‘Focused Deterrence’: Works in practice,but does it work in theory?

Aims of the Paper

In 2004, Freedman, commenting upon debates on deterrence within the fields of Security Studies and criminology, noted ‘… how little they draw upon each others’ work’ (Freedman, 2004, p. 60). Whether consciously or not, some have responded to Freedman’s observation and sought to meld findings from the study of deterrence across both disciplines (see e.g. Stein and Levi, 2014; Rascoff, 2014; Rid, 2012). In a statement of particular relevance to this paper, Braithwaite (2005) noted that the concept of deterrence in the realm of security is ‘…more dynamic compared to the static models that dominate criminology (p. 98)’.

This paper is rooted in two categories of deterrence study from across the disciplines of Security Studies and criminology:

1. The deterrence of terrorism, increasingly studied following the 9/11 attacks. A rich terminology has been developed to describe the theory base of terrorism deterrence.

The paper seeks to take the terminology that has developed in Security Studies around deterrence in general and around measures intended to control terrorism in particular and to put that lexicon to use in developing a descriptive framework suitable for application to Focused Deterrence. It is hoped that the descriptive framework advanced will help interested parties hypothesise and test potential causal mechanisms thereby improving the use of Focused Deterrence. It is not suggested that this paper will deliver the ultimate framework for discussing, devising, assessing and evaluating focused approaches to reducing crime. It is a conceptual piece attempting to establish a mindset. This paper is intended to light some lamps on a path that others might take, further illuminate, widen and improve.

The paper has four parts. The first briefly reviews the literature relating to Focused Deterrence and the deterrence of terrorism. The second sets out the language employed in approaches to reducing terrorism and considers its potential for use in relation to Focused Deterrence. The third uses this language to describe activities within the Focused Deterrence model. The fourth consists of a discussion with potential implications for practice.

Focused Deterrence

Along with debates about the effects of certainty, celerity and severity of punishments, the concepts of general and specific deterrence have engaged criminologists for many years.

A more recent addition to deterrence studies in criminology is Focused Deterrence. Focused Deterrence was first deployed in Boston, USA, in the mid-1990s and was operationalised in that city under the name of ‘Operation Ceasefire’. It was a response to the horrifically high levels of firearm-enabled injuries and deaths amongst young men. Focused Deterrence is a problem-oriented approach (Goldstein, 1979, 1990; Kennedy, 2019). As with other problem-oriented approaches, Focused Deterrence strategies are crafted following identification and detailed analysis of the presenting problem. Analyses in Boston prior to the first use of Focused Deterrence revealed, amongst other things, that the problem was highly concentrated both geographically and demographically. For young male members of the gangs identified in Boston at the time, there was a one in seven chance of dying from gun violence over a 9 year period (Kennedy 2011).

The Focused Deterrence approach was also referred to as a ‘pulling levers’ strategy. This term was used to describe a strategy where all ‘levers’ available to the police and partner agencies were ‘pulled’ in attempts to change offending behaviour. Police, partner agencies and community members told gang members that support in the form of offers of education, training and employment would be given to those who wanted to change their life-styles. However, gang members were also told that every ‘lever’ would be ‘pulled’ to target and disrupt those gangs that continued to use firearms. Moreover, rather than targeting only those doing the shooting, the attack would be on all members of the group. To do this, there would be punishment for all offences committed and al rules broken, including those normally ignored as minor transgressions such as public consumption of alcohol (Kennedy, 1997, 2009; Kennedy et al., 1996). Gang members were particularly vulnerable to the Focused Deterrence approach due to their typical ‘cafeteria’ style of offending, a term which recognises the versatility of offending displayed by gang members who often exploit the menu of offending opportunities presented to them (Klein, 1995, p. 68). There was no licence to commit other types of offences for offenders who eschewed gun violence. Such offending received a traditional enforcement approach. However, for those linked to gun violence the threatened special response was consistently applied.

The Focused Deterrence strategy appeared to be stunningly successful (Kennedy, 1997; Kennedy et al., 1996, 1997, 2001). In the period immediately prior to the use of Focused Deterrence, Boston suffered approximately 100 homicides per year. Operation Ceasefire was implemented in 1996. In 1999, 31 homicides were recorded, an impressive reduction (Kennedy, 2011, p.75). The Boston strategy has been replicated in numerous locations. By 2018, the approach had been implemented in 84 U.S. cities and other countries including Scotland, Turkey and Brazil. (Braga et al. 2018). Though best known for its use against firearm-enabled offending involving groups, the Focused Deterrence approach has also been used in relation to other crime types such as drug-dealing (Saunders et al., 2016) and domestic violence (Sechrist and Weil, 2018).

Braga and Weisburd performed a systematic review of the effects of Focused Deterrence strategies upon crime and reported ‘The basic findings of our review are very positive’ (Braga and Weisburd, 2012, p. 347). However, in common with other commentators on the subject of Focused Deterrence (e.g. Corsaro et al., 2013) the authors highlighted the importance of identifying potential mechanisms at play that could contribute to observed effects (pp. 349–351). Further calls for studies of theoretical mechanisms underpinning Focused Deterrence were made by the same authors in a later study (Braga and Weisburd, 2014). A further systematic review by Braga et al. in 2018, whilst finding that Focused Deterrence appeared to be effective also noted the lack of understanding of the operant mechanisms in successful initiatives The authors stated, ‘Nearly all of the Focused Deterrence program evaluations included in this review could be described as “black box” evaluations where it is uncertain which program elements were most important in generating observed crime reduction effects’ (Braga et al., 2018, pp. 35–36). The need to address the ‘black box’ of Focused Deterrence has been echoed by other authors. (e.g. Schnobrich-Davis and Gardner, 2018; Sullivan and Lugo, 2018; Trinkner, 2019). In short, evaluators were of the opinion that Focused Deterrence is effective, but nobody knows why due to a failure to identify causal mechanisms. The authors of this paper were reminded of a saying attributed to various sources: ‘It works in practice, but does it work in theory?’

Deterrence of terrorism

During the Cold War, nuclear deterrence was a binary situation. Deterrence either worked or it did not and if it had failed the failure would have been all-too obvious. In considering potential responses to terrorism, a major question faced by policymakers and researchers alike lies in whether terrorist acts can be deterred, given the fanaticism of some of the actors. Most experts in the field believe that they can, though acceptable results are more likely to represent marginal gains. Imperfect deterrence is better than no deterrence at all (see Lieberman, 2020; Trager and Zagorcheva, 2005; Wenger and Wilner, 2012; for useful overviews). In this respect, reducing the threat posed by terrorism is more akin to the reduction of crime where a world with less crime is preferable to a world with more crime, but the inevitability of some level of crime is accepted (Rid, 2012; Vinson, 2015).

The nature of measures required to reduce terrorist incidents will differ from those used to manage Cold War-era threats. Recognising that fanatical actors are likely to be difficult to deter directly, some have highlighted the importance of disaggregating the components of terrorist groups and their supporters to identify players who may be more ‘deterrable’ (Davis and Jenkins, 2002). Abrahms makes the point: ‘Terrorism may be deterrable even though most of its practitioners are not’ (2014, p.5). By way of example, finding ways to deter outwardly respectable figures from providing covert financial or other logistical assistance to terrorist groups could lead to a reduction in incidents through starving fanatics of support.

The Concept of Preventive Influence

During the post-Cold War period, the potential for broadening the concept of deterrence in Security Studies has been subject of debate, much of which has been informed by the emergence of the new threats previously described. The debate is neatly summarised by Mazarr et al. (2018). The narrowest concept holds that military means are the way to achieve deterrence. A broader concept accepts that means beyond the military may be necessary to deter an adversary. Both the narrow and broader concepts are based in threats that aggression will result in punishment and/or denial of the benefits sought. A third and even broader approach to deterrence allows for combining threats with positive inducements to forestall aggression. In relation to this much-broadened approach, Mazarr et al. acknowledge that ‘… the threat-plus-inducement understanding of deterrence can obscure the boundaries of the concept’ (Mazarr et al. 2018, p. 4).

Arguments in favour of mixing inducements and threats to affect an adversary’s cost-benefit calculation are attractive; whether such an approach should be viewed as deterrence is not. Writing some years before Mazarr et al. (2018), others raised concerns about the conceptual boundaries of deterrence. Davis and Jenkins in seeking to develop potential responses to terrorism found the concept of deterrence to be ‘…too narrow to use as an organising principle…’ (2002, p. 9). Recognising the potential value of blending inducements with threats they promoted the use of influence measures to reduce the threat from terrorism. In this, they shied away from stretching the concept of deterrence beyond the commonly accepted understanding that it involves the threat of negative consequences. In similar vein, Jeffrey Knopf writing later held that deterrence ‘…arises mainly because an actor expects a course of action to lead to a negative outcome’ (Knopf, 2009, pp. 38–39). In asserting this, Knopf drew upon the everyday use of language, believing that nearly every lay person could define deterrence and that those definitions would be likely to revolve around threats and not positive inducements (p. 34). In common with Davis and Jenkins (2002), Knopf chose to view deterrence as ‘…a form of influence…’ (2009, p. 34) within an overarching concept of ‘preventive influence’. Preventive influence ‘… might also be pursued through the promise of rewards or offer of reassurances’, but for Knopf, such measures represented positive inducements as opposed to deterrent measures (p. 38). Davis, Jenkins and Knopf all took the view that deterrence is but one part of a range of possible measures to exert influence, a range which includes dialogue and inducements as well as threats. But what is influence? Mazarr et al. (2021) found the term to be ‘…ill-defined and often employed in vague and imprecise ways’, where ‘Many definitions very explicitly note that influence involves achieving effects without force and often indirectly’. Other definitions are silent on whether force may be used to exert influence (p. 12). Though widely used in everyday language, researchers have tussled with the seemingly simple term (see e.g. Santa Maria, 2013).

The authors of this paper offer a way out of the definitional thicket surrounding the term ‘influence’ by asserting that force may be used to exert influence, without contending that all use of force represents an attempt to exert influence. It is the reasoning behind the use of force and the intended effects that dictate whether force is being deployed in pursuit of influence. Schelling (2008) when discussing measures capable of impacting upon the behaviour of adversaries differentiated between ‘brute force’ and measures which act upon the mind of the adversary. Brute force physically removes freedom of action. Measures which act upon the mind allow the subject to choose the way in which they respond in the face of such measures. Schelling paid particular attention to two approaches capable of acting upon an adversary’s decision making: ‘deterrence’ and something for which he invented a new term: ‘compellence’. In Schelling’s view, deterrence is the use of threats to convince an adversary not to do something; compellence is the use of threats and/or limited force to cause an adversary to stop doing something. The use of force within the concept of compellence demonstrates a capability to inflict further pain and aims to convince an adversary to change behaviour. This is influence. Overwhelming an adversary and removing all freedom of choice is not. It is the use of brute force. It should be noted, however, that totally overwhelming an adversary by way of brute force might influence the future behaviour of onlookers, both intentionally (‘Don’t mess with us!’) and unintentionally (‘Don’t mess with them; look what they can do!’). Whether influence results from intended or unintended outcomes, the key point is that the mechanism of effect is psychological and not physical. Sawyer (2021) is supportive of this stance in stating that when influence is the intended outcome ‘…the adversary retains a degree of volition rather than being forced into compliance or being eliminated altogether’ (p. 105). The use of force ‘…to get inside the adversary’s decision calculus and manipulate its perception of the costs and benefits’ is an attempt to influence if the intention is to convince an adversary to change behaviour (Sawyer, 2021, p.104).

A further example of the use of limited force with a view to influencing the decision making of adversaries is the concept of Cumulative Deterrence. Cumulative Deterrence may be defined as ‘…the simultaneous use of threats and … force over the course of an extended conflict’ (Almog, 2004, p.8), where the intention is to produce a deterrent effect in the future due to responding consistently as previously threatened. Consistent responses with attendant messaging that force has been used as a response to provocation and will only occur if certain lines are crossed may accumulate to produce cumulative deterrence over time. The distinction between compellence and cumulative deterrence is fine but still exists. Limited force that influences an adversary to stop doing something at the time is compellence. Cumulative deterrence is the result of the repeated use of limited force over time and building a reputation for consistently following through on threats. Both compellence and cumulative deterrence use force not to remove freedom of action but to target decision making in an adversary. Under the argument advanced to this point, they are influence measures.

The terminology of Preventive Influence

Effective use of pulling levers strategies to reduce crime and identify causal mechanisms for any effects observed requires access to a richer terminology than is currently used in criminology. Aims, methods and potential mechanisms of change to be activated in relation to any intervention must be accurately described. Tests of efficacy for hypothesised mechanisms must be devised if mechanistic explanation of pulling levers approaches is to be achieved within an assessment and evaluation regime. ² It was stated earlier in this paper that terminology developed within the discipline of Security Studies in relation to reducing terrorism could be exploited with a view to achieving these ends.

Figures 1 and 2 draw upon relevant Security Studies literature and represent efforts to summarise terminology relevant to the concept of preventive influence. Figure 1 consists of a lexicon of methods for exerting influence. Figure 2 is a lexicon of deterrence. Conceptually, Figure 2 is an expansion of the deterrence section of Figure 1. Whilst deterrence is shown as only one tool within the influence toolkit (Figure 1), it has many variants which are briefly ‘unpacked’ in Figure 2. It is not contended that the terminology presented by way of the two lexicons is exhaustive. It is contended that the two interlinked lexicons represent a step towards developing an analytical framework to provide support to those seeking to identify, activate, assess and evaluate causal mechanisms to improve the use of preventive influence, in the guise of Pulling Levers, in the field of crime reduction.OPEN IN VIEWER

OPEN IN VIEWER

Figure 2.

A lexicon of deterrence.

Figure 1 has been devised drawing upon the work of several authors (Davis, 2012a, 2012b; Davis and Jenkins 2002; Davis et al. 2021; Knopf, 2009; Schelling, 2008). The measures within Figure 1 should not be viewed as operating in isolation from each other. The intention of a well-formulated strategy of influence would be to blend measures as required to achieve the desired result.

As can be seen from Figure 1, deterrence is one way to exert preventive influence. Deterrence, however, comes in many forms. Figure 2 consists of a lexicon drawn from the deterrence literature in Security Studies, with an emphasis on the literature surrounding approaches to deterring terrorism. Some of the terms within the lexicon are listed in ‘family’ groups (A to I) to highlight where similar deterrent mechanisms might be intended to be activated even though different terms may be used to describe the deterrence type. Illustrative references are provided within the table. These references by no means represent the sum total of the scholarship surrounding the items listed. They merely seek to raise awareness in the reader of the types of deterrence subject of discussion in the Security Studies realm.

Describing the ‘Pulling Levers’ approach using the terminology advanced

The core ‘Pulling Levers’ model is summarised below (Kennedy, 1997, 2006, 2009, 2019; Kennedy et al., 1996). Though having been used against other crime problems as mentioned above (e.g. open-air drug markets and domestic violence), the ‘Pulling Levers’ approach in criminology is best known as a response to gun crime. In recognition of this, commentary has been added to stages of the model to illustrate how the terminology advanced earlier can be used to accurately describe measures intended to reduce firearm-enabled violence. Where a deterrence type can be described by more than one term, only the first term in each ‘family’ group within Figure 2 is used.

The core ‘Pulling Levers’ model as used against gun violence

(a) Develop understanding of the problem in the local context by way of a multi-agency partnership.

During this stage of the model, contextual intelligence is gathered to develop Tailored Deterrence approaches based upon identification of contextual ‘centres of gravity’ where the application of pressure is likely to generate behavioural change. For example, if violence is linked to the local criminal economy, that economy may be viewed as a point of focus for deployment of influence measures to change behaviour.

(b) Use communities, victims and those close to potential offenders to describe the harms caused by the problem behaviour and to communicate that things will change.

Prior warning of introduction of the new regime represents attempts at Dissuasion of potential offenders. Engagement with key groups within affected communities seeks to undermine any belief amongst offenders that their behaviour is acceptable to the community. This seeks to establish Deterrence via Delegitimisation.

(c) Make offers of support in relation to employment, training, housing etc., to those involved in the proscribed activity should they wish to change their ways.

Offers to provide support to those directly involved in proscribed behaviour may be viewed as Inducements (providing rewards for behavioural change) and/or Co-Option (bringing an offender into the law-abiding community). The existence of schemes and arguments for taking up offers of support may convince those on the periphery to avoid becoming involved in offending, thereby representing a successful use of Persuasion. The existence of such offers supports the concepts of Deterrence via Delegitimisation.The ground is taken from below the feet of the offender who justifies their behaviour by stating ‘I do this because there is nothing else for me’.

(d) Directly and repeatedly communicate with the identified players that a failure in them to change their behaviour will attract a premium multi-agency enforcement response. Those giving up the prohibited behaviour will attract a standard enforcement response for any other types of offending behaviour.

Such communications seek to influence offender behaviour by way of Deterrence by Punishment. In addition, through differentiating responses between types of offending in a harm-reduction approach, such approaches seek to achieve Narrow Deterrence. Some types of offending might still occur and be addressed by traditional responses, but proscribed behaviours will not be tolerated and will receive a premium enforcement response. The long-term goal is the establishment of Deterrence via Norms, whereby the proscribed behaviour is not even considered as an option by offenders.

(e) Punish any instances where the prohibited behaviour is exhibited through taking action for any offences committed or rules broken, exploiting the versatility of offending displayed by those involved.

This element of the model relies upon Cross-Domain Deterrence. It recognises that whilst it may not be possible to prosecute offenders for committing the offences within the proscribed offending ‘domain’ (i.e. gun crimes), behaviour in other offending ‘domains’ that are easier to detect makes offenders vulnerable to influence via law enforcement. Action may be taken for drug offences, parole violations, motoring offences etc. If violence is being pursued in relation to drug supply, disruption of that market could be undertaken to deny the violent the financial benefits they seek: Deterrence by Denial.

Swift and consistent application of the threatened punishments for every transgression represents Compellence, which could be expected to contribute to the generation of Cumulative Deterrence over time.

(f) Target punishment action on those posing the greatest risk and offenders associated with them, with a view to coercing offenders into controlling the behaviour of their more problematic associates.

Threatening to punish key players is classic Deterrence by Punishment. Extending that threat to those associated with the most dangerous offenders invokes Indirect Deterrence: applying deterrent pressures to those who might not pose a direct threat but who can be coerced into applying deterrent pressure to those who do pose a threat.