The use of stop and search has reduced in recent years in London, but there have been recent calls to increase its use. Matteo Tiratelli argues that a new study finds little, if any evidence, to suggest increased stop and search can reduce levels of violent crime.

• As violent crime has risen, there’s been a backlash against recent reductions in ‘stop and search’
• Using ten years of Metropolitan Police data, a large number of possible associations between stop and search under different powers and different crime types were tested. The central finding is that the effect of stop and search on crime is marginal, at best
• If we are interested in policy tools which will reduce the overall level of crime, particularly violent crime, then there’s not much evidence to suggest that forcing/empowering officers to do extra searches on each patrol is going to be effective

Stop and search has been a controversial topic over the last few years. In 2014 the then Home Secretary, Theresa May, told MPs that as many as 250,000 street searches were probably carried out illegally last year and called for significant reductions in their use. In London these changes were already being championed by the commissioner of the Metropolitan Police, Sir Bernard Hogan Howe, who boasted that they had been reduced by almost a third. But, recently, as violent crime has risen, there’s been a backlash. Last year the new Met chief, Cressida Dick, called for more stops and searches. And earlier this year London’s Mayor, Sadiq Khan, went back on his election pledge and revealed that police would be ‘significantly’ increasing stop and search in London.

There are many different angles to approach stop and search form. Reports have shown a startling disparity in the likelihood of black and Asian men being stopped and searched. There have also been investigations into its impact on communities, on trust in the police and its possible role in the 2011 Riots. But, it’s rare to see people explicitly assess whether changes in the level of stop and search deter people from committing crimes. This is the implicit, ‘common-sense’ idea that lies behind recent attempts to use the overall level of stop and search as a policy tool to reduce crime. But, despite a feeling amongst many officers and policy-makers that ‘it must have an effect’, there’s not much evidence to base these policies on.

Testing ‘common-sense’ assumptions

Our study aimed to test this ‘common-sense’ assumption. Using ten years of Metropolitan Police data (2004-2014), grouped in months and weeks within each London Borough, we tested a large number of possible associations between stop and search under different powers and different crime types. The central finding is that the effect of stop and search on crime is marginal, at best.

Although you can never prove a null hypothesis, there’s precious little evidence of a meaningful effect. We find some associations, particularly suggesting that stop and search might be reducing the number of recorded drug offences, but the overall picture is of tiny and inconsistent effects. Given recent trends in London, we were particularly interested in the connection between stop and search and violent crime. Looking initially at non-domestic violent crime we found no real evidence of an effect. The tiny association between section 1 and section 47 (weapon) searches showed that a 10% increase in stop and search would lead to a 0.01% decline in crime, but this effect disappeared when we looked across months and other search powers. When we tested the same models using ambulance incident data for calls related to ‘stab/shot/weapon wounds’, we found no significant effects whatsoever. This all suggests that, if there is any association between the overall level of stop and search and crime, it is likely to be at the outer margins of social and statistical significance.

This finding echoes earlier studies. A Home Office report into the impact of Operation BLUNT 2 (a knife crime initiative involving a large increase in section 60 searches in some Metropolitan Police borough) found no effects. There were similar findings from a variety of other studies looking at New York, London and Chicago, all of which we describe in our paper. If this is the evidence base for Sadiq Khan’s policy proposals, then it’s not a strong one.

Matching the evidence with policy

What should we conclude from this? If we are interested in policy tools which will reduce the overall level of crime, particularly violent crime, then there’s not much evidence to suggest that forcing/empowering officers to do extra searches on each patrol is going to be effective. But, this was never the legal justification for stop and search in the first place. The question should not be about if token stops and searches will deter potential offenders, but whether each and every stop is legally and operationally justified.