Security and the Drop in Violence between Acquaintances and Strangers

Security and the Drop in Violence between Acquaintances and Strangers

Our current research again draws on the CSEW. For some types of violence, the potential significance of security improvements is obvious. Armed robbery, for exam- ple, is likely to have been reduced by improved security (banks are the clearest example), but the CSEW is no help here. There may also have been improvements in security in some traditional violence hotspots, such as city centres and bars, but again the CSEW is limited in what it can tell us. What can be examined through the CSEW are patterns of relatively risky behaviours and their precautionary counter- parts that may explain the falls in violence amongst affected groups. Going out less, drinking less and avoiding certain risky behaviours or lifestyles can be measured using the CSEW. While quite significant changes to lifestyle and routine activities appear to have occurred, it is possible that this was after crime began its major decline, which would suggest that their effect was to consolidate the effects of secu- rity. Further research is needed to uncover how changes to lifestyle and routine activities may have helped to reduce violence.

Part of our security hypothesis suggested that general falls in crime (including violent crime) followed from the effects of the reduction in debut crimes that mark the onset of criminal careers, which typically include a wide range of offending including violence. These indirect effects of security on criminal careers are dis- cussed next.

FROM REDUCTIONS IN CRIME EVENTS TO REDUCTIONS IN CRIMINALITY

The security hypothesis outlined above proposes that improvements in security have had a knock-on effect on criminal careers. As the term ‘debut crimes’ suggests, this expectation is rooted in previous research that identified the types of crime that tend to mark the start of criminal careers (Svensson, 2002). Theft of motor vehicles was found to be especially significant. Previous research has also found that prolific offenders tend to be generalists rather than specialists (Piquero et al.,

2014). Hence, if ‘debut crimes’ are inhibited by making them more difficult, as in the case of vehicle theft, this would stifle further criminal activities. If security improvements were reducing crime by inhibiting the onset of criminal careers, we would expect a greater effect among younger adolescents who are new to offend- ing, and for the effect to be less among older offenders. This could be measured by looking at rates of onset and continuance of offending in different age groups.

CSEW data, which relate to victimization patterns, are not of much use in tracking changes in criminal involvement. We therefore looked to other sources and turned to data from the criminal justice system and found a set of data on arrests in the USA that spanned several decades (making it easier to access and use for this purpose than criminal statistics for England and Wales). Arrest data are not perfect but are preferable to data on prosecutions or convictions, where deci- sions about how to process those arrested may change over time and influence patterns independently of changes in actual criminal behaviour. The patterns we found in US arrest data accorded with our expectations, as shown in Figure 10. 5 (see Farrell et al., 2015, 2016).

CONTEXTS, MECHANISMS, SECURITY, CRIME DROPS AND DATA SIGNATURES

In conducting detailed analyses of our data, we have attempted to specify the pre- cise mechanisms through which particular outcome patterns would be produced in the contexts in which the security upgrades have been introduced. For example, in looking at car security devices we have distinguished which crimes electronic immo- bilisers might reduce (theft of cars, with no direct effect on theft from cars). We have looked for specific ‘data signatures’ that would be expected were our security hypotheses to be warranted (Farrell et al., 2016). We find, indeed, that immobilisers reduce the theft of vehicles but not theft from them (Farrell et al., 2011a, b).

In relation to domestic burglary, we have looked to see whether changing patterns of modus operandi are those that would be expected were security to be reducing the rate of burglary. Such changes include much greater reductions in forced entry as against entry that does not require security to be overcome (for instance, where a door or window has been left open). Figure 10. 6 shows that there has indeed been a much greater fall in burglaries requiring forced entry in England and Wales, as meas- ured by the CSEW, when compared to burglaries that did not require forced entry (Farrell et al., 2016).

We have sometimes been surprised where patterns have not accorded with expec- tations and this has led us to further theory development, which should then be open to test, as in the example of household alarms mentioned earlier.