Targeting the Most Harmful Offenders for an English Police Agency: Continuity an...

Like most police forces in England and Wales, Northamptonshire Police have traditionally been place-based in most of its crime analyses, focusing on the profiling of demand and resources by jurisdictional boundaries. In 2017, the national police inspectorate (HMICFRS) recommended that police forces rethink their approach to offender management, with particular regard to the most prolific and dangerous offenders. Up to this point, these programmes had been primarily driven by considerations of crime volume. The recent development of the Cambridge Crime Harm Index (CCHI) afforded an opportunity to examine this issue from the perspective of the total harmfulness of each offender and not just the frequency of their offending (Sherman et al. 2016b). This article reports on the first application of the CCHI to a population of detected offenders over a multi-year period.

In this analysis, we aim to build a better understanding of long-term, multi-year offending patterns by assigning weighted harm scores to police recorded crime data using the CCHI. Our aim is to contribute better evidence to decisions about how to set priorities for allocating scarce police resources. Our premise is that understanding the patterns of both harm levels and frequency of offending among the known population of offenders is a key part of evidence-based targeting. Moreover, such targeting depends heavily on systematic evidence of offending patterns from systematic tracking of the entire population of detected offenders. These aims lead us to address our overall research question: How concentrated is the total harm of detected offences among the complete list of detected offenders in a given year in an English police agency, and how consistent is the list of highest-harm “felonious few” individuals from one year to the next?

This general question is comprised of three analytic questions:

1. 1.

   What are the distributions and trends for the harm and frequency of detected crime across all offenders and repeat offenders, both within each year and year on year?

2. 2.

   Does offending among detected offenders escalate, remain steady or de-escalate in harm, frequency or intermittency over time?

3. 3.

   Is there a small felonious few subset of high-harm offenders in any given year, and if so do the individuals in that group stay constant or change substantially over time?

Our major limitation in answering these questions is a lack of access to prison records, which could enable calculations across offenders that are adjusted for time at risk in the community. Thus, our questions are all qualified by the premise of analysing people who were at large in society for at least part of each year that they come to police notice as suspects in new crimes.

Previous research on offender tracking highlights just how important it is to understand patterns of offending. Research focused on onset, prevalence, escalation, seriousness and cessation can help to identify desistance and termination opportunities that can arguably be targeted through crime control policy (Brantingham and Brantingham 2013). Chronic offenders are seen as an extreme within the general offending population and are variously characterized as habitual or “dangerous”. Yet offenders who commit a high volume of offences over extended periods of time (Blumstein et al. 1986) are not necessarily the most serious offenders, and the most serious are not necessarily the most frequent. Long-term offending research from developmental studies has focused primarily on frequency. That question is certainly important, but it leaves a gap in knowledge about short-term patterns of offending to which police can allocate resources. That gap can help criminology to consider whether there is a small group of offenders who are responsible for the majority of harm in short periods of time in any community and, if so, whether they are persistently and consistently harmful.

Tracking Offenders Over Time

Why criminals start offending, why criminals stop offending and what they do in-between have long been of interest to researchers. A notable early study that focused on life course offending is the Philadelphia Birth Cohort Study (Wolfgang et al. 1972). This research traced the developments of boys from birth until eighteen and found that 18% of the offenders could be defined as “chronic recidivists” and were responsible for more than half of the total offences committed. The “Pareto curve” concentration of the volume of crime events among a small proportion of offenders has had a substantial impact of criminal justice policy. It has also led to the identification of such concentrations in other dimensions of crime, in which a small number of units account for a large proportion of total crime event volume (Sherman 2007). For example a small proportion of all locations experience most crimes (Sherman et al. 1989), a small proportion of victims suffer the majority of crime harm (Farrell 1995; Dudfield et al. 2017) and a small proportion of all offenders have the most criminal convictions (Farrington et al. 2013).

The phenomenon of high concentration among a small part of a population is often described as “the power few” (Sherman 2007), but this phrase has been attacked by members of the public who find the phrase too complimentary to offenders who cause so much harm. More recent discussions of this group have denoted its members as the felonious few, as distinct from the “miscreant many” (Sherman et al. 2016a). In this article, we use interchangeably the terminology of “power few” (in reference to the statistical concept of a skewed distribution as a “power curve”) and the felonious few, in order to acknowledge a transition in terminology in the field of evidence-based policing. Yet whether the concentration being described is measured by offending volume or by the sum of its severity can make a big difference in the meaning of the concept. The total volume of an individual’s crimes, as recent evidence shows (Barnham et al. 2017; Bland and Ariel 2015), is not a good predictor of the total harm from an individual’s crimes.

That recent evidence comes from jurisdiction-wide offender census measures, rather than from the birth cohort studies that tended to reach the opposite conclusion. In the London birth cohort of some 400 males, for example, Farrington (1978) concluded that violent offenders tended to be more frequent or chronic than non-violent offenders. Summarizing the issues, Piquero (2000) stated that across a range of research, offenders who commit one or two crimes will be less harmful than those who continue to offend frequently. Yet the Philadelphia birth cohort study, in contrast, found that by 30 years of age, chronic offenders were no more dangerous than infrequent offenders. Multiple offending did not imply more serious offences, and increasing career length did not mean an escalation in offences (Weitekamp et al. 1995). In any case, much larger populations can reveal more granularity in differences across offenders, which could explain the different conclusions drawn from different methods. A large population of frequent offenders may display much wider variation in the total seriousness of their crime history.

Specialist Offenders and Generalist Offenders

Another dimension of criminal offending research seeks to explore categories of the “specialist offender” and the “generalist offender”. Those who offend in the same or similar category of offence over time are described as a specialist offender. “Generalist offenders” are those who are involved in a mixture of offence classes, also known as “offence switching” (Brame et al. 2004, p. 202). It is speculated by developmental criminologists that frequent offenders are more likely to be flexible (generalist) in their offending types, as opposed to less frequent offenders (Loeber and LeBlanc 1990; LeBlanc and Loeber 1998). This is supported by Chaiken and Chaiken and Chaiken’s (1982) research, which was reanalysed by Spelman (1994), finding that offenders with multiple crime types in their record committed about twice as many crimes as specialists. This led to the conclusion that the more frequent the offender, the more versatile they are across crime categories.

Escalation

Other researchers have addressed the question of whether more frequency or variety leads to escalation from less serious to more serious offences. Liu et al. (2011), for example, reviewed the 1953 birth cohort from the England and Wales Offender Index up until 1999. They found evidence to support both de-escalation being associated with ageing and escalation being associated with higher numbers of convictions. Yet whether de-escalation of seriousness applies to all offenders, or excludes a high-harm subset that could be called a felonious few, remains unclear from prior research. The lack of clarity is largely due to the small sample sizes of cohort studies, as well as the discordance between cohorts and the full population police must contend with. None of the developmental studies, to date, have examined entire populations embracing all age groups. If only for that reason alone, it is important to explore changes in distributions of offending populations within police jurisdictions over time.