Does Stop and Search Deter Crime? Evidence From Ten Years of London-wide Data

Despite recent declines in its use stop and search remains one of the most controversial powers vested in the police in the United Kingdom. Yet there has been surprisingly little research into its effectiveness as a tool of policing, significantly impeding debate around this topic. In this paper we used ten years of police, crime and other data from London to investigate the potential effect of stop and search on crime. Using lagged regression models and the natural experiment created by the sudden increase and subsequent decrease in the use of s60 powers in the capital, we show that the effect of stop and crime is likely to be marginal, at best. While there is some association between stop and search and crime (particularly drug crime), claims that this is an effective way to control and deter offending seem misplaced. We close the discussion, however, by arguing that this is somewhat to miss the point, and for two reasons. In a legal sense the key issue is that each and every stop is justified in itself, not because it has some putative wider effect on crime. In a sociological sense, our findings support the idea that stop and search is a tool of social control widely defined, not crime-fighting, narrowly defined. Introduction Use of police powers to stop and search (S&S) members of the public has fallen significantly in England and Wales over the last few years. The number of recorded searches in 2014/15 was approximately 541,000 down by 58% from a peak of almost 1 million in 2008/9 (Home Office 2015). Yet use of the power remains a controversial issue. The decline in recorded searches has not be accompanied by a similar reduction in the ethnic disproportionality in their application; in 2014/15, people who identifying as Black or Black British were still 4 times more likely to be searched than their white counterparts (Home Office 2015). S&S can still trigger significant reactions from individuals and groups who experience or observe its use, and wider social and political debates, as current disproportionalities entrench and interact with previous evidence of bias and discrimination. Our focus in this paper is not on ethnic disproportionality in S&S, which has been well documented elsewhere (Equality and Human Rights Commission 2010; Bowling and Phillips 2007; Quinton 2015; Bradford and Loader 2016). Nor are we concerned with the wider social and cultural ‘meaning’ of stop and search – as a tool of social control (Choongh 1997; Bradford and Loader 2016), for example – although we return to this question in the conclusion. Rather, we are concerned with whether S&S deters crime. This is a salient issue for two inter-connected reasons. First, despite reductions in its use, S&S remains one of the most widely used formal powers of the police. Yet very little is known about its effect on crime (Delsol 2015), particularly in a UK context. Second, it remains a commonplace of media and police accounts of S&S – perhaps particularly in relation to the recent reduction in its use – that S&S ‘must’ have an effect on crime (and therefore that restricting use of the powers risks an increase in crime). While the law governing S&S tends to revolve around the investigation of crime, there is no doubt that police officers and many observers of police activity take a broader view, believing and S&S has a deterrent effect. This makes consideration of its likely effect on crime a pressing policy concern – yet one which has rarely been addressed with appropriate empirical data. To address this question we use 10 years’ worth of S&S, crime, and other data from London, aggregated at the borough level, and we take two distinct analytical approaches. First, we utilize fixed effect regression models estimating the lagged effect of S&S on crime. Second, interrupted time series analysis is used to explore the potential effect of the sudden rise in the use of ‘suspicion-less’ or authorised searches that occurred from 2007 to 2011. To anticipate the results, we find that S&S has only a very weak and inconsistent association with crime. While there is some correlation, most notably in relation to drug offences, we conclude that the deterrent effect of S&S is likely to be small, at best. Setting the scene The ‘power’ of S&S in England and Wales actually comprises a range of powers governed by a number of pieces of legislation that enable officers to search for a range of items. Most well-known is section 1 of the Police and Criminal Evidence Act 1984 (hereafter s1), which enables searches for stolen goods and a range of prohibited items (such as offensive weapons). Additional powers include those granted under: section 23 of the Misuse of Drugs Act 1971 (s23) for controlled substances; section 47 of Firearms Act 1968 (s47) for firearms; section 60 of the Criminal Justice and Public Order Act 1994 (s60) to prevent acts of serious violence; and section 44 of the Terrorism Act 2000 (s44, since repealed) to prevent acts of terror. For the purposes of this paper we distinguish between two broad groups of powers – those that require officers to have ‘reasonable suspicion’ before conducting a search (s1, s23 and S47) and those that do not but, instead, require them to be authorised to carry out searches in a defined area for specified time period (s60 and s44). This is an important distinction in terms of the way the powers are used, their purpose in law, and, more importantly, in terms of the level of debate that surrounds them – s44, in particular, was one of the most controversial powers vested in the police in modern times (Shiner and Delsol 2015). Although S&S has come be to be one of the most emblematic, indeed arguably foundational powers, of the police in England and Wales, the extent of its uses varies markedly across time and place (Bradford 2017). The data used in this paper cover London during the period 2004 to 2014. By way of an introduction, Figure 1 shows that while recorded crime declined gradually but consistently over this period, S&S showed marked variations month-on-month and over the 10-year study period. The recorded use of reasonable suspicion searches increased between 2004 and 2010 and then went into steady decline. Authorised searches that do not require officers to have reasonable grounds were rarely used before the middle of 2007; use then increased markedly, reaching a peak in 2008, after which they too went into steady decline. After a further peak in August 2011 (which coincided with the London riots), usage again became rare. Figure 1. Trends in searches and crime (Metropolitan Police, 2004-14) Note: Susceptible crime = those crimes that are susceptible to detection by S&S. The deterrent effect of S&S We are concerned in this paper with the deterrent effect S&S may exert on potential offenders; particularly whether a marginal change in the S&S rate in an area can be linked to a subsequent change in the crime rate in the same area. It is worth reiterating that this is not the legal justification for most searches. While the authorised search powers are framed in legislation explicitly in terms of prevention, the more commonly-used reasonable suspicion powers are framed as investigatory tools aimed at apprehending the perpetrators of crimes that have already been or are in the process of being committed (a point we return to in the conclusion). However, considering the mechanisms by which S&S might reduce crime, and given the significant conceptual overlap between apprehension and deterrence as well as the nature of public and policy debate around S&S (very often framed in terms of deterrence), we believe that this is a justifiable starting point for our analysis. Deterrence theory outlines distinctions between different ‘varieties’ of deterrence. In the case of S&S the most pertinent relate, first, to the certainty of apprehension as a lever for deterring 1 PACE Code of Practice A states that the purpose of S&S is “to enable officers to allay or confirm suspicions about individuals without exercising their power of arrest (Home Office 2014: para 1.4). potential offenders. It is usually argued that what deters people from committing a crime is not the severity of any punishment that may ensue, nor the speed with which it will be delivered (celerity), but their perception of how likely they are to be caught if they were to commit the offence (Apel and Nagin 2011, Nagin 2013; see also Pratt et al. 2009). If it is to have a deterrent effect, S&S, must make acts of crime appear riskier to potential offenders by suggesting that they are likely to be caught if they do break the law. Second, deterrence theory rests on the contrast between specific and general deterrence: to whom is the risk of apprehension communicated? Specific deterrence functions at the individual level, referring here to the effect of S&S experiences on offenders – and others – who have been interdicted by police. A proportion of those searched will be arrested or handed an alternative sanction such as a cannabis warning or fixed penalty notice. Having been caught ‘red-handed’, it naturally seems possible that these individuals might be deterred from committing a crime in the future: being searched and possibly receiving a penalty at one time may change their risk calculus in relation to future potential acts. Likewise, simply being stopped and searched, even when one has not broken any law, may have a future deterrent effect on one’s behaviour. General deterrence, by contrast, operates at a wider level, and refers to the effect S&S awareness might have on the behaviour of the general population – as potential offenders – who see or hear about this type of police activity or who merely know the police can carry out S&S. Witnessing or having some knowledge about S&S may shift people’s risk perceptions, leading them to believe, for example, that police are effective ‘sentinels’ (Nagin et al. 2015) who are capable of apprehending offenders. Deterrence t