25 Jul 2011

Does fewer police really mean more crime?

With a savings target of about £1.6bn, police managers are planning a 10% reduction in the number of officers in England and Wales by 2015.  Last week Her Majesty's Inspectorate of Constabulary (HMIC) released the first reliable estimate of the effect on individual police forces, along with a research paper (PDF) suggesting tentatively that this reduction could lead to an increase in property crime of around 3%.

The HMIC analysis estimates that the police workforce in England and Wales will be reduced between March 2010 and March 2015 by 16,200 police officers, 1,800 Police Community Support Officers and 16,100 police staff (total 34,100).  This would take the overall size of the workforce back to the 2003/04 level, with police officer numbers reverting to 2001/02 levels.

Just under a third of the 34,100 total reduction has already taken place.  According to the latest Home Office statistics on police service strength, there were 139,110 full-time equivalent (FTE) police officers in the 43 police forces of England and Wales as at the end of Q1 2011.  This was a decrease of 3.2 per cent or 4,625 officers compared to a year earlier.

The following graph shows the trend in police numbers since 1994/95, projected to 2014/15:


The issue of whether reducing police numbers -- and in particular reducing the number of front-line police officers -- will actually make a material difference to the crime rate is rather sensitive for the Government.

There is a broad consensus among criminologists that police numbers are not a significant variable affecting rates of crime generally, when compared with socio-demographic and economic drivers.  This is not to say that the police have no effect on crime statistics, simply that policing initiatives and specific decisions on the use of police resources make more of a difference than the number of available officers.

The problem for the Government and the Home Office in particular is that this reality is somewhat removed from the public perception.   As discussed in an earlier post, crime statistics are highly politicised; we cannot be entirely confident in the accuracy of police recorded crime figures, and most members of the public tend to subjectively over-estimate the scale of crime as a problem in UK society.

To a large extent this problem is driven by inaccurate and unrepresentative reporting of crime in the popular press.  However over the past decade politicians from all of the major parties have been keen to trumpet their support for increases in police numbers ("more bobbies on the beat") as if that alone would provide greater security from the risk of crime.

The key research paper (PDF) released on the HMIC website was written by Ben Bradford, an LSE Fellow at the Methodology Institute, London School of Economics.  The paper is short and highly readable, and consists mainly of a review of 13 past studies that look at the relationship between police numbers and/or arrests and crime.

The author has worked with the London Metropolitan Police on a number of research projects, but based on his other writings there is no obvious reason to suppose he is partial to any particular agenda.  Bradford has written on his blog about the process that led to his findings in the paper, and notes that some of the evidence challenged his preconceptions.  He also makes it clear that the evidence for a correlation between police numbers and crime is suggestive rather than conclusive.

The debate over whether fewer officers equals more crime is really focussed almost entirely on street-level property offences such as burglary, robbery, vandalism and vehicle theft.  These offences can arguably be discouraged by a front-line police presence in the community, provided the ratio of officers to population and geographic area is sufficient.

Any effect will likely apply to property crime against local businesses as well as householders, although this relationship would be difficult to track due to patchy recording of crimes against business premises -- and businesses are more likely than householders to compensate by increasing their physical security, making use of private security staff, and managing their risk economically through recourse to insurance.

Police involvement in countering white-collar property crimes such as fraud and cyber intrusion is the work of specialised units and unlikely to be directly affected by an across-the-board reduction in police numbers.  These crimes already receive attention from regulatory agencies and funded national initiatives in addition to regional police resources.  We should also anticipate increased funding of specialised police units by the private sector, as for example the recent announcement of a fraud unit within the City of London police to be funded by the insurance industry.

Similarly, while we can expect anti-terrorism work to remain a leading police priority, the national security services will remain extensively funded.  Anti-terrorism initiatives tend to be intelligence-led and reduced police numbers will only present a problem in the aftermath of a serious incident -- a rare occurrence, and a resource concern principally for particular urban forces rather than police managers in general.

We may reasonably anticipate problems in the policing of protest marches and events, especially if there is a significant upswell in social unrest as funding cuts to other public programmes feed through over the next couple of years.  As protest groups become more organised and familiar with police tactics, the availability of sufficient numbers of trained officers will have a material effect on the ability of forces to maintain control of events.  This is an issue for the Met Police especially of course, but also for regional forces in light of the increase in activity by the far-right English Defence League and similar groups.

As Bradford notes, the evidence for an association between police numbers and violent crime is weaker.  Violent crimes are usually committed on impulse and few offenders consider the likelihood of detection.  In practice however a reduction in police numbers may reduce the availability of officers to respond adequately to public disorder problems associated with pub closing times on the high street, and to domestic violence calls -- inevitably the attritional offences that are already the lowest priority will bear the brunt of reduced resources.

Ultimately however it is unlikely that the HMIC will be able to develop any more definitive evidence that reducing police numbers, on the scale under discussion, will seriously increase crime rates.  Current methodologies for recording crime statistics are not robust enough to support detailed analysis.

Cuts to police funding are part of a broader reduction in public spending, and the effect on policing outcomes will therefore be buried by the broader disruptive effects of cuts to housing, education and social welfare programmes.  Given that police numbers have increased steadily over the past ten years, the reductions in force numbers under discussion should be manageable when viewed on a national scale.  At the local level however there is scope for additional scrutiny of the effects on specific policing initiatives.