Skip to content

A stocktake of crime statistics in England and Wales


After identifying an error in published tables a minor revision has been made to this release. The error related to data on the proportion of people who have been victims of plastic card fraud for the last two years in the time series (Oct-12 to Sep-13 and Oct-13 to Sep-14).  Revisions have been made to figure 14 in the statistical bulletin (also contained in reference table 01. Bulletin Tables - Crime in England and Wales, Year Ending September 2014) and to the short story: A stocktake of crime statistics in England and Wales.

23 April 2015 at 11:00 am

Trends that have been driving crime rates and the complexities around crime statistics

Every quarter, when we release new statistics, there is considerable interest in whether crime has gone up or down. The two main sources: the Crime Survey for England and Wales (CSEW) and police recorded crime, both have strengths and limitations and together provide a more comprehensive picture of crime than could be obtained from either series alone. However, neither source provides a picture of total crime.

While having two data sources is valuable it can also cause confusion, especially when they appear to have conflicting findings. This article outlines the complexities around these statistics and looks at some of the factors that have been driving trends. It also discusses what we can deduce has been happening to crime and examines trends for some individual offences.

For the population and offences it covers, the CSEW provides a valuable measure, on a consistent basis, of trends over time.  The survey covers a broad range of victim-based crimes experienced by the resident household population. It includes crimes that have been reported to the police, as well as those that haven’t been. However, it does not cover all crime. We are planning to extend the coverage of the CSEW so that it includes offences of fraud and cyber crime against the household population in future surveys. The survey doesn’t capture some serious, yet relatively low volume, crimes such as murder, rape and other sexual offences. Nor, being a survey of the household population, does it cover offences against other types of victims, such as shoplifting experienced by retailers, or so called “victim-less” crimes, such as drug possession and motoring offences, like speeding. This infographic shows what is covered in the survey.

While many of the offences not included in the survey are covered in the police recorded crime series, those statistics are restricted to crimes that have been reported to the police and subsequently recorded by them. There has been recent criticism on the quality of crime recording by the police and so trends need to be interpreted with caution.

The latest estimate from the CSEW is that 7 million crimes were committed against households and resident adults in the last year, compared with a peak of 19 million in 1995. This is lowest level since the survey began in 1981 and continues a downward trend seen since the mid 1990s. This is not unique to England and Wales; there have been similar trends across the western world. This figure of 7 million becomes more meaningful when related to the size of the population and broken down to show how the levels of victimisation (the proportion of households or people that have fallen victim) vary by crime type and across the population. Therefore, to assess whether or not crime levels are changing it is more useful for decision-makers to consider individual measures alongside each other, rather than focusing on a single headline figure. In addition, a focus on a single figure risks missing more important findings when looking at specific crime types.

Figure 1 shows comparative victimisation rates for a range of property crimes included in the survey. It illustrates how the long-term reduction in crime, measured by the survey, has changed the proportion of households and adults falling victim to crime.

Figure 1: CSEW property crime victimisation, 1995 and year ending September 2014.

Figure 1: CSEW property crime victimisation, 1995 and year ending September 2014.


  1. Vehicle-relate theft victimisation rates relate to vehicle-owning households only.
  2. Bicycle theft victimisation rates relate to bicycle-owning households only.

Download chart

The latest figures show households are now around a third as less likely to be a victim of burglary than in 1995 when such crime peaked, with 3 in 100 households falling victim in the year ending September 2014, compared with around 9 in 100 in the 1995 survey.  Around 4 in 100 vehicle-owning households experienced a vehicle-related theft1, compared with around 20 in 100 in 1995.

These trends in burglary and vehicle-related thefts are thought to reflect what criminologists describe as ‘target hardening’, making it more difficult for criminals to commit crime, for example by improving household and vehicle security. It may also reflect a change in the value of goods and their attractiveness to criminals.

People are now far more likely to be carrying valuable goods on their person. This is shown in the CSEW, where the value of items stolen in theft from the person offences, has doubled over the past decade2. It is thought that the theft of smartphones was a key driver in a rise of theft from the person offences recorded by the police in each of the last three years. The subsequent introduction of new security features by manufacturers is thought to have lessened the attractiveness of such devices, leading to the recent downward trend in theft from the person offences in the latest findings.

Advances in technology and the rise of the internet have provided new opportunities for criminals to commit crime. This has raised questions as to whether the fall in conventional crimes, as described above, has simply been replaced by new types of crime that are not yet well measured by the statistics. While fraud and cyber crime are not currently included in the headline CSEW estimates, the survey has included questions on bank and credit card fraud since the 2005/06 CSEW. The survey showed a rise in the level of victimisation, followed by a reduction after chip and pin technology was introduced (another example of ‘target hardening’). The latest CSEW estimates show that 5.0% of card owners were victims of card fraud in the last year, showing no change compared with the previous year (Figure 2). This level of victimisation is higher than more traditional offences, such as theft from the person (1%) and robbery (0.2%). 

Figure 2: Proportion of CSEW plastic card users who had been a victim of plastic card fraud in the last year, 2005/06 to year ending September 2014

Figure 2: Proportion of CSEW plastic card users who had been a victim of plastic card fraud in the last year, 2005/06 to year ending September 2014

Download chart

As far as the household population is concerned, this suggests some of the fall in CSEW crime since 1995 has moved into these new types of crime not included in the headline estimates; but not at a volume sufficient to eradicate the substantial falls described above.  There is work ongoing to incorporate measures of fraud and cyber-crime into the main Crime Survey estimates.

What do other sources tell us has been happening to fraud and cyber-crime? There have been improvements in administrative data on fraud and cyber-crime. Since April 2013 Action Fraud, a public facing national reporting centre, has taken responsibility for the centralised recording of fraud offences, which were previously recorded by individual police forces. The latest figures show around 213,000 fraud offences were reported to Action Fraud in the year ending September 2014. To put this in context, this is equivalent to 4 recorded offences per 1,000 population; over twice the rate of theft from the person and four times the rate of robbery recorded by the police.  Following the transition to Action Fraud there was an initial rise in the volume of fraud recorded by them. However, it is difficult to judge to what extent this increase in recorded fraud offences reflected improved recording practices, an increase in public reports, or a rise in actual criminality. More recent data shows levels of fraud reported to Action Fraud have remained generally flat over the past 18 months. 

Action Fraud refer all fraud offences recorded by them to the National Fraud Intelligence Bureau (NFIB), based at the City of London Police who are the national lead for investigating such offences. The NFIB also receive reports direct from industry bodies, the latest data shows they received 391,000 offences in the year ending September 2014. The NFIB analyses all reports of fraud referred to them and distributes those with investigative leads to forces to follow up. These data sources will only include crimes that have been referred to the police and it is not known whether levels of unreported fraud have been rising or not.

As physical violence is a crime that can’t be committed on-line we can have more confidence that reductions in levels of violent crime reflect real changes in society. The latest survey estimates that around 2 in every 100 adults were a victim of violent crime, compared with around 5 in 100 adults in 1995. These averages mask significant variation by type of violence and across the population. For example, men aged 16-24 were most likely to suffer violence at the hand of strangers (4.6% in the 2012/13 survey), while older women were least likely (0.1% of women aged 75 and over). In contrast, younger women were most likely to experience domestic violence, with older men least likely (for example 12.5% of women aged 20 to 24 compared with 2.4% of men aged 55-59). More detail can be found in the latest Focus on Violent crime and Sexual Offences 2012/13 publication.

The survey does not cover homicide (that is murder, manslaughter and infanticide3), but police recorded crime is thought to provide a comprehensive picture of such crimes, and is not subject to the same concerns around the quality of recording as is the case for some other offences. There has been a fall in the number of homicides over the last decade, with the number in the latest year (507 homicides) the lowest since 1977. To put these figures in context, the number of homicides in 2012/13 was three times lower than deaths arising from transport accidents (Focus on Violent crime and Sexual Offences 2012/13 publication).

In contrast with the latest figures from the survey, the police recorded crime series show a 16% rise in the volume of violence. What is behind this apparent discrepancy? There is good evidence to suggest that the rise in recorded violence reflects changes in police practice, rather than levels of crime. It is known that violent offences are more prone to subjective judgement by the police about whether or not to record the offence. There has been a renewed focus in the last 12-18 months on the quality of crime recording and a drive to improve the police service’s compliance with national standards for recording. This is thought to have led to a greater proportion of reported crimes being recorded. Nearly all of the police forces (40 out of 43) recorded an increase in violent crimes compared with last year, with the largest volume increase shown by the Metropolitan Police Service. It has been suggested that the rise in London is partly attributable to a greater proportion of reports of violence being recorded by the police as crimes than previously. Further evidence that levels of violent crime have actually been falling is provided by several health data sources on violent-related attendances or admissions at hospital. These patterns are consistent with the trend shown by the CSEW.

Indirect evidence of a possible change in police recording practice can be seen by looking at trends in comparable crimes included in both the CSEW and police recorded crime. In the case of the CSEW this comparison is restricted to those offences that respondents say have been reported to the police. Figure 3 shows that in recent years there had been a growing gap between the two sources, suggesting a reduction in police compliance with national recording standards. However, the gap between the two data series appears to be closing. This is thought to have been driven by the improvement in the recording of violent offences since Her Majesty’s Inspectorate of Constabulary (HMIC) carried out a national inspection of crime data integrity.

Figure 3: Police recorded offences as a percentage of the CSEW estimates (in comparable sub-set)

Figure 3: Police recorded offences as a percentage of the CSEW estimates (in comparable sub-set)

Download chart

Recording issues are also thought to have played a part in the large rise in sexual offences in the last year (up 22% from the previous year). When auditing a sample of records from police forces for the 12 months ending September 2013 the HMIC inspection found a high level of under recording of sexual offences. Action taken by forces in the interim to addresses this weakness is likely to have led to improved recording. It is also likely that other factors have contributed to the rise, including increased willingness of victims to come forward to report such crimes in the wake of high profile investigations and inquiries of child sexual abuse.

Crime is a complex picture and the statistics differ in their utility and coverage according to the type of crime and type of victim. Work is on-going to develop and test questions covering fraud and cyber-crime for future inclusion in the CSEW. The latest progress of this work can be found in the methodological note: Work to extend the Crime Survey for England and Wales to include fraud and cyber-crime (107.1 Kb Pdf) . By improving our measures of crime we aim to provide better statistics to inform better decision making by those who use our data.


  1. Vehicle-related theft includes: theft from vehicles, theft of vehicles and attempted thefts of and from vehicles.

  2. See Table 7.4 in Nature of Crime Tables, 2013/14 – Personal and other theft

  3. Infanticide applies to infants aged under 12 months killed by the mother while of a disturbed mind.

Categories: Crime and Justice, Crime, Crime Trends, Crime in England and Wales
Content from the Office for National Statistics.
© Crown Copyright applies unless otherwise stated.