resist comparisons between forces on the key issue of crime levels in urban areas.

The independent Statistics Commission has recently criticised the crime statistics for England and Wales. In December 2005 it said: "We think there is a need to explore alternative ways to convey trends in crime concisely and unambiguously - whilst being mindful of the need to avoid adding to public confusion." In January 2006 the Home Office announced a major cross-party review of how crime statistics are compiled and published to increase public understanding of crime trends.

A better way to present statistics on crime may be to measure crime by city. Cities are understandable geographical units in a way that local authority areas or police force areas are not. They also contain the highest levels of crime.

A similar American project, by the researchers Kathleen O'Leary Morgan and Scott Morgan, now in its twelfth year of publication, has shown that presenting crime data by city can have a powerful incentive effect on local agencies. The Prosecutor of Camden County, New Jersey, the city with highest crime in the last two years, has commented that the publication of the rankings "helped to coalesce the Camden Police Department, an array of law enforcement agencies at the county, state and federal levels, the Camden community and its neighbours to focus on the crime problem with new vigour and to find solutions."

Reform has therefore compiled data on levels of recorded crime in 2005 for each city in England and Wales with a population of over 100,000. Information where necessary for specific offences was obtained using Freedom of Information requests to police forces.

The crimes covered were: murder, rape, assault, burglary, robbery and vehicle crime. These were chosen to allow accurate comparisons and to follow the precedent set by the American edition. These offences were deemed to be readily understandable and popularly associated with the main types of crime perceived as a threat to one's property or person. And so while it does not cover some high frequency crimes such as criminal damage or drug offences, it can nevertheless be taken as a starting point. Together the offences surveyed typically represent a third (34 per cent) of all property crime and more than two thirds (69 per cent) of all violent crime.

Crime levels were divided by population totals for the town or city in question (drawn from the ONS mid-year population estimates) to give a rate per 1,000 (for robbery, burglary, vehicle crime and overall crime), per 10,000 (for rape) and per 100,000 population (for murder).

The results produced the following key findings:

  • Dramatic variation between the best and worst performers. At 105 crimes per 1,000 population, Nottingham had almost five times the level of crime as the safest town in the rankings: Colchester, which recorded 22 crimes per 1,000 population, and almost twice the average crime rate of 57 crime per 1,000 population for all the places surveyed.
  • Dramatic variation between towns of similar size. Nottingham's crime rate can be contrasted with the much better performance of towns of a similar size (approximately 275,000 people) in other parts of the country, such as Stockport (49 per 1,000) or Newcastle (45 per 1,000).

In terms of the different categories of individual offences, the results found:

  • Nottingham had the highest number of murders per 100,000 population, followed by Southend. Brighton and Hove and Cambridge had no murders in 2005.
  • Portsmouth had the highest number of rapes per 10,000 population, followed by Middlesbrough. Exeter and Swansea had the lowest.
  • Leicester had the highest number of assaults per 1,000 population, followed by Bradford. Colchester and York had the lowest.
  • Nottingham had the highest number of burglaries per 1,000 population, followed by Bradford and Hull. Colchester had the lowest rate.
  • Manchester had the highest number of robberies per 1,000 population, followed by London and Nottingham. Rotherham and Poole had the lowest.
  • Nottingham had the highest rate of vehicle crime per 1,000 population, followed by Manchester. Ashford and Colchester had the lowest.

It could be argued that the London boroughs should be included as separate "cities" in the overall ranking since each has a population of over 100,000, and they are incorporated in this way in Table 13.

  • The results for London also show a wide variation between the safest boroughs (Sutton, Kingston upon Thames and Richmond), and the most dangerous (Westminster, Hackney, Islington and Southwark).
  • When London boroughs are included in the overall ranking for serious offences, five out of the ten worst performing areas, and 11 out of the twenty most dangerous towns and cities in the rankings are in London.

These rankings can be compared with the Home Office's data on crime by Crime and Disorder Reduction Partnerships (CDRP) and the performance assessments for police forces. The CDRP data shows that local authorities such as Nottingham, Hull, Bristol and many London boroughs have comparatively high levels of crime across a range of offences. The performance assessments for police forces for "reducing crime" give Nottinghamshire Police and the Metropolitan police poor results and Essex Police an excellent result.

More importantly, the evidence contained in this report makes clear that while some areas of the country have relatively low, or at least not exceptional levels of crime, a number of our larger towns and cities remain high crime areas with a wide variation between the worst and the safest localities which is all too easily masked by national figures.

The findings suggest that the Home Office's key target - that the crime in high crime areas should fall more quickly than in other areas - is insufficiently challenging. It is, however, local rather than central initiatives that will have the greatest impact on crime.

As in Camden, New Jersey, the data presented here should provoke much greater efforts on the parts of local communities, and their police forces, in areas of high crime. Such efforts could focus on understanding the good practice of forces such the Essex Police and techniques pioneered by police forces in the United States.

These efforts would be greatly accelerated if the police were made accountable for their performance. Reform has previously argued that local police authorities do not make forces accountable to their communities, with the result that there is little incentive to improve performance. New forms of accountability, alongside better and clearer information on levels of crime in urban areas, should drive the change in police performance that many British cities need.