You are on page 1of 396

Benoit 

Leclerc · Ernesto U. Savona
Editors

Crime
Prevention
in the 21st
Century
Insightful Approaches for Crime
Prevention Initiatives
Crime Prevention in the 21st Century
Benoit Leclerc • Ernesto U. Savona
Editors

Crime Prevention
in the 21st Century
Insightful Approaches for Crime Prevention
Initiatives
Editors
Benoit Leclerc Ernesto U. Savona
School of Criminology and Criminal Justice Università Cattolica del Sacro
Griffith University Cuore-Transcrime
Brisbane, QLD, Australia Milano, Italy

ISBN 978-3-319-27791-2 ISBN 978-3-319-27793-6 (eBook)


DOI 10.1007/978-3-319-27793-6

Library of Congress Control Number: 2016943526

© Springer International Publishing Switzerland 2017


This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature


The registered company is Springer International Publishing AG Switzerland
Acknowledgements

We would like to thank the contributors to this volume for assisting in the peer
review process of the manuscripts included in this volume. We would also like to
thank Peter Bell, Rick Brown, Jesse Cale, Ron Clarke, Marcus Felson, Mike
Maxfield, Carlo Morselli, Janet Ransley, Danielle Reynald, and Clifford Shearing
for their assistance in the peer review process.

v
Contents

1 Crime Prevention in the Twenty-First Century ................................... 1


Benoit Leclerc and Ernesto U. Savona

Part I The Offenders’ Perspective on Crime Prevention


2 Learning from the Offenders’ Perspective on Crime Prevention ....... 9
Scott Jacques and Elizabeth Bonomo
3 Learning About Crime Prevention from Aborted Crimes:
Intrapersonal Comparisons of Committed
and Aborted Robbery ............................................................................. 19
Marie Rosenkrantz Lindegaard and Wim Bernasco
4 Learning from Offenders: Some Iatrogenic Effect
of Crime Prevention Measures .............................................................. 35
Martin Gill

Part II Crime Script Analysis


5 Script Analysis and the SCRIPT Acronym .......................................... 49
Benoit Leclerc
6 An Examination of Sexual Offenses Against Women
by Acquaintances: The Utility of a Script Framework
for Prevention Purposes.......................................................................... 59
Yi-Ning Chiu and Benoit Leclerc
7 Expanding the Script Analytic Approach Using Victim
Narratives: Learning About Robberies of Taxi Drivers
from the Drivers Themselves ................................................................. 77
Martha J. Smith

vii
viii Contents

Part III Crime Mapping and Spatial Analysis


8 Crime Mapping and Spatial Analysis ................................................... 101
Michael Townsley
9 Mapping Crime Prevention: What We Do and Where
We Need to Go ......................................................................................... 113
Martin A. Andresen

Part IV Social Network Analysis


10 Social Network Analysis and Crime Prevention .................................. 129
Giulia Berlusconi
11 Using Social Network Analysis to Design Crime Prevention
Strategies: A Case Study of Methamphetamine Manufacture
and Trafficking ........................................................................................ 143
David A. Bright
12 Uncovering the Social Pecking Order in Gang Violence ..................... 165
Jasmin Randle and Gisela Bichler
13 Static and Dynamic Approaches of a Drug Trafficking Network ...... 187
Andrea Giménez-Salinas Framis and Sonia Fernández Regadera

Part V Agent-Based Modelling


14 Formal Models of the Crime Event: Agent-Based Modelling
in Support of Crime Prevention ............................................................ 215
Daniel Birks
15 Investigating Prevention by Simulation Methods ................................ 235
Charlotte Gerritsen and Henk Elffers

Part VI Crime-Proofing Legislations


16 Proofing Legislation Against Crime as Situational
Prevention Measure ................................................................................ 247
Ernesto U. Savona
17 Crime Proofing of Legislation in Practice: The Case
of the 2014 EU Tobacco Products Directive.......................................... 275
Stefano Caneppele
18 A New Victorian Era: Getting Crime Out of Commercial Sex .......... 293
Adam Masters and Adam Graycar

Part VII Technologies of Crime Prevention


19 Technology, Opportunity, Crime and Crime Prevention:
Current and Evolutionary Perspectives ................................................ 319
Paul Ekblom
Contents ix

20 Passive WiFi Radar: A New Technology for Urban


Area Surveillance .................................................................................... 345
Kevin Chetty and Karl Woodbridge
21 Micro UAV Crime Prevention: Can We Help Princess Leia? ............. 359
Matthew Ritchie, Francesco Fioranelli, and Hervé Borrion
22 Technology for Crime and Crime Prevention:
A Supply Side Analysis ........................................................................... 377
Graham Farrell and Nick Tilley

Index ................................................................................................................. 389


Contributors

Martin A. Andresen School of Criminology, Institute for Canadian Urban


Research Studies, Simon Fraser University, Burnaby, BC, Canada
Giulia Berlusconi Università Cattolica del Sacro Cuore—Transcrime, Milan, Italy
Wim Bernasco Department of Spatial Economics, Netherlands Institute for the
Study of Crime and Law Enforcement (NSCR), Vrije Universiteit Amsterdam,
Amsterdam, Netherlands
Gisela Bichler Center for Criminal Justice Research, California State University,
San Bernardino, CA, USA
Daniel Birks Griffith Criminology Institute, Griffith University, Brisbane, QLD,
Australia
Elizabeth Bonomo Department of Criminal Justice and Criminology, Georgia
State University, Atlanta, GA, USA
Hervé Borrion UCL Department of Security and Crime Science, London, UK
David A. Bright School of Social Sciences, University of New South Wales,
Sydney, NSW, Australia
Stefano Caneppele School of Criminal Sciences, University of Lausanne,
Lausanne, Switzerland
Kevin Chetty Department of Security and Crime Science, University College
London, London, UK
Yi-Ning Chiu School of Criminology and Criminal Justice, Griffith University,
Brisbane, QLD, Australia
Paul Ekblom Design Against Crime Research Centre, Central Saint Martins,
University of the Arts London, London, UK
Department of Security and Crime Science, UCL, London, UK
Applied Criminology Centre, University of Huddersfield, Huddersfield, UK

xi
xii Contributors

Henk Elffers NSCR & VU University Amsterdam, Amsterdam, The Netherlands


Graham Farrell Center for Criminal Justice Studies, School of Law, University of
Leeds, Leeds, UK
Francesco Fioranelli UCL Department of Electronic and Electrical Engineering,
London, UK
Andrea Giménez-Salinas Framis Instituto de Ciencias Forenses y de la Seguridad,
Universidad Autonoma de Madrid, Madrid, Spain
Charlotte Gerritsen NSCR & VU University Amsterdam, Amsterdam, The
Netherlands
Martin Gill Perpetuity Research & Consultancy International (PRCI) Ltd, Kent, UK
Adam Graycar School of Social & Policy Studies, Flinders University, Adelaide,
South Australia, Australia
Scott Jacques Department of Criminal Justice and Criminology, Georgia State
University, Atlanta, GA, USA
Benoit Leclerc School of Criminology and Criminal Justice, Griffith University,
Brisbane, QLD, Australia
Marie Rosenkrantz Lindegaard Netherlands Institute for the Study of Crime and
Law Enforcement (NSCR), Amsterdam, Netherlands
Adam Masters Transnational Research Institute on Corruption, The Australian
National University, Canberra, ACT, Australia
Jasmin Randle Center for Criminal Justice Research, California State University,
San Bernardino, CA, USA
Sonia Fernández Regadera Instituto de Ciencias Forenses y de la Seguridad,
Universidad Autonoma de Madrid, Madrid, Spain
Matthew Ritchie UCL Department of Electronic and Electrical Engineering,
London, UK
Ernesto U. Savona Università Cattolica del Sacro Cuore-Transcrime, Milano,
Italy
Martha J. Smith School of Community Affairs, Wichita State University, Wichita,
KS, USA
Nick Tilley Department of Security and Crime Science, University College
London, London, UK
Michael Townsley School of Criminology and Criminal Justice and Griffith
Criminology Institute, Griffith University, Nathan, QLD, Australia
Karl Woodbridge Department of Electronic and Electrical Engineering, University
College London, London, UK
About the Editors

Benoit Leclerc is Associate Professor of Criminology and Criminal Justice at


Griffith University, Brisbane, Australia. Before joining Griffith, he was involved in
research and clinical work with adolescent and adult sex offenders at the Philippe-
Pinel Institute of Montréal for 7 years and received his Ph.D. from the Université of
Montréal, Canada. His research interests include script analysis, environmental
criminology and situational crime prevention, and sexual offending. He is the lead
investigator of a project funded by the Australian Research Council on the effective-
ness of situational prevention to prevent sexual offenses. Key publications include
articles in Criminology, Criminal Justice & Behavior, Child Abuse & Neglect,
British Journal of Criminology, Journal of Research in Crime and Delinquency and
Sexual Abuse: A Journal of Research and Treatment as well as books entitled
Cognition and Crime: Offender Decision-Making and Script Analyses (co-edited
with Richard Wortley – Routledge) and Pathways of Sexual Aggression (co-edited
with Jean Proulx, Eric Beauregard and Patrick Lussier – Routledge).

Ernesto U. Savona is Director of TRANSCRIME, Joint Research Centre of the


Università Cattolica del Sacro Cuore, Milan. He is Editor-in-Chief of the European
Journal on Criminal Policy and Research (Springer). Since 2003, he has been pro-
fessor of Criminology at the Università Cattolica del Sacro Cuore in Milan. From
1986 to 2002, he was professor of Criminology at the Faculty of Law at the
Università degli Studi di Trento. From 1971 to 1986, he was associate professor at
the Faculty of Statistical Sciences at the Università “La Sapienza” in Rome. He is
Past President of the European Society of Criminology (2003/2004) and Chair for
the term 2011–2012 of the Global Agenda Council on Organized Crime of the
World Economic Forum. Professor Savona is a member of the European Commission
experts group on policy needs for data on crime and criminal justice and of the
expert group on Firearms. His research interests and key publications include books
and articles on organized crime, money laundering, and corruption. He has devel-
oped for the 6th EU Research Framework Programme the crime-proofing approach
aimed at preventing opportunities for crime inadvertently created by regulation.

xiii
Chapter 1
Crime Prevention in the Twenty-First Century

Benoit Leclerc and Ernesto U. Savona

Even though statistics tend to show a decline of crime generally in the last decade
(Tonry, 2013), there are a number of crimes that represent and/or are perceived as a
serious and constant threat to people’s safety and the society. Terrorism is one
example. The emergence of the Islamic State in the summer of 2014 was alerted by
the media rather than by intelligence analysts. Like the 9/11 terrorist attack, this
seems to be an incident from which a lesson can be learned on the difficulty of pre-
dicting and preventing terrorism at a global level. The development of the Islamic
State was a gradual process that required time in recruitment, funding and arming.
It appears that there is no tangible explanation to date why intelligence agencies of
many governments, private businesses involved in the oil industry, organisations
selling weapons and others did not inform authorities of their reference points of the
signal that a new terrorism was emerging in the north of Iraq. Assuming that this
information was in circulation, difficulties in managing, interpreting and translating
data coming from different sources may have slowed down the process leading to an
alert from authorities. On the face of it, we have to admit that the prevention of ter-
rorism has failed to date.
Terrorism, like other forms of crime, raises doubts on the efficacy of the tradi-
tional instruments of criminal justice system (i.e. investigation-prosecution-
conviction) to prevent crime. It can be argued that the criminal justice system is not
well equipped to combat a crime in which offenders are ready to sacrifice their life
to kill other people. In the scale of complex or emerging criminal phenomenon, ter-
rorism is arguably the most difficult to prevent because the costs and benefits of

B. Leclerc (*)
School of Criminology and Criminal Justice, Griffith University,
176 Messines Ridge Road, Mt Gravatt Campus, Brisbane, QLD 4122, Australia
e-mail: b.leclerc@griffith.edu.au
E.U. Savona
Università Cattolica del Sacro Cuore-Transcrime, Largo Gemelli 1, 20123 Milano, Italy
e-mail: ernesto.savona@unicatt.it

© Springer International Publishing Switzerland 2017 1


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_1
2 B. Leclerc and E.U. Savona

committing this crime are distorted by ideology. Therefore, it is clear that


evidence-based knowledge on terrorism patterns to reduce the opportunities for its
development is needed. Years ago, Clarke and Newman (2006) opened an interest-
ing path to preventing terrorism via a situational prevention framework. But there is
still very little formally and rigorously evaluated evidence (Lum, Kennedy, &
Sherley, 2006–2009) available, due partly to the neglect of evaluation by the author-
ities pouring vast sums into security and partly to the (fortunate) rarity of attacks. A
recent ongoing project based at the University of Huddersfield (see www.hud.ac.uk/
research/researchcentres/acc/projects/reducing-serious-crime-and-terrorism-at-
mmpts/) seeks to advise practitioners on countering terrorism at multi-modal pas-
senger terminals whilst coping with the absence of direct impact evidence. By
elaborating detailed attack scripts, as illustrated by Clarke and Newman (2006), and
drawing on tested theory and practice from general situational crime prevention to
suggest theoretically and practically plausible interventions to reduce opportunities,
the risk of a terrorist attack may be reduced.
Taking the example of terrorism further, society cannot rely on apprehending and
incarcerating offenders to address most crime problems. Human rights, economic
and social costs associated with incarceration and victimisation are becoming expo-
nential (e.g., Makkai & Prenzler, 2011). In this context, how to inform crime pre-
vention initiatives emerges as a priority. Over the years, a number of innovative
approaches—some more technical than others—to learn about crime patterns for
guiding and supporting crime prevention initiatives have been developed. Some of
them have been relatively popular; however other approaches have not benefited
from the same interest. These approaches have one strength in common–their poten-
tial for crime prevention purposes is immense. In addition, they are complementary
and versatile enough to be applied in different contexts for different crimes. One of
these approaches taps into a research area referred as offender-based research
(Bernasco, 2010; Jacques & Bonobo, 2016). We believe that offender-based research
is a valuable approach because offenders are used as the main source of data to
understand and inform prevention practitioners on offending patterns that could
otherwise be neglected. A second approach is crime script analysis, which the main
purpose is to break down crime-commission processes to identify a wide range of
intervention points for prevention purposes (Cornish, 1994; Leclerc & Wortley,
2013). A third approach is crime mapping and spatial analysis, which years ago
became famous for orienting policing efforts. It is believed that this approach has
more to offer than identifying crime ‘hotpots’ (Anselin, Griffiths, & Tita, 2008;
Townsley 2016). A fourth approach is social network analysis, where the main pur-
pose is to tap into criminal networks. Again the versatility of this approach indicates
that it has more to offer to crime prevention (Morselli, 2014). A fifth approach is
agent-based modelling (Birks, 2016), which should become in the future one of the
main tools through which the potential of prevention initiatives could be assessed
before implementation. The sixth approach emerges through the lens of law—
namely crime-proofing legislations, which can be used as a technique to support
prevention initiatives applied on the field (Savona, 2006). In this era, an edited col-
lection on innovative approaches for crime prevention would be incomplete without
1 Crime Prevention in the Twenty-First Century 3

a section that covers the technologies of crime prevention. A seventh and final
approach is the design and use of technologies to protect potential targets from
crime (e.g., Farrell, Tseloni, Mailley, & Tilley, 2011).
This edited collection brings together a series of original contributions written by
who we believe are innovative thinkers dedicated to crime prevention. This collec-
tion is divided according to the list of approaches above. Each section focuses on
one particular approach, which includes one theoretical chapter to introduce the
approach of interest but also empirical/applied contribution(s) to showcase how it
can be applied. Specifically, theoretical contributions aim to introduce and explain
the approach and discuss under which circumstances it could better assist crime
prevention initiatives. Empirical/applied contributions are used to illustrate the
potential of the approach for crime prevention through its application to a particular
crime problem. Obviously not all possible crime prevention approaches are covered
in this collection. Efforts have been made to include approaches that we thought
would be most relevant for the future of crime prevention at the time this collection
was initiated.
Chapters 2–4 focus on offender-based research. Chapter 2, by Jacques and
Bonobo, introduces offender-based research by illustrating five ways through which
this approach can be used to inform crime prevention. Jacques and Bonobo make a
strong case of the relevance of offender-based research for crime prevention. They
conclude by discussing the choices involved in conducting offender-based research
for the purposes of situational crime prevention. In Chapter 3, Lindegaard and
Bernasco illustrate the potential of offender-based research for crime prevention by
examining why prospective offenders sometimes do not perpetrate their crimes they
anticipated to commit. Using a sample of robbers, they seek to distinguish aborted
robberies from those that are committed as well as the mechanisms and the ratio-
nales underlying aborted robberies. Subjective reasons for aborting anticipated rob-
beries include expected police and bystander interventions. Gill, in Chapter 4,
borrows the term ‘iatrogenic effects’ from medical science to showcase the impor-
tance of learning through offenders why a number of prevention measures do not
work—the ‘iatrogenic effects’ of crime prevention. Gill argues that ‘iatrogenic
effects’ have much to offer to get an understanding of crime prevention failures.
The emphasis of Chapters 5–7 is on crime script analysis. In Chapter 5, Leclerc
presents an overview of crime script analysis and then promotes its practice for
understanding and preventing crime events. Leclerc uses the SCRIPT acronym
(systematic, compatible, resilient, insightful, practical and theoretically grounded)
to showcase the qualities inherent to script analysis for crime prevention purposes.
Chiu and Leclerc, in Chapter 6, seek to better understand the crime-commission
processes involved in sexual offenses against women by acquaintances and identify
key intervention points by using crime scripts. A script comprising eight stages is
identified through court transcripts, and prevention measures are derived by using
the problem analysis triangle. In Chapter 7, Smith uses crime scripts from the per-
spective of taxi drivers as victims of robbery. Through data obtained from taxi driv-
ers, three main types of scripts are identified. Then Smith discusses potential
situational prevention measures to prevent robberies, or limit their potential negative
4 B. Leclerc and E.U. Savona

impact on taxi drivers, by drawing on the 25 situational prevention techniques clas-


sification designed by Cornish and Clarke (2003).
Chapters 8 and 9 focus on crime mapping and spatial analysis. In Chapter 8,
Townsley completes a review of crime mapping and spatial analysis and outlines the
modern ‘criminology of place’ inspired by Sherman and its utility for prevention.
Throughout his review, Townsley discusses the origins of this approach, its strengths
and weaknesses as well as its future for crime prevention. Andresen, in Chapter 9,
also retraces, to some extent, the origins of crime mapping. However, Andresen
takes a different angle and examines crime mapping through its main methodology,
that is, kernel density mapping. Andresen also discusses local spatial analysis, argu-
ing that this method can be easily applied to crime problems and be highly benefi-
cial for crime prevention.
In Chapters 10–13, the spotlight is turned on social network analysis. In Chapter 10,
Berlusconi introduces the approach of social network analysis and focuses on its
value for crime prevention. She argues that this approach can improve our under-
standing of social processes underlying crime problems but also assist law enforce-
ment agencies in enforcing crime. As an example, this approach can help assess
different disruption strategies on criminal groups. Chapter 11, by Bright, presents an
application of social network analysis to the development of crime prevention strat-
egies for drug trafficking. Specifically, a network of methamphetamine manufacture
and trafficking is analysed through a combination of social network analysis and
crime scripts. Bright discusses prevention initiatives drawing from both the crime-
commission process and the criminal network that supports and carries out the
activities in the crime-commission process. In Chapter 12, Randle and Bichler apply
social network analysis to a network of 158 street gangs operating in Los Angeles,
which offers insights into violence reduction strategies. Randle and Bichler present
five ways through which this approach increases opportunities for crime prevention,
which include identifying where to aim efforts to quell internal rivalries. Chapter 13
presents an application of social network analysis by Giménez-Salinas Framis and
Fernández Regadera. These authors examine a cocaine trafficking network in Spain
in both a static and dynamic context. Giménez-Salinas Framis and Fernández
Regadera’s analysis highlights the advantages of using this approach as a tool to
predict the most relevant players in a criminal network at the start of an investiga-
tion, as well as to design ways to neutralize and dismantle criminal networks.
Chapters 14 and 15 emphasize agent-based modelling. In Chapter 14, Birks
starts with examining how computational agent-based models can assist crime pre-
vention initiatives through testing and refining crime event theories—a critical
dimension to address before the implementation and evaluation of situational crime
prevention techniques. Chapter 15 sees Gerritsen and Elffers discussing the use of
agent-based and laboratory simulation methods for investigating preventive mea-
sures against crime. Gerritsen and Ellfers then distinguish anticipatory prevention
from mitigating prevention, making a case of the importance of agent-based studies
for evaluating different prevention measures.
Chapters 16–18 examine the relevance of crime-proofing legislations for crime
prevention. In Chapter 16, Savona introduces crime-proofing legislation as a
1 Crime Prevention in the Twenty-First Century 5

situational prevention measure to reduce the opportunities for crime inadvertently


created by legislation. Savona retraces the origins of this approach by framing it
inside the European Union approach for crime prevention developed in the middle
of 2000. Savona argues that proofing legislation against crime could be effective to
reduce criminal opportunities sometimes created through the increasing number of
laws and their complexity. Chapter 17, by Caneppele, presents a case study on
crime-proofing legislation on the new EU Tobacco Products Directive to showcase
the relevance of this approach. Caneppele’s examination indicates that a provision,
such as the ban on menthol and slim cigarettes, significantly increases the risks of
an enlargement of the illicit market. The implementation of the menthol ban, how-
ever, was planned only after a phase-out period of 4 years so that it is impossible to
determine whether and how much this study persuaded European regulators.
Masters and Graycar present another case of crime-proofing legislation in
Chapter 18. Their study investigates commercial sex and the effects of decriminal-
ization in the state of Victoria, Australia. Masters and Graycar analyse how success-
ful is the process of decriminalizing commercial sex in preventing its associated
criminality and whether this process will prevent or increase such criminality.
Chapters 19–22 focus on the technologies of crime and the relevance of this
approach for crime prevention. In Chapter 19, Ekblom introduces the approach by
seeking to link it to crime science, including situational crime prevention. Ekblom
further looks at the practicalities of adopting an evolutionary approach to preven-
tion but also the weaknesses of purely technological approaches to crime preven-
tion. He argues that understanding the evolution of technology can block but also
create crime opportunities. Chapter 20, by Chetty and Woodbridge, focuses on the
use of WiFi signals for crime prevention purposes. Chetty and Woodbridge argue
that WiFi signals could form the basis of a surveillance system based around a
technology known as passive radar for detecting and tracking people. The authors
present this technology and how it works, and outline potential applications for
preventing and detecting crime. In Chapter 21, Ritchie, Fioranelli and Borion
examine the technology of micro unmanned aerial vehicles in the context of crime.
They contextualized the implications of this technology by discussing some issues
to be faced by society as a result of its growing use (and potentially misuse), and
point out a number of tasks that law enforcement would need to achieve in order
to prevent crime committed via this technology. Lastly, they demonstrate the pos-
sibility of using a conventional radar system to detect micro unmanned aerial
vehicles flying, which has direct implications for the prevention of drone crimes.
Lastly, Chapter 22 by Farrell and Tilley discusses the impact of technology on
crime. Farrell and Tilley examine technology in terms of its intended and unin-
tended effects on the supply of crime opportunities. In their analysis, they use the
acronym of DAPPER (default, aesthetically neutral, has a powerful preventive
mechanism, is principled and acceptable to all except offenders, effortless to
engage, and rewarding in cost-benefit terms) to present what would be the most
elegant technology for effectively reducing the supply of crime opportunities.
Farrell and Tilley argue that the best security does not produce a fortress society
but rather improves our quality of life.
6 B. Leclerc and E.U. Savona

References

Anselin, L., Griffiths, E., & Tita, G. (2008). Crime mapping and hot spot analysis. In R. Wortley &
L. Mazerolle (Eds.), Environmental criminology and crime analysis. Cullompton, UK: Willan.
Bernasco, W. (Ed.). (2010). Offenders on offending: Learning about crime from criminals.
Cullompton, UK: Willan.
Birks, D. (2016). Formal models of the crime event: Agent-based modelling in support of crime
prevention. In B. Leclerc & E. U. Savona (Eds.), Crime prevention in the 21st century:
Insightful approaches for crime prevention initiatives. New York: Springer.
Clarke, R. V., & Newman, G. R. (2006). Outsmarting the terrorist. Westport, CT: Praeger Security
International.
Cornish, D. B. (1994). The procedural analysis of offending and its relevance for situational pre-
vention. In R. V. Clarke (Ed.), Crime prevention studies (Vol. 3). Monsey, NY: Criminal Justice
Press.
Cornish, D. B., & Clarke, R. V. (2003). Opportunities, precipitators and criminal decisions: A reply
to Wortley’s critique of situational crime prevention. In M. J. Smith & D. B. Cornish (Eds.),
Theory for practice in situational crime prevention (Crime Prevention Studies, Vol. 16).
Monsey, NY: Criminal Justice Press.
Farrell, G., Tseloni, A., Mailley, J., & Tilley, N. (2011). The crime drop and the security hypothesis.
Journal of Research in Crime and Delinquency, 48, 147–175.
Jacques, S., & Bonobo, E. (2016). Learning from the offenders’ perspective on crime prevention.
In B. Leclerc & E. U. Savona (Eds.), Crime prevention in the 21st century: Insightful
approaches for crime prevention initiatives. New York: Springer.
Leclerc, B., & Wortley, R. (Eds.). (2013). Cognition and crime: Offender decision making and
script analyses (Crime Science Series). London, UK: Routledge.
Lum, C., Kennedy, L., & Sherley, A (2006–2009). The effectiveness of counter-terrorism strate-
gies: A Campbell systematic review. Fairfax, VA: The Campbell Collaboration.
Makkai, T., & Prenzler, T. (2011). The nature and prevalence of crime. In H. Hayes & T. Prenzler
(Eds.), An introduction to crime and criminology (3rd ed.). Frenchs Forest, NSW: Pearson
Education.
Morselli, C. (Ed.). (2014). Crime and networks. New York: Routledge.
Savona, E. U. (2006). Double thematic issue on: Proofing EU legislation against crime. European
Journal on Criminal Policy and Research, 12, 177–178.
Tonry, M. (2013). Understanding crime trends in Italy and elsewhere. In S. Caneppele &
F. Calderoni (Eds.), Organized crime, corruption and crime prevention: Essays in honor of
Ernesto U. Savona. New York: Springer.
Townsley, M. (2016). Crime mapping and spatial analysis. In B. Leclerc & E. U. Savona (Eds.),
Crime prevention in the 21st century: Insightful approaches for crime prevention initiatives.
New York: Springer.
Part I
The Offenders’ Perspective on Crime
Prevention
Chapter 2
Learning from the Offenders’ Perspective
on Crime Prevention

Scott Jacques and Elizabeth Bonomo

Introduction

To learn about crime, including how to prevent it, criminologists should go to the
source. As a collective, criminals are the preeminent source of knowledge on crime.1
They have a firsthand perspective on what motivates the decision to break the law,
obey it, or abort an in-progress offense. Moreover, offenders are experts on the tech-
nical process involved in committing crime, such as how to steal a vehicle or rob a
pedestrian. And while police and victimization statistics, among other kinds, are
indicative of the crime rate, only the collective knowledge of offenders can truly tell
us how much crime is committed.
In some sense, offenders are not only the “problem” under study but also the
spring from which an answer will be found. This notion serves as part of the basis
for “offender-based research,” or “OBR” for short. Broadly defined, OBR refers to
scholarly work that entails analyzing data collected directly from criminals
(Bernasco, 2010). Examples include everything from the survey of students used by
Hirschi (1969) in Causes of Delinquency to experiments involving offenders (e.g.,
Decker, Wright, & Logie, 1993) and to ethnographies such as Shaw’s The Jack-
Roller (1930) and Sutherland’s (1937) The Professional Thief.
Because offenders are uniquely situated to inform criminology, the potential of
OBR for improving crime prevention is enormous. In this chapter, we outline five
ways that OBR can be used to inform understanding of crime prevention, more

1
Of course, few if any individuals fill a single criminological role. In one situation, for example, a
person may act as an offender, but in another situation be a victim or guardian. Yet for the sake of
simplicity, in this chapter we refer to one role at a time.
S. Jacques (*) • E. Bonomo
Department of Criminal Justice and Criminology, Georgia State University,
P.O. Box 4018, Atlanta, GA 30302, USA
e-mail: sjacques1@gsu.edu

© Springer International Publishing Switzerland 2017 9


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_2
10 S. Jacques and E. Bonomo

specifically situational crime prevention or “SCP.” Each of the lessons is illustrated


with an example from the OBR literature. We conclude by discussing the choices
involved in conducting OBR for the betterment of SCP.

Situational Crime Prevention

The theoretical premise of SCP is that crime may be prevented by reducing the
characteristics of situations that facilitate offending (Clarke, 2009). The practical
purpose, then, of SCP is to provide a theoretical framework for deducing ways to
block crime opportunities by manipulating the specific situational characteristics
that generate them (Clarke, 2009). To do so, SCP draws on the rational choice and
routine activity perspectives (Clarke & Cornish, 1985; Cohen & Felson, 1979).
The rational choice perspective suggests that an individual’s decision on how
to act is guided by bounded assessments of potential benefit and cost (Bentham,
1988 [1789]). According to this perspective, an individual is more likely to com-
mit an offense when the perceived benefit of doing so is greater or the perceived
risk of incurring cost is smaller (Clarke & Cornish, 1985). Also, a person is more
likely to commit an offense when its perceived utility—i.e., benefits minus
costs—is greater than the utility of obeying the law. Examples of benefits include
money, status, and pride; examples of costs include fines, diminished respecta-
bility, and shame.
Key to the rational choice perspective is that the decision to offend takes place
within a concrete situation, defined as a particular time and place. This is crucial
because situations vary in the opportunities they hold for crime (Cohen & Felson,
1979). One aspect of opportunity is the potential benefit and cost associated with a
particular situation. For instance, the opportunity for robbery is greater when there
is a potential victim with a lot of rather than no cash in possession. And typically,
the opportunity for crime is smaller when a police officer is nearby instead of out of
sight. The second aspect of opportunity is whether the situation presents the mini-
mal elements involved in committing an offense. For example, a motivated offender
cannot rob someone if no one is around to victimize, and a drug distributor cannot
sell his or her product if alone on a deserted island. While such limitations may
seem so obvious as to not even warrant mentioning, their obviousness shows that
opportunity is a crucial component in offending.
Returning to SCP, this framework is based on the notion that a person is less
likely to commit a crime when there is less opportunity, meaning the crime’s poten-
tial benefit is smaller, risk is greater, or the minimal elements to do so are absent.
SCP is more than an academic pursuit, however, as its purpose is to guide real-life
crime prevention efforts. On that note, SCP practices may be categorized as “strate-
gies” or “techniques.” SCP “strategies” are broad, or abstract, notions of how to
introduce discrete managerial and environmental change that helps reduce crime
opportunities (Clarke, 2009). The five strategies of SCP are (1) increasing the effort
and (2) risk involved in lawbreaking and (3) reducing the reward, (4) excuse, and
2 Learning from the Offenders’ Perspective on Crime Prevention 11

(5) provocation to do so (Clarke, 2009). SCP “techniques,” on the other hand, are
specific procedures based on a given strategy. Examples of the five strategies listed
above are, respectively, target hardening, introducing burglars alarms, disrupting
markets, curtailing disputes, and displaying signs that specify the law (for further
examples, see Clarke, 2009).

Five Ways to Learn from OBR About SCP

OBR’s distinguishing feature is analyzing data collected directly from offenders.


The motivation for this research procedure is the assumption that offenders are
uniquely situated to inform criminology. Criminals have personal knowledge of
how often they commit crime; what motivates offenses; and how they are commit-
ted. Moreover, offenders know how often their crimes are prevented; what thwarts
offenses; and how so. And therefore, OBR can contribute to knowledge about
SCP. There are at least five ways to learn from OBR about SCP.
Perhaps the most obvious way is to conduct experiments on, make observations
of, or ask questions of offenders that directly speak to the effectiveness of a particu-
lar SCP technique. An example of the first lesson is provided by Leclerc, Wortley,
and Smallbone (2011). They drew on data collected from a sample of adult offend-
ers who sexually abused children to examine the efficacy of potential victims’ self-
protection techniques. The respondents reported that the most productive way for
children to prevent sexual contact was to tell the offender that such contact was not
wanted or to say “no,” which are examples of reducing the “excuse” to commit a
crime—namely that the victim was an active participant. According to the offend-
ers, more than half of the victims successfully avoided sexual contact by employing
these techniques. The participants also reported that fighting back and yelling for
help were the least effective methods of resisting a sex offender. Though these two
techniques are meant to deter offenses by increasing the required effort and risk, the
participants stated that these techniques were only successful at warding off an
offense in 11.8 % and 4 % of cases, respectively. As the authors note, the implication
of these findings is that assertiveness training may be the most valuable intervention
point for reducing sex offenses against minors.
A second way to learn about SCP from OBR is by making inferences about what
kind of SCP techniques to invent and employ based on criminals’ descriptions of
their motives and methods of offending. Copes and Cherbonneau (2006), for
instance, examined the methods used by auto thieves to steal keys that unlocked the
door to vehicles they targeted. While the researchers did not directly examine any
particular SCP technique, they drew on their findings to craft suggestions for how to
reduce this crime. One finding and concomitant implication is that because offend-
ers are able to steal keys to vehicles, owners should not treat mechanical vehicle
security measures (e.g., a car alarm)—which are meant to increase the risk or effort
of offending—as sufficient to protect against theft. Additionally, owners should take
strides to protect their keys. Of course, leaving keys out in the open is a bad idea, but
12 S. Jacques and E. Bonomo

it is also important to recognize that some hiding places are better than others. As
the authors note, “spare keys are better protected in the home, no matter how well
they are hidden within the vehicle … [and] it is safer to hide valuables, including
spare car keys, in places away from common searching areas. These safer areas
include basements, utility rooms, guest rooms and children’s bedrooms” (Copes &
Cherbonneau, 2006, p. 930). In the language of SCP, what Copes and Cherbonneau
propose is that auto theft can be prevented by increasing the effort involved in
obtaining keys.
Third, OBR can be used to inform SCP by using qualitative findings to refine
what is learned from quantitative analyses. An example relates to Weisburd and col-
leagues’ (2006) analysis of whether geographically targeted crime prevention
efforts cause displacement to nearby, untargeted areas. To test this idea, two areas
with high levels of street-level crime and disorder were subjected to increased police
patrol, which is a way of increasing the risk of offending. Those areas as well as two
neighboring areas were monitored during an experimental period to see whether
offending in the target area “moved around the corner” to the neighboring areas.
The researchers’ quantitative results, which were based on systematic social obser-
vations of the areas, suggest that the increased policing did not displace crime from
one area to the other, but rather reduced offending in both the targeted and nearby
locales. Wisely, the researchers also included an ethnographic component in their
investigation that entailed interviewing offenders. Among other findings, the quali-
tative data showed that some offenders chose a new method of crime commission
rather than quit altogether or move elsewhere. For example, “[r]esearch subjects
began prearranging dates by means of phone or beepers …; quizzing potential cli-
ents to ensure they were not police officers; [and,] disguising their looks and engag-
ing in stealthy solicitation” (Brisgone quoted in Weisburd et al., 2006, p. 581).
These adaptations likely reduced the overall rate of offending because, one, they
take extra time and, two, they are less obvious and therefore less likely to attract
potential customers. Thus, the qualitative findings suggest that increased policing
may have reduced offending not only by increasing its risk but also by increasing
the effort involved.
Related to the above is a fourth way in which OBR can add to understanding
of SCP: by finding out how offenders elude crime prevention efforts. From the
perspective of control agents, the preferred result of implementing an SCP tech-
nique is that offenders choose to stop breaking the law or, if they persist, are
incapacitated via institutionalization. The next best result is for persistent
offenders to commit fewer offenses—a process known as restrictive deterrence
(Gibbs, 1975; Jacobs, 1996). This commonly happens with drug dealers, as
many of them choose to make fewer sales in exchange for diminished risk
(Jacobs, 1999; Jacques & Wright, 2015). For example, selling from known drug
“hot spots” attracts drug buyers but also provokes attention from police; instead,
then, some dealers legitimatize their presence by selling at places known for
non-criminal activities, though this likely reduces their sales as well. St. Jean
(2007) interviewed a Chicago dealer who explained how this tactic undercuts
suspicions:
2 Learning from the Offenders’ Perspective on Crime Prevention 13

The corner stores, fast food joints, check cashing joints, and bus stops are all places where
people be for a purpose. So how you going to tell me I have no reason to be there, or that I
am here to sell dope. You don’t know that. I may be here to wash my clothes, buy a bottle
of water or something else from the corner store … But I can only convince you about this
if the businesses are there. So you see, this is why we hang out in these parts because ain’t
no telling exactly that you here to get or sell dope. (p. 124)

An implication of such an illicit business practice is that police, business owners,


and other stakeholders should work collaboratively to increase surveillance in these
areas. St. Jean (2007) suggests bringing all stakeholders together, both formal and
informal agents, to create a shared surveillance approach that would make “blend-
ing in” more difficult for offenders.
The fifth manner in which OBR can be used to shed light on SCP is by uncover-
ing the unintended, negative consequences of particular techniques (see also
Merton, 1936). Duneier’s (1999) classic ethnography, Sidewalk, illustrates how
policing measures intended to decrease crime may inadvertently increase some
offense types. The technique in question is order maintenance policing (OMP).
Based on broken windows theory (Kelling & Wilson, 1982), OMP is geared toward
preventing serious crimes like robbery and burglary by increasing the risks
involved in minor crimes and disorderly conduct such as littering, loitering, pan-
handling, and fare evasion (see, e.g., Bratton, 1998). Though OMP has been found
to substantially reduce serious crime (Zimring, 2011), it is by no means a perfect
method of crime control. In addition to complaints that it is practiced in a discrimi-
natory fashion (Gelman, Fagan, & Kiss, 2007), another problem is that some
efforts aimed at reducing specific offenses wind up increasing others. Based on
observations of and interviews with unlicensed street vendors, Duneier (1999)
shows how police officers’ attempts to control this type of so-called disorder led to
others. For instance, police would confiscate vendors’ property if left unattended
on the street, which was a way of increasing the risk of this quasi-legal business.
This posed a serious dilemma to vendors when they needed to use the restroom, as
they could not legally stay with their property and go to the bathroom at the same
time. Instead of jeopardize losing their vending materials, venders opted to urinate
in the street, albeit in a relatively secretive manner. Thus, police attempts to reduce
one type of disorder, street vending, had the effect of increasing another, public
urination. Such a finding suggests several practical implications: at the punitive
end is increasing the risk of public urination, whereas “harm reduction” approaches
include relaxing restrictions on time away from one’s vending booth and putting a
public lavatory nearby.

Conducting OBR for the Sake of SCP

As outlined and illustrated above, there are at least five ways in which OBR can
inform understanding of SCP: (1) by directly determining what works to reduce
crime; (2) generating findings that are suggestive of what prevention measures to
14 S. Jacques and E. Bonomo

invent and employ; (3) refining understanding of why a given prevention method
reduces crime; (4) figuring out how offenders get around particular prevention
measures; and (5) gathering information on not only the positive but also the unin-
tended, negative outcomes of prevention procedures. The first and second ways
inform the extent to which SCP strategies and techniques reduce crime. The third
and fourth lessons also do so, but their unique contribution is determining the
exact mechanism behind a significant or null effect. And the fifth way of learning
involves uncovering the consequences of SCP measures that are less visible until
the offenders’ perspective is taken into account. Thus, these five ways of learning
not only garner knowledge of “what works” in reducing crime—a purely empiri-
cal question—but also illuminate the theoretical forces behind such effects, why
some methods do not work, and the inadvertent but important problems that may
result as well.
Returning to the point made in this chapter’s introduction, OBR is uniquely able
to shed light on these issues. To be clear, there is no doubt that data obtained from
other sources, such as law enforcement officials and victims, can be used to inform
crime prevention. However, if the analytic focus is crime (not “policing,” “victim-
ization,” or whatnot), these other sources are more likely than OBR to result in
indirect or incomplete information. For instance, police statistics are merely a proxy
for crime because what they really measure is the amount of law enforcement
(Black, 1970). And while victim statistics are useful for knowing the extent and
nature of some crime types, like burglary and robbery, they are largely useless for
victimless crimes, which include drug distribution and prostitution. What is more,
offenders are better positioned than others to inform understanding of why and how
they are affected by crime prevention measures. Police, victims, or other non-
offenders may acquire that information somehow, but the source of that knowledge
will always be an offender. There is no getting around it: research with criminals is
the most direct route to information about crime, including how to prevent it.
For OBR to inform SCP, first OBR has to be conducted. In doing so, two major
decision points always arise: What procedure should be used to sample offenders
and should quantitative or qualitative data be obtained? Before closing, we provide
some thoughts on the respective merits of the options, and also briefly discuss the
limitations of OBR broadly.
The two major procedures of sampling offenders are via formal and informal
channels. A formal channel is any government institution that houses or keeps a
population list of (potential) offenders, such as a roster of persons in jail, prison,
high school, or on parole or probation (see, e.g., Copes & Vieraitis, 2012;
Papachristos, Meares, & Fagan, 2012). An informal channel is any route to offend-
ers that does not involve relying on a government institution (see, e.g., Wright &
Decker, 1994, 1997). The formal and informal channels have opposite difficulties.
The major difficulty associated with the formal channel is gaining permission of the
“gatekeeper.” By gatekeeper, we mean an individual who controls access to an insti-
tution and its members. If a gatekeeper’s permission is not obtained, it is practically
impossible to use the institution to facilitate research. With informal sampling, find-
ing offenders and convincing them to participate is the major challenge because
2 Learning from the Offenders’ Perspective on Crime Prevention 15

there is no population list to draw on. For that reason, a researcher who depends on
informal sampling will often recruit offenders who also happen to be friends, fam-
ily, coworkers, or students, and from there build a snowball sample (see, e.g.,
Jacques & Wright, 2008). The difficulty posed by this sampling procedure is that
even the most well-connected researcher will eventually exhaust his or her network
of criminal ties. And for some researchers, such a network will be practically non-
existent to begin with and thereby preclude research getting off the ground. One
way around this particular problem is for a researcher to tap into new networks, such
as by going to a known drug hot spot and trying to recruit there (see, e.g., Jacobs,
1999). Yet that poses another problem, namely the chance of being victimized in the
course of conducting research (Jacobs, 2006; Jacques & Wright, 2010).
Whether a researcher uses a formal or informal channel to sample offenders is a
crucial choice for some research questions. If the major goal of research is some-
thing like obtaining numbers generalizable to a particular population, clearly it is
best to make use of available population lists. A researcher who uses an informal
channel to recruit participants usually cannot make claims about statistical general-
izability because, for one, the true population of offenders is unknown and, even if
it was, informal sampling is rarely if ever truly random sampling (but see Copes,
Jacques, Hochstetler, & Dickinson, 2015). However, it will be better to use an infor-
mal sampling channel if the major goal of the research is more along the lines of
“Why do some offenders not get arrested?” Obviously, it is not possible to study
offenders who are yet to be arrested by sampling individuals who are in jail, prison,
or listed on a parole or probation roster. These are just a couple of examples meant
to illustrate how the different channels of sampling offenders have different strengths
and limitations (for further details, see Copes et al., 2015). If OBR is to reach its
potential for informing SCP, both sampling channels should be used.
Another important consideration for researchers is whether to use quantitative or
qualitative methods. At the simplest level of conceptualization, quantitative research
is based on numbers and qualitative research is not (Jacques, 2014). Analytically, a
major difference between the two is that quantitative research can be used to pro-
duce statistical statements about the significance of effects, whereas qualitative
research can only be used to create theoretically generalizable ideas (Small, 2006).
Thus, quantitative research may often be the better option for determining what SCP
techniques significantly reduce crime. However, the process that goes into conduct-
ing quantitative research often—though not necessarily—requires constraining a
researcher’s attention to a relatively narrow list of preconceived concerns; if this
limitation is not self-imposed, the numbers can become distorted by such things as
the ordering or wording of questions. Qualitative research, on the other hand, is
relatively free from these constraints because the validity of findings is not a statisti-
cal matter. In other words, the limitation of qualitative research is also a virtue in
that it allows for the introduction and discussion of previously unconsidered topics.
Moreover, this freedom facilitates an expanded and more detailed collection of
information. Answers are not boiled down to a number but rather are meant to be
spelled out in detail: exactly why and how a crime was committed, for instance.
Thus, qualitative research may be preferred when the goal is to refine understanding
16 S. Jacques and E. Bonomo

of why a given SCP method reduces crime, figuring out how offenders get around
particular SCP measures, or gathering information on their unintended, negative
consequences.
This chapter has presented a sunny picture of OBR, but this method certainly has
limitations (see Bernasco, 2010; Copes et al., 2015). Perhaps the most widely circu-
lated criticism is “Why should we believe what offenders tell us?” As Richard
Wright puts it, we should expect offenders to lie for the very reason that they are
offenders. In other words, people who break the law are untrustworthy. What is
more, offenders may be motivated to distort the truth because revealing too much is
risky. While there is likely some truth to such assertions, the same can often be said
of victims, police, and other sources of criminological data. For example, victims
may lie to gain justice, and police officers may lie to protect their colleagues. All of
this is to say that offenders are not the only party with “good” reasons to tell less
than the whole truth. Nonetheless, the lessons of OBR for SCP are brought into
question to the extent that offenders are providing researchers with fiction.
In conclusion, we would like to emphasize that the five lessons of OBR for SCP
should not be thought of as independent ventures, but rather as a curriculum. Each
of the ways OBR contributes to knowledge of SCP is useful in its own right, but also
because each lesson informs the totality. Whether, for example, an SCP technique
effective in reducing crime is important, but knowing the full utility of a technique
also depends on knowing what, if anything, are its unintended, negative conse-
quences. And figuring out how offenders circumvent particular SCP techniques
may be used to improve or build the repertoire of effective methods available to
control agents. In short, unlocking the full potential of OBR for informing SCP will
require research geared toward all of its lessons.

References

Bentham, J. (1789/1988). The principles of morals and legislation. Amherst, NY: Prometheus Books.
Bernasco, W. (Ed.). (2010). Offenders on offending: Learning about crime from criminals.
Cullompton, UK: Willan Publishing.
Black, D. (1970). Production of crime rates. American Sociological Review, 35, 733–748.
Bratton, W. (1998). The turnaround: How America’s top cop reversed the crime epidemic.
New York, NY: Random House.
Clarke, R. V. (2009). Situational crime prevention: Theoretical background and current practice. In
M. D. Krohn, A. J. Lizotte, & G. P. Hall (Eds.), Handbook on crime and deviance (pp. 259–
276). New York, NY: Springer.
Clarke, R. V., & Cornish, D. B. (1985). Modeling offenders’ decisions: A framework for research
and policy. Crime & Justice, 16, 147–185.
Cohen, L., & Felson, M. (1979). Social change and crime rate trends: A routine activity approach.
American Sociological Review, 44, 588–608.
Copes, H., & Cherbonneau, M. (2006). The key to auto theft: Emerging methods of auto theft from
the offenders’ perspective. British Journal of Criminology, 46, 917–934.
Copes, H., Jacques, S., Hochstetler, A., & Dickinson, T. (2015). Interviewing offenders: The active
vs. inmate debate. In H. Copes & M. Miller (Eds.), Routledge handbook of qualitative crimi-
nology. London, UK: Routledge.
2 Learning from the Offenders’ Perspective on Crime Prevention 17

Copes, H., & Vieraitis, L. M. (2012). Identity thieves: Motives and methods. Boston, MA:
Northeastern University Press.
Decker, S., Wright, R., & Logie, R. (1993). Perceptual deterrence among active residential bur-
glars: A research note. Criminology, 31, 135–147.
Duneier, M. (1999). Sidewalk. New York, NY: Farrar, Straus and Giroux.
Gelman, A., Fagan, J., & Kiss, A. (2007). An analysis of the New York City Police Department’s
‘Stop-and-Frisk’ policy in the context of claims of racial bias. Journal of the American
Statistical Association, 102, 813–823.
Gibbs, J. P. (1975). Crime, punishment, and deterrence. New York, NY: Elsevier.
Hirschi, T. (1969). Causes of delinquency. Brunswick, NJ: Transaction Publishers.
Jacobs, B. (1996). Crack dealers and restrictive deterrence: Identifying narcs. Criminology, 34,
409–431.
Jacobs, B. (1999). Dealing crack. Boston, MA: Northeastern University Press.
Jacobs, B. (2006). The case for dangerous fieldwork. In D. Hobbs & R. Wright (Eds.), The sage
handbook of fieldwork (pp. 157–168). Thousand Oaks, CA: Sage Publications.
Jacques, S. (2014). The quantitative–qualitative divide in criminology: A theory of ideas’ impor-
tance, attractiveness, and publication. Theoretical Criminology, 18, 317–334.
Jacques, S., & Wright, R. (2008). Intimacy with outlaws: The role of relational distance in recruit-
ing, paying, and interviewing underworld research participants. Journal of Research in Crime
& Delinquency, 45, 22–38.
Jacques, S., & Wright, R. (2010). Dangerous intimacy: Toward a theory of violent victimization in
active offender research. Journal of Criminal Justice Education, 21, 503–525.
Jacques, S., & Wright, R. (2015). Code of the suburb: Inside the world of young middle-class drug
dealers. Chicago, IL: University of Chicago Press.
Kelling, G. L., & Wilson, J. Q. (1982). Broken windows: The police and neighborhood safety.
Atlantic Monthly, 249, 29–38.
Leclerc, B., Wortley, R., & Smallbone, S. (2011). Victim resistance in child sexual abuse: A look
into the efficacy of self-protection strategies based on the offender’s experience. Journal of
Interpersonal Violence, 26, 1868–1883.
Merton, R. K. (1936). The unanticipated consequences of purposive social action. American
Sociological Review, 1, 894–904.
Papachristos, A. V., Meares, T. L., & Fagan, J. (2012). Why do criminals obey the law? The influ-
ence of legitimacy and social networks on active gun offenders. Journal of Criminal Law &
Criminology, 102, 397–440.
Shaw, C. R. (1930). The jack-roller: A delinquent boy’s own story. Chicago, IL: University of
Chicago Press.
Small, M. L. (2006). ‘How many cases do I need?’ On science and the logic of case selection in
field-based research. Ethnography, 10, 5–38.
St. Jean, P. K. B. (2007). Pockets of crime: Broken windows, collective efficacy, and the criminal
point of view. Chicago, IL: University of Chicago Press.
Sutherland, E. H. (1937). The professional thief. Chicago, IL: University of Chicago Press.
Weisburd, D., Wyckoff, L. A., Ready, J., Eck, J. E., Hinkle, J. C., & Gajewski, F. (2006). Does
crime just move around the corner? A controlled study of spatial displacement and diffusion of
crime control benefits. Criminology, 44, 549–592.
Wright, R., & Decker, S. (1994). Burglars on the job: Streetlife and residential break-ins. Boston,
MA: Northeastern University Press.
Wright, R., & Decker, S. (1997). Armed robbers in action: Stickups and street culture. Boston,
MA: Northeastern University Press.
Zimring, F. E. (2011). The city that became safe: New York’s lessons for urban crime and its con-
trol. New York, NY: Oxford University Press.
Chapter 3
Learning About Crime Prevention
from Aborted Crimes: Intrapersonal
Comparisons of Committed and Aborted
Robbery

Marie Rosenkrantz Lindegaard and Wim Bernasco

Introduction

In this chapter we demonstrate that a lot can be learned about crime and crime pre-
vention by studying aborted crimes: crimes that were not committed despite being
anticipated by the prospective offenders. Getting to know why offenders sometimes
decide to call off a crime they have anticipated is an uncommon but potentially use-
ful way to gain insights into why certain crime prevention measures may actually
work. Our discussion of the potential of offender-based research is based on a study
that included an intrapersonal comparison of committed and aborted robberies. Our
aim in this chapter is twofold. The first aim is to report our findings about aborted
robberies. The second aim is to elaborate on the strengths and weaknesses of our
research design, and thereby provide recommendations for future studies. We will
use our study to answer the following three substantive questions:
1. What distinguished robberies that were aborted from robberies that were
committed?
2. What mechanisms explain why some robberies get aborted and others do not?
3. What reasons do offenders provide for aborting robberies?
We draw on interviews with 104 incarcerated and non-incarcerated robbery
offenders, in which we collected information about aborted and committed robber-
ies both involving the same offender. In our analysis we use answers to both open

M.R. Lindegaard (*)


Netherlands Institute for the Study of Crime and Law Enforcement (NSCR),
De Boelelaan 1105, 1081 HV Amsterdam, Netherlands
e-mail: Mlindegaard@nscr.nl
W. Bernasco
Department of Spatial Economics, Netherlands Institute for the Study of Crime
and Law Enforcement (NSCR), Vrije Universiteit Amsterdam, Amsterdam, Netherlands

© Springer International Publishing Switzerland 2017 19


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_3
20 M.R. Lindegaard and W. Bernasco

and closed questions about characteristics of the committed and aborted robberies
and the situations in which they occurred. These answers include the reasons for
aborting robbery that the offenders themselves provided during the interviews.
Although not all findings are easily applicable in a situational crime prevention
program, we believe that the method of analyzing aborted crime offers great poten-
tial for combatting crime, as it can potentially highlight situational factors that pre-
vent crimes about to be committed.

Counterfactual Strategy

Because experimental designs are seldom feasible in the study of crime, a good
alternative counterfactual strategy is important for explaining crime. An appropriate
counterfactual strategy seeks situations without crime that are similar to situations
with crime. This attempt underlies a variety of criminological studies of conditions
in which crime is decreasing or even absent, including the explanation of decreasing
crime statistics (Tonry, 2014), periods of peace in gang conflicts (Vargas, 2014),
de-escalating behavior during aggressive incidences in public drinking spaces
(Levine, Taylor and Best 2011), nonviolent methods in drug market conflicts
(Jacques, 2010), and desistance from crime over the life course (Maruna, 2001).
Common to these studies is the wish to explain crime and criminal decision making
by understanding conditions for non-criminal behavior. They obviously differ in
terms of their unit of analysis: periods with high and with low crime figures, periods
in the life course with and without criminal activities, and behavior during criminal
and non-criminal events.
Where studies of criminal behavior over the life course tend to focus on differences
in criminal behavior across various periods in the life of the same person (e.g.,
Blokland & Nieuwbeerta, 2005), studies of the effects of situational characteristics on
criminal behavior tend to compare situational characteristics involving different per-
sons; that is, they make between-person comparisons. For instance, in a study of drug
market conflicts, Jacques (2010) aimed at explaining the variable conditions of violent
and nonviolent retaliation without making intrapersonal comparisons. To identify the
situational characteristics of lethal assaults, Ganpat, van der Leun, and Nieuwbeerta
(2013) compared court case descriptions of lethal and nonlethal cases, each involving
both different offenders and different victims. Levine, Taylor, and Best (2011) focused
on explaining why conflicts in public drinking places sometimes escalated into physi-
cal aggression and sometimes de-escalated. Graham et al. (2006) observed conflicts in
bars with the aim of understanding why some conflict situations ended up with physi-
cal aggression while others did not. While the aim of these four studies was to under-
stand the situational circumstances for violent and nonviolent behavior, none of them
compared behavior of the same person in different situations. This design may con-
found personal and situational characteristics, and is therefore not the most rigorous
approach for assessing either personal or situational causes of criminal behavior.
In criminology only a handful of studies of situational characteristics used intra-
personal comparisons of criminal and non-criminal behavior. Phillips (2003)
3 Learning About Crime Prevention from Aborted Crimes… 21

applied a matched-case control, comparing violent and nonviolent management in


two conflicts from the same period in the life of an individual. Lussier, Leclerc,
Healey, and Proulx (2008) analyzed crime switching patterns over time of the same
offender. Leclerc, Lussier, and Deslauriers-Varin (2014) analyzed offending strate-
gies between different sexual offenses of the same offender. Hewitt and Beauregard
(2013) compared different levels of aggression in a series of crime events involving
the same offender. Bernasco, Ruiter, Bruinsma, Pauwels, and Weerman (2013) ana-
lyzed situational characteristics of moments of offending with moments of non-
offending of the same persons. The study we address in this chapter aimed at
explaining the effects of situational characteristics on the decision of an individual
to either commit or abort an anticipated robbery.
Our counterfactual strategy thus included two elements. The first element was
that the analysis did not compare a committed robbery with any other situation, but
with a situation that in many aspects was quite similar: an anticipated robbery that
was aborted. The second element was the intra-person comparison: we did not com-
pare a committed robbery to any other aborted robbery, but to an aborted robbery
involving the same individual.

The Present Study

In our study of robberies, we compared three different types of robberies carried out
by the same person: (1) robberies with threat, (2) robberies with physical violence,
and (3) anticipated robberies that were aborted. Respondents were asked to describe
the most recent robbery they committed for each category and to focus on robberies
that took place within a 5-year period prior to the interview. In the present discus-
sion, the first two types are joined and treated as “committed robberies” in order to
compare their situational characteristics with the third type: robberies that were
never committed. The definition of robbery was broad enough to include street rob-
beries, commercial robberies, and home robberies.

Aborted Robberies

We defined aborted robberies as robberies that the respondent had anticipated com-
mitting but decided not to carry out. Our counterfactual case was therefore defined
in relation to the decision-making process: respondents had a target on their mind,
but for whatever reason they decided not to rob the particular target. The character-
istic for an aborted robbery was that the potential victims were (and probably have
always remained) unaware that they were about to get robbed. This rather broad
definition of anticipated robberies included robberies that were planned well in
advance, but got aborted seconds before they were supposed to take place. It also
included robberies that were considered on the spot, but not carried out for some
22 M.R. Lindegaard and W. Bernasco

reason. It also included anticipated robberies that were being discussed among
friends at home, but had not yet been planned.
We interviewed 104 male respondents in the Netherlands who claimed that they
had experiences with committing a robbery. A minority (28) was not incarcerated
and was recruited via a snowball sample on the streets of Amsterdam. The other 76
were recruited in adult prisons (41) and juvenile facilities (35). Each interview
focused on all three types of robbery situations. Since the non-incarcerated respon-
dents found it difficult to concentrate as long as required to go through three types
of robbery events, we decided to leave out, if necessary, aborted robberies among
the non-incarcerated respondents. In practice that meant that we only succeeded
getting information about aborted robberies from five non-incarcerated offenders.
For the purpose of this chapter our sample is therefore 81 respondents.
All respondents who were asked to provide details about an aborted robbery were
able to do so without hesitation. In fact, remembering robberies that they had antici-
pated but decided not to perpetrate seemed easier than recalling the ones that had been
actually carried out. The amount of detail about the situation in which the robbery was
supposed to have taken place varied largely depending on where in the decision-mak-
ing process the respondent had decided not to carry out the robbery. In most cases there
had not yet been established any contact with the victim. Our definition of aborted
robberies therefore differed from “unsuccessful robberies,” where the offender did not
manage to get away with any valuables. In our definition unsuccessful robberies were
“committed robberies.” Aborted robberies were anticipated but had not yet been started.

Recruitment

The non-incarcerated respondents were recruited through a recruiter known to the third
author. The recruiter earned 20 Euros per recruited respondent. The non-incarcerated
respondents were paid 50 Euros for an interview. The incarcerated respondents were
asked for participation face to face by the first author and were paid 30 Euros per inter-
view. The response rate was 66.5 % (see Lindegaard, Bernasco, and Jacques, 2014 for
details about non-response). The only criterion for participation was explained as “hav-
ing experience with committing robberies.” Respondents were never asked to reveal
their real names and were promised confidentiality. The research design and contents
of the questionnaire and interview protocol were reviewed and approved by the ethics
committee of the Law Faculty of Vrije Universiteit Amsterdam.

Sample and Interviews

For the purpose of this chapter we describe 155 committed robberies and 81
aborted robberies. Robberies referred to were committed on the street, in shops and
other types of retail businesses, and in private homes. Respondents were asked to
3 Learning About Crime Prevention from Aborted Crimes… 23

focus on the most recent robberies within the three categories of robberies (threat
only; physical violence; aborted) and only on robberies committed within 5 years
prior to the interview or to their incarceration. Interviews lasted between 20 min
and 3 h depending on the number of robberies described by respondents, their
willingness to discuss the robberies in details, and their ability to reflect on their
experiences. Interviews were recorded and transcribed verbatim. Accounts of 30
robberies were cross-checked with media coverage descriptions and consensus
was found in all cases. Official records were not consulted.
Even though the interviews were semi-structured, all interviews focused on the
same key topics following the same order of attention. Focus was given to occur-
rences during the event rather than on the motivations of the offender. Each descrip-
tion of a robbery event started out with open questions where the respondent was
asked to describe the robbery as if it was a movie unknown to the interviewer. This
open description was followed by closed questions about specific situational char-
acteristics that we specifically wanted to identify (for further reflection on content
of interviews, establishment of rapport, and personal characteristics of the respon-
dents, see Lindegaard et al., 2014).

Analysis

To analyze the differences between committed and aborted robberies, we drew on


answers to the closed questions, and used a case-control design in which 14 char-
acteristics of an aborted robbery were compared to those of one or two robberies
committed by the same offender. This design helped to assure that any differences
discovered cannot be attributed to stable personal characteristics of the offender
(because both cases involved the same offender) and must therefore be related to
differences between situations. We estimated univariate and multivariate fixed-
effect logit models to assess the effect of the characteristics on whether the rob-
bery was committed or aborted. The 18 variables refer to where the robbery was
supposed to take place (residence or not), when it was supposed to take place
(during darkness or daylight), various aspects of planning (whether it was planned
longer than a day, hour or minute, whether clothing, method, neighborhood, street,
target, and time had been selected, whether tips had been provided), whether co-
offenders were involved, whether multiple victims were involved, whether a vic-
tim was known to the offender, whether the victim was known to be a criminal,
and whether the offender was under the influence of drugs including alcohol,
cannabis, or hard drugs.
The description of mechanisms that might explain the decision of the offender to
abort the robbery builds on hypotheses developed upon the completion of the inter-
views. It still needs verification through extensive analysis of all interview transcrip-
tions. Our findings of the subjective reasons for why the respondent decided to abort
the robbery are based on analysis of answers to the open question “what was the main
reason that you did not commit this robbery?” We coded the answers into seven
24 M.R. Lindegaard and W. Bernasco

different categories: presence of police, moral, bystanders, physical, fear of victim, co-
offenders, and doubt. Although the description of mechanisms was tentative, the rea-
sons provided by the respondents were representative for the sample but clearly
suffered from the problem of being post-rationalizations of their decision-making pro-
cess. We decided to present both types of findings because they illustrate some of the
potential strengths and weaknesses of learning about crime by focusing on aborted
crimes.

Findings

Characteristics of Aborted Robberies

Table 3.1 presents the result of the 18 univariate fixed effects logit models that link
characteristics of potential robbery situation to their outcome, i.e., whether they
were committed or aborted. Even in a univariate analysis, most of the characteristics
did not seem to have any systematic and significant effect on whether the robbery is
committed or aborted. The three exceptions were whether the robbery was commit-
ted in a home (more likely to be committed), whether it was planned long in advance
(more likely to be aborted), and whether the victim was known to the offender (less
likely to be committed).
Because the univariate regression models did not account for correlations
between the situational characteristics of anticipated robberies, a multivariate
model was required to tease out which were the major characteristics that deter-
mined the outcome. Using initially all 18 variables listed in Table 3.1, a multi-
variate model was determined that optimally described the situational
characteristics that affected whether an anticipated robbery was committed or
aborted. The results demonstrated that when the anticipated robbery was a home
robbery (OR = 2.83, p < 0.10) it was more likely to be committed, and when it
was planned more than 1 h in advance (0.37, p < 0.05) it was more likely to be
aborted.

Suggested Mechanisms Potentially Explaining Aborted


Robberies

We analyzed descriptions of the decision-making process of both committed


and aborted robberies in order to explain the two findings about aborted robber-
ies: (1) home robberies were less likely to get aborted than commercial and
street robberies; and (2) robberies that were planned for more than an hour were
more likely to get aborted than robberies that were planned for less than an hour.
We illustrate these suggestive mechanisms with quotes from the respondents.
3 Learning About Crime Prevention from Aborted Crimes… 25

Table 3.1 Relations between committing (as opposed to aborting) and characteristics of
anticipated robberies. Estimates of 18 separate univariate fixed effect logit models
Characteristics of anticipated robbery Odds ratio Robberies Persons
Location is home 2.22* 198 79
Dark during robbery 1.66 182 73
Planning longer than 1 minute 1.35 202 81
Planned longer than 1 hour 0.44** 202 81
Planning longer than 1 day 1.02 202 81
Clothes chosen 1.28 202 81
Method chosen 1.62 202 81
Target chosen 0.78 202 81
Neighborhood chosen 0.74 199 80
Street chosen 0.78 199 80
Time chosen 1.45 195 78
Tip provided 1.35 200 80
Co-offender(s) 0.81 200 80
More victims 2.00 115 48
Known victim 0.34** 111 47
Criminal victims 1.49 202 81
Any drugs (including alcohol) 1.72 200 80
Alcohol 2.12 200 80
Cannabis (hash, marihuana) 1.00 200 80
Hard drugs (ecstasy, cocaine, heroin) 4.65 202 81
Source: Bernasco et al. (2013) (drug use variables inserted here)
* = p < .10; ** = p < .05

Illustration of Home Robberies

The descriptions by our respondents suggested that they were less likely to abort
home robberies than from other kinds of robberies because (1) victims of home rob-
beries were aware of being robbed as soon as contact was established with the
offender, whereas with other kinds of robberies they could still withdraw after con-
tact was established with the victim; and (2) they expected home robberies to be
more messy than other types of robberies, meaning that “unexpected” occur-
rences—things that did not occur according to the plan—were a part of the package
when going for a home robbery, whereas unexpected events in street robberies and
commercial robberies were seen as a reason to abort.
Respondent 86 described a situation where he carried out a home robbery despite
a whole range of unexpected events (victim came downstairs, he was naked, he
refused to tell where the money was). Instead of considering backing off and giving
up due to these events, he described using extensive physical aggression:
Respondent: yes, yes, and we were inside, everyone was silent, everyone was scared, lots of
adrenaline, mainly for me, and they went searching. By the garden door they had like a
curtain with beads or something and that made a noise. So that man came downstairs. He
was all naked. I know it sounds weird but it really happened miss.
26 M.R. Lindegaard and W. Bernasco

Interviewer: Yeah, right


Respondent: He came downstairs all naked. I stood next to a ‘soldier’ [gang member]
and I looked at him [victim]. And I was like shit!
Interviewer: And he was really naked?
Respondent: He was naked!
Interviewer: Oh …
Respondent: And I was, I was … like I just had a blackout, you know. When I saw that
naked man my first reaction was I walked up to him. I hit him eight times with a hammer. On
his face miss! And he fell down on the ground but he kept trying to stand. That soldier, that
soldier, he tried to help my two Dutch friends. Afterwards they told me, yes they were scared,
they did nothing, they just stood there like statues. That’s why he couldn’t help me. So I had
to struggle alone with that man. But I kept saying, like really aggressively: “Shut up, shut the
fuck up! Where is the money?” Really quietly but very aggressively he replied: “I have noth-
ing! I have nothing!” I hit him again with that hammer, on his back, on his head, towards his
ear, and it felt as if my whole hand was covered by blood. I was completely covered by
blood. But it wasn’t really like that. And he… and I realized that my three friends found
nothing. That soldier of mine, he picked up a box and we thought the money would be in
there. So they left the house and I stayed with that … I kept pushing that man towards the
ground, pushing him down, right. And when I … when I saw my last friend run out of the
door I also left.

Respondent 67 explained why he carried out a home robbery that was unexpect-
edly difficult. According to him it was impossible to back off because the victims were
already aware that they were getting robbed. Instead of leaving the place when they
were faced with obstacles they took extraordinary risks, and eventually got caught:
So we started putting on gloves. The boy, young know the one I told you about who was
specialized in opening doors, he was unable to do it. But I thought I could also do it. I had seen
it on TV. So he tried, tried, tried. It was like a corridor apartment, right. And in corridor apart-
ments, neighbors more quickly hear what’s going on. And it was already three o’clock in the
morning. We tried to open the door. It didn’t work out, you know what I mean? But I thought:
“Hey, we’re already here. We already revealed who we are, right, we’re not leaving with
empty hands”. So we tried to force it. Force it, you? And, those guys, they weren’t afraid. We
tried, and we had to make more noise, The boy said: “OK, at some point I’m going to push
and then you have to break that door.” So we kicked the door. People inside already heard
everything for a long time. We try and try and we don’t get that door open. The kitchen win-
dow was next to the door. We broke the kitchen window, and opened it, and we came inside
like that. One boy had a gun, you know, I had a crowbar in my hand, and another one, the boy
who could open doors said: “Hey, can I also have a crowbar in my hand?” So we went inside.

During street robberies respondents were more likely to deal with unexpected
obstacles by aborting from robberies and looking for new targets. Respondent 58
described how he together with co-offenders kept changing the target after they had
put their mind to doing a robbery:
Because you wait, you keep waiting, you keep waiting, and you think … by every person
you think: “OK, now?”, “No, no, no, the next one, the next one, the next one”. So you
remain in a kind of tension, you remain in the tension, yes. That sucks! The tension really
sucks! Because you fight against your own adrenaline. You also fight against your will!
Because you want to do it now. And then someone [co-offender] says: “No, no, no, wait,
wait, wait”. So you get a little angry at the other one that says: “No, why not that one?” “No,
no, no, wait a little, because there’re two other people coming.” “Yes, let’s go! Let’s go and
get those two!” “No, no, no, wait, wait, wait.” Then he says, “Yes, now, now!” I say, “Yes,
yes, now I don’t want it” And we just continue like that.
3 Learning About Crime Prevention from Aborted Crimes… 27

Illustration of Planning Robberies

According to our respondents’ descriptions, when robberies were planned more


than an hour in advance, and something unexpected happened, they would rather
abort than changing the target. Respondent 44 described how he, for weeks in
advance together with co-offenders, had planned a robbery of a large compartment
store in detail. One of the employees in the store had provided them with informa-
tion and was going to open the safe for them while pretending to be unaware of the
robbery. They were staying in a hotel outside the city and had rented a bus for trans-
portation. When they arrived at the scene, there were police everywhere:
It was a large branch of HEMA [Dutch department store]. We drove there. And apparently
other people also knew about it. There was a police car in front of the door, and all kinds of
agents. Someone had probably informed them and they knew about it. And then we left.

Changing the target would require weeks of planning. Robberies that were
planned for less than an hour were commonly referred as “going hunting,” like
described by respondent 58 above. “Going hunting” meant wanting to do a robbery
without having identified the target yet. While being in a state of looking for oppor-
tunities, they were flexible about their targets and therefore less likely to abort from
a robbery they anticipated. Long periods of anticipation, as for robberies planned
more than an hour before the event, also seemed to create levels of doubt that were
more difficult to overcome than robberies that were decided upon on the spot.
Respondent 57 described how he was in doubt about a robbery he had planned for
weeks. He would normally use drugs to suppress his doubt. In this particular case
he had not used drugs and was looking for excuses to avoid committing the
robbery:
At least I know we were sitting somewhere. With the three of us. I think it was the same two
guys. And we were waiting. And eventually … Yeah, I also didn’t feel it like at all. I had
also not used drugs, and I wanted to leave. And then I eventually… I was also the driver of
the car, so they depended on me, and then I think I called my sister. Then I said, you need
to send me this SMS, then I changed the name to my mother’s name, just like ‘mom’ in my
phone. And then I made her send me a SMS: “You have to get home now. Something bad
happened. I want you to come now.” Then I used that phone to say: “Look, a SMS from my
mother,” I said, “I must go home, we have to leave.” And then they immediately came
along, because they agreed. They could see from the SMS like, “yes, I really have to go”.
Well, then we left [laughing].

Respondent 43 described how he was hanging out with his friends at home when
deciding to go somewhere to rob someone. The period of “hunting” was very short
for them. He claimed that they simply went to the local shopping street and grabbed
the first person they met:
Respondent: I just told those guys: “We’re robbing someone”. We were talking and I
grabbed the guy. I got his wallet, and ran away.
Interviewer: And you were just chilling out? Or what did you do, I mean, when you
came up with the idea?
Respondent: We were smoking a joint, watching TV, like that.
Interviewer: So you think it was also related to like boredom?
28 M.R. Lindegaard and W. Bernasco

Respondent: Yes, boredom


Interviewer: … that you had nothing to do. And did you just get the first possible person
or how did that go?
Respondent: Just someone, just someone random.
Interviewer: Yeah?
Respondent: Yes, it didn’t matter who it was

Subjective Reasons for Aborted Robberies

When the respondents had described the situations of aborted robberies we asked
them to provide the main reason for why they had aborted the robbery. This inter-
pretation afterwards provided insights into reasons that were not asked in the closed
questions. Where our findings of the comparison of the situational characteristics of
committed and aborted robberies showed that the location of the robbery and the
length of planning mattered for the likelihood to abort, the offenders provided a
larger amount of characteristics that according to their memory had mattered for
why they had aborted from committing an anticipated robbery. Despite our doubt
about whether these reasons were merely post-rationalizations of the event than to
the real considerations that made them decide to abort, we provide their answers
because they might be useful to include in a more systematic fashion in future
research about aborted events. We describe the reasons, divided into a limited num-
ber of categories, below: presence of police, moral concerns, bystanders, physical
circumstances, fear of victim, co-offenders, and doubt. The number of times that the
reasons were mentioned by the respondents is presented in Table 3.2.

Presence of Police

This category of reasons included perceptions that the risk of getting arrested during
the robbery was too great. Respondent 31 described how they realized that they
were being observed on cameras and therefore decided to abort the robbery:
We wanted to do it. First, we were planning to come by car. Then we would be fast, fast.
Around the corner, there was a garage. There you could stop the car. But, yeah, when we

Table 3.2 Subjective reasons Reasons aborted # %


for aborted robberies
Other 18 22.2
Police 17 21.0
Moral 13 16.0
Bystanders 9 11.1
Physical 10 12.3
Fear of victim 7 8.6
Co-offenders 7 8.6
Total 81 100.0
Source: Bernasco et al. (2013)
3 Learning About Crime Prevention from Aborted Crimes… 29

wanted to do it, it didn’t work out because there were cameras around there. From the police
station. The police station was behind the garage. We didn’t expect that there would be so
many cameras around. And then, yeah, then we eventually decided not to do it.

Moral Concerns

This category included reasons related to moral concerns such as not wanting to rob
people in the presence of children or not wanting to rob older people or women.
Respondent 33 explained how he had aborted a robbery when he realized that the
victims would be female:
Respondent: I don’t want that, I didn’t want that, no. You know, everything that has to do
with women, even if it’s a whore, it doesn’t matter. It’s a person, you know, it’s a woman.
Women you need to… I don’t want to hurt them, man. Women are a bit sacred to me.
Interviewer: What was the main reason that you didn’t want to do the robbery? That was

Respondent: Women, man.
Interviewer: Had to do it with the profession?
Respondent: Yes, also. OK, you know, we do … Some prostitutes who have to work,
you know, they get forced, and then you just steal their money. You make it a lot harder for
them. No man! I just refused to do that!

Bystanders

This category included reasons like there were too many bystanders or the bystand-
ers were too dangerous. Respondent 42 explained how there were too many custom-
ers in the store they wanted to rob. According to him, the risk that one of them
would play what was referred to as “hero” was too high:
Respondent: There were a lot of people and I thought like: No man, today is a busy day. I
don’t go.
Interviewer: Many customers?
Respondent: Yes, and I don’t want customers to be there when I enter. Maybe someone
will try to play the hero or something like that. Then I have to do something. And I’m not
planning to get caught for murder, or for attempted murder, no man. I only come for the
money. Not to hurt someone, you know what I mean?

Physical

Reasons included in this category were circumstances that made it practically impos-
sible to commit the robbery. For instance, that the victim was not present at the chosen
time of the event or did not have the goods intended as prey. Practicalities such as not
having the car necessary for the robbery, or as in the case of respondent 75 not having
the computer that enabled them to turn off the alarm, were included in this category:
Respondent: we wanted to rob a bank, everything was arranged, all the stuff was in place,
car, masks, guns, the grinder for the lockers, everything, but some things were not there.
30 M.R. Lindegaard and W. Bernasco

The code of the … like one of those devices that you can use to switch off the alarm, like a
bypass kind of machine. And that wasn’t there. But we anyway went there and then it was
supposed to come and then we would do it the next day. Waiting, waiting, waiting, waiting
and then we didn’t do it.
Interviewer: And then what?
Respondent: Then we didn’t do it, because it was too late, and the money was already
there … At that specific moment there would be enough.
Interviewer: So the money was gone?
Respondent: Yes
Interviewer: Ok. And why was that machine, that bypass machine not there?
Respondent: It had to get picked up in Germany and the one who was supposed to do it
got arrested. We heard that later. They were doing a large investigation in Germany on those
people. And they had … that’s why we didn’t get it.

Fear of Victims

In this category reasons related to being afraid of the victim were included. For
example, doubt about their ability to convince the victim to cooperate was given as
reason to abort. Respondent 37 described it in the following way:
Yes, let’s say, yeah, you don’t know what kind of, you don’t know what kind of man it is.
Imagine it’s suddenly a kick boxer, or something like that. That you suddenly get beaten up
there.

Co-offenders

In this category reasons related to the behavior of the co-offenders were included.
Typically the respondent described not being certain that the co-offender would do what
they had agreed on. Respondent 33 explained that he realized that his co-offenders were
too afraid to carry out the robbery and that made him decide not to do it anyway:
We didn’t even check it out. We just left. Because the other guys, they were shitting in their
pants. They were scared, you know. They got afraid. So then I already thought: I’m not
going to do anything together with you.

Doubt

This category included less specific reasons like not daring to do it anyway, not feel-
ing like doing it, and the feeling that it took too long. Respondent 83 explained how
he was looking for a suitable target but kept focusing on possible obstacles because
he actually did not dare to do it:
Then you go hunting on the street, to see if you can find a person who is easy to rip off, who
is easy to get, right. But, yeah, then you actually don’t have the guts to do it anyway. […]
Then you sit and watch. Then you think … actually I don’t dare anyway.
3 Learning About Crime Prevention from Aborted Crimes… 31

Discussion

This chapter reported on a study comparing committed and aborted robberies. Its
purpose was to illustrate what can be learned from analyzing crimes that were antic-
ipated but not carried out. Whereas our study did not have a direct crime prevention
purpose, studies of aborted crimes potentially offer insights that could be beneficial
for such purposes. By learning about why offenders abort anticipated crimes, we
might get to know what kind of prevention measures could facilitate aborting antici-
pated crimes. We attempted to answer three questions, namely the following: (1)
What distinguishes aborted from committed robberies? (2) What mechanisms
explain why some get aborted and others not? (3) What reasons do offenders pro-
vide for aborting robberies?
We based our conclusions on interviews with 104 robbery offenders who reported
on a total of 256 robbery situations, including 81 aborted robberies. Aborted robber-
ies were operationalized as robberies that the respondent had anticipated commit-
ting but decided not to carry out: the respondent had a target on his mind but decided
not to rob that particular target, leaving the potential victim unaware of the aborted
robbery.
In the statistical analysis of situational characteristics of aborted crimes we car-
ried out an intrapersonal comparison of committed and aborted robberies involving
the same offender. Where few previous studies of situational characteristics of crime
were based on intrapersonal comparisons, our study included such comparisons.
The effects we found in terms of committed and aborted robberies were therefore
unrelated to relatively stable personal characteristics, such as amount of robbery
experience. One likely hypothesis about the way personal characteristics potentially
could influence our findings is that offenders were more likely to abort from robber-
ies in the beginning of their career due to lack of experiences, whereas they later
were more likely to commit anticipated robberies because they became more pro-
fessional. Since we know that the time span between the committed and aborted
robberies reported in our study was short (respondents were asked to describe the
most recent robbery for each category), it is unlikely that the personal characteris-
tics of the respondent changed significantly between the committed and aborted
robbery described.
In relation to the first question about situational characteristics of aborted crimes
we found that two characteristics were significant for aborted crimes: (1) home rob-
beries got aborted less often than street and commercial robberies; and (2) robberies
planned for more than an hour were more likely to get aborted than robberies that
were planned for shorter periods of time.
In relation to the second question about mechanisms that could explain the deci-
sion to abort the crime, we proposed that our respondents were less likely to abort
from home robberies than other kinds of robberies for two reasons. First, victims of
home robberies were aware of being robbed as soon as contact was established with
the offender, and that made it difficult to withdraw. Second, offenders expected
home robberies to include “unexpected” occurrences—things that did not occur
32 M.R. Lindegaard and W. Bernasco

according to the plan. Unexpected events were seen as part of the package when
going for a home robbery whereas with street and commercial robberies they were
not seen as part of a script and therefore a reason to abort. Similar mechanisms
seemed to be at stake with regard to the finding that long planned robberies were
more likely to get aborted than shortly planned robberies: long planning was
described as making the offender unable to deal with unexpected events. When rob-
beries were planned for more than an hour, offenders described being unable to
define alternative targets if things did not go as expected. Robberies that were
described as a part of a “hunting process” were more likely to take place because the
script for going hunting implied being flexible about the target from the start.
In relation to the third question about subjective reasons for why respondents had
aborted an anticipated robbery, we found that several characteristics, not included as
closed questions, were mentioned as important for the decision to abort: presence of
police, moral concerns, presence of bystanders, physical circumstances, fear of vic-
tim, co-offenders, and doubt. Apart from the presence of bystanders, all these rea-
sons were new to us. In our statistical comparison of aborted and committed
robberies the presence of bystanders did not have a significant effect. When asking
our respondents if any bystanders were present while focusing on specific robberies,
they generally claimed to be so focused on getting the job done that they were
unaware of whether anyone was present and possibly watching them. When being
asked to reflect on the most important reason for the decision to abort bystanders
were anyway mentioned by some of the respondents. Despite the fact that our sam-
ple for aborted robberies was limited (we analyzed 81 aborted robberies), analysis
of the presence and behaviors of bystanders as potential prevention measures seems
worthwhile to investigate in future research. Even though the subjective reasons
provided by respondents potentially suffered from the problem of being post-
rationalizations more than factors that influenced the actual decision-making pro-
cess, the reasons might be relevant to include in future analysis of prevention
measures.
The three types of insights offered in this chapter differ in terms of epistemologi-
cal value. The statistical analysis of situational characteristics, significant for the
decision to abort, is strong because of the intrapersonal comparison: we know that
the tendencies to commit home robberies more often than street and commercial
robberies and to abort from long planned robberies compared with shortly planned
robberies are not the effect of personal preferences but of situational characteristics.
The weak aspect of this type of analysis was the fact that we included a very limited
amount of situational characteristics, leaving out factors that we were unaware of
but nevertheless might be the most relevant for the decision-making process of
offenders.
While our analysis of the mechanisms that might explain why offenders abort
from anticipated robberies was limited in scope, an analysis of open descriptions by
the offenders potentially helped explain findings of the statistical analysis while
being inductively driven: it took the experiences of the respondents as a starting
point for the focus of the analysis without uncritically reproducing their own expla-
nations as significant reason for their behavior.
3 Learning About Crime Prevention from Aborted Crimes… 33

The analysis of the subjective reasons provided by the respondents clearly suffered
from the last mentioned epistemological problem: what respondents claimed to be the
most significant reason was reproduced in the analysis. The strength of this kind of
data was that it was clearly inductively driven and thereby potentially interesting for
the development of hypotheses. The weakness was that it was highly questionable
whether the remembered reasons really guided the actual behavior in the situation
itself. The open coding of the descriptions of the decision-making process of the
aborted and committed robberies aimed at explaining the mechanisms does not suffer
from similar problems of interpretation. Apart from the obvious time-consuming
nature of such analysis, a possible issue lay in the interpretation of the person coding
the data; such interpretation problems are, however, possible to check, control, and
eventually minimize by using multiple coders (Miller, 2014).
The fact that we defined “robberies” and “aborted robberies” relatively broadly,
including street, home, and commercial robberies, and defined “anticipated rob-
bery” as everything from “discussing targets with friends at home” to “standing in
front of the shop door of a robbery planned in detail” made the amount of possible
significant differences so endless that little came out as significant in the analysis.
Recommendations for future research aiming at comparing “events” with “non-
events” therefore include defining both the event and the non-event more narrowly.
Reasons for aborting home robberies might be so different from aborting street rob-
beries that including them in the same analysis makes little sense. Similarly, reasons
for not committing a robbery that was discussed with friends at home are likely
quite different from the reasons not to do it when standing in full equipment in front
of the retail shop door. In other words, future attempts to compare committed and
aborted robberies should probably attempt to compare cases that are similar in many
more aspects than just being anticipated robberies.
The most important conclusion to draw based on our chapter is that it is possible
to gain relevant information about crimes that never took place by asking offenders
to describe such incidents. Getting to know more about aborted crimes from the
perspective of offenders is promising in terms of gaining better understanding about
the usefulness of crime prevention measures. We hope that other researchers will
make use of this approach for such purposes.

References

Bernasco, W., Ruiter, S., Bruinsma, G. J. N., Pauwels, L. F., & Weerman, F. M. (2013). Situational
causes of offending: A fixed-effects analysis of space-time budget data. Criminology, 51(4),
895–926.
Blokland, A. A. J., & Nieuwbeerta, P. (2005). The effects of life circumstances on longitudinal
trajectories of offending. Criminology, 43(4), 1203–1240.
Ganpat, S. M., van der Leun, J., & Nieuwbeerta, P. (2013). The influences of event characteristics
and actors behaviour on the outcome of violent events. British Journal of Criminology, 53(4),
685–704.
34 M.R. Lindegaard and W. Bernasco

Graham, K., Tremblay, P. F., Wells, S., Pernanen, K., Purcell, J., & Jelley, J. (2006). Harm, intent,
and the nature of aggressive behavior measuring naturally occurring aggression in barroom
settings. Assessment, 13(3), 280–296.
Hewitt, A., & Beauregard, E. (2013). Offending patterns of serial sex offenders: Escalation, de-
escalation, and consistency of sexually intrusive and violent behaviours. Journal of Investigative
Psychology and Offender Profiling, 11(1), 57–80.
Jacques, S. (2010). The necessary conditions for retaliation: Toward a theory of non‐violent and
violent forms in drug markets. Justice Quarterly, 27(2), 186–205.
Leclerc, B., Lussier, P., & Deslauriers-Varin, N. (2014). Offending patterns over time: An exami-
nation of specialization, escalation and de-escalation in the commission of sexual offenses. In
A. Blokland & P. Lussier (Eds.), Sex offenders: A criminal career approach. Oxford, UK:
Wiley.
Levine, M., Taylor, P. J., & Best, R. (2011). Third parties, violence, and conflict resolution: The
role of group size and collective action in the microregulation of violence. Psychological
Science, 22(3), 406–412.
Lindegaard, M. R., Bernasco, W., & Jacques, S. (2014). Consequences of expected and observed
victim resistance for offender violence during robbery events. Journal of Research in Crime
and Delinquency, 52(1), 32–61.
Lussier, P., Leclerc, B., Healey, J., & Proulx, J. (2008). Generality of deviance and predation:
Crime switching and specialization patterns in persistent sexual offenders. In M. Delisi &
P. Conis (Eds.), Violent offenders: Theory, public policy and practice (pp. 97–140). Boston,
MA: Jones and Bartlett Publishers.
Maruna, S. (2001). Making good. Washington, DC: American Psychological Association.
Miller, J. (2014). Epistemology and rigor in criminological research: An explanatory journey. In
S. Ruiter, W. Bernasco, & W. Huisman (Eds.), Eenvoud en Verscheidenheid: Liber amicorum
voor Henk Elffers (pp. 491–502). Amsterdam: NSCR and VU University.
Phillips, S. (2003). The social structure of vengeance: A test of black’s model. Criminology, 41(3),
673–708.
Tonry, M. (Ed.). (2014). Why crime rates fall and why they don’t (Crime and justice, Vol. 43).
Chicago, MI: University of Chicago Press.
Vargas, R. (2014). Criminal group embeddedness and the adverse effects of arresting a gang’s
leader: A comparative case study. Criminology, 52(2), 143–168.
Chapter 4
Learning from Offenders: Some Iatrogenic
Effect of Crime Prevention Measures

Martin Gill

This chapter begins by highlighting some limitations of using data derived from
offenders. It then moves on to introduce, briefly, some of the limits of crime preven-
tion measures, and progresses to the main part of the chapter focusing on just one
area where offenders’ insights can be important, that is, in understanding the inher-
ent weaknesses of crime prevention approaches. Given that crime prevention is
ubiquitous, offenders have to make sure that it does not work in a way that will
impede them. As this is a large topic and space is limited, I want to look at this in a
very defined area.
First, I want to look beyond traditional crime prevention measures and instead
look at the role of processes, more broadly policy and procedures which are com-
mon in a business environment (another area where crime prevention studies are
sparse), and show how these are crucial for stopping crime when done well and in
facilitating crime when done badly. Often crime prevention is built into processes
designed for other purposes, for example a process for claiming and checking
expenses may include elements designed to reduce dishonesty, by for example
requiring receipts.
Second, I want to explore the ways in which crime prevention measures may
facilitate crime. Recently crime prevention has been given a boost in that there is
strong evidence that a major contributory factor to the reduction in crime witnessed
across much of the industrialised world has been the implementation of good crime
prevention (Farrell, Tilley, Tseloni, & Mailley, 2011; van Dijk, Tseloni, & Farrell,
2012). Yet, as I hope to show, according to offenders, many crime prevention

The author would like to thank Paul Ekblom, Charlotte Howell, Sarah Webb and two anonymous
referees for advice on this chapter.
M. Gill (*)
Perpetuity Research & Consultancy International (PRCI) Ltd,
11a High Street, Tunbridge Wells, Kent TN1 1UL, UK
e-mail: m.gill@perpetuityresearch.com

© Springer International Publishing Switzerland 2017 35


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_4
36 M. Gill

measures are often easily overcome. In medical science ‘iatrogenesis’ is used to


describe an adverse effect of medical treatment. This has long been recognised as a
widespread cause of harm including death (see, Illich, 1974; Starfield, 2000). In
practice the iatrogenic effect is wide ranging and includes quite complex relation-
ships between the intervention and the outcome. For example, some measures can
cause side effects but these can be preferable to not having the treatment at all; brain
surgery for example may cure or paralyse and is a calculated risk rather than an
unintended consequence. Slightly differently, iatrogenic effects also include error
and negligence—resulting from the poor decisions and actions of the intervener. In
this chapter I would like to explore the iatrogenic effect of a type of crime preven-
tion from the perspective of the offender. It is not possible in the space available to
explore all the possible effects—although that would be a laudable aim—rather to
focus on just some aspects in the context of a specific type of measure, process.

Some Limitations to Learning from Offenders

There are methodological problems in researching offenders. Put simply, how can
we be sure that what offenders tell us about their criminality is the truth? After all
why would they want to help inform crime prevention efforts not least if they plan
to offend again? Indeed, looking at things from this angle there would appear to be
a reason to lie to make it more difficult for preventers in the future and easier for
themselves rather than the other way around. Moreover, many offenders are by
(legal) definition dishonest, certainly of instrumental crimes, and that further fuels
concerns that they may not tell the truth. In fact much has been written about the
difficulties inherent in the process of collecting, interpreting and using what offend-
ers have to say (for an excellent set of discussions, see Bernasco, 2010).
Moreover, we don’t know whether there are some offences where offenders’
accounts could be deemed more reliable. On the face of it, reflections on circum-
stances where some degree of rationality is evident, that is where to some extent
offenders weigh up the pros and cons of whether to commit an offence and different
ways of conducting it, have more to offer crime preventers. After all if they weigh up
the pros and cons they may perhaps be stopped or deflected in some way. But perhaps
those who are remorseful; who are aiming to desist from committing offences again;
and who see merit in others learning from their experience would be good targets for
interviews if they can be identified? This again warrants more research.
While the limitations of deriving data from offenders need to be acknowledged,
there are some examples of offenders’ insights being used to inform and guide prac-
tice (see Decker, 2005; Gill, 2005; Jacques & Reynald, 2012, Levi, 1998; and see
Gill, 2011a, 2011b). Moreover, there are many studies of how offenders behave that
provide a foundation for assessing what has or can work although much of the
research is on a fairly narrow range of offences and especially burglary (Cromwell &
Olson, 2004; Hearnden & Magill, 2004; Nee & Meenaghan, 2006; Wright & Decker,
1994), robbery (Gill, 2000; Matthews, 2002; Wright & Decker, 1997) and shop theft
4 Learning from Offenders: Some Iatrogenic Effect of Crime Prevention Measures 37

(Gill, Bilby, & Turbin, 1999; Hayes, 2006; Tilley, 2010). Similarly research has been
published on offenders’ perspectives of crime prevention measures although this is
limited to a fairly narrow range of measures and in particular CCTV (for a recent
review see Taylor & Gill, 2014) and intruder alarms (see Bennett & Wright, 1984).
The key premise of situational crime prevention1 is to make crime less likely and
this includes making the target less attractive or more robust against crime attempts
(see Cornish & Clarke, 2005). In effect providing the guidance for making a target less
attractive for offenders or helping to make the guardian more capable as routine activ-
ity theory would have it (Cohen & Felson, 1979). The problem for crime prevention is
that often the target is attractive to offenders even though attempts have been made to
harden it. Offenders’ accounts provide an opportunity to learn more about what are
effectively crime prevention failures, be that of concept (the wrong measure was intro-
duced) or implementation (the measure did not work properly) (Tilley, 2009).
Offenders can tell us precisely how they were able to overcome different mitiga-
tion measures, and specifically how some strategies designed to prevent offending
actually worked in their favour, that is, measures designed to be crime preventers
that actually become crime facilitators, what is in practice the iatrogenic effect. This
is an under-researched area.

Learning from Offenders

I would like to begin with just a few short vignettes about lessons I have learnt from
research work with offenders which help illustrate the iatrogenic effect of crime
prevention measures, at least from the perspective of the offender.
On one occasion I was with a shop thief as he recreated his offences and high-
lighted some of the strengths and weaknesses of different crime prevention mea-
sures in stores (see Gill, 2007). The store manager had advised me that the addition
of a ‘very tall big fat man’ as a security guard at the back exit was a significant
deterrent since he would ‘scare’ many thieves away. The thief I was with that day
found many easy opportunities in store. As he approached the exit he saw the
guard and reacted immediately, ‘what about that, what a bit of luck, a big fat git
who I can outrun’. As he later explained, he was not looking for a fight or for that
matter any form of confrontation and that he believed that the guard could not run
very fast made it easier to escape quickly and that was his priority. In short, the
iatrogenic effect was caused by the misperception that a big guard posed a deter-
rent effect; in this case it didn’t and that the offender viewed the guard as an
advantage was an unintended consequence (albeit that many other offenders may
well have been deterred).
On another occasion I interviewed 13 thieves about their attitudes to goods that
had a label attached stating ‘only sold in (store name)’. The crime prevention theory

1
The techniques can be found here: http://www.popcenter.org/search/?cx=0168173356798859758
49%3Agiidughzfro&q=25+Technique
38 M. Gill

for the tag was that it would make goods less attractive to thieves because they would
have to spoil the packaging to remove the label; anyone seeing the label/tag would be
more likely to recognise that the goods were stolen and that would render them
harder to sell; and if the police found them in the possession of thieves they would
more likely be under suspicion (and the tags also enable the police to return the
goods because the owners would be easily identified). In fact none of the thieves said
that the label/tag would put them off, either because they stole for self-use or because
they believed that their buyers would not be concerned they were stolen. But three
said that the label would be an advantage. Their argument was that buyers knew that
the goods they sold were stolen, not least because they were cheap, that was not an
issue, but they did worry that they may be counterfeit. The label was good evidence
that the goods were not fake (they were originally sold in a high street store) and that
made them easier to sell (and sometimes for more money). In this case iatrogenesis
was caused by the unexamined consequences of the design of a security measure.
In fact all crime prevention measures have the potential for an iatrogenic effect. A
high fence can make it more difficult for intruders to gain access, but if they do get
past it they can be hidden from view while they execute the next stage of their offence;
CCTV in organisational settings can serve security purposes but cameras can also lull
staff and others into a false sense of safety, and they stop looking for suspicious activ-
ity assuming that those who watch cameras will manage it (Beck and Willis, 1995;
Taylor & Gill, 2014); a sign stating that money on the premises is kept in a safe is an
advertisement that it is not easily accessible but it does indicate that there is money on
the premises; if an offender knows how to access the contents of a safe this is good
information (Gill, 2000). What is key in each example is to understand the balance
between the success of the measure, say in reducing crime or the fear of it against its
iatrogenic effects. There are many more examples. One good way of learning how
offenders see things differently to non-offenders (on this issue see the discussion on
‘affordance’ by Pease, 2006) is by collecting insights from offenders themselves.
In this chapter I would like discuss ways in which insights from offenders can
help understand the iatrogenic effects of processes (and in so doing draw parallels
with what Cornish & Clarke (2005) calls ‘rule setting’).

Offender Insights into Process Failures

Ben2 had a gambling habit which was out of control, and it came to a point where,
as he said, ‘I owed a lot and nothing was enough’. He worked in hospitality, and
there were at least two characteristics of his job that facilitated crime; he had a rea-
sonably autonomous role and he had access to cash, indeed he was responsible for
banking the money collected at the end of a shift. He would be expected to count it
and within a few days pay it into a bank. As he reflected:

2
Ben, Adam and Carl (all pseudonyms) were interviewed in prison on 15 January 2014, and Cyril
on the 21 September 2012.
4 Learning from Offenders: Some Iatrogenic Effect of Crime Prevention Measures 39

I thought, look I can borrow this for a few days and try my luck and the first time it worked.
I borrowed a thousand and I doubled it (via gambling) and I was able to pay the money in
and I still had some to myself … I saw it as borrowing the cash for now and intending to
pay it into the bank later but it got out of control … I did take the money but I was planning
to pay it back.

Whether this view was genuine or self-deception is unclear, but what was evi-
dent was that he had both the opportunity of access to money and a genuine belief
that he was unlikely to get caught at least in the short run. He learnt from
experience:
The first time I took the money I did not pay it back in until 3 weeks later and they didn’t
check, and then I realised that actually they don’t check immediately. You can always say
you forgot, on an occasional basis you can.

His perception of the opportunity emerged gradually. His employer’s oversight


of the system was weak. Had it identified non-payers and chased them immedi-
ately it is possible that Ben’s actions would have been thwarted before they
started:
after about a year they saw I was taking the money. I was gambling hoping to pay it back
but not doing so. For a year and they did not see that I was doing that.

A problem (gambling), an opportunity (access to cash), a skill set to obtain it (in


this case obtained as part of normal business duties) and a belief in avoiding capture
(in the absence of any credible accounting system and supervisor check) provided
what Levi (2008) terms the ‘slippery slope’. Clearly the employer fell foul of a
number of golden situational crime prevention rules, and for that matter good risk
management and personnel management generally. The iatrogenic effect was
reflected in at least inadequate accountable systems and poor supervisory oversight
(it seemingly took a year before they noticed the problem). It was fuelled by the lack
of a clear mechanism for Ben to get help with his problems all crystallised in losses
for the firm, and of course for the public purse in dealing with the aftermath, and
there were enormous social and personal costs for Ben:
My wife was really shocked and my child who is 7 does not know. I will tell him one day
but not yet, I just speak to him on the phone occasionally which upsets me of course.

Ben discovered iatrogenic effects in processes as he went along. The following


provides another example.
Adam was 39 at the time of the interview and had committed offences while run-
ning his own business. He was serving his first prison sentence, for what was his
first conviction, 3 years and 9 months for conspiracy to commit fraud and money
laundering. Adam was already involved in a scam that enabled him to offer clients
of his mail marketing lower prices than competitors when he discovered another
lucrative criminal opportunity that would eventually lead to his downfall:
I would take letters, hundreds of thousands say or millions and be responsible for posting
them. Now in the beginning (the mail delivery company) did not count the mail so I found
out I could say I had 1 million letters but there were really 2 million. Here was what I saw
as a business opportunity, I took 2 million letters from the client who paid for 2 million
40 M. Gill

letters, and I told (the mail delivery company) there were only 1 million and I only paid for
1 million. So as long as the (the mail delivery company) did not check I was ok.

For the mail delivery company the process of counting the number of mail items
was woefully inadequate. On only one occasion did they confront him with an
‘error’ and he immediately apologised and passed it off as a mistake. The iatrogenic
effects of this process facilitated Adam’s fraud, as he said, ‘they lost millions of
pounds for work they had not processed properly’. In the next example the victim
was complicit in the offence.
Carl was an independent financial advisor and mortgage broker, in his late 40s,
at the time of the interview. He too had not been to prison before (he had one previ-
ous conviction as a teenager) but was convicted of 15 counts of conspiracy to
defraud for which he was sentenced to 3 years and 4 months. Put simply Carl
obtained mortgages for people by fabricating supporting evidence; this might
include inflating people’s incomes and forging references:
I would say it would be naïve that I was not aware of breaking the law but it was incestuous
and it was common practice with bankers as they are happy to lend if they have a paper trail,
it made the whole process feel normal. So I knew I was bending the rules and felt justified
because everyone else was doing it. To me it was just business, just a way of earning a living
and that is what I did.

He claimed many offences dated back to days when such practices were more
commonplace:
My bosses showed me how to do it. I worked for a company that was owned by a bank and
my bosses showed me how to do false references … Say you have a client you have helped
and instead of charging them a fee I would say will you do a reference? And then the bank
wrote to them and they provided a reference. In (the) old days this was the way it worked
and a lot of the times I would do this.

Carl claimed that the victims were in fact what Ekblom (2011) would call ‘crime
promoters’ in that they increase the chances of an offence taking place:
I did take advantage of the lender but this did not concern me, because business develop-
ment managers from each bank or building society would come and say we have a load of
money and they would say do a.b.c and this would be about them circumventing their sys-
tems, with our help, to meet their targets. So when convenient they turned the tap on to
make things easy for you, and then the tap went off in harder times. They were as complicit
as the mortgage brokers they did business with. I never once felt guilty about them. And the
reason is they were more than complicit they were aiding and abetting and actively promot-
ing practices. As long as they had the hurdles jumped they did not care, they did not check
the validity of what was said they just needed me to say I had done what was necessary.

If the interviewee is right then at least some employees in some financial institu-
tions it seems flouted the rules governing how mortgages should be granted and
proactively engaged in illegality. The weak oversight and/or the poor crafting of the
rules were responsible in part at least for an iatrogenic effect that resulted in a crime
taking place. It is perfectly possible of course that the financial institutions weighed
up the risks and decided that there was a greater benefit in more profits in facilitating
these offences (especially if those who obtained mortgages this way were good pay-
ers). A parallel here would be retailers who display their goods openly to promote
4 Learning from Offenders: Some Iatrogenic Effect of Crime Prevention Measures 41

sales and in so doing make them attractive to thieves; they seemingly calculate that
there are benefits in losing some goods to theft by the increased sales that result
from open display (for discussion of retail security see Bamfield, 2012; Beck, 2009).
More of this later.
The final interviewee was Cyril who was in his mid 60s when he was interviewed
and was serving his first prison sentence, 7½ years for multiple (‘hundreds’ of)
charges of fraud and money laundering to which he pleaded guilty. Cyril was a self-
employed accountant and claimed to have never committed any prior offences. He
believes that he had a midlife crisis fuelled in part by a difficult marriage, but his
offences were facilitated by a sudden and surprising opportunity.
As an accountant Cyril along with a local solicitor he knew well was given a
power of attorney over the estate of a wealthy individual. Together they man-
aged her finances, so when bills arrived Cyril and the solicitor would settle them
ensuring that they jointly approved all payments from the account. Then one
day an invoice arrived from the nursing home which needed to be settled straight
away. The solicitor was away and so Cyril alone authorised payment and it was
paid. Suddenly he realised that he could make withdrawals without a
cosignatory:
I was in my fifties, and had never committed offences. Being made power of attorney was
not a big deal, I had done it before. I was working on this case for 2 years and there was no
problem. When the … solicitor was on holiday and I did it on my own, and no one checked,
it just made me think. It was all quite sudden. I would take £30,000 a month after that and
no one checked.

He benefited from weak oversight by the solicitor:


I spoke to the solicitor once a month but he did not worry about it and had no grounds to
suspect anything. I knew the solicitor for 15 years … he got in trouble but he was not
involved or aware of anything I was doing … The solicitor was really just negligent.

And the lack of oversight by financial institutions helped. As Cyril noted the
‘lack of checks or notice of anything suspicious by financial institutions was a help’,
and that ‘I could have suffered from an audit at any of the financial institutions’,
both accounts he drew money from as well as those he paid the money into.
He was well known at the financial institution where the account was held
and this is probably why his actions did not raise suspicions there. The solicitor
did not check financial updates he was provided with most probably because he
knew Cyril and trusted him; there were missed opportunities to identify Cyril’s
scam. He exploited several factors: slack procedures at that time in the financial
sector in withdrawing funds; his reputation as a local accountant who could be
trusted; skills to make the accounts look ‘normal’ despite his illegal withdraw-
als; and slack procedure on the part of the solicitor in offering effective due
diligence.
Cyril noted that ‘it was so easy, it was simple’. Although he knew he would be
caught when ‘the old lady died’, at that point he went on the run with £300,000 in
his account and lived the ‘high life’ until he was caught just as his money was run-
ning out.
42 M. Gill

Discussion: What Offenders Can Tell Us About Crime


Prevention

Offenders are well placed to tell us about the strengths and weaknesses of different
crime prevention approaches; it is perhaps surprising that we have not sought their
views more often on this issue not least to better understand how different types of
measures can be improved. In this chapter I have sought to show how offenders’
insights can help explain how a type of measure, business processes, can, when
poorly conceived and/or managed, work in offenders’ interests. I have borrowed the
medical term ‘iatrogenic effect’ which refers to the unintentional harmful effects of
medical intervention and applied it to crime prevention measures to highlight ways
in which crime prevention interventions designed for bona fide reasons unintention-
ally fail with harmful effects in that they work in the interests of offenders. I pro-
vided four different examples of how processes designed (at least in part) to prevent
crime actually facilitated it. The causes of the iatrogenic effect were seemingly a
combination of error, negligence and unintended side effects in the ways processes
were conceived, implemented and monitored.
Each of the examples illustrates points in more than one way, but in summary
processes were poorly conceived and therefore they were flouted even by those
who had created them (and therefore Carl was able to generate income for his
employer because breaching procedures often worked in the employer’s inter-
ests); they were poorly implemented so that they did not guide behaviour in the
way that was intended (so Adam was able to include more items than he was
paying for and was illegally defrauding his client of income); they were poorly
monitored (and so Ben was able to avoid paying in cash and as a consequence
illegally deprived his employer of income it was due); and because of poor pro-
fessional practice (and this enabled Cyril to capitalize on the solicitor failing to
follow proper procedure and benefited from lax practices in the financial institu-
tions he dealt with).
The study of crime prevention has started to draw upon knowledge within a
much broader range of disciplines influenced by the emergence of ‘crime science’
(Ross, 2013 and see LeClerc & Wortley, 2013). Yet in some areas, and crime in
organisations is a case in point, there has been insufficient attention paid to the role
of process in preventing crime. Indeed, there is growing awareness that good crime
prevention in business (at least) is in part about building security into processes
designed for other purposes; that way you can engage users and ensure that crime
prevention is doable rather than getting in the way of doing business (when done
well). The other way around is important too. For example, a study of security in the
supply chain (Peleg-Gillai, Bhat, & Sept, 2006) found many broader additional ben-
efits to good security including the following:
• Improved product safety
• Improved inventory management
• Improved supply chain visibility
4 Learning from Offenders: Some Iatrogenic Effect of Crime Prevention Measures 43

• Improved product handling


• Process improvements
• More efficient customs clearance process
• Speed improvements
• A reduction in problem identification time and higher customer satisfaction
This chapter has attempted to highlight research in one small area and has raised
more questions than it has answered. Certainly, more thought is needed about the
scale of the iatrogenic effect when applied to crime prevention. For example, does
the fence reduce crime overall by impeding enough motivated and resourceful
offenders that are unable to get over or through it, or increase it by providing cover
for illicit activity? Have the processes reduced loss or harm overall, or increased it?
What improvements can be made to achieve a better overall impact? Research is
also needed to catalogue the different unintended harmful consequences of different
types of crime prevention measures. Earlier I argued that each crime prevention
measure is likely to have harmful consequences but more research is needed on the
ways these are generated and then exploited for measures operating in different
contexts. Just as this chapter has considered the iatrogenic effects from offenders, so
there may be lessons from a focus on crime preventers, guardians, victims and so on
as is being explored with those assessing another way of looking at criminality via
script analysis (see LeClerc & Wortley, 2013). Most importantly, more research is
needed on the nature of the iatrogenic effect itself. As noted, that it exists does not
mean that a measure is redundant or undesirable; an iatrogenic effect is not neces-
sarily an adverse effect. It may be, as is the case in medicine, that the effects are
sometimes not harmful and/or that the benefits of the measure outweigh any nega-
tive impacts, and we need to better understand what the consequences would be
without the measure(s) in place.
This chapter has shown that offenders commit crime (in part) because:
• Measures designed to stop crime can sometimes work to offenders’ advantage.
• Measures that are often not recognised as crime prevention measures (but are
crucial to the crime prevention process) can fail and make crime easier and
possible.
• Weaknesses in crime prevention measures don’t necessarily mean that they are
undesirable; this will only sometimes be the case.
• Offenders are often well placed through, for example, their workplace routines to
exploit opportunities. That people needed to be given autonomy to do their jobs
and in so doing this creates weaknesses that offenders can exploit illustrates the
quite complex and varied nature of iatrogenic effects.
Evidence from the limited insights permissible in this chapter suggests that
errors in assessing risks, in matching the mitigation measures to vulnerabilities, and
in implementing the wrong measures or the right ones but doing it badly all play a
part. Those who design and manage crime prevention measures can be negligent
and in a variety of ways such as making bad decisions, not understanding side
effects, or introducing unnecessary measures via ignorance, carelessness and error
44 M. Gill

(see also Ekblom, 2011; Pawson & Tilley, 1997). It is a stark lesson to realise that
measures designed to prevent crime may actually facilitate it. Little is known about
the extent to which this happens or the ways in which it does, or indeed what the
range of harmful effects are although offenders’ insights afford the opportunity to
learn more. While it has long been known that crime prevention measures fail, a
better understanding of the precise reasons why this is the case are crucial, and
offenders have the potential to generate important insights.

References

Bamfield, J. (2012). Shopping and crime. Basingstoke: Palgrave.


Beck, A. (2009). New loss prevention. Basingstoke: Palgrave.
Beck, A., & Willis, A. (1995). Managing the risk to safe shopping. Leicester: Perpetuity Press.
Bennett, T., & Wright, R. (1984). Burglars on burglary. Aldershot: Gower.
Bernasco, W. (Ed.). (2010). Offenders on offending. Cullompton: Willan.
Cohen, L., & Felson, M. (1979). Social change and crime rate trends: A routine activity approach.
American Sociological Review, 44(4), 588–608.
Cornish, D., & Clarke, R. (2005). Opportunities, precipitators and criminal decisions’, in Theory
for Practice. In Situational Crime Prevention, Crime Prevention Studies (Vol. 16). Monsey,
NY: Criminal Justice Press.
Cromwell, P., & Olson, J. (2004). Breaking and entering: Burglars on burglary. Belmont, CA:
Wadsworth.
Decker, S. H. (2005). Using offender interviews to inform police problem solving. Washington,
DC: US Department of Justice.
Ekblom, P. (2011). Crime prevention, security and community safety using the 5Is framework.
Basingstoke: Palgrave.
Farrell, G., Tilley, N., Tseloni, A., & Mailley, J. (2011). The crime drop and the security hypothesis.
Journal of Research in Crime and Delinquency, 48, 147–175.
Gill, M. L. (2000). Commercial robbery: Offenders’ perspectives on security and crime prevention.
London: Blackstone Press.
Gill, M. (2005). Learning from fraudsters. Protiviti. www.perpetuitygroup.com/prci/publications.html
Gill, M. (2007). Learning from fraudsters: Reinforcing the message. www.perpetuitygroup.com/
prci/publications.html
Gill, M. (2011a). Learning from offenders’ accounts of their offending. Prison Service Journal,
194, 27–32.
Gill, M. (2011b). Should we expect more frauds in a recession? Views from both fraud managers
and fraudsters. International Journal of Law, Crime and Justice, 39(3), 204–214.
Gill, M., Bilby, C., & Turbin, V. (1999). Retail security: Understanding what deters shop thieves.
Journal of Security Administration, 22(1), 29–40.
Hayes, R. (2006). Shoptheft. In M. Gill (Ed.), The handbook of security (pp. 408–422). London:
Palgrave.
Hearnden, I., & Magill, C. (2004). Decision-making by burglars: Offenders’ Perspectives. Findings
no 249. London: Home Office.
Illich, I. (1974). Medical nemesis: The expropriation of health. London: Calder & Boyars.
Jacques, S., & Reynald, D. M. (2012). The offender’s perspective on prevention: Guarding against
victimization and law enforcement. Journal of Research in Crime and Delinquency, 49,
269–294.
LeClerc, B., & Wortley, R. (Eds.). (2013). Cognition and crime. London: Routledge.
4 Learning from Offenders: Some Iatrogenic Effect of Crime Prevention Measures 45

Levi, M. (1998). The craft of the long-firm fraudster: Criminal skills and commercial responses. In
M. Gill (Ed.), Crime at work (Vol. 2, pp. 155–168). Leicester: Perpetuity Press.
Levi, M. (2008). The phantom capitalists: The organisation and control of long-firm fraud.
Aldershot: Ashgate.
Matthews, R. (2002). Armed robbery. Collumpton: Willan.
Nee, C., & Meenaghan, A. (2006). Expert decision making in burglars. British Journal of
Criminology, 46, 935–949.
Pawson, R., & Tilley, N. (1997). Realistic evaluation. London: Sage.
Pease, K. (2006). No through road: Closing pathways to crime. In K. Moss & M. Stephens (Eds.),
Crime reduction and the law (pp. 50–66). Abingdon: Routledge.
Peleg-Gillai, B., Bhat, G., & Sept, L. (2006). Innovators in supply chain security: Better security
drives business value. Stanford: The Manufacturing Institute, Stanford University.
Ross, N. (2013). Crime: How to solve it, and why so much of what we’re told is wrong. London:
Biteback.
Starfield, B. (2000). Is US health really the best in the world? Journal of the American Medical
Association, 284(4), 483–485.
Taylor, E., & Gill, M. (2014). CCTV: Reflections on its use, abuse and effectiveness. In M. Gill
(Ed.), The handbook of security (2nd ed., pp. 705–726). London: Palgrave.
Tilley, N. (2009). Crime prevention. Collumpton: Willan.
Tilley, N. (2010). Shoplifting. In F. Brookman, M. Maguire, H. Pierpoint, & T. Bennett (Eds.),
Handbook of crime (pp. 48–67). Uffculme, Devon: Willan.
van Dijk, J., Tseloni, A., & Farrell, G. (Eds.). (2012). Closing the doors: New perspectives on the
international crime falls. Basingstoke: Palgrave.
Wright, R., & Decker, S. H. (1994). Burglars on the Job: Streetlife and residential break-ins.
Boston: Northeastern University Press.
Wright, R., & Decker, S. H. (1997). Armed robbers in action: Stickups and street culture. Boston:
Northeastern University Press.
Part II
Crime Script Analysis
Chapter 5
Script Analysis and the SCRIPT Acronym

Benoit Leclerc

Introduction

Crime script analysis is more popular than ever in criminology. This popularity did
not occur overnight. Derek Cornish demonstrated the relevance of adapting the con-
cept of the script to the study of crime-commission processes in 1994; yet initially,
crime scripts attracted little attention from criminologists. Since 2006, however,
there has been an increased interest in crime scripts worldwide. Indeed, criminolo-
gists are now attentive to the benefits of crime scripts. This chapter briefly covers
the concept of scripts, recent research on crime scripts, benefits of offender self-
reported data for crafting crime scripts, and the strengths of crime script analysis.
The main objective of this chapter is to promote the practice of using script analysis
in criminology.

The Concept of Scripts and the Contribution of Crime Scripts

Peoples’ brains are inhabited by different forms of schemas (Augoustinos, Walker, &
Donaghue, 2006). A schema is a cognitive structure that contains assumptions and
expectations about the social world based on past behavior and experience. The main
objective of schemas is to help people navigate in the social world. For instance,
every person has accumulated knowledge about himself/herself, other people, and
what sequence of actions to follow in order to achieve desired goals. People also have
schemas to provide them with guidance on how to interact with other people.
Accordingly different forms of schemas exist (e.g., self-schemas, person schemas,

B. Leclerc (*)
School of Criminology and Criminal Justice, Griffith University,
Mt Gravatt Campus, Brisbane, QLD 4111, Australia
e-mail: b.leclerc@griffith.edu.au

© Springer International Publishing Switzerland 2017 49


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_5
50 B. Leclerc

relational schemas, event schemas) (Augoustinos et al., 2006; Baldwin, 1992; Fiske
& Taylor, 1991). For instance, self-schemas refer to the knowledge structure people
have of themselves, which is linked to a person’s traits and past behaviors and thus
provides this person with a sense of identity. A person schema is a knowledge struc-
ture that contains personality prototypes of other people the person knows either
personally or by reputation. This structure helps people categorize others according
to their dominant personality traits and then to interact with them according to their
own personality. Relational schemas—called interpersonal scripts—are cognitive
structures that assist people in their interrelations with others (Baldwin, 1992; 1995).
Specifically, interpersonal scripts are a stereotyped relational pattern found in a par-
ticular context—what behaviors tend to be followed by what responses—and are
used to interpret social situations and the behavior of others. Finally, an event schema,
commonly referred as a script, is a knowledge structure that organizes the sequence
of actions to adopt in a particular context. For a more detailed coverage on the origins
of scripts, the reader is invited to read Cornish’s piece published in Crime Prevention
Studies in 1994.
Schank and Abelson (1977) illustrated the concept of scripts by using the now
classic example of eating in a restaurant, that is, the sequence of actions that must
be taken by the customer of a restaurant. The customer must enter the restaurant;
wait to be seated; get the menu; order; eat; get the cheque; pay; and exit. Applied to
crime, the script concept makes the crime-commission process much easier to iden-
tify and understand. Cornish (1994) initially borrowed the script concept in order to
provide a framework to systematically investigate and identify all of the stages of
the crime-commission process of a specific crime, the decisions and actions that
must be taken at each stage, and the resources required for successful completion of
the crime. This template is highly relevant for crime prevention as it assists crimi-
nologists in identifying additional intervention points for prevention besides those
just associated with the “doing” of crimes.
In criminology, there are three innovations that should be noted in the way crime
scripts have been recently used. First, scripts have now been applied to identify the
process of crimes that may have been perceived in the past as irrational and thus
unpreventable through the means of situational crime prevention. This body of
work reinforces the notion that offenders are making decisions before, during, and
after committing their crime, even if these decisions can be made in a few seconds.
For example, interpersonal crimes such as child sexual abuse and sexual offenses
committed by stranger serial offenders have been reconstructed through scripting,
which highlighted the rational nature of these crimes and, as a result, the potential
for situational prevention (Beauregard, Proulx, Rossmo, Leclerc, & Allaire, 2007;
Leclerc, Smallbone, & Wortley, 2013; Leclerc, Wortley, & Smallbone, 2011).
Equally important, the utility of scripts for understanding “organized” crimes has
also been demonstrated by several scholars showing that scripts can capture the
crime-commission process of a criminal organization’s activities (e.g., Chiu,
Leclerc, & Townsley, 2011; Hiropoulos, Freilich, Chermak, & Newman, 2013;
Savona, 2010) and the specific role that certain actors can play within these organi-
zations (Morselli & Roy, 2008).
5 Script Analysis and the SCRIPT Acronym 51

Second, the exercise of thinking broadly, yet specifically, about situational


prevention through mapping potential measures onto crime scripts has been com-
pleted by a number of scholars in the past few years (e.g., Brayley, Cockbain, &
Laycock, 2011; Chiu & Leclerc, 2016; Chiu et al., 2011; Clarke & Newman,
2006; Hiropoulos et al., 2013; Leclerc et al., 2011; Savona, Giommoni, &
Mancuso, 2013; Thompson & Chainey, 2011). This exercise is consistent with
Cornish’s ultimate goal to enhance the effective use of situational crime preven-
tion initiatives. In this sense, this recent body of work represents a vital contribu-
tion to the field of crime prevention.
Third, to further enlarge the reach of situational prevention, the notion of crime
scripts has been thought of from the perspective of other actors—victims and poten-
tial guardians (Leclerc, 2013a; 2013b). A study completed by Smith (2010; 2016)
on taxi drivers who were involved in a robbery or attempted robbery revealed how
potential victims could protect themselves through situational prevention. Recently,
Leclerc and Reynald (2015) designed a theoretical script model of the intervention
process of capable guardians in public places and showed how situational preven-
tion measures could be applied to facilitate the intervention process of guardians.
Ekblom (2012) also looked at the interaction between offenders and preventers
through the lens of a script approach by using the concept of script clashes.

Offender Self-Report and Crime Specificity

From which source can we extract the most valuable information for crime scripts?
Bullock, Clarke, and Tilley (2010) suggested a number of potential data sources that
are useful for crime script analysis, including materials obtained from police investi-
gations (see Savona et al., 2013); offender self-reports (see Leclerc et al., 2011;
2013); victims (Smith, 2010; 2016); and court transcripts (see Chiu et al., 2011; Chiu
& Leclerc, 2016). Although each source represents a way to collect data on crime
scripts, and a combination of them may present the ideal situation for a researcher,
the most informative method is to talk to offenders themselves. The main reason for
this is that offenders are the ones performing the actions that we are seeking to ana-
lyze. Therefore, there is no better source for this information than offenders them-
selves. This taps into the rationale underlying and driving offender-based research
(Jacques & Bonobo, 2016). Only offenders can provide a detailed account of their
actions. For instance, there are stages of the crime-commission process for which
sources other than offender self-reports would not be useful such as when the
offender contemplates, plans, selects, and approaches the target. If data are missing
for these stages, it is harder to reconstruct the script. This point is vital for situational
crime prevention practice because the script builds on the offender’s perspective in
the design of prevention measures. Of course, this point can also be made for victims
(in the case of victims’ scripts) and guardians (in the case of guardians’ scripts) as
well (Leclerc, 2013a).
52 B. Leclerc

When interviewing offenders we can investigate the decisions and motives


underlying their actions at each stage of the script, which can be difficult to achieve
using other data sources, even police data. Recent offender-based research (or eth-
nographic research), using a script approach (e.g., Copes, Hochstetler, &
Cherbonneau, 2012; Jacques & Bernasco, 2013), has provided additional insights
about how offenders think, and why they think the way they do in the process of
committing crimes. We can also capture how offenders interact with the environ-
ment throughout the crime-commission process and if there are any particular fea-
tures of the situations with which they interact that can either prevent them from
completing their script or facilitate offending. Consistent with this point, unsuccess-
ful scripts, that is, attempted offenses, can be investigated as well in order to reveal
what stopped offenders from offending (on the idea of aborted crimes, see
Lindegaard & Bernasco, 2016; Lindegaard, Bernasco, Jacques, & Zevenbergen,
2013). Critically, and along these lines, it is also possible to investigate what offend-
ers think of the effectiveness of particular control measures.
Offender self-reports are not without limitations of course but the benefits clearly
outweigh the limitations in that offenders represent a unique data source for crime
research (Bernasco, 2010). A number of criminologists criticize this method, assum-
ing that offenders will not report reliable or valid information on their crimes.
Limitations that can affect the quality of self-reported data include poor memory
recall, cognitive distortions to minimize or deny particular aspects of offending, the
inability to understand the questions asked during the interview, lack of willingness
to participate in the research, lack of time given to complete the interview, the desire
to present a favorable image of oneself, and an inadequate context in which the
interview is conducted. These limitations will affect the quality of self-reported
data. However, there are also techniques used by researchers to limit the negative
impact of these factors, such as building an authentic rapport with the offender;
investigating the most recent offense; designing clear and simple questions; using
incentives; choosing a comfortable, safe, and private location for interviews; using
psychological instruments to gain insights into, or control for, particular issues such
as social desirability. Regardless of research limitations and the techniques used to
lower their impact on the quality of data, the reality is that these limitations apply to
any research using self-report data regardless of the nature of the population or the
phenomenon under study. Under these circumstances, the nature of the data sought
by the researcher should drive the research and the method/investigation, not poten-
tial limitations and confounding factors.
How can situational prevention best operate from a script approach? One part of
the answer relates to crime specificity. Crime specificity is a principle strongly
advocated in situational crime prevention (Clarke, 2008). This principle refers to
how specific the crime under study should be. Clarke recommends focusing on very
specific categories of crime because each crime involves unique motivations, unique
decision-making processes, and different situational factors that will all have an
impact on how a crime will be committed (Cornish & Clarke, 2008). Clarke even
argues that situational prevention can succeed only when a detailed analysis of very
specific categories of crime is conducted. This point is analogous to Cornish’s
5 Script Analysis and the SCRIPT Acronym 53

(1994) idea of using crime scripts for situational prevention. Cornish argues that
crime script tracks are the most useful scripts for situational prevention because
they are the most specific to a particular crime. Essentially, there are four levels of
generality of crime scripts, that is, the metascript, the protoscript, the script, and
script tracks. The metascript level comprises all crimes from the same classification
(e.g., sexual offending and burglary). The protoscript is the level at which different
subgroups of a specific crime are distinguished (e.g., sexual offenses against chil-
dren versus sexual offenses against adults). At the script level, the crime is subdi-
vided into categories according to a situational dimension relevant for prevention
(e.g., domestic versus public crime location). The track level provides a more
detailed study of a crime-commission process within a particular set of circum-
stances and therefore is the level at which situational crime prevention can best
operate. Indeed, a script track is the level at which the interaction between the
offender and the environment in which the crime is committed is given most consid-
eration. An example of this is the commission of sexual offenses against boys in
youth-oriented organizations as a crime-specific track (Cornish, 1998).

The Strengths of Crime Scripts Using the Acronym


of “SCRIPT”

The strengths of script analysis are numerous. I will only discuss some of these
strengths here—others are also likely identifiable. In the tradition of problem-
oriented policing and situational prevention (e.g., Clarke, 1999; Clarke & Eck,
2005; Farrell & Tilley, 2016), I have developed an acronym to tap into the strengths
of crime script analysis. I have simply used “SCRIPT” (systematic, compatible,
resilient, insightful, practical, theoretical) as an acronym for that purpose. The
reader should note that Borion (2013) has previously looked at the qualities that
empirically generated crime scripts should possess in order to ensure that they are
useful for computer modeling approaches and crime prevention. In this chapter, I
take a different approach and look at the inherent qualities of the analysis of scripts
rather than the qualities that empirical scripts should have for practical purposes.
First, crime script analysis is systematic. It has the power to break down a com-
plex sequence of actions (e.g., the crime-commission process) into a series of steps,
which provides a structure and therefore simplifies it (Cornish, 1994). In other
words, this analysis simplifies complex behavioral processes through a systematic
approach. It is arguably impossible to study crime-commission processes taken as a
whole because they include different stages, each involving different decisions,
actions, and/or situations and even sometime different actors.
Second, crime script analysis is compatible with the perspective of any actors
involved in crime events. Script analysis is not limited to the offenders’ point of view
to prevent crime but can involve other actors in crime events as well (Leclerc, 2013a;
2013b). The victim’s perspective can be adopted to uncover the script for avoiding
victimization successfully (see Smith, 2010; 2016). A researcher can also think of
54 B. Leclerc

crime through the lens of capable guardians and identify the script for intervening
during the commission of a crime in a particular context (Leclerc & Reynald, 2015).
Police scripts can also be examined to facilitate successful and safer interventions
(Dwyer, Graesser, Hopkinson, & Lupfer, 1990). The bottom line is that script analysis
can be adopted from angles other than the offenders’ for crime prevention purposes.
Third, crime script analysis is resilient. It is resilient to time, the evolution of crime,
criminal organizations, offenders’ methods for committing crime, and the adaptation
of the crime prevention field of research in response to crime evolution, which is not
necessarily the case for a number of situational prevention techniques (see Ekblom &
Tilley, 2000). Crime scripts are and will always be applicable to the study of crime-
commission processes simply because any goal-directed sequence of human behavior
can be scripted regardless of its form or content and when it is examined.
Fourth, crime script analysis is insightful. Script analysis has the power of reveal-
ing not only the actions adopted by offenders at each step of the crime-commission
process but also the decision-making process underneath those actions and perhaps
even the motives driving behaviors (Cornish & Clarke, 2002). Scripts can also cap-
ture the behavioral habits or routines of repeat offenders when they carry out par-
ticular forms of crime as well as the range of actors involved, if any, in criminal
activities. Script analysis can be informative well beyond the actions of offenders.
Fifth, crime script analysis is practical. Consistent with the essence of situational
prevention, script analysis has the power of generating a wide range of intervention
points for situational prevention because it breaks down the crime-commission pro-
cess into a series of stages (Cornish, 1994). Given that the most valuable aspect of
script analysis is disrupting the crime-commission process, two aspects of script
analysis are important for achieving this goal. First, more intervention points means
more opportunities for disrupting the crime-commission process before completion.
Second, as script analysis may allow us to uncover at which stages offenders could
be most effectively stopped, it has the potential to allow us to direct situational
efforts to these points in particular.
Sixth, the concept of scripts is in its foundation theoretical. It is theoretically
sound and grounded. As discussed above, it is deeply rooted in cognitive science
and social psychology (Fiske & Taylor, 1991; Schank & Abelson, 1977; in criminol-
ogy, see Cornish, 1994). This theoretical foundation provides a neat and comple-
mentary framework from which to approach the study of crime, criminal decision
making, and importantly crime prevention.
In terms of weaknesses, there are important obstacles to the effective use of
crime script analysis. Perhaps the major shortcoming is related to the nature of data
available to study crime-commission processes, not the analysis of scripts
themselves. Moreto and Clarke (2013) illustrated this point well in their discussion
of the potential of crime script analysis to combat wildlife crime. Wildlife crime
involves a complex interplay of a range of actors across different countries and, as
a result, getting access to quality data in the first place is an important challenge.
Moreto and Clarke pointed out that scripts could be useful for further understanding
and preventing wildlife crime but getting access to relevant data may be simply too
difficult in this case.
5 Script Analysis and the SCRIPT Acronym 55

Another potential weakness may be the terminology used in crime script analysis
research. Cornish has used different terminologies to refer to the stages of crime-
commission processes. In 1994, Cornish borrowed the terminology used in cogni-
tive psychology in his adaptation of scripts to crime (e.g., instrumental initiation,
instrumental actualization). In recent research, Cornish and others have used differ-
ent terminology perhaps simpler or perceived as more applicable for criminologists
(e.g., initiation, continuation, completion) (Cornish & Clarke, 2008). For instance,
Savona et al. (2013) followed this approach and used a different terminology than
cognitive psychology to uncover the stages of human trafficking (e.g., recruitment,
transportation, exploitation, escape). In the end, the researcher should potentially
choose the terminology that best suits the purpose of the research. Perhaps it should
be kept in mind that the terminology used in cognitive psychology keeps us closer
to the theoretical relevance of scripts for providing insights into human decision
making in general.

What Is the Future of Script Analysis in Criminology?

To make the best of script analysis, quality data are essential, which means that there
is a need to collect data from offenders as well as from a range of other actors
involved in crime events depending on the type of scripts under study. Getting access
to self-report data can represent an important challenge in the current context where
research budgets are cut down by governments and access to incarcerated offenders
is getting more difficult than ever due to how institutions operate. This research con-
text does not encourage future and young researchers to undertake research with
offenders. However, other data sources can (and should) be used to uncover crime
scripts and young scholars should be encouraged to use these methods if necessary.
In the tradition of offender-based research, there is also the possibility of studying
active offenders (e.g., Jacobs, 1999; Jacques & Bernasco, 2013; Wright & Decker,
1994, 1997) although this may prove challenging with certain offender populations,
such as sexual and homicide offenders. Script analysis has the advantages of being
systematic, compatible, resilient, insightful, practical, and theoretically grounded. As
pointed out by Cullen (2011), more and more there is a need to study how offenders
think and how crimes are committed in order to better understand and reduce crime.
I am hopeful that young researchers will seriously take this fact into consideration
and see script analysis as a research avenue to make a significant contribution to
crime, its prevention, and criminology more broadly.

References

Augoustinos, M., Walker, I., & Donaghue, N. (2006). Social cognition: An integrated introduction
(2nd ed.). London: Sage.
Baldwin, M. W. (1992). Relational schemas and the processing of social information. Psychological
Bulletin, 112, 461–484.
56 B. Leclerc

Baldwin, M. W. (1995). Relational schemas and cognition in close relationships. Journal of Social
and Personal Relationships, 12, 547–552.
Beauregard, E., Proulx, J., Rossmo, D. K., Leclerc, B., & Allaire, J.-F. (2007). Script analyses of
the hunting process of serial sex offenders. Criminal Justice and Behavior, 34, 1069–1084.
Bernasco, W. (2010). Offenders on offending: Learning about crime from criminals. Cullompton,
UK: Willan.
Borion, H. (2013). Quality assurance in crime scripting. Crime Science, 2, 1–12.
Brayley, H., Cockbain, E., & Laycock, G. (2011). The value of crime scripting: Deconstructing
internal child sex trafficking. Policing, 5, 132–143.
Bullock, K., Clarke, R. V., & Tilley, N. (2010). Introduction. In K. Bullok, R. V. Clarke, & N. Tilley
(Eds.), Situational prevention of organized crimes (pp. 1–16). Devon: Willan.
Chiu, Y.-N., & Leclerc, B. (2016). An examination of sexual offenses against women by acquain-
tances: The utility of a script framework for prevention purposes. In B. Leclerc & E. U. Savona
(Eds.), Crime prevention in the 21st century: Insightful approaches for crime prevention initia-
tives. New York: Springer.
Chiu, Y.-N., Leclerc, B., & Townsley, M. (2011). Crime script analysis of drug manufacturing in
clandestine laboratories: Implications for strategic intervention. British Journal of Criminology,
51, 355–374.
Clarke, R. V. (1999). Hot products: Understanding, anticipating and reducing demand for stolen
goods (Police research series paper, Vol. 112). London: Home Office.
Clarke, R. V. (2008). Situational crime prevention. In R. Wortley & L. Mazerolle (Eds.),
Environmental criminology and crime analyses (pp. 178–194). Cullompton, UK: Willan.
Clarke, R. V., & Eck, J. (2005). Crime analysis for problem solvers in 60 small steps. Washington,
DC: US Department of Justice Office and Community Oriented Policing Services. www.pop-
center.org.
Clarke, R. V., & Newman, G. R. (2006). Outsmarting the terrorist. Westport CT: Praeger Security
International.
Copes, H., Hochstetler, A., & Cherbonneau, M. (2012). Getting the upper hand: Scripts for manag-
ing victim resistance in carjackings. Journal of Research in Crime and Delinquency, 49,
249–268.
Cornish, D. B. (1994). The procedural analysis of offending and its relevance for situational pre-
vention. In R. V. Clarke (Ed.), Crime prevention studies (Vol. 3). Monsey, NY: Criminal Justice
Press.
Cornish, D. B. (1998). Regulating lifestyles: A rational choice perspective. Paper presented at the
7th International Seminar on Environmental Criminology and Crime Analysis, Barcelona,
Spain.
Cornish, D. B., & Clarke, R. V. (2002). Analyzing organized crimes. In A. R. Piquero & S. G.
Tibbetts (Eds.), Rational choice and criminal behavior: Recent research and future challenges
(pp. 41–63). New York, NY: Routledge.
Cornish, D. B., & Clarke, R. V. (2008). The rational choice approach. In R. Wortley & L. Mazerolle
(Eds.), Environmental criminology and crime analysis. Cullompton, UK: Willan.
Cullen, F. (2011). Beyond adolescence-limited criminology: Choosing our future: The American
Society of Criminology 2010 Sutherland address. Criminology, 49, 287–330.
Dwyer, W. O., Graesser, A. C., Hopkinson, P. L., & Lupfer, M. B. (1990). Application of script
theory to police officers’ use of deadly force. Journal of Police Science and Administration, 17,
295–301.
Ekblom, P. (2012). Happy returns: Ideas brought back from situational crime prevention’s explora-
tion of design against crime. In G. Farrell & N. Tilley (Eds.), The reasoning criminologist:
Essays in honour of Ronald V. Clarke (Crime science series, pp. 163–198). Cullompton, UK:
Willan.
Ekblom, P., & Tilley, N. (2000). Criminology, situational crime prevention and the resourceful
offender. British Journal of Criminology, 40, 376–398.
Farrell, G., & Tilley, N. (2016). Technology for crime and crime prevention: A supply side analysis.
In B. Leclerc & E. U. Savona (Eds.), Crime prevention in the 21st century: Insightful approaches
for crime prevention initiatives. New York, NY: Springer.
5 Script Analysis and the SCRIPT Acronym 57

Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York, NY: McGraw-Hill.
Hiropoulos, A., Freilich, J., Chermak, S., & Newman, G. (2013). Cigarette smuggling and terror-
ism financing: A script approach. In B. Leclerc & R. Wortley (Eds.), Cognition and crime:
Offender decision-making and script analyses (Crime science series, pp. 186–208). London,
UK: Routledge.
Jacobs, B. (1999). Dealing crack: The social world of streetcorner selling. Boston, MA:
Northeastern University Press.
Jacques, S., & Bernasco, W. (2013). Drug dealing: Amsterdam’s Red Light District. In B. Leclerc
& R. Wortley (Eds.), Cognition and crime: Offender decision-making and script analyses
(Crime science series, pp. 120–139). London, UK: Routledge.
Jacques, S., & Bonobo, E. (2016). Learning from the offenders’ perspective on crime prevention.
In B. Leclerc & E. U. Savona (Eds.), Crime prevention in the 21st century: Insightful approaches
for crime prevention initiatives. New York, NY: Springer.
Leclerc, B. (2013a). New developments in script analysis for situational crime prevention: Moving
beyond offender scripts. In B. Leclerc & R. Wortley (Eds.), Cognition and crime: Offender
decision-making and script analyses (Crime science series, pp. 221–236). London, UK:
Routledge.
Leclerc, B. (2013b). Script analysis for crime controllers: Extending the reach of situational pre-
vention. In S. Caneppele & F. Calderoni (Eds.), Organized crime, corruption, and crime pre-
vention—Essays in honours of Ernesto U. Savona. New York, NY: Springer.
Leclerc, B., & Reynald, D. (2015). When scripts and guardianship unite: A theoretical script model
to facilitate intervention by capable guardians. Security Journal. doi:10.1057/sj.2015.8.
Leclerc, B., Smallbone, S., & Wortley, R. (2013). Interpersonal scripts and victim reaction in child
sexual abuse: A quantitative analysis of the offender-victim interchange. In B. Leclerc &
R. Wortley (Eds.), Cognition and crime: Offender decision-making and script analyses (Crime
science series, pp. 101–119). London, UK: Routledge.
Leclerc, B., Wortley, R., & Smallbone, S. (2011). Getting into the script of adult child sex offend-
ers and mapping out situational prevention measures. Journal of Research in Crime and
Delinquency, 48, 209–237.
Lindegaard, M., & Bernasco, W. (2016). Learning about crime prevention from aborted crimes:
Intra-personal comparisons of committed and aborted robbery. In B. Leclerc & E. U. Savona
(Eds.), Crime prevention in the 21st century: Insightful approaches for crime prevention initia-
tives. New York, NY: Springer.
Lindegaard, M. R., Bernasco, W., Jacques, S., & Zevenbergen, Z. (2013). Posterior gains and
immediate pains: Offender emotions before, during and after robberies. In J.-L. Van Gelder,
H. Elffers, D. Reynald, & D. Nagin (Eds.), Affect and cognition in criminal decision making
(Crime science series). London, UK: Routledge.
Moreto, W. M., & Clarke, R. V. (2013). Script analysis of the transnational illegal market in endan-
gered species: Dream and reality. In B. Leclerc & R. Wortley (Eds.), Cognition and crime:
Offender decision-making and script analyses (Crime science series, pp. 209–220). London,
UK: Routledge.
Morselli, C., & Roy, J. (2008). Brokerage qualifications in ringing operations. Criminology, 46,
71–98.
Savona, E. U. (2010). Infiltration by Italian organized crime (Mafia, N’drangheta and Camorra) of
the public construction industry. In K. Bullok, R. V. Clarke, & N. Tilley (Eds.), Situational
prevention of organized crimes (pp. 130–150). Devon: Willan.
Savona, E. U., Giommoni, L., & Mancuso, M. (2013). Human trafficking for sexual exploitation
in Italy. In B. Leclerc & R. Wortley (Eds.), Cognition and crime: Offender decision-making
and script analyses (Crime science series, pp. 140–163). London, UK: Routledge.
Schank, R., & Abelson, R. P. (1977). Scripts, plans, goals and understanding: An inquiry into
human knowledge. Hillsdale, NJ: Erlbaum.
Smith, M. J. (2010). Constructing crime scripts from victim and near-victim narratives. Paper pre-
sented at the Annual Meeting of the American Society of Criminology, United States, San
Francisco.
58 B. Leclerc

Smith, M. J. (2016). Expanding the script analytic approach using victim narratives: Learning
about robberies of taxi drivers from the drivers themselves. In B. Leclerc & E. U. Savona
(Eds.), Crime prevention in the 21st century: Insightful approaches for crime prevention initia-
tives. New York, NY: Springer.
Thompson, L., & Chainey, S. (2011). Profiling illegal waste activity: Using crime scripts as a data
collection and analytical strategy. European Journal on Criminal Policy and Research, 17,
179–201.
Wright, R., & Decker, S. (1994). Burglars on the job: Street life and residential break-ins. Boston,
MA: Northeastern University Press.
Wright, R., & Decker, S. (1997). Armed robbers in action: Stickups and street culture. Boston,
MA: Northeastern University Press.
Chapter 6
An Examination of Sexual Offenses Against
Women by Acquaintances: The Utility
of a Script Framework for Prevention
Purposes

Yi-Ning Chiu and Benoit Leclerc

Introduction

According to community prevalence studies, approximately 14–25 % of adult


women have experienced rape (Kilpatrick, Edmunds, & Seymour, 1992; Koss,
1993). Research shows that the majority of rape cases are committed by offenders
known to the victim, with current partners and acquaintances comprising the bulk of
the offenders (Kilpatrick et al., 1992; Koss, Dinero, & Seibel, 1988; Myhill & Allen,
2002; Tjaden & Thoennes, 2000). The chapter uses court data to apply crime script
analysis to sexual offenses against women by acquaintances (i.e. non-stranger
offenders previously known to the victim, including partners, friends, relatives and
acquaintances), in order to promote understanding of the crime-commission pro-
cesses involved via script analysis and to identify potential points of intervention.

An Overview of Previous Research

In the absence of data on the crime-commission process of acquaintances sex


offenders specifically, we present a summary using the literature on sexual offenses
against women in general. In regards to victim selection, offenders listed factors
such as victim vulnerability, location or availability and general physical appear-
ance as main reasons for selecting their victims (Beauregard, Rossmo, & Proulx,
2007; Stevens, 1994). Offending strategies (sometimes referred to as modus ope-
randi) were measured in different ways, ranging from simple blitz (sudden forceful
attack) versus con methods (gaining the victim’s confidence gradually in various

Y.-N. Chiu, Ph.D. (*) • B. Leclerc, Ph.D.


School of Criminology and Criminal Justice, Griffith University,
Mt Gravatt Campus, Brisbane, QLD 4111, Australia
e-mail: winnie.chiu@griffithuni.edu.au; b.leclerc@griffith.edu.au

© Springer International Publishing Switzerland 2017 59


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_6
60 Y.-N. Chiu and B. Leclerc

ways) (Atkenson, Calhoun, & Morris, 1989; Ullman & Knight, 1991, 1992), to
more sophisticated classifications based on patterns detected, such as home intruder,
tracker and lurker (Beauregard, Rebocho, & Rossmo, 2010). Other variables exam-
ined in relation to offending strategies include the number of offenders involved,
breaking and entering and the use of violence or tools (including weapons, vehicles
or disguises), which could be seen as ‘facilitating hardware’ for the crime (Gibson,
Linden, & Johnson, 1980). Weapons are used as an instrument to leave no choice to
victims but to perform sexual behaviours on offenders regardless of the offender-
victim relationship (Leclerc & Cale, 2015). Offender violence showed a curvilinear
relationship with the degree of closeness of the victim-offender relationship (Ullman
& Siegel, 1993), where acquaintance offenses are generally less violent than
stranger ones; however, this excludes offenses by intimates and family members
(Koss et al., 1988). Offenders previously known to the victim tend to be more likely
to seek interaction with the victim after the attack (Bownes, O’Gorman, & Sayers,
1991). Studies that questioned offenders on their motivations found revenge as the
main motivation (Hale, 1997; Mann & Hollin, 2007). Other motives include power,
anger, seeking sexual gratification to enhance positive mood and sexual pleasure.

Theoretical Background

Rational Choice Approach

The rational choice approach posits that potential offenders are people who make
deliberate and purposive decisions to commit crime, with the ultimate aim to benefit
themselves. This approach embodies the shift of focus from the offender to the
crime event (Cornish & Clarke, 2008). Instead of viewing crime as a stable trait or
predisposition of the individual (Waldo & Dinitz, 1967), the offender in this
approach is perceived as a person in a state of readiness for crime and acts on crimi-
nal opportunity (Clarke & Cornish, 1985). It is the offender’s decisions that are seen
to drive criminal behavior. These decisions are considered in an environment of
both opportunities and constraints, which can facilitate or prohibit potential crimes.
Although this process may not always involve conscious and explicit decision-
making, it contains some degree of rationality, meaning these decisions are not ran-
dom or unpredictable and can be reconstructed (Brantingham & Brantingham,
1978; Clarke & Cornish, 1985; Cusson, 1983; Walsh, 1978; Willmer, 1970).
The rational choice perspective does not attempt to decipher the nature of the
offender, nor explain the origins of criminal behaviour. It has a more pragmatic
application: the prevention and disruption of criminal actions. The rational choice
perspective provides a conceptual framework for situational crime prevention.
Crime can be discouraged by increasing the efforts and risks necessary to commit
the crime, and by reducing the associated rewards. Focus is given to the interaction
of the offender with the immediate environment during crime-commission, and the
role of opportunity and situational factors. The more specific the crime, the better
6 An Examination of Sexual Offenses Against Women by Acquaintances… 61

potential that exists for crime prevention and analysis. With a higher level of speci-
ficity, more detailed understanding and tailored prevention can be applied. This
specificity is important for crime scripts.

Crime Script Analysis

Using the rational choice approach as a theoretical framework, Cornish (1994) bor-
rowed the concept of scripts from cognitive psychology to develop a systematic
framework to decipher the extent of crime-commission processes involved in crime.
Crime scripts were formed to capture the step-by-step accounts of procedures
before, during and after crime-commission, including any actors, equipment, loca-
tions, resources, decisions and activities involved. Crime scripts enhance the under-
standing of crime-commission, because the crime is perceived as a process rather
than a single event and breaks down the crime into stages that can be analysed
(Cornish, 1994). Once performed, the actions involved form a cognitive script,
which becomes a subconscious template for future behaviour—meaning crime
scripts have the ability to predict criminal behaviour. It therefore becomes a useful
tool to (1) break down and examine the crime-commission process, and (2) develop
potential situation prevention strategies at every stage of the crime.
Currently other crime scripts exist in criminological literature for sexual offenses
(Beauregard, Proulx, Rossmo, Leclerc, & Allaire, 2007; Deslauriers-Varin &
Beauregard, 2010; Leclerc, Wortley, & Smallbone, 2011). Beauregard and col-
leagues focused on either target selection or the ‘hunting process’ of offenders.
These studies examined serial sexual offenses committed by strangers. Leclerc et al.
(2011) used scripts for situational prevention purposes but with a sample of child
sexual offenders. Therefore, there is currently no script that permits the breaking
down of the entire crime-commission process of acquaintance-perpetrated sexual
offenses for prevention purposes.

Aims of the Study

Although crime scripts were originally developed in part to provide a roadmap in


which situational prevention can be matched to each step of the crime-commission
process (Cornish, 1994), only a few studies have actually used crime scripts for this
purpose (e.g. Chiu, Leclerc, & Townsley, 2011; Clarke & Newman, 2006; Leclerc
et al., 2011). Prevention strategies may reveal potential weak points in the script, by
identifying areas where the crime-commission process can be disrupted by reduc-
ing associated opportunities and increasing risks (Levi & Maguire, 2004). Despite
the recent use of crime scripts in sexual offenses, it has not yet been applied to
acquaintance sexual offenses against women, which is the most common context in
which sexual violence is perpetrated against women (Myhill & Allen, 2002). In this
62 Y.-N. Chiu and B. Leclerc

study, we use a crime script framework to investigate sexual offenses against


women by acquaintances. The first aim is to better understand the steps involved in
sexual offenses against women by acquaintances—despite the large body of
research on this subject, few studies examine and break down the crime event
beyond victim resistance. The second aim is to demonstrate the applicability of
script analysis for the purposes of crime prevention by completing the exercise of
identifying potential situational crime prevention measures to disrupt the script
process.

Method

Procedure and Sample

Content analysis of court transcripts relating to sexual offenses by acquaintances in


Queensland, Australia, was conducted. The use of court data has both weaknesses
and strengths. It is possible for court data to be subject to inaccuracies such as
incorrect recording, lack of details and, in particular, focus on only certain aspects
of the event (Porter, 2008). This means that information relevant to the analysis
might not always be present. However, the strengths of court data outweigh these
weaknesses, including many problems inherent to other data collection methods
such as interviews or questionnaires (Berg, 2007; Porter, 2008). Interviews and
questionnaires are vulnerable to biases such as subject recollection and subject
perception distortion and matters such as truth and accuracy. On the other hand,
archival court data has undergone more stringent legal scrutiny and generally based
on multiple sources of evidence, such as witness statements, forensic evidence and
offender statements. In effect, court data is more reliable than information collected
from a single source as potential biases are countered by other pieces of evidence
(Porter, 2008). Another advantage of this data is that it is a form of open-source
data, which is easily accessible to the public and therefore involve fewer confiden-
tiality and ethical issues. Information collection is therefore fast and inexpensive,
which is a valuable quality when compiling crime scripts and in producing relevant
research that can catch up to the dynamic nature of crimes and shifting policies
(Cornish & Clarke, 2002).
Public online records in Austlii databases were searched using a Boolean search
operator in all Queensland cases. To avoid overlaps with child sexual abuse, the
search terms ‘Rape NOT child’ were used. Boolean searches match these search
terms by free text within documents, which contained the judge’s sentencing
remarks. The results were then filtered to ensure that they fit the necessary param-
eters and that cases did not overlap. In total 23 cases were selected within the fol-
lowing parameters: acquaintance rape, adult female victim (over 16 years of age),
hands-on offenses, male offenders, occurred in Queensland and were detailed
enough for analysis. This did not exclude cases that involved more than one offender.
Case details are listed in the Appendix.
6 An Examination of Sexual Offenses Against Women by Acquaintances… 63

Data and Analytic Strategy

The analytic strategy involved a further three-step process. The first summarisation
step broke down the court data into the following categories: Actors, Background,
Relationship, Circumstances, Approach Method, Locations, Victim reaction,
Offender reaction, Sexual acts, Exit, Threats/Force, Tools/Transport, Post-Action
and Time. The events were listed in dot point form (i.e. summarised) then broken
down (i.e. location, victim reaction, threat). Secondly the data was coded and tabu-
lated, as indicated in Table 6.1. The categories were determined in part by existing
literature on sexual crimes, including drawing from offender approach methods
such as con, surprise, blitz (Atkenson et al., 1989; Lea, Hunt, & Shaw, 2011; Park,
Schlesinger, Pinizzotto, & Davis, 2008) as well as the child sex offending script
(Leclerc et al., 2011). For example, victim resistance methods have often been bro-
ken down into forceful verbal, non-forceful verbal and physical resistance catego-
ries (Beauregard, Lussier, & Proulx, 2005; Ullman & Knight, 1995). The last step
was script development. The above information gradually coalesced into a complete
script, represented by a web of different stages, actions and actors involved in time
sequence. The common patterns and points in each case were identified and com-
pared. Although not every case proceeded through every stage listed in the final
script due to the individual circumstances of each case, the final script represents the
general pattern from the 23 cases.

Results

Script and Actions

The final script comprises of eight different stages, including offender(s) and victim
prehistory, setting, instrumental actualisation, offender approach method, continua-
tion, interaction and crime completion, post-action and exit (see Fig. 6.1). The first
few steps involve the crime-setup phase, which is followed by the crime completion
phase.
1. Offender/s and Victim Prehistory
Offender and victim prehistory is unique to acquaintance rape in that they
share both a previously established relationship and socially historical events
leading up to the crime. The most common relationship between offenders and
victims was friend or acquaintance. The second most common type of relation-
ship was partner/ex-partner.
2. Setting
Factors surrounding the setting of the crime including the circumstances of
their meeting, the location involved, the time of day and also any surrounding
third parties who may have been present initially or during the course of the
crime. The most common location involved a private indoor area, more specifically
64 Y.-N. Chiu and B. Leclerc

Table 6.1 Data categorisation from court transcripts


Variables Categories
Social elements
Actors involved Offender, victim, co-offenders, third parties (victim’s children;
partners; other relatives and other known—friends, acquaintances,
co-workers, housemates)
Relationship between Date, partner or ex-partner (including husband, boyfriend, de facto
victim and offender partner, and sex friend), relative, friend/acquaintance, co-worker or ex
co-worker, neighbour, family member’s spouse
Circumstances Date rape, break in, group outing, visitation
Approach method
Con Offered or asked for assistance/ride/information, etc.
Solicitation for sex
Wanted to show something to victim
Bribed victim with money, drugs, etc.
Surprise Victim was asleep
Lay in wait/snuck up on victim
Blitz Used weapon
Used overwhelming force or violence
Setting
Location Offender’s home, victim’s home, vehicle, other residence, other
building, outdoors
More than one crime Yes/no
scene
Time Daylight, darkness
Interaction
Victim reaction Forceful verbal resistance (screamed, yelled/yelled for help),
non-forceful verbal resistance (said no, pleaded, used threats), cried,
physical resistance (fought back, tried to escape), no victim reaction,
compliance, called police, negotiation
Offender reaction Ignored, used force, used threats, used violence, negotiated, ceased to
demand
Sexual acts Kissing, grabbing/hugging, fondling, digital penetration, vaginal
penetration, anal penetration, cunnilingus, fellatio, masturbation,
suffocation, choking, beating/slapping
Other variables
Disruption Third party intervened/disrupted, victim escaped
Exit Offender left scene, victim left scene, offender dropped victim home,
victim dropped offender home, victim passed out
Threats/force Threats used, violence used, weapon used
Tools/transport Weapon, car, condom, gag, bindings, disguise, other (e.g. alcohol,
victim’s keys, removed car door handle)
Post-action Apologies, threats, tried to comfort/be intimate with victim, tried to
ask victim on a date, told victim not to tell anyone, stole from victim,
told victim not to move
(continued)
6 An Examination of Sexual Offenses Against Women by Acquaintances… 65

Table 6.1 (continued)


Variables Categories
Other Potential ‘triggers’ (date/sexual advance rejection, prior relationship
break-up), prior alcohol consumption (victim, offender), prior drug
consumption, offender removed clothing, offender made victim
remove clothing, moderate-to-severe victim injury, alcohol consumed
during offense, drugs involved during offense
Note: The term ‘trigger’ refers to factors or offender-victim prehistory that may result in the
offender being in a heightened emotional state (e.g. stress/anger), and is not intended to imply
causality

Fig. 6.1 Crime script for sexual offenses against women by acquaintances
66 Y.-N. Chiu and B. Leclerc

either the victim or offender’s home. By far the most common circumstances
leading up to sexual offenses by acquaintances was some form of visitation. This
visitation could occur either where the offender visited the victim, or vice versa,
or even where one visited another person living in the same house as the other
(for example, a roommate). Other studies have also shown that sexual offenses
involving known offenders are more likely to occur indoors (Jones, Wynn,
Kroeze, Dunnuck, & Rossman, 2004; Porter & Alison, 2004), with the majority
of these events occurring within the victim’s or assailant’s home (Stermac, Du
Mont, & Kalemba, 1995). Contrary to stranger sexual offenses (e.g. Beauregard
& Leclerc, 2007), only a few cases involved the use of multiple offense locations,
which involved transporting the victim to another site. The cases overwhelmingly
occurred at night, and almost half involved third parties.
3. Instrumental Actualization
The main element that was instrumental to the crime was isolation of the
victim. Even though many cases involved the presence of third parties in the
house, the victim was isolated in one part of the house before the sexual assault
took place. In some cases, the offender waited until the victim was alone (e.g. the
victim walked into another room or others left room), or actively isolated her
from nearby third parties. In other instances they were initially alone in the set-
ting itself. Another factor examined was the concept of potential ‘triggers’ such
as rejection or prior relationship break-up. It is important to note that the word
‘trigger’ may imply cause, which is not the intended meaning in this instance.
This trigger refers to the inherent history between the offender and victim, in
which may result in the offender being in a heightened emotional state such as
stress or anger at the time.
4. Offender Approach Method
Offenders most commonly approached the victim through various con meth-
ods (especially offering or asking for assistance and information, followed by
direct solicitation for sex), then surprise (in particular where the victim was
asleep) and finally blitz. The use of extreme violence and weapons was rare in
this sample. This supports findings that acquaintance sexual offenses are less
likely to involve physical force and violence than stranger offenses (Stermac
et al., 1995; Woods & Porter, 2008).
5. Continuation
The continuation of the crime may involve the use of tools, or removal of
clothing. Mostly the offenders removed the victim’s clothing themselves rather
than forcing the victim to take off their clothing. The tools mentioned in these
cases included car, weapon, condom, bindings and gag, mask and other.
6. Interaction and Crime Completion
Instead of a sequence where one step clearly preceded the other, it was
found that offender reaction, victim reaction and sexual acts formed an inter-
active relationship with each other (see Leclerc, Smallbone, & Wortley, 2013).
For instance, offender reaction was shaped in part by victim reaction, which
in turn affected the acts performed. The cyclical pattern resulted in this
6 An Examination of Sexual Offenses Against Women by Acquaintances… 67

interaction reoccurring over and over again in the sequence of events.


Offenders tended to react by using force and threats, or by ignoring the victim.
In some instances, violence, negotiation or ceasing the assault occurred.
Victim reaction ranged from physical resistance, forceful and non-forceful
verbal resistance, to compliance and no reaction. Due to the nature of the data,
being court cases, most of the cases resulted in more severe sexual outcomes,
namely vaginal penetration.
7. Post-Action
Post-actions completed by the offender in some instances included a variety
of behaviors, ranging from telling the victim not to tell anybody what happened,
to apologies, threats, comforting the victim, stealing money from the victim,
asking the victim on a date or telling the victim not to move so the perpetrator
could avoid apprehension. The nature of actions may be shaped by the relation-
ship between the victim and offender and their prehistory. This is consistent with
literature showing that offenders previously known to the victim are significantly
more likely to seek interaction with the victim after the attack (Bownes et al.,
1991).
8. Exit
The exit involves the stage where either the victim or offender left the
scene. Only in a few cases did the victim escape. Interestingly, despite the
number of cases where third parties were present at some point, it was rela-
tively rare for third parties to intervene or disrupt the crime. This could be in
part due to the instrumental actualization stage, where the offender success-
fully isolated the victim from others; therefore, even if potential guardians
were in the same house they were unaware the assault took place. Because of
this, the few instances where intervention occurred were completely by acci-
dent (for example, calling out to offender), where they did not intentionally
disrupt the offense. This finding is inconsistent with previous studies that
found the presence of a bystander to reduce the probability of completed
sexual offenses (Clay-Warner, 2002). However, it is of note that in many
instances the third parties were isolated from the victim in a different section
of the house during the actual assault, and in such circumstances may not
have been categorized as direct ‘bystanders’ in previous studies. These third
parties were present in the house, but not at the offense scene and therefore
were ineffective as guardians.

Implications for Prevention

A broad range of potential prevention strategies are identified and broken down
according to each crime script stage in Table 6.2. Suggestions that are specific to
certain offender-victim relationships are labelled. The potential situational
Table 6.2 Crime script and potential intervention points for sexual offenses against women by acquaintances
68

Script stages Manager-place Guardian-target Handler-offender


1. Offender/s and • Establish clear boundaries • Public awareness of
victim prehistory • General education regarding acquaintance rape and rape definitions and
potentially risky circumstances potential contexts
Partners: (e.g. date rape,
• Understanding the partner’s attitudes and unable to give
perceptions about rape will bring more awareness in consent is
the event of any ‘triggers’ (e.g. break-ups, fights, constituted as rape)
saying no) Partners:
• Be aware that changes in the relationship (e.g. • Pre-establish
rejection, break-ups) may influence partners’ actions boundaries, discuss
and potentially alter the dynamics and safety around concepts regarding
them rape perceptions
Relatives: and definitions,
• Be aware that offenses can happen even among acceptable and
relatives non-acceptable
behaviors, possible
scenarios, etc.
2. Setting • Maintaining visual transparency (e.g. Acquaintances/friends: • Presence of friends/
leaving door or window open) • Meet in a group rather than alone with potential family can act as
• Increase general lighting and visibility offender handlers
• Meet in public areas • Designate a friend to look out for each other through
• Meet in familiar areas the night
• Designate curfews • Identify potential escape routes and strategies
• Patrols of hotspots (e.g. clubs, bars) • Control levels of intoxication
• CCTV • Organize exit (e.g. designated driver if drinking,
Partners: prepare car or taxi)
• Change house keys after a negative
break-up
Relatives:
• Lock bedroom doors, even if staying over
Y.-N. Chiu and B. Leclerc

at a relative’s house
6

Script stages Manager-place Guardian-target Handler-offender


3. Instrumental • Have guardians present in location • Encourage general awareness of environment • Increase public
actualization • Encourage women to trust their instincts when awareness in terms
reacting to dangerous/uncomfortable situations, and of guardianship and
get out the role and
Acquaintances/friends: practices required,
• Do not accept drinks that not prepared/overseen by by encouraging
oneself, or left unattended young people to
• Within reason, keep to areas where others are actively look out for
present to avoid isolation (e.g. wandering off at each other and
parties, sleeping over alone at male friend’s house, identify
sleeping on the sofa after drinking at a party) circumstances in
• Adopt a buddy system with another female which women can
• Have a system of having someone check via phone be vulnerable
or text if one has returned home safely after a night
out
• Maintain the attitude/habit among friends to look
after females in group settings
Relatives:
• Within reason limit isolation or opportunity for
offenses (e.g. drinking alone with male in-laws)
4. Offender • General house security (locks, alarms) • Victim resistance education • Remove excuses—
approach method • Security beside bed in the event of attacks • Education regarding common coercive strategies general education
while sleeping and break-ins (e.g. phone, and ways to respond regarding emotional
alarm) • Call to alert nearby guardians coercion and rape
• Education/encouragement against peer pressure, and
An Examination of Sexual Offenses Against Women by Acquaintances…

to develop resilience
5. Continuation • House security design if in victim’s home • Wear secure clothing that is difficult to remove
(e.g. panic button, speed dial for help)
(continued)
69
70

Table 6.2 (continued)


Script stages Manager-place Guardian-target Handler-offender
6. Interaction and Friends/acquaintances: • Victim self defense programs • Remove excuses—
crime completion • Encourage guardians (e.g. parents, • Victim protective strategies (e.g. pepper spray, widespread
roommates) to regularly check up on guests personal alarms) messages against
• Victim education regarding self protection in sexual sexual crimes (e.g.
assault situations ‘no means no’)
7. Post-action • Education regarding sexual crime, blackmail and
guilt techniques
8. Exit • Keep physical and forensic evidence intact • Counselling of victims regarding shame/blame
• Emotional counselling
• Establish support hotlines
• Encourage reporting
• Education of friends/family regarding victim
support
Note: Where suggestions have not been specified as applying particularly to one type of relationship, they are generalized and may be adopted by all offender-
victim relationships
Y.-N. Chiu and B. Leclerc
6 An Examination of Sexual Offenses Against Women by Acquaintances… 71

prevention measures are categorized using the problem triangle developed by


Clarke and Eck (2005). The problem triangle is comprised of three elements
(offender, place and victim) that must converge for any crime to occur. These three
elements form the inner triangle and are regulated by corresponding elements in the
outer triangle: the handler, manager and guardian. In other words, the handler super-
vises the offender, the manager supervises the crime setting and the guardian super-
vises the victim. Criminal opportunities are more likely to occur in the absence of
capable guardianship or place management (Felson, 2008). According to the ratio-
nal choice perspective (Clarke & Cornish, 1985; Cornish & Clarke, 1986), in these
circumstances the offender may exploit opportunities when benefits are perceived
to outweigh the cost.
In light of this analysis, the most effective intervention points appears to be
within the crime set up phase, in particular the setting and instrumental actual-
ization stages. Some situational prevention strategies in the setting can be
adopted by making the physical environment more secure in terms of visibility
and guardianship. A public setting could potentially benefit from general light-
ing, CCTV and patrols of hotspots. It may also be possible to designate cur-
fews, meet in public areas and in groups depending on the context, ask a friend
to look out for her (and vice versa) throughout the night, and pre-organize an
exit strategy (for example, having a designated driver). In private settings, sim-
ple security measures such as locks and the presence of others as guardians may
be applied. Strategies for the instrumental actualization stage can target public
awareness such as educating the public about acquaintance rape, guardianship
and the potential contexts in which it can occur. This has the effect of removing
excuses for the offender, educating people to become more effective guardians,
and informing victims of potentially risky circumstances that young women
may not be aware of. Women can also be encouraged to trust their instincts
regarding uncomfortable situations rather than succumb to peer pressure, and
be taught to be generally more aware and alert of the environment. Other guard-
ianship practizes that can be encouraged are meeting in a group rather than
alone, designating friends to look out for each other, controlling levels of intox-
ication while out, to identify potential escape routes and strategies and organiz-
ing the method of exit—for example, having a designated driver, or preparing
a car or taxi.
A main factor found to be instrumental to the acquaintance crime script is iso-
lation. The whole crime-commission process and crime achievement is hinged on
the offender either actively isolating the victim, or finding the victim alone.
Although potential guardians were present in almost half of these events, this
supervision was rendered useless once the victim was isolated with the offender
in any part of the house. Perhaps a key point in public and victim education is to
stress that even in a setting where others are present there are pockets of opportu-
nities for isolation and crime-commission. Therefore, even in a setting with other
people present, it is always necessary to maintain some level of vigilance and
awareness of the surroundings. As these opportunities are more likely to occur in
72 Y.-N. Chiu and B. Leclerc

a private residence, it is somewhat safer to conduct meetings in public areas with


a group. Also the safety of women in shared living situations should be treated
cautiously especially when hosting visitors or parties. In summary, these prac-
tizes have the effect of hardening the target (through victim education), training
the public to become better guardians, and removing excuses for offenders.
Improving the potential of third parties as guardians through public education is
important, as in many instances third parties were nearby, however, they failed to
act as effective guardians. Despite being in the same house, third parties were
often unaware the assault took place and did not intervene. This has implications
for guardianship. The presence of bystanders in sexual offenses (especially sex-
ual offenses by acquaintances), and the context and relationship of bystanders to
the scene, is largely neglected in literature. This is a phenomenon that needs fur-
ther exploration and analysis from a situational viewpoint. Although bystander
intervention may be a potential tool in the primary prevention of sexual offenses,
more information is needed to guide prevention programs (Banyard, 2008).
Finally, an important point is that in order to suggest more applicable preven-
tion strategies, the nature of the offender-victim relationship should be consid-
ered. An important factor for prevention is the need for crime specificity, as
conceptualized by the rational choice perspective (Cornish & Clarke, 2008);
strategies should be tailored to more specific situations for maximum effective-
ness—for example different strategies could apply for partners compared to
friends. This has been tentatively applied to the prevention table. For partners
and ex-partners, some suggested strategies include pre-establishing boundaries
and discussing concepts. By talking to partners regarding their perceptions of
acceptable and non-acceptable behavior, definitions of rape and possible
scenarios, it is possible to gain a better understanding of their attitudes. This will
enhance victim awareness in the event of any ‘triggers’ (e.g. break-up, saying
‘no’). Women should also be aware that changes in the relationship (e.g. rejec-
tion) may influence offender actions and potentially alter the level of safety once
experienced around them. Changing keys to the house after a bad break-up is
also a potential physical safety and prevention measure. For relatives and in-
laws some suggestions include locking bedroom doors even when staying over
at a relative’s house, maintain a sense of vigilance by being aware that offenses
can happen even among relatives and (within reason) limit opportunities for
offenses. In the context of friendships, is may be also possible to limit opportu-
nities for sexual offenses within reason (e.g. avoid sleeping over alone at a male
friend’s house, or avoid sleeping off alcohol on a friend’s sofa after drinking at a
party). Other potential strategies include adopting guardianship in the form of
other friends or a buddy system with another female; maintain the attitude or
habit among friends to always take extra care to look after females in group set-
tings. Finally for acquaintances it may be possible to avoid being alone with
acquaintances until getting to know them better, avoid being isolated even in
group situations, watch alcohol intake, leave if the situation seems uncomfort-
able and pre-arrange personal transport or a designated driver in outings.
6 An Examination of Sexual Offenses Against Women by Acquaintances… 73

Despite the applicability of situational crime prevention strategies to sexual


offending, the context of sexual offenses against women by acquaintances pres-
ents special challenges. In fact, human relationships and histories can be intri-
cately complex, and the advertised concept of women needing to trust their
instincts in this instance is problematic. Firstly, it is unclear, and secondly,
instincts can be misleading given that the victim may have had many prior safe
interactions alone with the offender. Due to these complexities, there is currently
a lack of clear and defined protective strategies for women in the context of sex-
ual offenses in established relationships. There are few realistic ‘one size fits all’
prevention strategies for acquaintance rape; solutions should be tailored to the
specific circumstances and offender-victim relationships. An initial attempt has
been made to address this in Table 6.2 by including strategies tailored to specific
relationships at each stage of the script. However, this is an area that clearly needs
further exploration and research.

Conclusion

In summary this study has explored the crime-commission processes involved


in sexual offenses against women by acquaintances and demonstrates the appli-
cability of script analysis in crime prevention by identifying key intervention
points through crime scripts. The value of crime scripts is manifold: scripts
conceptualize crime as a whole, assisting the understanding of the entire crime
process in all of its components and stages. This understanding is key to crime
prevention, and it is particularly valuable in sexual offenses against women
because there are currently gaps in academic knowledge regarding the complete
processes involved during the crime event. Another strength of crime scripts is
its value as an analytical tool to break down the crime and map out all the
stages. This provides potential to introduce prevention strategies at every stage
and identify key points to crime disruption. The analysis can also operate at dif-
ferent levels of specificity, and it can account for the interactive components of
the actors involved, and how this interaction evolves (e.g. it cycles, reacts, esca-
lates and de-escalates). This makes crime scripts a suitable tool to examine
interpersonal offenses. As there is much more potential for crime scripts to be
applied further in this area, in sexual offenses and as well as other offense types,
future directions may involve considering SOW scripts at different levels of
specificity, or examining offenses committed by unknown offenders specifi-
cally. This study did not account for non-apprehended offenders; it is possible
that offenders who have been caught and prosecuted may operate differently to
offenders that have not been caught. Indeed, offenders who have not been
caught represent a major gap in current research due to the difficulty of obtain-
ing data relevant to this type of sample. These are areas that could be explored
in future studies.
74 Y.-N. Chiu and B. Leclerc

Appendix

List of case numbers used in analysis

Number Case citation


1 QCA 9
2 QCA 12
3 QCA 26
4 QCA 30
5 QCA 48
6 QCA 61
7 QCA 67
8 QCA 76
9 QCA 100
10 QCA 102
11 QCA 116
12 QCA 125
13 QCA 142
14 QCA 198
15 QCA 211
16 QCA 226
17 QCA 244
18 QCA 265
19 QCA 279
20 QCA 290
21 QCA 339
22 QCA 350
23 QCA 501

References

Atkenson, B. M., Calhoun, K. S., & Morris, K. T. (1989). Victim resistance to rape: The relation-
ship of previous victimization, demographics and situational factors. Archives of Sexual
Behavior, 18(6), 498–507.
Banyard, V. L. (2008). Measurement and correlates of prosocial bystander behaviour: The case of
interpersonal violence. Violence and Victims, 23, 83–97.
Beauregard, E., & Leclerc, B. (2007). An application of the rational choice approach to the offend-
ing process of sex offenders: A closer look at the decision making. Sexual Abuse: A Journal of
Research and Treatment, 19, 115–133.
Beauregard, E., Lussier, P., & Proulx, J. (2005). The role of sexual interests and situational factors
on rapists’ modus operandi: Implications for offender profiling. Legal and Criminological
Psychology, 10, 265–278.
Beauregard, E., Proulx, J., Rossmo, K., Leclerc, B., & Allaire, J.-F. (2007). Script analysis of the
hunting process of serial sex offenders. Criminal Justice and Behavior, 34, 1069–1084.
6 An Examination of Sexual Offenses Against Women by Acquaintances… 75

Beauregard, E., Rebocho, M. F., & Rossmo, K. (2010). Target selection patterns in rape. Journal
of Investigative Psychology and Offender Profiling, 7, 137–152.
Beauregard, E., Rossmo, K., & Proulx, J. (2007). A descriptive model of the hunting process of
serial offenders: A rational choice perspective. Journal of Family Violence, 22, 449–463.
Berg, B. L. (2007). Qualitative research methods for the social sciences. Boston, MA: Pearson
Education.
Bownes, I., O’Gorman, E., & Sayers, A. (1991). Rape—A comparison of stranger and acquain-
tance assaults. Medical, Science and Law, 31, 102–109.
Brantingham, P. J., & Brantingham, P. L. (1978). A theoretical model of crime site selection. In
M. Krohn & R. Akers (Eds.), Crime law and sanctions (pp. 105–118). Beverly Hills, CA: Sage.
Chiu, Y.-N., Leclerc, B., & Townsley, M. (2011). Crime script analysis of drug manufacturing in
clandestine laboratories: Implications for prevention. British Journal of Criminology, 51,
355–374.
Clarke, R. V., & Cornish, D. B. (1985). Modeling offenders’ decisions: A framework for research
and policy. In M. Tonry & N. Morris (Eds.), Crime and justice: An annual review of research
(Vol. 6). Chicago, IL: University of Chicago Press.
Clarke, R. V., & Eck, J. E. (2005). Crime analysis for problem solvers in 60 small steps. Washington,
DC: Office of Community Orientated Policing Services, US Department of Justice.
Clarke, R. V., & Newman, G. R. (2006). Outsmarting the terrorists. Westport, CT: Praeger Security
International.
Clay-Warner, J. (2002). Avoiding rape: The effects of protective actions and situational factors on
rape outcomes. Violence and Victims, 17, 691–705.
Cornish, D. B. (1994). The procedural analysis of offending and its relevance for situational preven-
tion. In R. V. Clarke (Ed.), Crime prevention studies (Vol. 3). Monsey, NY: Criminal Justice Press.
Cornish, D. B., & Clarke, R. V. (1986). The reasoning criminal: Rational choice perspectives on
offending. New York, NY: Springer.
Cornish, D. B., & Clarke, R. V. (2002). Analyzing organized crimes. In A. Piquero & S. Tibbetts
(Eds.), Rational choice and criminal behaviour. London, UK: Routledge.
Cornish, D. B., & Clarke, R. V. (2008). The rational choice perspective. In R. Wortley &
L. Mazerolle (Eds.), Environmental criminology and crime analysis. Cullompton, UK: Willan.
Cusson, M. (1983). Why delinquency? Toronto: University of Toronto Press.
Deslauriers-Varin, N., & Beauregard, E. (2010). Victim’s routine activities and sex offenders’ tar-
get selection scripts: A latent class analysis. Sexual Abuse: A Journal of Research and
Treatment, 22, 315–342.
Felson, M. (2008). Routine activity approach. In R. Wortley & L. Mazerolle (Eds.), Environmental
criminology and crime analysis. Cullompton, UK: Willan.
Gibson, L., Linden, R., & Johnson, S. (1980). A situational theory of rape. Canadian Journal of
Criminology, 22, 51–65.
Hale, R. (1997). Motives of reward among men who rape. American Journal of Criminal Justice,
22, 101–119.
Jones, J., Wynn, B., Kroeze, B., Dunnuck, C., & Rossman, L. (2004). Comparison of sexual
assaults by stranger versus known assailants in a community-based population. American
Journal of Emergency Medicine, 22, 454–459.
Kilpatrick, D. G., Edmunds, C. N., & Seymour, A. K. (1992). Rape in America: A report to the
nation. Arlington, VA: National Victim Center.
Koss, M. P. (1993). Detecting the scope of rape: A review of prevalence research methods. Journal
of Interpersonal Violence, 8, 198–212.
Koss, M. P., Dinero, T. E., & Seibel, C. A. (1988). Stranger and acquaintance rape. Psychology of
Women Quarterly, 12, 1–24.
Lea, S. J., Hunt, L., & Shaw, S. (2011). Sexual assault of older women by strangers. Journal of
Interpersonal Violence, 26, 2303–2320.
Leclerc, B., & Cale, J. (2015). Weapon use and sexual abuse outcomes: A multivariate and con-
junctive analysis of sexual offenses against women. Security Journal, 28, 54–70.
76 Y.-N. Chiu and B. Leclerc

Leclerc, B., Smallbone, S., & Wortley, R. (2013). Interpersonal scripts and victim reaction in child
sexual abuse: A quantitative analysis of the offender-victim interchange. In B. Leclerc &
R. Wortley (Eds.), Cognition and crime: Offender decision-making and script analyses.
London, UK: Routledge.
Leclerc, B., Wortley, R., & Smallbone, S. (2011). Getting into the script of adult child sex offenders
and mapping out situational prevention measures. Journal of Research in Crime and
Delinquency, 48, 209–237.
Levi, M., & Maguire, M. (2004). Reducing and preventing organised crime: An evidence-based
critique. Crime, Law & Social Change, 41, 397–469.
Mann, R. E., & Hollin, C. R. (2007). Sexual offenders’ explanations for their offending. Journal of
Sexual Aggression, 13, 3–9.
Myhill, A., & Allen, J. (2002). Rape and sexual assault of women: The extent and nature of the
problem (Home Office research study, Vol. 237). London, UK: Home Office.
Park, J., Schlesinger, L. B., Pinizzotto, A. J., & Davis, E. F. (2008). Serial and single-victim rapists:
Differences in crime-scene violence, interpersonal involvement, and criminal sophistication.
Behavioral Sciences and the Law, 26, 227–237.
Porter, L. E. (2008). Using archival data and multidimensional scaling to explore leadership:
Examples from group crime. Issues in Forensic Psychology, 8, 33–44.
Porter, L., & Alison, L. (2004). Behavioral coherence in violent group activity: An interpersonal
model of sexually violent gang behavior. Aggressive Behaviour, 30, 449–468.
Stermac, L., Du Mont, J., & Kalemba, V. (1995). Comparisons of sexual assaults by stranger and
known assailants in an urban population of women. Canadian Medical Association Journal,
153, 1089–1094.
Stevens, D. J. (1994). Predatory rapists and victim selection techniques. The Social Science
Journal, 31, 421–433.
Tjaden, P., & Thoennes, N. (2000). Full report of the prevalence, incidence, and consequences of inti-
mate partner violence against women: Findings from the National Violence Against Women Survey.
Washington, DC: National Institute of Justice and Centers for Disease Control and Prevention.
Ullman, S. E., & Knight, R. A. (1991). A multivariate model for predicting rape and physical injury
outcomes during sexual assaults. Journal of Consulting and Clinical Psychology, 59, 724–731.
Ullman, S. E., & Knight, R. A. (1992). Fighting back: Women’s resistance to rape. Journal of
Interpersonal Violence, 7, 31–43.
Ullman, S. E., & Knight, R. A. (1995). Women’s resistance strategies to different rapist types.
Criminal Justice and Behavior, 22, 263–283.
Ullman, S. E., & Siegel, J. M. (1993). Victim-offender relationship and sexual assault. Violence
and Victims, 8, 121–134.
Waldo, G. P., & Dinitz, S. (1967). Personality attributes of the criminal: An analysis of research
studies, 1950–1965. Journal of Research in Crime and Delinquency, 4, 185–202.
Walsh, D. (1978). Shoplifting: Controlling a major crime. London, UK: Macmillan.
Willmer, M. A. P. (1970). Crime and information theory. Edinburgh: Edinburgh University Press.
Woods, L., & Porter, L. (2008). Examining the relationship between sexual offenders and their
victims: Interpersonal differences between stranger and non-stranger sexual offenses. Journal
of Sexual Aggression, 14, 61–75.
Chapter 7
Expanding the Script Analytic Approach
Using Victim Narratives: Learning About
Robberies of Taxi Drivers from the Drivers
Themselves

Martha J. Smith

Introduction

Driving a taxicab can be a difficult and dangerous job. Drivers work long hours, often in
high-stress traffic environments, and face conditions that put them at risk for being
crime victims. Some of these conditions include working alone, at night and in high-
crime areas; picking up strangers—who may be drunk or high on drugs—and taking
them to isolated or unfamiliar places; and carrying cash (see OSHA, 2010). Offenders
may harass them, seek to avoid payment, assault them, rob them, or even kill them.
Burgel, Gillen, and White (2014) looked at taxi driver victimization in a Western city in
the USA and calculated the prevalence of robbery and assault for each year worked: this
was 6 % and 14 %, respectively. In the previous 12 months, 8 % of these drivers reported
they had experienced a robbery, 12 % reported an assault, and 6 % were assaulted with a
weapon. Although relatively rarer over a driver’s career than physical assaults, robberies
and attempted robberies can have serious consequences for victims, such as injuries,
enhanced fear of crime, time off work, and loss of the stolen property.
This chapter looks at taxi driver robbery and attempted robbery victimization, using
the script analytic approach developed by Cornish (1994a, 1994b). This approach can
accommodate more than crime scripts from the offender’s point of view; it can be used
to describe daily activities of individuals (or groups of individuals) and the decisions
and actions required to carry them out. Three main types of scripts are discussed here:
(1) traditional offender crime scripts, (2) potential-victim scripts, and (3) actual-victim
scripts. The scripts set out here are built from information provided by taxi drivers
operating in Cardiff, Wales, and include drivers’ reactions to offenders’ actions during
the unfolding of these incidents. Comments from the taxi driver victims on their reac-
tions and other crime prevention measures are also examined.

M.J. Smith (*)


School of Community Affairs, Wichita State University,
Box 135, 1845 N. Fairmount St., Wichita, KS 67260-0135, USA
e-mail: martha.smith@wichita.edu

© Springer International Publishing Switzerland 2017 77


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_7
78 M.J. Smith

Making Taxi Driving Safer and More Secure

Historically, the taxi industry has been subject to a wide range of regulations about
vehicle characteristics and driver qualifications and practices relating to, for
example:
• The presence of a partition (a shield or screen) separating the driver and the
passengers
• Limits on the number of passengers
• The type of pick-up condition allowed (e.g., on the street, on a rank, or by
pre-booking)
• The use of a meter and standardized fares
• The display of the driver’s identification or license
• Checks on the criminal and driving histories and the knowledge of drivers
• Driver screening of passengers prior to pick up using factors such as race, ethnic-
ity, appearance, or destination
With the exception of the partition, most of these regulations primarily protect
passenger interests.
In the past 15–20 years driver-protection devices, and regulations requiring them,
have become more common. For example, the City of Toronto (2014) requires secu-
rity cameras and emergency lighting in taxis and permits global positioning systems
(GPS). There is evidence that both partitions and cameras can be effective against at
least some types of taxi driver victimizations, such as assault (Stone & Stevens,
2000, looked at partitions) and homicide (Chaumont Menéndez et al., 2013, looked
at both cameras and partitions).
Not all driver protections, however, are related to devices or are likely to be regu-
lated; many depend on the experience and actions of the drivers themselves. Smith
(2005) identified 33 potential measures that could be used to help prevent robberies
against taxi drivers, but only about one-third would be likely to be subject to legisla-
tive oversight. Even where legislation exists, it is not always adhered to by drivers.
For example, drivers often ignore provisions that require them to pick up all pas-
sengers regardless of appearance or destination (see Gambetta & Hamill, 2005, who
studied taxi drivers in Belfast, Northern Ireland, and New York City). Gambetta and
Hamill (2005) examined the behavioral cues drivers used to assess their customers’
trustworthiness (including passenger screening) and explored how effective the
most commonly used techniques were in taxi settings.
The present analysis also looks at driver reactions to dangerous situations, but it
focuses on driver reactions to unfolding crimes as they are occurring—referred to
here as “crimes-in-action.” This type of analysis requires a dynamic tool—the script
analytic approach—which allows researchers to link offender actions to taxi driver
(potential victim) reactions in the sequence in which they occur (see Leclerc,
Smallbone, & Wortley, 2014). With this tool, crime prevention reactions by drivers,
as well as common patterns of crime commission by offenders, can be compared.
Leclerc and his colleagues (2014) used Baldwin’s (1992) theoretical concept of the
7 Expanding the Script Analytic Approach Using Victim Narratives… 79

“interpersonal script” to explore patterns of interaction between adult offenders and


their child victims of sexual abuse at the stage where offenders were seeking to gain
the cooperation of their victims. Here, the reported incidents are grouped together
for the script analysis according to the similarity of the settings and the unfolding
actions of the offenders. Drivers’ actions and reactions are included within these
scripts. Later, these reactions are discussed in terms of their potential for limiting
loss of property and driver injury.

Expanding the Script Analytic Approach

Crime Scripts

The script analytic approach was developed by Cornish (1994a) to describe the
process of crime commission by setting out the full array of actions of the offender
from preparation to completion. Nine stages of crime commission were identified:
preparation, entry, pre-condition, instrumental pre-condition, instrumental initia-
tion, instrumental actualization, doing, post-condition, and exit. Offender scripts
look at the crime in terms of the offender’s actions: they describe each of the sequen-
tial actions performed by the offender as part of the unfolding crime. Cornish’s
script analytic approach also involves linking situational crime prevention (SCP)
measures to each stage of the crime as potential intervention points. For crime pre-
vention purposes, tracks, the lowest and most specific category in Cornish’s hierar-
chy of scripts, are particularly useful to explore due to their crime and situational
specificity. The scripts used here are examples of crime tracks because they detail
specific crimes discussed by victims.

Building Offender Crime Scripts from Victim Accounts

Cornish (1994a) listed a number of ways in which crime scripts could be elicited,
one of which was from victim reports. A more direct way of gaining this informa-
tion, however, is from talking to the victims themselves. Of course, these accounts
are subject to all the usual problems related to the recall of past events, particu-
larly the failure to remember some crime details and the addition of other plausi-
ble details to fill in where knowledge or memory falls short (see discussion in
Porter & Alison, 2004, of problems related to accounts of crime incidents and
Baldwin, 1992, for research about how people “remember” expected details not
included in a presented script). It is also likely for some relatively rare crimes,
such as robbery in particular settings, that the number of victims may be limited.
Nevertheless, using victim narratives to build crime scripts has a number of
advantages: (1) victims may be more willing to talk about the crime event than
offenders, particularly if it was a unique experience for them; (2) they may
80 M.J. Smith

remember the actions in more detail than an offender who may have committed
many other similar crimes; and (3) they may be easier to find or gain access to. In
addition, victims may be able to provide a fine level of detail about the sequencing
of the interactions between themselves and the offender during the crime event.
For example, Guerette and Santana (2010) were able to gather information about
victim self-protective behaviors (VSPBs) before, during, or after one aspect of
reported rape and robbery events (i.e., victim injury) using data from the National
Crime Victimization Survey.
The scripts described in this section are not true “victim” scripts because they do
not use the actions of the victim as the main sequence of actions described. The
victim in these offender scripts is, however, both the source of the information for
building the script and an actor (and reactor) in the unfolding drama.1 “Potential-
victim” and “actual-victim” scripts are, however, described below.

Adding Reactive Tactics, Routine Precautions,


and Other SCP Measures

When examining a dynamic unfolding crime-in-action, it may be useful to distin-


guish three types of situational crime prevention (SCP) measures (see Cornish &
Clarke, 2003) that are available to a potential crime victim: SCP measures that are
in place and operating, routine precautions (Felson & Clarke, 1995; Stanko, 1990),
and what are called here “reactive tactics” (see also Smith, 2004, cited in Tilley,
2005)2. For example, security cameras in taxi vehicles (which may act as a place
manager) or partitions between passengers and drivers (which may harden the tar-
get) are SCP measures that perform their preventive function when in place and
operational. They may act to prevent crime both before and during a crime incident
without any independent actions by the potential victim. Other SCP measures, like
taking routine precautions, such as hiding money (concealing targets) or keeping
the driver’s window closed when stopped (target hardening), may require constant
updating to be in place prior to a potential or unfolding crime as a preventive mea-
sure. For example, money that builds up over a night of driving must be continu-
ously hidden (to conceal it as a target) and windows rolled down for some purpose
must be rolled back up (so the target may be hardened). On the other hand, “reactive
tactics” only come into existence (or manifest themselves) when a crime is in prog-
ress. Calling for assistance from other drivers (to extend guardianship) or not
resisting a robber (to reduce emotional arousal) are preventive measures that are

1
These scripts were called “victim scripts” in an earlier version of this chapter (Smith, 2010). The
current analysis suggests that this term is more usefully applied to two different types of scripts,
potential-victim scripts and actual-victim scripts (see discussion below).
2
For the present discussion, reactive tactics are considered to be a form of SCP. However, they
differ from traditional SCP and routine precautions because reactive tactics involve an identified
offender. Further analysis is needed to clarify the relationships among these types of measures.
7 Expanding the Script Analytic Approach Using Victim Narratives… 81

carried out as a robbery is occurring. Of course, calling for assistance requires the
existence of a radio or telephone to be able to be used in reaction to the crime. While
a crime may be prevented if a potential offender knows that a potential victim has a
telephone, the mechanisms by which the prevention occurs are different in these
two situations. The telephone, in the second situation, while an SCP measure, would
not be considered a reactive tactic. For the present analysis of an unfolding and
dynamic crime-in-action, it is helpful to have terms that can distinguish the different
time frames (and mechanisms) involved with some types of preventive measures. In
addition, it is important to note that information on reactive tactics can only be
adequately gathered from the victims themselves, as it is here, since they are in the
best position to know what offender actions they faced, what they actually did to try
to avoid harm, and why they did it.
Reactive tactics are particularly useful to explore in the context of crimes against
taxi drivers in places where other, more sophisticated and expensive (and perhaps
more effective) SCP measures are not yet widely used or available, or are not
required by taxi companies or government regulations. Even when SCP measures
such as partitions, cameras, and GPS tracking are present and in use, drivers may
still carry out reactive tactics. Their effectiveness needs to be explored.

Potential-Victim Scripts and Actual-Victim Scripts

Scripts can set out other behavioral routines besides crimes because all such activi-
ties involve alternative courses of action, decisions and the actions needed to carry
them out. These are referred to here as “daily activity scripts.” In a sense, daily
activity scripts are always potential-victim scripts, as a crime may occur while
actors are going about their daily activities (see Cohen & Felson, 1979). It is only
when the daily activity scripts become a series of stages that are populated by pos-
sible crime alerts, however, that it becomes useful to consider them as potential-
victim scripts. For taxi drivers, these potential-victim scripts are usually taxi trip
scripts, and these can differ, for example, according to time of day, where the pas-
sengers are picked up, who they are, whether they are drunk or high on drugs, and
where they are going. Each point along the taxi trip carries with it possibilities for
different types of crime to occur. Identifying the risky action points along this trip
script can be useful for helping drivers learn about when they may need to have
prevention measures ready.3
Actual-victim scripts are daily activity scripts in which crimes have occurred.
They can help identify points at which crimes may occur in the future and are used
to construct potential-victim scripts. An example of an actual-victim script is set out
here (see Table 7.5).

3
For example, several types of hypothetical taxi trips, and the attendant crimes that might occur at
each stage of these trips, were discussed with the Cardiff taxi drivers interviewed for this research
during the feedback sessions held in 2002.
82 M.J. Smith

Questions to Be Addressed

Two main questions are addressed in this chapter:


1. What were the reactions of taxi drivers to offender actions during robbery
incidents?
2. What types of crime prevention measures were used in these incidents?

Methods

Study Site

The data analyzed here were collected as part of a 2002 study of how taxi drivers in
Cardiff, Wales, protected themselves against crime. Drivers in Cardiff faced a number
of challenges in that period. The City Centre was the main recreation site, with a num-
ber of large clubs and numerous pubs and restaurants (Thomas & Bromley, 2000),
where alcohol-related violence was often a problem (Maguire, Nettleton, Rix, &
Raybould, 2003). Cardiff had two types of taxi services: (1) hackney cabs that are
licensed to pick up fares off the street (known as “black and white” cabs) and (2)
private-hire cabs that are licensed to pick up pre-booked fares. In 2002, there were
1545 licensed taxi drivers in Cardiff, with around 500 (almost one-third) driving
hackney vehicles.

Sample Selection and Data Collection

The researcher interviewed 83 taxi drivers working in Cardiff. Subjects, who


were identified using a snowball sample of referrals and through direct contact
by the researcher, completed an informed consent form and were paid for their
time.4 The semi-structured interview instrument was based on previous taxi
driver research and on the initial driver interview, with questions added periodi-
cally as the study progressed. The interviews ranged from 50 min to 4 h and
averaged 90 min. Most of the interviews were carried out during the day in the
driver’s cab5 around Cardiff Civic Center, thus providing privacy, convenience
and interviewer safety. Using the cab as the interview setting may have assisted

4
This research was not subject to a formal institutional review board protocol. Nevertheless, the
researcher followed the general format required by US universities in that period. A copy of the
form used is available from the author upon request. Subjects were paid 20 pounds sterling per
hour.
5
Dr. Lesley Noaks suggested this venue to help protect the interviewer’s safety.
7 Expanding the Script Analytic Approach Using Victim Narratives… 83

drivers’ memories of events. All but two of the interviews were recorded and, of
these 81 interviews, five were not transcribed and three were partially
transcribed.6 In Table 7.1, the demographic and taxi-related characteristics for
the sample are set out.

Identifying Robbery Events

A robbery was defined as the taking of property from the driver or the cab through
the threat or use of force. Attempted robberies included any incidents in which the
offender tried to take something from the driver or taxi using actual or threatened
force, but did not succeed. A “near” robbery was an incident where the driver indi-
cated that it was very likely that a robbery had been about to occur but did not or was
abandoned. The interview transcriptions, recordings, or notes were reviewed to
identify any robbery, or “near” robbery, incidents in which the driver provided

Table 7.1 Characteristics of the Cardiff taxi drivers interviewed


All interviewed drivers Victims of completed, attempted,
(N = 83) or “near-miss” robberies (N = 13)a
Ageb Mean = 44.0 years (N = 80) Mean = 46.8 years (N = 12)
Range = 23–63 years Range = 30–63 years
Gender Males = 77 (92.7 %) Males = 12 (92.3 %)
Females = 6 (7.2 %) Females = 1 (7.7 %)
Ethnicity/racec White/Welsh = 61 (73.5 %) White/Welsh = 10 (76.9 %)
Other = 22 (26.5 %) Other = 3 (23.1 %)
Type of taxi vehicle Hackney = 57 (68.9 %) Hackney = 9 (69.2 %)
driven most recentlyd Private hire = 26 (31.3 %) Private hire = 4 (30.8 %)
Years as a Cardiff taxi Mean = 9.7 years Mean = 8.8 years
driver Range = 0.5–30 years Range = 1.33–22 years
a
Includes drivers in the crime-track analysis and one driver who was the victim of two robberies
but did not provide enough information about the crimes to enable them to be scripted
b
Three drivers were not asked their ages. One was a robbery victim
c
Only the first ethnic/racial identification provided is reported. “White/Welsh” also includes
British, Celtic, Church of England, European, and Spaniard. “Other” includes African, Black
African, Caribbean, East African, Gypsy, Hindu, Indian, Iraqi, Jewish, Pakistani, Sikh, and
Somali
d
Four drivers were inactive (Private hire = 2, Hackney = 2). None was a robbery victim

6
Notes were taken at the first interview. A low battery prevented another interview from being
recorded. Extensive notes were made when this was discovered, and crime incidents were coded.
Length of the interviews and language difficulties prevented some interviews from being tran-
scribed. All of the interviews that were not transcribed were coded for crime victimization
information.
84 M.J. Smith

enough details to build at least a partial script.7 This process resulted in the identifi-
cation of four completed robberies, eight attempted robberies, and one near miss.
Two of these incidents (one completed and one attempted robbery) involved a single
driver. Table 7.1 shows that robbery victims were very similar to the wider group of
83 drivers.

Building Crime and Daily Activity Tracks

Cornish’s (1994b) nine-stage template was used for coding all the crime incidents
as “individual tracks.” After the individual tracks were coded, robbery incidents
were grouped as “crime tracks.” During this stage of the analysis, the incidents were
examined to see if the offender approaches to street robbery victims identified by
Smith (2003)—confrontations (immediate demands), blitzes (immediate violence),
cons (distractions), or snatches (quick taking of property)—could be applied to the
taxi settings described here. These categories were not found to be directly appli-
cable although some of these actions did occur in the course of the taxi robberies.
The approaches used by offenders in a taxi robbery may have been different because,
unlike street robbers, all but two of the offenders here spent time in the cabs as pas-
sengers prior to the “doing” stage of the incident. Nevertheless, these approaches
did highlight the importance of grouping individual tracks together by considering
both the sequence in which the actions needed for a robbery (a demand for, or taking
of, property and the threat or use of force) occurred and the nature of these actions.

Results

Four tracks of taxi driver robbery incidents were identified (see Tables 7.2, 7.3, 7.4,
and 7.5). Tables 7.2, 7.3, and 7.4 contain offender tracks built from victim narra-
tives. Table 7.5 is an actual-victim track also based on accounts provided by victims.
In these tables and in the text, a driver’s exact wording is indicated with quotation
marks and a code identifying the driver (e.g., Robbery Victim #4 is “[RV4]”).
Although not always specifically stated by the drivers, it appears that all but one of
the scripted incidents involved a pickup from a pub or club, or late night from the
city center.

7
One driver, who was the victim of two attempted robberies, did not provide enough information
about either incident to build a robbery script. He did provide information about the reactive tactics
he used: (1) he “pushed one offender out of the car”; (2) “the other one had gone out of the car and
come round to this side of the car, like … and I just drove away with him hanging half in and half
out of the car and I finally pushed him out as well” [RV13]. His demographic information is
included in Table 7.1.
7 Expanding the Script Analytic Approach Using Victim Narratives… 85

Table 7.2 Individual track of a “near-miss” robbery—initially friendly approach—female driver


(N = 1)a
Driver reactions and reactive
Scene/function Stages in the event tactics
PREPARATION
ENTRY One man got in.
PRE-CONDITION Initially said he was going to a
northern part of Cardiff [but later
changed destination].
INSTRUMENTAL Got to a block of flats.
PRE-CONDITION Owed £8.10.
INSTRUMENTAL NON-PAYMENT SCRIPT: “‘I’m not stupid. This is
INITIATION “‘I’m just going in, love, to good-bye, isn’t it?’” [RV4]
get the money.’” [RV4]
“‘No, come to the door with “Oh, it might be genuine, so I
me if you like.’” [RV4] went to the door.” [RV4]
The area opens up into a DID NOT GO TO ISOLATED
passageway. “‘You can come up PLACE:
here with me if you like.’” [RV4] “‘No, it’s alright. I’ll wait here.’”
[RV4]
INSTRUMENTAL BECOMES ROBBERY SCRIPT: CONCEALED MONEY:
ACTUALIZATION “‘Have you got your moneybag in “‘No, I haven’t got one. I just
the car?’” [RV4] came out. What money I’ve got
is in my pocket, just my float.’”
[RV4]
“‘Oh, alright then.’” [RV4]
DOING
POST-CONDITION
EXIT Ran through the passageway and
out through the other side.
a
This track is an offender script based on an account provided by a single victim
Note: A driver’s exact wording is indicated with quotation marks and a code identifying the driver
(e.g., Robbery Victim #4 is “[RV4]”). Summary remarks (shown in all capital letters) highlight the
reactive tactics used

Table 7.2 presents an individual track involving a “near-miss” robbery of a


female driver. In that track, the offender (who may have been a “runner”—the term
used in Cardiff for non-payers) conned the driver into leaving the car and then asked
where her money bag was. The driver, who later said she should not have left the
car, quickly reacted by saying she had just come out (meaning she had little money
to be stolen). The offender then left through a passageway without demanding the
money and without injuring the driver.
Table 7.3 sets out an offender track built from descriptions of two events. In both
incidents, a group of male passengers made intimidating remarks to the driver, sug-
gesting they would beat him up or mug him when he stopped. In this track, both
drivers reacted by trying to get to safety, in one case by attracting police attention.
The police directed the driver in the first incident to move his cab so it was not
blocking the road and then let the offenders go. One of the offenders in the second
86

Table 7.3 Robbery track—intimidating threats used (N = 2; completed = 1, attempted = 1)


Driver reactions or reactive tactics (taken or
Scene/function Stages in crime mentioned)
PREPARATION
ENTRY Picked up 3 men at rank in center of town (1st incident).
Picked up 3 “skinheads” at a nightclub (2nd incident)
PRE-CONDITION 2nd incident: Gave false destination [Were actually going to a
remote area].
INSTRUMENTAL
PRE-CONDITION
INSTRUMENTAL 1st incident: “When the journey commenced they basically said GOT TO SAFETY
INITIATION they were going to turn me upside down and shake me like a money ATTRACTED POLICE ATTENTION:
box … Wanted to beat me up or mug me.” [RV3] “Deliberately went through red lights to get
back to town … to catch the police’s attention.”
[RV3]
Blocked the road [in an unspecified way].
2nd incident: Used profanity. GOT TO SAFETY:
“No CB. No radio. Get to garage. They know
me there.” Said: “I want fuel.” [RV12]
Turned to get to petrol station. Got out of car.
M.J. Smith
7

INSTRUMENTAL 1st incident: “The 3 actually stood out of the car and they wanted STAYED IN CAR
ACTUALIZATION to beat me up.” [RV3]
1st incident: The police came up—“attitude to me was get out Drove away.
of here and he told me to leave the scene.” [RV3]
2nd incident: Physically attacked by two. Chased driver. All GOT TO SAFETY:
passengers got out of the car. Got inside.
“Man opened gate quick. I got in there. Locked
the door.” [RV12]
DOING 2nd incident: One offender went to taxi and stole something. “I haven’t seen yet [2 months later] what they
take.” [RV12]
POST-CONDITION 1st incident: Looked in the mirror and saw what
happened. [Police let them go. No charges.]
“It’s very scary.” [RV3]
EXIT 2nd incident: They left. Did not call the police.
This track is an offender script based on accounts provided by victims
Note: A driver’s exact wording is indicated with quotation marks and a code identifying the driver (e.g., Robbery Victim #4 is “[RV4]”). Summary remarks
(shown in all capital letters) highlight the reactive tactics used
Expanding the Script Analytic Approach Using Victim Narratives…
87
88

Table 7.4 Robbery track—initially friendly offender—attack with little/no warning (N = 8; completed = 3, attempted = 5)
Scene/function Stages in the event Driver reactions or reactive tactics (taken or mentioned)
PREPARATION Incident 8: Passenger had 15 pints prior to pick up.
ENTRY Incidents 1, 2, 3, 4, 5, 6, 7: Picked up 2 or 3 passengers.
Incident 8: Single man was well mannered, well dressed.
PRE-CONDITION Incidents 1, 2, 3, 5: Going to an area or estate known for One driver asked for money up front.
criminal activity.
Incident 7: Passengers gave the driver the fare.
Incident 8: Going to a distant destination. Driver did not ask for money up front.
INSTRUMENTAL Incidents 2, 5: One passenger sat in front with at least one
PRE-CONDITION in the back.
Incidents 1, 2, 3: Conversation was friendly, not hostile.
INSTRUMENTAL Incidents 1, 4: Dropped off other passenger(s).
INITIATION Incident 8: Stopped on the way and he “came back as a different Would have flashed his lights for the police, if he had seen
person” [had been pleasant initially] [RV2] them.
Incident 8: Asked to stop in isolated dead end. Needed to go Driver got out of the car as well.
in house to get money for the fare. Got out of the car.
INSTRUMENTAL Incidents 3, 8: Passenger (in the front or only) said “Give me
ACTUALIZATION your money.”
Incidents 1, 2, 3, 4, 5, 6: Passenger in the back used the physical
force or showed a knife.
Incidents 1, 4, 8: Attacker grabbed around the driver’s neck.
Incidents 1, 2, 5: Attacker had a knife. (In one incident the DENIED BENEFITS:
passenger took the driver’s hand before putting a knife to his Incident 5: Told attacker he knew his relative and was able
throat.) to prove it.
Driver noted he could not call because his phone was out
of range. He had no shield and no alarms.
M.J. Smith
7

DOING Incident 1: Attacker shoved driver. COMPLIED WITH ROBBER:


Incident 2: Took the money out of driver’s pocket (£ 20 note). Incidents 1, 2 3: Complied and did not use force in return.
Incident 3: Passenger held the driver from behind. CONCEALED BULK OF MONEY:
Incident 1: Gave them a bag with lower denomination bills
and change.
Incident 4: Asked for money. DENIED BENEFITS:
Fought them off. Punched him. Shouted at him.
Incident 6: Tried to take money out of driver’s pocket. Unclear what the driver’s response was, but this was an
attempted robbery.
Incident 7: Passenger paid first and then said he had a gun GOT TO SAFETY:
and demanded money. Driver pulled around (with the would-be robber hanging
onto the car) and drove to the police station.
Incident 8: Passenger pushed him to the ground. [Dropped credit
card and other ID.]
POST-CONDITION Incidents 1, 3, 5, 6, 8: Attackers ran away. Incident 1: Pushed the panic button. Waited for the police.
Incident 2: Did not run after (followed advice given by
taxi-driver relative). Called the police, but they could not
find attacker.
EXIT Incident 8: Driver chased attacker. Went to the police.
[Attacker later paid fare and court fees.]
(LATER) Overall, all but two drivers called the police or reported the
crime at the station.
Only one person was apprehended.
Expanding the Script Analytic Approach Using Victim Narratives…

This track is an offender script based on accounts provided by victims


Note: A driver’s exact wording is indicated with quotation marks and a code identifying the driver (e.g., Robbery Victim #4 is “[RV4]”)
Summary remarks (shown in all capital letters) highlight the reactive tactics used
89
Table 7.5 Track of taxi trip to drug sales area with overlapping robbery by person outside the taxi (N = 2; completed = 1, attempted = 1)
Stages of a taxi trip Actions of the passenger Actions of third-party offender Driver reactions or reactive tactics
Ready for a fare
Pickup location
Passenger(s) features
Entry into car One man got in.
Destination Went to a known drug area. [1st incident: Driver picked up someone he knew.]
[2nd incident: Driver was not told exact destination.]
Atmosphere in car
Change in destination or route
Interim stop(s)/driver waiting Went into a residence. 1st incident: Man approached and asked “So I gave him a light.” [RV8]
in the car for a light. [Did not see other man.]
1st incident: Man opened driver’s door.
Both incidents: Knife was shown. COMPLIED WITH ROBBER And CONCEALED
2nd incident: Man asked for money. BULK OF MONEY:
1st incident: Driver had hidden most of his earnings.
Gave him what was in his pocket.
2nd incident: Driver said he had just come out and
gave him his bag of change with some notes.
1st incident: Man stuck hand through Got out of the car, but the men were gone.
open window, attempted to remove key, [Later, man with knife plead guilty.]
but it broke off.
2nd incident: Came out, hit the 2nd incident: “Off we went.” [RV5]
robber, grabbed the money, and
got back in the car.
Stop at destination
Payment of fare
Exit vehicle
This track is an actual-victim script based on accounts provided by victims
Note: A driver’s exact wording is indicated with quotation marks and a code identifying the driver (e.g., Robbery Victim #4 is “[RV4]”). Summary remarks (shown in
all capital letters) highlight the reactive tactics used
7 Expanding the Script Analytic Approach Using Victim Narratives… 91

incident took something from the cab once the driver had fled to an open business.
Neither driver in this track was injured.
In Table 7.4, a single robbery track was built from eight robbery incidents in
which the passengers were initially congenial and not intimidating. For seven of
these incidents, the attack or demand came while the driver was in the car. For these
seven incidents, the driver had no warning that a robbery was unfolding until (at the
initial actualization) the passenger in the back of the car (in five cases) grabbed him
around the neck (three cases), physically held him (one case), put a knife to his
throat (one case), or took his hand (one case). In two of these cases, the assailant
displayed a knife right after the initial actualization of the crime. The other two rob-
bery events involved the passenger in front demanding money initially (one case)
and the passenger, after paying the fare, claiming to have a gun (one case). The
driver reported that an offender verbally demanded money in only two of the eight
cases in this track.
Seven of the eight drivers in this track reported using reactive tactics. Only three
incidents involved completed robberies: in one, the driver handed his change bag
over to the robbers; in another, the offenders took 20 pounds from a driver’s pocket;
and in the last incident, it was not clear how the robbers got the driver’s money. In
four other incidents, the driver reported reactive tactics other than compliance. One
driver, who had been grabbed around his neck, fought back and shouted. No knife
was used in that incident and he was the only driver who reported being injured
(later having a stiff neck). A second driver who had a knife at his throat talked his
way out of the robbery by telling the offender he knew one of his relatives. In a third
noncompliance incident, in which the passenger (offender) paid the fare and then
produced a gun and asked for money, the driver was able to move his car, with the
offender holding on, and drive to the police station. In the last such incident, the
driver’s taxi had both a partition and GPS tracking, unlike any of the other drivers
in this track, but he was outside his vehicle when the passenger asked for his money,
grabbed him, and pushed him to the ground. The passenger then dropped cards
identifying him at the scene and fled. This case went to court, with the passenger
reportedly paying his fare and court costs.8 Drivers in five of the other incidents
described in this track also contacted the police, but either no one was found or they
heard no more from the police after the event.
Table 7.5 uses a different type of script to set out the two robbery events in which
drivers were attacked from outside the car by a third party while waiting for their
passenger to return. This is an actual-victim script that uses a taxi trip as the orga-
nizing daily activity to frame the other actions. At the stage when the driver was

8
This incident (Incident 8) was initially similar to the incidents in this track with its non-threaten-
ing passenger; yet, it differs from the other seven incidents in a number of respects. For example,
the demand for money and the physical attack by the passenger occurred outside of the taxi and
may have been part of a non-payment offense in which the driver exited his vehicle in pursuit of
the fare, with the passenger’s demand for money and use of force occurring either as part of that
crime or as an attempt to gain even more from the situation, perhaps as an afterthought. When an
incident deviates from other incidents in a track, it may be useful to classify it as a separate track.
It was not separated here in order to illustrate this point.
92 M.J. Smith

waiting for the passenger, an offender approached the car and a robbery event
unfolded. In one incident, the driver was approached for a light by a third party who
then produced a knife, got the money, and attempted to remove the car key. The
driver reacted by complying with the robber, which resulted in a completed robbery,
but he provided only part of his takings. In the other incident, the driver also reacted
by complying with the robber’s demands when a knife was displayed, and he too
was able to conceal the bulk of his takings for the night, by claiming he had just
started his shift. The returning passenger hit the robber, took back the money bag,
and gave it to the driver before the taxi left the scene. Because the victim got his
money back, this was classified as an attempted robbery.

Discussion

The crime tracks built from the taxi driver narratives were useful for showing the
dynamic nature of the crimes-in-action and for identifying the reactive tactics used
by drivers. They are the type of “interpersonal script” (Leclerc et al., 2014) that can
be used to evaluate the interactions of offenders and victims in taxi driver robbery
incidents. For example, in looking at the unfolding crimes, as a robbery got closer
to the “doing” stage (at either the instrumental initiation or instrumental actualiza-
tion stage), the reactive tactics focused on avoiding two potential outcomes—i.e.,
loss of money and harm to the driver, the two elements central to the definition of a
robbery. One successful strategy for protecting the driver’s money involved hiding
or concealing money and convincing the offender that there was no more money
available (concealing the target). Similarly, drivers who were able to get away from
the offenders avoided further harm (target removal), as did one who did not go to an
isolated place (thereby improving natural surveillance possibilities). In addition,
staying in a secured car when the offenders were outside (target hardening) and not
resisting the robber (reducing emotional arousal) may have protected drivers from
injury. On the other hand, trying to attract police attention (strengthening formal
surveillance) may have been a useful tactic for driver protection in one case, but the
police officer who responded was not supportive of the driver and quickly let the
intimidating passengers go. Similarly, relying on physically forceful reactive tactics
may have been useful for one of the drivers included here but is problematic.
Pushing and hitting tactics can be difficult to perform, may not prevent the robber
from gaining the money or doing further harm to the driver (particularly if a weapon
is present), and may lead to complaints about taxi-driver roughness. While the
driver who responded with force was the only driver who reported an injury, it is
unclear, even using the interactive script approach, whether his injury would have
been less severe if he had not resisted. It is also unclear whether the driver would
have responded with force if a knife had been present, an issue that was not addressed
in the interview.
It is not possible from the information gathered from these victims to know just
what prevention measures would have kept these drivers from becoming victims. It
7 Expanding the Script Analytic Approach Using Victim Narratives… 93

is possible, however, to note what the drivers said in response to these incidents and
to discuss briefly how prevention measures available now for taxis might have been
used in these robbery events. For example, drivers, who commented on vehicle
partitions (target hardening), were divided in their views on whether they would
have liked to have one in their cabs. Two said they would and one said it might have
prevented it (“it’s that big word ‘if’” [RV11]). Two other drivers did not want parti-
tions because passengers (presumably “runners”) can do damage if locked into
London cabs (which contain both partitions and automatic door locks) or because
passengers would be put off by the presence of a partition. In the UK, partitions are
found in London cabs, which are expensive to buy or rent, and cannot always be
easily fitted to other cars; one driver noted he had checked to try to get a partition
fitted but could not. Just as a partition did not protect one of the drivers who got out
of his vehicle and was then hit by his passenger (in Table 7.4), partitions might not
have protected the two drivers approached by non-passengers (in Table 7.5) or the
woman who got out of her car to try to collect her fare (in Table 7.2). These driver
accounts help explain why partitions are resisted by some drivers and when they
might not be effective. Cameras (strengthen formal surveillance) and GPS tracking
devices (extend guardianship) were not commonly available in the UK during this
period. One cab company in Cardiff had GPS tracking at that time, but it was used
primarily for dispatching drivers to jobs. Like partitions, these devices hold promise
for preventing some robberies (cameras) or assisting responders (GPS tracking), but
they are expensive and are not likely to be installed by drivers unless required.
Most of the actual changes drivers described making following these incidents
related to practices rather than equipment although one driver said that he got a cell
phone (extending guardianship) after the incident. Following the robbery or near
robbery incidents, drivers reported that they
• Became more aware or “more judgmental” [RV3]
• Operated on a different side of town or did not take fares to the part of town
where the incident occurred (“even if they look honest enough” [RV3])
• Screened potential passengers by type of person and number late at night
• Afterwards always got the money up front (at the beginning of the trip)
• Never said how long they had been out
• Stopped working late nights (at least temporarily or less often)
• Hid their money
• Carried something (that could be used for self-protection in an attack)
• Had less faith in the police
Some of these types of reactions to actual or potential victimization (e.g., pas-
senger screening practices) may make it difficult for many late-night riders to get
cabs, causing resentment against drivers, or may lead, in some cases, to sanctions
by taxi regulators.
To place these preventive measures within a broader theoretical context, it is use-
ful to classify them using the 25-techniques table developed for SCP (Cornish &
Clarke, 2003) (see Table 7.6). This table draws upon research on taxi protections
used by drivers and summarized in Smith (2005), as well as some of the techniques
Table 7.6 Possible situational crime prevention techniques to prevent robberies of taxi drivers or limit their negative effects
94

Increase the effort Increase the risks Reduce the rewards Reduce provocations Remove excuses
1. Target harden 6. Extend guardianship 11. Conceal targets 16. Reduce frustrations 21. Set rules
• Safety shields (A) • GPS system and alarm (A) • Limit change to small bills (A) and Stress • Regulate the
• Keep windows • Info. on 24-h locations (A) • Never flash cash (D) • Distribute rules at industry with
rolled up (D) • Radio, alarm code, or “open mike” (C) • “I just started.” “It’s been key supervised crime
• Stay inside • Keep dispatch informed (C) slow.” “Just ticking over.” (D) ranks (R) prevention in
cab—unless • CB radio and cell phone (D) • Hide cash or keep it in more mind (R)
escaping attack from • Use buddy system (C) (D) than one place (D)
within (D)
2. Control access to 7. Assist natural surveillance 12. Remove targets 17. Avoid disputes 22. Post
facilities • Good lighting in cab (A) • Use farecards (R) • Require knowledge instructions
• Limit # of • Ranks near 24-h stores or other • Decal re: limited money (A) tests (R) • List rules,
passengers (R) locations (R) • Allow credit or debit cards (C) • Ask for money up regulations and
• Pre-book all bar • Keep windows clear (A) (D) front (A) fares in
pickups (C) • Use street lighting well/Avoid dark • Use safe or drop off money (C) passenger area
• Screen passengers places (D) (D) of taxi (A)
(C) (D) • Travel main routes/Avoid alleys and • Look for escape opportunity (D)
• Central door locks dead ends (D)
(D)
3. Screen exits 8. Reduce anonymity 13. Identify property 18. Reduce emotional 23. Alert
• Ask “rowdies” to • Keep passenger trip book (A) • Use cab-locator system (GPS) arousal conscience
get out (D) • Caller ID (C) (A) • Exclude violent • Post-notices re:
• Have regular riders (C) (D) drivers (vet) (R) community
• Control the industry awareness and
to reduce driver “cab safe”
abuses (R) programs at
• Do not resist robber ranks and in
(C) (D) cabs (R)
M.J. Smith
7

Increase the effort Increase the risks Reduce the rewards Reduce provocations Remove excuses
4. Deflect offenders 9. Utilize place managers 14. Disrupt markets 19. Neutralize peer 24. Assist
• Limit seating • Security cameras in cab (A) pressure compliance
options (R) (D) • Supervise ranks (R) • Ask troublemakers • Use multiple
• Get destination up • Second person up front with driver in a group to get out payment
front (R) (D) (A) (D) systems (A)
5. Control tools/ 10. Strengthen formal surveillance 15. Deny benefits 20. Discourage 25. Control drugs
weapons • Attract police attention (D) • Notice culprit’s description (D) imitation and alcohol
• Don’t carry a • CCTV at ranks (R) • Resist attacker (D) • Censor details of • Increase
weapon (A) • Trouble lights on cabs (R) • Disable-vehicle capacity (C) modus operandi in passenger
• Authorized police stops (A) (D) press releases and sobriety prior to
• Decoy police (as drivers) (R) • Inside trunk release (C) (D) interviews (A) pick up (R)
• First-aid kit (A) • Keep crime • Prohibit drug or
prevention tips alcohol
during training “in consumption in
house” (A) cab (A)
Adapted from Smith (2004), which was reproduced in Tilley (2005)
Note: Sources controlling implementation of measure: D = Driver (reactive tactics used by drivers in these robbery incidents, in italics); R = Regulator/Police/
City; C = Company/Owner; and A = Any Source: Driver/Company/Owner/Regulator/Police/City (if all likely to be needed, in italics)
Expanding the Script Analytic Approach Using Victim Narratives…
95
96 M.J. Smith

discussed by Cardiff drivers in the larger 2002 study. It also classifies techniques
that the drivers reported using during, and after, the robbery (or near robbery) inci-
dents reported here (which appear in italics). After each included measure, there is
also a notation for the type of party that might be in the best position to implement
this measure—the driver (D), the vehicle owner or dispatch company (C), the taxi
regulator, police, or the local governmental authority (the Council) (R), or any of
these parties (A). These parties may be encouraged to use (or discouraged from
using) these SCP measures by the types of super controllers described by Sampson,
Eck, and Dunham (2010), including organizational, contractual, regulatory, and
market incentives. Further analysis is needed. Nevertheless, this table clearly shows
that there is a need for involvement by a variety of parties to help make the taxi
environment safe for drivers.

Conclusions

The script layout allowed easy comparisons of the unfolding sequences of these
robberies to be made across individual tracks (involving one crime event) and
crime tracks (involving two or more crime events). The results provide support
for using interpersonal scripts, which link offender actions and driver reactions,
for training new drivers. This could help drivers consider a wider variety of poten-
tial cues for discerning the unfolding of a larger number of crime scripts and
assess the effects of using various potential reactive tactics in response to these
cues. In terms of these robberies, one driver said that he was surprised by what
occurred—e.g., his view that pickups from addresses were safe was challenged—
while another referred to advice he had been given from a relative who was a
driver about how to handle a robbery. Having formal training by experienced driv-
ers who discuss unfolding crime scripts could be one way of spreading “practice”
wisdom gained the hard way. Because, among the 83 drivers interviewed, com-
plaints about the lack of support for taxi drivers among police and the local gov-
ernment regulators of taxis were common, it might also be very useful for these
groups to meet and discuss these problems using scripts based on their experi-
ences. They could explore patterns of police or regulator lack of action (as in
Table 7.3, where a police officer would not let the driver explain why he blocked
traffic and then let the intimidating passengers go without holding them), as well
as situations in which taxi drivers may be asking for responses from police and the
local authority that cannot be met.
This research was limited by the small number of robbery events available from
this group of taxi drivers, which may be a good sign for driver safety in Cardiff, but
may reduce the applicability of the findings to another setting, with a wider variety
of robberies than were described here. The research would also have benefited from
more detailed information about the robbery events themselves. Part of this paucity
of information may relate to the nature of robbery events, particularly when seen
from the perspective of the victim. Robberies of taxi drivers by their passengers may
7 Expanding the Script Analytic Approach Using Victim Narratives… 97

not involve (or perhaps require) a great deal of initiating interaction between the
offender and victim. In this respect taxi driver robberies appear to be like street rob-
beries (Smith, 2003), with victims still taken by surprise despite the general atten-
tiveness of drivers to passenger cues for inappropriate or criminal behaviors (see
Gambetta & Hamill, 2005). In addition, robberies may, in general, involve less
offender-victim interaction than other crimes. For example, assaults, such as those
that arise from a dispute or an unsuccessful con (such as a “runner” incident), or a
sexual assault in which the victim is “groomed” (e.g., Leclerc, Wortley, & Smallbone,
2011), may yield more details and cues at each stage. A comparison of robbery and
assault scripts using a single group of victims, however, would help answer whether
this lack of detail results from the robbery process itself or from untapped knowl-
edge held by victims.
Future researchers should be able to gather more information about crimes
against taxi drivers for script building if they systematically ask victims about (1)
the type of vehicle the driver was using at the time of the incident, (2) where and
when the pick-up and drop-off occurred, (3) the condition of the passengers, and (4)
interactions among passengers and between passengers and the driver. It may also
be helpful for victims to be presented with an empty script template that they can
complete with the assistance of the interviewer, as occurred with the calendars used
in Horney and Marshall (1991) to find out more about an offender’s crime commis-
sion and lifestyle. Detailed time and place information would allow researchers to
build maps and compare the information to drivers’ perceptions so that risky places
can be identified and made safer.

Acknowledgements This chapter is based on research supported by a grant from Safer Cardiff
Ltd. It represents the views of the author and does not necessarily represent the views of Safer
Cardiff Ltd., Cardiff University, Wichita State University, or any of the individual taxi drivers
interviewed. The author is grateful to Dave A’Hern, Barbara Natasegara MBE, Bernadette Bowen-
Thomson, Sally Chahrour-Jackson, Elizabeth Renton, Rosemary Howls, Paul Shone of Cardiff
County Council, Inspector Tony Rees of South Wales Police, Terry Evans and Mr. Korash of the
Hackney Cab Association, my former colleagues in the School of Social Sciences at Cardiff
University, the dedicated group of transcribers, and all of the taxi drivers.

References

Baldwin, M. W. (1992). Relational schemas and the processing of social information. Psychological
Bulletin, 112(3), 461–484.
Burgel, B. J., Gillen, M., & White, M. C. (2014). Work-related violence experienced by urban taxi
drivers. American Journal of Industrial Medicine, 57(12), 1377–1385. doi:10.1002/ajim.22397.
Chaumont Menéndez, C., Amandus, H., Damadi, P., Wu, N., Konda, S., & Hendricks, S. (2013).
Effectiveness of taxicab security equipment in reducing driver homicide rates. American
Journal of Preventive Medicine, 45(1), 1–8.
City of Toronto. (2014). Toronto Municipal Code, Chapter 545, Article VIII, §545-142, B. and
C. Retrieved April 22, 2015, form http://www.toronto.ca/legdocs/municode/1184_545.pdf
Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 44(4), 588–608.
98 M.J. Smith

Cornish, D. B. (1994a). Crimes as scripts. In D. Zahm & P. Cromwell (Eds.), Proceedings of the
International Seminar on Environmental Criminology and Crime Analysis (pp. 30–45). Coral
Gables, FL: Florida Criminal Justice Executive Institute, University of Miami.
Cornish, D. B. (1994b). The procedural analysis of offending and its relevance for situational
crime prevention. Crime Prevention Studies, 3, 151–198.
Cornish, D. B., & Clarke, R. V. (2003). Opportunities, precipitators and criminal dispositions: A reply
to Wortley’s critique of situational crime prevention. Crime Prevention Studies, 16, 41–96.
Felson, M., & Clarke, R. V. (1995). Routine precautions, criminology, and crime prevention. In
H. D. Barlow (Ed.), Crime and public policy: Putting theory to work (pp. 179–190). Boulder,
CO: Westview Press.
Gambetta, D., & Hamill, H. (2005). Streetwise: How taxi drivers establish their customers’ trust-
worthiness. New York, NY: Russell Sage.
Guerette, R. T., & Santana, S. A. (2010). Explaining victim self-protective behavior effects on
crime incident outcomes: A test of opportunity theory. Crime and Delinquency, 56(2), 198–
226. doi:10.1177/0011128707311644.
Horney, J., & Marshall, I. H. (1991). Measuring lambda through self-reports. Criminology, 29(3),
471–495.
Leclerc, B., Smallbone, S., & Wortley, R. (2014). Interpersonal scripts and victim reaction in child
sexual abuse: A quantitative analysis of the victim-offender interchange. In B. Leclerc &
R. Wortley (Eds.), Cognition and crime: Offender decision making and script analysis
(pp. 101–119). London: Routledge.
Leclerc, B., Wortley, R., & Smallbone, S. (2011). Getting into the script of adult child sex offend-
ers and mapping out situational prevention measures. Journal of Research in Crime and
Delinquency, 48(2), 209–237.
Maguire, M., Nettleton, H., Rix, A., & Raybould, S. (2003). Reducing alcohol-related violence
and disorder: An evaluation of the ‘TASC’ Project (Home Office Research study, Vol. 265).
London: Home Office Research, Development and Statistics Directorate.
Occupational Safety and Health Administration (OSHA). (2010). OSHA fact sheet: Preventing
violence against taxi and for-hire drivers. Retrieved April 22, 2015, from http://www.osha.gov/
Publications/taxi-driver-violencefactsheet.pdf
Porter, L. E., & Alison, L. J. (2004). Behavioural coherence in violent activity: An interpersonal model
of sexually violent gang behavior. Aggressive Behavior, 30(6), 449–468. doi:10.1002/ab.20047.
Sampson, R., Eck, J. E., & Dunham, J. (2010). Super controllers and crime prevention: A routine
activity explanation of crime prevention success and failure. Security Journal, 23(1), 37–51.
doi:10.1057/sj.2009.17.
Smith, J. (2003). The nature of personal robbery (Home Office Research study, Vol. 254). London:
Home Office Development and Statistics Directorate.
Smith, M. J. (2004). Routine precautions used by taxi drivers: A situational crime prevention
approach. Paper presented at the American Society of Criminology meeting, Nashville,
Tennessee, November.
Smith, M. J. (2005). Robbery of taxi drivers (Problem-oriented policing guide, Vol. 34).
Washington, DC: US Department of Justice, Office of Community Oriented Policing Services.
Smith, M. J. (2010). Constructing crime scripts from victim and potential-victim narratives. Paper
presented at the American Society of Criminology meeting, San Francisco, California, November.
Stanko, E. (1990). Everyday violence: How women and men experience sexual and physical dan-
ger. London: Pandora.
Stone, J. R., & Stevens, D. C. (2000). Effectiveness of taxi partitions: Baltimore, Maryland, case
study. Transportation Research Record, 1731, 71–78. doi:10.3141/1731-09.
Thomas, C. J., & Bromley, R. D. F. (2000). City-centre revitalization: Problems of fragmentation
and fear in the evening and night-time city. Urban Studies, 37(8), 1403–1429.
doi:10.1080/00420980020080181.
Tilley, N. (2005). Crime prevention and system design. In N. Tilley (Ed.), Handbook of crime
prevention and community safety (pp. 266–293). Cullompton, UK: Willan.
Part III
Crime Mapping and Spatial Analysis
Chapter 8
Crime Mapping and Spatial Analysis

Michael Townsley

What Is Spatial Analysis and Where Does It Come from?

It is important to stress that spatial approaches to crime prevention are comprised of


two parts: identifying the pattern or spatial concentration (“hotspots") and taking
action against the problem (“policing"). To use a medical analogy, patients need to
be first diagnosed before they can be treated. In some cases a reprieve from symp-
toms may be sufficient for recovery, but often a detailed understanding is required
for treatment to be effective. In the case of crime prevention it is possible to simply
increase the police presence in hotspot areas and hope that this is enough to deter
offenders in the aggregate. But, as I will argue below, this is a short-term and often
detrimental approach. In the classic primary-secondary-tertiary classification
(Brantingham & Faust, 1976), the approach described here is a mix of secondary
and tertiary crime prevention.
The theoretical foundations of this approach are rooted in environmental crimi-
nology (Wortley & Mazerolle, 2008). This school of thought differs from others in
criminology in its focus on crime events and in the role that opportunities in the
immediate environment play in influencing criminal decision-making. It focuses on
the where, when and how of crimes and is largely unconcerned with the why of
crimes. Environmental criminology is a triumvirate of approaches: (1) rational
choice theory (Cornish & Clarke, 1986, 2008) explains how offenders decide
whether or which criminal opportunity to exploit in a given environment; (2) routine
activity theory (Cohen & Felson, 1979; Felson, 2008) explains the circumstances in
which offenders and victims come into contact in space and time and (3) geometry
of crime theory (Brantingham & Brantingham, 1981) describes offenders' spatial
awareness and hunting patterns.

M. Townsley (*)
School of Criminology and Criminal Justice and Griffith Criminology Institute,
Griffith University, Mt Gravatt Campus, Nathan, QLD 4111, Australia
e-mail: m.townsley@griffith.edu.au

© Springer International Publishing Switzerland 2017 101


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_8
102 M. Townsley

Rational Choice Theory

Rational choice theory (Cornish & Clarke, 1986, 2008) assumes that offender
decision-making is similar in kind to decision-making in other contexts. Decision-
makers are assumed to be utility maximisers, prefering to use less energy or time to
achieve a given goal than alternatives. Cornish and Clarke assume that offenders
weigh up the costs and benefits of exploiting criminal opportunities; if the benefits
outweigh the costs, then that opportunity is likely to be exploited. For example, is
acquisitive crime the benefits usually translate into the perceived economic yield
(the items stolen during shoplifting, say) and the costs are usually partitioned into
the percieved risks of apprehension and the percieved effort required to successfully
exploit the opportunity (e.g. time taken to travel to the crime scene, equipment
required, skills or experience needed, accomplices).
This approach plays an important role in determining the most appropriate means
to disrupt or diminish criminal opportunities, but also in anticipating how offenders
will react to crime prevention. Rational choice theory provides the foundation for situ-
ational crime prevention (Clarke, 1983, 1995, 2008), an approach that considers how
a setting or environment can be changed in order to reduce the quality and quantity of
criminal opportunities. In this sense rational choice theory provides the framework for
treating a crime problem. Our analysis might reveal that a hotspot is caused by an
abundance of easy criminal opportunities, so manipulating the environment so that
those opportunities are harder to exploit is likely to have a beneficial impact on crime.
There are considerable examples of the utility of rational choice theory, and by
extension situational crime prevention (Clarke, 2012). A simple example is found in
the security hypothesis articulated by Farrell, Tilley, Tseloni, and Mailley (2011).
They demonstrate how the introduction of electronic immobilisers and central lock-
ing ushered in sustained, national reductions in vehicle crime.

Routine Activity Theory

Routine activity theory was proposed by Cohen and Felson (1979). This approach
says that crime is possible when a motivated offender and a suitable target come
together in some place. Crime is prevented if a target is protected by capable guard-
ian (friends of a person or owner of an object), or a place manager is regulating the
setting (door staff at a nightclub), or offenders are controlled by handlers (parents,
teachers, peers) or some combination of the above controllers.
The routine activity approach was a novel contribution to our understanding of
crime because it suggested that everyday activities contribute to how offenders and
victims came to converge. Importantly, the convergence rate of these two actors
rather than their absolute numbers that determined the amount of crime.1

1
In other words, increases in crimes could be the result of more offenders or increasing motivation
(of a fixed number of offenders) or both. Yet, routine activity theory implies that offender preva-
8 Crime Mapping and Spatial Analysis 103

Routine activity theory explains why crime is unevenly distributed throughout


the year (McDowall, Loftin, & Pate, 2011), week (Kurland, Johnson, & Tilley,
2014) and day (Rengert, 1997). For this reason it is of central relevance in the diag-
nosis of crime problems because it highlights how the rhythm of (non-criminal)
social activities contributes to criminal opportunity.

Geometry of Crime Theory

The geometry of crime theory takes an explicit spatial perspective on crime pat-
terns. Briefly, it says that individuals have certain locations, nodes, they are inti-
mately familiar with and spend considerable amount of time at (e.g. school,
domicile, work). They also travel between nodes in regular predictable routes,
paths. The combination of nodes and paths for an individual is termed an awareness
space; it could be thought of as the sum total of the spatial cognition of a person.
The implication of these straightforward concepts are that they set out what the
offenders' search space who are looking for criminal opportunities. They can only
exploit opportunities they are aware of, and these are limited to their awareness
space. For most people they will commit crime in the nearby vicinity of nodes or
paths in their awareness space.
How can this be used in analysis? While an individual awareness space is of
academic interest, consider if we could gather the awareness spaces of an entire
community. Some nodes would be shared by a small number of people, those living
in the same household, but others would be shared by many people, such as public
transport hubs, schools and shopping centres. This means is that opportunities
located near nodes shared by many people will be scrutinised by more people and
good opportunities to commit crime (if present) will be exploded quickly. Criminal
opportunities featuring in few awareness spaces are unlikely to be exploited or will
be exploited at a far lower rate.
This then leads to predictions about where crime hotspots will emerge and why.
The Brantinghams talk about two types of hotspots, crime generators and crime
attractors (Brantingham & Brantingham, 1999). The key distinction between the
two is that generators are the result of an abundance of targets (shopping centres,
say) and attractors are the results of abundance of offenders who travel to locations
that they know provide good opportunities for crime (night-time entertainment pre-
cincts, open-air drug markets).
Some commentators depict environmental criminology as a response to an
excessive focus by researchers on criminality and treating offenders as fundamen-
tally different from non-offenders. While this may be true, environmental criminol-
ogy is underpinned by knowledge conceived outside criminology. Two concepts
that directly inspired the thinking of environmental criminologists are:

lence and propensity can remain fixed and crime can increase if the offender-victim convergence
rate is increased.
104 M. Townsley

• Tobler's first law: Everything is related to everything else but near things are
more related than distant things (Tobler, 1987). The composition of your house-
hold is probably more similar to your neighbour's household than a random
household on the other side of town. Commercial premises are more likely to be
located next to other commercial premises than residential dwellings. In fact
Tobler's first law applies to land use in all communities—spaces set aside for
residential, human services, commercial and light industry all tend to cluster into
distinct zones.
• least effort principle: Humans will choose the path of least resistance in order to
achieve goals (Zipf, 1965). Individuals need to move across space in order to
achieve goals such as work, education and recreation. Each of these requires a
decision about where and how to travel. The social and physical environment, in
turn, either facilitates or restricts the free movement of individuals. For instance
minorities prefer to avoid white majority communities (Rengert & Wasilchick,
1990), bridges act as connectors and rivers act as obstacles (Clare, Fernandez, &
Morgan, 2009). However, everyone has finite resources to achieve their goals, so
trips that require fewer resources than others will be preferred (all other things
being equal). Long trips will only be taken if the payoff is greater then the addi-
tional resources required to make it (Townsley & Sidebottom, 2010; Vandeviver,
Van Daele, & Vander Beken, 2014).
How do these two concepts underpin environmental criminology? If property
types and demographics cluster in space, then whatever combination of factors
make a house or person an attractive or lucrative target will similarly tend to be
spatially concentrated. Because humans move in predictable ways, much of this
spatial configuration produces a predictable convergence between offenders and tar-
gets/victims, thus leading to criminal opportunities. The consistency of these con-
vergences gives rise to spatial clustering of crimes.
To illustrate the point, consider Ratcliffe's temporal constraint concept (Ratcliffe,
2006). He describes a scenario in which an individual has discretionary time before
some fixed activity, a child attending school in his example. If the journey to school
takes 10 min and the child leaves home 20 min before school starts, she has 10 min
free time. Ratcliffe then shows that there is a “zone of roaming" the child could
cover with their free time and still observe their fixed activity. The practical applica-
tion of the concept are that houses around schools are more exposed to individuals
with discretionary time than other properties. Yet this increased exposure only
occurs at certain times—increased vigilance is only required for a fraction of the
day. The same can be said of properties around public transport, shopping precincts
and other crime generators (Brantingham & Brantingham, 1995). The broader point
is that temporal constraints are interwined as spatial contrainsts; there is only some
far one can travel with a fixed period of discretionary time.
To return to the medical analogy, how do we understand treatment? How does
locating a spatial concentration of crime help us prevent crime? Merely locating a
crime pattern or problem is insufficient for its alleviation. Some amount of analytic
work needs to take place in order for the underlying cause of the problem pattern to
8 Crime Mapping and Spatial Analysis 105

be identified. One approach used for this task in problem-oriented policing


(Goldstein, 1990; Scott, Eck, Knutsson, & Goldstein, 2008). Problem-oriented
policing considers the fundamental unit of law enforcement agencies to be problems
rather than dealing with individual matters, arrests, or offenders. Problems are sim-
ply groups of crime events that have some common characteristic. For example it
may be all burglaries on a housing estate or a series of motor vehicle thefts from a
shopping mall. By grouping many crime events, analysts can identify common
characteristics of the crime commission process and target weak links in order to
prevent or interrupt this process. Burglaries on public housing estates may reveal a
lack of physical damage to doors or windows, implying that premises are left inse-
cure when they are unoccupied. Promoting increased vigilance may be an effective
means of curbing the problem.

How and Under Which Circumstances Can Spatial Analysis


Work Best for Informing Prevention Initiatives?

There are four attributes that enable the context for spatial approaches of crime
prevention to be effective. The first is accurate and reliable data. In the law enforce-
ment context, recorded crime data or calls for service data are the most viable source
of crime information. However, analysts need to bear in mind that the utility of
recorded or reported data varies by crime type. For instance drug offences are a very
good indicator of police activity but not a reliable indicator of drug use. Crimes such
as burglary and vehicle crimes usually trigger an insurance claim, so these are fairly
well reported. Even so, there is likely a sample bias here as well. Low-income
households or students sharing accommodation may not take out contents insurance
and so would not have a financial incentive to report burglaries. Importantly, due to
Tobler's law these types of households are likely to clustered in space further obscur-
ing diagnosis of crime problems. This point was illustrated by Ceccato and Lukyte
(2011) in the city of Vilnius (Lithuania). They showed large differences in spatial
patterns of crime between police recorded crime data and a large victimisation
survey.
Other data are important also. Because the theoretical framework focuses on the
opportunity structure of crime, analysts need to depict this, ideally comprising the
population at risk and its major correlates. For some crimes with static targets (e.g.
residential burglary, vandalism) this is fairly straightforward because many com-
munities have digitised property parcels, land use, street networks, public buildings
and other infrastructure. This backdrop can be used to depict the population-at-risk.
For crimes with mobile targets (e.g. motor vehicle theft, street robbery) the popula-
tion at risk is much more difficult to credibly establish. Spatial concentrations may
simply indicate a greater ambient target population. See Andresen and Jenion
(2010) and Andresen (2011) for a detailed treatment of determining ambient target
population.
106 M. Townsley

The second determinant is use of an explicit theory. Explicit theories are


necessary because they (1) help generate hypotheses, (2) suggest which factors (i.e.
data) should be included in the analysis and (3) have undergone a process of for-
malisation. The benefit of this last attribute is that explicit theories are subject to
testing and falsification.
Environmental criminology is the appropriate body of theories to use in crime
analysis because they alone attempt to explain attributes of crimes. Police leaders
and analysts tend to operate under implicit theories, i.e. a world view often
informed by personal experience. The problem with this type of knowledge is that
it can differ by individual; two officers with 30-year experience can draw different
conclusions in the same situation. In addition, this tacit knowledge is extremely
difficult to apply in a falsifiable way. This is not the same as saying that experience
counts for naught, just that effective analysts incorporate explicit theories into their
analytical approach.
Eck (1997) illustrates the point, arguing that crime maps and analytic products
are often theory-free and, therefore, difficult to interpret. Crime mapping software
makes it easy to combine many datasets, but the map maker cannot include all geo-
graphic datasets—the resulting map would be too crowded and tax the reader. For
instance, the City of Chicago currently lists over 300 different geospatial datasets
available for analysis. So which should be included and which omitted? Those that
have theoretical relevance, according to Eck. Explicit theories are desirable because
they outline which factors are most likely to be associated with crime. Implicit theo-
ries, on the other hand, are too nebulous to discriminate between relevant factors.
The third requirement is a logical conclusion of the first two: some means of
generating and testing hypotheses. Combining appropriate theory with valid data
will permit an analyst to probe the problem and hopefully uncover the cause of the
problem. This is the final step in the “diagnosis" phase of crime analysis. Ekblom
(1998) and Weisel (2003) both describe this stage as a process of documenting the
problem, data collection and analysis. Ekblom’s (1998) compendium question is a
useful guide to frame this step (p. 11):
the core of the analysis can be summarised in a ‘compendium’ question: what offences
occurred; where, when, under what circumstances and by what method were they commit-
ted; and who or what was the victim or target? The answer to the question can contribute to
an understanding of the crime pattern, the criminal opportunities that underlie it, and the
scope for prevention.

The fourth and final component is an understanding of how to prevent crime. It


may seem banal but many senior ranking police officers know little about prevent-
ing crimes (Ratcliffe, 2008a). Popular policing tactics—random patrolling, rapid
response and arrest policies—have little empirical support with respect to crime
control (Goldstein, 1979, 1990; Sherman et al., 1997). Even tactics to increase
police visibility, such as saturation patrolling or high-visibility patrolling, have lim-
ited effectiveness. In their systematic review of drug law enforcement methods,
Mazerolle, Soole, and Rombouts (2007) found that intensive policing operations
had little impact on drug problems. But empirical evidence of effective pragmatic
8 Crime Mapping and Spatial Analysis 107

tactics are available, notably focused patrolling of high risk areas underpinned by
effective forecasting (Fielding & Jones, 2012; Fitterer, Nelson, & Nathoo, 2014;
Nix, 2015; Sorg, Haberman, Ratcliffe, & Groff, 2013, 2014, Sorg, Wood, Groff, &
Ratcliffe, 2014).

What Are the Strengths and Weaknesses of Spatial Analysis?

One of the advantages of spatial approaches to understanding crimes is that all


aspects of human behaviour have an inherent geographical component to them:
everything has to take place somewhere. Spelman and Eck (1989) found that crime
was most heavily concentrated at places, more than among offenders and victims.
Forty percent of all crime is experienced by the top 10 % of victims, 50 % of crime
is committed by the top 10 % of offenders and 60 % of crimes are hosted in the worst
10 % of places.
With reasonably accurate data and modern analytic techniques and mapping
technologies it is relatively straightforward to explore and identify spatial patterns
of crime. The ease of generating patterns, ironically, makes analysis harder because
spurious relationships is easier to unearth. Humans are sensitive to patterns and are
easily fooled: pareidolia is a well-known example of seeing faces in inanimate
objects such as Jesus in a grilled cheese sandwich, and apophenia is the broader
tendency of seeing patterns or connections in random or meaningless data. The
easier it is to generate data and compare and contrast relationships, the more likely
spurious results will arise. Unfortunately the standard of analytic training in most
law enforcement agencies is insufficient for the increased complexity introduced by
the spatial domain (Goldstein 2003).
A further problem introduced by spatial analyses is the well-known Modifiable
Area Unit Problem (MAUP) (Openshaw 1983). MAUP refers to the tendency for
observed spatial patterns to be influenced by the administrative boundaries used for
aggregation. For instance, Ratcliffe (2008a) describes how police beats in the UK
had their origins in parish catchment areas, something presumably independent of
the opportunity structure for crimes. Analysts have a surplus of administrative
boundaries they can use to frame their analysis, including electoral wards, post-
codes, neighbourhoods and a range of statistical boundaries. None of these have any
clear link with the underlying potential for crime and its opportunity surface.
Importantly if analysts were to shift between these, they are highly likely to produce
different depictions of the spatial distribution of crimes.
Another problem with crime analysis particularly as it relates to prevention is
regression to the mean. In the simplest terms, when an area is extreme at one
point in time, it is likely to be closer to the average in the next time period.
Consider the scenario where a hotspot is identified and resources deployed for
prevention. If in the next time period the area is no longer considered “hot" it is
difficult for an analyst to discriminate between effective practice and regression
108 M. Townsley

to the mean. Because hotspots are areas with extreme crime counts/rates
regression to the mean is a pervasive competing explanation that needs to be
included in any evaluation. Seasonal trends are another potential opportunity for
regression to the means to obscure effectiveness. Resources deployed at the sea-
sonal peak will appear to be more effective than they really are. The best remedy
or protection against regression to the mean is using comparison areas, places
that did not receive the preventative action and are unlikely to experience spill-
over from the action area. For the seasonality example, observing declines in the
action and comparison areas will highlight the subordinate contribution made by
the tactics selected.
There is a risk that police may sour community relations because hotspot polic-
ing can favour intensive responses that just suppress the problem. Merely increasing
patrolling or arrests will not only be ineffective (Telep and Weisburd 2012) but the
legitimacy of police will be diminished. Generic responses, that impact many com-
munity members, are likely to suppress symptoms rather than address causal mech-
anisms. A better approach is to use highly specific, targeted approaches underpinned
in a procedurally fair way (Tyler 1990).
What are the solutions to these problems? Evidently better training for analysts
is a must. Particularly training in inferential techniques (assessing the likelihood of
observing patterns based on some expectation of crime levels), generating and test-
ing hyptheses, as well as both process and outcome evaluations (Eck 2002; Goldstein
2003; Townsley 2013; Townsley et al. 2011).
In addition, it is not just analysts that require support, but also police leaders.
It remains a rarity in modern law enforcement agencies that selection criteria
for promotion includes crime prevention as a core competency in police execu-
tives (Ratcliffe 2008a). It is much more common that senior executives have an
investigative background and leaders are often completely unaware that crime
reduction can include situational approaches that prevent crime from occurring
in the first place, instead of relying on deterring offenders through the use of the
criminal justice system.

What Is the Future of Spatial Analysis in Criminology?

Spatial analysis and crime mapping have recently captured the attention of scholars
and law enforcement alike. With respect to preventing crime it represents arguably
the most active area of policing-focused criminological research today. The reasons
for this trend are that data collection and storage have become increasingly cheaper
and easier. In addition, the focus of the approach, the location of crimes, is informa-
tion the police systematically collect. Building on efforts law enforcement routinely
are engaged has lead to a rapid uptake.
But what of the future of spatial analysis and crime mapping? At the start of this
chapter I stressed that this approach is comprised of two parts: the identification of
8 Crime Mapping and Spatial Analysis 109

hotspots and deploying resources to address them. The future of this approach is
best assessed by looking at the limitations of each part.
In terms of identifying spatial concentration of crimes (diagnosis), the historical
factors outlined above suggest that analysis will continue to be considered an impor-
tant and vital role in law enforcement agencies. Thanks to increasing (and cheaper)
computing power, it is becoming easier to deploy advanced statistical techniques
that were impractical even a short time ago. The benefits of big data (Grant 2012;
Manyika et al. 2011) and data science (Stanton 2012) movements suggest that the
focus on reliable and valid data will not go away.
Turning to the prevention (treatment), here I am less hopeful. Analysts typically
do not receive the training necessary to make inferences about data. Many intelli-
gence products are at the level of describing the criminal environment rather than
interpreting it (Ratcliffe 2008a; Townsley et al. 2011). Simply describing a crime
problem is akin to a physician limiting diagnosis to merely cataloguing a patient's
symptoms. Analysis requires a process of developing an explanation of how the set
of symptoms is related to the health problem.
This training gap is not limited to analysts, however. Police leaders and decision
makers in law enforcement agencies also need exposure to effective crime preven-
tion knowledge (Ratcliffe 2008b). Policing tactics still largely comprise of conven-
tional policing approaches (random patrolling, high visibility patrolling, rapid
response, specialised units) that have an extensive body of evidence suggesting that
they are ineffective.
This training and knowledge gap is not new. Scholars have made this point many
times for many years (Clarke and Goldstein 2003; Goldstein 1979; Scott et al.
2008). It is the primary factor Goldstein identified, reflecting on 20 years of research,
limiting the wider adoption of problem-oriented policing (Goldstein 2003). What is
required is for police to rethink their understanding of the unit of analysis for police
work. Goldstein's central claim was that police organisations should move away
from viewing their work as the processing of individual crimes/arrests/calls for ser-
vice and instead aggregate common criminal matters into problems. This claim is
still relevant today. There is a need to move away from offender-centricism, to
place-orientation and problem-solving, away from prosecutions, towards opportu-
nity reduction (Weisburd et al. 2010).

References

Andresen, M. A. (2011). The ambient population and crime analysis. The Professional Geographer,
63(2), 193–212.
Andresen, M. A., & Jenion, G. W. (2010). Ambient populations and the calculation of crime rates
and risk. Security Journal, 23(2), 114–133.
Beirne, P. (1987). Adolphe quetelet and the origins of positivist criminology. American Journal of
Sociology, 92(5), 1140–1169.
Brantingham, P. J., & Brantingham, P. L. (1981). Notes on the geometry of crime. In P. J.
Brantingham & P. L. Brantingham (Eds.), Environmental criminology (pp. 27–54). London,
UK: Sage Publications.
110 M. Townsley

Brantingham, P. J., & Faust, F. L. (1976). A conceptual model of crime prevention. Crime &
Delinquency, 22(3), 284–296. doi:10.1177/001112877602200302.
Brantingham, P. L., & Brantingham, P. J. (1995). Criminality of place: Crime generators and crime
attractors. European Journal of Criminal Policy and Research, 3(3), 5–26.
Brantingham, P. L., & Brantingham, P. J. (1999). A theoretical model of crime hot spot generation.
Studies on Crime and Crime Prevention, 8, 7–26.
Ceccato, V., & Lukyte, N. (2011). Safety and sustainability in a city in transition: The case of
Vilnius, Lithuania. Cities, 28(1), 83–94.
Clare, J., Fernandez, J., & Morgan, F. (2009). Formal evaluation of the impact of barriers and con-
nectors on residential burglars’ macro-level offending location choices. The Australian and
New Zealand Journal of Criminology, 42(2), 139–158.
Clarke, R. V. (1983). Situational crime prevention: Its theoretical basis and practical scope. In
N. Morris & M. Tonry (Eds.), Crime and justice: A review of research (Vol. 4, pp. 225–256).
Chicago, IL: University of Chicago Press.
Clarke, R. V. (1995). Situational crime prevention. In M. Tonry & D. P. Farrington (Eds.), Crime and
justice: A review of research (Vol. 19, pp. 91–150). Chicago, IL: University of Chicago Press.
Clarke, R. V. (2012). Opportunity makes the thief really? And so what? Crime Science, 1, 3.
Clarke, R. V. G. (2008). Situational crime prevention. In R. Wortley & L. Mazerolle (Eds.),
Environmental criminology and crime analysis (pp. 178–194). Cullompton: Willan
Publishing.
Clarke, R. V. G., & Goldstein, H. (2003). Thefts from cars in centre-city parking facilities: A case
study in implementing problem-oriented policing. In J. Knutsson (Ed.), Problem-oriented
policing: From innovation to mainstream (Vol. 15, pp. 257–298). Monsey, NY: Criminal
Justice Press.
Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 44(4), 588–608.
Cornish, D. B., & Clarke, R. V. (2008). The rational choice perspective. In R. Wortley &
L. Mazerolle (Eds.), Environmental criminology and crime analysis (pp. 21–47). Cullompton:
Willan Publishing.
Cornish, D. B., & Clarke, R. V. G. (1986). The reasoning criminal: Rational choice perspectives
on offending. New York: Springer.
Eck, J. E. (1997). What do those dots mean? Mapping theories with data. In D. L. Weisburd &
T. McEwen (Eds.), Crime mapping and crime prevention (Vol. 8, pp. 377–406). Monsey, NY:
Criminal Justice Press.
Eck, J. E. (2002). Assessing responses to problems: An introductory guide for police problem-
solvers. Washington, DC: Office of Community Oriented Policing Studies.
Ekblom, P. (1998). Getting the best out of crime analysis. London, UK: Home Office.
Farrell, G., Tilley, N., Tseloni, A., & Mailley, J. (2011). The crime drop and the security hypothesis.
Journal of Research in Crime and Delinquency, 48(2), 147–175.
Felson, M. (2008). Routine activity approach. In R. Wortley & L. Mazerolle (Eds.), Environmental
criminology and crime analysis (pp. 70–77). Cullompton: Willan Publishing.
Fielding, M., & Jones, V. (2012). ‘Disrupting The Optimal Forager’: Predictive risk mapping and
domestic burglary reduction in Trafford, Greater Manchester. International Journal of Police
Science & Management, 14(1), 30–41.
Fitterer, J., Nelson, T., & Nathoo, F. (2014). Predictive crime mapping. Police practice and
research: An international journal, 16, 121–135. doi:10.1080/15614263.2014.972618.
Goldstein, H. (1979). Improving policing: A problem-oriented approach. Crime and Delinquency,
25(2), 236–258.
Goldstein, H. (1990). Problem-oriented policing. New York: McGraw-Hill.
Goldstein, H. (2003). On further developing problem-oriented policing: The most critical need, the
major impediments and a proposal. In J. Knutsson (Ed.), Problem-oriented policing: From
innovation to mainstream (Vol. 15, pp. 13–48). Monsey, NY: Criminal Justice Press.
Grant, E. (2012). The Promise of Big Data. The promise of big data, pp. 15–43.
8 Crime Mapping and Spatial Analysis 111

Kurland, J., Johnson, S. D., & Tilley, N. (2014). Offenses around stadiums: A natural experiment on
crime attraction and generation. Journal of Research in Crime and Delinquency, 51(1), 5–28.
Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., et al. (2011). Big data: The
next frontier for innovation, competition and productivity. McKinsey Global Institute.
Mazerolle, L., Soole, D., & Rombouts, S. (2007). Drug law enforcement: A review of the evalua-
tion literature. Police Quarterly, 10(2), 115–153. doi:10.1177/1098611106287776.
McDowall, D., Loftin, C., & Pate, M. (2011). Seasonal cycles in crime, and their variability.
Journal of Quantitative Criminology, 28(3), 389–410.
Nix, J. (2015). Predictive policing. In R. G. Dunham & G. P. Alpert (Eds.), Critical issues in polic-
ing: Contemporary readings (7th ed., pp. 275–287). Long Grove, IL: Waveland Press.
Openshaw, S. (1983). The modifiable areal unit problem. Norwich, UK: Geo Books.
Ratcliffe, J. H. (2006). A temporal constraint theory to explain opportunity-based spatial offending
patterns. Journal of Research in Crime and Delinquency, 43(3), 261–291.
Ratcliffe, J. H. (2008a). Intelligence-led policing. Cullompton: Willan Publishing.
Ratcliffe, J. H. (2008b). Knowledge management challenges in the development of intelligence-
led policing. In T. Williamson (Ed.), The handbook of knowledge-based policing: Current con-
ceptions and future directions (pp. 205–220). Chichester: John Wiley & Sons.
Rengert, G. F. (1997). Auto theft in central Philadelphia. In R. Homel (Ed.), Policing for pre-
vention: Reducing crime, public intoxication and injury (pp. 199–220). Monsey, NY:
Criminal Justice Press.
Rengert, G. F., & Wasilchick, J. (1990). Space, time and crime: Ethnographic insights into resi-
dential burglary. Unpublished report to the National Institute of Justice. Philadelphia, PA:
Temple University.
Scott, M. S., Eck, J. E., Knutsson, J., & Goldstein, H. (2008). Problem-oriented policing and envi-
ronmental criminology. In R. Wortley & L. Mazerolle (Eds.), Environmental criminology and
crime analysis (pp. 221–246). Cullompton: Willan Publishing.
Sherman, L. W., Gartin, P., & Buerger, M. (1989). Hot spots of predatory crime: Routine activities
and the criminology of place. Criminology, 27(1), 27–56.
Sherman, L. W., Gottfredson, D., MacKenzie, D., Eck, J. E., Reuter, P., & Bushway, S. (1997).
Preventing crime: What works, what doesn’t, what’s promising. Washington, DC: National
Institute of Justice.
Sorg, E. T., Haberman, C. P., Ratcliffe, J. H., & Groff, E. R. (2013). Foot patrol in violent crime
hot spots: The longitudinal impact of deterrence and posttreatment effects of displacement.
Criminology, 51(1), 65–101.
Sorg, E. T., Wood, J. D., Groff, E. R., & Ratcliffe, J. H. (2014). Boundary adherence during place-
based policing evaluations: A research note. Journal of Research in Crime and Delinquency,
0022427814523789.
Spelman, W., & Eck, J. E. (1989). Sitting ducks, ravenous wolves and helping hands: New
approaches to urban policing. Public Affairs Comment, 35(2).
Stanton, J. (2012). An introduction to data science. New York, NY: Syracuse University.
Telep, C. W., & Weisburd, D. (2012). What is known about the effectiveness of police practices in
reducing crime and disorder? Police Quarterly, 15(4), 331–357.
Tobler, W. (1987). Experiments in migration mapping by computer. Cartography and Geographic
Information Science, 14(2), 155–163.
Townsley, M. (2013). Making crime mapping more inferential. In: D. Weisburd & G. J. N. Bruinsma
(Eds.), Encyclopeadia of criminology and criminal justice (p. (Accepted 1 Nov, 2012)).
New York: Springer.
Townsley, M., & Sidebottom, A. (2010). All offenders are equal, but some are more equal than
others: Variation in journeys to crime between offenders. Criminology, 48(3), 210–222.
Townsley, M., Mann, M., & Garrett, K. (2011). The missing link of crime analysis: A systematic
approach to testing competing hypotheses. Policing, 5(2), 158–171.
Tyler, T. R. (1990). Why people obey the law: Procedural justice, legitimacy, and compliance. New
Haven, CT: Yale University Press.
112 M. Townsley

Vandeviver, C., Van Daele, S., & Vander Beken, T. (2014). What makes long crime trips worth
undertaking? Balancing costs and benefits in burglars’ journey to crime. British Journal of
Criminology, azu078.
Weisburd, D., Telep, C. W., Braga, A. A., & Groff, E. R. (2010). The importance of place in polic-
ing: Empirical evidence and policy recommendations. Brottsförebyggande rådet (BRÅ)
Weisel, D. L. (2003). The sequence of analysis in solving problems. In J. Knutsson (Ed.), Problem
oriented policing: From innovation to mainstream (Vol. 15, pp. 115–146). Monsey, NY:
Criminal Justice Press.
Wortley, R., & Mazerolle, L. (Eds.). (2008). Environmental criminology and crime analysis.
Cullompton: Willan Publishing.
Zipf, G. K. (1965). Human behavior and the principle of least effort: an introduction to human
ecology. New York: Hafner.
Chapter 9
Mapping Crime Prevention: What We Do
and Where We Need to Go

Martin A. Andresen

Introduction

Crime prevention is an activity that can take many forms. Crime prevention initia-
tives can be considered developmental, community-based, rehabilitation, policing,
deterrence-punishment, and through modifications of criminal opportunities often
referred to as crime prevention through environmental design—see Tilley (2005)
and other chapters in this volume for discussions of the various forms of crime pre-
vention. Mapping crime prevention is most often associated with the latter of these
types of crime prevention initiatives, but may also be part of policing and community-
based initiatives. Most of the discussions below are most closely aligned with situ-
ational crime prevention (Clarke, 1980, 1983, 1997, 2012).
Mapping for the purposes of crime prevention is instructive because it shows
where the criminal event problem is, but it is also instructive for the purposes of
evaluating various crime prevention initiatives. In this chapter, I briefly cover the
state of the art in mapping for crime prevention and then move into some consider-
ations for the future of mapping crime prevention. I hope to show that adding
sophistication to mapping crime prevention does not necessarily increase its com-
plexity, but adds a significant amount of information that is instructive for under-
standing the criminal event problem at hand, identifying the appropriate crime
prevention initiative, and performing an evaluation of crime prevention initiatives
that can inform future crime prevention activities.

M.A. Andresen (*)


School of Criminology, Institute for Canadian Urban Research Studies, Simon Fraser
University, Burnaby, BC, Canada, V5A1S6
e-mail: andresen@sfu.ca

© Springer International Publishing Switzerland 2017 113


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_9
114 M.A. Andresen

State of the Art: What We Do to Map Crime Prevention

Though one may find a number of different methods to map crime for prevention
purposes, the most common forms of doing so are dot maps and (kernel) density
maps (Santos, 2013). Dot maps can take two forms: (1) one dot on the map for
every (criminal) event; and (2) graduated dots such that as the number of (crimi-
nal) events at each location increases the dot gets larger. This technique can be
instructive when there are not many dots to place on the map but very quickly
becomes difficult to interpret as the number of criminal events increases. Kernel
density maps are based on dots on a map but create a surface representing the
intensity of dots at any location in the study area. There are other basic and
advanced techniques available, but are seldom used when mapping for crime pre-
vention (Santos, 2013).
The kernel density hot spot map is calculated as follows, and shown in
Fig. 9.1. Though the details of how the calculation is made can vary, the general
formation begins by placing a grid over the study area. For each cell in that grid
a value is calculated that represents the intensity of criminal event activity. From
each cell, a circle is drawn (the diameter of this circle is called the bandwidth)
and the criminal events (points) that fall within that circle are counted and used
to calculate the kernel, or the “height” of the hot spot. Generally speaking, the
more criminal events within the circle, the higher the kernel will be for that cell.
This is repeated for every cell within the grid placed over the study area that can
then be used to generate a map showing hot spots: hot spots are usually identified
by the presence of a number of cells that have high kernel values, as can be seen
in Fig. 9.2a.

Fig. 9.1 Kernel density calculations. Source: Andresen (2014)


9 Mapping Crime Prevention: What We Do and Where We Need to Go 115

Fig. 9.2 Kernel density, single versus dual Kernel. Source: Andresen (2014)

Considerations for Mapping Crime Prevention

No discussion of the consideration for the mapping of anything, including crime pre-
vention, would be complete without a brief discussion of the ecological/atomistic fal-
lacy and the modifiable areal unit problem. The ecological/atomistic fallacy refers to
the improper inference that can be made when analyzing spatial data, including maps,
and the modifiable areal unit problem refers to how the results of an analysis can
change when you change the spatial unit of analysis. We discuss each, briefly, in turn.
The ecological/atomistic fallacy was formally identified by Robinson (1950) when
he identified the conditions that must hold in order to claim that relationships found at
116 M.A. Andresen

one unit of analysis (the neighborhood, for example) would be the same for another
unit of analysis (the individual, for example). He then went on to state how there is no
evidence for these conditions to hold true in any analysis—Openshaw (1984a) showed
that the differences between individual and area-level relationships could not be
known, a priori. As such, the ecological fallacy occurs when a researcher or practitio-
ner states that what is true of the whole is also true of all its parts; the atomistic fallacy
occurs in the opposite direction such that what is true of the parts is also true of the
whole. Of course, the fallacy is rarely, if ever, committed so overtly. Rather, an analy-
sis (for crime prevention, for example) is undertaken using census level data (dissemi-
nation areas of census block groups) and then the researcher or practitioner attributes
the relationships found to the individuals who live in that census area. This is actually
an easy fallacy to commit, but it is also an easy fallacy to avoid.
The modifiable areal unit problem (MAUP) also emerges from the aggregation of
data. In the census, for example, data are aggregated from individuals to census tracts
and other spatial units of analysis for subsequent analysis. The trouble is that the
aggregations of these data are usually deemed arbitrary in the sense that they do not
represent “natural units” of analysis. This can emerge in two primary ways: the scale
problem and the zoning problem. The scale problem emerges from analyzing different
sized spatial units of analysis: census block groups, census tracts, and neighborhoods,
for example. The zoning problem emerges when the size and dimension of the spatial
units does not change but they are placed differently on the map—think of dragging
census tract boundaries over by one block, or more. In both cases, different sets of
individuals are aggregated into spatial units of analysis. The problem with these modi-
fiable areas is that the results of any analysis can change, often unpredictably.
Openshaw (1984b) showed that any desired results could be obtained through the
modification of areal units, and Fotheringham and Wong (1991) showed that statisti-
cal results were not reliable and that their results were “depressing”—Wooldredge
(2002) did have more promising results in a spatial crime analysis context.
This does not mean that mapping for crime prevention is inherently problematic.
But it does mean that when undertaking mapping for crime prevention that the
research and/or practitioner must be aware that what they are doing and any results
that they find may be an artifact of the spatial units of analysis that they are using.
Consequently, the use of at least two spatial units of analysis should be undertaken
to (hopefully) show that results are not sensitive to the spatial unit used in the
analysis. It would be recommended to start with the smallest unit of analysis avail-
able (the discrete criminal event point, if possible) and then aggregate, as neces-
sary. This is important because neither the researcher nor the practitioner would
want to improperly identify a location for a crime prevention initiative.

The Limitations of Kernel Density Mapping

As discussed above, kernel density mapping is one of the most common forms of
crime mapping that is used in crime prevention. Though instructive for its use to iden-
tify crime hot spots, like any analytical technique, it has its limitations. Two of these
limitations are discussed here.
9 Mapping Crime Prevention: What We Do and Where We Need to Go 117

First, kernel density mapping, in its most common form, uses only criminal
event data in order to identify hot spots. But what is a hot spot? John Eck and col-
leagues define as hot spot as “an area that has a greater than average number of
criminal or disorder events, or an area where people have a higher than average risk
of victimization” (Eck, Chainey, Cameron, Leitner, & Wilson, 2005, p. 2). This defi-
nition has two components to it: (1) greater than average number of criminal events,
and (2) higher than average risk. These can be two very different phenomena.
If the researcher or practitioner is concerned with the prevention of the volume
of crime, then a kernel density map that only uses criminal event data is just fine.
However, because criminal events occur when motivated offenders and suitable tar-
gets converge in time and space without the presence of capable guardians (Cohen
& Felson, 1979), where there are more convergences there are more criminal events.
As such, just because there are a lot of criminal events does not mean that a person
is actually at a high risk of victimization. This is quite common in central business
districts and has been known for decades (Schmid, 1960a, 1960b). As a conse-
quence of this, if the intention of a crime prevention initiative is to reduce the risk
of victimization then mapping only criminal events may be problematic. This is
why we calculate crime rates: so we can compare rates across time and space when
we know the population at risk of victimization changes.
In order to control for this in kernel density mapping, if it is a concern, a dual
kernel must be calculated. This is a very simple technique to perform with the only
complication being the need for population at risk data that varies across space at a
spatial resolution that is appropriate for the crime prevention initiative. This is rela-
tively simple when mapping for crime prevention at the level of the city because
reliable data are available for the resident population at the dissemination area (cen-
sus block group) that will give a reasonable approximation of the population at risk,
but is not always a good representation of the population at risk—see Andresen and
Jenion (2010) for a critique of such data. An example of how this impacts the hot
spot map is shown in Fig. 9.2, for violent crime in Vancouver, British Columbia,
Canada, using the ambient population that is more appropriate for considering vio-
lent crime but is not as readily available as census data.
We can see from Fig. 9.2 that using a single-kernel (criminal event data only)
violent crime has a hot spot that is located within the central business district (the
peninsula at the northern portion of the city) that exhibits a distance decay pattern
as one moves away from the central business district. Using this information may
lead the researcher or practitioner to target the high-density area for a crime preven-
tion initiative. This is not incorrect because that is where the volume of violent
crime is occurring, but when considering the dual kernel map that hot spot all but
disappears. The location with the highest density is still in the central business dis-
trict, but no longer in the center of the peninsula; this high-density area is now on
the edge of the central business district itself and closer to skid row. Crime preven-
tion initiatives imposed in the center of the central business district would be very
different from those being implemented on the edge of skid row in any given city.
The second issue regarding kernel density mapping that needs to be made clear is
that the method itself is not technically appropriate for criminal event data. This is
because kernel density mapping is a surface generation technique for geographically
118 M.A. Andresen

continuous data. An example of geographically continuous data is temperature:


temperature is everywhere but it is expensive to measure temperature everywhere.
So, we measure temperature in a number of discrete locations (this would be deter-
mined based of the appropriate sampling model for the study) and then kernel den-
sity mapping or some other technique would be used to re-create the continuous
surface.
The question we need to ask is whether or not this is a useful technique for iden-
tifying crime prevention initiative locations. If we are mapping for the purposes of
crime prevention we must necessarily properly identify the correct locations to
implement those initiatives. Based on the discussion above regarding how the kernel
density map is created, it should be clear that a value can be calculated at a location
that does not have (and never has had) any criminal events because criminal events
occurred within the bandwidth of that location. Additionally, because of the band-
width criminal events may be counted more than once such that any hot spots will
appear to be larger than they really are. Indeed, this is the nature of kernel density
mapping: all of the data are smoothed over a surface to generate an easy-to-interpret
map. However, if the information that is generated in the map is not correct, or at
least misleading, its value for the prevention of crime must be questioned. A prob-
lem that may emerge because of this is the inefficient use of limited resources for
crime prevention.
This discussion should not leave the reader convinced that kernel density mapping
is not instructive. On the contrary, such a mapping technique is quick, simple, and easy
to implement and may provide very useful information to the researcher and the prac-
titioner. It is important, however, to know the limitations of the method in order to be
able to interpret the output (maps) with caution. One way to do this is to employ more
than one method of mapping crime for the purposes of identifying the potential loca-
tion for crime prevention initiatives, or evaluating those crime prevention initiatives.

The Importance of Local Spatial Analysis for Mapping Crime


Prevention: The Importance of Where

There are two aspects of local spatial analysis that are discussed here in the context
of mapping crime prevention. First, when considering the possibility of crime dis-
placement, an example of “local” versus “global” analyses is provided—crime dis-
placement is a well-known and heavily researched area in the field of crime
prevention (Barr & Pease, 1990; Eck, 1993; Hesseling, 1994; Weisburd et al., 2006).
And second, an example is shown regarding changes in spatial patterns of crime that
uses both kernel density mapping and a local spatial analytical technique.
The first example with regard to crime displacement is in the context of a police
foot patrol that took place in North Vancouver, British Columbia, Canada. Andresen
and Lau (2014) undertook an evaluation of this police foot patrol initiative and
found that there was a drop in the calls for police service of approximately 17 %,
with most of the reduction in criminal events revolving around mischief and com-
mercial burglary. These authors considered the primary patrol area and a surrounding
9 Mapping Crime Prevention: What We Do and Where We Need to Go 119

area to investigate the possibility of crime displacement from the police foot patrol.
The surrounding area was considered as a whole to measure crime displacement.
However, the surrounding area was relatively large such that it may be difficult to
identify statistically significant increases in the number of criminal events; the pri-
mary patrol area was 0.90 square kilometers, whereas the potential displacement
area was approximately 3 square kilometers. In order to address this concern,
Andresen and Malleson (2014) performed a local analysis of crime displacement on
this police foot patrol initiative and considered 32 units of analysis for crime dis-
placement instead of just 1 unit of analysis.
Andresen and Malleson (2014) used a locally based spatial point pattern test
developed by Andresen (2009) to test the similarity of the spatial patterns of crimi-
nal events before and after the police foot patrol as well as identify the locations in
which the concentrations of crime increased or decreased. As such, this test allowed
the researchers to find out which of the concentrations of criminal events changed
even though it is known that criminal events decreased in both the primary patrol
area and the potential displacement area.
In their analysis, Andresen and Malleson (2014) found that the spatial crime pat-
terns had changed as a result of the police foot patrol. This, of course, is an expected
result because of a police presence in approximately one-quarter of the study area—
commercial burglary did not exhibit much spatial pattern change but the primary
patrol area is where the vast majority of the commercial land use area is located so
not much change could occur. When considering all crime types aggregated together,
little information could be obtained from their results. Some places exhibited
increases in the concentrations of criminal events and other exhibited decreases;
moreover, these increases and decreases occurred in both the primary patrol area
and the potential displacement area. The primary result of interest when considering
all criminal events is that there appeared to be more increases in the concentrations
of criminal events at the border areas of the primary patrol area where less time will
be spent by those who undertake the police foot patrol. In the context of mischief,
the crime type that exhibited the most significant decrease in activity, there was a
moderate indication that the spatial pattern of this crime type shifted away from the
primary patrol area into a small number of areas.
The importance of thinking local should be self-evident from these results.
Global evaluation of the police foot patrol indicates that criminal events decreased
with no evidence for crime displacement. However, the spatial pattern of the remain-
ing number of criminal events had shifted to particular areas. As such, mapping
crime prevention at the local level can allow the researcher and/or practitioner to
identify next steps for the prevention of further criminal activity.
The second example is also in the context of the police foot patrol in North
Vancouver, Canada. In this example, Andresen (2015) shows how the use of the
more standard mapping technique, kernel density mapping, does not provide much
insight regarding any changes in the spatial patterns of criminal events, but another
local spatial statistical technique (local Moran’s I) does provide some insight. The
maps of these two techniques are shown in Fig. 9.3 and represent mischief, the
crime type most impacted by the police foot patrol—the results are similar for other
crime types that Andresen (2015) analyzed.
120 M.A. Andresen

Fig. 9.3 Kernel density, single versus dual Kernel. Source: Andresen (2015)

Comparing Fig. 9.3a, b, very little information can be obtained regarding any
change in the spatial pattern of criminal events. There is a hot spot in both maps at
the southern portion of the study area (contained within the primary patrol area)
with moderate density locations shifting around slightly. If anything is to be identi-
fied, it is the emergence of a moderate intensity “hot” spot in the west side of the
study area. Comparing Fig. 9.3c, d, there also is not a lot of change that occurs.
However, the information is far more specific in the areas of potential concern. The
location in the western section of the study area has actually lost its statistical sig-
nificance as a hot spot (criminal event cluster) and the criminal event cluster area in
the southern portion of the primary patrol area has expanded. This is far more spe-
cific, and curious, information provided than the kernel density maps and may be
because the kernel density maps smooth out the criminal events, as discussed above.
This example shows that the use of one crime mapping technique, particularly in
the context of evaluating a crime prevention initiative, is probably not a good idea.
This does not show that one technique is wrong and one technique is correct, but
that more than one method of analysis should be undertaken to investigate/evaluate
crime prevention. If there is consistency between the multiple methods then the
9 Mapping Crime Prevention: What We Do and Where We Need to Go 121

researcher and/or practitioner can have confidence in the results. If not, caution
must be undertaken before further action is taken.

Sophistication Without Complexity

In their seminal article that outlined the fundamental elements of a criminal event,
Cohen and Felson (1979) stated that a crime occurs when a motivated offender, a
suitable target, and the lack of a capable guardian converge in space and time. As
such, understanding space (where) and time (when) are critical for understanding
crime patterns and, hence, crime prevention. Given that mapping is fundamentally
spatial (it is temporal as well as we shall see below), mapping crime prevention
must pay particular attention to where and when crime occurs.
Similar to the use of local analysis, just knowing where criminal events occur
may be misleading, or at least limited, when trying to understand the crime
problem. As discussed further below, with an example, knowing when criminal
events occur is particularly important. This may seem to be an odd statement in
the context of mapping for crime prevention but it is far more important than
usually thought.
As outlined by Hirschfield (2005), it is not just where and when criminal events
occur that matters, but the combinations of the two: in those places where criminal
events occur, when are they occurring, and at those times when criminal events are
occurring, where are they occurring? These two different questions may lead to very
different interventions. For example, if a location of criminal activity is to be tar-
geted because it is a “known” problem, in order to get the best results from a crime
prevention intervention that intervention should be developed and implemented
considering the timing of those criminal events. If assaults are the problem and
there is a drinking establishment in the area, the timing of most of the assaults will
probably be Friday and Saturday evenings. As such, a crime prevention initiative
that (intentionally or unintentionally) targets a time frame that considers when teen-
agers are out of school (weekdays, 3–6pm) may not be effective. Mapping different
time frames can help with this identification.
It is also possible that a high volume of criminal activity occurs on particular
days and at particular times. If this is the case, all criminal events (or those of inter-
est by the researcher and/or practitioner) that occur within that time range should be
selected and mapped for prevention purposes. It is quite possible that the locations
that emerge will be different than where they would be expected. Perhaps not, but
this needs to be investigated properly to prevent the misuse of scarce resources in
crime prevention. Such different forms of analyses provide much more sophistica-
tion to the mapping of crime prevention without becoming complicated. More maps
will have to be generated, but this extra information will be most instructive for the
prevention of crime.
122 M.A. Andresen

The Importance of Temporal Aggregation for Mapping Crime


Prevention: The Importance of When

As stated above, routine activity theory highlights the importance of the conver-
gence in space and time for understanding criminal events. Of course, when Cohen
and Felson (1979) were discussing these concepts they were speaking of very spe-
cific moments when motivated offenders and suitable targets converged because
they were interested in understanding specific criminal events, and the aggregation
of those events to understand crime patterns over time. However, we must also
consider time when mapping crime, for crime prevention or not. Most often, this
temporal consideration involves a time frame of criminal events to place on a map:
do we map crime for a particular year, particular season, or a particular month?
This choice obviously depends on the context of the investigation, but it does have
some implications for mapping crime prevention that need to be explicated.
The literature that most commonly investigates these implications is that of sea-
sonality and crime. Such investigations have gone back to early nineteenth century
France (Quetelet, 1842) and find that a seasonal pattern is quite often present—see
Andresen and Malleson (2013) for a discussion of the different studies that do and
do not exhibit seasonality in their criminal event data. If seasonality is present, there
is usually a peak of some degree during the summer months that is explained using
routine activity theory: during the summer months the weather is nicer (warmer and
drier), children are out of school, and people take vacations that all leads to an
increase in motivated offenders and suitable targets converging in time and space
with the lack of capable guardians (Andresen & Malleson, 2013). Because of this
phenomenon, anyone evaluating a crime prevention initiative must consider the
timing of that initiative because measuring the presence of before and after effects
could simply be because of changing of the seasons. But there is another consider-
ation when mapping crime prevention.
The other consideration is that crime patterns may change at different times of the
year, or even different days of the week. As such, when mapping where criminal events
are occurring, the researcher or practitioner must also consider when criminal events are
occurring if s/he wishes to know the appropriate place (and time) to implement a crime
prevention initiative, as discussed above. Unfortunately, there is very little research in
this area to call on in order to show its importance. Overall, this research shows that
during the peak season(s) for criminal events, there is a particular spatial pattern that
emerges: criminal events appear to increase disproportionately in areas/neighborhoods
that are of low socioeconomic status (Breetzke & Cohn, 2012; Ceccato, 2005; Harries
& Stadler, 1983; Harries, Stadler, & Zdorkowski, 1984). Though instructive, this may
not have any particular implications for mapping crime prevention because criminal
events are most often over-represented in these areas anyway.
In an investigation of changing spatial patterns for different seasons and a variety of
crime types, Andresen and Malleson (2013) found that the spatial patterns of crime were
quite dissimilar at different times of the year—only sexual assault and robbery at small
spatial units of analysis (dissemination areas, equivalent to the census block group) had
little change in their spatial patterns from season to season. As shown in Fig. 9.4,
9 Mapping Crime Prevention: What We Do and Where We Need to Go 123

Fig. 9.4 Spatial changes in crime patterns, summer versus yearly aggregate. Source: Andresen
and Malleson (2013)
124 M.A. Andresen

criminal events increased during the summer in very predictable places: the beach,
the central business area, large parks, large shopping centers, and the summer
fair (PNE).
This analysis shows that if crime prevention is truly the goal during a crime
prevention initiative—it could be community building, for example—the time
frame chosen to identify problem areas is very important. Do you want to decrease
criminal events based on criminal events that occur over the course of the year? Or
do you want to focus on criminal events that occur during a particular time of year
because that is when most of the criminal events occur? It should be clear that the
choice made will impact the effectiveness of the crime prevention initiative, because
if the “wrong” or “inappropriate” criminal events are mapped, that crime prevention
initiative may be implemented incorrectly because the prevention of crime is highly
situational (Clarke, 1980, 1983, 1997, 2012).
The example shown here has used a rather coarse temporal unit of analysis, the
season. Criminal events are also known to have different temporal patterns by day
of the week (more assaults during the weekend, for example) and even within the
day (residential burglaries tend to occur during the day and commercial burglaries
during the night) (Andresen, 2014). If spatial patterns of crime vary based on these
temporal units of analysis, the implications for crime prevention become even
stronger. If we are to map for the purposes of crime prevention we must map the
temporal dimension properly as well.

Concluding Thoughts

This chapter considers the importance of mapping for crime prevention. Rather than
simply discussing how to undertake that mapping within any particular software
program, some considerations for what should be done are discussed as well as
some cautions for the various techniques are discussed.
I try to cover aspects of mapping for crime prevention that do not involve any
specialist knowledge. Of course some training in mapping software is neces-
sary, but no specialist training is required. All of these techniques are readily
available in packaged software, most of which in the most common ArcGIS
software program.
The most important aspect of these discussions to consider is that one does not
have to get into complicated analyses when mapping for crime prevention. However,
more sophisticated analyses will prove to be instructive to understand the local
crime problem as well as provide more precise guidance as how to address that
problem with a crime prevention initiative. Mapping for crime prevention is a fruit-
ful endeavor that can very easily aid in the effective use of scarce resources for the
prevention of crime.
9 Mapping Crime Prevention: What We Do and Where We Need to Go 125

References

Andresen, M. A. (2009). Testing for similarity in area-based spatial patterns: A nonparametric


Monte Carlo approach. Applied Geography, 29(3), 333–345.
Andresen, M. A. (2014). Environmental criminology: Evolution, theory, and practice. New York,
NY: Routledge.
Andresen, M. A. (2015). Identifying changes in spatial patterns from police interventions: the
importance of multiple methods of analysis. Police Practice and Research, 16(2), 148–160.
Andresen, M. A., & Jenion, G. W. (2010). Ambient populations and the calculation of crime rates
and risk. Security Journal, 23(2), 114–133.
Andresen, M. A., & Lau, K. C. Y. (2014). An evaluation of police foot patrol in Lower Lonsdale.
Police Practice and Research, 15(6), 476–489.
Andresen, M. A., & Malleson, N. (2013). Crime seasonality and its variations across space.
Applied Geography, 43, 25–35.
Andresen, M. A., & Malleson, N. (2014). Police foot patrol and crime displacement: A local analy-
sis. Journal of Contemporary Criminal Justice, 30(2), 186–199.
Barr, R., & Pease, K. (1990). Crime placement, displacement and deflection. Crime and Justice: A
Review of Research, 12, 277–318.
Breetzke, G. D., & Cohn, E. G. (2012). Seasonal assault and neighborhood deprivation in South
Africa: Some preliminary findings. Environment & Behavior, 44(5), 641–667.
Ceccato, V. (2005). Homicide in San Paulo, Brazil: Assessing spatial-temporal and weather varia-
tions. Journal of Environmental Psychology, 25(3), 307–321.
Clarke, R. V. G. (1980). Situational crime prevention: Theory and practice. British Journal of
Criminology, 20(2), 136–147.
Clarke, R. V. (1983). Situational crime prevention: Its theoretical basis and practical scope. Crime
and Justice: An Annual Review of Research, 4, 225–256.
Clarke, R. V. (1997). Situational crime prevention: Successful case studies (2nd ed.). Monsey, NY:
Criminal Justice Press.
Clarke, R. V. (2012). Opportunity makes the thief. Really? And so what? Crime Science, 1, Article
3.
Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 44(4), 588–608.
Eck, J. E. (1993). The threat of crime displacement. Criminal Justice Abstracts, 25(3), 527–546.
Eck, J. E., Chainey, S., Cameron, J. G., Leitner, M., & Wilson, R. E. (2005). Mapping crime:
Understanding hot spots. Washington, DC: National Institute of Justice, Office of Justice
Programs, United States Department of Justice.
Fotheringham, A. S., & Wong, D. W. S. (1991). The modifiable areal unit problem in multivariate
statistical analysis. Environment and Planning A, 23(7), 1025–1044.
Harries, K. D., & Stadler, S. (1983). Determinism revisited: Assault and heat stress in Dallas.
Environment & Behavior, 15(2), 235–256.
Harries, K. D., Stadler, S., & Zdorkowski, R. (1984). Seasonality and assault: Explorations in
inter-neighborhood variation, Dallas 1980. Annals of the Association of American Geographers,
74(4), 590–604.
Hesseling, R. B. P. (1994). Displacement: A review of the empirical literature. Crime Prevention
Studies, 3, 197–230.
Hirschfield, A. (2005). Analysis for intervention. In N. Tilley (Ed.), Handbook of crime prevention
and community safety (pp. 629–673). Cullompton, UK: Willan Publishing.
Openshaw, S. (1984a). Ecological fallacies and the analysis of areal census data. Environment and
Planning A, 16(1), 17–31.
Openshaw, S. (1984b). The modifiable areal unit problem. CATMOG (Concepts and Techniques
in Modern Geography) (vol. 38). Norwich: Geo Books.
Quetelet, L. A. J. (1842). A treatise on man and the development of his faculties. Edinburgh: W.
and R. Chambers.
126 M.A. Andresen

Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American


Sociological Review, 15(3), 351–357.
Santos, R. B. (2013). Crime analysis with crime mapping (3rd ed.). Thousand Oaks, CA: Sage
Publications.
Schmid, C. F. (1960a). Urban crime areas: Part I. American Sociological Review, 25(4), 527–542.
Schmid, C. F. (1960b). Urban crime areas: Part II. American Sociological Review, 25(5),
655–678.
Tilley, N. (2005). Handbook of crime prevention and community safety. Portland, OR: Willan
Publishing.
Weisburd, D., Wyckoff, L. A., Ready, J., Eck, J. E., Hinkle, J. C., & Gajewski, F. (2006). Does
crime just move around the corner? A controlled study of spatial displacement and diffusion of
crime control benefits. Criminology, 44(3), 549–591.
Wooldredge, J. (2002). Examining the (ir)relevance of aggregation bias for multilevel studies of
neighborhoods and crime with an example of comparing census tracts to official neighbor-
hoods in Cincinnati. Criminology, 40(3), 681–709.
Part IV
Social Network Analysis
Chapter 10
Social Network Analysis and Crime
Prevention

Giulia Berlusconi

Introduction

Social network analysis (hereafter SNA) has been increasingly adopted by both
criminologists and law enforcement agencies to study crime. It refers to the analysis
of the patterns of social interactions among actors and how these patterns influence
individual behaviors (Wasserman & Faust, 1994). In criminology, SNA has proven
to be a valuable tool for the study of personal and neighborhood networks and their
influence on crime and for the analysis of criminal groups.
This chapter discusses the network approach in criminology and the various
areas of the application of SNA in the criminological context. It describes how SNA
is a valuable tool not only for research but also for law enforcement purposes.
Within law enforcement agencies, SNA methods and techniques have mainly been
used for the purpose of crime enforcement. In the context of this contribution, it is
argued that social network methods and techniques have great value for crime pre-
vention as well.

Crime and Social Network Analysis

Less than two decades ago, Nigel Coles (2001) suggested that it’s not what you
know, but who you know that counts (emphasis in the original). Coles was convinc-
ingly arguing for the adoption of SNA in the study of crime and, in particular, of
criminal groups. Indeed, “social network analysis has the very real potential to
uncover the complexities of criminal networks” (Coles, 2001, p. 581). This poten-
tial exists, according to the author, not only in a set of methodological techniques

G. Berlusconi (*)
Università Cattolica del Sacro Cuore—Transcrime, Milan, Italy
e-mail: giulia.berlusconi@unicatt.it

© Springer International Publishing Switzerland 2017 129


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_10
130 G. Berlusconi

but also in two valuable insights: (1) every individual is part of a larger social sys-
tem whose participants influence his behavior and (2) the pattern of interaction with
other actors often follows certain rules or regularities (Coles, 2001; see also
Galaskiewicz & Wasserman, 1994; Knoke & Kuklinski, 1991). Analyzing such
regularities could help unpack criminal groups, their activities, and organizations.
While suggesting the great potential of SNA for studying organized crime, Coles
(2001) complained about the failure of criminologists to adopt SNA’s concepts and
methods to better understand criminal groups. Ten years later, in a review of the
applications of SNA in criminology, Papachristos (2011) sadly acknowledged that
only a very limited number of papers adopting network analysis had been published
in criminology journals, compared to sociology and public health papers. This
neglect was even more striking considering the number of network-related concepts
and images that permeate criminological theories (e.g., social control, peer
influence).
The skepticism by criminologists towards SNA may be partially due to the fact
that it implies a departure from most traditional regression-oriented approaches that
assume independence among the variables (Emirbayer, 1997; Papachristos, 2011).
SNA considers social actors as interacting units. Furthermore, these interactions
shape individuals’ behaviors, including deviant ones. Therefore, from a theoretical
point of view, SNA emphasizes the interdependence among actors rather than their
independence. As a consequence, the independence assumption as the basis of most
causal analyses using regression models is no longer valid and new methods have to
be applied (Papachristos, 2011).
From a methodological point of view, SNA encompasses a range of techniques
rooted in mathematical graph theory (Carrington, Scott, & Wasserman, 2005;
Wasserman & Faust, 1994). Social networks consist in a set of nodes (or vertices)
and the edges (or lines) among them. The former can be represented by individuals,
groups, or even countries, while the latter exist when a relation is present between
two nodes. Edge connecting nodes may have different meanings (e.g., friendship,
exchange of information, co-offending) and may have various properties. Indeed,
the weight or the direction of the edges may be specified, for example, by counting
the number of telephone calls between two actors or by recording the sender and the
receiver of the call (Scott, 2000; Wasserman & Faust, 1994).
In their reviews of studies applying SNA in criminology and its methods and data
collection strategies, various authors have demonstrated the versatility of this
approach (Calderoni, 2014b; Carrington, 2011; Grannis, 2014; Haynie & Soller,
2014; McGloin & Kirk, 2010; Papachristos, 2006, 2011; Piquette, Smith, &
Papachristos, 2014; Radil, 2014). Carrington (2011) identifies three main areas of
application of SNA in criminological research.
First, SNA can be adopted to research the influence of personal networks on
crime and, more generally, on delinquent behavior. This category includes studies
on the consequences that personal networks may have on juvenile delinquency or,
in fewer cases, on adult criminality (e.g., Ennett et al., 2006; Haynie, 2001, 2002;
Kreager & Haynie, 2011; McGloin, 2009; Payne & Cornwell, 2007). Such analyses
use social networks as independent variables to explain crime across the population.
10 Social Network Analysis and Crime Prevention 131

Indeed, the structure of social relations (i.e., the attributes of personal networks) is
considered an explanatory variable of crime and delinquent behavior in addition to
individual characteristics (Haynie & Soller, 2014; Papachristos, 2011).
A second category of studies adopting SNA in criminological research encom-
passes the analyses of neighborhood networks and their influence on crime
(Carrington, 2011). Recognizing the importance of the “neighborhood effects” on
crime (Sampson, Morenoff, & Gannon-Rowley, 2002), these studies adopted SNA
to assess the facilitating and deterring effects that the structural properties of
neighborhood networks have on crime and delinquency rates (Soller & Browning,
2014). Recent studies also focused on interactions among neighborhoods and the
effects of these interactions on crime by integrating spatial and social network
methods. Drawing on the idea that “observable outcomes in one neighborhood are
partly the product of social actions and activities that can stretch beyond local
communities” (Radil, 2014, p. 4995; see also Morenoff, Sampson, & Raudenbush,
2001), some authors have started to model proximity not only in geographical
terms but also in social network terms (e.g., Tita & Radil, 2010). Levels of crime
in a neighborhood are thus influenced by features of both geographically proxi-
mate locations and places connected by social ties (e.g., rivalries among gangs)
(Radil, 2014).
Finally, SNA has been adopted to explore and model the organization of crime
(Carrington, 2011). Street gangs (e.g., McGloin, 2007; Papachristos, 2006), terrorist
groups (e.g., Krebs, 2001; Rothenberg, 2002), and organized crime groups (e.g.,
Bright, Caitlin, & Chalmers, 2012; Calderoni, 2012; Campana, 2011; Morselli,
2009; Natarajan, 2006; Varese, 2013), as well as illicit markets and co-offending
networks (e.g., Bichler, Schoepfer, & Bush, 2015; Heber, 2009; Malm, Bichler, &
Nash, 2011), have been analyzed through the lens of SNA. In recent years, scholars
have also adopted SNA to study online networks (e.g., Décary-Hétu, 2014a; Décary-
Hétu & Dupont, 2012). In this context, the network is analyzed as a dependent vari-
able and social network techniques are adopted to describe the criminal group and
its main structural properties (Papachristos, 2011). Indeed, SNA makes it possible
to seek, rather than assume, the structure of a criminal group (Morselli, 2009).
Therefore, within this approach, all types of organizations, from decentralized to
hierarchical, are conceived as networks of collaborating criminals, and the struc-
tural properties of such networks are subject to analysis through social network
measures.
As Carrington (2011, p. 244) notes, “this research tends to be exploratory and
descriptive rather than theory-testing,” although SNA also allows scholars to test
hypotheses from the literature and compare criminal networks across countries or
markets (Calderoni, 2014b; Papachristos, 2011). For instance, the idea that hierar-
chy does not play a pivotal role in mafia associations in specific circumstances such
as international drug trafficking can be tested and possibly supported by empirical
studies (see Calderoni, 2012). In recent years, scholars have started to adopt statistic
models for social networks to identify the social processes underlying illicit net-
work formation and development as a consequence of both endogenous and exog-
enous factors (Bichler & Franquez, 2014; Boivin, 2014; Everton & Cunningham,
132 G. Berlusconi

2014; Berlusconi, Aziani, & Giommoni, 2015). Therefore, SNA also fosters an
understanding of the etiology and dynamics of criminal groups and illicit trade
networks.
The adoption of SNA in criminology has proven to be a valuable tool for various
reasons. First, it can help refine criminological theories founded upon the idea that
personal—and neighborhood—networks play a role in the etiology of deviance and
crime (Papachristos, 2014). Second, it helps understand complex organizations such
as organized crime and terrorist groups. As a consequence, academics and law
enforcement agencies have adopted SNA for crime enforcement purposes
(Calderoni, 2014b).

Social Network Analysis and Crime Enforcement

If scholars were skeptical about the adoption of SNA to study crime, others
directly involved in criminal intelligence recognized its potential for the analysis
of criminal networks (Coles, 2001). In this context, SNA has been considered a
valuable tool almost exclusively to study criminal groups for the purpose of
crime enforcement.
Since the 1970s, law enforcement agencies have increasingly applied link analy-
sis “to portray the relationships among suspected criminals, to determine the struc-
ture of criminal organizations, and to identify the nature of suspected criminal
activities” (Harper & Harris, 1975, p. 157). Link analysis is adopted for both tactical
and strategic intelligence analysis, as it allows the identification of connections
among individuals using information on activities, events, and places (Sparrow,
1991a; Strang, 2014; Van der Hulst, 2009). The output of this type of analysis is a
two-dimensional representation of actors and the relations among them. The stron-
ger the relationship between two actors, the closer they will be displayed in the
graph representing the network (McAndrew, 1999).
Link analysis has the advantage of visually representing criminal networks start-
ing with the law enforcement data. However, this analysis is not interested in the
mechanisms behind the formation and persistence of criminal networks.
Furthermore, since it does not entail any mathematical computation, the way con-
nections among actors are represented in the graph are likely to influence the under-
standing of the network under analysis. For instance, actors in the center or at the
top of the graph may be interpreted as central actors or leaders regardless of their
values of degree or betweenness centrality (Klerks, 2001; McGrath, Blythe, &
Krackhardt, 1997; Van der Hulst, 2009).1 Instead, SNA enables the analysis of the
structural properties of criminal networks through a set of measures based on

1
Degree and betweenness measure an actor’s centrality within a network. Degree centrality mea-
sures the number of nodes with which each node is connected. Betweenness centrality measures
the extent to which a node lies on the shortest path between any two other nodes (Wasserman &
Faust, 1994).
10 Social Network Analysis and Crime Prevention 133

mathematical graph theory. Network properties are thus calculated rather than
inferred from a graph representing the criminal group (Davis, 1981; McAndrew,
1999).
Since the 1980s, the adoption of SNA for strategic analysis and enforcement of
criminal groups has been advocated by several authors, including law enforcement
analysts (Davis, 1981; Lupsha, 1980, 1983). In the following years, the adoption of
SNA to criminal intelligence was also promoted by scholars researching organized
crime. A common argument for its adoption was that its techniques allow an in-
depth analysis of the internal configuration of criminal groups and are thus valuable
for research, intelligence, and investigation. However, in most cases, such claims
were not supported by empirical analyses of criminal organizations (Ianni & Reuss-
Ianni, 1990; McAndrew, 1999; Sparrow, 1991b; Van der Hulst, 2009).
Van der Hulst (2009) identified various fields of crime enforcement in which
SNA could provide valuable insights. SNA could help identify the key actors to be
removed from the network to achieve destabilization and predict the impact of their
removal as a consequence of an arrest by law enforcement agencies. Social network
techniques could also help identify aliases through the analysis of actors with simi-
lar patterns of connections, especially in the case of large investigations, and they
could provide evidence for prosecution. Furthermore, SNA may help identify poten-
tial defectors according to their position in the network (Faulkner & Cheney, 2015).
SNA is currently adopted by law enforcement agencies. Duijn and Klerks (2014)
describe the Dutch experience and the benefits for intelligence and investigation.
According to the authors, SNA is particularly useful in guiding operational intelli-
gence projects with the aim to identify strategies to target and disrupt criminal net-
works. For instance, the analysis of the topology of a network enables to define the
targeting strategy that is likely to lead to the maximum of network disruptions (Xu
& Chen, 2008).
Despite evidence of positive experiences with the adoption of network tech-
niques for intelligence and investigation, law enforcement analysts also show some
skepticism towards SNA, mainly because they do not observe any significant
advantage offered by current applications to crime enforcement, especially for long-
term investigations in which police agencies managed to gather detailed knowledge
on the suspects from different sources (e.g., wiretapping, background checks)
(Calderoni, 2014b). Nonetheless, SNA can be useful for intelligence collection. In
particular, it can aid ongoing investigations in identifying key individuals and sub-
groups within a larger network of co-offenders, and suggesting effective strategies
for network disruption (Strang, 2014).
For instance, Calderoni (2014a) described how SNA can be useful to identify mafia
bosses with limited information. Contrary to most network studies on criminal groups,
the author retrieved the network structure using information on meetings among co-
offenders, which is easily accessible also at the preliminary stages of a criminal inves-
tigation and is not conditioned by a court order. In the context of a hierarchical
organization such as the Italian ‘Ndrangheta, individual positions within the network
(e.g., degree and betweenness centrality scores) can help identify the leaders of the
criminal group with limited information and resources (Calderoni, 2014a).
134 G. Berlusconi

SNA also supports strategies to disrupt criminal organizations (Strang, 2014).


The benefit of the adoption of network techniques is twofold: the vulnerabilities of
criminal networks to different types of attacks can be identified, and destabilization
strategies can be selected and tested through simulations. Indeed, not all networks
are equally vulnerable to attacks. Disruption strategies cannot be similarly applied
to all criminal organizations; rather, they should be established by considering the
variations in the structure of criminal groups (Malm & Bichler, 2011; Malm,
Bichler, & Van De Walle, 2010; Xu & Chen, 2008).
SNA enables analysts to assess the topological features of criminal networks and
thus their vulnerability to attacks based on both network measures and individual
characteristics. Through the simulation of different types of attacks, it has been
demonstrated that the removal of bridges (i.e., actors with high betweenness cen-
trality scores) is likely to cause more damage to criminal networks. Random attacks,
i.e. the random removal of nodes, are instead not likely to cause the disruption, or
even the fragmentation, of such networks (Keegan, Ahmed, Williams, Srivastava, &
Contractor, 2010; Xu & Chen, 2008).
SNA has proven to be a useful tool for crime enforcement. Indeed, it provides a
set of measures that enable law enforcement agencies to identify actors with a
prominent position within a criminal network and assists them in selecting the best
strategies for network disruption. Recent developments in the application of SNA in
criminological research suggest that it can be a valid instrument for crime preven-
tion as well.

Moving Forward: SNA and Crime Prevention

Besides favoring advancements in research on crime and supporting law enforce-


ment agencies during criminal investigations, SNA can assist in preventing crime
and emergent crime problems (Strang, 2014). Recent studies provide good exam-
ples of the value of SNA for crime prevention.
If we consider social networks as independent variables to explain crime and
victimization across the population, SNA can be adopted to study crime epidemics
and identify individuals who are more likely to be involved in gunshot episodes. If
networks are analyzed as dependent variables, the adoption of a network approach
can help understand the etiology and dynamics of criminal groups and networks.
SNA enables analysts to predict leadership roles within criminal organizations, as
well as the implications of disruption strategies, thus avoiding network reorganiza-
tion and the committing of more crimes. Furthermore, the adoption of a network
approach to gang-related violence can help to identify targeted prevention strategies
to reduce homicides and nonlethal shootings.
SNA has been used to explain how personal networks influence delinquent
behavior. Similarly, SNA can help identify individuals who are more likely to be a
victim of gunshot injuries as a consequence of both personal characteristics and
those of their social networks (Papachristos, Braga, & Hureau, 2012). The traditional
10 Social Network Analysis and Crime Prevention 135

criminological approach to crime epidemics—i.e., the dramatic increase in crime in


a specific location and within a restricted period of time—evaluates an individual’s
risk of victimization based on a series of individual, situational, and community risk
factors. However, crime is highly concentrated within populations and neighbor-
hoods characterized by the presence of such risk factors (Papachristos, 2011).
Papachristos and colleagues (Papachristos et al., 2012; Papachristos, Wildeman,
& Roberto, 2015) explored the relationship between social networks and the risk of
gunshot injury and demonstrated how SNA can help assess an individual’s risk of
being a crime victim or an offender, by analyzing his/her personal network. Evidence
of a relationship between social distance and gun victimization was found in a study
conducted in Boston’s Cape Verdean community. Indeed, “the closer one is to a
gunshot victim, the greater the probability of one’s own victimization, net of indi-
vidual and other network characteristics” (Papachristos et al., 2012, p. 999). Another
study conducted in Chicago confirmed the association between the presence of gun-
shot victims in one’s social network and his/her probability of victimization
(Papachristos et al., 2015).
These studies demonstrate the utility of SNA to understand the risk of gun vio-
lence in urban areas and identify proper prevention strategies. Recognizing that the
risk of victimization is highly concentrated within communities and associated with
specific behaviors such as co-offending allows to direct prevention efforts towards
specific segments of the population, instead of targeting the population at large or
high-risk neighborhoods. SNA may thus support violence reduction strategies by
redirecting resources to specific locations and segments of the population
(Papachristos et al., 2012, 2015).
SNA can also be adopted to predict an individual’s future involvement in gang-
related activities after his/her participation in a murder. McCuish, Bouchard, and
Corrado (2015) studied a homicide co-offending network whose members were part
of the Canadian BC gang and found that homicide offenders were not recruited
within the co-offending network, but they had high-ranking positions after the
homicides. Therefore, involvement in gang homicides appears to be relevant for the
criminal career of gang members. Following the criminal trajectories of adolescents
involved in homicides after their release could help concentrate investigative efforts
and adopt preventive measures targeting these individuals.
The adoption of a network approach can also help in understanding the structural
properties of criminal groups and networks, as well as their etiology and dynamics.
SNA can be used not only to identify leadership roles within criminal groups but
also to predict such roles at the early stages of an investigation. Calderoni expanded
on a previous study (Calderoni, 2014a) and demonstrated how using only informa-
tion on meeting attendance, which is more easily accessible than, for example, wire-
tap records, makes it possible to predict criminal leadership and thus prevent
criminal groups from operating (Calderoni, 2015).
The author analyzed a criminal group belonging to the ‘Ndrangheta through the
network based on attendance at mafia meetings. The investigation was divided into
four time periods of increasing duration. A logistic regression model was run for
each of the four periods with leadership (leader = 1) as the dependent variable and
136 G. Berlusconi

centrality measures as the independent variables, controlling for a prosecution bias


and the number of meetings attended by each actor. The results showed that after the
first year of investigation and only 34 meetings, it was possible to successfully
identify the role of 75.6 % of all actors involved in the criminal investigation
(Calderoni, 2015).
SNA may thus assist law enforcement agencies in the early identification of
relevant actors and in the selection of individuals to be wiretapped. It could also
enable “the adoption of preventive measures with the aim of hindering the activi-
ties of criminal organizations” (Calderoni, 2015, p. 105). Indeed, by identifying
the leaders in advance, it would be possible to concentrate investigative efforts
and allocate resources for the electronic surveillance of a limited number of
persons, as well as to adopt personal preventive measures or apply special
surveillance orders.
The development of statistical models to study network evolution may help
criminologists go beyond mere descriptive analyses and understand the dynamics
of criminal groups. SNA is adopted to support ongoing criminal investigations
with the purpose of identifying potential network vulnerabilities and strategies for
disruption (Strang, 2014). However, little information is available on the conse-
quences that different disruption strategies may have on the structure and activi-
ties of criminal groups (Duijn & Klerks, 2014). Traditional enforcement strategies
targeting leaders are not always applicable, especially in the case of loose net-
works of collaborating criminals. Moreover, the removal of critical nodes does
not automatically entail an increase in the vulnerability of a criminal organization,
or its disruption, because network flexibility and high turnover may reduce the
effects of law enforcement targeting (Bright, Greenhill, & Levenkova, 2014;
Carley, Lee, & Krackhardt, 2002; Décary-Hétu, 2014b; Morselli, Giguere, &
Petit, 2007). Hence, the impact of law enforcement interventions is not necessar-
ily negative because it may result in better adaptation of the targeted criminal
group rather than its disruption (Ayling, 2009).
SNA can identify the structural properties of criminal networks and their
changes over time. By analyzing how criminal networks recover from arrests
and other law enforcement interventions (e.g., drug or asset seizures), SNA can
provide insights into the adaptation of criminal groups to law enforcement tar-
geting. Therefore, social network techniques may help assess the implications
of different disruption strategies and select those that limit network reorganiza-
tion and the resulting committing of more crimes (Berlusconi, 2014; Bright,
2015). Indeed, analyzing criminal networks and their changes over time as a
consequence of law enforcement interventions enables analysts to predict the
impact of the arrest of key actors on the structure and activities of such net-
works (Ianni & Reuss-Ianni, 1990).
Recently, longitudinal modeling techniques have been applied to analyze the
relationship between the structure and resilience of criminal networks and law
enforcement strategies. Everton and Cunningham (2014) demonstrated how
approaches that use non-coercive means can shape the structure of criminal net-
works and make them more vulnerable to further attacks which target critical nodes.
10 Social Network Analysis and Crime Prevention 137

Similarly, dynamic modeling has been applied to illicit trade networks, such as the
gun trade (Bichler & Franquez, 2014). Indeed, SNA can help identify the
consequences of various strategies to disrupt criminal groups and other types of
illegal networks, such as online child exploitation networks and other clandestine
online networks (Keegan et al., 2010; Kila & Bouchard, 2015).
Finally, SNA can help map the rivalries among gangs and the related violence for
the purpose of understanding the patterns of gang-related violence and predicting
future conflicts (Piquette et al., 2014). For instance, Descormiers and Morselli
(2011) analyzed how gang-level attributes (e.g., ethnicity) and other factors (e.g.,
proximity of gang turf) help to anticipate conflicts among gangs in Montreal.
Similarly, Papachristos (2009) focused on retaliation among gangs in Chicago and
described how patterns of gang-related violence can be explained by prior conflict
relations and the position of the gangs within rivalry networks.
Other studies modeled the effects of both geographic and network processes on
gang violence. They found that both spatial proximity and prior conflicts (i.e., rival-
ries among gangs) influence the patterns of gang violence (Brantingham, Tita,
Short, & Reid, 2012; Papachristos, Hureau, & Braga, 2013; Radil, Flint, & Tita,
2010; Tita & Radil, 2011). These studies consider networks as the dependent vari-
able, whereas explanatory variables include properties of the network (e.g., reci-
procity), of the nodes (e.g., size of the gang, ethnicity), and of the location (e.g.,
poverty level in a neighborhood), as well as spatial proximity of gangs’ turfs and
other social processes (Papachristos, 2011).
Analyses of gang-related violence adopting a network approach may suggest
interventions at the gang level that could prevent specific groups from perpetrating
violence against rival gangs (Papachristos, 2009). For instance, gang injunction
policies may be revisited to incorporate the results from the network analysis of
gang rivalries and alliances and to consider the event of a reorganization of the
gangs targeted by injunctions. Insights into the effects of intervention strategies may
also be acquired through the adoption of network models that simulate gang-related
violence and predict future events (Hegemann et al., 2011; Tita, Butts, Valasik, &
Brantingham, 2012).

Conclusions

Criminological studies adopting a network approach suffer from a number of limi-


tations and methodological problems. Much research is still exploratory or descrip-
tive in nature. Furthermore, scholars often rely on law enforcement data characterized
by missing information and fuzzy network boundaries (Berlusconi, 2013;
Carrington, 2011; Malm et al., 2010; McGloin & Kirk, 2010; Morselli, 2009; Von
Lampe, 2009).
Nonetheless, SNA has proven its value in refining criminological concepts and
theories to aid the understanding of social processes behind crime problems, and to
assist law enforcement agencies in enforcing and preventing crime. Future
138 G. Berlusconi

developments will hopefully go beyond a descriptive approach and will rely on


more complete and new data sets, thus being able to guide policy decisions and
crime prevention programs.

Acknowledgments I would like to thank Dr. Francesco Calderoni, Dr. Benoit Leclerc, and the
anonymous reviewer for their valuable comments on earlier versions of the manuscript.

References

Ayling, J. (2009). Criminal organizations and resilience. International Journal of Law, Crime and
Justice, 37(4), 182–196.
Berlusconi, G. (2013). Do all the pieces matter? Assessing the reliability of law enforcement data
sources for the network analysis of wire taps. Global Crime, 14(1), 61–81.
Berlusconi, G. (2014). Law enforcement and criminal network resilience: The impact of law
enforcement action on the structure of mafia-related drug trafficking networks (PhD Thesis).
Università Cattolica del Sacro Cuore, Milano. Retrieved from http://hdl.handle.net/10280/2457
Berlusconi, G., Aziani, A., & Giommoni, L. (2015). Does size matter? The determinants of heroin
flows in Europe. Presented at the 7th Annual Illicit Network Workshop, Montreal.
Bichler, G. M., & Franquez, J. (2014). Conflict cessation and the emergence of weapons supermar-
kets. In A. J. Masys (Ed.), Networks and network analysis for defence and security (pp. 189–
215). Cham: Springer.
Bichler, G. M., Schoepfer, A., & Bush, S. (2015). White collars and black ties: Interlocking
social circles of elite corporate offenders. Journal of Contemporary Criminal Justice, 31(3),
279–296.
Boivin, R. (2014). Risks, prices, and positions: A social network analysis of illegal drug trafficking
in the world-economy. International Journal of Drug Policy, 25(2), 235–243.
Brantingham, P. J., Tita, G. E., Short, M. B., & Reid, S. E. (2012). The ecology of gang territorial
boundaries. Criminology, 50(3), 851–885.
Bright, D. A. (2015). Identifying key actors in drug trafficking networks. In G. M. Bichler & A. E.
Malm (Eds.), Disrupting criminal networks: Network analysis in crime prevention (pp. 66–88).
Boulder, CO: FirstForumPress, a division of Lynne Rienner Publishers.
Bright, D. A., Caitlin, E. H., & Chalmers, J. (2012). Illuminating dark networks: A social network
analysis of an Australian drug trafficking syndicate. Crime, Law and Social Change, 57(2),
151–176.
Bright, D. A., Greenhill, C., & Levenkova, N. (2014). Dismantling criminal networks: Can node
attributes play a role? In C. Morselli (Ed.), Crime and networks (pp. 148–162). New York and
London: Routledge.
Calderoni, F. (2012). The structure of drug trafficking mafias: the ‘Ndrangheta and cocaine. Crime,
Law and Social Change, 58(3), 321–349.
Calderoni, F. (2014a). Identifying mafia bosses from meeting attendance. In A. J. Masys (Ed.),
Networks and network analysis for defence and security (pp. 27–48). Cham: Springer.
Calderoni, F. (2014b). Network analysis of organized criminal groups. In G. Bruinsma &
D. Weisburd (Eds.), Encyclopedia of criminology and criminal justice (pp. 4972–4981).
New York, NY: Springer.
Calderoni, F. (2015). Predicting organised crime leaders. In G. M. Bichler & A. E. Malm (Eds.),
Disrupting criminal networks: Network analysis in crime prevention (pp. 89–110). Boulder,
CO: FirstForumPress, a division of Lynne Rienner Publishers.
Campana, P. (2011). Eavesdropping on the Mob: The functional diversification of Mafia activities
across territories. European Journal of Criminology, 8(3), 213–228.
10 Social Network Analysis and Crime Prevention 139

Carley, K. M., Lee, J.-S., & Krackhardt, D. (2002). Destabilizing networks. Connections, 24(3),
79–92.
Carrington, P. J. (2011). Crime and social network analysis. In J. Scott & P. J. Carrington (Eds.),
Sage handbook of social network analysis (pp. 236–255). London: SAGE Publications.
Carrington, P. J., Scott, J., & Wasserman, S. (Eds.). (2005). Models and methods in social network
analysis. Cambridge: Cambridge University Press.
Coles, N. (2001). It’s not what you know - it’s who you know that counts: Analysing serious crime
groups as social networks. British Journal of Criminology, 41(4), 580–594.
Davis, R. H. (1981). Social network analysis: An aid in conspiracy investigations. FBI Law
Enforcement Bulletin, 50(12), 11–19.
Décary-Hétu, D. (2014a). Information exchange paths in IRC hacking chat rooms. In C. Morselli
(Ed.), Crime and networks (pp. 218–230). New York, NY: Routledge.
Décary-Hétu, D. (2014b). Police operations 3.0: On the impact and policy implications of police
operations on the Warez scene. Policy & Internet, 6(3), 315–340.
Décary-Hétu, D., & Dupont, B. (2012). The social network of hackers. Global Crime, 13(3),
160–175.
Descormiers, K., & Morselli, C. (2011). Alliances, conflicts, and contradictions in Montreal’s
street gang landscape. International Criminal Justice Review, 21(3), 297–314.
Duijn, P. A. C., & Klerks, P. (2014). Social network analysis applied to criminal networks: Recent
developments in Dutch law enforcement. In A. J. Masys (Ed.), Networks and network analysis
for defence and security (pp. 121–158). Cham: Springer.
Emirbayer, M. (1997). Manifesto for a relational sociology. American Journal of Sociology,
103(2), 281–317.
Ennett, S. T., Bauman, K. E., Hussong, A., Faris, R., Foshee, V. A., Cai, L., et al. (2006). The peer
context of adolescent substance use: Findings from social network analysis. Journal of
Research on Adolescence, 16(2), 159–186.
Everton, S. F., & Cunningham, D. (2014). Terrorist network adaptation to a changing environment.
In C. Morselli (Ed.), Crime and networks (pp. 287–308). New York and London: Routledge.
Faulkner, R. R., & Cheney, E. R. (2015). Defection from a fraud network. In G. M. Bichler & A. E.
Malm (Eds.), Disrupting criminal networks: Network analysis in crime prevention (pp. 111–
128). Boulder, CO: FirstForumPress, a division of Lynne Rienner Publishers.
Galaskiewicz, J., & Wasserman, S. (1994). Advances in the social and behavioural sciences from
social network analysis. In S. Wasserman & J. Galaskiewicz (Eds.), Advances in social network
analysis: Research in the social and behavioural sciences (pp. xi–xvii). London: SAGE
Publications.
Grannis, R. (2014). Social network analysis and the measurement of neighborhoods. In
G. Bruinsma & D. Weisburd (Eds.), Encyclopedia of criminology and criminal justice
(pp. 4964–4972). New York: Springer.
Harper, W. R., & Harris, D. H. (1975). The application of link analysis to police intelligence.
Human Factors: The Journal of the Human Factors and Ergonomics Society, 17(2), 157–164.
Haynie, D. L. (2001). Delinquent peers revisited: Does network structure matter? American
Journal of Sociology, 106(4), 1013–1057.
Haynie, D. L. (2002). Friendship networks and delinquency: The relative nature of peer delin-
quency. Journal of Quantitative Criminology, 18(2), 99–134.
Haynie, D. L., & Soller, B. (2014). Network analysis in criminology. In G. Bruinsma & D. Weisburd
(Eds.), Encyclopedia of criminology and criminal justice (pp. 3265–3275). New York: Springer.
Heber, A. (2009). The networks of drug offenders. Trends in Organized Crime, 12(1), 1–20.
Hegemann, R. A., Smith, L. M., Barbaro, A. B. T., Bertozzi, A. L., Reid, S. E., & Tita, G. E. (2011).
Geographical influences of an emerging network of gang rivalries. Physica A: Statistical
Mechanics and its Applications, 390(21–22), 3894–3914.
Ianni, F. A. J., & Reuss-Ianni, E. (1990). Network analysis. In P. Andrews & M. B. Peterson (Eds.),
Criminal intelligence analysis (pp. 67–84). Loomis, CA: Palmer Enterprises.
140 G. Berlusconi

Keegan, B., Ahmed, M. A., Williams, D., Srivastava, J., & Contractor, N. (2010). Dark gold:
Statistical properties of clandestine networks in massively multiplayer online games. In
Proceedings of the Second IEEE International Conference on Social Computing (pp. 201–
208). Minneapolis, MN.
Kila, J., & Bouchard, M. (2015). Vulnerabilities in online child exploitation networks. In
G. M. Bichler & A. E. Malm (Eds.), Disrupting criminal networks: Network analysis in
crime prevention (pp. 153–176). Boulder, CO: FirstForumPress, a division of Lynne
Rienner Publishers.
Klerks, P. (2001). The network paradigm applied to criminal organizations: Theoretical nitpicking
or a relevant doctrine for investigators? Recent developments in the Netherlands. Connections,
24(3), 53–65.
Knoke, D., & Kuklinski, J. H. (1991). Network analysis: Basic concepts. In G. Thompson,
J. Frances, R. Levacic, & G. Mitchell (Eds.), Markets, hierarchies and networks: The co-
ordination of social life (pp. 173–182). London: SAGE Publications.
Kreager, D. A., & Haynie, D. L. (2011). Dangerous liaisons? Dating and drinking diffusion in
adolescent peer networks. American Sociological Review, 76(5), 737–763.
Krebs, V. E. (2001). Mapping networks of terrorist cells. Connections, 24(3), 43–52.
Lupsha, P. A. (1980). Steps toward a strategic analysis of organized crime. The Police Chief, (May
1980), 36–38.
Lupsha, P. A. (1983). Networks versus networking: Analysis of an organized crime group. In G. P.
Waldo (Ed.), Career criminals (pp. 59–87). Beverly Hills, CA: SAGE Publications.
Malm, A. E., & Bichler, G. M. (2011). Networks of collaborating criminals: Assessing the struc-
tural vulnerability of drug markets. Journal of Research in Crime and Delinquency, 48(2),
271–297.
Malm, A. E., Bichler, G. M., & Nash, R. (2011). Co-offending between criminal enterprise groups.
Global Crime, 12(2), 112–128.
Malm, A. E., Bichler, G. M., & Van De Walle, S. (2010). Comparing the ties that bind criminal
networks: Is blood thicker than water? Security Journal, 23(1), 52–74.
McAndrew, D. (1999). The structural analysis of criminal networks. In D. Canter & L. Alison
(Eds.), The social psychology of crime: Groups, teams and networks (pp. 51–94). Aldershot:
Ashgate Publishing.
McCuish, E. C., Bouchard, M., & Corrado, R. (2015). The search for suitable homicide co-
offenders among gang members. Journal of Contemporary Criminal Justice, 31(3), 319–336.
McGloin, J. M. (2007). The organizational structure of street gangs in Newark, New Jersey: A
network analysis methodology. Journal of Gang Research, 15(1), 1–34.
McGloin, J. M. (2009). Delinquency balance: Revisiting peer influence. Criminology, 47(2),
439–477.
McGloin, J. M., & Kirk, D. S. (2010). Social network analysis. In A. R. Piquero & D. Weisburd
(Eds.), Handbook of quantitative criminology (pp. 209–224). New York: Springer.
McGrath, C., Blythe, J., & Krackhardt, D. (1997). The effect of spatial arrangement on judgments
and errors in interpreting graphs. Social Networks, 19(3), 223–242.
Morenoff, J. D., Sampson, R. J., & Raudenbush, S. W. (2001). Neighborhood inequality, collective
efficacy, and the spatial dynamics of urban violence. Criminology, 39(3), 517–558.
Morselli, C. (2009). Inside criminal networks. New York, NY: Springer.
Morselli, C., Giguere, C., & Petit, K. (2007). The efficiency/security trade-off in criminal networks.
Social Networks, 29(1), 143–153.
Natarajan, M. (2006). Understanding the structure of a large heroin distribution network: A quan-
titative analysis of qualitative data. Journal of Quantitative Criminology, 22(2), 171–192.
Papachristos, A. V. (2006). Social network analysis and gang research: Theory and methods. In J. F.
Short & L. A. Hughes (Eds.), Studying youth gangs (pp. 99–116). Lanham, MD: AltaMira Press.
Papachristos, A. V. (2009). Murder by structure: Dominance relations and the social structure of
gang homicide. American Journal of Sociology, 115(1), 74–128.
10 Social Network Analysis and Crime Prevention 141

Papachristos, A. V. (2011). The coming of a networked criminology? Using social network analy-
sis in the study of crime and deviance. In J. McCord (Ed.), Beyond empiricism: Institutions and
intentions in the study of crime (pp. 101–140). New Brunswick, NJ: Transaction Publishers.
Papachristos, A. V. (2014). The network structure of crime. Sociology Compass, 8(4), 347–357.
Papachristos, A. V., Braga, A. A., & Hureau, D. M. (2012). Social networks and the risk of gunshot injury.
Journal of Urban Health: Bulletin of the New York Academy of Medicine, 89(6), 992–1003.
Papachristos, A. V., Hureau, D. M., & Braga, A. A. (2013). The corner and the crew: The influence
of geography and social networks on gang violence. American Sociological Review, 78(3),
417–447.
Papachristos, A. V., Wildeman, C., & Roberto, E. (2015). Tragic, but not random: The social con-
tagion of nonfatal gunshot injuries. Social Science & Medicine, 125(Jan. 2015), 139–150.
Payne, D. C., & Cornwell, B. (2007). Reconsidering peer influences on delinquency: Do less prox-
imate contacts matter? Journal of Quantitative Criminology, 23(2), 127–149.
Piquette, J. C., Smith, C. M., & Papachristos, A. V. (2014). Social network analysis of urban street
gangs. In G. Bruinsma & D. Weisburd (Eds.), Encyclopedia of criminology and criminal jus-
tice (pp. 4981–4991). New York, NY: Springer.
Radil, S. M. (2014). Spatial models and network analysis. In G. Bruinsma & D. Weisburd (Eds.),
Encyclopedia of criminology and criminal justice (pp. 4991–4997). New York, NY: Springer.
Radil, S. M., Flint, C., & Tita, G. E. (2010). Spatializing social networks: Using social network
analysis to investigate geographies of gang rivalry, territoriality, and violence in Los Angeles.
Annals of the Association of American Geographers, 100(2), 307–326.
Rothenberg, R. (2002). From whole cloth: Making up the terrorist network. Connections, 24(3), 36–42.
Sampson, R., Morenoff, J., & Gannon-Rowley, T. (2002). Assessing “Neighborhood Effects”:
Social processes and new directions in research. Annual Review of Sociology, 28, 443–478.
Scott, J. (2000). Social network analysis: A handbook (2nd ed.). London: SAGE Publications.
Soller, B., & Browning, C. R. (2014). Neighborhood effects and social networks. In G. Bruinsma
& D. Weisburd (Eds.), Encyclopedia of criminology and criminal justice (pp. 3255–3265).
New York, NY: Springer.
Sparrow, M. K. (1991a). Network vulnerabilities and strategic intelligence in law enforcement.
International Journal of Intelligence and Counterintelligence, 5(3), 255–274.
Sparrow, M. K. (1991b). The application of network analysis to criminal intelligence: An assess-
ment of the prospects. Social Networks, 13(3), 251–274.
Strang, S. J. (2014). Network analysis in criminal intelligence. In A. J. Masys (Ed.), Networks and
network analysis for defence and security (pp. 1–26). Cham: Springer.
Tita, G. E., Butts, C. T., Valasik, M., & Brantingham, P. J. (2012). Rivalries and reciprocity: A
relational event model of gang violence. Presented at the 68th American Society of Criminology
Annual Meeting, Chicago.
Tita, G. E., & Radil, S. M. (2010). Making space for theory: The challenges of theorizing space and
place for spatial analysis in criminology. Journal of Quantitative Criminology, 26(4), 467–479.
Tita, G. E., & Radil, S. M. (2011). Spatializing the social networks of gangs to explore patterns of
violence. Journal of Quantitative Criminology, 27(4), 521–545.
Van der Hulst, R. C. (2009). Introduction to social network analysis (SNA) as an investigative tool.
Trends in Organized Crime, 12(2), 101–121.
Varese, F. (2013). The structure and the content of criminal connections: The Russian Mafia in
Italy. European Sociological Review, 29(5), 899–909.
Von Lampe, K. (2009). Human capital and social capital in criminal networks: Introduction to the
special issue on the 7th Blankensee Colloquium. Trends in Organized Crime, 12(2), 93–100.
Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications. New York,
NY: Cambridge University Press.
Xu, J. J., & Chen, H. (2008). The topology of dark networks. Communications of the ACM,
51(10), 58–65.
Chapter 11
Using Social Network Analysis to Design
Crime Prevention Strategies: A Case Study
of Methamphetamine Manufacture
and Trafficking

David A. Bright

The conceptualization of organized crime and terrorist groups as networks


represents a shift away from characterizations of such groups as strict hierarchies
in which power and authority is concentrated in “leaders” or “bosses.” Research
converges on the notion that such groups are loosely affiliated and flexible (e.g.,
Morselli, 2009b; Williams, 2001), without a rigid hierarchy (but see Calderoni,
2012, 2014) and in which ties are being continually formed and broken over time
(e.g., Bright & Delaney, 2013; Carley, Dombroski, Tsvetovat, Reminga, &
Kamneva, 2003). Methodologically, the use of social network analysis (SNA) for
the study of criminal networks has grown in popularity over the last decade or so.
SNA has been employed by researchers to describe and analyze the structure and
operation of groups of individuals involved in organized crime and terrorism.
Researchers have utilized SNA to study a variety of criminal endeavors that are
characterized by group activity and criminal associations, for example, terrorism
(Everton, 2012; Krebs, 2002), drug trafficking (Bright, Hughes, & Chalmers, 2012;
Natarajan, 2000), motor vehicle theft and ringing operations (Morselli & Roy,
2008), Outlaw Motorcycle Gangs (Morselli, 2009a), and human trafficking
(Mancuso, 2013). SNA is also used by law enforcement agencies to facilitate intel-
ligence collection and analysis.
Much previous research that has used SNA to examine criminal networks has
focused on the social structure of networks (e.g., density and centrality) and on the
identification of key or important actors in criminal networks. Research results typi-
cally have implications for law enforcement interventions aimed at disrupting and
dismantling criminal networks (e.g., Bright, Greenhill, & Levenkova, 2014; Everton,
2012). For example, the targeted arrest of the most well-connected individuals or
key brokers may have the most significant impact on the overall structure of the

D.A. Bright (*)


Faculty of Arts and Social Sciences, School of Social Sciences, University of New South
Wales, Sydney, NSW 2052, Australia
e-mail: David.Bright@unsw.edu.au

© Springer International Publishing Switzerland 2017 143


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_11
144 D.A. Bright

network (Bright et al., 2014; Xu & Chen, 2009). SNA can also guide the collection
of intelligence by prioritizing surveillance targets (Schwartz & Rouselle, 2009).
One limitation of these detection and disruption strategies is that they can be trig-
gered only when a network is formed and once law enforcement agencies have col-
lected sufficient intelligence on the actors and connections between them.
Comparatively little attention has been paid to the capacity for SNA to reveal crime
prevention strategies which may be effective against organized crime groups such as
those involved in trafficking illicit drugs. In this chapter, I present a case study of a
methamphetamine manufacture/trafficking network and analyze the network using
SNA in conjunction with crime script analysis (CSA). The aim of the analysis is to
generate a set of crime prevention strategies. Although the crime prevention strate-
gies developed are directed specifically at the manufacture and trafficking of meth-
amphetamine, the strategies are likely to be generalizable to the manufacture and
trafficking of other synthetic drugs and to other forms of organized crime (e.g.,
people trafficking, illicit firearm trafficking, money laundering).
SNA is an analytic technique that conceptualizes social groups as networks.
SNA conceptualizes individuals in the network as nodes and the links or relation-
ships between individuals as ties. SNA can produce metrics about the network as a
whole (density, centralization), at the subgroup level (e.g., clustering, cliques) and
at the actor level (e.g., centrality). Centrality scores can provide information about
the relative influence, authority, and power of actors in the network. SNA can assist
in the identification of key actors in criminal networks. Key actors include those
who are connected to many other actors (high degree centrality) and those who act
as brokers between pairs of actors (high betweenness centrality). In criminal net-
works actors who are high on degree centrality may also be vulnerable by virtue of
the fact that they are so well connected and therefore more visible. The most strate-
gically positioned actors may therefore be those who are low on direct directedness
(low degree centrality) and high on brokerage potential (high betweenness central-
ity; see Morselli, 2010; Calderoni, 2014).
In this chapter, I use SNA as the primary method to describe the social connec-
tions and identify key actors in a case study of a criminal network. CSA is used as
a complementary strategy to describe social connections across crime scenes.
Crime scripts (Cornish, 1994) refer to the sequence of activities that must be acted
out to successfully commit a crime. The crime script approach disaggregates
organized crime into a series of connected events called “crime scenes” or
“niches” (Ekblom, 2003). As crime scripts usually comprise multiple scenes, and
can include several alternate iterations in each scene (Morselli & Roy, 2008), they
facilitate flexibility and innovation across the crime commission process. Crime
scenes or niches depend on interconnections between offenders (Ekblom, 2003);
each crime scene is supported by the loose, flexible networks of actors who popu-
late the scenes, collect the necessary resources, skills, and knowledge, and under-
take the activities necessary to fulfill the scripts (Bright & Delaney, 2013; Morselli
& Roy, 2008). SNA can be used in conjunction with CSA to identify and highlight
these types of relationships and associations (for example, see Bright & Delaney,
2013; Morselli & Roy, 2008).
11 Using Social Network Analysis to Design Crime Prevention Strategies… 145

Within crime scenes, actors may specialize in particular roles, with some inter-
changeability between roles (Bright & Delaney, 2013). As the number of roles
required tends to increase with the complexity of the crime, social networks become
an important source of the skills, resources, equipment, knowledge, and labor
required offender cooperatively complete the crime commission process. The pri-
mary sets of actors are likely to organize themselves into social networks that facili-
tate the exchange of skills and resources and which can minimize risks. The process
of forming branching networks facilitates the accumulation of both human capital
(skills, resources, knowledge) and social capital (connectedness, combining
resources in a way that facilitates the crime). Within each crime scene, actors will
choose situations/context in which they expect to be subjected to minimal (if any)
scrutiny and therefore low risk, or be able to corrupt officials.

Case Study: Methamphetamine Manufacture and Trafficking

Organized Crime and Crime Prevention Strategies

Drug law enforcement agencies tend to prioritize interventions that produce arrests,
convictions, and seizures (Chiu, Leclerc, & Townsley, 2011). Unfortunately, these
priorities and the exigencies of law enforcement practice mean that there is usually
no systematic long-term collection or analyses of data on the structures and pro-
cesses characteristic of criminal groups. Furthermore, reactive interventions such as
crackdowns, raids, search and seizure, and buy-busts (Mazerolle, Soole, & Rombouts,
2006) only target offenders who are already known to police and often only after
crimes have been committed. Law enforcement agencies rarely analyze criminal
groups beyond the immediate needs of operational interventions. To facilitate crime
prevention approaches against organized crime, scholars are now beginning to
explore the potential for social network analysis and related approaches to devise
crime prevention strategies for organized crime. There is some extant scholarly work
on crime prevention focused on organized crime. This work is used as a springboard
for the development of crime prevention strategies using the results of SNA.
Crime prevention is well established within criminology, but has mostly
focused on street crime and violent crime, rather than organized crime (for an
overview, see Sutton, Cherney, & White, 2013). According to Ekblom (2003), the
design of crime prevention strategies should proceed in three phases: (1) a com-
prehensive description and analysis of the crime problem; (2) a precise articula-
tion of scenes, actors, skills, and resources; and (3) consideration of an extensive
range of options to undermine the crime script. Although this process was designed
for more traditional crimes committed by individuals (e.g., property crime),
Ekblom (2003) shows that it can be applied to the development of crime preven-
tion strategies directed at organized crimes such as drug trafficking. Ekblom’s
approach is used in this chapter to apply results of social network analysis to the
development of crime prevention strategies.
146 D.A. Bright

Organized crime has two distinct features that have the potential to frustrate
attempts to design effective crime prevention strategies (Cornish & Clarke, 2002;
Levi & Maguire, 2004; Hancock & Laycock, 2010). Firstly, organized crime scripts
tend to be complex and can involve the sharing of necessary resources and equip-
ment, a variety of locations, extended timeframes, a number of different actors, and
a range of activities (i.e., a complex crime script). The complexity inherent in orga-
nized crime such as drug trafficking renders difficult any attempt to generate crime
prevention strategies, especially those relating to specific situations or circum-
stances. Secondly, organized crime is characterized by the existence of an organized
group of offenders who link together to share resources, skills, and experience that
facilitate the enactment of the crime script. The focus on criminal networks is rec-
ognition that in organized crime, individuals are usually part of an extended net-
work that is involved in crime commission across the supply chain. In the case of
drug trafficking this may involve gathering necessary supplies, bringing together
actors with requisite skills, contracting specialists, and seeking customers. Organized
crime can then be conceptualized as involving two levels for the purposes of the
design of crime prevention strategies: (1) the crime commission process; (2) crimi-
nal networks.
In this two-level conceptualization of organized crime, the primary criminal act
is comprised of a sequence of scenes or niches that are connected across time and
place, and are dependent on the existence of a loosely connected network of actors
who bring together and exchange the required skills, knowledge, resources, and
the capacity to exploit crime opportunities as they arise (see Morselli, 2010;
Schwartz & Rouselle, 2009). Due to the complexity inherent in organized crime
such as drug trafficking, there is usually a division of labor, and the crime commis-
sion process relies on the exchange of skills and resources. From one crime scene
to the next the situations and targets (i.e., skills, resources, actors) change.
Therefore, different subgroups of actors are likely to be involved in the different
crime scenes. This has important implications for the development of prevention
strategies. Prevention strategies targeted at vulnerabilities in crime scripts/scenes
may prevent offending by increasing risks significantly so that offenders are no
longer willing to assume requisite risks, or drive changes in offender behavior as
they attempt to circumvent such risks. Prevention strategies seek to undermine the
successful implementation of activities in crime scenes by reducing opportunities
and/or increasing risks.
Another challenge for crime prevention against organized crime is that for
crime prevention strategies to be determined, usually the “target” must be identi-
fied. However, in the case of organized crime, the target is more difficult to iden-
tify. CSA and SNA can be used to overcome this challenge. Through the use of
CSA and SNA, the key questions regarding target specification become: (1) What
are the skills, resources, and knowledge that are required to successfully imple-
ment the crime scene?; (2) Who are the actors involved and what are the charac-
teristics of the connections or relationships between the actors?; and (3) What are
the characteristics of the network as a whole and how does this relate to the crime
commission process?
11 Using Social Network Analysis to Design Crime Prevention Strategies… 147

The associational/networked structure of organized crime facilitates the


intertwining of legitimate business and criminal actors (Papachristos & Smith,
2013). The exchanges are dependent on the types of social structures and links
within the network. Criminal networks expand their size and reach as new actors
join the network. New actors may join a network because they possess specialized
skills or can facilitate access to necessary resources (see for example, Bright &
Delaney, 2013). New criminal network ties may occur following chance meet-
ings, build on existing relationships (e.g., familial ties), or occur via active recruit-
ment of actors with desired skills and resources. “Offender convergence settings”
(Felson, 2006) are locations at which offenders in a criminal network can find
new recruits and accomplices (e.g., prison).
The process of offending can therefore be described both horizontally (across
scenes) and laterally (the criminal networks). Therefore, crime prevention strategies
targeting organized crime should include strategies across the two levels: (1) interven-
tions to impede or restrict the crime commission process (including resources, loca-
tions, activities); and (2) interventions designed to limit the formation and growth of
criminal networks. Given the co-location of crime scripts and criminal networks, the
use of SNA and CSA as a combined analytic strategy may be particularly well suited
to guiding crime prevention strategies across the two broad target areas specified
above. Indeed, some previous research has combined SNA and CSA to study drug
trafficking networks (e.g., Morselli & Roy, 2008; Bright & Delaney, 2013). Prevention
strategies built by examining CSA and SNA would seek to target vulnerabilities across
the crime script sequence by impacting on the crime commission process and on the
network of offenders supporting the crime commission process.
In the next section, I present a case study of a methamphetamine manufacture and
trafficking network that operated in Australia in the 1990s. The structure and operation
of the network is analyzed using SNA and CSA with the aim of developing crime pre-
vention strategies. Results will show how SNA can be used to identify crime preven-
tion strategies aimed at preventing the manufacture and trafficking of methamphetamine,
and the growth of large, profit-motivated methamphetamine trafficking operations.

Method

The first task was to collect relational and actor data on the criminal network1. The
criminal network was engaged in the manufacture and trafficking of the synthetic drug,
methamphetamine. Methamphetamine is produced by transforming precursor chemi-
cals (e.g., pseudoephedrine) into methamphetamine via any one of a variety of chemi-
cal processes. Data was extracted from files of the Office of the Director of Public
Prosecutions (ODPP) in NSW, Australia. File information included details of a police
investigation and subsequent prosecution across an 8-year period. Files included

1
Data was used in a previous study on the evolution of criminal networks; see Bright & Delaney,
2013).
148 D.A. Bright

transcripts of telephone intercepts and listening devices, internal communications such


as summaries and memos, documents indicating charges laid, summaries of facts,
criminal history records, police and witness statements, informant statements, physical
surveillance reports, trial transcripts, and case synopses.
As the initial aim of this research was to investigate the evolution of the network
across time, the 8-year period was divided into four 2-year time periods (Time 1 to
Time 4; see Bright & Delaney, 2013). For the purpose of the current chapter, I report
only the early phase (time period 1, 1991–1992) and late phase (time period 4,
1997–1998). Two time periods are sufficient in order to describe the network’s
structure and to document the evolution of the network across time.
Each actor was allocated an alphanumeric identifier in which the prefix indicated
the period in which they first entered the network. Binary relational data was
extracted to develop an adjacency matrix in which a “1” indicated the presence of a
relationship and a “0” indicated the absence of a relationship.

Crime Scene Analysis

Previous research (Ritter, Bright, & Gong, 2012; Bright & Delaney, 2013; Chiu
et al., 2011) has studied the crime scenes or supply chain steps required to manufac-
ture and traffic methamphetamine in Australia. For the purposes of this chapter I use
the four steps identified by Bright and Delaney (2013) as this is the most parsimoni-
ous version and simplifies the analyses. Future research could attempt to replicate
the results documented here using the crime script analysis of Chiu et al. (2011) in
which seven steps are used. The four crime scenes used in this chapter are:
1. Sourcing precursor chemicals and laboratory equipment: Precursor chemicals
can be obtained in a number of ways including importation, diversion from legit-
imate chemical industries, and the bulk purchase of “cold and flu” medications
that contain pseudoephedrine. Laboratory equipment can be purchased from
chemical/laboratory equipment retailers, or regular kitchen equipment can be
used as a substandard alternative.
2. Manufacture of methamphetamine in clandestine laboratories: Several different
chemical processes or “recipes” can be employed to convert precursor chemicals
into methamphetamine. These conversion processes vary in complexity and the
requisite knowledge and skills required.
3. Wholesale level distribution of methamphetamine: Following manufacture,
methamphetamine is packaged into kilogram and multiple kilogram amounts
and sold to wholesale distributors. The “pure” methamphetamine produced in
the laboratory is usually diluted or “cut” with impurities prior to being sold at
wholesale level. Wholesale dealers then “cut” the drug further, divide the larger
amounts into smaller amounts, and sell to retail level dealers.
4. Retail level distribution of methamphetamine: Retail level dealers divide the
drug into single dosage units, usually following further “cutting,” and then sell
to illicit drug users.
11 Using Social Network Analysis to Design Crime Prevention Strategies… 149

File information was also used to allocate each actor to a role within the network.
Roles included “precursor providers” who sourced the necessary chemical precur-
sors, and “wholesale dealers” who sold methamphetamine in kilogram and multiple
kilogram quantities. When actors appeared to have more than one role, the primary
role played by the actor was determined.
Actors were also allocated to the crime scene in which they participated. For
example, some workers were provided assistance at manufacture sites, whereas oth-
ers provided support in the wholesale dealing scene.

Social Network Analysis

SNA metrics were calculated using the Visone software package (Brandes &
Wagner, 2004). Standardized degree centrality and betweenness centrality scores
were calculated for each actor in both the “early phase” and “late phase” of the
network. Degree centrality indicates the proportion of the other actors in the net-
work to whom an actor is connected. Betweenness centrality is a measure of the
extent to which an actor is positioned on the shortest paths between pairs of other
actors in the network and is usually considered a measure of brokerage potential.

Results

Figures 11.1 and 11.2 display network maps of the “early phase” and “late phase”
of the network respectively. Viewing the two graphs simultaneously reveals that the
network expanded in size between the early phase and the late phase. The network
originated with M1.13 who was “cooking” methamphetamine for his own use. He
also supplied the drug to his group of friends and associates. Expansion of the net-
work occurred when M1.13 was introduced to M1.5 who was able to supply large
amounts of precursor chemicals. M1.5 introduced M1.13 to M1.16 who also had
access to large amounts of precursor chemicals. With this increased supply, M1.13
could manufacture larger amounts of methamphetamine.
During the early phase of the operation, M1.13 took a proportion of the drug for
his own use. Associates and friends of M1.13 contributed to the network by provid-
ing their homes as sites for manufacture, and provided assistance at laboratory sites
(e.g., conveying equipment, checking on the site). In return, they received drugs and
money.
M2.3 joined the network in around 1993 and he established ties with metham-
phetamine cooks, workers, and dealers. M1.16, M1.5, and M2.3 were provided with
large amounts of precursor chemicals by M2.5 who was the owner of a chemical
supply business. By the late phase, the network had evolved into a large metham-
phetamine manufacture business, with several clandestine laboratory sites. It
included wholesale and retail niches involving many actors.
150 D.A. Bright

Fig. 11.1 Network map: Early phase of the criminal network

Fig. 11.2 Network map: Late phase of the criminal network


11 Using Social Network Analysis to Design Crime Prevention Strategies… 151

Results of the SNA show that the network became less dense and more centralized
over time. In the “early phase” the network comprised 20 actors and 29 links with a
density of 15.26 % and an average degree of 2.90. By the “late phase,” the network
had expanded to 57 actors and 144 ties with a density of 8.71 % and an average
degree of 4.97. Centralization in the early phase of the network was 41.52, while in
the late phase network it was 67.30. It should be noted that centralization and den-
sity are influenced by the number of actors in the network, which renders these
metrics difficult to interpret when comparing networks of different size. Average
degree is a more reliable measure. While the network appeared to become less
dense and interconnected across time, some actors became clear hubs (high degree
centrality) in the network.
Visual inspection of the network map of the early phase of the network reveals
one large chain-like component and two disconnected dyads. M1.13 (the metham-
phetamine cook) had the highest degree centrality score and was directly connected
with more than half of the network (degree centrality = 0.53). M1.18 and M1.19
were next highest on degree centrality, connected to about one third of other actors
in the network (degree centrality = 0.32). M1.13 also had the highest score on
betweenness (0.36) suggesting he was an important broker in the flow of informa-
tion and resources across the network. In the late phase, M2.3 and M1.16 were hubs
(high centrality), connected to 74 % and 51 % of network actors respectively. M2.3
was also high on betweenness centrality (0.66), suggesting he was an important
broker in the late phase of the network.
In the early phase of the network, the network map shows that the methamphet-
amine cook (M1.13) was a well-connected actor and was also a broker in the net-
work. A residence provider (M1.18) and a laborer were both high on betweenness
centrality (brokerage). The most strategically positioned actor (high brokerage, low
degree) was M1.3—an actor whose role is unknown.2
In the late phase of the network, two wholesale dealers (M1.16 and M2.3) were
high on degree and betweenness. One precursor provider was high on both between-
ness and degree centrality. Other high degree/betweenness actors in the late phase
network include laborers (M4.31) and retail dealers (M2.4, M4.32). A strategically
positioned actor was M1.1 a worker in the manufacture scene who appeared to be a
broker across key components of the crime script (e.g., between precursor providers
and cooks and between cooks and wholesale level dealers).
In criminal networks, being high on degree centrality may be a vulnerability
because well-connected actors are more visible. On the other hand, brokerage is
arguably advantageous and does not necessarily indicate high visibility.
Strategically positioned actors are therefore those who are high on betweenness
centrality yet low on degree centrality (see Morselli, 2010; Calderoni, 2014).
Figures 11.3 and 11.4 show that strategically positioned actors (low degree and
high brokerage; see Morselli, 2010) were M1.3 in the early phase and M1.1 in the

2
The “unknown” status may be a result of his high brokerage and low degree. Law enforcement
may not have identified this actor as important either because he was not well connected and/or
because he appeared to play only a minor role.
152 D.A. Bright

Fig. 11.3 Strategic positioning: Early phase of the criminal network

late phase of the network. The most visible actors were M1.13 in the early phase
network and M2.3 and M1.16 in the late phase network.
The combined crime script analysis and SNA results are displayed graphically in
Figures 11.5 and 11.6. Three crime scenes were operational in the early phase of the
network (precursor attainment, manufacture, wholesale dealing) while four crime
scenes were present in the late phase (precursor attainment, manufacture, wholesale
dealing, retail dealing). There was no retail dealing in the early phase of the net-
work, whereas by the late phase there was a retail-dealing niche comprising 21 retail
level dealers. This appears to reflect a marked shift in the activities of the network:
from manufacture for personal use and supply to friends, to a focus on profit and
increased manufacture and supply.

Crime Scenes and Roles

Examining roles and crime scenes in more detail reveals some changes between the
early phase network and the late phase network. In the early phase of the network,
one actor was supplying precursor chemicals, whereas by the late phase, four actors
were supplying chemical precursors. Two of the four appeared to be operating
11 Using Social Network Analysis to Design Crime Prevention Strategies… 153

Fig. 11.4 Strategic positioning: Late phase of the criminal network

somewhat independently of each other, while the other two were connected to each
other. In the early phase of the network, the manufacture of methamphetamine was
conducted by only one cook (M1.13), while in the late phase of the network there
were four methamphetamine cooks, with each one operating at different lab sites.
This suggests that there was an expansion of manufacture operations between the
early and late phases of the network. Multiple independent laboratory sites increase
overall security; even if one site is detected by law enforcement, other sites remain
to continue manufacture. This increases the need for more individuals with special-
ized skills in cooking methamphetamine as the lab sites operated independently of
each other.
In the early phase, five individuals provided premises and three actors supported
the operation with their labor, while in the late stage there was only one premises
provider (connected only to a precursor chemical supplier) and three individuals
contributing their labor. In the early phase, only one actor was involved in wholesale
distribution (M1.15) while in the late phase 13 actors were involved in wholesale
distribution, 8 as wholesale dealers who were supported by 5 actors who completed
associated tasks such as transporting drugs.
In the late phase of the network, two roles emerge that were not present in the early
phase: corrupt officials and standover men. These new roles may have emerged in
response to increased security concerns in the newly developed profit-driven context.
154 D.A. Bright

Fig. 11.5 Network map with crime scripts and roles showing betweenness centrality: Early phase
of the criminal network

Standover men were connected only to retail level and wholesale level dealers, sug-
gesting that the necessary security or enforcement activities related primarily to the
sale of drugs rather than to other aspects of the operation such as precursor attainment
and manufacture. Corrupt police officers were connected only to a few actors: one
was connected to two of the methamphetamine cooks and precursor suppliers, while
the other two were connected to the most central actor, a wholesale dealer. The corrupt
police officers assisted network actors by tampering with evidence and providing net-
work actors with information regarding the police investigation.
The SNA, complemented by CSA, has aided in a comprehensive description of
the crime commission process and the criminal network per Ekblom (2003). I turn
now to consider the next step in Ekblom’s approach: the design of a range of crime
prevention strategies.
11 Using Social Network Analysis to Design Crime Prevention Strategies… 155

Fig. 11.6 Network map with crime scripts and roles showing betweenness centrality: Late phase
of the criminal network

Development of Prevention Strategies

Prevention strategies can be devised by building on the results of the analyses. Crime
prevention strategies are usually designed to increase effort, increase risk, decrease
rewards, remove excuses, and reduce provocations (Cornish & Clarke, 2003).
Drawing on existing frameworks for crime prevention in organized crime (e.g.,
Ekblom, 2003; Hancock & Laycock, 2010), I categorize prevention strategies into
two primary categories: (1) crime prevention strategies targeted at the crime com-
mission process; and (2) crime prevention strategies targeted at the establishment
156 D.A. Bright

and expansion of criminal networks. While the focus of the strategies is on the
manufacture and trafficking of methamphetamine, many of the prevention strategies
may be broadly applicable to trafficking of other drugs (especially synthetic drugs),
and to other types of organized crime activity.

Crime Prevention Strategies Targeting the Crime Commission


Process

There are four main crime scenes or niches comprising the crime commission pro-
cess for the manufacture and trafficking of methamphetamine (Bright & Delaney,
2013; Chiu et al., 2011; Ritter et al., 2012): (1) the acquisition of precursor chemi-
cals and laboratory equipment, (2) the manufacture of methamphetamine, (3)
wholesale distribution, and (4) sale of end product at retail level. As detailed above,
within each scene actors play specific roles in order to enact the activities required.
Crime prevention efforts are likely to be more effective when the interventions
occur early in the crime commission sequence: in the case of methamphetamine
manufacture and trafficking, the most effective interventions are likely to be those
against the crime scripts involving obtaining precursors and equipment and manu-
facture of methamphetamine. Therefore, the crime prevention strategies outlined
below focus on these first two scripts in the crime commission sequence. Prevention
strategies focus on the activities undertaken in each scene and the roles played by
actors.

Obtaining Chemical Precursors and Equipment

Legislative regulation of precursor chemicals and laboratory equipment can be used


to increase the efforts and risks associated with attempts to obtain chemicals and
equipment for illicit purposes. These types of precursor controls have been shown
to be effective, at least in the short term (e.g., McKetin, Sutherland, Bright, &
Norberg, 2011). For example, purchasers may be required to provide photographic
identification at point-of-sale, and real-time purchase tracking databases can moni-
tor purchase patterns (e.g., Project STOP in Australia; see Ransley et al., 2012). In
the criminal network case study presented above, only a small number of individu-
als were involved in the acquisition of precursors, so restricting access may signifi-
cantly constrain the network’s capacity to gain access to these required resources.
Restrictions on access to quality laboratory equipment increase the efforts
required to obtain equipment and may cause network actors to use substandard
equipment that produce lower quality end product and/or reduced yields or that
increase the risks of explosions and chemical burns (e.g., Sexton, Carlson, Leukefeld,
& Booth, 2006).
The establishment of cooperative partnerships between potentially vulnerable
companies (e.g., chemical wholesalers, pharmacies) and law enforcement can also
11 Using Social Network Analysis to Design Crime Prevention Strategies… 157

increase the risk of attempting to obtain precursor chemicals and equipment for
illicit purposes. Such partnerships should include the implementation of systems
that facilitate real-time information sharing between companies and law enforce-
ment agencies.
Rewards may also be diminished by specifically targeting chemicals that pro-
duce large and high quality yields, and those that rely on lower risk, relatively easy
cooking methods (see Ritter et al., 2012). Faced with increased effort and risk,
actors may decide to cease production or to use more risky, difficult or untested
cooking methods.

Manufacture

Prevention strategies targeting manufacture can increase effort by restricting access


to the necessary instructions or “recipes” for producing methamphetamine. For
example, regulation and monitoring can be used to shut down Internet web sites that
promulgate such recipes. Prevention strategies can also increase the risks of manu-
facture by enhancing natural surveillance by place managers. For example, training
could be provided to real estate agents and Parks and Wildlife officers to identify
warning signs of methamphetamine production sites. Community educational pro-
grams (e.g., advertisements, mail drops) can also educate community members to
identify signs of methamphetamine laboratories that are operating in residential
premises. Increased vigilance of residents can increase the risk of detection and
require increased efforts to disguise production.
Reducing rewards of manufacture can be facilitated by disseminating fake reci-
pes (Chiu et al., 2011) that produce either an inert substance or significantly lower
yields of end product. Rewards can also be reduced by early detection of clan labs
before the full manufacture process has been completed. As precursor chemicals
and equipment are expensive, the loss of a full production cycle can represent a
large loss to criminal network actors (Gong, Ritter, Bright, & Doran, 2012).

Roles

A number of important roles were identified within each script across the crime
commission process. For example, in the methamphetamine manufacture and
trafficking process, precursor providers and residential premises providers were
important in providing the resources required for the manufacture process.
Prevention strategies should be implemented that restrict the ability of network
actors or other individuals to assume these roles. This can be achieved by increas-
ing the effort and risk required to obtain the necessary resources. As access to
such resources becomes more restricted, it is likely that fewer individuals will be
able to fill the roles. As there are fewer individuals in the “criminal pool” with
access to such resources, it becomes increasingly difficult for network actors to
locate actors who can fill these roles.
158 D.A. Bright

Corrupt officials play an important role in providing information to network


actors on law enforcement strategy and intervention and/or undermining judicial
procedures. In the methamphetamine manufacture and trafficking network described
above, a corrupt police officer arranged for the fingerprints of one of the network
actors to be removed from an item in evidence. The implementation of anti-
corruption measures in areas vulnerable to potential corruption by organized crime
groups (e.g., police, law firms) can increase the effort and increase risk for criminal
actors to locate officials who are willing to assume such a role (e.g., Huther & Shah,
2000). Education, monitoring, and auditing of law enforcement officers and others
who work in vulnerable employment contexts, which may be vulnerable to corrup-
tion, may also help to prevent corruption and increase the risks posed by detection
(Huther & Shah, 2000; Matei & Matei, 2009). Furthermore, increasing both formal
and informal sanctions and punishments imposed on public officials found guilty of
corruption that is related to organized crime activities can inflate the perceived risks
of engaging in corruption and make efforts to collaborate with organized crime
groups more difficult.
It may also be possible to reduce the quantum of financial rewards flowing to
criminal network actors from certain roles. For example, increased monitoring to
prevent tampering of evidence means that public officials who have been corrupted
by organized crime groups are unable to provide anticipated rewards. Financial
rewards and profit can be undermined by focusing attention on those roles related to
financial transactions and profit. For example, financial transactions should be mon-
itored to detect suspicious transactions in order to increase risks of detection for
those involved in illicit financial transactions.

Crime Prevention Strategies Targeting the Criminal Network

From the results of the SNA and CSA, six areas were identified which may provide
intervention points for crime prevention strategies: (1) Strategic recruitment of
actors with the necessary skills, resources, and experience is necessary to ensure the
completion of the crime commission process; (2) Associations and contacts are
critical for the expansion and adaptability of the network across time; (3) Connections
between legitimate and illegitimate actors are also important to the crime commis-
sion process (e.g., legitimate businesses that sell chemicals may be at risk of coop-
eration or infiltration by organized crime groups; (4) Brokers are key to the operation
of the network and to the flow of information and resources through the network; (5)
Some actors are positioned strategically in the network with low degree centrality
and high betweenness centrality. These actors occupy key brokerage roles and
remain relatively less visible as they have fewer direct connections with other actors;
(6) The network structure as a whole was characterized by loose connections, low
density, and flexibility across time (e.g., fluctuations in centralization). These net-
work level features facilitate the crime commission process and enhance security,
flexibility, and adaptation across time. Each of these six intervention points are
11 Using Social Network Analysis to Design Crime Prevention Strategies… 159

discussed below with recommendations for crime prevention strategies against the
criminal network.

Recruitment of Actors with Specific Skills, Resources, Knowledge

Restrictions or regulations on chemicals and equipment inhibit the capacity of indi-


viduals to fill particular roles (e.g., precursor supplier) and can increase the effort
required to recruit actors with the required set of skills, knowledge, and/or resources.
For example, such interventions may increase search time that is required to recruit
actors with the required skills, knowledge, and resources. Network actors may need
to take increased risks in order to identify and recruit individuals who have the
required characteristics (e.g., by looking beyond trusted associates) (for example,
see Gambetta, 2009). Prevention strategies should also aim to increase the risks
involved in the recruitment process such as increasing the penalties for those who
provide important skills and resources for the purpose of facilitating illicit activities
(e.g., supply of precursors, manufacture) rather than a traditional focus on those
who traffic end product (i.e., retail level dealers).
Profits can be restricted by limiting access to those who provide important skills
and resources. Network actors may therefore need to connect with actors with sub-
optimal skills and resources. Actors who can supply only suboptimal skills (e.g., in
methamphetamine manufacture), knowledge (e.g., of law enforcement strategies),
and resources (e.g., precursor chemicals) are likely to generate lower yields of end
product and consequently reduced profits. Faced with increased effort and risk and
decreased profit, actors may decide to cease manufacture and trafficking activities.

Associations

Associations or links between actors are critical in preserving network structure.


The growth and expansion of networks relies on the formation of new connections
between individuals over time as illustrated by the case study described above.
“Offender convergence settings” have been identified as important to the process of
the growth of criminal networks (Felson, 2006). Increasing the effort required to
generate new ties may impact the capacity of the network to grow and adapt. One
potential prevention strategy is to limit access to offender convergence settings in
order to inhibit the capacity to generate new links. For example, prisons may offer
fertile ground to locate people who are willing to engage in criminal enterprise (e.g.,
Cullen, Jonson, & Nagin, 2011). Interventions within prisons such as gang interven-
tions or internal practices that limit contact between potential associates may limit
these opportunities. The risks inherent in making new associations might also be
increased through the use of non-association or anti-consorting legislation and
increased sanctions for associating with known or potential criminal associates
(e.g., (Ayling, 2011). Nonetheless, this type of legislation is controversial (see
Ayling & Broadhurst, 2012).
160 D.A. Bright

Criminal Cooperation Between Legitimate and Illegitimate Actors

The SNA identified connections between legitimate businesses and public officials
on the one hand and criminal actors on the other. These connections facilitated
access to an industrial chemist who provided precursor chemicals to the group and
a police officer who interfered with evidence and provided intelligence to network
actors regarding law enforcement strategy and interventions. Prevention strategies
can be employed to limit opportunities for connections between legitimate business
and criminal associates. For example, the effort required to create such partnerships
can be increased through regulation of legitimate businesses that operate in poten-
tially vulnerable industries (e.g., chemical companies, laboratory equipment supply
companies). As mentioned previously, purchasers or precursor chemicals can be
required identification at point-of-sale and databases can be maintained to record all
purchases. Risks can also be increased through the regulation and auditing of
employees, strengthening formal surveillance and informal surveillance of legiti-
mate businesses that are identified as being vulnerable, and education campaigns
targeting employees (e.g., on issues such as integrity, ethics, potential risks).

Brokers

Brokers are those actors who are high on betweenness centrality and play an impor-
tant role in the flow of information and resources through a criminal network (2009b).
The results of analyses suggest that targeting crime prevention strategies at brokers
and potential brokers may be effective in undermining the structure and functionality
of criminal networks. Limitations on access to key resources restrict opportunities for
potential brokers to exploit structural holes (Burt, 1992) between those who wish to
exchange such resources. It may also be possible to identify potential or emerging
brokers based on network position or personal characteristics (Burt, Jannotta, &
Mahoney, 1998; Morselli, 2005). Once emerging brokers are identified, the risks of
detection can be inflated by placing them under increased scrutiny and surveillance.
Fewer opportunities for brokerage can translate into lower rewards both for the
individual brokers and for other network actors who rely on brokers. False informa-
tion can be disseminated to undermine the value of brokers to the criminal network.
Individuals who transmit false or misleading information or who facilitate the
exchange of poor quality resources are of less value to the network and may fail to
successfully occupy brokerage position (Morselli, 2005).

Strategic Position

Actors who are strategically positioned in criminal networks are those who have
relatively high betweenness centrality scores and relatively low degree centrality
scores (Morselli, 2010). They are less visible to law enforcement agencies by virtue
of the low degree but are able to facilitate the transmission of information and
11 Using Social Network Analysis to Design Crime Prevention Strategies… 161

resources due to their high betweenness scores. Prevention strategies could focus on
making it more difficult for actors to achieve such a strategic position in criminal
networks, perhaps by driving them into direct connections with others. Actors who
have many direct connections with other actors are high in degree centrality and are
more visible and vulnerable (increased risks of detection). For example interven-
tions that limit the capacity of actors to remain anonymous or increase the need for
physical meetings between individuals could reduce the number of indirect connec-
tions, and increase direct connections across the network. Such an intervention
would limit the potential for actors to occupy strategic positions within the network.
This could be accomplished through the use of telephone/email surveillance and by
regulation of mobile and other telecommunications devices to require identification
details and impair the capacity of actors to remain anonymous. Counterintuitively,
limits on actors ability to occupy strategic positions could be achieved by facilitat-
ing trust within the network. If actors believe they can trust those around them, they
may be more willing to connect directly rather than through intermediaries.
Increased penalties could be instated for individuals who facilitate or broker
access to resources and equipment required for manufacture. Law enforcement
focus on financial transactions, increased monitoring of suspicious transactions, and
the confiscation of money/profits from organized crime groups can lead to decreased
profit (reward). If actors are concerned that their profits will be confiscated, poten-
tial rewards fall significantly and this may be sufficient to convince some actors to
cease their participation in the criminal enterprise.

Network Structure

Highly centralized networks are those in which there are a proportionately few hubs
(actors with high degree centrality) who are characterized by very many connections to
other actors. A highly centralized network—one with centers of gravity formed around
hubs with clusters of actors surrounding the hub—may be more vulnerable to law
enforcement interventions such as surveillance and arrests. Highly centralized criminal
networks can be dismantled relatively quickly by removing the central actors (hubs) and
then following links to all other actors (e.g., Bright et al., 2014; Xu & Chen, 2009).
Interventions that promote network centralization will increase the risks inherent in
being part of a criminal network and therefore increase the risks of detection and pros-
ecution (see Baker & Faulkner, 1993). This could be achieved by policies that under-
mine trust so that small, dense clusters are formed with only a few trusted associates.

Limitations of the Data and Analyses

There are a number of limitations of the data and the analysis that should be borne
in mind. First, the data set may contain errors or missing data that have the poten-
tial to bias the results. Furthermore, centrality scores can be artificially influenced
162 D.A. Bright

by the focus of law enforcement efforts. High centrality scores may reflect the
amount of information collected on some actors rather than the actors’ connectivity.
Secondly, the effectiveness of the suggested strategies is largely unknown, although
some have already been evaluated and demonstrated to be efficacious (e.g., precur-
sor controls). Thirdly, as with all crime prevention strategies, there may be some
displacement to other crime. In the case of criminal networks involved in drug
trafficking, prevention strategies that drive groups away from drug manufacture/
trafficking may simply lead to a shift in focus to other illicit activity, e.g., gun traf-
ficking. Finally, the generalizability of the findings and the crime prevention strate-
gies to other criminal networks and other types of organized crime is unknown and
requires further research.

Conclusions

This chapter has demonstrated the utility of SNA and CSA to facilitate the design of
crime prevention strategies against criminal networks. A framework was used to
facilitate the design of crime prevention strategies across two levels in organized
crime groups: (1) crime prevention process; (2) criminal networks. The strategies
aim to undermine the crime commission process and to inhibit the formation and
growth of criminal networks.

References

Ayling, J. (2011). Pre-emptive strike: How Australia is tackling outlaw motorcycle gangs.
American Journal of Criminal Justice, 36(3), 250–264.
Ayling, J., & Broadhurst, R. (2012). The suppression of organised crime: New approaches and
problems (Best practice in policing and security, pp. 37–55). London: Palgrave Macmillan.
Baker, W. E., & Faulkner, R. R. (1993). The Social Organization of Conspiracy: Illegal Networks
in the Heavy Electrical Equipment Industry. American Sociological Review, 58(6), 837–860.
Brandes, U., & Wagner, D. (2004). Visone—Analysis and visualisation of social networks. In
M. Junger & P. Mutzel (Eds.), Graph drawing software (pp. 321–340). Berlin: Springer-Verlag.
Bright, D. A., & Delaney, J. J. (2013). Evolution of a drug trafficking network: Mapping changes
in network structure and function across time. Global Crime.
Bright, D. A., Greenhill, C., & Levenkova, N. (2014). Dismantling criminal networks: Can node
attributes play a role. In C. Morselli (Ed.), Crime and networks. New York, NY: Routledge.
Bright, D. A., Hughes, C. E., & Chalmers, J. (2012). Illuminating dark networks: A social network
analysis of an Australian drug trafficking syndicate. Crime, Law, and Social Change.
Burt, R. S. (1992). Structural holes. Cambridge, MA: Harvard University Press.
Burt, R. S., Jannotta, J. E., & Mahoney, J. T. (1998). Personality correlates of structural holes.
Social Networks, 20(1), 63–87. doi:10.1016/S0378-8733(97)00005-1. http://dx.doi.org.
Calderoni, F. (2012). The structure of drug trafficking mafias: The ‘Ndrangheta and cocaine.
Crime, Law and Social Change, 58(3), 321–349.
Calderoni, F. (2014). Identifying Mafia Bosses from Meeting Attendance Networks and Network
Analysis for Defence and Security (pp. 27–48): Springer.
11 Using Social Network Analysis to Design Crime Prevention Strategies… 163

Carley, K. M., Dombroski, D., Tsvetovat, M., Reminga, J., & Kamneva, N. (2003). Destabilizing
dynamic covert networks. Paper presented at the 8th International Command and Control
Research and Technology Symposium., National Defence War College, Washington, DC.
Chiu, Y., Leclerc, B., & Townsley, M. (2011). Crime script analysis of drug manufacturing in
clandestine laboratories. British Journal of Criminology, 51, 355–374.
Cornish, D. B. (1994). The procedural analysis of offending and its relevance for situational pre-
vention. In R. V. Clarke (Ed.), Crime prevention studies (pp. 151–196). Monsey, NY: Criminal
Justice Press.
Cornish, D. B., & Clarke, R. V. (2002). Analyzing organised crimes. In A. R. Piquero & S. G.
Tibbetts (Eds.), Rational choice and criminal behavior: Recent research and future challenges
(pp. 41–62). New York, NY: Routledge.
Cornish, D. B., & Clarke, R. V. (2003). Opportunities, precipitators and criminal decisions: A reply
to Wortley’s critique of situational crime prevention. In M. J. Smith & D. B. Cornish (Eds.),
Theory for practice in situational crime prevention (Crime Prevention Studies no. 16,
pp. 41–96). Monsey, NY: Criminal Justice Press.
Cullen, F. T., Jonson, C. L., & Nagin, D. S. (2011). Prisons do not reduce recidivism the high cost
of ignoring science. The Prison Journal, 91(3 suppl), 48S–65S.
Ekblom, P. (2003). Organised crime and the conjunction of criminal opportunity framework. In
A. Edwards & P. Gill (Eds.), Transnational organised crime (pp. 28–41). London: Routledge.
Everton, S. F. (2012). Disrupting dark networks. New York, NY: Cambridge University Press.
Felson, M. (2006). The Ecosystem for Organized Crime. HEUNI Paper No 26. HEUNI. Helsinki.
Gambetta, D. (2009). Codes of the underworld: How criminals communicate. Princeton: Princeton
University Press.
Gong, W., Ritter, A., Bright, D., & Doran, C. (2012). How profitable is methamphetamine dealing
in Australia? Drug and alcohol dependence, 122(3), 208–212.
Hancock, G., & Laycock, G. (2010). Organised crime and crime scripts: Prospects for disruption.
In K. Bullock, R. V. Clarke, & N. Tilley (Eds.), Situational Prevention of Organised Crime
(pp. 172–192). Cullompton, Devon: Willan Publishing.
Huther, J., & Shah, A. (2000). Anti-corruption policies and programs: A framework for evaluation
(Vol. 2501): World Bank Publications, Washington, DC.
Krebs, V. E. (2002). Mapping networks of terrorist cells. Connections, 24(3), 43–52.
Levi, M., & Maguire, M. (2004). Reducing and preventing organised crime: An evidence-based
critique. Crime, Law and Social Change, 41, 397–469.
Mancuso, M. (2013). Not all madams have a central role: Analysis of a Nigerian sex trafficking
network. Trends in Organized Crime.
Matei, L., & Matei, A. (2009). Corruption in the public organizations. Towards a model of cost-
benefit analysis for the anticorruption strategies. Transylvanian Review of Administrative
Sciences, 27, 145–171.
Mazerolle, L., Soole, D. W., & Rombouts, S. (2006). Street-level drug law enforcement: A meta-
analytical review⋆. Journal of Experimental Criminology, 2(4), 409–435.
McKetin, R., Sutherland, R., Bright, D. A., & Norberg, M. M. (2011). A systematic review of
methamphetamine precursor regulations. Addiction, 106(11), 1911–1924.
Morselli, C. (2005). Contacts, opportunities, and criminal enterprise. Toronto, ON: University of
Toronto Press.
Morselli, C. (2009a). Hells Angels in springtime. Trends in Organized Crime, 12, 145–158.
Morselli, C. (2009b). Inside criminal networks. New York, NY: Springer.
Morselli, C. (2010). Assessing vulnerable and strategic positions in a criminal network. Journal of
Contemporary Criminal Justice, 26, 382–392.
Morselli, C., & Roy, J. (2008). Brokerage qualifications in ringing operations. Criminology, 46(1),
28.
Natarajan, M. (2000). Understanding the structure of a drug trafficking organisation: A conversa-
tional analysis. In M. Natarajan & M. Hough (Eds.), Illegal Drug Markets: From Research to
Prevention Policy. Monsey, NY: Criminal Justice Press.
164 D.A. Bright

Papachristos, A. V., & Smith, C. M. (2013). The embedded and multiplex nature of Al Capone. In
C. Morselli (Ed.), Crime and networks. New York, NY: Routledge.
Ransley, J., Mazerolle, L., Manning, M., McGuffog, I., Drew, J. M., & Webster, J. (2012). Reducing
the methamphetamine problem in Australia: Evaluating innovative partnerships between
police, pharmacies and other third parties. Canberra, Australia: NDLERF Monographs.
Ritter, A., Bright, D. A., & Gong, W. (2012). Evaluating drug law enforcement interventions
directed towards methamphetamine in Australia. Canberra: National Drug Law Enforcement
Research Fund.
Schwartz, D. M., & Rouselle, D. A. (2009). Using social network analysis to target criminal net-
works. Trends in Organised Crime, 12, 188–207.
Sexton, R. L., Carlson, R. G., Leukefeld, C. G., & Booth, B. M. (2006). Patterns of illicit metham-
phetamine production (“cooking”) and associated risks in the rural south: An ethnographic
exploration. Journal of Drug Issues, 36(4), 853–876.
Sutton, A., Cherney, A., & White, R. (2013). Crime prevention: Principles, perspectives and prac-
tices. Cambridge: Cambridge University Press.
Williams, P. (2001). Transnational Criminal Networks. In D. F. R. John Arquilla (Ed.), Networks
and netwars: The future of terror, crime, and militancy. Santa Monica, CA: RAND.
Xu, J., & Chen, H. (2009). Untangling criminal networks: A Case Study Intelligence and Security
Informatics (pp. 958–958).
Chapter 12
Uncovering the Social Pecking Order in Gang
Violence

Jasmin Randle and Gisela Bichler

Introduction

Estimates suggest that the more than 33,000 gangs operating in the USA are respon-
sible for generating nearly 48 % of the violent crime, most of which occurs in major
cities and suburban areas (FBI, 2012). The situation in Los Angeles serves as an
illustration; gang-initiated violence accounted for 57 % of the city’s 515 homicides
in 2005 (Winton, 2005). While many street gangs are involved to some extent in the
trafficking of illicit goods and services (e.g., illegal drugs and prostitution), criminal
enterprise does not appear to be the sole impetus of violence. For instance, Houston
and Anglin (1995) show that less than 5 % of the inter-gang-related deaths in Los
Angeles, California, were related to drug trafficking. This finding raises questions
about the nature of gang conflict.
A new thesis about gang violence has begun to emerge from the evidence gener-
ated by recent studies. Drawing upon social network theory, proponents argue that
the behavior of gang members is constrained (and enabled) by the social network
within which the individual and their group are emeshed (e.g., Descormiers &
Morselli, 2011; McCuish, Bouchard, & Corrado, 2015; Papachristos, 2009;
Papachristos, Hureau, & Braga, 2013; Papachristos, Wildeman, & Roberto, 2015).
Gang members do not kill because they are poor, black, or young or live in a socially disad-
vantaged neighborhood. They kill because they live in a structured set of social relations in
which violence works its way through a series of connected individuals (Papachristos,
2009: 75).

J. Randle, M.A. • G. Bichler, Ph.D. (*)


Center for Criminal Justice Research, California State University,
San Bernardino, 5500 University Parkway, San Bernardino, CA 92407, USA
e-mail: gbichler@csusb.edu

© Springer International Publishing Switzerland 2017 165


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_12
166 J. Randle and G. Bichler

Social networks place “adversaries in positions where each must attempt to


defend, maintain, or repair their reputation” (Papachristos, 2009: 76). Building from
the extensive literature on violent gang member interactions (see for example
Decker & Curry, 2002; Klein & Maxson, 2006; Wilkinson & Fagan, 1996),
Papachristos (2009) argues that patterns of networked violence emerge from the
aggregation of individual-level disputes associated with institutionalized conflict
among competing groups (rivalries), retribution for perceived harms or threats (reci-
procity), and the need to avoid subjugation (local hierarchies). To this, McCuish
et al. (2015) add that the struggle for control of the group may trigger intra-group
violence that may eventually lead to the evolution of new factions. And, as these
factions jostle for position within the consortium, conflict can emerge between
groups that are otherwise thought to be in allegiance (Decker & Curry, 2002;
Descormiers & Morselli, 2011). In short, violence spreads like a contagion through
the network as individuals react to the behavior of others.
The structure of violent inter- and intra-gang conflict has been investigated
within a single gang (e.g., McCuish et al., 2015), within an identifiable neighbor-
hood (e.g., Tita & Radil, 2011), and across entire cities, i.e., Boston (Papachristos
et al., 2013), Chicago (Papachristos, 2009; Papachristos et al., 2013), Montreal
(Descormiers & Morselli, 2011), Newark (McGloin, 2005), and St. Louis (Decker
& Curry, 2002). However, to the best of our knowledge, the structure of street gang
violence has not been investigated at a regional level. The present study extends this
line of inqury by investigating the social structure of inter- and intra-gang violence
among 158 gang subsets operating in Los Angeles County, California, as revealed
by prosecutions over a 9-year study period (January 1, 2002—December 31, 2010).
Our primary aim is to expose the regional structure of violence among historic
rivals—the Bloods and the Crips. By doing so, this study provides direction for
county-level anti-violence suppression efforts.
This chapter begins with a brief review of the relevant literature: this section
provides a synopsis of the research examining city-level investigations of inter- and
intra-group violence. We then describe how the violence network was mapped from
court cases, before presenting the results. The chapter concludes with a brief discus-
sion of the implications that our findings have for regional anti-gang suppression
efforts.

Gang Rivalries

Bloods and Crips

One of the most well-known, ongoing gang rivalries is thought to exist between two
predominantly African-American street gangs—the Bloods and the Crips.
Explanations for why the groups formed vary. It is suggested that troubled urban
youth bonded in their quest for “fun,” whereas others argue that young
12 Uncovering the Social Pecking Order in Gang Violence 167

African-Americans disenfranchised and excluded from what were known as white


organizations, such as the boy scouts, gathered to generate “family ties.” Irrespective
of their origins, these gangs and their intense rivalry began in Los Angeles in the late
1960s and early 1970s. Since then, the gangs have branched out, generating a con-
sortium of sorts, with subsets claiming multiple areas and appearing in different cit-
ies. To illustrate, estimates suggest that there are more than 1100 gangs in 115 cities
throughout the USA with “Blood” or “Crips” in their name (Howell, 2012,p. 13).
As the gangs grew and subdivided or sprung up spontaneously in new communi-
ties, the likelihood of inter-group violence also increased due to competition over
resources and reputation (Howell, 2012, p. 18). While in some areas consortium
alliances continue to foster positive associations (e.g., Descormiers & Morselli,
2011), in other cities intra-consortium fighting predominated (e.g., Decker & Curry,
2002). It came to a point in which Bloods were no longer fighting only against
Crips; rather, they fought amongst one another as well. Bloods attacked Bloods and
Crips victimized Crips. Recent inquiry into the nature of gang rivalries reveals how
violence diffuses through these networks of connected groups.

Networked Violence

Coined as hyperdyadic contagion by Christakis and Fowler (2009), various phe-


nomena, such as information, emotions, and disease, have been shown to routinely
pass through social networks, moving from person-to-person like an infection.
Individuals on the receiving end will react to what they have learned or experienced,
and this response will be influenced by the structure of the social neighborhood
within which each person is embedded. Their reaction will then facilitate additional
ripple effects back towards the origin and forward, infecting new people. Applied to
gang violence, studies show that risk for involvement in violence also spreads in this
fashion (e.g., Papachristos et al., 2013, 2015). When a gang member becomes
embroiled in a dispute, that begets an injury or perceived harm to reputation or sta-
tus, the individual (or group) will react in some fashion.
Papachristos (2009) likens the use of violence among street gangs to Ivan Chase’s
(1980) description of how chickens establish a pecking order. Acts of aggression
and displays of dominance are used to establish social order. By winning a conflict,
one chicken subjugates another. Subjugation then leads to a decline in social status.
To halt one’s fall to the bottom of the social order, the loser of the conflict then
attacks another, weaker than themselves. If not, the subjugated chicken may become
the victim of additional attacks. Among street gangs, aggressive posturing and use
of violence helps to establish, maintain, and elevate social status. This works when
the use of violence is publicly known—this means that information travels beyond
the dyad of the offender and victim to the rest of the network. In short, violence
spreads like a contagion through the network as individuals react to the behavior of
others (Papachristos, 2009).
168 J. Randle and G. Bichler

Previous studies found evidence of two types of reaction, each generating


identifiable network structures.
1. Reciprocity between dyads. Perceived harm may be reciprocated in an act of
retaliatory violence: since individual gang members are enmeshed within a
group with institutional memory, retaliation may occur, even if the original com-
batants are dead or jailed. Thus, long-standing rivalries are likely to generate
reciprocal attacks as groups struggle for dominance. In Chicago for example,
reciprocal attacks accounted for 37 % of gang-related homicides (Papachristos,
2009) and in a subsequent study, reciprocity was found to be a significant predic-
tor of fatal and nonfatal gunshot injuries in Boston and Chicago (Papachristos
et al., 2013).
2. Partial transitivity among triads. Violence can also spread outward to those not
directly involved, as the aggrieved party attempts to save face in a status-
conferring use of aggression (Papachristos, 2009). The need to avoid subjugation
can result in different network structures that reflect local hierarchies among
triads. For instance, we may see a domino effect, where the victimized person or
group, responds by attacking another: in network terms, this chain-like structure
is referred to as a directed or transitive tie. Additionally, groups may exhibit high
levels of aggression by attacking many other groups (e.g., high outward two-star
formations or outdegree centrality) or a group may come to be repeatedly victim-
ized (e.g., inward two-star formations or indegree centrality). For instance,
Descormiers and Morselli (2011) and Papachristos et al. (2013) uncovered
highly skewed distributions with a small number of groups exhibiting signifi-
cantly higher outdegree and indegree scores.

Intra-Group Conflict

The use of aggressive posturing and violent conflict is not restricted to competitions
for dominance between rival gangs. In fact, several studies document the existence
of intra-gang or intra-consortium violence. For example, Decker and Curry (2002)
found that gang-related homicides in St Louis occurred more often among groups
with the same affiliation: 61.5 % of Crips were killed by other Crips. Descormiers
and Morselli (2011) offer a glimpse into this phenomenon. Through interviews with
20 detained youth gang members, they obtained the information needed to map the
structure of positive and negative associations among 35 groups active in Montreal,
of which 31 % were Bloods subsets and 37 % were Crips. As expected, the majority
of alliances were intra-gang (88 %) and conflict tended to be reported between con-
sortiums (77 %). However, intra-group conflict was found to occur when subsets
were observed to be acting in a fashion that contradicted group norms. Descormiers
and Morselli (2011: 307) report an interviewee stating:
If they represent the same color than our gang, well they better represent it well. If they
don’t do things right, we will have to sort them out (Dawggi, member of the Bloods affili-
ated gang, AYB)
12 Uncovering the Social Pecking Order in Gang Violence 169

Intra-subset violence occurs when members of a group have a personal dispute


or cliques within the group vie for control. Evidence suggests that some leadership
disputes arise when previously incarcerated individuals return to reclaim their posi-
tion or are perceived as doing such (e.g., Papachristos, 2009: 93). For instance,
McCuish et al. (2015) constructed a co-offending network of criminal activity com-
mitted over 14 years by 137 people who were part of the BC gang or who co-
offended with gang members. Then, they examined two murders orchestrated by the
acknowledged leader of the gang. While the focus of this study was on the change
in social status that occurred with the use of violence and the characteristics of the
individuals recruited to participate in the crimes, the motive for each crime is infor-
mative. The victim of the first murder was a rival gang member and this event served
to catapult the offending group into a position from which they came to dominate
drug trafficking activity in the area. The second murder, however, was intra-group:
a prominent member was targeted for setting up his own criminal operation.
Interestingly, the murder occurred after the head of the gang was released from
incarceration.
Sometimes these internal conflicts are intensive enough to split the group, result-
ing in two new factions. If the groups maintain part of the original identity, that is,
they continue to see themselves as Crips, then any ensuing conflict between the
newly formed subsets would be classed as intra-gang conflict between different
subsets.

Spatial Structure of Rivalries

Geographic patterns of street gang violence have been widely studied and the gen-
eral finding is that greater density, particularly when there is competition for a stake
in an illicit market, correlates with more violence (e.g., Taniguchi, Ratcliffe, and
Taylor 2011). The conventional explanation is that these patterns of spatial concen-
tration emerge when competing groups claim turf in close proximity. For example,
Robinson et al. (2009) studied the density of gangs in Los Angeles County to test
the hypothesis that homicide rates would be higher in areas with a greater number
of nearby potential rivals. Excluding zip codes with low populations and vast uni-
versity campuses, they found that areas with higher concentrations of street gangs
were directly correlated with elevated homicide levels. Zip codes that contained
more than ten street gangs in a 2-mile radius (of zip code’s center point) accounted
for 40 % of all the homicides in Los Angeles County; yet, these zip codes contained
only 16 % of the County’s population (Robinson et al., 2009).
However, when we consider that conflict diffuses through a network, a more
nuanced understanding of gang violence begins to surface. Through a series of stud-
ies, Tita and colleagues investigated the nature of gang violence among criminally
active street gangs claiming territory in Hollenbeck, a neighborhood in Los Angeles
known for violent gang rivalries (e.g., Tita & Radil, 2011; Brantingham, Tita, Short,
& Reid, 2012). Three critical findings emerge from this body of work.
170 J. Randle and G. Bichler

1. Spatial patterns of homicide were nonrandom, showing some degree of


clustering.
2. Yet, the most violent rivalries often involved groups claiming turf in noncon-
tiguous areas and rivalries accounted for more violence than geographic
proximity.
3. Inter-gang violence tends to concentrate in areas along a competition boundary,
located between the two rival groups, at equidistance from their home turf or
gang set space.
What Tita and colleagues add to the discussion of networked gang-violence is
that it is important to understand how social interactions, in this case rivalry between
groups, can offer an alternative explanation for spatial patterns of violence.

Current Study

One gap in the extant network-oriented literature is that aggressive posturing and
violence is also used to dominate the local neighborhood through attacks on non-
gang involved people. Asserting dominance within the community is important for
several reasons. A few highly visible attacks on non-gang citizens may cast an
image of omnipresence across the community which may far exceed the group’s
actual control or influence. It follows that this generalized threat could serve the
group well, i.e., acts of public violence may intimidate community members and
decrease interference in group activities that might cut profits gained from illicit
enterprise. Moreover, we know that gang membership is inherently unstable.
Irrespective of the reason why people average only a few years of active involve-
ment, public displays of violence market the group’s prowess and may serve to
attract new members. When the person attacked is not known to be a gang member,
this may have a greater impact on potential recruits that are unaware of the real size,
capabilities, and inter-gang reputation of the group. While Papachristos (2009)
acknowledges that a substantial amount of networked gang violence involves non-
gang-affiliated people, the current study is among the first to investigate how map-
ping these attacks helps us to understand violence networks.
The second contribution made by this chapter is that to our knowledge, this is the
first attempt to map the network structure of violence accruing from the historical
Bloods/Crips rivalry across a region. By examining street gang violence through
this lens, we are able to examine three suppositions about the nature of networked
violence that are drawn from prior studies of city-level violence.
1. Gang violence will reflect the historic rivalry between the Bloods and the Crips.
Most conflict will involve inter-gang violence among subsets, suggesting that
continued retaliation (reciprocal attacks) may be associated with institutional-
ized conflict. It is also expected that local hierarchies will be observed where
some subsets come do subjugate their rivals. Subjugation of subsets may be
observed as a knock-on or domino effect (Group A attacks Group B, and Group
12 Uncovering the Social Pecking Order in Gang Violence 171

B attacks Group C), or a transitive pattern wherein one group attacks many
others or is victimized by many others. These patterns would suggest that an
inter-gang pecking order exists.
2. Local hierarchies will reveal intra-gang conflict. Some subsets will be more
aggressive toward others within the gang as factions vie for dominance over
their counterparts. Subjugation of subsets within a consortium may be observed
as a knock-on or transitive pattern suggesting an intra-gang pecking order exists.
3. Conflict is unlikely to spill across city boundaries, as groups are more apt to
compete with more spatially proximate rivals, and if inter-city violence does
errupt, it is unlikely that groups are from adjacent cities.
If found to be scale invariant, this study offers additional support for the thesis
that gang violence spreads through social networks.

Case Generation

We used information extracted from California court cases in which one or more of
the defendants was suspected or known to be a Blood or Crip gang member at the
time of the incident. To be included in the study
1. The case involved at least one charge/conviction for assault with a deadly
weapon, attempted homicide, or homicide.
2. The defendant was tried as an adult.
3. The violent event occurred in Los Angeles County between January 1st 2002 and
December 31st, 2010 (a period of 9 years).
Eligible court cases were found through LexisNexis, an electronic library that
provides information about federal and state cases. Due to the nature of recording
practicies for California, most of the cases identified through LexisNexis were
appeals for homicides or attempted murder. Thus, the findings reported here cannot
be generalized to less serious violence. The search protocol began with the names
of each active subset (a.k.a. factions) of the Bloods and Crips. This list was gener-
ated with the assistance of two detectives working in jurisdictions with entrenched
gang activity: one detective was from the Inglewood Police Department, and the
other, worked for the Hawthorne Police Department. Each detective modified a list
of Blood and Crip subsets that was retrieved from the Los Angeles County Street
Gangs website (www.streetgangs.com) by adding new factions and removing
groups that were defunct or inactive since January 2002. In addition to ensuring that
the lists were current and complete, detectives were also asked to note which sub-
sets were rivals. The two lists were consolidated and discrepancies resolved prior to
initiating the case identification process. Our search uncovered 284 cases that satis-
fied the inclusion criteria, and these violent crimes involved 439 offenders (includ-
ing defendants and accomplices) and 384 victims. Please note that that the sample
is primarily male: less than 5 % of offenders/victims were female.
172 J. Randle and G. Bichler

Network Generation and Description

Subset-to-subset network. We captured the network of all inter- and intra-gang vio-
lence by aggregating events to the gang subsets. The subset associated with each
defendant (the primary defendant and up to three accomplices) was linked to the
subset associated with each victim. This means that the network is weighted to
reflect the number of victimizations. In total, there were 158 gang subsets, of which
there were: 38 Bloods subsets (24 %), 109 Crips subsets (70 %), and 10 subsets with
no Bloods or Crips affiliation (6 %). Individuals with no-known gang affiliation
were assigned a unique research identification code. This strategy ensured that non-
gang members are not treated as a single group as this would bias the centrality
measures used (see Malm, Bichler, & Nash, 2011). In each dyad, the linkage origi-
nates with a defendant and ends with a victim. The subset-to-subset network had 43
components joining 251 unique entities (37 % of these entities are victims with no-
known gang subset affiliation) with 625 total victimizations (including reflexive ties
that indicate intra-group conflict). Constructing the network in this fashion permits
analysis of the hierarchical structure and magnitude of subset initiated violence.
City-to-city network. A geographic network was produced by aggregating all
gang-affiliated offender-victim pairs to the city associated with the home “turf” of
the subset.
Linking the city associated with the home turf of the defendants’ gang subset to
the city associated with the victims’ claimed gang territory permits an assessment
of where intra- and inter-gang conflict crosses city boundaries. For example, if an
individual that identifies with the Black P-Stones attacks an individual that identi-
fies with the Carver Park Compton Crips, there will be a link between Los Angeles
City and Compton, as the Black P-Stones are based in Los Angeles and Carver Park
Compton Crips are located in Compton. Since both the offender and victim must be
associated with a gang to be included in this analysis, the network contains only 206
links. Of the 15 cities1 identified, the city-to-city network included 12 cities arranged
in a single component.

Structural Variables

Rivalry. Reciprocity occurs when an offender later becomes the victim in another
attack (see the node labeled i in Table 12.1 where the solid line depicts the initial
attack and the dotted line illustrates the retaliation).
A word of caution is warranted. A time window effect may reduce the number
of reciprocal attacks identified at the beginning and end of the study period; that is,

1
Gang subsets were from Altadena, Antelope Valley, Athens, Carson, Compton, Florence, Gardena,
Hawthorne, Inglewood, Long Beach, Los Angeles, Lynwood, Pasadena, Pomona, and Santa
Monica.
12 Uncovering the Social Pecking Order in Gang Violence 173

Table 12.1 Variables used to uncover the structure of inter- and intra-gang violence
Structural variable Visual representation
Reciprocity

Dominance
Degree centrality Outdegree Indegree
Transitive
structures
Chain-like (directed
or transitive ties)
Two-star formations Outward two-star Inward two-star

Table 12.2 Distribution of cases


Cases Years
2002 2003 2004 2005 2006 2007 2008 2009 2010 AVG
Percent 8.4 9.5 12.5 22.0 12.5 12.1 9.2 8.8 5.1 11.1
Percent 17.4 7.7 23.5 25.0 20.6 18.2 28.0 12.5 21.4 19.4
btw rivals

events occurring at the beginning of the study period may have been in response to
a violent attack occurring before January 2002, but this connection would not have
been captured in the data. Reciprocity may also be affected by several other issues
that are associated with a prosecutorial effect. For example, the length of time
needed to try a major case may push it beyond the study period; plea bargaining
may reduce the number of cases that go to trial; events within a retaliatory sequence
of violence may involve misdemeanors rather than serious felonies; and, some vio-
lent attacks remain undetected or unsolved. Table 12.2 reports that the distribution
of cases peaks in 2005 and that the time window effect may present a greater threat
at the tail end of the interval with only 5.1 % of cases occurring in the final year of
the study period. However, if rivalries are the primary impetus of reciprocated
violence, and the temporal boundary of the study was a problem, we would not
have found a slightly higher than average amount of rivalry cases in 2010. Thus,
our ability to observe reciprocity should not be unduly compromised by a time
window effect.
Dominance. Dominant groups can be identified with two sets of statistics—degree
centrality and transitivity. Degree centrality counts the relative number of links that
each actor possesses (Hanneman & Riddle, 2005). In this study, the links between
actors are directed, permitting us to calculate outdegree and indegree centrality.
Outdegree centrality assesses the number of ties that an actor extends toward
others. Considering the way the networks were generated for this study, a subset
174 J. Randle and G. Bichler

with a high outdegree centrality score is considered to initiate more attacks than
other groups in the network. Since violence may occur among subset members,
reflexive ties capture intra-group violence. Indegree centrality counts the number of
ties received from others. As used here, subsets with high indegree scores are vic-
timized more often and/or by more groups.
By comparing outdegree and indegree centrality scores, we can identify the
dominant groups: those subsets that offend more often than they are attacked.
However, direct aggression alone does not fully reveal the dominance of a group.
Another facet of dominance can be uncovered by examining transitive relations.
Transitivity is a higher order structure involving links among sets of three actors
(Holland & Leinhardt, 1976; Wasserman & Faust, 1994). Here, we are interested in
potential transitivity or “intransitive” structures among triads, that is, sets of three
subsets.
Two forms of potential transitivity are of interest. Examined together, the exis-
tence of chain formations (directed or transitive ties) and two-star formations
(inward and outward) highlight the presence of local hierarchies within the network
(see Table 12.1). Chain-like formations among gang subsets occur when one subset
victimizes another, who in turns attacks a third group. Two-star formations occur
when a group attacks two other groups (outward), or alternatively, when one group
is victimized by two other subsets (inward). Combined, these patterns suggest that
within local circles, some groups are more dominant than others. If these patterns
are observed to illustrate intra-group conflict and reciprocated ties are absent, then
the hypothesis that gang violence is fueled by the Blood/Crip rivalry would not be
supported.
For the city-to-city network, ties represent the total number of gang-on-gang
violent incidents (incidents involving non-gang victims are excluded from this anal-
ysis). Again, in- and outdegree centrality identify the recipients and exporters of
violence, this time at the city level; reflexive ties are indicative of within city con-
flict; reciprocal ties show cities that may be locked in battle or experiencing gang
migration; and, transitivity is taken to reflect an inter-city hierarchy. UCINET 6 was
used to create the networks and calculate all of the structural statistics (Borgatti,
Everett, & Freeman, 2002).

Results

Subset Violence

Rivalry. Table 12.3 reports on reciprocity. Most of the conflicts observed, 67.2 %,
were between subsets and people from the community with no-known gang affilia-
tion. The Crips were more likely to attack non-gang individuals. The Crips were
also involved in more inter-gang attacks and intra-gang conflict.
Turning to gang-on-gang conflict, 62 % of attacks were intra-gang (51.7 % same
gang and 10.2 % same subset) as opposed to between different gangs (37.6 %).
12

Table 12.3 Reciprocity of violence within and across groups


Within group Across groups
Intra-subset Intra-gang Inter-gang Gang and community
All attacks
Nature of attacks 11 Bloods 28 Bloods-on-Bloods 28 Bloods → Crips 146 Bloods → non-gang
10 Crips 78 Crips-on-Crips 35 Crips → Bloods 268 Crips → non-gang
3 Bloods → other 2 non-gang → Bloods
9 Crips → other 5 non-gang → non-gang
2 other → Bloods
Number of victimizations 21 106 77 421
Percent of all attacks (N = 625) 3.36 % 16.96 % 12.32 % 67.36 %
Percent of gang-on-gang attacks 10.24 % 51.71 % 37.56 % –
(N = 205)
Uncovering the Social Pecking Order in Gang Violence

Reciprocated violence
Reciprocated conflicts 5 Bloods 1 Bloods-on-Bloods 1 Bloods ←→ Crips –
7 Crips 3 Crips-on-Crips
Number of victimizations 22 26 3 –
Percent of all attacks (N = 625) 3.52 % 4.16 % 0.48 % –
Percent of gang-on-gang attacks 10.73 % 12.68 % 1.46 % –
(N = 205)
175
176 J. Randle and G. Bichler

Table 12.4 Difference in average degree centrality scores between Bloods and Crips
Bloods (n = 38 subsets) Crips (n = 109 subsets)
Outdegree
Mean (SD) 5.68 (8.00) 3.67 (4.82)
Min. to Max. 0 to 45 0 to 36
Difference of means 2.01
T-test (1 tailed) Bloods>Crips 0.05a
Indegree
Mean (SD) 2.05 (2.99) 1.06 (1.94)
Min. to Max. 0 to 12 0 to 14
Difference of means 0.99
T-test (1 tailed) Bloods>Crips 0.02a
Difference (Outdegree—Indegree)
Mean (SD) 3.63 (6.73) 2.61 (4.61)
Min. to Max. −8 to 33 −12 to 31
Difference of means 1.03
T-test (1 tailed) Bloods>Crips 0.16
a
Significant at the p < 0.05 level

Only a small proportion of these attacks generated reciprocal violence. The largest
percent of reciprocated attacks occurred within the gang, either between subsets
(12.7 %) or within the group (10.7 %). Put another way, of the 51 reciprocated vic-
timizations, about 94 % were within group. In the 9-year period examined, two
reciprocated relationships were notably unbalanced:
• Insane Crips caused 5 victimizations and the Rolling 20s Crips attacked twice.
• Rollin 60s Neighborhood Crips generated 12 victimizations and the 83 Gangster
Crips returned one act of violence.

Dominance. Comparing average centrality scores, we see that even though there
are more Crips groups in the sample, the Bloods are involved in more violence.
Table 12.4 reports difference of means T-tests that indicate that subsets associated
with the Bloods are significantly more likely to victimize others and to be attacked.
By calculating the difference between outdegree and indegree centrality, it is
possible to see whether, on balance, one gang is more aggressive than the other.
This means that they would victimize more than they are victimized. While the
T-test fails to achieve significance, on average, Bloods subsets cause more violence
than they receive.
Table 12.5 names the subsets with the highest levels of aggression. Since it is
important to consider overall violence and gang-on-gang violence, two sets of sta-
tistics were generated. To compare structural statistics generated from different net-
works, the outdegree centrality scores are normalized. This accounts for the
difference in network size—the full victimization network contains 625 victimiza-
tions and the gang-on-gang network includes 205 attacks.
At first glance it would appear that the Black P-Stone Bloods are the most aggres-
sive subset in Los Angeles County during the 9-year period examined. However,
12 Uncovering the Social Pecking Order in Gang Violence 177

Table 12.5 Comparison of the five most aggressive subsets from each gang
Rank Bloods Crips
Subset name NrmOutDeg Subset name NrmOutDeg
All violent attacks (N = 625)
1 BLACK P-STONE 0.926 ROLLIN 60S 0.741
BLOODS NEIGHBORHOOD CRIPS
2 BOUNTY HUNTER 0.391 INSANE CRIPS 0.432
BLOODS
3 INGLEWOOD 0.350 GRAPE STREET WATTS 0.370
FAMILY BLOODS CRIPS
4 PASADENA 0.247 EAST COAST CRIPS 0.247
DENVER LANE
BLOODS
5 PUEBLO BISHOP 0.247 TRAGNIEW PARK CRIPS 0.226
BLOODS
Gang-on-gang conflict (N = 205)
1 BLACK P-STONE 0.538 ROLLIN 60S 1.075
BLOODS NEIGHBORHOOD CRIPS
2 FRUIT TOWN PIRUS 0.336 INSANE CRIPS 0.336
3 BOUNTY HUNTER 0.269 ROLLIN 40S 0.269
BLOODS NEIGHBORHOOD CRIPS
4 CRENSHAW MAFIA 0.202 EAST COAST CRIPS 0.269
GANGSTER
BLOODS
5 PASADENA 0.202 ACACIA BLOCK CRIPS 0.269
DENVER LANE
BLOODS
Note: Bold text highlights subsets that are not present in the top five for both networks

removing non-gang-associated victims reveals that among gang subsets, the Rollin
60s Neighborhood Crips are the most aggressive group. While the Black P-Stone
Bloods remain the most aggressive subset of the Bloods, their normed outdegree
centrality score is half of the Rollin 60s Neighborhood Crips.
Many of the highest ranking subsets remain in the top 5, irrespective of the way
in which the network is configured. Of note, two groups swap out for each gang.
Depicted in bold text, the Inglewood Family Bloods and the Pueblo Bishop Bloods
are dominant when all attacks are considered, but they drop from a rank in the top 5
when only gang-on-gang violence is examined. This means that these groups have
a high ranking because of the number of non-gang victims they target. A similar
pattern is apparent for two Crips subsets.
To complete this investigation of local dominance we count the number of
directed lines (transitive ties that form chain-like sequences) and star formations
(inward and outward). About 24 % of victimization paths are directed lines or
chain-like sequences of violence (Table 12.6). The clear majority of potentially
transitive ties are outward 2 stars where one subset victimizes two other groups.
Since many of the victims were individuals with no known gang affiliation, this
178 J. Randle and G. Bichler

Table 12.6 Transitive patterns for the entire network and gang-on-gang conflict
Full network Gang-on-gang
(N = 625) conflict (N = 205)
Structure Visualization Count Percent Count Percent
Potentially transitive
Directed line 324 24.02 78 41.71
Outward 888 65.83 50 26.74

Inward 63 4.67 32 17.11

With reciprocity
Outward with one 64 4.74 17 9.09
reciprocal tie
Inward star with one 10 0.74 10 5.35
reciprocal tie
Fully reciprocal star 0 0.00 0 0.00
formation

Total 1349 100.00 187 100.00

abundance of outward star structures may be inflated since individuals were not
combined into groups.
Selecting only conflict among subsets, we find a considerably different structure.
About 42 % of the violence occurs in a directed line and 27 % exhibits an outward
star. This finding lends support for the notion that gang violence contributes to the
development of a pecking order with some groups subjugating others. Unable to
retaliate, the victimized group turns on another, perhaps weaker subset. While this
simple structure predominates, we see more reciprocal ties when only gang-on-
gang violence is considered.
Subset Violence Visualized. Figures 12.1 and 12.2 illustrate the patterns of
violence described above. This type of visualization is called a digraph because it
depicts who offends. Symbols are color coded to differentiate between Bloods
(dark grey circles), Crips (light grey circles), individuals (white squares), and other
gangs (very light grey squares). Symbol size varies by the number of attacks that
originated from the group (outdegree centrality). Arrowheads indicate the direction
of attack. To make the image easier to read, arrowheads are resized to reflect
the number of victimizations (similar to the lines). Grey lines with a label of “r”
illustrate channels of reciprocated violence. The looping lines indicate reflexive ties.
Reflexive ties show intra-subset conflict.
Figure 12.1 presents the main component and Fig. 12.2 depicts the 42 smaller
components. The main component is the largest interconnected group where every
Fig. 12.1 Main component illustrating subset-to-subset violence. Arrowheads note the direction
of attack and line thickness indicates the number of victim/offender dyads. Symbol size varies by
outdegree centrality and grey ties illustrate reciprocal and reflexive violence

Fig. 12.2 Subset violence among smaller components. Arrowheads note the direction of attack
and line thickness indicates the number of victim/offender dyads. Symbol size varies by outdegree
centrality and grey ties illustrate reciprocal and reflexive violence
180 J. Randle and G. Bichler

subset and non-gang-affiliated victim is connected by at least one incident. Often,


structural analysis focuses on this part of the network.
One of the most striking features of Fig. 12.1 is that the subsets are not equal.
Some subsets were much more active during the 9-year study period. The most
active group (largest dark circle) was the Black P-Stones, but their violent reputation
comes from attacking more non-gang individuals than other gangs or subsets. In
fact, they victimized one other Bloods group, and five Crips subsets (one group
reciprocated). In turn, they were attacked by two different Crips subsets. We draw
attention to this conflict because it is representative of the overall pattern of violence.
Rather than being dominated by entrenched, mutual conflicts, there is a greater
tendency to be involved in violent chains or local conflict. Figure 12.2 brings this
point out even further.
Most of the smaller components (64.3 %) contain at least one star formation. Yet,
there are no proper transitive triads showing direct connections among sets of three
nodes. Full transitivity is missing. To achieve complete transitivity we would see a
dense network where each group is connected to all other groups. Our overall den-
sity is 0.37 % and the gang-on-gang network exhibits 1.20 % density. While these
values are lower than reported elsewhere (e.g., Papachristos, 2009; Descormiers &
Morselli, 2011), the results are consistent with expectations. Gang violence net-
works are relatively sparse. The two-star formations are connected in the main com-
ponent with chains of violence. These directed or transitive ties are infrequently
found in the smaller components. Smaller components are characterized by two-star
formations.
Thus far, this analysis suggests that serious gang violence in Los Angeles County
shares some of the same properties as city-level violence.
1. Few attacks are reciprocated—24.9 % of gang-on-gang attacks, but only 8.2 % of
all attacks. While Papachristos (2009) found that the majority of homicides were
not reciprocal, his study did uncover a higher portion of reciprocated violence
(35 %).
2. While there is evidence of inter-gang conflict, about 62 % of gang-on-gang
attacks involve subsets of the same gang. Generally, there is more conflict within
gangs than between Bloods and Crips. Our finding is consistent with Decker and
Curry (2002).
3. Most attacks (67.4 % of all victimizations) involve individuals without any
known gang affiliation. A similar finding was obtained by Papachristos (2009).
One finding, that to some extent contradicts prior research (see for example
Papachristos et al., 2013, 2015), is that there is evidence of local hierarchical struc-
tures, as identified by two-star formations and transitive ties. The next question to
investigate is whether this conflict is exported between different cities.
12 Uncovering the Social Pecking Order in Gang Violence 181

COMPTON LONG BEACH

ATHENS INGLEWOOD

LYNWOOD

CARSON

LOS ANGELES
ANTELOPE VALLEY

PASADENA

HAWTHORNE

SANTA MONICA ALTADENA


POMONA FLORENCE
GARDENA

Fig. 12.3 Gang subset conflict among cities. Symbol size denotes outdegree centrality; line width
shows the number of offender/victim dyads, and arrowheads indicate where violence was exported
to (direction of attack)

City-to City Violence

Precise geographic information was not available for all gang subsets included in
this study. Consequently, subsets were aggregated to the city in which their primary
turf was located. Figure 12.3 illustrates this aggregation of violence using similar
techniques as the prior figures: symbol size and line width vary to reflect the number
of victimizations; arrowheads designate the direction in which the violence flows;
double-headed arrows identify reciprocated attacks; loops indicate internal conflict;
and, straight lines depict exported violence.
The City of Los Angeles is a clear focal point for gang-related violence in the
county. While LA exports violence to several cities, gangs from other cities do
victimize people from LA. Cities with long-standing, entrenched gang activity—
Inglewood, Long Beach, and Compton—are also involved in a high level of vio-
lence. Reciprocal ties exist between LA and Inglewood and LA and Long Beach.
Hawthorne, Santa Monica and Gardena are isolated (not connected to the main
component) because cases involving subsets located in these cities concerned non-
gang affiliated people during the study period. These patterns reflect
prior neighborhood-level research by Tita and colleagues (e.g., Tita & Radil, 2011)
showing that conflict between groups is not strictly a matter of turf proximity—Los
Angeles borders Inglewood but not Compton or Long Beach.
182 J. Randle and G. Bichler

What Does This Mean for Crime Prevention?

Gang Injunctions

Injunctions are often used in California to suppress street gang violence. These
court orders are behavioral prohibitions that are lodged against a gang. Typically,
the stipulations prohibit gang members and their associates from participating in
specified activities within designated areas. The first gang injunction was
imposed in 1987 in the City of Los Angeles against the Playboy Gangster Crips.
Currently, there are more than 40 permanent injunctions against 72 gangs in the
city of Los Angeles alone. Arguments against the use of injunctions posit that
these civil restrictions are served in a manner that assumes guilt prior to provid-
ing proof, while arguments in support of injunctions suggest that they restrict
public association among gang members that intimidates other community
members and engenders further criminal activity (Felson, 2006).
Permanent gang injunctions target 17 Blood and Crip gang subsets in the City of
Los Angeles (City Attorney of Los Angeles, 2013). Table 12.7 lists 13 of these
subsets that appeared in the observed networks. We report the percentage of victim-
izations included in this study that occurred prior to and after the filing of the injunc-
tion. Groups ranked in our list of the top five most aggressive subsets are noted. This
ad hoc analysis raises questions about the influence of gang injunctions.
No clear pattern exists. The permanent injunction against the Black P-Stone
Bloods, one of the groups observed to victimize the most, may have reduced their use
of extreme violence following the court order: we find that 76 % of the victimizations

Table 12.7 Sample of permanent injunctions against LA City gang subsets


Percent of victimizations included
Subset Date injunction filed in this study
Before injunction After injunction
Bloods
Black P-Stone Bloodsa Sept. 21, 2006 76 24
Bounty Hunter Bloodsa Dec. 2, 2003 21 79
East Side Pain Oct. 8, 2008 100 0
Swan Bloods Dec. 15, 2009 100 0
Crips
42 Gangster Crips Apr. 7, 2005 0 100
43 Gangster Crips Apr. 7, 2005 0 100
48 Gangster Crips Apr. 7, 2005 0 100
Geer Gang Crips Sept. 22, 2006 22 78
Grape Street Watts Cripsa May 25, 2005 50 50
Main Street Crips Dec. 15, 2009 100 0
Rollin 40s Neighborhood Cripsa Apr. 10, 2008 33 67
Rollin 60s Neighborhood Cripsa Nov. 24, 2002 3 97
School Yard Crips Sept. 22, 2006 100 0
a
Ranked in the top five as the most violent subsets in this study
12 Uncovering the Social Pecking Order in Gang Violence 183

we examined took place before the gang injunction was filed and only 24 % occurred
after. However, with enhanced penalties for gang activity and facing the ban against
congregating, the group may have simply incorporated additional sub-subsets to avoid
crime prevention efforts. Franchising or fracturing to extend territory is not uncom-
mon. For instance, the East Coast Crips have 11 sub-subsets located in four different
cities, the Neighborhood Crips have at least 5, and the Hoover Crips have 8 sub-sub-
sets in the City of Los Angeles. It is also plausible that injunctions cause groups to
reinvent themselves with new or modified names to avoid crime control efforts.
Nonetheless, the same cannot be said for the Rollin 60s Neighborhood Crips—the
Crips group with the highest record of observed violence. We find that 3 % of their
observed attacks occurred prior to the gang injunction and 97 % after the injunction.
Arguably the number of events may have increased after the gang injunction, not
because the group became more violent, rather, officers may have become more
focused on the criminal activities of this gang. In turn, the increased attention from
law enforcement would inevitably generate more cases for the prosecutor’s office.

Strengthening Intervention Strategies

Improving the effectiveness of behavioral prohibitions like injunctions, and their


successor focused-deterrence, may require rethinking the basis upon which the
behavioral limitations are made. As argued by Kennedy, Braga, and Piehl (1997),
tapping into the network is critical to the success of these types of anti-crime efforts.
For a focused-deterrent to work, you need to know which groups and individuals to
target. You also need to identify the community stakeholders with the most social
capital with offenders, and thus the greatest potential to affect change (Kennedy
et al., 1997). Mapping out the collective social network through a group audit, and
applying network analytic techniques to uncover its structural features, is pivotal to
this endeavor (Sierra-Arevalo & Papachristos, 2015).
McGloin and Rowan (2015) succinctly argue that network analysis aids the
development of gang prevention strategies in three ways. First, social network anal-
ysis can be used to reveal rivalries and alliances, thereby highlighting specific
groups for attention during intervention planning and implementation. Second,
understanding the network’s structural characteristics, specifically the degree of
network cohesion in tandem with potential brokerage positions, is necessary for
identifying pressure points upon which to leverage initiatives that employ focused-
deterrence. And third, extending network generation procedures to include non-
gang members will expand the list of potential leverage points, such as finding those
with the most social capital or influence to include in the intervention effort.
We argue that it is also important to avoid setting narrow constraints on the
boundary of the network. That is, restricting group inclusion to jurisdictional bound-
aries or maps of gang set space is likely to generate structural holes in the group’s
collective social network. Various factors contribute to broadening the geographic
expanse covered by a social network. For instance, population migration extends
networks that may once have been spatially constrained. Also, people rarely spend
184 J. Randle and G. Bichler

all of their time in a single set of places, and with widespread vehicle access and an
extensive freeway system, the mobility potential in Southern California is excep-
tional. Another factor to consider is serving time in prison. Large, regional facilities
offer economies of scale, but they also bring people together from different com-
munities who might not have met otherwise. When released, these individuals
return to their home neighborhoods, and their newly extended social network may
help them to break earlier geographic constraints. Thus, social networks, and the
origin of the conflicts they produce, may transcend community borders.
In light of the results of this study, and for the reasons argued above, it would
appear that the injunction policy in Los Angeles should be revisited. Five directives
accrue from this research.
1. Network analysis should be included in the investigatory process to identify criti-
cal aspects of behavioral prohibitions. Emphasis should be placed on identifying
internal conflict and prohibitions enacted to quell said conflicts. Injunctions should
include rivals, such as other factions or subsets of the same gang, as well as allies.
2. Target the most active groups—both in terms of attacks on citizens with no
known gang affiliation and gang-on-gang violence.
3. The network observed in this study suggests that local hierarchies are predomi-
nant in the structure of extreme violence, mostly within the same gang, between
subsets. Targeting the most dominant groups is critical as they originate a greater
proportion of violence and fuel the process of contagion.
4. Social structures are dynamic. Groups reinvent themselves, they split and expand, and
the primary players change. Thus, permanent injunctions must be regularly amended.
5. Generating collaborative injunctions, that interlink several jurisdictions, will
provide greater effectiveness when the underlying social network of gang con-
flict is used as the basis for forming cooperative agreements among criminal
justice agencies. Successful regional efforts do not necessarily need the partici-
pation of adjacent communities—the social proximity of gang activity is more
important than spatial proximity.

Limitations

This study is not without its limitations. First, we focused on felonious vio-
lence resulting more often in extensive court action (appeals). Berg (2012) argues
that some characteristics of gang-related violence are the same in all jurisdictions:
(1) they tend to involve firearms, (2) occur in public places, (3) victimize males
between 15 and 19, and (4) often revolve around retaliation (2012). The last point is
of interest. Retaliation for an injustice does not mean that the same level violence,
if any violence at all, is reciprocated. Retaliation can take many forms. Using only
extreme instances of violence—assault with a deadly weapon, attempted homicide
and homicide—excludes a wide range of behavior that, if included, would better
capture the structure of subset conflict. It follows that future investigations of gang
violence should include a range of offending behavior.
12 Uncovering the Social Pecking Order in Gang Violence 185

Second, conflict is dynamic and aggregate patterns are the outcome of individual
posturing and competition for control of the group and to gain a better position
within the network of subsets. McCuish et al. (2015) found that by studying a single
group over time, they were able to document how engaging in violence, within and
external to the group, elevated the career trajectory of some gang members.
Interpreted in light of our findings, this suggests that any effort to study gang vio-
lence, should consider the trajectory of conflict. Evidence about the group’s net-
work, allies, affiliates, and rivals is time sensitive, and thus, dynamic network
analysis is needed to understand the trajectory of the group’s structure. Several stud-
ies by Papachristos and colleagues illustrate this point (e.g., Papachristos, 2009;
Papachristos et al., 2013, 2015). On a related note, our study used a time window
spanning 9 years. Subsequent investigations of the rivalry between the Bloods and
Crips could employ a number of narrow intervals to improve temporal sensitivity.

Conclusion

Network analysis provides a series of theoretically derived tools that can significantly
strengthen our understanding of criminal behavior so as to improve the effectiveness
of our crime prevention efforts. Most prosecuted cases of gang violence in Los Angeles
County were observed to involve intra-gang rivalries and gang attacks on individuals
with no known gang involvement. By revisiting a historic and widely accepted rivalry
between the Bloods and the Crips, this study demonstrated that the actual social struc-
ture of gang violence does not fully conform to our preconceptions.

References

Berg, N. (2012, January 31). The 5 U.S. cities with the worst gang violence. Retrieved October 24,
2013, from The Atlantic Cities website: http://www.theatlanticcities.com/
neighborhoods/2012/01/5-us-cities-worst-gang-violence/1095/
Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). Ucinet 6 for windows: Software for social
network analysis. Harvard, MA: Analytic Technologies.
Brantingham, P., Tita, G. E., Short, M. B., & Reid, S. E. (2012). The ecology of gang territorial
boundaries the ecology of gang territorial boundaries. Criminology, 50(3), 851–885.
doi:10.1111/j.1745-9125.2012.00281.x.
Caldwell, B. (2010). Criminalizing day-to-day life: A socio-legal critique of gang injunctions.
American Journal of Criminal Law, 37(3), 241–290.
Chase, I. D. (1980). Social process and hierarchy formation in small groups: A comparative per-
spective. American Sociological Review, 45(6), 905–924.
Christakis, N. A., & Fowler, J. H. (2009). Connected: The surprising power of our social networks
and how they shape our lives. Boston: Little, Brown.
City Attorney of Los Angeles, (2013). Gang Injunctions [Fact sheet]. Retrieved October 24, 2013,
from Los Angeles City Attorney website: http://www.atty.lacity.org/index.htm
Decker, S. H., & Curry, D. G. (2002). Gangs, gang homicides, and gang loyalty: Organized crimes
or disorganized criminals. Journal of Criminal Justice, 30, 343–352.
Descormiers, K., & Morselli, C. (2011). Alliances, conflicts, and contradictions in Montreal’s
street gang landscape. International Criminal Justice Review, 21(3), 297–314.
186 J. Randle and G. Bichler

Felson, M. (2006). Crime and nature. Thousand Oaks, CA: Sage.


Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods. Riverside, CA:
University of California, Riverside.
Holland, P. W., & Leinhardt, S. (1976). The statistical analysis of local structure in social networks.
Sociological Methodology, 7, 1–45.
Houston, H., & Anglin, D. (1995). The epidemic of gang-related homicides in Los Angeles County
from 1979 through 1994. Journal of the American Medical Association, 274(13), 1031.
Howell, J. C. (2012). Gangs in America's communities. Thousand Oaks, CA: Sage Publications.
Kennedy, D. M., Braga, A. A., & Piehl, A. M. (1997). The (un)known universe: Mapping gangs
and gang violence in Boston. In D. Weisburd & T. McEwen (Eds.), Crime mapping and crime
prevention, crime prevention studies (Vol. 8, pp. 219–237). Monsey, NY: Criminal Justice
Press.
Klein, M. W., & Maxson, C. L. (2006). Street gang patterns and policies. New York, NY: Oxford
University Press.
Malm, A., Bichler, G., & Nash, R. (2011). Co-offending between criminal enterprise groups.
Global Crime, 12(2), 112–128.
McCuish, E. C., Bouchard, M., & Corrado, R. (2015). The search for suitable homicide co-offenders
among gang members. Journal of Contemporary Criminal Justice., 31(3), 319–336.
McGloin, J. M. (2005). Policy and intervention considerations of a network analysis of street
gangs. Criminology and Public Policy, 4(3), 607–636.
McGloin, J. M., & Rowan, Z. (2015). Street gangs and co-offending networks. In G. Bichler &
A. Malm (Eds.), Disrupting criminal networks: Network analysis in crime prevention (Crime
Prevention Studies, Vol. 28, pp. 9–26). Boulder, CO: First Forum Press. division of Lynne
Rienner Publishers, Inc.
Papachristos, A. V. (2009). Murder by structure: Dominance relations and the social structure of
gang homicide. American Journal of Sociology, 115(1), 74–128.
Papachristos, A. V., Hureau, D. M., & Braga, A. A. (2013). The corner and the crew: The influence
of geography and social networks on gang violence. American Sociological Review, 78(3),
1–31.
Papachristos, A. V., Wildeman, C., & Roberto, E. (2015). Tragic, but not random: The social con-
tagion of nonfatal gunshot injuries. Social Science & Medicine, 125, 139–150.
Robinson, P. L., Boscardin, W., George, S. M., Teklehaimanot, S., Heslin, K. C., & Bluthenthal,
R. N. (2009). The effect of urban street gang densities on small area homicide incidence in a
large metropolitan county, 1994–2002. Journal Of Urban Health, 86(4), 511–523. doi:10.1007/
s11524-009-9343-x.
Sierra-Arevalo, M., & Papachristos, A. V. (2015). Applying group audits to problem-oriented
policing. In G. Bichler & A. Malm (Eds.), Disrupting criminal networks: Network analysis in
crime prevention (Crime Prevention Studies, Vol. 28, pp. 27–46). Boulder, CO: First Forum
Press. division of Lynne Rienner Publishers, Inc.
Taniguchi, T. A., Ratcliffe, J. H., & Taylor, R. B. (2011). Gang set space, drug markets, and crime:
Violent and property crimes around drug corners in Camden. Journal of Research in Crime and
Delinquency, 48(3), 327–363.
Tita, G., & Radil, S. (2011). Spatializing the social networks of gangs to explore patterns of vio-
lence. Journal of Quantitative Criminology, 27(4), 521–545. doi:10.1007/s10940-011-9136-8.
Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications. New York,
NY: Cambridge University Press.
Wilkinson, D. L., & Fagan, J. (1996). The role of firearms in violence ‘scripts’: The dynamics of
gun events among adolescent males. Law and Contemporary Problems, 59, 55–89.
Winton, R. (2005, May 13). L.A. Home Turf for hundreds of neighborhood criminal groups. Los
Angeles Times, Local.
Chapter 13
Static and Dynamic Approaches of a Drug
Trafficking Network

Andrea Giménez-Salinas Framis and Sonia Fernández Regadera

Introduction

In recent years, social network analysis has been amply used to develop social stud-
ies, especially those aimed at understanding covert networks. These types of net-
works are formed by players with broadly shared incentives to conceal their
activities and connections, and to keep at least some of them away from the public
eye (Robins, 2009). The lack of information about the different components of these
networks and their internal workings makes it harder to investigate and dismantle
them. In this regard, SNA, initially devised to study social groups, structures and
dynamics (Moreno, 1953), has proven to be a very useful tool to identify or predict
network structures and their components, to identify their weak points and to design
prevention strategies which are more successful, efficient and tailored to the specific
internal structure of each network. This is why it has been mainly used for juvenile
gangs (McGloin, 2005; Xu & Chen, 2008), criminal organisations (Bruinsma &
Bernasco, 2004; Kleemans & Van De Bunt, 2003; Morselli, 2009; Morselli &
Giguère, 2006; Morselli & Roy, 2008; Natarajan, 2006) and terrorist groups (Krebs,
2001; Carley, 2003). Most authors agree that it is a promising tool for investigation,
criminal intelligence and the study of criminal networks (Sparrow, 1991; Krebs,
2001; Morselli, 2009, Bright, Greenhill, Reynolds, Ritter, & Morselli, 2014b).
In relation to organised crime, apart from the traditional mafia organisations that
can still maintain hierarchical structures (Calderoni, 2012; De La Corte & Giménez-
Salinas Framis, 2010; Varese, 2013), most studies on non-mafia-type organisations
have revealed that their structures are more segmented, they are normally very flex-
ible and resistant (Benson & Decker, 2010), they have limited functional hierarchy

A.G.-S. Framis (*) • S.F. Regadera


Instituto de Ciencias Forenses y de la Seguridad, Universidad
Autonoma de Madrid, Madrid, Spain
e-mail: Andrea.gimenezsalinas@inv.uam.es

© Springer International Publishing Switzerland 2017 187


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_13
188 A.G.-S. Framis and S.F. Regadera

and they are able to reorganise rapidly (Ronfeldt & Arquilla, 2001; Williams, 2001;
Kleemans & Van De Bunt, 2003; Bruinsma & Bernasco, 2004; Morselli, 2009).
These types of networks especially include those dedicated to drug trafficking
(Adeler, 1993; Desroches, 1999; Gimenez-Salinas, 2014; Natarajan, 2000, 2006;
Zaitch, 2002). This is why the effectiveness and applicability of SNA to study these
criminal networks are important.
SNA does not only provide a graphical representation of the networks, but also
allows us to analyse their structure to support and expedite police investigations
(Krebs, 2001). Specifically, it helps detect subgroups within a network, uncover their
patterns of interaction and identify their key members. Although this can also be
done using other tools or methods, SNA brings to light aspects that are not easy to
detect at first sight (Morselli & Roy, 2008). Conversely, SNA enables police or any
intelligence analyst to identify the most central actors in the network in relational
terms (Schwartz & Rouselle, 2009). The police can thus monitor them and act in the
most effective way to weaken the network and its human and material resources. But
its value for investigators does not end here, as proven by Van der Hulst (2009). SNA
addresses seven dimensions that are valuable for investigations: scenario building,
risk analysis and threat assessments, hypothesis testing, network destabilisation and
identification of code names, matching roles and substitutes, support decisions and
deployment of intelligence assets, and evidence for prosecution.
Criminal networks have traditionally been analysed within a static dimension (at
a single point in time) but a dynamic perspective may also be valuable for preven-
tion purposes. Although some empirical approaches have been used to explore the
dynamics of networks, this area still requires further empirical research (Bright &
Delaney, 2013). Criminal networks are social groups subjected to internal and exter-
nal pressures, mainly because the exchange of illicit commodities requires covert
action. The alignment of actors in networks can change when new members join,
members are detained, due to expert requirements (Williams, 2001), or because of
changes caused by external pressures due to geographical expansion, market-driven
strategies, etc. The dynamic perspective of networks is extremely useful for preven-
tion and intelligence because, apart from better understanding internal network
structures, knowing how a network is growing or evolving over time and how it
responds to internal or external challenges are crucial to design effective strategies
to minimise its efficiency, or to test theoretical hypothesis about network flexibility
or responsiveness (Bright & Delaney, 2013).
Studies on dynamics in criminal networks can be classified as three approaches:
descriptive techniques to detect structural changes in social networks by comparing
structural properties over time; statistical approaches to detect, describe and explain
the changes; and simulation methods that rely on multi-agent technology to analyse
network dynamics. In this regard, Carley, Lee, and Krackhardt (2003) used a simula-
tion method to analyse the evolution of terrorist groups. Bright and Delaney (2013)
used a fourth visualisation method through which structural changes can be detected
by network maps. Further studies on network evolution and dynamics will provide
new insights into network continuity and ways to force network discontinuance.
This chapter analyses the value of SNA to investigate these networks. Using a
cocaine trafficking network investigated and dismantled in Spain several years ago
13 Static and Dynamic Approaches of a Drug Trafficking Network 189

as a sample, our analysis is twofold: from a static perspective, we focus on one level
of the organisation, for which we will identify its structure and main players; and
from an individual perspective, we analyse certain individual attributes of members
of the group and link them to their structural role. Then, we highlight the qualities
of SNA as a predictive tool within a dynamic view of the network. We use a descrip-
tive approach to explain the evolution of a network during the investigation. This
type of longitudinal analysis reveals how this dynamic approach is promising to
identify central actors early and prioritise initiatives to neutralise and dismantle
criminal networks.

Methodology

This study would not have been possible without the collaboration of the judicial
police unit of the Spanish Guardia Civil, which gave us access to the investigation
of a drug trafficking group in Spain.1 The network used as a sample for our analysis
is a cocaine trafficking network that was uncovered in Madrid (Spain). This group
brought Colombian cocaine to Spain through Uruguay or Brazil, to then adulterate
and distribute it in Madrid. It is a mid- or intermediate-level group (Adeler, 1993;
Natarajan & Belanger, 1998; Reuter & Haaga, 1989) between the Colombian
wholesalers and the small retailers in Madrid. It is a typical cocaine trafficking
organisation operating in Spain.
We collected quantitative and qualitative information about the drug trafficking
network investigated between 2006 and 2009 using three specifically designed
spreadsheets. One for data regarding the criminal network, another for members of
the criminal group and the third for relations and contacts for SNA purposes. The
spreadsheets are described further below.
(a) Data on criminal organisations: This spreadsheet included 76 variables, which
were divided into four main dimensions: (a) characteristics of the criminal
organisation (b) details of the illicit market and (c) instrumental or transverse
activities carried out. The variables collected are shown in the table below
(Table 13.1).
(b) Data on members: This spreadsheet was prepared to collect information about
members, including the following variables: age, sex, nationality, level of
education, family status, job situation, legal or illegal residency status (for for-
eigners), nationality, role in the organisation and police record.
(c) Data on contacts between members: This spreadsheet was prepared to collect
information about the number of contacts between network members. Contacts
were selected from two main sources of information: meetings corroborated by

1
Spanish police reports include ample data of each investigation and the characteristics of each
organisation and its members. In particular, each police report contains the techniques used by the
agents in the criminal investigation: entry and search, telecommunications interception, asset sei-
zure, photographic reconnaissance, interrogation, shadowing, etc. During the data collecting and
analyses process, the identity of each subject under investigation was anonymous, thus their right
to privacy was safeguarded.
190 A.G.-S. Framis and S.F. Regadera

Table 13.1 Criminal network data collection variables


Research dimensions Variables
Features of the organisation Number of individuals under arrest
Objective of the organisation in Spain
Structure
Location of the main headquarters
Division of labor
Collaboration with other groups
Technical and complex means deployed
Use of experts
Security means used
Police infiltration
Use of legal enterprises
Illegal market Dedication to more than one market
Main illegal market
Second illegal market exploited
Modus operandi
Revenue
Geographical location
International activity
Instrumental activities Corruption in Spain and abroad
Obstruction of justice in Spain and abroad
Money laundering activities
Violence within and outside organisations
Homicides committed
Other instrumental activities

the police during the investigation and the telephone contacts in the police files
(wiretaps). These contacts were recorded in a template which stated when they
took place, who the caller and the person receiving the call were and the dura-
tion of the call. The UCINET 62 software was used to carry out SNA.
Data from police investigations is difficult to come by and is a valuable source of
information to know more about covert networks. In addition, wiretaps are frequently
used to carry out SNA (Berlusconi, 2013; Natarajan, 2000; Varese, 2013) and can
supplement ethnographic research while posing no risk to the researcher. However,
police data also have major limitations as repeatedly referred to in literature: (a) data
from police investigations are incomplete and biased by the investigation and the
actions of the investigators (Krebs, 2002 & Sparrow, 1991); (b) the barriers between
the legal and illicit activities of the actors who are considered members makes it dif-
ficult to identify the members who are permanently linked to the network (Xu &
Chen, 2005) and (c) concealed networks tend to be dynamic and this fact is not
apparent from the information collected, which generally reflects static data and
realities (Bright & Delaney, 2013). Therefore, we can use the results of this investiga-
tion to exemplify the potential of static and dynamic analyses of criminal networks,
notwithstanding the fact that the initial information about these covert networks does
not accurately reflect reality but rather provides the closest approach to it.

2
Borgatti, S.P., Evertt, M. G. & Freeeman, L.C. (2002).
13 Static and Dynamic Approaches of a Drug Trafficking Network 191

Results

We now analyse the data from the investigation of the cocaine trafficking network in
Spain from a static and a more dynamic perspective. The detailed snapshot of the
network using SNA provides an insight into the network to uncover dynamics and
relationships that are not found using traditional approaches of study of criminal
organisations based on hierarchical structures (Gimenez-Salinas et al., 2012; Morselli,
2009). This type of static analysis is useful insofar as we have complete information
about the network, and its value increases proportionally the more information we
have. However, for the purpose of criminal investigation it is more useful to have
predictive tools to allow us to steer the investigation towards more relevant and vul-
nerable points for the network and to devise ways to neutralise it. Therefore, we will
show how a dynamic analysis of the network offers predictive advantages in criminal
investigation and in the analysis of intelligence. We now analyse the static approach.

Static Analysis of the Network and Potential of SNA

Analysing social networks provides information about different static dimensions


that can be classified in two main groups. The first is relational and enables the
analysis of existing links between the actors that form part of the network and pro-
vides structural conclusions about the same. Under this dimension, we will analyse
the general structure of the network, its main components, the core-periphery mem-
bers and subgroups. The second dimension is individual and provides an insight into
the network’s components, in this case the actors that belong to the network. With
regard to actors, we have collected a series of personal attributes and qualities that
supplement the information about the structure of the network (Robins, 2009) and
may be valuable to analyse the projection of the network in the future and its vulner-
abilities. For this study we were only able to collect the socio-demographic attributes
or characteristics of the actors, but other types of attributes may be more valuable for
prevention purposes. Since we had access to material and economic resources of the
network’s actors, we have been able to better assess their weight within the network
(Bright, Greenhill, Reynolds, et al., 2014b; Schwartz & Rouselle, 2009). Moreover,
with information about their roles, motivations and psychological aspects we could
have designed future strategies to reduce personal or collective incentives to remain
in the network and promote individual or collective network discontinuance.

The Network’s General Structure

Through a better understanding of the structural properties of a criminal network we


know more about the network’s topography, and can identify vulnerabilities and
devise destructive actions against it. The network’s general structure provides indi-
cators about a network’s vulnerabilities or resistance. In this respect, from a
192 A.G.-S. Framis and S.F. Regadera

structural perspective, there are two types of networks: exponential and scale
free networks (Albert, Jeong, & Barabasi, 1999). Exponential networks are structur-
ally homogenous, with most nodes having approximately the same number of links
(not many). On the other hand, scale-free networks are inhomogeneous, most nodes
having proportionately few links and a minority having a large number of links
(hubs). These structures have a significant impact on resilience. Scale-free networks
are much more vulnerable and less resistant because the network can be dismantled
by removing well-connected nodes (hubs) (Bright, Greenhill, & Levenkova, 2014a).
In the case of exponential networks, many more nodes need to be removed for the
same result. In other words, removing nodes with few connections has little impact
on the general structure.
When SNA was carried out on our drug trafficking network, we found a centralisa-
tion of 6.4 % and a density score of 15 %, with a standard deviation of 1.58.3
Comparing this to networks with a higher (Barnes, 1972) (80 %) or lower (Natarajan,
2006) (12.3 %) density, shows that it is not a dense network. In terms of centralisation,
it is not very centralised but this could be because the network has three subgroups.
Three nodes centralise most of the contacts (hubs). As Fig. 13.1 shows, JPPM, SNRM
and JJTE, who according to the investigation are the leaders of the network’s three
subgroups,4 comprise most of the contacts. In this case, if they are detained, most of the
network’s activity is neutralised because of contact centralisation. Apart from these
actors, the two leaders of the wholesale and importer group (SNRM and JPPM) are the
most central actors in terms of contacts and centralise most of the network’s contacts,
while the leader of the cocaine retail group (JJTE) comes a distant third. Therefore, the
network would be more efficiently disrupted by removing SNRM or JPPM.
If we analyse the core of the overall network5 and its peripheral members, we
see that the core is formed by 11 members (PCCP, JPPM, ELM, SNRM, HAMS,
WGV, CPGT, HPM, JDD, ARG and FR) of a total of 27 subjects investigated,
which means that 40.7 % of the members are in general strongly related within the
structure and have a similar relationship pattern coordinating tasks. The following
table provides general information about the actors of the network’s core, which is
comprised by seven members of the importer group (PCCP, JPPM, HPM, ARG,
FR, ELM and JDD) and four of the wholesaler groups in Spain (SNRM, HAMS,
WGV and CPGT).
Table 13.2 provides a description of the core and peripheral actors and their sex,
age, nationality and role. Apart from the main operational activity of the group

3
These measures are not the same as those provided in Gimenez-Salinas (2014) because the mem-
bers included in both networks are different. This article only includes the members that collabo-
rate with the organisation on a permanent basis.
4
The police reports show that JPPM is the retail group leader, SNRM is the wholesaler group
leader and JPPM is the importer group leader.
5
The method used is a core peripheral analysis that predicts the degree of coreness or closeness to
each actor’s core. This method exposes two types of nodes, namely a cohesive subgraph (the core)
in which actors are connected to each other in some maximal sense, and a class of player that is
loosely connected to the cohesive subgraph by a lack of maximal cohesion with the core (Borgatti
& Evertt, <CitationRef CitationID=”CR12” >1999</Citation Ref>). Actors in the core can coordi-
nate their actions and those in the periphery cannot.
13 Static and Dynamic Approaches of a Drug Trafficking Network 193

involved in the drug trafficking market (importer, wholesaler and retailer), the role of
all the actors in the organisation according to their main task has also been identified.
Their role in the groups has been divided into three categories: leadership, coordina-
tion and operation.6 As we can see, the police files (i.e. the description in the reports)
reveal that most of the core members carry out coordination tasks while the periph-
eral members mainly carry out operational and collaboration tasks. Exceptionally,
JJTE, the leader of the cocaine retailer group, appears as a peripheral member.

Table 13.2 Description and roles of core and peripheral actorsa


CORE members
Subgroup Name Sex Age Country Role
Importer PCCP F 41 Ecuador Operation (money transfer)
JPPM M 60 Ecuador Leadership
HPM M 50 Bolivia Operation (collaboration tasks)
ELM M 43 Colombia Coordination
FR M Ecuador Coordination
ARG M Colombia Operation (transport preparation)
JDD M 65 Spain Coordination
Wholesaler SNRM M Leadership
HAMS M 39 Colombia Coordination
WGV M 45 Colombia Coordination
CPGT F 44 Colombia Coordination
Peripheral members
Subgroup Name Sex Age Country Role
Importer OS M Colombia Operation (transport)
LIRP F Operation (collaboration tasks)
FRC M Operation (collaboration tasks)
JB M Operation (collaboration tasks)
JEAB M Operation (infrastructure supplier)
Wholesaler YPMG M Colombia Coordination
JHLM M
JE M Operation (collaboration tasks)
JO M Operation (collaboration tasks)
VR M Operation (collaboration tasks)
ERC M 41 Colombia Operation (collaboration tasks)
Retailer JJTE M 49 Spain Leadership
JV M Operation (collaboration tasks)
RB M Operation (collaboration tasks)
JG M Operation (collaboration tasks)
RAJH M Operation (collaboration tasks)
a
Some columns are empty because the police investigation did not provide this information

6
This category was created after checking the police files and comparing the information in them with
the information about their role in the organisation. Leadership includes group leaders according to
police interpretation and role information; coordination includes secondary positions in the organisa-
tion and management roles; and operation includes all other roles: transportation, debt collection, etc.
194 A.G.-S. Framis and S.F. Regadera

Figure 13.1 is provided in full for a complete picture of the network, which
shows subgroups of a network and the contacts of each of its actors in different
colours. As the table has been prepared using Mode 1 matrices with valued data, the
lines are thicker when the actors have been in contact more than once. As we had
access to the frequency of meetings between the actors and to telephone records, we
opted to use valued data matrices to provide the most complete sample of the
communications.

JV OS FRC
RB JJTE
ARG
JEAB JDD
LIRP
RAJH
JG HAMS FR
SNRM
JPPM
PCCP
CPGT
HPM
ERC JB
GYPMG ELM

VR JHLM
WGV JO JE

wholesaler Importer Colombia Retailer

Fig. 13.1 The network’s general structure

The Network’s Main Players

Apart from the general structure, SNA can be used to identify the most strategic
nodes of a network in terms of power or brokerage position (Morselli, 2010). The
best connected actors are those that play a central or intermediary role in the rela-
tionship with the other the actors; therefore, if they are identified early, investigators
or analysts can manage the surveillance and detection resources in a more precise,
efficient and effective way.
The most common measures used to identify the important players in a network
are centrality and intermediation (betweenness), but eigenvector centrality and
closeness centrality can also be measured using SNA. These measures work as fol-
lows: (a) Degree centrality measures the number of direct contacts surrounding a
node, which is usually associated with the leader of the group. (b) Betweenness
centrality measures the significance of a players’ intermediation in a network (how
often a node appears in the shortest route connecting two other nodes) (Sparrow,
1991). This position is also called Broker (Morselli, 2009). (c) Closeness centrality
measures how many steps on average it takes for an individual to reach everyone
else in the network. (d) Eigenvector centrality measures the extent to which actors
are connected to other well-connected actors.
Table 13.3 provides information about the actors in the network and the above-
mentioned measures. As we can see, JPPM has highest degree of centrality and
betweenness in the network. He is the leader of the cocaine importing group from
13 Static and Dynamic Approaches of a Drug Trafficking Network 195

Colombia to Spain. As regards geodetic distance, he is also the most prominent,


which makes him the most central in the network because he is connected to other
members with a high level of centrality. However, in relation to degree of interme-
diation, SNRM is the most prominent actor. He is the leader of the wholesale
group in Spain and an intermediary between the suppliers or importers in Colombia
and the retailers in Spain. His intermediary role is clear because he operates as a
link between the importer and the local distributors, which shows his capacity to
control the connectivity flow in the network and puts him in the position of
Broker’s edge (Morselli, 2009). If we had to decide which of the two actors (JPPM
or SNRM) to remove from the network to weaken it, the best option would be to
remove the one with the highest degree of intermediation due to his capacity to
disconnect the rest of the network (the two subgroups—importer and distributor).
In this regard, Bright, Greenhill and Levenkova (2014a) shows that, from different
simulation strategies (random attack, degree attack, weight attack and mixed strat-
egy), targeting hubs, especially in scale-free networks, is the most effective strat-
egy to disrupt networks, which can be supplemented by mixing node attributes.

Table 13.3 Centrality and intermediation measures of the network’s actors


Member Degree NBetweenness NEigenvector NCloseness
JPPM 119.000 8.798 90.072 3.310
SNRM 96.000 11.312 76.821 3.312
JDD 35.000 1.772 46.639 3.275
ELM 22.000 1.722 23.888 3.295
HPM 22.000 0 31.586 3.266
HAMS 24.000 1.684 26.358 3.293
WGV 22.000 0 29.939 3.266
CPGT 6.000 0.201 5.799 3.289
FR 12.000 0 14.911 3.266
PCCP 9.000 0 10.366 3.266
JG 3.000 0 4.083 3.266
ARG 8.000 0 11.995 3.268
YPMG 6.000 0. 452 3.420 3.275
JEAB 5.000 0 7.271 3.270
JB 5.000 0 5.398 3.287
LIRP 2.000 0 2.019 3.266
VR 2.000 0 2.722 3.266
OS 2.000 0 2.422 3.268
JHLM 2.000 0 1.385 3.268
JE 2.000 0 2.722 3.266
ERC 2.000 0 1.385 3.268
JO 1.000 0 1.361 3.266
FRC 1.000 0 0.826 3.230
JG 3.000 0 4.110 3.599
JV 0 0 0.027 1.887
RB 0 0 0.027 1.887
RAJH 0 0 0.027 1.887
196 A.G.-S. Framis and S.F. Regadera

Figure 13.2 shows these measures in a graph; the subject with the highest degree
of centrality (JPPM) is represented with a circle and the subject with the highest
degree of intermediation (SNRM) is represented with a triangle.

JV
RB OS
FRC
ARG
RAJH JEAB
JJTE
JDD
JG
LIRP
HAMS
FR
SNRM JPPM

PCCP

JB HPM
CPGT
ELM

ERC

JHLM
YPMG
JE
JO
VR
WGV

Wholesaler Importer Colombia Retailer Centrality Betweenness

Fig. 13.2 Degree of centrality and betweenness in the network

Individual Attributes

In order to fully understand covert networks, we should not only focus on the organ-
isational side of things; further information about the individual characteristics,
attributes and motivations of actors is also very important (Robins, 2009) and can be
used to suppress and disrupt networks. Schwartz and Rouselle (2009) recommend
using individual attributes and link weights to develop law enforcement intervention
and intelligence priorities. Bright, Greenhill and Levenkova (2014a) state that to
target and dismantle criminal networks more effectively, law enforcement should
consider node-level features, such as the roles played by individuals in the network
in addition to centrality scores.
The individual factors that contextualise the network are manifold, but the
following are the main ones:
– Socio-demographic characteristics such as sex, age, nationality and marital
status.
– Resources: Capacity, resources or skills (Natarajan, 2006; Bright & Delaney,
2013), possessing information, social capital (Schwartz & Rouselle, 2009),
experience as positive aspects and weaknesses, limitations or shortcomings as
negative aspects.
13 Static and Dynamic Approaches of a Drug Trafficking Network 197

– Psychological factors: Attitudes, beliefs, motivations as the cognitive driving


force behind the collective action (Robins, 2009) and psychological and psy-
chotic features attributable to specific actors or roles (Bovenkerk, 2000).
– Possessions, resources or forms of economic capital that can be exchanged.
– Individual behaviour that could be the answer to opportunities or limitations
offered by the structure of the network or organisation of which the individual is
part (Morselli, 2005).
To take one example from our cocaine trafficking network, we have chosen one
of the most common socio-demographic factors: sex. Figure 13.3 represents our
network as distributed by sex. Females represent 1 % of the members of the network
as there are only three women. The first of these women (CPGT) is part of the
importer group and is SNRM’s wife and a relative of YPMG, who is her nephew.
Her role in the group is very important because she manages a telephone booth
centre for the group from which money is sent out of Spain and her intermediation
role with the distributor group is clearly shown in Fig. 13.4. The other two women
(LIRP and PCCP) are part of the distributor group. The information available about
LIRP is scarce and all we know is that she held meetings with JPPM and ELM, but
PCCP is a known collaborator of JPPM, who lives with ELM and her role is to send
money to South America through telephone booth centres or money transfer service
companies. The role of women in organised crime has been traditionally underesti-
mated (Fiandaca, 2007) and SNA analysis could shed some light on the important
role of women and female spouses in illegal activities.

JV
RB OS
FRC
RAJH ARG
JEAB
JJTE JDD
JG LIRP
HAMS
FR
SNRM JPPM
PCCP
JB HPM
CPGT
ELM
ERC

JHLM
YPMG
JE
JO
VR
WGV

Men Women

Fig. 13.3 Network description by sex


198 A.G.-S. Framis and S.F. Regadera

Position in the Network and Individual Attributes

This chapter has analysed two types of characteristics which relate to the individu-
als that make up a network: their position in the network which depends on the
global volume of contacts and their personal attributes, as a means of supplementing
information about each subject. Taking a step further we can also combine both sets
of information to focus on the network’s inside knowledge. For instance, we may
want to know whether or not the most important actors in the network have legiti-
mate jobs (Morselli, 2009) apart from their involvement in organised crime, or if
they are mainly engaged in illicit activities or, on the contrary, if the main players
have a longer or shorter police record than other members of the network. We will
now illustrate this using our network and comparing the roles of the actors in the
network with the highest degree of centrality. This graph was published in another
article (Gimenez-Salinas, 2014) to prove that the traditional vision of roles or of
higher hierarchical positions does not always coincide with the position of the actors
in the network, taking into account the ties between them through the ARS. In our
network, represented in Fig. 13.4, there seems to be no difference between a hierar-
chical vision and that provided by SNA because the leaders, classed as such accord-
ing to the information in the investigation, have the highest degree of centrality in
the network. However, the article did reveal other organisations in which actors that
were coordinators, at a secondary hierarchical level, had a higher degree of central-
ity, even though their position of power with regard to their contacts was much
higher than that of the leaders.

Leaders Coordinators Operative roles


RB ARG
OS
JPPM

JDD

FR
JV
HAMS

JJTE
RS

PCCP
YPMG
HPM
RAJH SNRM
WGV
ELM

JG
VR A

EDRS
CPGT
JE
JEAB
JO GA

Fig. 13.4 Degree of centrality of the actors and roles in the organisation. Source: (Gimenez-
Salinas, 2014)
13 Static and Dynamic Approaches of a Drug Trafficking Network 199

Dynamic Analysis of the Networks

Thus far our analysis has been transverse, providing a static picture of our network
within a specific period of time. But this represents a fictitious reality as we are deal-
ing with live social systems, which are changing and dynamic over time. Criminal
organisations, understood as social networks, evolve over time, at least in three
ways (a) they are stable when they retain the same structural characteristics and the
same important players within a period of time; (b) they are changing when their
form and initial configuration evolve internally; and (c) they change and evolve as a
result of an exogenous factor that threatens their initial form and configuration (e.g.
police arrests) (Morselli & Petit, 2007).
In the field of criminal organisations, it is harder to discern the evolution of net-
works because it is difficult to access information in real time. However, it is essen-
tial to grasp the networks from their evolutionary dimension to find out how they
change over time and how they react to external changes such as police arrests,
changes to illicit markets or financial supervision of the their economic resources
(Bright & Delaney, 2013).
As we cannot always have access to information to know the temporal development
of the network, with a view to preventing and improving intelligence related to these
organisations, all efforts should focus on being able to predict and act early with the
least possible information. To this end, dynamic network analysis is useful as a way of
anticipating and identifying the weaknesses of important players whose removal from
the organisation will diminish its efficiency in the future. SNA allows us to carry out
this dynamic analysis even though there are other programmes that may be more effi-
cient in this respect, such as those linked to Dynamic Network Analysis: DNA (Carley,
2003) or Simulation Investigation for Empirical Network Analysis (SIENA).
We will carry out a simple analysis to examine the evolution of the cocaine traf-
ficking network, taking into account the data obtained in the two-year criminal
investigation. The static analysis carried out up until this point has given us some
insight into the network as a whole, but we have to wait until the end of the investi-
gation to reach conclusions about the network. So, how useful then is SNA as a
predictive or anticipation tool? If SNA only works to know more about the internal
workings of organisations once they have been dismantled, we cannot use it to
anticipate or devise early ways of taking them down and dismantling them. In other
words, we cannot use it as a preventive tool but rather just to give us more insight
into organisations. Therefore, the static perspective of the network needs to be
supplemented with a dynamic approach to appreciate how it changes over time,
detect the most relevant players in the network early to prioritise the efforts to iden-
tify and neutralise the members with the greatest impact on the network’s survival
and potential (Carley, 2003).
We will illustrate this form of dynamic analysis using the abovementioned
drug trafficking network. The information provided up until this point is the com-
plete set of information about the network after the investigation. We will now
200 A.G.-S. Framis and S.F. Regadera

divide the period of police investigation into four phases (times) and we will
analyse the network’s evolution with the information about contacts (wiretaps
and meetings) between actors that was available at the end of each phase of the
police investigation. This process will enable us to know which players were the
most relevant at each of the stages and check whether, before the investigation
ended, it was possible to identify the key (central) members of the network. First
of all, we will provide the network’s structure measures over the four phases of
the investigation and then we will analyse the key actors in terms of degree and
betweenness over time.
Table 13.4 provides general measures regarding the network in each of the
four phases of the investigation. Centralisation increases over time, as average
degree over the two-year period. On the contrary, density increases except at the
final stage.

Table 13.4 Descriptive data of network structure in the four phases of the investigation
Network structure Time 1 (n = 7) Time 2 (n = 16) Time 3 (n = 17) Time 4 (n = 27)
Centralisation 4.3 % 3.6 % 5.2 % 6.4 %
Average degree 0.885 (Nrm. 1.769 (Nrm. 3.370 7.849
0.248) 0.347) (Nrm. 0.318) (Nrm. 0.444)
Density 17 % 34 % 63 % 15 %

Figures 13.5, 13.6 and 13.7 show the different phases of the investigation. The
investigation lasted two years and we have divided the investigation into four phases
of six months each. To provide a better understanding of the evolution of each
actor’s centrality measures (degree and betweenness) during the investigation,
Tables 13.5 and 13.6 provide a summary.

First Phase of the Investigation

Figure 13.5 shows how the investigation started with the surveillance of the mem-
bers of the importer group from Colombia (highlighted in yellow). The first contact
with the wholesale group also occurred at this stage but on a secondary basis. This
group exposes, as tables number 5 and 6 show, JPPM (head of the importer group)
as the member with the highest degree of centrality and betweenness. This first
phase also exposes, although to a much lesser extent with respect to the rest of the
network, the only subject of the wholesale group that we later discover is the head
of the wholesale network in Spain (SNRM). Of the three most important leaders of
the groups, two of them have already appeared within six months of the start of the
investigation.
13 Static and Dynamic Approaches of a Drug Trafficking Network 201

ELM LIRP

JDD
JPPM

SNRM
HPM
JB

Wholesaler Importer Colombia

Fig. 13.5 Information about the network as at 6 months of the start of the investigation

Second Phase of the Investigation

Figure 13.6 below shows the contacts detected within the first year of the investiga-
tion. This chart already distinguishes the two different subgroups; the wholesale
subgroup in Spain is complete and the importer network has not been completely
identified. The players with the highest degree of centrality are JPPM (head of the
importer group), ELM (one of the main coordinators of the Importer group) and
SNRM (head of the wholesale group, who emerges with a lesser degree of centrality
and intermediation (see tables 13.5 and 13.6). A contact of the third subgroup also
appears at this stage, the retailer and, in reality, the leader of the group. For the sec-
ond time consecutively, the first contact of the group is in fact the leader.

JHLM YPMG ELM


ERC
CPGT
RB
JEAB
JJTE WGV JPPM

HAMS JB LIRP
SNRM

HPM
JDD

Wholesaler Importer Colombia Retailer

Fig. 13.6 Information about the network as at 12 months of the start of the investigation

Third Phase of the Investigation

Figure 13.7 shows the investigation after 18 months. The chart clearly shows the main
players and almost all the groups in full. According to the centrality and intermedia-
tion measures, the main players begin to appear at this stage (JPPM, head of the
wholesale group; SNRM, head of the importer group). JDD appears in third place,
when in the final graph he drops down to fifth place and carries out an operative-type
role in the network. Then come ELM and HPM, who do carry out coordination tasks.
202 A.G.-S. Framis and S.F. Regadera

JJTE

JB
LIRP

JEAB HAMS ERC


ARG

JHLM
SNRM
FRC JDD HPM
JPPM

OS
YPMG
WGV CPGT
ELM

Wholesaler Importer Colombia Retailer

Fig. 13.7 Information about the network as at 18 months of the start of the investigation

The fourth and last phase of the investigation exposes the network in full as
shown at the start of this chapter (Fig. 13.8).

JV OS FRC
RB JJTE
ARG
JEAB JDD
LIRP
RAJH
JG HAMS
FR
SNRM JPPM
PCCP
CPGT
HPM
ERC JB
YPMG ELM

VR JHLM
WGV JO JE

Wholesaler Importer Colombia Retailer

Fig. 13.8 Information about the network as at 24 months of the start of the investigation

A dynamic analysis of the position of the actors reveals that, throughout the
investigation, the key players in the final network appear with high centrality and
intermediation measures from the start. Tables 13.5 and 13.6 show how JPPM
comes first in terms of centrality form the start, followed by ELM who is his
right-hand man and group coordinator. 18 months after the start of the investigation,
SNRM comes first as head of the wholesalers group. However, due to his capacity
to intermediate between the importers group and the retailers, he comes first in
terms of intermediation after one year and remains in that position until the
investigation concludes.
Table 13.5 Degree distribution among members investigated during each of the four phases of the investigation
Phase 1 (6 months) Phase 2 (12 months) Phase 3 (18 months) Phase 4 (24 months)
Rank ID Deg. Rank ID Deg. Rank ID Deg. Rank ID Deg.
1 JPPM 16.000 1 JPPM 21.000 1 JPPM 59.000 1 JPPM 119.000
2 ELM 13.000 2 ELM 19.000 2 SNRM 27.000 2 SNRM 96.000
3 JDD 6.000 3 SNRM 12.000 3 JDD 27.000 3 JDD 35.000
4 JB 5.000 4 JDD 10.000 4 ELM 19.000 4 HAMS 24.000
5 SNRM 3.000 5 HAMS 8.000 5 HPM 11.000 5 WGV 22.000
6 LIRP 2.000 6 JB 5.000 6 HAMS 6.000 6 HPM 22.000
7 HPM 1.000 7 JJTE 3.000 7 JB 5.000 7 ELM 21.000
8 JEAB 2.000 8 ARG 4.000 8 FR 12.000
9 YPMG 2.000 9 JEAB 2.000 9 PCCP 9.000
10 JHLM 2.000 10 LIRP 2.000 10 ARG 8.000
11 RB 2.000 11 CPGT 2.000 11 JEAB 6.000
12 CPGT 2.000 12 YPMG 2.000 12 CPGT 6.000
13 ERC 2.000 13 WGV 2.000 13 YPMG 6.000
13 Static and Dynamic Approaches of a Drug Trafficking Network

14 HPM 1.000 14 ERC 2.000 14 JB 5.000


15 WGV 1.000 15 JHLM 2.000 15 JJTE 4.000
16 LIRP 0.000 16 OS 1.000 16 JG 4.000
17 FRC 1.000 17 OS 2.000
18 ERC 2.000
19 JHLM 2.000
20 LIRP 2.000
(continued)
203
Table 13.5 (continued)
204

Phase 1 (6 months) Phase 2 (12 months) Phase 3 (18 months) Phase 4 (24 months)
Rank ID Deg. Rank ID Deg. Rank ID Deg. Rank ID Deg.
21 VR 2.000
22 JE 2.000
23 RB 1.000
24 JO 1.000
25 JV 1.000
26 FRC 1.000
27 ERAJH 1.000
A.G.-S. Framis and S.F. Regadera
Table 13.6 Degree distribution among members investigated during each of the four phases of the investigation
Time 1 (n = 7) (6 months) Time 2 (n = 16) (12 months) Time 3 (n = 17) (18 months) Time 4 (n = 27) (24 months)
Rank ID Betw. Rank ID Betw. Rank ID Betw. Rank ID Betw.
1 JPPM 0.588 1 SNRM 5.569 1 SNRM 7.039 1 SNRM 12.186
2 ELM 0.196 2 JPPM 1.863 2 JPPM 4.372 2 JPPM 9.772
3 JB 1.765 3 JDD 2.286 3 HAMS 3.343
4 HAMS 1.020 4 ELM 0.907 4 JJTE 2.640
5 ELM 0.843 5 JB 0.163 5 JDD 1.885
6 ELM 1.722
7 YPMG 1.024
8 CPGT 0.201
13 Static and Dynamic Approaches of a Drug Trafficking Network

9 JG 0.107
205
206 A.G.-S. Framis and S.F. Regadera

Discussion and Conclusions

The example of the cocaine trafficking network analysed in this chapter has given
us an insight into some of the advantages of using SNA to find effective strategies
to dismantle and reduce group capacities. SNA is particularly suitable for flexible
and covert networks, where members do not have a fixed position or role and are
more difficult to arrest. SNA does not just offer a graphic representation but has
numerous other uses to improve the effectiveness of the arrests of members of crim-
inal networks and decrease their capacity over time.
A structural analysis of the group gives us an insight into the centralisation, den-
sity of contacts and whether the same network has different subgroups with a higher
density of contacts. This is extremely relevant to identify weaknesses and forms of
effective disruption. Our cocaine trafficking network clearly revealed the existence
of three subgroups which carried out importing, wholesale and retail activities. It is
debatable whether these groups are different but, given that cocaine distribution
constitutes a chain that goes from the countries of origin to its retail sale in the
destination countries, if we treat the small networks in a fragmented way, we are not
being entirely efficient; we need to see the network globally and identify its weak-
nesses or limitations to break the illicit market chain (Malm & Bichler, 2011). Thus,
by knowing the internal structure, the core-peripheral members in relational terms,
and which groups have a higher degree of contacts and the density of the subgroups,
we can more easily identify weaknesses to be targeted to dismantle the network. We
have seen that depending on the security priorities imposed on group suppression,
the network can be destabilised by first seeking out the main core or the broker
subgroup. In our case, the wholesaler subgroup is the powerful group because it
connects importers and retailers, so priorities are driven by these intermediary
groups (Gimenez-Salinas, 2014; Gimenez-Salinas et al., 2012; Natarajan, 2006). In
terms of key actors, JPPM and SNRM were the most central and broker actors, and
removing SNRM was essential due to his higher degree of betweenness.
SNA also enables an individual evaluation of the members of the network, which
provides further information about the network’s weaknesses. Knowing aspects
such as roles, economic, technical or material resources available, the personal
motivations or beliefs of the members etc. provides more qualitative information to
devise law enforcement strategies that are more efficient and specifically tailor-
made to the type of network at hand. We have used our case to show member distri-
bution by sex as an example of individual attributes. This distribution shows the
intermediation and connection roles that many women play in these criminal groups.
The role of women in these groups is gradually becoming more important, as
revealed by many investigations (Beare, 2010; Campbell, 2008; Caroll, 2002;
Fiandaca, 2007; Requena, De Juan, Gimenez-Salinas, & De La Corte, 2014), where
we can find women in management and coordination positions. Traditionally,
women have played a secondary role in terms of their visibility, but yet have an
increasingly significant role as efficient intermediaries and collaborators in tasks
13 Static and Dynamic Approaches of a Drug Trafficking Network 207

such as transport, money laundering or concealment activities. In this case, SNA


brought to light the important role of three women, which is something that is not
always visible at first sight. They play a key role as administrators of the group’s
funds and their connection with crime is explained by their personal relationship or
cohabitation with the main leaders.
At an individual level, cross-checking information about the individual attributes
and central roles in the network is useful as a means of triangulating information
and identifying information that does not appear at first sight. As proven by Bright,
Greenhill, Reynolds et al. (2014b), the mixture of centrality measures and actor
attributes (roles, etc.) provides powerful information and tools to dismantle and
reduce network efficiency in the future. In this regard, we have provided an example
adding individual attributes or roles to the centrality measures of each member.
Often, a traditional vision of the network leads to neutralising actors whose hierar-
chical position is more significant. However, by analysing their level of contacts and
connecting them to other personal attributes, we can identify greater capacities in
certain actors, and a greater accumulation of social capital (Von Lampe, 2009;
Bright et al., 2012; Schwartz & Rouselle, 2009) and resources on which law enforce-
ment actions can focus.
Networks are dynamic in their form, size and internal connections. Therefore, in
order to be useful from a law enforcement perspective, SNA should focus on detect-
ing vulnerable elements with scarce elements of information in order to anticipate
change and movements without having to wait for the investigation to conclude.
Dynamic approaches can also be useful to understand changes and the evolution of
networks over time and reactions to external or internal changes. In this regard,
simulation methods proposed by K.M. Carley provide promising results for net-
work disruption. The dynamic analysis provided in our example is just a descriptive
analysis to show broader possibilities of understanding network evolution. Analysing
the cocaine trafficking network in a dynamic way from the very start of the investi-
gation revealed two important things: firstly, that the main players who would later
become central actors and intermediaries in the network already appear as main
players at the start of the investigation (at least two of them). Secondly, the first
contacts of the different subgroups include, in two of the phases, the most important
persons or the leaders. Although these conclusions need to be supported with future
investigations, it does highlight this tool’s power of identification at the early stages
of investigations. The type of analysis carried out is straightforward. In fact, one the
most promising outlines on the road to predicting the development of networks is
provided by another more advanced and specific tool for this purpose: the DNA
(Carley, 2003) or SIENA7(Simulation Investigation for Empirical Network Analysis)
that use simulation models to make predictions and help intelligence agents decide
their priorities or disruption strategies.

7
Snijders et al. (2008).
208 A.G.-S. Framis and S.F. Regadera

Finally, despite the potential explanatory prowess of SNA, it has weaknesses,


especially when applied to covert networks. The first is the sources of information
available to investigate these types of networks. These networks remain hidden for
protection so it is very difficult to identify the type and number of contacts of the
members. As referred to in this chapter, wiretap data are one of the most used
sources during an investigation. Even if these data are valuable and reliable
(Berlusconi, 2013), they are also biased as regards the samples of contacts from
which global networks can be built. The data collected by the police will never be a
genuine full sample of contacts and their collection will always be influenced by the
investigation and its driving strategic interests. Missing data are very often a prob-
lem in these investigations, although they can be handled differently to uncover the
real structure of the network (Baber, Stanton, Howard, & Houghton, 2009;
Berlusconi, 2013; Marchette & Priebe, 2008). In our opinion this limitation affects
any criminal investigation. This area only studies the part of reality that comes to
light in the police investigations. This is a limitation for the results of this study and
especially for the example of the dynamic network. It is worth noting, in line with
Sparrow (1991), that network structure and the centrality measures of the actors can
easily be more a reflection of the focus of the law enforcement investigation than of
structure position in the network. As Bright & Delaney, (2013) suggests, there is a
risk that evolution in the network may be artificial and thus so would be the resulting
intelligence collected. Nevertheless, some centrality measures are quite robust
under conditions of missing nodes and edges (Berlusconi, 2013; Borgatti, Carley, &
Krackhardt, 2006) and wiretap data collected can be supplemented with other
sources of information (reports, surveillance, etc.), which ultimately reduces the
probability of the network being influenced by the focus of the investigation (Bright
& Delaney, 2013).
Despite these limitations and others resulting from the size of the sample used, the
results highlighted the potential of SNA as a predictive and investigation tool to help
investigators devise more effective strategies to dismantle networks. In this regard,
the implications in practice are the following: (a) SNA is a useful method of intelli-
gence and investigation to better understand the internal workings of networks as
well as their limitations and weaknesses, all of which can be used to reduce their
efficiency. (b) SNA is a tool that reveals the central actors of each network, in refer-
ence to their contacts which, in addition to the information regarding their attributes
and resource capacity, constitutes very valuable information to reduce the group’s
potential in the long term and help make decisions about the main objectives to
defuse. (c) Finally, dynamic analysis of networks should be explored further as a way
of predicting and anticipating the evolution of networks. Further investigation is
required to understand the evolution of networks in time, the effects of external and
internal changes and how to reduce the long-term capacity of networks at an early
stage. This would give investigators and intelligence agents better prediction tools.

Acknowledgements We would like to thank the judicial police unit of the Spanish Guardia Civil
for giving us access to data. We are also very grateful to all those who reviewed this chapter for
their comments and valuable suggestions.
13 Static and Dynamic Approaches of a Drug Trafficking Network 209

References

Adeler, P. A. (1993). Wheeling and dealing: An ethnography of an upper-level drug dealing and
smuggling community (2nd ed.). New York: Columbia University Press.
Albert, R., Jeong, H., & Barabasi, A. L. (1999). Diameter of the world-wide web. Nature, 401,
130–131.
Baber, C., Stanton, N., Howard, D. & Houghton, R. J. (2009). Predicting the structure of covert
networks using genetic programming, cognitive work analysis and social network analysis.
Report NATO.
Barnes, J. A. (1972). Social networks. Reading: Addison-Wesley.
Beare, M. (2010). Women and Organize Crime. Report n. 13, Research and National Coordination.
Organized Crime Division.
Benson, J. S., & Decker, S. H. (2010). The organizational structure of international drug smug-
gling. Journal of Criminal Justice, 38, 120–138.
Berlusconi, G. (2013). Do all the pieces matter? Assessing the reliability of law enforcement data
sources for the network analysis of wire taps. Global Crime, 14(1), 61–81.
Borgatti, S. P. (2006). Identifying sets of key players in a social network. Computational &
Mathematical Organization Theory, 12(1), 21–34.
Borgatti, S. P., Carley, M., & Krackhardt, D. (2006). On the robustness of centrality measures
under conditions of imperfect data. Social Networks, 28, 124–136.
Borgatti, S. P., & Evertt, M. G. (1999). Models of core/periphery structures. Social Networks, 21,
375–395.
Borgatti, S. P., Evertt, M. G., & Freeeman, L. C. (2002). Ucinet for windows:Software for social
network analysis. Harvard: Analytic Technologies.
Bovenkerk, F. (2000). Wanted: ‘Mafia boss’-Essay on the personology of organized crime. Crime
Law and Social Change, 33(3), 225–242.
Bright, D. A., & Delaney, J. J. (2013). Evolution of a drug trafficking network: Mapping changes
in network structure and functions across time. Global Crime, 14(2–3), 238–260.
Bright, D. A., Hughes, C. E., & Chalmers, J. (2012). Iluminating dark networks a social network
analysis of an Australian drug traffickin syndicate. Crime, Law and Social Change, 57,
151–176.
Bright, D. A., Greenhill, C., & Levenkova, N. (2014a). Dismantling criminal networks. Can attri-
butes play a role? In C. Morselli (Ed.), Crime and networks (pp. 148–162). New York, NY:
Routledge.
Bright, D. A., Greenhill, D., Reynolds, M, Ritter, A., Morselli, C. (2014b). The use of actor-level
attributes and centrality measures to identify key actors: A case study of an australian drug
trafficking network. Journal of Contemporary Criminal Justice. doi:10.1177/1043986214553378.
Bruinsma, G., & Bernasco, W. (2004). Criminal groups and transnational illegal markets: A more
detailed examination on the basis of social network theory. Crime, Law and Social Change, 41,
79–94.
Calderoni, F. (2012). The structure of drug trafficking mafias: The ‘Ndrangheta and cocaine.
Crime, Law and Social Change, 58(3), 321–349.
Campbell, H. (2008). Female drug smugglers on the US-Mexico border: Sex, crime and empower-
ment. Antroplogical Quarterly, 81(1), 233–267.
Carley, K. M. (2003). Dynamic network analysis. In R. Breiger & K. M. Carley (Eds.), Summary
of the NRC workshop on social network modeling and analysis. Washington: National Research
Council.
Carley, K. M., Lee, J., & Krackhardt, D. (2003). Destabilizing networks. Connections, 24, 79–92.
Caroll, R. (2002, May 28). The rise of Godmothers. The Guardian. http://www.theguardian.com/
world/2002/may/28/gender.uk
De La Corte, L., & Giménez-Salinas Framis, A. (2010). Crimen.org. Evolución y claves de la
delincuencia organizada. Barcelona: Ariel.
210 A.G.-S. Framis and S.F. Regadera

Desroches, F. (1999) Wholesale drug dealers, paper presented to panel on ‘The structure and oper-
ation of illegal commodity markets’. Annual Meeting of the American Society of Criminology.
Toronto.
Fiandaca, G. (2007). Women and the mafia: Female roles in organized crime structure. Palermo:
Springer Publication.
Gimenez-Salinas, A. (2014). Illegal networks or Criminal Organisations: Structure, power and
facilitators in cocaine trafficking structures. In C. Morselli (Ed.), Crime and networks (pp. 131–
148). New York, NY: Routledge.
Gimenez-Salinas, A. Roman, H., Nieto, L, Martinez, A. J. & Fernandez, S. (2012). Estructuras y
liderazgo en cuatro redes españolas dedicadas al tráfico de drogas. Revista Española de
Investigación Criminológica, 10.
Kleemans, E. R., & Van De Bunt, E. G. (2003). The social organization of human trafficking. In
D. Siegel, H. G. Van de Bunt, & D. Zaitch (Eds.), Global organized crime. Trends and develop-
ments (pp. 97–104). Dordrecht: Kluwer.
Krebs, V. E. (2001). The network paradigm applied to criminal organizations. Connections, 24,
53–65.
Krebs, V. E. (2002). Mapping networks of terrorist cells. Connections, 24(3), 43–52.
Malm, A., & Bichler, G. (2011). Networks of collaborating criminals: Assessing the structural vul-
nerability of drug markets. Journal of Research in Crime and Delinquency, 48(2), 271–297.
Marchette, D. J., & Priebe, C. E. (2008). Predicting unobserved links in incompletely observed
networks. Computational Statistics & Data analysis, 52, 1373–1386.
McGloin, J. M. (2005). Policy and intervention considerations of a network analysis of street
gangs. Criminology and Public Policy, 4(3), 607–636.
Moreno, J. L. (1953). Who shall survive? Foundations of sociometry, groups psychotherapy and
sociodrama. New York, NY: Beacon House Ltd.
Morselli, C. (2005). Contacts, opportunities and criminal enterprise. Toronto, ON: Presses de
l’Université de Toronto.
Morselli, C. (2009). Inside criminal networks. New York, NY: Springer.
Morselli, C. (2010). Assessing vulnerable and strategic positions in a criminal network. Journal of
Contemporary Criminal Justice, 26(4), 382–392.
Morselli, C., & Giguère, C. (2006). Legitimate strengths in criminal networks. Crime, Law and
Social Change, 45(3), 185–200.
Morselli, C., & Petit, K. (2007). Law-enforcement disruption of a drug importation network.
Global Crime, 8(2), 109–130.
Morselli, C., & Roy, J. (2008). Brokerage qualifications in ringing operations. Criminology, 46(1),
71–98.
Natarajan, M. (2000). Understanding the structure of a drug trafficking organization: A conversa-
tional analysis. Crime Prevention Studies, 11, 273–298.
Natarajan, M. (2006). Understanding the structure of a large heroin distribution network: A quan-
titative analysis of qualitative data. Journal of Quantitative Criminology, 22, 171–192.
Natarajan, M., & Belanger, M. (1998). Varieties of upper-level drug dealing organizations: A
typology of cases prosecuted in New York City. Journal of Drug Issues, 28(4), 1005–1026.
Requena, L., De Juan, M., Gimenez-Salinas, A., & De La Corte, L. (2014). A psychosocial Study
on crime and gender: Position, Role and Status of women in a sample of Spanish Criminal
Organizations. International Journal of Social Psychology, 29(1), 121–149.
Reuter, P. H., & Haaga, J. (1989). The organization of high level drug markets: An exploratory
study. Santa Monica, CA: Rand.
Robins, G. (2009). Understanding individual behaviours within covert networks: The interplay of
individual qualities, psychological predispositions, and network effects. Trends in Organized
Crime, 12, 166–187.
Ronfeldt, D., & Arquilla, J. (2001). Networks, netwars and the fight for the future. First Monday,
6, 10.
13 Static and Dynamic Approaches of a Drug Trafficking Network 211

Schwartz, D. M., & Rouselle, T. (2009). Using social network analysis to target criminal networks.
Trends in Organized Crime, 12, 188–207.
Snijders, T. A. B., Steglich, C. E. G., Schweinberger, M., & Huisman, M., (2008). Manual of
SIENA version 3.2. ICS. University of Groningen. Groningen: Department of Statistics,
University of Oxford.
Sparrow, M. K. (1991). The application of network analysis to criminal intelligence. An assess-
ment of the prospects. Social Networks, 13, 251–274.
Van der Hulst, R. C. (2009). Introduction to social network analysis as an investigative tool. Trends
in Organized Crime, 12, 101–121.
Varese, F. (2013). The structure and the content of criminal connections: The Russian Mafia in
Italy. European Sociological Review, 29(5), 899–909.
Von Lampe, K. (2009). Human capital and social capital in criminal networks: Introduction to the
special issue on the 7th Blankensee Colloquium. Trends in Organized Crime, 12(2), 93–100.
Williams, P. (2001). Transnational criminal networks. In J. Arquilla & D. Ronfeldt (Eds.), Networks
and netwars: The future of terror, crime and militancy (pp. 61–97). Santa Monica, CA: RAND.
Xu, J., & Chen, J. (2005). Criminal network analysis and visualization. Communication of the
ACM, 48(6), 101–107.
Xu, J., & Chen, H. (2008). The topology of dark networks. Communications of the ACM, 51,
58–65.
Zaitch, D. (2002). Trafficking cocaine: Colombian drug entrepreneurs in the Netherlands. The
Hague, Netherlands: Kluwer.
Part V
Agent-Based Modelling
Chapter 14
Formal Models of the Crime Event:
Agent-Based Modelling in Support
of Crime Prevention

Daniel Birks

Introduction

Designing, implementing and evaluating crime prevention interventions is a


necessary but resource intensive process. Such efforts typically involve collating
data describing a particular crime problem, analysing those data, designing and
implementing strategies which aim to address identified problems, and carrying out
appropriate evaluation of the impacts of implemented strategies (Clarke, 1997).
Moreover, to be confident about the effectiveness of a particular intervention, repli-
cation is required to establish the consistency of observed effects, and explore how
context impacts on intervention outcomes (Shadish, Cook, & Campbell, 2002). In
this chapter I discuss how agent-based models (ABM1) may support initial efforts in
this area, by providing computational laboratories free from logistical, ethical and
financial constraints in which formal models of crime theory and the mechanisms of
causation they posit can be tested, refined and potential crime prevention interven-
tions that draw upon them prototyped.

Agent-Based Models

The ABM is a type of computational model devised to aid in the analysis of com-
plex systems comprising autonomous heterogeneous entities that act and interact in
rich and non-linear ways (Bonabeau, 2002). Many real world systems are complex,

1
For the sake of brevity the acronym ABM is used to refer to both agent-based model(s) and agent-
based modelling.
D. Birks (*)
Griffith Criminology Institute, Griffith University,
176 Messines Ridge Road, Mt Gravatt, Brisbane, QLD, Australia
e-mail: d.birks@griffith.edu.au

© Springer International Publishing Switzerland 2017 215


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_14
216 D. Birks

but one of the most studied is human society. Harnessing significant advances in
computational capacity, the last three decades have seen a range of social sciences
explore and capitalise on the unique strengths of ABMs to better understand how
the action and interaction of individual members of the population translate into
observed societal outcomes (Hedström, 2005). This approach has been demon-
strated to be particularly useful where the intricacies of cause and effect exhibited
by complex social systems are difficult to formalise through written theory, lack
reliable individual level data to describe them, and are bereft of elegant analytical
solutions capable of estimating system behaviour at the individual level (Epstein,
1999).
In order to describe the agent-based analytical approach to human systems it is
useful to contrast it against more traditional solutions often applied within the social
sciences.2 Analysis of societal phenomena often proceeds via a top-down approach
(Epstein, 1999). This typically involves the observation of relationships between
aggregate level system properties (e.g. area level affluence and educational attain-
ment; residential mobility and crime rates), some statistical measure of fit between
these properties, and subsequent generalised inference about lower order interac-
tions between individual entities. An important limitation of such top-down
approaches is that they often struggle to adequately quantify the mechanisms by
which observed system properties are related. This can be particularly problematic
where an understanding of these mechanisms is desirable to facilitate some form of
intervention that aims to produce or disrupt particular outcomes.
The ABM seeks to provide a complementary approach to study societal phenom-
ena that operates from the bottom-up; exploring how individual action and interac-
tion influence aggregate system properties (Epstein & Axtell, 1996). ABMs are
used to simulate artificial worlds that mimic social systems at a manageable level of
complexity, providing an appropriate analogue to the system of interest, which is
easier to both manipulate and study than the system itself (Gilbert & Troitzsch,
2005). These worlds are populated with simulated decision makers commonly
referred to as agents. Agents are designed to perceive, reason and act based on cur-
rent theory or empirical evidence from the field of interest. By simulating these
actors, and their repeated actions and interactions with both one another and the
environment they inhabit, researchers gain insight into the aggregate properties that
result from particular individual level configurations. Thus, while the statistical
model measures fit between aggregate system properties, ABMs offer formalised,
abstracted models through which the emergence of system relations from individual
level action can be systematically explored (Epstein & Axtell, 1996; Hedström &
Swedberg, 1998).
ABMs can be used to support a broad range of significant research efforts
(Epstein, 2008). While acknowledging this diversity, here we discuss two distinct
but related applications of ABMs that may support crime prevention in important

2
This contrast simply serves description but should not imply competition between approaches.
Any social scientist worth his or her salt can clearly see the advantage of triangulating multiple
methods in the pursuit of better understanding the complexity of society.
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 217

ways; (1) in the development of models that explore the explanatory capacity of
mechanisms proposed by criminological theory; and (2) in predicting the likely
outcomes of manipulations to such mechanisms through intervention. Both these
approaches are now briefly summarised.

Explanatory Agent-Based Models

Explanatory ABMs seek to understand under what circumstances particular sys-


tem outcomes are likely to occur. Within the social sciences these models offer
means to assess the capacity of theories describing individual action in explaining
observed societal phenomena. In pursuing this goal, artificial environments are
created and populated with agents designed to reflect current theory in the field of
interest. An agent is a formal computational model of an autonomous decision-
making entity. Within the social sciences agents typically represent human deci-
sion makers, but may also be designed to simulate other autonomous entities such
as social groups, organisations or nations. Agents are bestowed with a series of
defining variables and behaviours. Variables allow agents to quantify their own
static characteristics (e.g. home location, ethnicity, and preference for mode of
transport) and dynamic state (e.g. wealth, memory of previous action, current
goal, and social network). Behaviours allow agents to modify that internal state
and the world around them. In explanatory ABMs these behaviours are typically
designed to formalise current theory, representing how we hypothesise individu-
als perceive, reason, and act in particular circumstances. In implementation, these
behaviours may take a range of forms, from simple condition-action rules to more
complex representations of cognitive functioning, such as the commonly used
belief–desire–intention model (Rao & Georgeff, 1995).
In an ABM agents act and interact with one another within some environment.
This environment can take a number of forms and play varying roles in a given
simulation. For instance, it might facilitate the actions of agents, defining locations
for them to congregate, paths for them to travel on, or providing resources for them
to gather. While model environments such as these may be spatially explicit they
may also be used to represent abstract environmental constructs, such that agents
interact within some social space where proximity relates to the convergence of peer
groups, social norms etc. Most importantly, model environments allow ABMs to
simulate situated and typically localised actions and interactions.
Once constructed, ABM can be used to carry out simulations in which agents
repeatedly act and interact based upon their behaviours and current situation. A
simulation is a specific instantiation of a particular ABM, which encompasses
configurations of both agents and the environment. As a simulation progresses,
researchers observe system properties that result from the action and interaction
of agents in this particular system configuration, comparing emergent properties
to what is expected by theory or, where data is available, to empirical findings. By
undertaking these comparisons explanatory ABMs allow researchers to test the
218 D. Birks

internal consistency of existing theories and moreover assess whether hypothe-


sised individual level behaviours are sufficient to explain known attributes of a
social system. This approach, labelled by some as a generative approach to social
science (Epstein, 1999), sees the ABM as a scientific instrument capable of testing
the sufficiency of hypotheses that propose potential explanations of observed
social phenomena.
In this way, the explanatory ABM acts as a formalised thought experiment, capa-
ble of demonstrating the aggregate consequences of proposed individual level
mechanisms. Where our understanding of individual action and interaction is lim-
ited, and/or there are difficulties in testing rival theoretical accounts of individual
action, this approach provides the means to explore propositions of theory in an
environment that is free from traditional logistical, ethical and monetary constraints
(Birks, Donkin, & Wellsmith, 2008; Eck & Liu, 2008a).
Notably, the processes involved in constructing ABMs can also be particu-
larly useful in supporting theorising about the system of interest. The implemen-
tation of an ABM requires theoreticians to be explicit about the mechanisms by
which they propose observable phenomena come about. This is a direct conse-
quence of the need to translate hypothetical mechanisms into formal models rep-
resented as algorithms and data structures that can be bestowed upon agents
within a given model. In many ways analogous to crime scripting (Brayley,
Cockbain, & Laycock, 2011), this process of theory formalisation often high-
lights logical inconsistencies in existing theories that would otherwise be over-
looked, in turn requiring theoreticians to refine their hypotheses concerning
mechanisms of interest.

Predictive Agent-Based Models

A further application of ABMs relates to their ability to estimate the impacts of


changes to a particular modelled system. Given a sufficiently validated model of
individual behaviour, ABMs provide a platform to systematically assess what sys-
tem properties are likely to result from novel system configurations (Gilbert &
Troitzsch, 2005). Thus, having constructed explanatory models that produce plau-
sible outcomes, we can develop subsequent “what-if” type scenarios that aim to
assess the likely system impacts resulting from changes to behavioural or environ-
mental conditions (Elffers & Gerritsen, 2016). To illustrate, one might devise a
model of traffic flow derived from individual accounts of driver behaviour and sub-
sequently examine the impact of several proposed changes to the street network
designed to alleviate congestion. Such prospective scenarios are particularly impor-
tant in studying those social systems where high levels of interdependence dictate
that seemingly small changes to individual level system components can have
diverse, significant and often unintended effects on aggregate societal outcomes
(tipping points etc.).
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 219

Strengths of Agent-Based Approaches

While explanatory and predictive ABMs are typically applied to address largely
distinct problems, they are both supported by a range of advantages that the ABM
confers on the analysis of complex social systems. These are now briefly
discussed.
First, ABM avoid assumptions about population homogeneity (Epstein & Axtell,
1996). Within an ABM each agent acts autonomously and can be created with its
own unique characteristics, preferences and behaviours. Thus, diverse populations
of decision makers can co-exist and be studied within a single model. This approach
is considerably more flexible than many traditional statistical or mathematical mod-
els where, for the sake of mathematical tractability, the characteristics of individuals
and their interaction are often assumed to remain constant across populations.
Second, as agents are situated in an explicit environment, ABMs permit the
explicit simulation of localised interactions. Agent decision-calculi can be devised
to incorporate both individual- and situation-specific information into decision-
making processes. This supports the simulation of bounded rationality, such that
agents can be devised to draw only on restricted information available at the time a
decision is made. This approach to modelling decision-making avoids unrealistic
assumptions often embedded in mathematical or statistical models that bestow
actors with access to complete information about the world in which they operate.
Third, ABMs are well suited to study changes in the longitudinal dynamics of
social systems (Gilbert & Troitzsch, 2005). ABMs simulate the progression of time
via discrete increments, often referred to as cycles. During each cycle agents act and
interact based upon their specified behaviours, which in turn, draw on agent’s local
circumstances and individual characteristics. Many thousands of these cycles may
occur as a simulation progresses. Thus, ABM is performed in a recursive fashion,
permitting the longitudinal examination of interconnected and interdependent sys-
tem elements over time. Such temporal dynamics are especially important for the
modelling non-equilibrium systems and phenomena such as tipping points, where
the accumulation of individual action over time can lead to rapid and significant
diversions in system behaviour (Grodzins, 1958).
Fourth, ABMs provide an intuitive platform for demonstrating complex concepts
to a broad audience. Models are specified through the conceptualization and for-
malisation of individual level behaviours. Specifying model constructs at this level
is considerably more accessible than complex mathematical abstractions (Bonabeau,
2002). This dictates that audiences need not be highly skilled in the field of ABM in
order to inform, interrogate and interpret models. And as a result, simulations can
be more easily supported by domain experts who often ask more pertinent questions
of models than those who are predominantly interested in developing them (Gilbert
& Troitzsch, 2005).
Fifth, ABMs offer researchers unrestricted powers of both observation and manip-
ulation. Within a given simulation all variables, at all scales of aggregation, can be
observed and analysed by the researcher. Moreover, researchers can systematically
220 D. Birks

manipulate all model configurations, for instance altering the decision-making rules
of an agent, or the environment in which such decisions are made. As a result of this
capacity ABMs permit a wealth of systematic simulated experiments to be carried
out3 free from constraints imposed on empirical experimentation. These properties
facilitate the creation of computational laboratories where researchers can construct
and carry out simulation experiments, holding a system configuration static while
systematically manipulating a single construct of interest (Epstein & Axtell, 1996;
Macy & Willer, 2002).
Given the minimal requirements of such experiments, thousands of simulation
experiments can be run at little cost, exploring the same mechanisms in a range of
circumstances, and at different spatial and temporal scales. In this way the robust-
ness of theoretical proposition can be explored. To illustrate, one might develop an
ABM that explored several distinct models of criminal decision-making, compar-
ing simulated crime generated by these models to observed distributions of crime,
and in turn, assessing which model best reproduces plausible outcomes. Given
such a behaviour, further experiments could be performed situating the same
mechanisms in varying environments, of varying sizes and configurations, and
over short and long time scales, further assessing generalisability across a range of
contexts.

Agent-Based Modelling and Computational Criminology

Since the mid-2000s there has been burgeoning interest in the application of
ABM within the field of criminology and in particular environmental criminol-
ogy’s study of crime events and patterns (see Brantingham & Brantingham, 2004
for an important initial effort in this area). This interest fundamentally stems
from the way in which environmental criminologists conceptualise crime as the
outcome of numerous spatio-temporally situated interactions between boundedly
rational potential offenders, victims and guardians, and the environment they
inhabit (Birks, Townsley, & Stewart, 2012). This depiction sees crime events and
patterns as the emergent outcomes of a complex system (Birks, Townsley, &
Stewart, 2014). It is under this conceptualisation that scholars have recently
applied agent-based methods to model and better understand crime event interac-
tions and their likely influences on observed crime phenomena, addressing a
variety of research problems.

3
Note that ABM can be constructed to be either deterministic or probabilistic. In probabilistic
ABM the results of a single simulation offer restricted insight into the system of interest. Multiple
replications of the same probabilistic models are typically run to assess the range of outcomes that
result from a particular model configuration. Also note that the random number generators used in
ABM can be seeded (this process generates the same trace of random numbers when a simulation
is instantiated) in order to facilitate non-deterministic but reproducible model outcomes.
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 221

Mirroring previously discussed explanatory and predictive applications,


agent-based studies within this field can be thought to support crime prevention
efforts (directly or indirectly) in at least two distinct ways:
1. Through the development of explanatory models that seek to aid in the specifica-
tion, testing and refinement of hypothesised mechanisms of crime that underlie
proposed crime prevention interventions;
2. Through the development of predictive models that estimate the likely impacts
of potential interventions designed to manipulate such mechanisms.
Elffers and Gerritsen (2016) focus on the second of these applications; providing
several examples of the simulation of potential crime prevention interventions and
in particular exploring the likely influence of varying configurations of guardian-
ship (e.g. Bosse, Elffers, & Gerritsen, 2010; Bosse & Gerritsen, 2010; Melo,
Belchior, & Furtado, 2006). In what follows, I concentrate on the first application—
the development of explanatory ABMs for theoretical testing, refinement and suc-
cession. In doing so, I argue that the ABM can generate foundational knowledge
about the likely mechanisms of crime, which subsequently can be used to inform
the design of interventions that aim to manipulate these mechanisms. Drawing on
applications within the field of environmental criminology, I concentrate on explan-
atory models aimed at understanding several theoretical accounts of crime that are
difficult to test through traditional experimentation to the level that would be desir-
able to best inform situational crime prevention.

Using Explanatory Agent-Based Models to Prototype


Criminological Theory in Support of Situational Crime
Prevention Intervention

A primary requirement of effective crime prevention is an appropriate under-


standing of the particular crime problem studied. If interventions are designed to
influence particular mechanisms of crime, they rely not only on the presence of
these mechanisms, but also a sufficient understanding of them to facilitate their
manipulation. Thus, the evaluation of existing crime prevention interventions
provides a test of crime event theories, capable of providing valuable feedback
that can be used to correct and refine not only policy and practice but also crimi-
nological theory (Tilley, 2002). Yet, in order that such efforts increase our the
understanding of crime event mechanisms, interventions rely on the constellation
of factors that include well-specified theory, its effective translation into interven-
tion, and high competence of intervention implementation and evaluation design
(Ekblom & Pease, 1995). The debate over how evaluations are best conducted in
order to maximise their overall utility for intervention and theoretical refinement
is ongoing. Sherman et al. (1998), for example, describe high rates of variability
in methodological quality observed amongst crime prevention evaluation studies.
Similarly, in discussing the general quality of crime prevention evaluation,
222 D. Birks

Ekblom and Pease (1995) suggest that standards are generally low. The most com-
mon types of evaluation being measures of crime pre- and post-intervention or
interrupted time series, in some cases with the use of a control area to account for
confounding effects.
Drawing on a range of theories from the field of environmental criminology,
situational approaches to crime prevention seek to reduce crime by reducing the
opportunities for crime (Clarke, 1980). Unlike much of traditional criminological
study, these approaches are predominantly concerned with crime events, suggesting
that the proximal mechanisms that influence observed crime events and patterns are
considerably more easy to both identify and manipulate than those distal causes that
influence criminal involvement over the life course (Cornish & Clarke, 1986).
Situational crime prevention calls for prevention strategies that are crime and prob-
lem specific (Clarke, 1995). Thus, optimal intervention design relies on the identifi-
cation and quantification of these proximal mechanisms involved in generating a
particular crime problem.
Currently, three core theoretical approaches often referred to as the opportunity
theories (Wortley & Mazerolle, 2008) underpin this study of crime events and pat-
terns; the routine activity approach (Cohen & Felson, 1979; Felson, 2008), rational
choice perspective (Clarke, 1980; Cornish & Clarke, 1986, 2008) and crime pattern
theory (Brantingham & Brantingham, 1978, 1981, 1993 2008). Collectively, these
theories provide depictions of the spatio-temporally situated processes by which
offenders find suitable targets in their environment, reason about them, and the role
that the dynamic environmental backcloth plays in influencing such interactions to
shape patterns of crime.
Diverse crime prevention efforts draw on the opportunity theories in the design
of interventions seeking to reduce victimisation. As discussed, the primary aim of
situational crime prevention is to remove or disrupt opportunities for crime by
manipulating the offender expected utility calculus set-out by the rational choice
perspective; increasing offender perceptions of risk and effort while minimising
perceptions of reward (Clarke, 1980). The field of Crime Prevention Through
Environmental Design similarly seeks to devise environmental morphologies that
influence and disrupt offender crime event decision-making with the hope of reduc-
ing crime (Jeffrey & Zahm, 1993). While such approaches have been aptly demon-
strated to be effective in reducing crime (Clarke, 1997), SCP interventions can still
result in divergent outcomes (Guerette & Bowers, 2009). Consequently, there is still
considerable work to be done in identifying those strategies most likely to be effec-
tive in varying contexts and in identifying new strategies aimed at reducing crime in
existing contexts.
To date, progress toward this goal has been encumbered by two key obstacles;
(1) a relative scarcity of reliable and representative micro-level data concerning
crime events, and (2) fundamental difficulties associated with undertaking con-
trolled experiments in the study of crime events. While the nature of these problems
are in no way unique to the study of criminological phenomena, their presence
ultimately restricts the testing, refinement and succession of opportunity based
accounts of crime.
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 223

To illustrate, both the routine activity approach and crime pattern theory describe
how spatio-temporal activity patterns of potential victims, offenders and guardians
influence the observed distribution of crime events. Our ability to collect representa-
tive data at the individual level proposed by these theories (i.e. the activities of
individuals) to provide an optimal test of their hypotheses is extremely limited.
Similarly, it is logistically infeasible to examine the internal mechanisms of offender
decision-making in situ and en masse, in pursuit of thorough tests of the rational
choice perspective. Furthermore, our ability to assess the influence of the ever-
changing environmental backcloth on both offenders and offending opportunities is
severely restricted by a lack of accurate micro level data, and our inability to sys-
tematically manipulate environmental contexts from which inferences concerning
the influence of certain environmental features might be drawn. As a result of these
problems, evaluations of individual level interventions are usually defined at the
area level, and from which data are unsuitable for testing the individual level mech-
anisms proposed by theory (Eck, 1995).
Further confounding these difficulties, theory depicts such mechanisms embed-
ded within or alongside one another. This interconnectedness dictates that even with
substantial insight into the offence process it is difficult to estimate the impacts of
individual mechanisms in order to best identify causal paths for optimal crime pre-
vention. Furthermore, such interconnectedness dictates that if a proposed mecha-
nism is erroneous or not adequately understood it is almost impossible to eliminate
as a potential explanation without quantifying the interactions that are occurring in
situ (Birks et al., 2012). Hence, even if the crime event is observed in all of its com-
plexity, and indeed those mechanisms proposed by theory are an accurate reflection
of reality, it is still difficult to disentangle the respective effects of each. As a result
of these problems a causal gap remains between observed intervention outcomes,
crime patterns, and proposed individual-level behaviours that interventions aim to
manipulate. This divide dictates that it is often difficult to accurately test theoretical
depictions of crime to the extent that would be desirable for tailored crime preven-
tion development (Birks et al., 2012).
In response to these difficulties, one potentially productive application of ABMs
is to provide a constraint free environment in which the explanatory capacity of
proposed crime event mechanisms can be systematically assessed (Birks & Elffers,
2014). In doing so, ABMs provides an instrument capable of exploring the afore-
mentioned divide between proposed crime event mechanisms and observed crime
events, in turn, supporting rigorous assessment of theories that inform crime pre-
vention intervention.
In pursuit of this goal, ABMs can be built to simulate populations of potential
offenders and victims whose behaviours are derived from mechanisms proposed by
the opportunity theories. For example, agents can be bestowed with the expected
utility calculus proposed by the rational choice perspective, or the spatially refer-
enced awareness space set out by crime pattern theory. Such models of offending
can then be validated by comparing crime generated by actors who behave accord-
ing to theory to a range of aggregate properties consistently observed in the empiri-
cal study of crime events. These include the observation that crime is both spatially
224 D. Birks

and temporally concentrated; that a small number of victims are subject to a


disproportionate amount of victimisation; and that a small number of offenders are
responsible for a disproportionate amount of offences. Such regularities of crime
represent the trace effects of whatever mechanisms are indeed operating in situ and
thus provide system properties against which simulation outcomes should be com-
pared for the purposes theory testing (Birks et al., 2012).
Where mechanisms proposed by theory are not sufficient to generate commonly
observed patterns of crime, they can be eliminated as a potential explanation, or
incrementally refined and tested again through further simulation studies. In this
way models identify those proposed mechanisms that are most likely to reflect real
world crime event mechanisms, and thus offer viable candidates for manipulation
when incorporated into the design of crime prevention interventions. In recent years
a number of studies have taken this approach with the aim of better understanding
the proximal mechanisms of crime and how they influence observed spatial, tempo-
ral and functional distributions of crime events and patterns. Several illustrative
examples are now described.
Groff (2007a, 2007b, 2008) presents a series of simulation studies utilising a
computational laboratory devised to assess the veracity and explanatory capacity of
key propositions of the routine activity approach. Groff (2007a) presents an initial
model in which agents representing police and potential offenders, victims and
guardians move randomly between intersections in an environment derived from
GIS data describing the road network of Seattle, Washington. Using this model a
number of experiments are performed where the time agents spend away from home
is systematically manipulated and the resulting levels of street robbery observed.
Results of these experiments are congruent with those predicted by the routine
activity approach—as time spent away from home increases so does the likelihood
of victimisation. Extending this work, Groff (2007b) presents a modified version of
the original model in which the spatial activities of agents are more developed.
Exploring several propositions of crime pattern theory, in this variant the impact of
differing conceptualisations of agent activity spaces are examined. Firstly, directed
movement behaviour is compared to the initial random movement model. Agents
operating under the “street directed” movement behaviour are allocated a number of
routine activity nodes and a series of paths that connect them via predetermined
routes. In addition, a further model variant is presented in which agents again utilise
the random movement behaviour but do so in an abstract environment that replaces
the Seattle street network with a uniform grid of intersections. Comparing these
model variants through simulation experimentation, time spent away from home is
held static and resulting crime trends are observed under each of the three concep-
tualisations of activity spaces. Results of these experiments provide fundamental
insights into the likely influence of actor spatial activities and their interaction with
common street geometries. Groff (2008) then presents a further model extension
which incorporates temporal constraints into the activities of agents, comparing
activity spaces conceptualised as simple, temporal and spatio-temporal. In the sim-
ple condition, agents move randomly throughout the street network for some pro-
portion of each day and stay at home for the remainder; in the temporal variant,
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 225

movement remains random but time away from home is specified by a series of
temporal constraints; in the spatio-temporal variant offenders are allocated routine
activity nodes, specified paths to those nodes, and a temporal schedule of activities.
A series of simulated experiments are then performed where the average time spent
away from home is manipulated and the resulting crime examined across all model
configurations. By comparing the simple model to the temporal and spatio-temporal
variants an appropriate counterfactual is established permitting the examination of
the effects of each configuration of hypothesised activity space. Results demon-
strate that several propositions of the opportunity theories concerning actor activi-
ties provide key explanatory capacity in understanding observed likelihood and
spatial distribution of crime events, and moreover that temporal and spatial con-
straints have differential effects on both the prevalence and distribution of crime
within the model.
Similarly, Wang, Liu, and Eck (2008) describe a simulation study of street rob-
bery underpinned by mechanisms of the routine activity approach and crime pattern
theory. Utilising a hybrid of ABM and cellular automata, the presented model situ-
ates agents within an environment drawn from the Cincinnati street network. Agents
represent three key entities: offenders, targets and places. Offender agents operate
under a number of theoretically derived behaviours that allow them to navigate their
environment, choose suitable targets and learn from past experience about suitable
crime places. Similarly, target agents are capable of traversing the environment and
adapting their behaviour to avoid locations where previous victimisation has
occurred. Place agents represent the locations at which crimes can occur and are
represented as cells on the modelled street network. Each place agent is associated
with a place manager designed to represent the crime controller described by Eck
(1994). The management effectiveness of a place is then reflected in a measure of a
place’s ability to control crime within it. Drawing on the routine activity approach,
when offender and target agents converge at a given place, offenders assess their
own motivation, the desirability of available targets, the capability of guardians
present and the management effectiveness of the current place in their decision to
offend. Using this model the authors initially demonstrate that the model produces
results congruent with those expected by the routine activity approach. Subsequently,
two simulated experiments are performed. The first demonstrates that bestowing
agents with temporally constrained routine activities generates hourly patterns of
offending similar to those observed in the empirical study of crime. The second
explores the impact of providing offender and potential victim agents with the abil-
ity to adapt to what is known about “good” crime places, and thus gravitate towards
or away from them respectively. Results of this experiment demonstrate that intro-
duction of the adaptation mechanism produces increases in levels of repeat victimi-
sation, and generates a distribution of crime across place agents that conforms to a
power curve congruent with that commonly observed in the empirical study of
crime (Spelman, 1995).
Brantingham et al. (2008) present an example application of their ABM platform
Mastermind that seeks to provide a formal modelling instrument for criminolo-
gists, policing agencies and city planners interested in the spatial and temporal
226 D. Birks

characteristics of crime in urban areas (Brantingham et al., 2005, 2008;


Brantingham, Glasser, et al., 2005a, 2005b). Focusing on motor vehicle theft in
this application, the Mastermind platform is used to explore the ramifications of
spatially referenced awareness spaces proposed by crime pattern theory. Performing
a series of experiments whereby offender agents are bestowed with different can-
didate representations of awareness spaces and placed within an environment mod-
elled on the downtown street network of Vancouver, British Columbia, the authors
examine how such candidate representations of offender spatial knowledge interact
with patterns of activity, target selection and subsequently crime occurrence. These
experiments demonstrate clear utility of ABM in considering the complex interac-
tions that result from individual offender and target behaviour that would other-
wise be difficult to capture, and quantify through traditional analytical means.
Importantly, they also highlight the insights that can be derived not only from
examining model outcomes, but also the longitudinal trace of simulated actors that
generate such patterns, which in turn increase our understanding of how individ-
ual-level behaviours can interact in complex and often unforeseen ways to generate
spatio-temporal patterns of crime.
More recently, Birks et al. (2012, 2014) explicitly discuss the approach of genera-
tive social science within criminology, presenting a computational laboratory
designed to assess the individual and cumulative explanatory capacity of three key
propositions of the opportunity theories in explaining several commonly observed
characteristics of crime. Creating agents whose movement, target selection and
learning behaviours are designed to reflect key propositions of the routine activity
approach, rational choice perspective, and crime pattern theory, the authors create a
series of theoretical “switches” that allow agent behaviours to be systematically
manipulated. These switches permit agents to be bestowed with the presence or
absence of a mechanism proposed by theory, thus providing a counterfactual against
which the impact of a given mechanism can be assessed. With the hope of under-
standing how these behaviours interact and subsequently influence crime events and
patterns, a series of simulation experiments are then undertaken. Following a tradi-
tional 2 × 2 × 2 experimental design the crime generated by offender agents operating
under each unique configuration of the three behavioural conditions are examined
and compared to three distinct regularities of crime derived from empirical studies.
Moreover, in order to assess the generalisability of observed outcomes, for each
experimental configuration 500 model replications are carried out examining the
impact of the same mechanism configurations across a unique population of agents
situated in a unique abstract urban environment. Results of these investigations con-
sistently demonstrate that the cumulative effects of routinised anchor-based move-
ment, utility-based target selection and spatial learning proposed by the opportunity
theories are sufficient to explain widely observed patterns of crime concentration,
repeat victimisation and offender mobility. Furthermore, the constructed computa-
tional laboratory is used to model both property (e.g. residential burglary—Birks
et al., 2012) and interpersonal (e.g. street robbery—Birks et al., 2014) victimisation,
demonstrating that the same mechanisms are capable of explaining a range of empir-
ical regularities of crime across two fundamentally distinct types of offending.
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 227

These studies provide key tests of the opportunity theory’s depiction of crime
that would be exceedingly difficult to undertake through traditional empirical
approaches. Importantly, their results provide confidence in the viability of proxi-
mal crime event mechanisms put forward by the opportunity theories and in doing
so support situational crime prevention efforts that seek to manipulate such mecha-
nisms to leverage reductions in crime. More generally they also provide a compel-
ling argument for the use of ABM in assessing the causal properties of crime event
mechanisms proposed by theory. Finally, they also deliver essential groundwork for
the incremental development of predictive models that seek to estimate how such
mechanisms might be best manipulated to disrupt crime events.
It is also important to note that while these brief examples deal solely with theo-
ries of environmental criminology in support of situational crime prevention, ABM
may support a wide range of theoretical, and subsequently applied, efforts within
many fields of criminological study. To illustrate such diversity, ABM might be
constructed to explore the evolution of illegal trade networks (Bichler & Birks,
2015), the spatial foraging of protected species and illegal poaching of them (Hill,
Johnson, & Borrion, 2013), the formation of organised criminal groups, the escala-
tion of arguments and the emergence of violent outcomes, and the impact of life
choices on criminal propensity, to name but a few.
The requirements of such agent-based investigations are simple: the presence of
proposed individual level mechanisms sufficiently explicit to be formalised, and
some known empirical regularities that such mechanisms purport to explain, and
against which the explanatory capacity of theory can be assessed. Here, I propose
that these requirements are the hallmark of all good theory irrespective of one’s
intention to implement an ABM. The application of ABM within criminology is
undoubtedly still in its infancy. Indeed, a considerable amount of foundational
research is required if criminologists are to harness the potential of ABM in increas-
ing our understanding of criminological phenomena, and subsequently in the design
of effective interventions aimed at reducing harm.
One potentially transformative application of ABMs within criminology is the
explicit simulation of crime prevention interventions (Groff & Birks, 2008). Once
appropriately validated explanatory models of crime are constructed, researchers
can introduce candidate crime prevention interventions into their simulated societ-
ies with the aim of assessing their likely impact on mechanisms of interest. In the
future, ABMs may be applied in this way to identify potential interventions that are
likely to produce the greatest reductions in crime while also minimising unintended
consequences. While such models cannot (and do not purport to) replace traditional
evaluation of implemented intervention, where resources are often limited, ABM
may permit prioritisation of interventions that are likely to be the most fruitful in
terms of crime reduction. In turn, allowing simulation studies to guide expensive
and resource intensive empirical investigation. While to date less research has been
carried out in this regard, several authors have begun to discuss and explore the pos-
sibilities of such predictive models for intervention prototyping (Bosse et al., 2010;
Bosse & Gerritsen, 2010; Elffers & Gerritsen, 2016; Groff & Birks, 2008; Malleson
& Brantingham, 2009; Melo et al., 2006).
228 D. Birks

With respect to situational crime prevention, the primary requirement for


simulation of intervention is an appropriate model of the proximal causes of crime
events. In order that we can have any level of confidence in models that aim to esti-
mate the impact of candidate situational crime prevention interventions, our explan-
atory models must be capable of producing outcomes congruent with those we
observe in reality. Within environmental criminology progress in this area is mov-
ing relatively quickly, with a range of models producing plausible patterns of crime
events (see above).
One important point with respect to such validation that should be mentioned
here concerns the equifinality principle (von Bertalanffy, 1968). Equifinality
acknowledges that the observed outcome state of any open system may be the result
of one or more of numerous potential input states, that is—many configurations of
mechanisms may be capable of explaining the same observed phenomena. As such,
ABMs may assess only the explanatory sufficiency of proposed mechanisms and
not their explanatory necessity.
Consequently, the agent-based model is capable only of assessing which mecha-
nisms can and cannot produce a specific outcome, not which one is producing that
outcome in reality. Note however that this observation applies to almost all forms of
analysis applied in the study of criminological phenomena. Perhaps most impor-
tantly then, ABMs can be used to falsify theoretical accounts of crime. Given the
vast array of potential interventions already on offer to practitioners which must rely
on some model of crime (albeit often a less explicit one than those studied through
ABM), this ability to separate those accounts that are capable of producing out-
comes in a particular context seems both an attainable and worthy goal.
Ultimately, following other disciplines that have adopted computational models,
the processes involved in producing practically robust and insightful predictive
models of intervention effectiveness will likely be both iterative and cyclical.
Following such an approach models are designed to depict criminological theory,
such models are then assessed against a range of empirical regularities of crime.
Once a model is capable of producing a range of plausible outcomes (the more the
better) simulations can be constructed using these behaviours and incorporating
candidate interventions. Subsequently, the impacts of such interventions are esti-
mated, and compared to (likely small scale) real world evaluations of interventions,
with observed disparities used to incrementally refine model constructs.
Accomplishing this goal will take time and involve forward planning in the design
of empirical studies that aim to best inform their simulated equivalents. Yet, as more
and more data describing crime events become available, and there is more and
more interest in the diversification of the tools brought to reduce crime, it is one that
I hope will be met.
At this point it is important to note that while ABM has the potential to inform both
criminology and crime prevention in new and unique ways, it is not without its chal-
lenges, several of which are now discussed. ABMs, by their very nature, seek to pro-
vide appropriate analogues of complex systems. As such, those who develop them
should always remain mindful of the need to manage model complexity. Models that
become as complex as the system they aim to mimic can become similarly difficult to
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 229

understand and interpret. Moreover, such models risk over-fitting to specific contexts,
which, in turn, may limit the generalisability of model outcomes.
While the appropriate level of complexity associated with any given model is
clearly driven by its intended applications (Axtell & Epstein, 1994), ABM should
be built in an incremental fashion. In doing so, simple models precede more com-
plex ones, such that increasingly nuanced system components are incorporated
piece by piece, allowing them to be appropriately tested and their impact on model
behaviour systematically assessed (Townsley & Birks, 2008). In this way, step
changes in model complexity are planned, and trade-offs between complexity and
utility can be addressed in the context of research questions that guide model devel-
opment. Ultimately, such an approach follows the principle that underlies all model
development, not just ABM—to seek the most parsimonious model that remains
capable of explaining the widest variety of known system characteristics.
A further challenge facing the application of ABM within criminology relates
to their position within the field. Currently, ABM are applied relatively rarely by
criminologists in comparison to other analytical techniques. As a result, peer
understanding of ABM methods is not widespread. As ABM can be applied in a
vast array of contexts, for a range of purposes, and in substantively different
ways, methods of standardised model description will likely be increasingly use-
ful as the application of ABM grows within criminology. As such, the develop-
ment of standardised frameworks to describe ABMs and enable model
peer-review, replication, comparison and extension are of key importance. Such
practices have been shown to promote transparency, understanding and accep-
tance of ABM in other fields (e.g. Grimm et al., 2010). Thus, it is from the adop-
tion or modification of these established methods that the most productive efforts
will likely originate.
Finally, in disseminating ABM research, authors must acknowledge that model
outcomes are necessarily a function of assumptions derived in the development of a
model. Consequently, bold claims relating to the insights derived from models
should be appropriately tempered and supported by theoretical and empirical study.
Here, it is only through the triangulation of multiple complementary methods that
our understanding of what are undoubtedly complex social systems is best
supported.

Conclusion

ABM offers a novel method by which mechanisms specified by criminological the-


ory can be examined, and tested in ways that are often infeasible through traditional
methods due to a range of well acknowledged constraints. Here, I have argued that
performing such computational investigations can support more effective crime pre-
vention strategies—in that a detailed understanding of the mechanisms involved in
the commission of crime events is required to inform optimal intervention
strategies.
230 D. Birks

The ABM currently sits in an interesting position relative to other research


methodologies within criminology. It can be informed by both quantitative and quali-
tative research. It can generate both direct theoretical and applied insights and enable
all-important cross-pollination of these increases in understanding. Given the special-
ist skills often required to develop such models (Eck & Liu, 2008b), the construction
of ABMs often brings together scholars from diverse backgrounds beyond criminol-
ogy, that include computer scientists, mathematicians, geographers, psychologists,
and moreover practitioners and policy makers who bring invaluable domain knowl-
edge. This is perhaps the most powerful lever the ABM can pull in advancing the
disciplines of criminology and crime prevention. Where we seek to understand and
address hard societal problems that span multiple traditional disciplinary boundaries,
unifying interdisciplinary perspectives to help address crime is undoubtedly a useful
enterprise. It is hoped that this chapter provides a demonstration of the rationale
underlying, and basic apparatus required of such future models. If nothing else, the
ABM’s focus on developing formal models of proposed crime event mechanisms
should promote rigour in the study of criminological phenomena, and ultimately sup-
port the development of interventions, which by necessity require such specificity in
order to be effective in reducing crime.

References

Axtell, R. L., & Epstein, J. M. (1994). Agent-based modeling: Understanding our creations. The
Bulletin of the Santa Fe Institute, 9(2), 28–32.
Bichler, G., & Birks, D. (2015). Exposed: Agent-based simulation of military weapons transfers.
Poster presented at the Center for Criminal Justice Research, California State University, San
Bernardino. doi:10.13140/RG.2.1.1630.9927.
Birks, D., Donkin, S., & Wellsmith, M. J. (2008). Synthesis over analysis: Towards an ontology of
volume crime simulation. In L. Liu & J. Eck (Eds.), Artificial crime analysis systems: Using
computer simulations and geographic information systems (pp. 160–191). Hershey, PA: Idea
Group.
Birks, D., & Elffers, H. (2014). Agent-based assessment of criminological theory. In G. Bruinsma
& D. Weisburd (Eds.), Encyclopaedia of criminology (pp. 19–32). New York: Springer.
Birks, D., Townsley, M., & Stewart, A. (2012). Generative models of crime: Using simulation to
test criminological theory. Criminology, 50, 221–254.
Birks, D., Townsley, M., & Stewart, A. (2014). Emergent regularities of interpersonal victimiza-
tion: An agent-based investigation. Journal of Research in Crime and Delinquency, 51(1),
119–140.
Bonabeau, E. (2002). Agent-based modeling: Methods and techniques for simulating human sys-
tems. Proceedings of the National Academy of Sciences of the United States of America, 99(3),
7280–7287.
Bosse, T., Elffers, H., & Gerritsen, C. (2010). Simulating the dynamical interaction of offenders,
targets and guardians. Crime Patterns and Analysis, 3(1), 51–66.
Bosse, T., & Gerritsen, C. (2010). An agent-based framework to support crime prevention. In
Proceedings of the Ninth International Conference on Autonomous Agents and Multi-Agent
Systems, AAMAS‘10 (pp. 525–532). ACM Press.
Brantingham, P., & Brantingham, P. (1978). A theoretical model of crime site selection. In
M. Krohn & R. Akers (Eds.), Crime, law and sanctions. Beverly Hills, CA: Sage.
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 231

Brantingham, P., & Brantingham, P. (1981). Notes on the geometry of crime. In P. Brantingham &
P. Brantingham (Eds.), Environmental criminology. Beverly Hills, CA: Sage.
Brantingham, P., & Brantingham, P. (1993). Environment, routine, and situation: Toward a pattern
theory of crime. In R. Clarke & M. Felson (Eds.), Routine activity and rational choice, advances
in criminological theory (Vol. 5). New Brunswick, NJ: Transaction Publishers.
Brantingham, P., & Brantingham, P. (2004). Computer simulation as a tool for environmental
criminologists. Security Journal, 17(1), 21–30.
Brantingham, P., & Brantingham, P. (2008). 5. Crime pattern theory. In R. Wortley & L. Mazerolle
(Eds.), Environmental criminology and crime analysis (p. 78). Cullompton: Willan.
Brantingham, P., Brantingham, P., & Glasser, U. (2005). Computer simulation as a research tool in
criminology and criminal justice. Criminal Justice Matters, 58.
Brantingham, P., Glasser, U., Jackson, P., Kinney, B., & Vajihollahi, M. (2008). Mastermind:
Computational modelling and spatiotemporal aspects of crime in urban environments. In L. Liu
& J. Eck (Eds.), Artificial crime analysis systems: Using computer simulations and geographic
information systems (p. 252–280). Idea Group Inc.
Brantingham, P., Glasser, U., Kinney, B., Singh, K., & Vajihollahi, M. (2005a). A computational
model for simulating spatial and temporal aspects of crime in urban environments. IEEE
International Conference on Systems, Man, and Cybernetics, 4, 3667–3674.
Brantingham, P., Glasser, U., Kinney, B., Singh, K., & Vajihollahi, M. (2005b). Modelling urban
crime patterns: Viewing multi-agent systems as abstract state machines. Proceedings of the
12th International Workshop on Abstract State Machines, 101–117.
Brayley, H., Cockbain, E., & Laycock, G. (2011). The value of crime scripting: Deconstructing
internal child sex trafficking. Policing, 5(2), 132–143.
Clarke, R. (1980). “Situational” crime prevention: Theory and practice. The British Journal of
Criminology, 20, 136–147.
Clarke, R. (1995). Situational crime prevention. In M. Tonry & D. P. Farrington (Eds.), Crime and
justice, volume 19 building a better society—Strategic approaches to crime prevention.
Chicago: The University of Chicago Press.
Clarke, R. (1997). Situational crime prevention: Successful case studies (2nd ed.). Guilderland,
NY: Narrow & Heston.
Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 44, 588–608.
Cornish, D., & Clarke, R. (1986). The reasoning criminal: Rational choice perspectives on offend-
ing. New York, NY: Springer.
Cornish, D., & Clarke, R. (2008). 2. The rational choice perspective. In R. Wortley & L. Mazerolle
(Eds.), Environmental criminology and crime analysis (pp. 21–47). Cullompton: Willan.
Eck, J. (1994). Drug markets and drug places: A case-control study of the spatial structure of illicit
drug dealing. Doctoral dissertation.
Eck, J. (1995). Examining routine activity theory: A review of two books. Justice Quarterly, 12(4),
983–997.
Eck, J., & Liu, L. (2008a). Contrasting simulated and empirical experiments in crime prevention.
Journal of Experimental Criminology, 4, 195–213.
Eck, J., & Liu, L. (2008b). Varieties of artificial crime analysis: Purpose, structure, and evidence
in crime simulations. In L. Liu & J. Eck (Eds.), Artificial crime analysis systems: Using com-
puter simulations and geographic information systems (pp. 413–432). Hershey, PA: Idea
Group.
Ekblom, P., & Pease, K. (1995). Evaluating crime prevention. Crime and Justice, 19, 585–662.
Elffers, H., & Gerritsen, C. (2016). Investigating prevention with simulation methods. In B. Leclerc
& E. U. Savona (Eds.), Crime prevention in the 21st century: Insightful approaches for crime
prevention initiatives. New York: Springer.
Epstein, J. (1999). Agent-based computational models and generative social science. Complexity,
4(5), 41–60.
Epstein, J. (2008). Why model? Journal of Artificial Societies and Social Simulation, 11(4), 12.
Retrieved December 10, 2014, from http://jasss.soc.surrey.ac.uk/11/4/12.html.
232 D. Birks

Epstein, J., & Axtell, R. (1996). Growing artificial societies: Social science from the bottom up.
Cambridge: MIT Press.
Felson, M. (2008). 4. Routine activity approach. In R. Wortley & L. Mazerolle (Eds.), Environmental
criminology and crime analysis (pp. 70–77). Cullompton: Willan.
Gilbert, N., & Troitzsch, K. (2005). Simulation for the social scientist (2nd ed.). Berkshire, UK:
Open University Press.
Grimm, V., Berger, U., DeAngelis, D. L., Polhill, J. G., Giske, J., & Railsback, S. F. (2010). The
ODD protocol: A review and first update. Ecological Modelling, 221(23), 2760–2768.
Grodzins, M. (1958). The metropolitan area as a racial problem. Pittsburgh: University of
Pittsburgh Press.
Groff, E. (2007a). Simulation for theory testing and experimentation: An example using routine
activity theory and street robbery. Journal of Quantitative Criminology, 23(2), 75–103.
Groff, E. (2007b). ‘Situating’ simulation to model human spatio-temporal interactions: An
example using crime events. Transactions in GIS, 11(4), 507–530.
Groff, E. (2008). Adding the temporal and spatial aspects of routine activities: A further test of
routine activity theory. Security Journal, 21(1–2), 95–116.
Groff, E., & Birks, D. (2008). Simulating crime prevention strategies: A look at the possibilities.
Policing, 2(2), 175–184.
Guerette, R. T., & Bowers, K. J. (2009). Assessing the extent of crime displacement and diffusion
of benefits: A review of situational crime prevention evaluations. Criminology, 47(4),
1331–1368.
Hedström, P. (2005). Dissecting the social. On the principles of analytical sociology. Cambridge:
Cambridge University Press.
Hedström, P., & Swedberg, R. (Eds.). (1998). Social mechanisms. An analytical approach to social
theory. Cambridge: Cambridge University Press.
Hill, J. F., Johnson, S. J., & Borrion, H. (2013). Potential uses of computer agent-based simulation
modeling in the evaluation of wildlife poaching. In A. Lemieux (Ed.), Situational prevention of
poaching. London: Routledge.
Jeffrey, C., & Zahm, D. (1993). Crime prevention through environmental design, opportunity
theory, and rational choice models. In R. Clarke & M. Felson (Eds.), Routine activity and ratio-
nal choice (pp. 323–350). London: Transaction Publishers.
Macy, M. W., & Willer, R. (2002). From factors to actors: Computational sociology and agent-
based modeling. Annual Review of Sociology, 28, 143–166.
Malleson, N., & Brantingham, P. L. (2009). Prototype burglary simulations for crime reduction
and forecasting. Crime Patterns and Analysis, 2(1), 47–65.
Melo, A., Belchior, M., & Furtado, V. (2006). Analyzing police patrol routes by simulating the
physical reorganization of agents (Multi-agent-based simulation VI, pp. 99–114). Berlin:
Springer.
Rao, A. S., & Georgeff, M. P. (1995). BDI-agents: From theory to practice. Proceedings of the
First International Conference on Multiagent Systems (ICMAS‘95).
Shadish, W., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs
for generalized causal inference. Belmont, CA: Wadsworth Cengage learning.
Sherman, L., Gottfredson, D., MacKenzie, D., Eck, J., Reuter, P., & Bushway, S. (1998). Preventing
crime: What works, what doesn’t, what’s promising. National Institute of Justice Research
Brief.
Spelman, W. (1995). Criminal careers of public places. In J. Eck & D. Weisburd (Eds.), Crime and
place (pp. 115–144). Monsey, NY: Criminal Justice Press.
Tilley, N. (2002). Introduction. In N. Tilley (Ed.), Evaluation for crime prevention (Vol. 14).
Monsey, NY: Criminal Justice Press.
Townsley, M., & Birks, D. J. (2008). Building better crime simulations: Systematic replication and
the introduction of incremental complexity. Journal of Experimental Criminology, 4(3),
309–333.
von Bertalanffy, L. (1968). General systems theory: Foundations, development, applications.
New York: George Braziller.
14 Formal Models of the Crime Event: Agent-Based Modelling in Support of Crime… 233

Wang, X., Liu, L., & Eck, J. (2008). Crime simulation using GIS and artificial intelligent agents.
In L. Liu & J. Eck (Eds.), Artificial crime analysis systems: Using computer simulations and
geographic information systems (pp. 209–225). Hershey, PA: Idea Group.
Wortley, R., & Mazerolle, L. (2008). Environmental criminology and crime analysis. Cullompton:
Willan.
Chapter 15
Investigating Prevention by Simulation
Methods

Charlotte Gerritsen and Henk Elffers

Introduction

From the point of view of crime prevention, crime can be envisaged as unfolding
itself in two phases in time: anticipation, which happens before the crime is
actually being executed, and execution, which happens during the actual com-
mitment of the crime. What is happening in the aftermath of crime (detection,
prosecution) is interesting in itself for criminologists, but has little relevance
for prevention.
Prevention takes a different guise in the anticipatory and execution phase of
crime. In this chapter we introduce the term anticipatory prevention which
refers to attempts to prohibit the idea that the necessary conditions for a crime
to occur will simultaneously materialise. Those necessary conditions are,
according to the routine activity framework (Cohen & Felson, 1979), the co-
occurrence in time and space of a motivated offender and a suitable target, in
the absence of a capable guardian.
When a crime is actually happening, anticipatory prevention has failed, but
we still can envisage measures that attempt to deflect the seriousness or indeed
interrupt the course of the crime. For such action we coin the term mitigating
prevention.
In this chapter we look into simulation methods for investigating the effects of
prevention, and we treat this topic separately for anticipatory and mitigating pre-
vention, as we argue that anticipatory prevention requires agent based simulation
methods, while mitigating prevention should bank on simulation in the sense of
studying potential victims or guardians in an artificial (laboratory) situation.

C. Gerritsen (*) • H. Elffers


NSCR & VU University Amsterdam, Amsterdam, The Netherlands
e-mail: CGerritsen@nscr.nl

© Springer International Publishing Switzerland 2017 235


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_15
236 C. Gerritsen and H. Elffers

Investigating Anticipatory Prevention by Agent Based


Simulation

Of late, we are enjoying the coming of age of agent based (ABM) simulation
(Epstein, 2006) as a research method in criminology (Brantingham & Brantingham,
2004). ABM-simulations build an artificial context in which artificial agents (reflect-
ing actors such as offenders, targets, guardians and others) iteratively “perform”
actions according to simple (sometimes deterministic, but often probabilistic) rules
that usually prescribe certain actions dependent on the context in which the agent is
located in space, relative to where other agents are.1 Recently some important steps
towards understanding the possible role of simulation studies in theory testing have
been made (Birks & Elffers, 2014; Birks, Townsley, & Stewart, 2012; Groff, 2007).
In that line of thinking, simulation is used for investigating which elements of a
theory are indispensable for producing—in an artificial context—regularities in
crime patterns, regularities that have been demonstrated empirically to have been
observed in real life. Such regularities are in this context often called “stylised
facts”. As such it is a theory development method. It investigates the structure of
explanatory theories, by demonstrating that some parts of a theory cannot be missed
in the sense that without them, established patterns in offending do not occur in the
artificial environment, while others may, at least in the artificial context, be superflu-
ous. Simulation is used, because analytical grasping the complex interactions
between large numbers of diverse agents defies human analytical shrewdness.
Notice, however, that although a particular theoretical construct may be capable of
generating some regularity, there may be other, non-modelled constructs that are
equally capable to do so. Likewise, the usefulness of simulation methods is, of
course, also bound to the theoretical paradigm in which it has been formulated. For
a more thorough discussion of these epistemological issues please refer to the litera-
ture cited above.
In between theory oriented use of simulation and practical application of ABM
simulation models, a number of development studies proposing simulation models
for certain crimes have been published (e.g. Bosse, Gerritsen, & Klein, 2010; Groff,
2008; Johnson, 2008; Liu, Wang, Eck, & Liang, 2005; Malleson, Evans, & Jenkins,
2009). Such models are indispensable for the next step, applying simulation to eval-
uate concrete measures.
In the present chapter, we look at the application of simulation methods as a
means for evaluation of preventive measures, in the line of Groff and Birks (2008).
Within an ABM-framework and using the routine activity approach, this requires

1
The classical (non-criminological) example is Schelling’s (1987) differential moving tendency
simulation: white and black pawns are randomly distributed on a chessboard. Repeatedly an arbi-
trary pawn may select to be relocated to a new square, where white pawns prefer squares next to
other white pawns just a little bit more than next to black pawns and vice versa. In no time (i.e. in
not too many iterations of the simulation) under this scheme almost all pawns are segregated in
black and white neighbourhoods. (Notice that this simulation can be easily done “by hand” and
does not need a computer).
15 Investigating Prevention by Simulation Methods 237

models with agents that are meant to imitate the behaviour of offenders, targets and
sometimes guardians.
Within a simulation model (existing or made for the purpose), a prevention mea-
sure can be implemented in the form of a change in the characteristics of various
agents (e.g. moves of offender agents are hampered or precluded around places with
high concentrations of targets; guardian agents get an ability to “observe” what is
happening farther away; or a number of “property targets” get a higher probability
of stopping an offender agent in successfully burgling that property). Evaluating a
prevention measure boils down to comparing the outcomes of a number of simula-
tion runs with and a number of simulations without the new characteristic built in.
Usually we understand how a change in a characteristic may change just the next
single move of an agent, but due to the complex interaction between various agents,
some of them having built in those new characteristics, further development of the
whole artificial society becomes opaque. It is exactly the—usually rather com-
plex—process that governs the whereabouts of all agents, offenders, targets and
guardians, and their associated actions, that lies behind the occurrence of crime and
hence behind the resulting spatio-temporal pattern of crime occurrence. Simulation
is foremost useful for disentangling the complexities of a multitude of interactions,
which of course does not imply that not some problems demand so many and such
complex interactions that also simulation methods start to become very and perhaps
too complex.
Notice that the more we already know about the theory in which a prevention
method becomes introduced, and the more we know about the validity of the simu-
lation model in which they are built in, the more useful a simulation will be. In a
well-understood case we are aware what differences can be accounted for by known
elements in the theory, so that we can concentrate on differences generated by the
newly built-in prevention elements.
Anticipatory prevention is, within this framework, every influence that precludes
co-occurrence in time and space of offender and target agents and the absence of the
third agent, guardians, or, expressing it differently, every influence on their respec-
tive movements that decreases the likelihood that this occurrence will take place.
Examples of anticipatory prevention are measures for decreasing the streams of
potential offenders, of potential targets and for increasing the streams of potential
guardians. We may also think of measures for deflecting streams in order to mini-
mise the likelihood of meeting of potential targets and offenders, and maximising
the likelihood of meeting of guardians with offenders and targets, increasing the
opportunities for guardians to observe their environment (cf. Felson, 2006).

Investigating Mitigating Prevention by Simulation Methods

The field of mitigating prevention has a quite different character from that of antic-
ipatory prevention (Reynald, 2011; van Bavel & Elffers, 2013). In the execution
phase of a crime, standard conditions of routine activity theory have been fulfilled,
238 C. Gerritsen and H. Elffers

and an offender has decided to actually commit a crime. Mitigating prevention


consists of measures that, during the execution of this intention, change the percep-
tion of the offender, such that he either interrupts the execution of the crime, or at
least changes tack and maybe scale down his crime. Notice that such a change is
leveraged by the perception of the perpetrator, during his execution of a crime, of
what is going on around him, especially with respect to target and possible guard-
ians (cf. Lindegaard, Bernasco, Jacques, & Zevenbergen, 2013). This is a funda-
mental difference compared to the anticipatory phase, where the central processes
are linked to the co-occurrence in space and time of various actors. This difference
is reflected in the dominant research paradigms in both cases. While ABM is not
completely absent,2 the principal research method for mitigating prevention is
placing actors in a simulated environment and observing their behaviour. Notice
that it is hard to stage such research with “real perpetrators”.3 Most research is
directed at the behaviour of victims and possible guardians, e.g. for training police
officers how to operate in order to de-escalate violence, in a “laboratory” situation
where (human) actors simulate the role of violent perpetrators, according to a
script. Such simulations may also take place using virtual reality, in which case the
virtual actors react on the behaviour of the human actor under scrutiny, based on a
given theory of criminal interaction.

Examples of Anticipatory Prevention

Let us present here a few examples of simulation studies on anticipatory prevention.4


Bosse, Elffers, and Gerritsen (2010) have looked into the effect of a number of
different surveillance strategies of (formal) guardians, within a routine activity para-
digm on burglary, which is thought to take place when a motivated offender finds an
insufficiently guarded attractive target. The dynamics become quite complex when
offenders are informed about the different and dynamic values of targets, deducted
from the popularity of targets under fellow offenders, and when guardians are
informed about where previous offences have taken place. The authors present four
“worlds” represented as a 5 × 5 grid and consisting of 25 static targets. The distribu-
tion of the attraction values of the targets differs per world (equal values, distributed
without structure, concentric ring structure and segregated). In these worlds
offenders and guardians move around according to some predefined behavioural

2
For example Gerritsen’s (2011) work on ABM modelling of aggression in crowds.
3
There is an interesting research tradition in bringing ordinary people in such environments and
look whether they will displaying criminal behaviour, as a function of environmental queues, e.g.
in tax evasion simulation (Webley, Robben, Elffers, & Hessing, 1991). Van Bavel (forthcoming) is
reporting on theft experiments with ordinary people in the role of offenders. The crux in that type
of experiments is of course how to manipulate the motivation of the prospective offenders.
4
Notice that authors cited here did not present their work in terms of anticipatory prevention,
which is a term introduced in the present chapter.
15 Investigating Prevention by Simulation Methods 239

rules. Offenders could move around randomly or they move with a high probability
toward the most attractive neighbour. Offenders had an attraction threshold; they
would only execute a burglary onto a target if the attraction value of that target
exceeds the threshold. Guardians could move around (1) randomly, (2) with a high
probability towards the most victimised neighbour or (3) a restricted version of
strategy, (2) with some restrictions to the area in which they could move around.
These three simulation strategies may be seen as the analogues of real life strate-
gies: random surveillance, global hot spot surveillance and local hot spot surveil-
lance. Offenders refrain from crime when a guardian is present at the same spot.
From the results of the simulations the authors inferred that it is better for guard-
ians to have a strategy to move to the most victimised targets than to move ran-
domly. Further, a clear difference between crime rates was observed, co-varying
with degree of spatial concentration of attraction values of the targets. Of course, it
is important to realise that these conclusions were drawn based on the assumptions
made by the authors in the development of the model.
The influence of different police patrol strategies on crime rates is an interest-
ing application of ABM. This topic has been explored by Melo, Belchior, and
Furtado (2006). In their article they describe a tool for assisting the investigation
of different strategies of spatial (re-)organisation of (formal) guardianship. They
apply the tool to the public safety domain and show the possibilities for preven-
tive policing. Their hypothesis is “that by knowing where the crime is currently
happening and its associated reasons, it is possible to make a more optimised
distribution of human resources and, consequently, to decrease the overall crime
rate” (pp. 101–102).
The society they created is represented by a grid of cells and consists of facilities
(these reflect commercial or entertainment businesses, and are also called notable
points or targets), offenders and police officers, all located in time and space.
Offenders are able to see a fixed number of cells around them. Police officers intend
to prevent crime and each police team has at least one route to patrol. Their presence
on this route and at the notable points prevents crime from happening.
Three different types of routes can be distinguished, namely short routes, wide
routes and critical routes. When the short route approach is used the team heads to
the most vulnerable point by creating a route that leads directly to this point and
others closest to it. A wide route means that teams take the longest possible path to
the most vulnerable point in their beat. Police teams allocated to the critical routes
approach need to cover notable points that are identified as having the highest crime
rate. The vulnerability of a notable point is calculated by comparing the number of
crimes that occurred and the number of crimes prevented. When the first outweighs
the second the location is considered to be vulnerable.
The main conclusion from this work is that regardless of the type of routes, route
organisation reduces the crime rates. A high level of mobility of the police makes
them visible and offers a sense of safety. They should re-evaluate vulnerability
often, but it is necessary that they spend some time (at least one “day”) at locations
in order to actually create the sense of safety. Unfortunately, the authors did as yet
not take the costs of moving and spatial reorganising into account.
240 C. Gerritsen and H. Elffers

Bosse and Gerritsen (2010) performed similar work and compared different
strategies for guardian movement in terms of their efficiency (costs) and effective-
ness (in terms of number of assaults). They compared reactive, anticipating and
hybrid approaches. In the reactive approaches, guardians react to the density of
offenders, or passers-by present at a certain location at an earlier time point, or to
the number of assaults that took place at that location in the recent past or over all
earlier time points. When the guardians take an anticipating approach they will
move to a location where they expect a high density of offenders or passers-by or a
high crime rate. The hybrid approaches are combinations of reactive and anticipat-
ing approaches namely (1) assaults that took place in the recent past and the density
of passers-by expected in the near future and (2) assaults that took place at all earlier
time points & the density of passers-by in the near future. The total amount of
assaults and the total costs of each of the approaches were calculated as well during
the entire simulation period. These approaches were tested in five different scenar-
ios (with different and changing attractiveness values of the hot spot locations). The
results show that the hybrid strategies are most effective but that the anticipatory
approaches are most cost-efficient.
Though the police strategy simulations discussed here show interesting and
promising results, they should not be overgeneralised. The simulations have been
performed with a limited amount of scenarios and using specific parameters and
simplified assumptions. The main goals of these approaches were to test the under-
lying theory and to attract the attention of policy makers by demonstrating the pos-
sibilities of ABM in this domain. In fact, the papers show that police strategy choice
may be a fruitful approach for anticipatory prevention, and they can be seen as an
exhortation to consider real life experiments now.
These studies all demonstrate that relatively simple interactions between offend-
ers, guardians and passers-by can result in complex dynamical patterns. When the
number of aspects increases it will be more difficult to oversee all patterns.
Computational methods like ABM can be beneficial in exploring these complex
systems.
The interaction between guardians, offenders and targets has been the research
interest of many other interesting simulation projects, among which research into
victims (Birks, Townsley, & Stewart, 2014; Johnson, 2008), pirates (Marchione,
Johnson, & Wilson, 2014), poachers (Yang, Ford, Tambe, & Lemieux, 2014), deter-
rence (van Baal, 2004), peer influence (Bosse, Elffers, and Gerritsen, 2010).

Examples of Mitigating Prevention

Mitigating prevention can be mostly found in applications which have de-escalation


of (consequences of) offending behaviour as a research focus. We mention two
research programmes in this respect.
In the STRESS project (which stands for Simulation-based Training of Resilience
in Emergencies and Stressful Situations), agent-based simulation is used to develop
15 Investigating Prevention by Simulation Methods 241

a virtual training model for people who work in the public safety domain
(http://stress.few.vu.nl). These people often have to deal with (verbal) aggressive
behaviour. The environment developed in the STRESS project helps them to learn
how to apply de-escalating communication techniques. The trainee is placed in a
scenario in the training environment and has to make decisions on how to proceed
in a certain situation. During the training the heart rate and skin conductance are
measured and together with the answers given during the training they provide input
for the feedback by the supervisor (de Man, 2014) and help make the training more
tailored to the specific needs of a trainee to ensure an optimal learning effect (Bosse,
Gerritsen, de Man, & Treur, 2013).
Another research project which focuses on de-escalating aggressive behaviour is
the recently started de-escalate project (Defusing Escalating Behaviour through the
Use of Interactive Light Scenarios, http://www.de-escalate.nl; de Kort et al., 2014).
While in the STRESS project the focus lies on intervention by a human being, the
de-escalate project uses the environment as the most important mitigating factor.
From Farrington and Welsh’s (2002) review it is known that street lighting influ-
ences crime levels. In this project it is hypothesised that light can play an important
role in de-escalating aggressive behaviour. The role of light is twofold: light can
alter the atmosphere and make a tense situation more relaxed, but light also has a
positive effect on the psychological and mental states related to aggression. The
main focus is on preventing or decreasing situations that unintentionally got out of
hand (or likely will in the near future) since it is hypothesised that light does not
have an influence on intentional aggressive behaviour. Before actually experiment-
ing in the field, the project intends to use simulation methods to explore various
complex interactions between several possibly aggressive agents under the influ-
ence of interactive light scenarios.

Suggestions for Future Research and Discussion

More and more criminologists find their way to using ABMs and other types of
simulation which is promising for potential applications. The main focus of crimi-
nologists who currently use ABMs in their research has primarily been on the
domain of environmental criminology with the interaction between potential offend-
ers, guardians and targets as main focus. There are so much more other interesting
opportunities in different criminological research areas. It is important that crimi-
nologists think outside the box they are so often trapped in.
Gathering and analysis of data is still the main focus in criminological research, but
the use of computer simulations has some additional and interesting benefits over
traditional methods. ABM helps gain more insight into (informal) theories, especially
when relationships are not clear at first sight, for example because the number of rel-
evant aspects is large or relationships between aspects are complex. Next to its use as
an analytical device for investigating the structures of theories, we feel that simulation
has also potential for actually investigating promising intervention strategies. In an
242 C. Gerritsen and H. Elffers

agent-based simulation it is relatively easy to manipulate parameters and see what the
consequences are. This is hardly possible in a real world setting (Birks & Elffers,
2014; Groff, 2007; Townsley & Birks, 2008; Townsley & Johnson, 2008). Further, by
using agent-based simulations it is possible to perform experiments that would be
unethical to do in real life. It is a promising alternative to test new policies before they
are actually implemented.
On the side of using simulated environments for investigating mitigating preven-
tion strategies, we also see some promising developments. Again, the major advan-
tage of such methods is that it is possible to bring people in threatening environments
without exposing them to real harm, as well as making it possible to re-iterate train-
ing in a cheap and reproducible way. It makes exercising possible for trainees,
sometimes even without involvement of others. Using a simulation for training pur-
poses should be considered as an addition to existing real life training, not as a
substitute of the latter.
Of course there are some limitations that need to be considered when using simu-
lations. A model is not identical to the real world, but is a stylised representation.
Using simulations in pre-evaluating prevention strategies should keep an eye open
for possible mismatch of models and reality. Conclusions drawn on the basis of a
given model are in fact based on the assumptions of the developers. That is of course
not different from other uses of models, but it does not harm to be aware of it.
We would relish to see the arrival of research programmes that look in an integrated
way to (1) simulation model development and testing on stylised facts, (2) pre-evaluat-
ing of prevention strategies, and (3) real-life evaluations of the measures that come out
of simulation studies as the most fruitful ones (Eck & Liu, 2008). We would also wel-
come simulations into more sophisticated versions of the interplay between controllers
of different types (handlers, guardians, place managers). Thus, we may exploit the
likely contributions of simulation, without being blind for its limitations. This seems to
us to be the way forward, both in anticipatory and mitigating prevention contexts.

References

Birks, D., & Elffers, H. (2014). Agent-based assessment of criminological theory. In G. Bruinsma
& D. Weisburd (Eds.), Encyclopaedia of criminology (pp. 19–32). New York: Springer.
Birks, D., Townsley, M., & Stewart, A. (2012). Generative models of crime: Using simulation to
test criminological theory. Criminology, 50(1), 221–254.
Birks, D., Townsley, M., & Stewart, A. (2014). Emergent regularities of interpersonal victimisation:
An agent-based investigation. Journal of Research in Crime and Delinquency, 51(1), 119–140.
Bosse, T., Elffers, H., & Gerritsen, C. (2010). Simulating the dynamical interaction of offenders,
targets and guardians. Crime Patterns and Analysis, 3(1), 51–66.
Bosse, T., & Gerritsen, C. (2010). An agent-based framework to support crime prevention. In
Proceedings of the Ninth International Conference on Autonomous Agents and Multi-Agent
Systems, AAMAS‘10 (pp. 525–532). New York: ACM Press.
Bosse, T., Gerritsen, C., de Man, J., & Treur, J. (2013). Effects of virtual training on emotional
response: A comparison between different emotional regulation strategies. In Proceedings of
the 7th International Conference on Brain and Health Informatics, BHI‘13 (Lecture notes in
artificial intelligence, pp. 21–31). Berlin: Springer.
15 Investigating Prevention by Simulation Methods 243

Bosse, T., Gerritsen, C., & Klein, M. (2010). Predicting the development of juvenile delinquency
by simulation. Berlin: Springer.
Brantingham, P. L., & Brantingham, P. J. (2004). Computer simulation as a tool for environmental
criminologists. Security Journal, 17(1), 21–30.
Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 44, 588–608.
de Kort, Y. A. W., IJsselsteijn, W. A., Haans, A., Lakens, D., Kainauskaite, I., & Schietecat, A.
(2014). De-escalate: Defusing escalating behaviour through the use of interactive light
scenarios. www.de-escalate.nl
de Man, J. (2014). Analysing emotional video using consumer EEG hardware. In Proceedings of
the 16th International Conference on Human-Computer Interactions, HCI‘14 (pp. 729–738).
Berlin: Springer.
Eck, J., & Liu, L. (2008). Contrasting simulated and real experiments in crime prevention. Journal
of Experimental Criminology, 4, 195–213.
Epstein, J. (2006). Generative social science: Studies in agent-based computational modeling.
Princeton: Princeton University Press.
Farrington, D. P., & Welsh, B. C. (2002). Effects of improved street lighting on crime: A systematic
review (Home Office Research study, Vol. 251). London: Home Office.
Felson, M. (2006). Crime and nature. Thousand Oaks, CA: Sage.
Gerritsen, C. (2011). Using ambient intelligence to control aggression in crowds. Proceedings
of the Fifth International Workshop on Human Aspects in Ambient Intelligence (pp. 53–56).
Groff, E. (2007). Simulation for theory testing and experimentation: An example using routine
activity theory and street robbery. Journal of Quantitative Criminology, 23(2), 75–103.
Groff, E. R. (2008). Characterizing the spatio-temporal aspects of routine activities and the geo-
graphic distribution of street robbery. In L. Liu & J. Eck (Eds.), Artificial crime analysis sys-
tems: Using computer simulations and geographic information systems (pp. 226–251).
Hershey, PA: Idea Group.
Groff, E. R., & Birks, D. (2008). Simulating crime prevention strategies: A look at the possibilities.
Policing, 1, 1–10.
http://stress.few.vu.nl
http://www.de-escalate.nl
Johnson, S. (2008). Repeat burglary victimisation: A tale of two theories. Journal of Experimental
Criminology, 4, 215–240.
Lindegaard, M. R., Bernasco, W., Jacques, S., & Zevenbergen, B. (2013). Posterior gains and
immediate pains: Offender emotions before, during and after robberies. In J. L. Van Gelder,
H. Elffers, D. Reynald, & D. Nagin (Eds.), Affect and cognition in criminal decision making:
Between rational choices and lapses of self-control (pp. 58–76). New York: Routledge.
Liu, L., Wang, X., Eck, J., & Liang, J. (2005). Simulating crime events and crime patterns in a RA/
CA model. Reading, PA: Idea Publishing.
Malleson, N., Evans, A., & Jenkins, T. (2009). An agent-based model of burglary. Environment and
Planning B: Planning and Design, 36, 1103–1123.
Marchione, E., Johnson, S. D., & Wilson, A. (2014). Modelling maritime piracy: A spatial
approach. Journal of Artificial Societies and Social Simulation, 17(2), 9. http://jasss.soc.surrey.
ac.uk/17/2/9.htnl.
Melo, A., Belchior, M., & Furtado, V. (2006). Analyzing police patrol routes by simulating the
physical reorganization of agents. In J. S. Sichman & L. Antunes (Eds.), Proceedings of the
6th International Workshop on Multi-Agent-Based Simulation (pp. 99–114). Berlin:
Springer.
Reynald, D. M. (2011). Guarding against crime: Measuring guardianship within routine activity
theory. Farnham, UK: Ashgate.
Schelling, T. (1987). Micro motives and macro behaviour. New York: W.W. Norton & Co.
244 C. Gerritsen and H. Elffers

Townsley, M. K., & Birks, D. J. (2008). Building better crime simulations: Systematic replication
and the introduction of incremental complexity. Journal of Experimental Criminology, 4(3),
309–333.
Townsley, M., & Johnson, S. (2008). The need for systematic replication and tests of validity in
simulation. In L. Liu & J. Eck (Eds.), Artificial crime analysis systems: Using computer
simulations and geographic information systems (pp. 1–18). Hershey, PA: Information Science
Reference.
van Baal, P. (2004). Computer simulations of criminal deterrence: From public policy to local
interaction to individual behavior. Den Haag, The Netherlands: BJU Boom Juridische
uitgevers.
van Bavel, M. (forthcoming). Do offenders heed guardians?
van Bavel, M., & Elffers, H. (2013). Experiments in guardianship research. In B. C. Welsh, A. A.
Braga, & G. J. N. Bruinsma (Eds.), Experimental criminology: Prospects for advancing sci-
ence and public policy (pp. 90–107). New York: Cambridge University Press.
Webley, P., Robben, H. S. J., Elffers, H., & Hessing, D. J. (1991). Tax evasion: An experimental
approach. Cambridge: Cambridge University Press.
Yang, R., Ford, B., Tambe, M., & Lemieux, A. (2014). Adaptive resource allocation for wildlife
protection against illegal poachers. In Proceedings of the 13th International Conference on
Autonomous Agents and Multiagent Systems, AAMAS (pp. 453–460). New York: ACM.
Part VI
Crime-Proofing Legislations
Chapter 16
Proofing Legislation Against Crime
as Situational Prevention Measure

Ernesto U. Savona

Introduction

The core idea underlying crime proofing is that legislation (intended in its broad
sense as all the norms that, at international, national and local level, regulate the
behaviours of citizens and organisations) may inadvertently produce the risk for
opportunities of crime.1
The first type of risk is non-compliance with certain regulations of the sector,
which include regulatory gaps or a lack of clarity that offers opportunities for crime.
There are different examples: the use of confusing terminology, the presence of
gaps in the regulation of a sector, an insufficient implementation of the regulation
(which creates differences among countries that can be exploited by criminals). The
second type of risks is side-effects, which “occur when legislation is created with a
certain purpose, but at the same time the legislation provides unwanted opportuni-
ties for crime”. These risks may occur when a sector is regulated too much or is
deregulated (Vander Beken & Balcaen, 2006, p. 321). Legislation may also create
risks if it “is ambiguous, provides additional opportunities for malfeasance, stokes
the motivation of persons whose positions expose them to such opportunities” (Levi
& Dorn, 2006, p. 253).

1
Different is the case of legislation that deliberately facilitates crime. It could happen that in heavy
criminal context legislators produce legislation with the aim of being easily exploited by
criminals.
E.U. Savona (*)
Università Cattolica del Sacro Cuore-Transcrime, Milano, Italy
e-mail: ernesto.savona@unicatt.it

© Springer International Publishing Switzerland 2017 247


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_16
248 E.U. Savona

Once established why and where this legislation produces such opportunities, the
prevention activity should consist in proofing it against crime. Framing it inside the
situational prevention approach means understanding the impact of the regulations
on real life (Russell & Clarke, 2006). This is not always easy to assess, considering
that the effects of the legislation could take time before they are seen.
This approach developed in the middle of 2000 in the framework of the EU
Commission, quite close to the Impact Assessment methodology, has been neglected
for a long time for different reasons that will be explained in this chapter.
The author believes that today, with the development of complex crimes, proof-
ing legislation against crime is a more relevant challenge than it was in the past and
could contribute to the reduction of these crimes.
Today the number of norms at different levels, and their complexity, is increas-
ing at international, national and local level. That means frequent overlapping at
the same level or at different levels of legislation. The complexity of these laws is
also an issue which is quite often addressed by scholars and policy-makers.
Recently the Italian National Authority against corruption declared that one of the
remedies against corruption is to simplify laws addressing vulnerable sectors,
such as procurement and construction (Amadore, 2015). As in many other instru-
ments of situational crime prevention, crime proofing of legislation targets spe-
cific crimes that violate trust, such as fraud and corruption. Exploiting loopholes
of the legislation means complying apparently with the norms on one side, and
intentionally violating their dispositions on the other. This could happen at indi-
vidual level or at a more systemic level when criminal organisations exploit the
loopholes of the legislation to achieve their own interests. In this way, these loop-
holes become opportunities for crime.

Crime Proofing of Legislation

These are the opportunities that crime proofing as an instrument of crime prevention
wants to reduce. It considers legislation for its criminogenic effects. Indeed, poorly
drafted legislation may offer unwanted illegal opportunities to criminals. If applied
to existing norms, this approach will lead to ex post crime proofing. When enacted
during the drafting of new legislation, it will be ex ante crime proofing. In order to
reduce such effects, crime proofing of legislation aims at pointing out existing (ex
post crime proofing) or future (ex ante crime proofing) opportunities for crime due to
legislation and at highlighting related interventions in order to proof it against crime.
In 2002, some authors underlined the necessity of introducing a formal crime
risk assessment into the legislative process (Albrecht & Kilching, 2002; Albrecht,
Kilching & Braun, 2002). Indeed, the EU established a steering committee in
2003 to advice on the development of a crime risk assessment process to be used
at the EU level. The Committee was assisted by some researchers, who produced
a literature review (Russell & Clarke, 2006). In 2006, Transcrime created an
initial methodology for proofing EU legislation (explained in the chapter),
proposed on the results of the European Union Crime Proofing Steering Group
16 Proofing Legislation Against Crime as Situational Prevention Measure 249

and of the Mechanisms for Assessing the Risk of Crime (MARC) project. 2This
methodology was initially developed by this project (Savona, 2006a; Savona,
Calderoni, Martocchia, & Montrasio, 2006) later tested through different case
studies in different fields (Savona, 2006b): corporate crimes (Levi & Dorn,
2006), businesses of insurance or private security (Dorn & Levi, 2006), interna-
tional banking (Curtol, Di Nicola, Martocchia, & Savona, 2006), mobile phones,
waste disposal, banking, pharmaceuticals (Vander Beken & Balcaen, 2006), pub-
lic procurement (Curtol, Pesarin, & Vander Beken, 2006) and waste management
(Dorn, Van Daele, & Vander Beken, 2007).
The crime proofing as it has been developed by the author and his colleagues
of Transcrime (Savona, Maggioni, Calderoni, & Martocchia, 2006b) consists of
two phases:
(a) Assessment of the risk that a legislative measure may produce unintended crim-
inal implications/consequences.
(b) Action to close the loopholes in the legislation, thereby “proofing” it against crime.
This process has four aims:
(a) To identify, if present, the unintended criminal implications/consequences of
existing or forthcoming legislation.
(b) To determine whether there is a crime risk, and if so, of what crime and of what
magnitude.
(c) To analyse pros and cons in terms of crimes arising from each policy option.
(d) To suggest textual changes to the legislation likely to reduce crime.

Crime Proofing in the European Union

The origin of crime proofing is almost lost among different initiatives around the
end of the past century: the Swedish Tax Legislation Offices (cited by Albrecht,
Kilchling, & Braun, 2002, p. 266) and different preliminary studies (Albrecht,
Kilchling, & Braun, 2002; Leppä, 1999). Its main development occurs when the
issue becomes politically relevant for the European Union. This issue was origi-
nally highlighted by the EU Institutions regarding possible actions to be taken to
monitor and control the infiltration of organised crime into the legitimate economy.
In 1980, the Gabert Report emphasised the relationships between legislation and
fraud, pointing out that complex laws may stimulate fraudulent behaviours and that
these behaviours were facilitated by asymmetries in the implementation of EU
policies (Gabert, 1984). After the Tampere European Council Presidency
Conclusions of 1999,3 which made broad statements on the need to integrate
aspects of crime prevention into legislation, several documents enacted by the EU

2
Special issue of the European Journal of Criminal Policy and Research, 12(3–4)
3
Conclusion No. 41 of the Tampere European Council, 15–16 October 1999.
250 E.U. Savona

Commission have emphasised the added value of a crime proofing strategy for
legislation (European Commission, 2000; Commission of the European
Communities, 2001a, 2001b, 2001c; European Crime Prevention Network, 2001,
p. 4). Establishing a crime proofing mechanism for the EU decision-making pro-
cess could have provided powerful instruments to prevent crime, by minimising
possible risks arising from the laws’ weakness, and therefore achieving greater
security and more opportunities for EU citizens. According to the European
Commission, “crime prevention includes all activities which contribute to halting
or reducing crime as a social phenomenon, both qualitatively and quantitatively,
either through permanent and structured co-operation measures or through ad hoc
initiatives. These activities are undertaken by all actors likely to play a preventive
role: local representatives, enforcement services and judicial system, social ser-
vices, education system, associations in the broad sense, industry, banks and the
private sector, research workers and scientists, and the general public, supported by
the media” (Commission of the European Communities, 2000, p. 7).
Several law reform projects, at international and national levels, may have an
impact on crimes. The screening of legislation for provisions that create or facilitate
opportunities for crime may significantly contribute towards minimising potential
crime risks and their implications in terms of costs and victims.
The importance of crime proofing in the EU crime prevention strategy was con-
firmed in the 2004 Assessment of the Tampere Programme, where the Commission
stated that “as regards crime prevention, efforts must be made to make goods and
services less vulnerable to crime. The crime-proofing of legislation must be an inte-
gral part of this effort” (Commission of the European Communities, 2004, p. 15).
For reasons that have not been clearly explained the issue of crime proofing has
been suddenly abandoned by the European Commission in the second part of the
2000. The author assumes that this was provoked by the conflicts between differ-
ent bureaucracies. DG Home4 launched the priority of crime proofing and was
authorised to intervene in all the policy sectors of the EU Commission that needed
new legislation such as Health or Environment. Pointing out crime implications in
future EU legislation could have scrambled the division of the labour and compe-
tences among different Directions giving to one (DG HOME) the authority to
modify incoming legislation. The problem was considered with the Impact
Assessment procedure, a mandatory process through which future EU legislation
should have been scrutinised (European Commission, 2009). Crime implications
were inserted in the social and environmental aspects—but quite often crime
aspects have not been considered by EU, as the case of Directive Illicit Tobacco
Markets addressed in the essay of Stefano Caneppele contained in this book.
Impact Assessment and Crime proofing could have been complementary if prop-
erly done. At that time there were all the elements for this action: Impact
Assessment for analysing the risk of crime connected to a forthcoming legislation
and crime proofing for reducing it modifying the laws.

4
Departments (Directorates-General) of Migrations and Home Affairs of the European
Commission. It prepares EU-level rules in these policy areas and watches over their application.
16 Proofing Legislation Against Crime as Situational Prevention Measure 251

Impact Assessment (IA) and Crime Proofing in the European


Context

Impact Assessment is “a set of logical steps to be followed when you prepare policy
proposals. It is a process that prepares evidence for political decision-makers on the
advantages and disadvantages of possible policy options by assessing their potential
impacts” (European Commission, 2009, p. 4).
The IA process aims at providing European Union policymakers with “sound
analysis supported by the best data available” (European Commission, 2009, p. 6).
It has been mainly designed to address the social, economic and environmental
impacts of envisaged proposals of the European Commission (European
Commission, 2009, p. 48). In this framework, crime issues are briefly considered.
The Commission’s Guidelines envisage the following questions to act as drivers
when addressing crime and terrorism issues related to different policy options
(European Commission, 2009, p. 36):
• “Does the option have an effect on security, crime or terrorism?
• Does the option affect the criminal’s chances of detection or his/her potential
gain from the crime?
• Is the option likely to increase the number of criminal acts?
• Does it affect law enforcement capacity?
• Will it have an impact on security interests?
• Will it have an impact on the right to liberty and security, the right to fair trial and
the right of defense?
• Does it affect the rights of victims of crime and witnesses?”
In some ways Impact Assessment if properly done through the inclusion of crime
issues could have been the right complement of the crime proofing exercise. If a
new law was in the process, assessing the risk of crime eventually produced by one
of the options considered could have been an alert signal given to regulators before
enactment. At the end, the decision if and how to modify the forthcoming legisla-
tion because of the risk of crime, could have been a more responsible decision. The
Impact Assessment procedure has become a routine activity at least inside the
European Commission and, for what is known to the author, the crime implications
of forthcoming legislation have been rarely considered at European level.

Assessing the Risk of Crime Related to Legislation and Proofing


It Consequently

Although not practised until now, Impact Assessment and Crime Proofing could
come together. Evaluating if past legislation has impacted upon crime or if the new
law will have an impact could help regulators to proof legislation against crime.
This could modify policy options (in the case of IA) and/or revise the form and the
252 E.U. Savona

content of the legislation in those parts that could produce opportunities for crime
when a law is enacted. How to merge the developing IA process with Crime
Proofing, which is almost neglected? What follows in this chapter is the tentative for
developing a methodology that combines the two. Reference should be done to the
manual Transcrime has developed in 2006 for the European Commission (Savona,
Maggioni, et al., 2006b).
As explained there are two phases: the risk assessment one and the proofing. The
Crime Risk Assessment process only aims at highlighting potential crime risks in
envisaged policy options that may affect the effectiveness of EU policies. The final
decision whether to adopt one policy option or another, implying different levels of
crime risk, will lie with the decision-making bodies that will evaluate the different
trade-offs between crime risks and the other impacts (social, economic and environ-
mental impacts) of the proposed action.

Assessing the Risk of Crime Due to Legislation

Every type of legislation may produce unintended criminal opportunities. In gen-


eral, when drafting norms, legislators focus on the main objectives of their piece of
legislation and do not consider the potential criminogenic effects of law. However,
such effects may be extremely relevant and may ultimately undermine the main
objectives of the law.
The following are examples that could be still not in place, that describe how
laws may inadvertently create crime risk.
Example 1: bearer shares legislation.
A bearer share is an instrument which contributes towards anonymity being
maintained inside the market, influencing the financial sector’s transparency. In
many countries their use has been restricted or prohibited, but in some cases they
are still allowed (for example, in Belgium bearer shares still exist, although there
are plans to ban them by 2007/2008). Bearer shares facilitate money laundering
activities, because they make it possible to convert illicit money into negotiable
and anonymous financial instruments. Moreover, as shares are controlling stakes
of corporations, bearer shares could be exploited to govern a company, whilst
remaining anonymous.
Example 2: shell companies.
Shell companies are companies which conduct either no business or minimal busi-
ness; their legitimate use is often to obtain financing prior to starting business
operations, or to minimise business and tax costs. Although legitimate, they are
often exploited to hide criminal activities, as they do not carry out any productive
activity but work as junctions for money transfers. Their use is enhanced in those
jurisdictions providing for anonymous banking and financial services, which allow
the identity of the beneficial owners behind the entities to be concealed. Such leg-
islations increase the opportunities for several criminal activities to be committed,
16 Proofing Legislation Against Crime as Situational Prevention Measure 253

ranging from fraud against creditors and tax crime to money laundering and the
financing of terrorism. All major financial scandal (such as Enron and Parmalat)
showed the involvement of shell companies in such massive frauds.
Example 3: suspicious transaction legislation.
According to anti-money laundering legislation, bank clerks have to report suspi-
cious transactions. However, in those areas where criminal organisations are pow-
erful and where intimidation or collusive behaviours are frequent, such a duty may
expose clerks to threats or violence. This was facilitated by weak legislation, not
providing effective measures to guarantee the anonymity of bank officials report-
ing suspicious operations. Such a loophole discouraged officials from reporting
and therefore facilitated the commission of money laundering and other illicit
operations.
Example 4: dividends-received deduction and criminal exploitation.
The dividends-received deduction system is meant to avoid double taxation of divi-
dends. Dividends that are already subject to taxation in the home country of the
company should not be taxed again in the country of the shareholder. According to
the Belgian Tax Code (articles 202 § 2 and 203), dividends paid to a domestic
company could, under certain conditions and within certain boundaries, be
deducted and excluded from taxation. There was an exception on this benefit for
dividends paid by a company that was established in a country in which the laws
on the payment of taxes was “seriously more advantageous” than in Belgium.
However, the wording “seriously more advantageous” was not further elaborated
by the legislator, which of course made it very difficult to know the exact scope of
application of the provision in the law.
Consequently, this system favoured a practice in which a company benefited
from the dividends-received deduction even for dividends that were barely taxed or
not taxed at all in foreign countries. Such ambiguity in the norm wording stimu-
lated a mechanism that finally led companies to evade taxation on dividends. As of
24 December 2002, this system was changed by law to prevent such abuses.
Example 5: turnover of officials.
Article 46 of Council Legislation (EC) no. 2157/2001, which creates a European
Company (Societas Europaea) fixes the term of office for members of the govern-
ing bodies. The rule explicitly provides for a term of 6 years, after which auditors
should quit their office, but it also says that “members may be reappointed once or
more than once for the period determined in accordance with paragraph 1”. The
consequence is that there is no effective duty of turn over: for example, an auditor
could be re-elected without limit. This loophole decreases the independence of
guardians, possibly leading to collusion and conflict of interests.
Example 6: controls in the financial system.
Art. 129 of the Italian Testo Unico Bancario (TUB) proved to be the weak point in
the Cirio case. Indeed, its unclear wording seemed to limit the powers of Bankitalia,
the Italian National Bank, to control only in the “primary market” of foreign bonds
(i.e. the market between the company issuing the bonds and the financial interme-
diaries/banks acquiring them). The secondary market (i.e. the market between
254 E.U. Savona

financial intermediaries/banks selling foreign bonds and private small savers)


seems to be outside the powers of control of Bankitalia. This uncertainty has been
exploited by banks in defrauding small savers by selling worthless bonds, since
personal operations with bank clients (“trattative personalizzate in relazione a stru-
menti finanziari richiesti dalla clientela”) were excluded from the competence of
Bankitalia.
Example 7: prohibitionism.
Prohibition strategies often produced criminal opportunities. Perhaps the most rel-
evant consequence of prohibitionism is the development of black markets. Many
documented cases demonstrate their generation following prohibition. The most
notable include prohibition of alcohol in the USA, whose consequence was an
enormous black market, fuelled by the smuggling of alcohol from neighbouring
countries, particularly Canada, and illegal production within the USA. Indeed, ille-
gitimate trade in prohibited and demanded products appears to be a foreseeable
consequence of prohibition, and it may be particularly extensive where significant
quantities of the product are available for diversion to illegitimate markets because
of local exemptions or because the product is legally available in neighbouring
jurisdictions. The unparalleled development of organised crime groups in this era
of prohibition has also been recognised.
Example 8: foot and mouth disease.
With the Animal Health Act of 1981 the UK established a compensation scheme
under which the owners of infected livestock are given compensation for the
destruction of their infected livestock. During the outbreak of Foot and Mouth
Disease (FMD), this mechanism has been allegedly exploited by those involved.
A range of allegations was received by government agencies, including the
deliberate infection of stock for the purposes of making claims, the making of
false and inflated claims and the commission of expenses fraud by administering
officials. The review highlighted weak controls as a contributory factor. Indeed,
farmers were allowed to select a valuer for their infected stock, a practice which
was considered to put “pressure” on valuers to increase their valuations. The
absence of independent and autonomous valuers was exploited to perform frauds
against the scheme.

How to Identify Unintended Crime Risks: A Step-Based


Checklist

The following checklist that has been initially prepared for European Union Officials
could be extended to other and different regulative experiences. It should be consid-
ered as an instrument through which Regulators could be able to explore possible
unintended criminal implications associated with legislation. Its purpose is to
explore whether a regulatory strategy contains any crime risk and, if so, to separate
such risk into its components.
16 Proofing Legislation Against Crime as Situational Prevention Measure 255

The checklist is structured on three steps, which allows for a gradual assessment.
It has been developed to answer two questions: (a) Does legislation produce any
unintended crime risk? And (b) if so, how is this crime risk structured? The first two
steps of the checklist will provide information in order to answer question (a), while
the third step will answer question (b).
The checklist follows the proportionate analysis principle: after each step, only
those legislations which show a relevant crime risk will continue to the following
steps. This decision is taken by the officials that are answering the checklist. Crime
risks are relevant when the answers to the questions of the same step highlight
actual criminal implication requiring further assessment.

Structure of the Checklist

The checklist is structured as follows:


STEP 1—General Crime Risk Indicators
This step includes some specific regulatory measures likely to stimulate an inter-
est in carrying out illicit activities. They address either criminals’ motivation or
opportunities to exploit legislation.
If a policy option falls into one of these risk indicators, it will be assumed to be
a relevant crime risk. Step 2 is required.
STEP 2—Coherence of Legislation and Market Vulnerability to Crime
This part takes two aspects into account:
• The overall coherence of legislation, on the assumption that ambiguous and
inconsistent policy strategies may contribute to the fragmentation and complex-
ity of the overall regulatory framework, thus causing overlaps and differing
implementation at national level. Such shortcomings may be exploited to cir-
cumvent, misuse and bend legislation to the criminals’ economic interests.
• The vulnerability of the regulated market sector to crime. Market vulnerability
plays a crucial role in exploring possible unwanted effects, as the same policy
may have different impacts according to the specific market addressed. What is
deemed as a relevant risk in one market may be insignificant in another.

Example

The introduction of new taxes may be an incentive for non-compliant and


fraudulent behaviours in a sector already infiltrated by criminal activities or
characterised by weak controls. On the contrary, in a strongly regulated sec-
tor, with effective controls and heavy sanctions for abuses, operators would be
less willing to run the risk of being detected, in return for the profit arising
from avoiding taxes.
256 E.U. Savona

The vulnerability of a market sector derives from its attractiveness and its acces-
sibility to potential criminals.
1. Attractiveness: it can be defined as the profitability of a market for criminals
minus the risk they run of being detected and punished:
ATTRACTIVENESS = PROFITABILITY − RISK OF DETECTION
(a) Profitability: it concerns the profit/gain/benefit a criminal could obtain by
committing an illicit activity, motivation.
(b) Risk of detection: it concerns the risk to criminals of being detected and
punished. The higher the risk of detection, the lower the attractiveness, and
hence the market is less vulnerable to crime.
2. Accessibility: it concerns the possibility of a market being entered by crimi-
nals, i.e. the presence of obstacles or barriers raised by the legislator to prevent
criminals from accessing the market. The more a market is provided with bar-
riers against access by illegitimate operators, the less vulnerable it is to crime.
Of course, barriers may produce negative effects on the market, but this issue
concerns the evaluation of other impacts produced by the policies which should
be considered in balance, as a trade-off with the crime risk assessment.
At the end of this assessment if any relevant crime risk has been envisaged the
policy option will pass to the third step.
STEP 3—Crime Components
The previous step discovers if any crime risk is associated with the policy option.
In case of a positive answer this step analyses the envisaged risk, entering into the
details of its components and exploring its magnitude. Criminological expertise is
required to answer the related questionnaire.

The Three-Step Checklist

The following paragraph contains the question in capital letters. Examples are pre-
sented in boxes.
STEP 1. General Crime Risk Indicators
QUESTION 1.1. DOES THE LEGISLATION INTRODUCE A NEW OR
MORE BURDENSOME OBLIGATION?
The introduction of new or heavier burdens increases the risk of non-compliant
behaviours, usually carried out through illicit activities. Waste disposal services, for
example, are characterised by burdens and costs finalised to reduce negative environ-
mental impacts linked to waste production. However, these burdensome obligations
are also an incentive to choose illegal channels for waste disposal. Waste producers
aim at minimising their costs, thus opting for waste firms which offer low costs,
without checking whether they are legitimate operators or criminal organisations, as
well as sometimes colluding with them. As waste disposal services are mainly pro-
vided by local firms, this favours criminal organisations, which act as unfair competi-
tors and create illegal oligopolies. Burdensome obligations may include: new
standards, new administrative requirements, additional procedures/bureaucracy.
16 Proofing Legislation Against Crime as Situational Prevention Measure 257

Example
Council Directive 2000/53/CE on End-Of-Life Vehicles promotes the recov-
ery and reuse of vehicles at the end of their life, by establishing a treatment
process which should improve the environmental performances of the eco-
nomic operators. This entails weighting the last owner of a vehicle with the
costs of its treatment and of a Certification of Destruction (COD). However,
such increased administrative burdens are likely to raise the rate of abandoned
vehicles and of illegal dumping activities.5

QUESTION 1.2. DOES THE LEGISLATION PROVIDE FOR TAX


ALLOWANCES AND/OR EXEMPTIONS?
Tax allowances stimulate avoidance behaviours, usually through deceptive
means for meeting the requirements to obtain the concession. Such measures may
include: tax relief, tax exemptions or tax deductions.

Example 1
Council Legislation 2007/2000/EC removed a tariff on sugar imported from
Serbia and Montenegro into the EU. This allowance was used to perform a
sort of “carousel fraud”: goods were being exported from the EU, using export
subsidies, and then freely re-imported from Serbia and Montenegro, finally
evading taxes.

Example 2
The EU duty suspension system was introduced as a part of the Internal
Market to improve trade between EU Member States (MSs). The unwanted
effect of this policy was the so-called “excise diversion funds” phenomenon.
According to this system, those trading excise goods and providing evidence
of their movement to other registered traders in the EU, or of their export
outside the EU, can have the payment of the excise duty suspended. This
would be payable whether the goods were moved to an unregistered trader or
for consumption. Such movements, together with the status of goods, were
recorded through the Accompanying Administrative Document (AAD).
However this mechanism has been abused to obtain the suspension: compa-
nies stimulated the trade of excise goods, which were distributed on the mar-
ket and not in a registered country, or exported and then returned them as
duty-free goods to the home market.

5
The case studies reported hereinafter are taken from Russell M., Clarke R., Government
Regulations and their Unintended Consequences for Crime: a Project to Develop Risk Indicators,
London 2003, report prepared by the Jill Dando Institute and Transcrime for the EU Crime
Proofing Steering Group.
258 E.U. Savona

These examples show how tax relief can increase motivation to commit crime.
Inadequate vetting of warehouses and standards as proof of movement, as well as
improper controls and inspections (reduced after the introduction of the duty sus-
pension system) were identified as added factors which favoured such fraud.
QUESTION 1.3. DOES THE LEGISLATION INTRODUCE BENEFITS SUCH
AS GRANTS, SUBSIDIES AND COMPENSATION SCHEMES?
Grants, subsidies, compensation schemes and benefits can often be an incentive
to adopt deceptive means to obtain such profits. For example differences in the level
of compensation for different benefits may lead to deceptive behaviour in order to
receive the most profitable benefit.
Fraudulent behaviours are the most likely offences stimulated by such measures,
potentially facilitated by corrupting the public officials in charge with the proceed-
ings. Such measures may include: subsidies, grants and compensation schemes.

Example 1
The UK foot and mouth compensation scheme increased the motivation for
committing fraud, exploiting the weak controls in the sector. The system pro-
vides compensation to the owners of infected livestock when they declare that
these have been destroyed. Some farmers exploited this compensation scheme
by various means: false or inflated claims, deliberate infection of their live-
stock to make claims and colluding with administering officials in defrauding
the government.

Example 2
The agriculture sector offers many examples of fraudulent behaviour aiming
to exploit government benefits, e.g. for organic food and farming. The
European Union promotes the development of organic farming, through tar-
geted funds. However, it was often found that false organic products are traded
as organic, as well as these funds being diverted toward non-organic
farming.

QUESTION 1.4. DOES THE LEGISLATION INTRODUCE TAX ON LEGAL


GOODS/SERVICES, OR INCREASE THE COST OF LEGAL GOODS/
SERVICES?
Taxation is traditionally associated with avoidance behaviours, which may also
include fraud. When taxation involves legal goods or services, increasing its cost,
parallel markets and smuggling are likely to be established, often taking advantage
of the existing differences between national tax systems. Fraud and corruption may
also occur, facilitating illicit trafficking. Such measures may include: taxes, customs
duties and excise.
16 Proofing Legislation Against Crime as Situational Prevention Measure 259

Example
Since the 1990s, the UK has introduced high excise on tobacco, aiming to
discourage tobacco consumption. However, the negative impact has been an
increase in tobacco smuggling, involving other European countries, in par-
ticular from the Netherlands, which are gates into UK from the continent for
illegal trade.6

QUESTION 1.5. DOES THE LEGISLATION PROHIBIT OR LIMIT THE


PRODUCTION AND/OR THE DISTRIBUTION OF A DEMANDED GOOD/
SERVICE?

Example
During the twentieth century countries produced laws prohibiting many goods
and services belonging to the category of “vice”, such as alcohol, drugs, pros-
titution, gambling. In each of these cases the experience shows that when one
of these goods/services is prohibited a correspondent illegal market emerges
to respond to the existing demand. The recent debate about the legalisation of
marihuana has been oriented to the goal of reducing the amount of crime con-
nected with the illegal traffic of this substance.

When limits to the distribution/production of a good are established, the demand


is higher than the supply. This stimulates the creation of black markets, where
illegitimate operators act as unfair competitors. Smuggling of goods, together with
counterfeiting and trademarks violations, are the main offences linked to such poli-
cies. Examples of limits to the production/distribution are: limits to export/import,
introduction of legal monopolies, more stringent trademark legislation, new stan-
dards of quality/labels and increases to the cost of production factors.
QUESTION 1.6. DOES THE LEGISLATION REMOVEALAW ENFORCEMENT
CAPACITY OR DECREASE FUNDING FOR LAW ENFORCEMENT ACTIVITY
OR IN ANY OTHER WAY WEAKEN A LAW ENFORCEMENT ACTIVITY?
The assumption is that a decrease in law enforcement activity is likely to produce a
consequent increase in the behaviour subject to enforcement. This happens either when
funding of the law enforcement agencies is reduced or when their effective competen-
cies and powers are weakened in some way. It should be pointed out that, even when a
law enforcement capacity has been improved, possible unwanted effects can be envis-
aged: for instance displacement of illicit activity may occur towards less controlled
sectors. Removal of a law enforcement capacity may include: decreased funds, moving
funds from one sector to another, reduction in powers and competencies.

6
To this purpose see the related study: Van Duyne P.C., Organizing cigarette smuggling and policy
making, ending up in smoke, in “Crime, Law and Social Change”, no. 39, 2003, pp. 285-317.
260 E.U. Savona

Example 1
The introduction of federal jurisdiction over tobacco smuggling in the USA
facilitated an increase in tobacco smuggling, due to a decreased involvement
of State law enforcement agencies. Indeed, the centralisation of law enforce-
ment activity reduced the effectiveness of controls.

Example 2
Unintended negative impacts have been associated to the strengthening of
USA–Mexico border enforcement in the 1990s. As organised smuggling
groups perceived a higher difficulty in moving between the countries, an
increase of illegal immigrants in the USA was registered, together with
border-crossing attempts in less protected areas. Another consequence was
the increased number of crossing deaths, due to the adoption of more clandes-
tine methods of transit from one country to the other.

QUESTION 1.7. DOES THE LEGISLATION INCREASE THE


DISCRETIONARY POWER OF OFFICIALS OR PROVIDE THEM WITH NEW
DISCRETIONARY POWERS?
Discretionary power entails taking decisions or actions without a substantial control
upon their advisability. Such powers usually concern the allocation of benefits or the
imposition of burdens, which increase the interest in corrupting officials. The main risks
associated with the empowerment of administering officials are a lower visibility due to
increased discretion, a higher risk of conflict of interests, the involvement of a lower
number of officials or a too chaotic ramification of competencies. Corruption is the most
likely offence arising from such legislation, linked with fraudulent activities aimed at
manipulating the final decisions. Examples of discretionary power are: possibility to
take decisions without need of motivating them, decisions/actions not subject to super-
vision of independent control body, decisions without predetermined guidelines.

Example
The number of goods classifications used by customs has been deemed to be
a risk factor for corruptive practices. The large number of possible classifica-
tions, which allows for different interpretations, favours the discretion of cus-
toms officials, e.g. giving them the power to classify a good into a lower-tax
category.

If at least one of the above questions is answered “YES”, step 2 is required.


STEP 2. Coherence of Legislation and Market Vulnerability to Crime—1.
Coherence of the Legislation
16 Proofing Legislation Against Crime as Situational Prevention Measure 261

QUESTION 2.1.1. IS THE LEGISLATION LIKELY TO PRODUCE


REGULATIVE OVERLAPS?
Every new legislative measure produces a series of impacts which go beyond the
specific regulated sector. Such interplays should be considered when drafting a pro-
posal, as unintended contradictions or mere overlaps with similar legislation may
occur. It might lead to the exploitation of these shortcomings in the whole regula-
tory framework to the purpose of bending the law for private gains. Indicators of
regulative overlaps are: excess of acts regulating the same field, contradictions
among the provisions of the same act or conflict between goals and norms.

Example
The recent Italian scandals, in particular the Parmalat and the Banca Popolare
Italiana (BPI) cases, highlight an example of regulatory uncertainty. In the lat-
ter scandal the BPI was investigated to find whether it acted illegally in attempt-
ing to buy its competitor Antonveneta and whether any illegitimate help of the
ex-governor of Bankitalia could be found. Regulatory overlaps have been dis-
covered in the identification of Bankitalia and Consob’s competencies. The for-
mer should control banking activities; the latter is in charge of monitoring the
stock exchange market. Both of them played a role in the Antonveneta business,
but the borders of their duties and responsibilities were not clear. Such a gap,
making the respective competencies uncertain, poses a great risk of the market
being compromised along with the involved operators.

QUESTION 2.1.2. IS THE LEGISLATION EASILY APPLICABLE WITHIN


THE JURISDICTION?
Laws should be implemented by national governments without causing ineffi-
ciencies or undesired effects. Possible consequences are, apart from delays in the
implementation, a fragmentary compliance and a messy regulatory framework. The
risk of law misuse and exploitation may occur accordingly. Indicators of difficult
applicability of legislation are: tight deadlines, high costs and long and complex
requirements and adaptation.

Example
The lack of a proper implementing structure and mechanism, as well as the
provision of unrealistic deadlines, may hamper Member States of the European
Union in the correct application of new measures. The same effect may be
produced when national budgets are insufficient or inadequate for meeting the
requests of the EU legislator.

QUESTION 2.1.3. IS THE LEGISLATION LIKELY TO PRODUCE


REGULATORY ASYMMETRIES AMONG DIFFERENT JURISDICTIONS?
Regulative asymmetries are one of the main factors associated with circumven-
tion of legislation. They may depend both on how the legislation is conceived and
262 E.U. Savona

on how it is drafted (e.g. a too generic content prone to many interpretations, a


vague language, problems in translation) and on unwanted effects due to the differ-
ent national legal systems (e.g. delocalisation or centralisation of specific compe-
tencies, the different balance between private and public levels, the welfare system).
Law and jurisdiction shopping, not illegal itself, may actually hide deceptive behav-
iours and illegitimate purposes. They are generally associated with tax and duties
evasion, but often work as an incentive for fraudulent conducts or as facilitators for
illegal practices linked to the business environment.
Factors likely to create asymmetries are: generic and ambiguous terms, norms
prone to different implementations, difficulties in translation.

Example
Off-shore jurisdictions are used not only to bypass tax burdens, but also as
privileged channels for the activities of organised crime groups. Money laun-
dering, financing of terrorism and trafficking of legal and illegal goods exploit
the lack of transparency offered by these legal systems. Asymmetries between
on-shore and off-shore jurisdictions are an incentive to choose less transpar-
ent channels, which offer anonymous banking services and profitable busi-
ness solutions.7

STEP 2. Coherence of Legislation and Market Vulnerability to Crime—2.


Market Vulnerability to Crime
QUESTION 2.2.1. IS THE MARKEt alREADY INFILTRATED BY
ECONOMIC AND ORGANISED CRIME? IF SO, WHICH CRIMES?
This is an indicator of market vulnerability: a market where criminal activities
are frequent (such as the tobacco market and the public procurement sector or waste
disposal market) is more exposed than others to further abuses.

Example
Criminals have preferred sectors for different reasons.8 They become vulner-
able when these markets are in areas where the presence of organised crime is
relevant. The existence of other crimes, such as extortion, organised crime-
related homicides, fraud and corruption, could be a good indicator of the
vulnerability of these markets.

7
To deepen the question see Transcrime, Euroshore-Protecting the EU Financial System from the
Exploitation of Financial Centres and Offshore Facilities by Organised Crime, Transcrime, Trento,
January 2000.
8
To deepen the example, see chapter “Organised crime investments in Europe” in Savona Ernesto
U. & Riccardi Michele (Eds.). 2015. From illegal markets to legitimate businesses: the portfolio of
organised crime in Europe. Final Report of Project OCP—Organised Crime Portfolio (www.
ocportfolio.eu). Trento: Transcrime—Università degli Studi di Trento.
16 Proofing Legislation Against Crime as Situational Prevention Measure 263

QUESTION 2.2.2. IS THE MARKET/SECTOR PROFITABLE FOR


CRIMINALS?
Profitability of a given market in relation to crime refers to the gains/profits a
criminal could obtain by carrying out illicit activities. It can be assessed by estimat-
ing how much added value the component of a market, i.e. production factors (capi-
tal, raw material, labour), product and structure (competitive, monopolistic, or
oligopolistic) would produce when subject to a criminal transaction. Such measures
may include: high economic value of the product, monopolistic structure, high
value of production factors, high demand and low supply.

Example
The fashion and pharmaceutical markets are profitable for counterfeiting
because of the high value of trademarks;
The waste market is profitable for criminal organisations, whose profits are
gained by avoiding the costs of the legitimate market. The tobacco market is
profitable for smuggling because of the high taxes levied on cigarettes.

QUESTION 2.2.3. IS THE MARKET PROVIDED WITH SYSTEMS/


STRUCTURES TO DETECT CRIMINALS?
This concerns the risk to criminals of being detected and consequently punished.
The main assumption is that the greater the likelihood that a criminal will be detected,
the less the likelihood that s/he will be willing to commit a crime. Therefore, a high risk
of detection reduces the attractiveness of a market to crime. Controls and security strat-
egies are the necessary instruments to increase the risk of detection; these may include:
mechanisms for the identification of legal/natural persons operating in the market, dis-
closure procedures, internal auditing to monitor managing powers, independent super-
vision authorities provided with clear and effective powers, requirements of
professionalism and probity for guardians and inspections and cross-checks.

Example
The waste disposal market is characterised by low visibility and weak con-
trols, due to the difficulty in monitoring waste movement from the producers
to the final site. In Italy this happens because of the local dimension of waste
disposal services connected to criminal oligopolies, which are able to prevent
effective controls. Paper controls, i.e. documents certifying the nature and
destination of waste during transit, contribute towards weakening the whole
system, by being exposed to falsification and abuse.

QUESTION 2.2.4. IS THE MARKET ACCESSIBLE TO CRIMINALS?


Accessibility of a market to crime means the objective capability of illegitimate new-
comers to infiltrate a given market. It depends on the obstacles and barriers raised by
legislation against illicit behaviours. The main assumption is that the more obstacles/
264 E.U. Savona

barriers rose against illegitimate entry into the market, the less it will be accessible to
criminals. Such measures may prevent companies with bad or less reliable finances
(thus more inclined to bypass legitimate channels) from accessing to the market. Barriers
may be of different natures, ranging from: background checks, standards of profession-
alism/competence/honesty, economic thresholds, administrative authorisations etc.

Example
Public Procurement legislations usually fix an economic threshold for compa-
nies which intend to compete for tenders. The minimum capital required aims
at selecting participants, on the assumption that possession of a high capital
should ensure high reliability. However, if the legislation does not apply also
to subcontractors the goal of the provision is defeated. Such a loophole may
be exploited by criminal organisations, which would be able to bypass the
economic barrier.

STEP 3. Crime Components—1. Crime


QUESTION 3.1.1. WILL THE AMOUNT OF CRIME VARY?
This refers to the increase in the number of crimes that could be indirectly
induced by the new legislation.

Example
Laws prohibiting or regulating sectors in between legal and illegal markets,
such as drugs and prostitution, could produce other indirect crimes beyond the
crimes of drug trafficking or prostitution, such as robberies, burglaries,
violence.

QUESTION 3.1.2. HOW WILL THE RISK OF BEING DETECTED


COMMITTING A CRIME VARY?
This refers to the expected effects of the considered legislation on the likelihood
of being detected (e.g. detection, and/or identification of the author and/or report to
law enforcement agencies and/or prosecuted). The risk of being detected is one of
the major deterring factors for criminals.

Example
Thanks to technological improvements, many new laws introduced electronic
controls in sensible sectors/markets (e.g. transports, food production). The
replacement of paper documents with electronic documents with appropriate
anti-forgery devices increase the risk of being detected for criminals and
should therefore deter them from forge certificates and documents.
16 Proofing Legislation Against Crime as Situational Prevention Measure 265

QUESTION 3.1.3. HOW WILL THE EXPECTED PROFIT FOR THE


CRIMINALS VARY?
This refers to the expected effects of the considered legislation on profits obtained
through fraud (i.e. reduction of costs and/or increase of revenues). If appropriate,
consider also non-monetary forms of profit.

Example
Legislation introducing expensive burdens or requirements on market opera-
tor may render very profitable a non-complying or falsely complying behav-
iour by enterprises.

STEP 3. CRIME COMPONENTS—2. Authors


QUESTION 3.2.1. WILL THE NUMBER OF AUTHORS VARY?
This refers to the expected effects of the considered legislation on the overall
number of criminals (reported and unreported). Laws that reduce controls or
increase the potential profits deriving from criminal activities will likely increase
the number of criminals. However, an increase of the amount of crimes may be
caused by an increase of the number of crimes per author.

Example
If a law regulating customs duties will create opportunities for fraud, it could
be possible that the number of fraudsters will increase. The same thing hap-
pens for corruption. When the opportunities for corruption increase due to a
law it could happen that in an environment affected by systemic corruption
more people take the risk (which seems low) of being corrupted or corrupt.

QUESTION 3.2.2. HOW WILL THE COMPLEXITY OF THE


ORGANISATIONAL STRUCTURE OF A CRIME VARY?
This refers to the expected effects of the considered legislation on the level of
organisation required to commit a fraud. In particular, lawmakers should assess if
the considered policy option/main action will affect the minimum number of crimi-
nals required to commit a fraud and/or the need of a coordinated organisation.
The complexity of the organisation required is a relevant element in the crimi-
nals’ choice to commit a crime. Indeed co-ordination of different agents as well as
recruitment and management of many people is particularly complex for illegal
organisation, since they cannot rely on legal instruments of argument settlement (no
access to contracts, tribunals, law enforcement system) in the illegal arena. The
more complex the organisation of frauds and other crimes become, the more diffi-
cult will be for criminal to commit such crimes.
266 E.U. Savona

Example
A piece of legislation introduces controls on the quality of food. Consequently,
it regulates the production processes and requires such processes to be certi-
fied by competent authorities. These provisions will indirectly increase the
organisational requirements to commit crimes in the food sector. Indeed,
criminal may have to bribe national authorities and/or forge the required cer-
tifications. Since such activities do not fall under the standard food production
activities, they will probably need the involvement of third persons in the
criminal activity, so increasing the complexity of the organisation.

QUESTION 3.2.3. HOW WILL THE INDIVIDUAL SKILLS/KNOWLEDGE


REQUIRED TO COMMIT A CRIME VARY?
This refers to the expected effects of the considered legislation on the overall
skills and knowledge needed to commit a fraud. If a sector requires considerable
skills and knowledge to operate, it will be less accessible to criminals.

Example
Electronic documents require higher skills to be forged than mere paper docu-
ments. Consequently, their introduction may increase the skills required to
commit forgery and other crimes.

QUESTION 3.2.4. FOR NATURAL PERSONS, HOW WILL THE


PROFESSIONAL REQUIREMENTS NEEDED TO COMMIT A CRIME OR
FACILITATING THE COMMISSION OF A CRIME VARY?
This refers to the expected effects of the considered legislation on the profes-
sions (accountants, lawyers, doctors, engineers etc.). Many roles have specific pro-
fessional requirements or other certifications in order to operate in a market/sector.

Example
Annual financial reports must be certified as true by a registered accountant,
who is recognised by her peers and through a professional body.

QUESTION 3.2.5. FOR LEGAL PERSONS, HOW WILL THE ECONOMIC/


LEGAL REQUIREMENTS NEEDED TO COMMIT A CRIME VARY?
This refers to the expected effects of the considered legislation on the economic
and/or legal requirements actually needed to commit a crime. Such requirements
may be expressly imposed by the law or ensue from the structure of the targeted
16 Proofing Legislation Against Crime as Situational Prevention Measure 267

market/sector. These issues may affect illegal enterprises/activities just as well as


the affect legal activities. Therefore it is important to investigate the impact of a
proposed legislation on the structure of the enterprises operating in the sector.

Example
In order to ensure appropriate guarantees for the customers and the solidity of
the banking system, banks have to fulfil with a series of requirements in order
to operate with the public, such as capital reserves. Any change in such
requirements will consequently modify the structure of the banking sector.
These modifications may impact on the sector in a positive (e.g. increasing
concurrence and transparency of the banking system) or negative way (e.g.
facilitating frauds or increasing the opacity of the market), thus affecting the
economic/legal requirements to commit credit frauds and other similar crimes.

STEP 3. CRIME COMPONENTS—3. Victims


QUESTION 3.3.1. HOW WILL THE AMOUNT OF VICTIMS (NATURAL
PERSONS) VARY?
This refers to the expected effects of the considered legislation on the overall
number of victimised people. The amount of victimised people is a very relevant
issue concerning the impact of a potential crime risk.

Example
After the occurrence of a disaster, emergency legislation allocating funds for
reconstruction is usually enacted, whose vulnerabilities are exploited
by organised crime in order to divert the aids to their convenience. On 23
November 1980 an earthquake struck Irpinia (Southern Italy) causing 2735
deaths. The emergency relief law was exploited by organised crime (camorra).
The level of criminal infiltration in the building industry rose, the buildings
were not reconstructed or were reconstructed late and/or badly, and other
crimes (killings, corruption and arson) were committed.

QUESTION 3.3.2. HOW WILL THE SOCIO-DEMOGRAPHIC


CHARACTERISTICS OF VICTIMS (NATURAL PERSONS) VARY?
This refers to the expected effects of the considered legislation on the main
socio-demographic (e.g. educational level, sex, age, ethical origins) characteristics
of the victims. This question aims at assessing any variation in the characteristics of
the victims of a crime. This issue may be relevant even if the overall amount of
expected victims is not likely to increase.
268 E.U. Savona

Example
The introduction of anti-pollution norms in Italy has force many enterprises to
move their activities in countries with more permissive environmental legisla-
tion. E.g. an enterprises producing aluminium coffee machines in Northern
Italy moved its production plants in Eastern Europe, where anti-pollution laws
are less strict and could easily be fooled. As a consequence, the people living in
such countries became victims of possibly illegal pollution and had to sustain
higher health care costs in countries with worse the public health systems.

QUESTION 3.3.3. WILL THE AMOUNT OF VICTIMS (LEGAL PERSONS)


VARY?
This refers to the expected effects of the considered legislation on the overall
number of victimised legal persons (e.g. profit/non-profit organisations, state and
governmental agencies).

Example
A change in industrial property legislation may not specify in a clear way
when a company’s trademark is imitating another registered trademark. Such
a piece of legislation would likely increase the number of companies suffering
from illicit trademark imitation.

QUESTION 3.3.4. HOW WILL THE ECONOMIC/LEGAL


CHARACTERISTICS OF VICTIMS (LEGAL PERSONS) VARY?
This refers to the expected effects of the considered legislation on the economic
(e.g. corporate size, national or multinational companies) legal (e.g. corporate struc-
ture) characteristics of the victims (legal persons).

Example
Legislation of the credit sector may require specific pre-existing assets in order
to obtain fund from banks or other sponsor. This mechanism may push newborn
enterprises to avoid the legal credit sector and resort to the usury market.

STEP 3. CRIME COMPONENTS—4. Costs/Harm


QUESTION 3.4.1. WILL THE TOTAL COSTS OF THE CRIME VARY?
An assessment of the economic impact of the proposed legislation on the
whole society will provide a very useful tool of comparison between envisaged
policy options.
16 Proofing Legislation Against Crime as Situational Prevention Measure 269

Example
Legislation regulating vulnerable markets should take in account direct and indirect
costs of crime, including the cost of anticipation and reaction usually not considered
by many policy options. In the case of prohibition of health-damaging substances
the emergence of an illegal market produces more crime and direct and indirect
costs for citizens that could be reduced increasing the costs of law enforcement.

QUESTION 3.4.2. HOW WILL PRIVATE COSTS VARY?


This refers to the expected effects of the considered legislation on the direct costs
suffered by victims. The assessment of such impact may be very relevant when
choosing the optimal policy option.

Example
Legislation decreasing controls on the pharmaceutical sector may affect the
product reliability and safety. This may impact on the cost suffered by victims
of frauds and counterfeiting of pharmaceutical products, since they may face
higher health care costs due to unexpected effect of medicines.

QUESTION 3.4.3. HOW WILL SOCIAL COSTS VARY?


This refers to the expected effects of the considered legislation on any non-
private cost, such as costs imposed on a whole sector/area or on society.

Example
Frauds against the EU common agricultural policy directly affect the EU and
MSs budget. However, these crimes also hinder the ultimate purposes of the
common agricultural policy, subtracting fund from deserving commercial
farms. The social costs include the losses and inefficiencies caused by this
diversion of funds from their original and legitimate goal.

The Crime Risk Assessment Procedure

The following diagram represents a procedure to assess the risk of crime related to a
forthcoming legislation that could have multiple options related to a single law or many
laws aiming at the same goal. The CRA process starts with the Initial Screening which
frames the laws or the policy options into seven different general risk indicators (out-
lined before at The Three-Step Checklist—STEP 1. General Crime Risk Indicators); if
the prospected legislation is found to be at risk, the procedure should go to the
Preliminary Crime Risk Assessment (PCRA), which assesses the potentially crimino-
genic policy options. If the PCRA highlights at least a medium level crime risk, an
Extended Crime Risk Assessment (ECRA), which will assess more deeply the crime
risks connected to the legislation and/or its options, will be recommended. The ques-
tions outlined before at step 1, 2, 3 are the guidelines of this procedure.
The following diagram illustrates the main steps of the CRA process:
270 E.U. Savona

Initial Screening (IS): does the


Step 1 proposal fall into any typology of the NO
regulation at risk?

YES

Preliminary Crime Risk Assessment


(PCRA): are there unintended criminal No further
Step 2 NO
activity required
risks arising from the envisaged policy
options?

YES

Extended Crime Risk Assessment


Step 3 (ECRA): assessment of the risks
outlined in step 2.

Proofing the Legislation in Practice9

The three steps outlined before could be adapted according to the typology of proof-
ing activity, mainly if it is ex post or ex ante. It is possible to know that a given
market is highly vulnerable because of its attractiveness and because of the strong
presence of organised crime, then the analysis of laws could come almost immedi-
ately looking at the characteristics that make this legislation at risk of producing
opportunities for crime. This is the specific activity of proofing the legislation.
In a recent experience this approach was applied by Transcrime in different
areas, in particular on a specific Directive and in a specific geographical and limited
area (the province of Trento in Italy). In this case the analysis was developed through
two steps. The former classifying those sectors considered at crime risk, and the
second looking at those laws that were regulating the different sectors under
observation.
Among the sectors considered were procurement, construction, transportation,
that in the Italian reality present some risks of being exploited by criminals and in
the Province of Trento were found as vulnerable (Transcrime, 2013).
For each of these sectors the following activities were developed:
1. Literature analysis and case studies, mainly drawn from judicial sources, of how
organised crime has exploited these sectors.

9
The methodology of the CPL has been applied, directly or indirectly by Transcrime in different
areas: evaluation of crime risk implications (Savona, Maggioni, Calderoni, & Martocchia, 2006a),
procurement (Curtol, Pesarin, et al., 2006), international banking (Curtol, Di Nicola, Martocchia, &
Savona, 2006) and waste management (Dorn et al., 2007). Recently, the CPL has also been applied
to the tobacco sector (Calderoni, Savona, & Solmi, 2012; Caneppele, Savona, & Aziani, 2013).
16 Proofing Legislation Against Crime as Situational Prevention Measure 271

2. The analysis of legislation regulating these sectors one by one.


3. Recommendations on how and where to reduce the vulnerability of the legisla-
tion (proofing it against crime).
The three levels of analysis were:
• CONSISTENCY
This step checks whether there are contradictions among the different sources
of legislation (national versus local). The assumption is that a legislation that
presents contradictions and overlapping of different sources could produce con-
fusion in the implementation of the law, becoming vulnerable to its exploitation
by criminals.
The following qualitative indicators were considered:

1. Volume of the legislation How many laws are regulating the same sector?
2. Internal Organisation Has the legislation an obscure and inconsistent structure (i.e. not
clear partitions and not logically articulated)
3. Conflicts and Are the single dispositions of the legislation conflicting among
contradictions in the them and/or conflicting with the general legislation of the sector?
same legal framework

• CLARITY
This step refers to the language used by texts of law. Generic and/or too wide and
ambiguous concepts could create misunderstandings with different typologies of
implementation. The assumption is that the less the clarity of the text and its
ambiguity, the more it could be exploited for criminal purposes.
The following qualitative indicators were considered:

4. Objectives Clarity Are the objectives of the legislation clearly understandable or


ambiguous?
5. Risk of misinterpretations Are there concepts and dispositions that could have different
and/or conflicting interpretations?
6. Definitions Is the legislation lacking definitions of concepts and/or key
words?

• CONTROL MECHANISM
This step evaluates whether the law provides adequate control mechanisms. The
assumption is that the less these mechanisms are present, the more the legislation
could be exploited for criminal purposes
The following qualitative indicators were considered:

7. Control Mechanisms Does the legislation provide mechanisms to control its correct
implementation?
For each of these seven indicators it was developed a qualitative analysis and given a
qualitative score following a scale from the Lowest vulnerability to the Highest: High–Low;
Low; Medium; Medium–High; High.
272 E.U. Savona

Conclusions

In the different applications of the crime proofing methodology there has been a
wide room for subjective interpretations by those who have implemented it. In the
three-step methodology that has been used for the European Union Directives the
evaluation has been done by a panel of experts in the sector of the legislation. The
initial idea of setting measurable indicators has been abandoned in favour of more
explorative and qualitative evaluations on the laws, on the vulnerability of the mar-
ket under legislation and its attractiveness. This approach developed by a panel of
experts without setting up benchmarks could be developed in favour of a more
quantitative and measurable approach with measurable indicators. This depends on
the sector considered and data available about crime environment. For example the
European and national legislation of procurement is a framework for single laws
that are produced and implemented at local level. There are areas inside the same
country where the risk of exploitation is different due to the presence of crime envi-
ronment and to socio-economic conditions that make this environment attractive to
crime. The same thing could occur for the construction industry that is an attractive
sector for organised crime infiltration in the legitimate businesses and its implemen-
tation could produce asymmetries, both among countries and inside the same coun-
try, more related to demand and supply mechanisms. This necessary flexibility in
evaluating whether past or new legislation could produce opportunities for crime
requests a panel of experts knowing areas and sectors that follow established guide-
lines rather than a quantitative assessment.
As for many applications of the situational prevention techniques proofing legis-
lation against crime is a methodology that varies across different sectors and areas
regulated by laws. It could be improved with more and more applications in differ-
ent sectors, areas and countries with different legislative systems. At least, if widely
practised, crime proofing could make policy makers aware of the risk that their
action could produce criminal consequences.

References

Albrecht, H.J., Kilching M. (2002). Crime Risk Assessment, Legislation, and the Prevention of
Serious Crime - Comparative Perspectives. European Journal of Crime, Criminal Law and
Criminal Justice, 10(1), 23–28.
Albrecht, H.-J., Kilchling, M., & Braun, E. (Eds.). (2002). Criminal preventive risk assessment in
the lawmaking procedure. Criminological Research Reports from the Max Planck Institute for
Foreign and International Criminal Law, Vol. 102, Freiburg i. Br.
Amadore, N. (2015). Raffaele Cantone: “Il diktat? Semplificare”. Press & Imprese, 2, 3–7.
Calderoni, F., Savona, E. U., & Solmi, S. (2012). Crime proofing the policy options for the revision
of the tobacco products directive. Transcrime—Joint Research Centre on Transnational Crime.
Caneppele, S., Savona, E. U., & Aziani, A. (2013). Crime proofing of the new tobacco products
directive. Transcrime—Joint Research Centre on Transnational Crime.
16 Proofing Legislation Against Crime as Situational Prevention Measure 273

Commission of the European Communities. (2000). Communication from the commission to the
European Parliament. The Prevention of Crime in the European Union. Reflection on Common
Guidelines and Proposals for Community Financial Support (No. COM (2000) 786). Brussels:
Commission of the European Communities.
Commission of the European Communities. (2001a). Communication concerning the Fraud proof-
ing of legislation and contract management (No. SEC (2001) 2029). Brussels: Commission of
the European Communities.
Commission of the European Communities. (2001b). COMMUNICATION FROM THE
COMMISSION TO THE COUNCIL AND THE EUROPEAN PARLIAMENT. Biannual update
of the scoreboard to review progress on the creation of an area of “freedom, security and jus-
tice” in the European Union. Brussels: Commission of the European Communities.
Commission of the European Communities. (2001c). Joint Report from commission services and
EUROPOL. Towards a European strategy to prevent organised crime (No. SEC(2001) 433).
Brussels: Commission of the European Communities.
Commission of the European Communities. (2004). Communication from the Commission to
Council and the European parliament, area of freedom, security and justice: Assessment of the
tampere programme and future orientations (No. COM (2004) 401). Brussels: Commission of
the European Communities.
Curtol, F., Di Nicola, A., Martocchia, S., & Savona, E. U. (2006). Regulation of offshore banking services
and financial entities. European Journal on Criminal Policy and Research, 12(3-4), 279–298.
Curtol, F., Pesarin, G., & Vander Beken, T. (2006). Testing the mechanism on EU public procure-
ment legislation. European Journal on Criminal Policy and Research, 12(3-4), 337–364.
Dorn, N., & Levi, M. (2006). Regulation of insurance and corporate security: Integrating crime
and terrorism seriousness into the analysis. European Journal of Criminal Policy and Research,
12(3-4), 257–277.
Dorn, N., Van Daele, S., & Vander Beken, T. (2007). Reducing vulnerabilities to crime of the
European waste management industry: The research base and the prospects for policy.
European Journal of Crime, Criminal Law and Criminal Justice, 15(1), 23–36.
European Commission. (2000). Communication from the commission to the council and the
european parliament the prevention of crime in the European Union Reflection on common
guidelines and proposals for Community financial support (No. Doc. 2000/0304 (CNS)).
Brussels: European Commission.
European Commission. (2009). Impact Assessment Guidelines (No. SEC (2009) 92). European
Commission.
Gabert, V. (1984). European Parliament Working Documents 1983-1984, Report on behalf of the
Budgetary Control of Fraud against the Community Budget (No. Document 1-1346/83).
European Parliament. Retrieved from http://aei.pitt.edu/61485/1/B1953.pdf
Leppä, S. (1999). Anticipating instead of preventing: Using the potential of crime risk assessment
in order to minimize the risks of organized and other types of crime. Helsinki.
Levi, M., & Dorn, N. (2006). Regulation and corporate crime: Managers and auditors. European
Journal of Criminal Policy and Research, 12(3-4), 229–255.
European Crime Prevention Network. (2001). First programme: Setting priorities and objectives
for the period July 2001—December 2002. Stockholm: European Crime Prevention Network.
Russell, M., & Clarke, R. (2006). Legislation and unintended consequences for crime. European
Journal of Criminal Policy and Research, 12(3-4), 189–211.
Savona, E. U. (2006a). Initial methodology for the crime proofing of new or amended legislation
at the EU level. European Journal of Criminal Policy and Research, 12(3-4), 221–228.
Savona, E. U. (2006b). The Crime Risk Assessment Mechanism (CRAM) for Proofing EU and
National Legislation against Crime. European Commission—Transcrime.
Savona, E. U., Calderoni, F., Martocchia, S., & Montrasio, S. (2006). Finalising the crime risk
assessment mechanism for the crime proofing activities of European legislation/regulation.
European Journal of Criminal Policy and Research, 12(3-4), 365–397.
Savona, E. U., Maggioni, M., Calderoni, F., & Martocchia, S. (2006a). A study on crime proof-
ing—evaluation of crime risk implications of the European commission’s proposals covering a
274 E.U. Savona

range of policy areas (Contract No. DG.JAI-D2/2004/05). Retrieved from http://www.tran-


scrime.it/wp-content/uploads/2013/11/Final_Manual-A_study_on_Crime_Proofing.pdf
Savona, E. U., Maggioni, M., Calderoni, F., & Martocchia, S. (2006b). A Study on Crime Proofing –
Evaluation of Crime Risk Implications of the European Commission’s Proposals Covering a
Range of Policy Areas. MANUAL. Retrieved from http://www.transcrime.it/wp-content/
uploads/2013/11/Final_Manual-A_study_on_Crime_Proofing.pdf
Transcrime. (2013). Crime proofing the policy options for the revision of the Tobacco Products
Directive. Provincia Autonoma di Trento.
Vander Beken, T., & Balcaen, A. (2006). Crime opportunities provided by legislation in market
sectors: Mobile phones, waste disposal, banking, pharmaceuticals. European Journal of
Criminal Policy and Research, 12(3-4), 299–323.
Chapter 17
Crime Proofing of Legislation in Practice:
The Case of the 2014 EU Tobacco Products
Directive

Stefano Caneppele

Introduction

Crime proofing of legislation (CPL) can be a valuable tool for crime prevention.
CPL makes policymakers aware of the unintended criminal opportunities that regu-
lation may generate. CPL can be applied to the forthcoming legislation (ex ante
evaluation) or to the legislation in force (ex post evaluation).
This chapter shows how CPL operates in practice. The case study regards the
new EU Tobacco Products Directive which has been finally adopted by the European
Council on the 14th March 2014. Transcrime applied the crime proofing methodol-
ogy to the draft version (submitted to the EU Parliament) in order to understand
whether the new regulation may unintentionally have generated criminal opportuni-
ties. The exercise indicated that the preliminary provisions relating to a ban of men-
thol and slim cigarettes in the EU Market: (a) could have increased the illicit tobacco
market by between 11.2 and 17 % for slim cigarettes and by 2.9 and 5.1 % for men-
thol cigarettes; (b) could have reduced the EU MSs tax revenue to between €1 and
more than €2 billion due to the illicit sales of menthol and slim cigarettes. Eventually,
the final version of the Tobacco Products Directive did not include the ban for slim
cigarettes.

S. Caneppele (*)
School of Criminal Sciences, University of Lausanne, Lausanne, Switzerland
e-mail: stefano.caneppele@unil.ch

© Springer International Publishing Switzerland 2017 275


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_17
276 S. Caneppele

Criminal Opportunities in the Tobacco Market: The Illicit


Trade in Tobacco Products

Criminal opportunities in the tobacco market refer to the illicit trade of tobacco
products (ITTP). The European Commission defines illicit trade as ‘any practice or
conduct prohibited by law and which relates to production, shipment, receipt, pos-
session, distribution, sale or purchase, including any practice or conduct intended to
facilitate such activity’ (European Commission, 2014, see art. 1 of the WHO
Framework convention on tobacco control). The ITTP comprises various illicit
activities (Calderoni, Savona, & Solmi, 2012) such as:

• Smuggling/contraband: The unlawful movement or transportation (including the


online sale) of tobacco products (genuine or counterfeit) from one tax jurisdic-
tion to another without the payment of applicable taxes or in breach of laws
prohibiting their import or export ‘contraband’.
• Counterfeiting: The Illegal manufacturing of a product which bears or imitates a
trademark without the owner’s consent.
• Cheap Whites, or Illicit Whites: Cigarettes produced legally in one country but
normally intended for smuggling into countries where there is no prior legal
market for them. Taxes in production countries are normally paid, while they are
avoided/evaded in destination countries.
• Bootlegging: Legally buying tobacco in a low-tax country and illegally reselling
it in a high-tax country.
• Illegal manufacturing: Cigarettes manufactured for consumption which are not
declared to the tax authorities.
• Unbranded tobacco: Manufactured, semi-manufactured and even loose leaves of
tobacco carrying neither labeling nor health warnings. It may be sold by weight
and consumed in roll-your-own cigarettes or in empty cigarette tubes (‘chop-
chop’ in some countries) or sold in the form of loose cigarettes in large plastic
bags (‘baggies’).

The EU Tobacco Products Directive

On the 19th December 2012, DG SANCO presented the proposal for a revision of
the TPD Directive and its accompanying impact assessment. The main objective
of the revision was ‘to improve the functioning of the internal market, while
ensuring a high level of health protection’ (European Commission, 2012a, p. 46,
2012b, p. 49). The impact assessment suggested preferred policy options in five
policy areas (European Commission, 2012a): (1) smokeless tobacco products
(STP) and extension of the product scope (i.e. nicotine-containing products (NCP)
and herbal products for smoking), (2) packaging and labeling, (3) ingredients/
17 Crime Proofing of Legislation in Practice… 277

additives, (4) cross-border distance sales and (5) traceability and security features
(European Commission, 2012a, 2). The proposal for a revision of the Tobacco
Products Directive (COM (2012) 788 final) was presented on the 19th December
2012. It received all the preferred policy options suggested in the impact assess-
ment. Table 17.1 summarises the main elements of the proposal which could have
changed the TPD currently in force.

Applying the Crime Proofing of Legislation to the TPD

CPL works through a crime risk assessment mechanism (CRAM). CRAM starts
with an initial screening (IS), it continues through a Preliminary Risk Assessment
(PRA) and it ends with an Extended Risk Assessment (ERA). Each step narrows
down the attention on the most at risk part of regulation which may generate crimi-
nal opportunities.

Initial Screening

Initial Screening (IS) is the first step in the Crime Risk Assessment (CRA) process.
Its aim is to select which part of the TPD should undergo a CRA process. Namely,
the IS scans all the articles and it identifies those which fall into CPL risk indicators
((1). Fee/Obligation; (2). Concession; (3). Grant, Subsidy or Compensation Scheme;
(4). Tax or Cost; (5). Availability Restriction; (6). Law Enforcement; (7). Regulatory
Power).
Overall, the IS highlighted that most of the articles of the TPD proposal fell
within the risk indicators concerning fees and obligations, imposing taxes or costs
or restricting the availability of the tobacco products (see Table 17.2). In other
words, most of the innovations recommended were considered at risk by the first
step of the crime proofing methodology. Since they impose fee/obligation or tax/
cost or restrict the availability of tobacco products, those measures could have gen-
erated opportunities for crime. Therefore it was necessary to further investigate
through the Preliminary Crime Risk Assessment (PCRA) the possible magnitude of
those crime risks.

Preliminary Crime Risk Assessment

The Preliminary Crime Risk Assessment (PCRA) analyses the vulnerability of the
tobacco market to crime and discusses possible crime risks arising from the preferred
policy options.
278 S. Caneppele

Table 17.1 TPD proposal for revision (December 2012): main changes
TITLE II—Tobacco products
Ingredients and Reporting of ingredients and emissions (Art. 5)
emissions (Chapter Manufacturers are obliged to electronically report ingredients (glossary)
I) of tobacco products in accordance with a common format and to provide
(Art. 3-6) supporting data (e.g. marketing reports).
Regulation of ingredients (Art. 6)
Prohibition of the use of additives in tobacco products which contain:
(a) vitamins or similar, or (b) caffeine and taurine or other stimulant
compounds, (c) colouring properties for emissions.
Prohibition of the use of flavourings in the components of tobacco
products such as filters, papers, packages, capsules or any technical
features allowing modification of flavour or smoke intensity (prohibition
of menthol cigarettes).
Tobacco products other than FMC, RYO and STP (i.e. cigars, cigarillos
and pipes) are exempted.
Labeling and Combined health warnings for tobacco for smoking (Art. 9)
Packaging Each unit packet and any outside packaging of tobacco for smoking shall
(Chapter II) carry combined health warnings which shall: (a) be comprised of a text
(Art. 7–14) warning and a corresponding colour photograph specified in the picture
library, (b) include smoking cessation, (c) cover 75 % of the external
area of both the front and back surface of the unit packet and any outside
packaging
Product Description (Art. 12)
The labeling shall not include any element or feature that: (a) promotes a
tobacco product by means that are false, misleading, deceptive or likely to
create an erroneous impression about its characteristics, health effects,
hazards or emissions; (b) suggests that a particular tobacco product is less
harmful than others or has health or social effects; (c) refers to flavour,
taste, any flavourings or other additives or the absence thereof; (d)
resembles a food product.
Cigarettes with a diameter of less than 7.5 mm shall be deemed to be
misleading (prohibition of slim cigarettes).
Traceability and Security Features (Art. 14)
EU tracking and tracing system. An EU tracking and tracing system at
packet level for tobacco products throughout the supply chain (excluding
retail) is introduced. Tobacco manufacturers shall conclude contracts with
independent third parties that provide data storage capacities for such
system ensuring full transparency and accessibility by member states at all
times (par. 6).
Security features. Adoption of security features against counterfeiting
and against illicit/cheap whites on all tobacco products (e.g. holograms).
Tobacco for oral Art. 15
use (Chapter III) No changes compared to the current TPD
(continued)
17 Crime Proofing of Legislation in Practice… 279

Table 17.1 (continued)


TITLE II—Tobacco products
Cross-border Art. 16
distance sales of Notification and age verification system. Retailers of tobacco products
tobacco products engaged in cross-border distance sales shall notify their cross-border
(Chapter IV) activities to the member states where the company has its headquarters
and where it intends to sell. Member states may require the retailer to
appoint a natural person, who ensures compliance with the TPD of
products delivered to customers in the Member States concerned. A
mandatory age verification system is foreseen.
Novel tobacco Art. 17
products (NTP) Duty of notification, labeling and ingredients requirements.
(Chapter V) Manufacturers and importers of tobacco products have to notify the
competent authorities of Member States of any novel tobacco product they
intend to place on the markets of the MS concerned. In addition NTP
placed on the market must respect the rules on labeling (health warnings
on both sides) and ingredients regulation (ban on products with
characterising flavours).
TITLE III—Non-
Tobacco Products
Nicotine- Art. 18
Containing Subject NCP above a certain nicotine threshold to the legislation on
Products (NCP) medicinal products
Subject NCP below a certain nicotine threshold to labeling requirements
Herbal Products Art. 19
for Smoking Mandatory health warnings
Under Article 14, TPD draft refers to delegated for (a) defining the key elements (such as duration,
renewability, expertise required, confidentiality) of the contract referred to in par. 6, including its
regular monitoring and evaluation; (b) defining the technical standards to ensure that the systems
used for the unique identifiers and the related functions are fully compatible with each other across
the Union and (c) defining the technical standards for the security feature and their possible rota-
tion and to adapt them to scientific, market and technical development

Vulnerability of the European Tobacco Market

The analysis of the vulnerability of the European tobacco market is based on two
dimensions. The first is how the tobacco market is attractive to crime (Table 17.3).
The second relates to the accessibility of the tobacco market to criminals (Table
17.4). Since this analysis has been carried out in our recent study (Calderoni et al.,
2012), this section provides only a summary of and an update on the results.

Crime Risks Arising from Specific Main Actions

Tables 17.5, 17.6, 17.7 and 17.8 combine, for each TPD section, the envisaged
crime risks and their probability. The probability assessment has been carried out
relying on secondary sources that could support the assessment.
280 S. Caneppele

Table 17.2 Measures of the TPD proposal falling within the seven risk indicators of the initial
screening
Risk indicators 1. Fee/Obligation
2. Concession 3. Grant, Subsidy
or Compensation Scheme 4. Tax
or Cost 5. Availability Restriction
Legal elements of 6. Law Enforcement 7.
the proposal Articles of the TPD proposal Regulatory Power
1 2 3 4 5 6 7
Ingredients and Mandatory reporting in • •
emissions (Art. 3–6) harmonised format (Art. 5)
Ban on characterizing flavours, •
including menthol (Art. 6)
Labeling and Mandatory enlarged picture • •
packaging (Art. warnings (Art. 9)
7–14) Impose stricter labeling (Art. 11) • •
regulation for smokeless tobacco
products
Ban of misleading colours, • •
symbols and slim FMC (Art. 12)
EU tracking and tracing system • •
and security features (Art. 14)
Cross-border Notification and age verification • •
distance sale of system (Art. 16)
tobacco products
(Art. 16)
Novel tobacco Duty of notification, stricter • • •
products (Art. 17) labeling and ingredient
regulation (Art. 17)
Nicotine-containing Subject NCP above a certain • •
products (Art. 18) nicotine threshold to the
medicinal products legislation,
and the remaining NCP to
labeling requirements (Art. 18)
Herbal products Health warnings (Art. 19) • •
(Art. 19)

Results of the Preliminary Crime Risk Assessment

The results of the preliminary crime risk assessment (PCRA) highlighted a high
risk that banning menthol and banning slim cigarettes may increase ITTP crimes
(Table 17.9). Both of them were included in the extended crime risk assessment.
While some of the new measures have a low ITTP Crime Risk Probability, the
PCRA did not determine the ITTP Crime Risk Probability for other provisions since
there was still a lack of scientific evidence and/or a lack of clarity on the implemen-
tation of the measures through delegated acts.
17 Crime Proofing of Legislation in Practice… 281

Table 17.3 Attractiveness of the tobacco market to crime


Levels of crime The ITTP is a large EU ITTP grew from 8.5 % of total
component of the EU consumption in 2007 (Joossens, Merriman,
tobacco market and its Ross, & Raw, 2009, 10) to 10.5 % in 2013
magnitude is increasing (KPMG, 2014)
A wide variety of actors are involved in ITTP
(some legitimate manufacturers in the
tobacco industry, organised or terrorist
groups, and other criminals)
Profitability The ITTP is in general Tobacco products are heavily taxed in the EU
extremely profitable MSs (80 % mean tax share) (DG TAXUD
2011, 6)
Cigarettes are the commodity with the
highest fiscal value per weight (Joossens,
1998, 149–150)
Risk of The risk of detection in the Sheer size of the illicit market
detection commission of the ITTP is Difficult detection of the ITTP
very low International cooperation against the ITTP is
still weak
The policing of the ITTP seems rarely to be a
priority for national law enforcement
agencies
The stigma normally associated with
criminal activity does not always apply to the
ITTP

Table 17.4 Accessibility of the tobacco market to criminals


Violence and/or corruption • The ITTP also competes with the legal market for
(modus operandi) tobacco products; the levels of violence should be
particularly low, or otherwise customers would move
rapidly to the legal sector
• The ITTP is frequently related to bribery and corruption
Exploitability of factors, products • Given the extent of the illicit market and the significant
or structures of the tobacco demand for illicit tobacco products, the factors, products
market for crime and structures of the tobacco market are exploitable for
criminal purposes

Extended Crime Risk Assessment

The PCRA identified the banning of menthol and slim cigarettes as having a high
probability of creating unintended crime opportunities in the tobacco market. In
order to assess the possible impact on crime, perpetrators, victims and costs,
Transcrime carried out a collateral study (Caneppele, Savona, & Aziani, 2013).
Since both policy options adopted the same measure (ban), that could have pro-
duced similar effects (eliminating the availability of menthol and slim cigarettes),
the extended crime risk assessment will be carried out jointly.
282 S. Caneppele

Table 17.5 Crime risks arising from ingredients and emissions


Art. 5 Mandatory reporting in harmonised format
Envisaged crime risks Crime risk probability
Transfer of the increased LOW—According to the Commission, ‘the costs for
reporting costs to final consumers introducing such a system on a mandatory basis would be
may increase retail prices, thereby marginal and largely off-set by the savings generated by the
boosting the ITTP in EU member use of one single format across the EU’ (European
states Commission, 2012a, p. 99)
Art. 6 Ban on characterising flavours, including menthol
Envisaged crime risks Crime risk probability
Legitimate tobacco products may LOW—For candy and fruity flavours. The ban on
become less palatable than illicit ingredients such as candy and fruity flavours may have a
products for consumers, who may marginal impact on the ITTP since the market share of such
decide to buy illicit FMC with products is negligible
characterising flavours and HIGH—For menthol FMC. According to Euromonitor
products with increased toxicity Euromonitor International, 2013 data ((2013), menthol
or addictiveness (since there is no FMC represents more than 10 % of the legal tobacco
legal alternative); this may market in five EU MSs (Finland, Poland, Sweden,
increase the ITTP in EU member Denmark, Estonia). Studies have frequently reported that a
states ban on, or strict limitation of, menthol cigarettes may
considerably increase the illicit trade in countries where
these products are popular and a relatively large demand
already exists (United States Trade Representative, 2011,
Compass Lexecon, 2011). A sudden limitation on the
availability of menthol cigarettes might unintentionally
create opportunities for the ITTP

Ban on Menthol and Slim Cigarettes

Globally, no country has adopted a ban on slim cigarettes. In regard to menthol ciga-
rettes, only Brazil banned all flavours and additives in tobacco products (Framework
Convention Alliance, 2012), but the impact of this decision on ITTP is still unknown
because the Brazilian law will come into force by September 2013. On the other
hand, the USA initially considered banning menthol cigarettes, but then discarded
this option. According to the US Government, ‘the sudden withdrawal from the
market of products to which so many millions of people are addicted would be dan-
gerous. First, there could be significant health risks to many of these individuals.
Second, it is possible that our health care system would be overwhelmed by treat-
ment demands […]. Third, the agency also believes that, given the strength of the
addiction and the resulting difficulty of quitting tobacco use, a black market and
smuggling would develop to supply smokers with these products. It also seems
likely that any black market products would be even more dangerous than those
currently marketed, in that they could contain even higher levels of tar, nicotine, and
toxic additives’ (United States Trade Representative, 2010, 7).
Consequently, there is no information and data available on the actual impact of
banning menthol and/or slim cigarettes on the ITTP. Therefore Transcrime devel-
oped a study to estimate the impact of menthol and slim cigarette ban on the ITTP.
17 Crime Proofing of Legislation in Practice… 283

Table 17.6 Crime risks arising from labelling and packaging


Art. 9 Mandatory enlarged picture warnings
Envisaged crime risks Crime risk probability
Transfer of the increased LOW—According to the Commission, the EU harmonisation will
labeling production costs reduce the one-off costs, while ongoing costs will be higher. ‘Costs
to final consumers may could increase by 1.3–1.5 % following the introduction of EU-wide
increase retail prices, pictorial warnings […]’ with ‘an annual cost increase for tobacco
which may boost the companies between €59 million and €68 million’ (European
ITTP in EU member Commission, 2012a, p. 86)
states IMPOSSIBLE TO DETERMINE—The increased standardisation
Increased standardised may favour counterfeiting but, at present, it is impossible to
packaging may facilitate determine to what extent. Undoubtedly, the adoption of an effective
counterfeiting and thus EU tracking and tracing system and of security features will
increase the ITTP in EU contribute to reducing the crime risks (see Policy Area 5)
member states IMPOSSIBLE TO DETERMINE—At present, it is difficult to
Increased standardised assess the impact of mandatory enlarged warnings on brand loyalty
packaging may weaken and consumer behaviours. Roland Berger’s study (2013, 6)indicates
brand relevance, so that that ‘the number of cigarettes sold is estimated to fall by 1.6 % in
consumers may decide to the premium segment and 1.2 % in the below premium segment’
switch to cheaper with also a displacement toward the ITTP
cigarettes, even illicit
ones (the so-called
downtrading)
Art. 11 Imposing stricter labeling regulation for smokeless tobacco products
Envisaged Crime Risks Crime Risks Probability
Transfer of the increased LOW—The increased cost appears limited, and therefore so too
labeling production costs does the probability of these behaviours occurring
to the final consumer may
increase retail prices,
thereby boosting the ITTP
in some member states
Art. 12 Ban of misleading colours, symbols and slim FMC
Envisaged crime risks Crime risk probability
Legitimate tobacco IMPOSSIBLE TO DETERMINE—For any promotional and
products may become less misleading element (slim FMC excluded). The shift to illicit
recognisable and less products depends on the importance that EU consumers give to the
attractive than illicit brand (through appealing colours, symbols and slim FMC).
products for consumers, Eurobarometer estimated that 23 % of EU consumers consider the
who may decide to buy packaging important in their choice of cigarette brand. The
illicit FMC with percentage varies widely: from 58 % in Slovakia to 10 % in
misleading colours, Luxembourg (European Commission, 2012a). However, the
symbols, slim FMC definition ‘promotional and misleading element’ is too generic,
with the consequence that it is impossible to determine whether a
sudden ban might encourage EU smokers to buy illicit products
HIGH—For slim FMC. According to Euromonitor Euromonitor
International, 2013 data ((2013a), slim FMC represent more than
10 % of the legal tobacco market in several EU MSs (Bulgaria,
Romania, Poland, Latvia, Greece, Estonia, Cyprus, Hungary,
Lithuania). A sudden ban on slim FMC in EU countries where their
consumption is popular may increase the demand for illicit slim
FMC
(continued)
284 S. Caneppele

Table 17.6 (continued)


Art. 14 EU tracking and tracing system
Envisaged crime risks Crime risk probability
Transfer of the increased IMPOSSIBLE TO DETERMINE—The key issue concerns costs.
internal costs to final Unfortunately, the European Commission, on not providing an
consumers may increase overall figure, admits that ‘it is not easy to fully estimate the costs’
retail prices, which may (European Commission, 2012a, p. 108). Traceability may
boost the ITTP indubitably contribute to the reduction of smuggling, and
counterfeiting as well. According to the European Commission, the
EU T&T system reduces illicit contraband by up to 30 % in 5 years
(estimate based on industry data). However, it cannot be assumed
that the same trend will be observed in the next years since T&T
systems cover the market shares controlled by the four main
tobacco companies which signed bilateral agreements with the
European Commission. In most of the EU tobacco markets, this
share is around 95 % of the entire market (Euromonitor
International, 2013). More information on the actual
implementation of the EU T&T system is needed
Art. 14 Security features
Envisaged crime risks Crime risk probability
Transfer of the increased IMPOSSIBLE TO DETERMINE—The key issue concerns costs.
internal costs to final Since the European Commission will determine technical standards
consumers may increase for the security features through delegated acts, it is not possible to
retail prices, thereby quantify the additional costs. The sole estimate available refers to
boosting the ITTP the costs of applying holograms on all cigarette packages. These
costs would amount to approximately €150 million, which may be
balanced by a 10 % reduction of counterfeiting in 5 years
(European Commission, 2012a, p. 110). Security features may
indubitably contribute to the reduction of counterfeiting.
Nevertheless, there is still a lack of scientific evidence and clarity
on the type of security features to be adopted

The assessment is based on the abovementioned study (see Caneppele et al., 2013),
and on general criminological assumptions about the likely evolution of (a) crime
risks, (b) characteristics of perpetrators, (c) characteristics of victims (e.g. tobacco
growers, tobacco manufactures, tobacco retailers, tobacco consumers) and (d) crime
costs/harms.

Crime

Would the amount of risk vary? The study suggested that the ITTP risk was likely
to increase especially in those countries where menthol and slim cigarettes were
highly popular. There was evidence that a proportion of menthol and slim con-
sumers would have moved to the ITTP, since there would be no alternative. The
growing demand for menthol and slim cigarettes would have generated an illicit
supply of menthol and slim FMC. Transcrime study found that the increase in the
17 Crime Proofing of Legislation in Practice… 285

Table 17.7 Crime risks arising from cross-border distance sales


Notification of the member states where the company has its headquarters and where it
intends to sell (Art. 16)
Envisaged crime risks Crime risk probability
Transfer of the increased LOW—The percentage of European tobacco e-buyers appears to be
internal costs to final very small (from 0 to 1 %) and problems already exist in enforcing
consumer may increase the current TPD on Internet sales (European Commission, 2012a).
retail prices, which may On the contrary, the proposal may reduce criminal opportunities in
boost the ITTP the internet market, which is associated with tax evasion, by
authorising only registered companies
Mandatory age verification system (Art. 16)
Envisaged crime risks Crime risk probability
Transfer of the increased LOW—The percentage of European tobacco e-buyers appears to be
internal costs to final very small (from 0 to 1 %) and problems already exist in enforcing
consumers may increase the current TPD on internet sales (European Commission, 2012a).
retail prices, thereby On the contrary, the proposal may reduce criminal opportunities in
boosting the ITTP the internet market, which is associated with under-age purchasing

level of the ITTP could have ranged between 2.9 and 5.1 % as a consequence of
the ban on menthol cigarettes, and between 11.15 and 16.95 % because of the ban
on slim cigarettes. Therefore the total increase would have been between 14.0 and
22.1 %. Apparently, the most affected countries would have been those near the
Eastern EU border, where the ITTP was already flourishing (European Commission,
2011; OLAF, 2012; EurActiv.com, 2013). Transcrime study showed that if the
bans on menthol and slim cigarettes were adopted, 12 out of 28 EU member states
would have been seriously affected by the growth of the ITTP (Fig. 17.1). The
significant ITTP growth across Eastern EU borders may have jeopardised the
efforts made by the European Commission to counter the phenomenon (European
Commission, 2011).
How would the risk of detection vary while engaging in the ITTP? The probable
increase in the levels of ITTP as a reaction to the banning of menthol and slim FMC
was likely to impact on the risk of detection. EU seizures of tobacco products were
relatively stable since 2005, and amounted to approximately 7–8 % of the estimated
illicit market (OLAF, 2012). According to Transcrime estimates, the banning of
menthol and slim cigarettes would have generated between 8.2 and 17.8 billion
extra illicit cigarettes per year. Without an increase in law enforcement resources
and capacities, this could have meant a possible drop in the incidence of EU sei-
zures over the estimated illicit market up to 5–6 %. Indeed, in an overall context of
budgetary constraints, it would have been difficult to envisage an increase in law
enforcement resources to prevent and police the ITTP increase. Therefore, EU
member states would have been unable to tackle the increased risks of growth by the
ITTP with a comparable increase in law enforcement. The ITTP’s growth would
have been concentrated along the Eastern EU border, which may have generated
further problems in terms of corruption of border police and customs officers
(European Commission, 2011). As result, the risk of detection for criminals would
have declined.
286 S. Caneppele

Table 17.8 Crime risks arising from novel tobacco products, nicotine-containing products, herbal
products
Duty of notification, stricter ingredient and labeling regulation for novel tobacco
products (Art. 17)
Envisaged crime risks Crime risk probability
Novel tobacco products should LOW—The probability of these behaviours occurring appears
comply with duty of to be scarce, at least in the short run. In fact, advertising
notification and with the same limitations on tobacco products seem to prevent from
limits imposed for ingredients successfully launching new products on a large scale in a
and labeling regulation. This relatively short time
could prevent tobacco
manufacturers from releasing
in Europe new tobacco
products which could be
counterfeit and/or smuggled
from outside EU
Subject NCP above a certain nicotine threshold to the medicinal product legislation and
the remaining NCP to labeling requirements (Art. 18)
Envisaged crime risks Crime risk probability
Products above a certain IMPOSSIBLE TO DETERMINE—NCP (in particular
nicotine threshold, which e-cigarettes) represent an emerging and fluid market, not yet
would require medical regulated by strict rules, which grew out of the anti-smoking
authorisation, may be legislation. The existence of an illicit market seems to emerge
excluded from the legal from some police operations carried out in Italy which
market and therefore favour seized—in the first part of 2013—more than 11 thousand of
the growth of an ITTP market e-cig kits and 692 refillers which did not report, fully or
for NCP partially, the required labels (La Stampa, 2013). The growth of
ITTP crime risks will depend also on how many NCP products
are authorised. In the meantime, several European
governments decided to take action against the e-cig (see for
example (Briggs, 2013; Sparks, 2013)). For all those reasons
and owing to the absence of information, it is currently
impossible to determine the level of ITTP NCP risks created
by the proposed directive
Health warnings on herbal products(Art. 19)
Envisaged crime risks Crime risk probability
Transfer of the increased LOW—The probability of these behaviours occurring appears
labeling production costs to limited since the labeling cost is low
final consumer may increase
retail prices, thereby boosting
the ITTP in some member
states

How will the expected profits vary for the perpetrators of the ITTP? The tobacco
market has a dual nature because it consists of a legitimate and an illegitimate part.
The banning of menthol and slim cigarettes would remove only the legitimate sup-
ply. Profits for illicit traders would increase as a result of a monopolistic position
which would enable them to determine prices without considering the legal bench-
marks. According to Transcrime study, the yearly expected extra profits can be esti-
mated at between €0.4 and €1.3 billion.
17 Crime Proofing of Legislation in Practice… 287

Table 17.9 Results of the preliminary crime risk assessment


Illicit trade in tobacco products (ITTP)
Crime risk probability
High Low Impossible to determine
Art. 5 Mandatory reporting in
harmonised format
Art. 6 Ban on characterising Menthol Candy and
flavours, including menthol fruity flavours
Art. 9 Mandatory enlarged increasing of standardisation and
picture warnings retail prices brand loyalty
Art. 11 Imposing stricter
labeling regulation for
smokeless tobacco products
Art. 12 Ban of misleading Slim Others
colours, symbols and slim
FMC
Art. 14 EU tracking and
tracing system and security
features
Art. 16 Cross-border distance
sale
Art. 17 Novel tobacco products
Art. 18 Nicotine-containing
products
Art. 19 Herbal products

70%
60%
50%
40%
30%
20%
10%
0%
Czech Republic

Netherlands
Germany

Lithuania

Romania
Belgium
Bulgaria

Denmark
Estonia

Hungary

Slovakia
Slovenia
Finland

Poland
Portugal

Sweden
Austria

France

Greece

Ireland

Latvia

Spain
Italy

UK

Menthol (Average value) Slim

Fig. 17.1 Estimated increase of ITTP market resulting from the combined effect of the ban of
menthol and slim cigarettes. Source: Caneppele, Aziani and Savona (2013)
288 S. Caneppele

Consequently, the envisaged bans are likely to increase significantly the ITTP’s
expected profits, as a result of higher returns on both menthol and slim cigarettes.

Perpetrators

Would the number of perpetrators vary? Increased levels of ITTP were likely to
create new opportunities for getting involved. The overall number of people actively
involved in the trade would probably increase. According to the latest estimates the
current number of ITTP perpetrators could vary between 100,000 and 150,000 peo-
ple (Transcrime, 2015). However, it would not have been possible to provide a clear
estimation of the growth.
How would the complexity of the organisational structure of ITTP vary? In the
case of menthol cigarettes, the ban envisaged would have a low impact on the com-
plexity of the organisational structure required to engage in the ITTP, since the
manufacturing of menthol cigarettes is similar to that of non-mentholated cigarettes
and does not require any special equipment. As already pointed out, ‘menthol may
be added at any of the following stages; spraying onto the final blend, through addi-
tion to the filter via a thread, or by application to the cigarette paper or the foil used
to wrap the cigarettes’ (European Commission, 2012b, 15).
In the case of slim cigarettes, the ban envisaged would have a limited impact on
the complexity of the organisational structure required to engage in the ITTP, since
the manufacturing of slim cigarettes requires minimal technical adaptations to the
standard tobacco machineries. Indeed, criminal investigations proved that ITTP per-
petrators were already engaged in the market (Ahec, 2012).
How would the individual skills/knowledge required for the ITTP vary? The bans
envisaged would have a limited impact on the individual skills required to engage in
the ITTP, since the manufacturing of menthol cigarettes is similar to that of non-
mentholated cigarettes, and manufacturing slim cigarettes requires minimal techni-
cal adaptation to the standard tobacco machinery.

Victims

Would the number of victims (namely, consumers) vary? The risk of increased levels
of ITTP implies that the number of consumers of illicit tobacco products may
increase. The only studies available on consumers’ attitudes to the banning of slim
(in Romania) and menthol (in Poland and in Finland) cigarettes revealed a similar
pattern, although with some limitations1 (Coelho & Arink, 2013; Coelho & Moore,
2013). If the bans were approved, 42 % of slim and between 51 % (Poland) and 75 %
(Finland) of menthol consumers would buy their products on the illicit market. In
the case of menthol cigarettes alone, this would mean that approximately 10 % of

1
In both studies, the experiments gave only the alternative between buying licit or illicit slim/
menthol cigarettes. They did not consider the hypothesis that a ban would induce some consumers
to quit smoking.
17 Crime Proofing of Legislation in Practice… 289

Polish consumers would move to the illicit market. Moreover, legitimate market
operators, especially the retail sector, are likely to be seriously damaged by the
availability of illicit products, which cannot be sold in the legal market.
How would the socio-demographic characteristics of victims-consumers vary?
Logically, menthol and slim FMC consumers will be most affected by the bans.
Unfortunately, there are no available studies on the socio-demographic characteris-
tics of European menthol and slim cigarette smokers. Therefore, owing to the lack
of available studies in the field, it is currently impossible to determine how the
characteristics of victims would vary.
Would the amount of victims/legal persons vary? As already mentioned, upstream
and downstream distributors would have been affected by the increase in ITTP as
consequence of the menthol and slim ban, since a part of their income may be trans-
ferred to illicit suppliers. As already noted, the increase would have not been the
same in all the EU countries, so that a significant impact was expected only where
menthol and slim cigarettes were popular.

Costs/Harm

Would the total cost of the crime vary? Surprisingly, only one study focused also on
the impact that menthol and slim bans would have on the illicit side of the tobacco
market (RolandBerger, 2013). This study argued that a combination of menthol and
slim bans together with the pack standardisation measures would generate high job
losses (from 70,000 to 175,000) and a strong drop in tax revenue (from €2.2 billion
to €5.0 billion). Moreover, the study foresaw a strong increase in illicit trade (25–
55 %) which would entail higher costs for law enforcement staff and a decrease in
the expected contraction in smoking prevalence, as well as an increase in the health
risks for smokers because counterfeit cigarettes were more dangerous (Levinson,
2011). Transcrime study, focusing on the ban of menthol and slim cigarettes, argued
that it seemed likely that, at least in the short-medium term, the levels of ITTP
would significantly rise in several European countries to meet the increased demand
for menthol and slim tobacco products. It estimated that EU MSs would lose tax
revenues amounting to between €1.082 and €2.259 billion per year due to the illicit
sales of menthol and slim FMC.
How would private costs for victims vary? As already mentioned, some illicit
products, such as ‘illicit whites’ and counterfeits, were frequently reported to be
significantly more dangerous to smokers’ health (Center for Regulatory Effectiveness,
2011; Levinson, 2011). Therefore, the banning of menthol and slim cigarettes would
generate higher health costs. In addition, the increased ITTP would significantly
affect legitimate tobacco growers (Polskie Radio, 2013), tobacco manufacturers,
and the retail sector, with the risk of lower revenues, loss of employment and
reduced fiscal revenues.
How would social costs vary? The actual impact on social costs of the menthol
and slim ban would depend on the actual increase in the ITTP. According to the
RolandBerger (2013), which also accounted for packaging standardisation, ITTP
290 S. Caneppele

would increase by 25–55 %, so that social costs would be remarkable in terms of job
losses and lost tax revenues. Transcrime study estimated that a menthol and slim
ban could seriously impact on the European legitimate tobacco market with an ITTP
growth by 14.1–22.1 %. In particular, the social costs would not be homogeneously
distributed. Eastern countries and Poland in particular, would suffer disproportion-
ate costs compared with those of the other EU member states.

Results of the Extended Crime Risk Assessment

The ECRA on the banning of menthol and slim cigarettes was supported by a col-
lateral study (see Caneppele et al., 2013). Notwithstanding data and studies limita-
tions, the assessment on menthol and slim bans suggested that:
– A ban on menthol cigarettes would increase the level of the ITTP in Europe by
between 2.9 and 5.1 %.
– A ban on slim cigarettes would increase the level of the ITTP in Europe by
between 11.2 and 17.0 %.
– In Europe, the total ITTP increase would be between 14.1 and 22.1 %.
– The yearly expected extra profits for ITTP perpetrators could be estimated at
between €0.4 and €1.3 billion (for both menthol and slim cigarettes).
– EU MSs would lose yearly tax revenues amounting to between €1.1 and €2.3 bil-
lion due to the illicit sales of menthol and slim cigarettes.
– The social costs of ITTP crime would not be homogeneously distributed. Eastern
countries and Poland in particular would suffer disproportionate costs compared
with those of other EU member states.

Conclusions

This chapter presented a case study on crime proofing of legislation. The application
to the Tobacco products Directive proposal revealed that some provisions, namely
the ban on menthol and slim cigarettes, would significantly increase the crime risks
of an enlargement of the illicit market. Eventually, the European Parliament did not
vote for the ban of slim cigarettes which—according to the analysis—would play a
key role in increasing ITTP risks. On the other hand, the entry into force of the men-
thol ban was also planned only after a phase-out period of 4 years.
It is impossible to say whether and how much this study persuaded European regu-
lators. However, this exercise proved the relevance of the CPL methodology in crime
prevention. Unfortunately, the impact assessment studies conducted in preparation of
new EU directives did not consider properly the criminal impact. Generally, deficien-
cies in the social impact analysis (which include crime) have also been highlighted by
the Impact Assessment Board (IAB). The IAB observed in its most recent report that
‘[d]espite the Board’s previous recommendations to thoroughly assess social impacts
17 Crime Proofing of Legislation in Practice… 291

the Board notes that there has been no progress in the initial assessments of these
impacts’ (2013, 27). Incorporating CPL in the policy cycle would make policy mak-
ers aware of the unintended criminal opportunities. It would also support them and
their decisions, providing a wider set of information on crime risks.

References

Ahec. (2012, December 10). МВР започна ”коледно прочистване” с Тотко Поршето.
Retrieved from http://dnes.dir.bg/news/tzvetan-tzvetanov-kostinbrod-antimafioti-nelegalna-
fabrika-tzigari-totko-porsheto-12592796
Briggs, H. (2013, June 12). E-cigarettes Face New Restrictions. BBC. Retrieved from http://www.
bbc.co.uk/news/health-22870301.
Calderoni, F., Savona, E. U., & Solmi, S. (2012). Crime Proofing the Policy Options for the
Revision of the Tobacco Products Directive. Trento: Transcrime—Università degli Studi di
Trento. Retrieved from http://transcrime.cs.unitn.it/tc/fso/pubblicazioni/AP/Transcrime-CP_
of_the_EU_TPD.pdf
Caneppele, S., Savona, E. U., Aziani, A. (2013). Crime Proofing of the New Tobacco Products
Directive. Trento: Transcrime—Università degli Studi di Trento. Retrieved from http://tran-
scrime.cs.unitn.it/tc/fso/pubblicazioni/AP/TPD%20-%20crime%20proofing%20of%20
the%20new%20tpd_4.pdf
Center for Regulatory Effectiveness. (2011). Center for Regulatory Effectiveness: An Inquiry into
the Nature, Causes and Impacts of Contraband Cigarettes. Center for Regulatory
Effectiveness—Merrill Lynch Bank of America.
Coelho, M., & Arink, M. (2013). The Impact of a Menthol Ban on Illicit Trade in Poland. SKIM.
Coelho, M., & Moore, T.-L. (2013). The Impact of a Ban on Menthol Cigarettes on Illicit Trade in
Finland. SKIM.
Compass Lexecon. (2011). Estimating Consequences of a Ban on the Legal Sale of Menthol
Cigarettes. Retrieved from http://www.fda.gov/downloads/AdvisoryCommittees/Committees
MeetingMaterials/TobaccoProductsScientificAdvisoryCommittee/UCM243622.pdf
DG TAXUD. (2011, July). Excise Duty Tables: Part III—Manufactured Tobacco: REF 1033 rev.3.
Retrieved from http://ec.europa.eu/taxation_customs/resources/documents/taxation/excise_
duties/tobacco_products/rates/excise_duties-part_iii_tobacco_en.pdf
EurActiv.com. (2013, 12 February). Tobacco Smugglers Thrive on EU’s Eastern Border. Retrieved
from http://www.euractiv.com/europes-east/experts-seek-ways-curb-smuggling-news-517716
Euromonitor International. (2013). Tobacco Dataset. Euromonitor International.
European Commission. (2011). Commission Staff Working Paper: Action Plan to Fight Against
Smuggling of Cigarettes and Alcohol Along the EU Eastern Border SEC(2011) 791 final.
European Commission. Retrieved from http://ec.europa.eu/anti_fraud/documents/preventing-
fraud-documents/eastern_border_action_plan_en.pdf
European Commission. (2012a). Commission Staff Working Document. Impact Assessment
Accompanying the Document: Proposal for a Directive of the European Parliament and of the
Council on the Approximation of the Laws, Regulations and Administrative Provisions of the
Member States Concerning the Manufacture, Presentation and Sale of Tobacco and Related
Products. Brussels. Retrieved from http://ec.europa.eu/health/tobacco/docs/com_2012_
788_ia_en.pdf
European Commission. (2012b). Proposal for a Directive of the European Parliament and of the
Council on the Approximation of the Laws, Regulations and Administrative Provisions of the
Member States Concerning the Manufacture, Presentation and Sale of Tobacco and Related
Products., Pub. L. No. COM(2012) 788 final (2012). Retrieved from http://ec.europa.eu/health/
tobacco/docs/com_2012_788_en.pdf
292 S. Caneppele

European Commission. (2014). Directive 2014/40/EU of the European Parliament and of the
Council of 3 April 2014 on the Approximation of the Laws, Regulations and Administrative
Provisions of the Member States Concerning the Manufacture, Presentation and Sale of
Tobacco and Related Products and Repealing Directive 2001/37/EC. Retrieved from http://
ec.europa.eu/health/tobacco/docs/dir_201440_en.pdf
Framework Convention Alliance. (2012, 23 March). Brazil Bans Flavoured Tobacco. Retrieved
from http://www.fctc.org/index.php?option=com_content&view=article&id=728:brazil-bans-
flavoured-tobacco&catid=233:product-regulation&Itemid=237
Impact Assessment Board. (2013). Impact Assessment Board Report—Report for the year 2012.
European Commission. Retrieved from http://ec.europa.eu/governance/impact/key_docs/docs/
iab_report_2012_en.pdf
Joossens, L. (1998). Tobacco Smuggling: An Optimal Policy Approach. In I. Abedian, R. van der
Merwe, N. Wilkins, & P. Jha (Eds.), The Economics of Tobacco Control: Towards an Optimal
Policy Mix (pp. 146–154). Cape Town: Applied Fiscal Research Centre, University of Cape
Town.
Joossens, L., Merriman, D., Ross, H., & Raw, M. (2009). How Eliminating the Global Illicit
Cigarette Trade Would Increase Tax Revenue and Save Lives. Paris, France: International
Union Against Tuberculosis and Lung Disease (The Union). Retrieved from http://global.
tobaccofreekids.org/files/pdfs/en/ILL_global_cig_trade_full_en.pdf
KPMG. (2014). Project Sun, A Study of the Illicit Cigarette Market in the European Union, 2013.
Results. Retrieved from http://www.kpmg.co.uk/email/06jun14/om014549a/pageturner/files/
assets/basic-html/page1.html
La Stampa. (2013, June 7). Sigarette elettroniche “tarocche”, in cinque mesi 800 mila sequestri.
LaStampa.it. Retrieved from http://www.lastampa.it/2013/06/07/italia/cronache/sigarette-
elettroniche-tarocche-in-cinque-mesi-mila-sequestri-RLtFM5RYL74AUIQ8ockRCN/pagina.
html
Levinson, B. (2011). An Inquiry into the Nature, Causes and Impacts of Contraband Cigarettes.
Washington, DC: Center for Regulatory Effectiveness.
OLAF. (2012). Roadmap: Stepping up Efforts to Fight Against Cigarette Smuggling—A
Comprehensive EU Strategy. Retrieved from http://ec.europa.eu/governance/impact/planned_
ia/docs/2013_olaf_001_fighting_cigarette_smuggling_en.pdf
Polskie Radio. (2013). Poland Opposes EU Menthol and Slim Cigarettes Ban. Polskie Radio dla
Zagranicy. Retrieved from http://www.thenews.pl/1/12/Artykul/125820,Poland-opposes-EU-
menthol-and-slim-cigarettes-ban-. Accessed 5 April 2016.
RolandBerger. (2013). The New Tobacco Products Directive—Potential Economic Impact.
Munich: RolandBerger.
Sparks, I. (2013, May 31). France to Ban Electronic Cigarettes from all Public Places and Subject
Them to the Same Tight Rules as Tobacco. Dailymail. Retrieved from http://www.dailymail.
co.uk/news/article-2333843/France-ban-electronic-cigarettes-public-places-subject-tight-
rules-tobacco.html
Transcrime (2015). The European Outlook on the Illicit Trade in Tobacco Products. Trento:
Transcrime—Università degli Studi di Trento.
United States Trade Representative. (2010, November). United States—Measures Affecting the
Production and Sale of Clove Cigarettes DS406: First Written Submission of the United States
of America. United States. Retrieved from www.worldtradelaw.net/wtodisputesubmissions/us/
DS406_USFirstWrittenSubmission.pdf
United States Trade Representative. (2011). United States—Measures Affecting the Production
and Sale of Clove Cigarettes DS406: Second Written Submission of the United States of
America. Office of the United States Trade Representative. Retrieved from http://www.
worldtradelaw.net/wtodisputesubmissions/us/DS406_USSecondWrittenSubmission.pdf
Chapter 18
A New Victorian Era: Getting Crime Out
of Commercial Sex

Adam Masters and Adam Graycar

Introduction

Approaches to social crimes differ widely around the world, with different cultures apply-
ing their own standards of what is, or is not, acceptable behaviour. This chapter explores
one type of social crime—commercial sex—and the effects of decriminalisation in the
state of Victoria, Australia. We ask how successful is decriminalisation in preventing the
criminality associated with commercial sex and whether decriminalisation will prevent or
increase such criminality.
In Australia, the states and territories have been changing their outlook on the
sale of sexual services (sex work), those who sell these services (sex workers) and
their working environment (brothels). As a multi-jurisdictional federation, there is
little consistency to the changes or how commercial sex is treated by the law and
police. For example, in the Australian Capital Territory and Victoria licenced broth-
els have operated for decades, while at the other end of the scale brothels in
Tasmania and South Australia remain criminalised. Even jurisdictions which have
taken a path toward the decriminalisation of sex work—in effect, preventing crime
by abolishing it—have many regulatory and criminal offences remaining on the
statutes books.
Decriminalisation is not legalisation. In Victoria, decriminalisation has created
additional regulatory offences for sex work, while retaining many of the criminal
offences associated with procuring, human trafficking, child exploitation and organised

A. Masters (*)
Transnational Research Institute on Corruption, The Australian National University,
22 University Avenue, Acton, ACT 2601, Australia
e-mail: adam.masters@anu.edu.au
A. Graycar
School of Social & Policy Studies, Flinders University,
GPO Box 2100, Adelaide, South Australia 5001, Australia

© Springer International Publishing Switzerland 2017 293


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_18
294 A. Masters and A. Graycar

crime (see Appendix). Before 1984, attempts to outlaw or eliminate prostitution


criminalised sex workers and created profitable opportunities for organised crime
(Albon, 2014; Costigan, 1984; Dobinson, 1992; Neave, 1985; Richardson, 1992).
Over the past 40 years, Victoria has progressively decriminalised the sex industry.
There has been a marked reduction of police activity related to sex work (see
Fig. 18.1). Decriminalisation devolved regulatory control to bodies in local and
state government. As a result, and despite the intentions of legislators, crime
remained a part of the Victorian commercial sex market as criminals exploited the
uncertainty among police and government agencies over who was ultimately
responsible for controlling the sex industry.
We conceptualise the Victorian sex industry as occupying a space alongside
other semi-illegal markets like sports doping (Paoli & Donati, 2013) or gambling
(Haberfeld & Sheehan, 2013). Braithwaite (2005) describes these as ‘markets in
vice’ and he includes over-eating, drugs, alcohol, tobacco and tax advice. However,
the market in the vice of over-eating does not qualify as a semi-illegal market.
Whether semi-illegal markets operate within the law is dependent on circumstances.
Unlike highly criminalised markets (e.g. illegal drugs) or predominantly lawful
markets (e.g. finance), semi-illegal markets occupy significant space on both sides
of the legal divide.
Shifting sex work from the criminal law to a more administrative environment
makes it a candidate for the process of crime-proofing legislation, a process under
theoretical development in Europe aimed at minimising the unintended crimino-
genic effects of legislative action (see Levi & Dorn, 2006; Morgan & Clarke, 2006;
Savona, 2006 and Savona’s chapter in this volume). Our analysis in this chapter is
framed by the work of Levi and Dorn (2006: 242), whose criteria for legislative/
regulatory crime risk assessment is part of the crime-proofing legislation project.
We have applied their approach retrospectively to the Victorian case to highlight
some considerations for jurisdictions contemplating change to the way sex work is
treated. Levi and Dorn (2006) developed their analytical approach with markets in
mind. It is therefore a suitable tool for analysing the regulation of a commercial sex
market.

Prostitution in Victoria: 1851–1984

As our title suggests, there is a new approach to the enforcement of criminal


laws based on social standards from another time. Queen Victoria (1831–1901)
lends her name to both the jurisdiction scrutinised herein, and the social values
of the era, which underpinned the laws prohibiting the commercialisation of
sex. The colony of Victoria was settled by Europeans in 1835, became a sepa-
rate colony from New South Wales in 1851, and a State of the Australian
Commonwealth in 1901. Throughout this period and too this day, prostitution—
defined as the exchange of sexual services for money—has not in and of itself
been illegal (Neave, 1985: 5; Pinto et al., 1990: 2). However, a series of legisla-
tive measures have been used by police and other authorities to control
18 A New Victorian Era: Getting Crime Out of Commercial Sex 295

Fig. 18.1 Prostitution in Victoria—offences 1977–2013


296 A. Masters and A. Graycar

Table 18.1 Historic legislation used for the control of prostitution in Victoria
Act Powers or offences
Vagrancy Act 1852 Arrest and detention for no lawful means of support
The Criminal Law and s.44 an offence ‘by force take away or detain her against her will
Practice Statute 1864 any woman of any age with intent to marry or carnally know
her’. See s.51
Conservation of Public Compelled prostitutes to be medically examined. Those found to
Health Act 1878 be diseased were sent to Lock Hospitals for confinement
Crimes Act 1891 s.14 and s.15 outlawed the procuring of adult women
Police Offences Act 1891 s.7(2) Importuning in a public place
Police Offences Act 1891 s.5 Living off earnings
(1907 amendments) s.6 Keeping a brothel
Police Offences Act 1891 s.26 Soliciting
(1928 amendments) s.30 Prostitutes assembling in ‘refreshment houses’
s.69(2) Prostitutes behaving riotously in public places
s.79 Pimping
s.80 Brothel keeping
Police Offences Act 1940 s.3 Criminalised single-woman brothels
Police Offences Act (1961 s.3 Loitering for the purposes of prostitution
amendment)
Crimes Act 1958 s.59 Procuring offences
Summary Offences Act 1966 s.18 Solicit or accost any person … loiter in a public place
(prostitute)
s.18A Loiter in or frequent any public place for the purposes of
inviting or soliciting any person to prostitute him/herself for
pecuniary reward (client)
Vagrancy Act 1966 s.10(1) Living off earnings of prostitution
s.11(1) Keeping or managing a brothel
s.12 Occupier permit premises to be used for prostitution
Source: Summarised from Perkins, 1991, Chap. 2

prostitution. Table 18.1 details the development of these laws in Victoria up


until 1984, the year in which the decriminalisation process became a substantive
project in the state. The year 1984 marked a pivotal moment when the Planning
(Brothels) Act 1984 (Vic) was passed.
The offences outlined in Table 18.1 gave authorities the power to control and
punish all those involved with prostitution. They reflected the morality of the era,
criminalising (1) street prostitution; (2) sex work in brothels; (3) offences related
to the premises used for prostitution; (4) living off the earnings of prostitution; (5)
brothel keeping and (6) procuring. The history and application of these laws have
been well detailed elsewhere (see Daniels, 1984; Dunstan, 1968; Frances, 2007;
Perkins, 1991). However, the broad categorisation is useful for our purposes in
mapping offences related to sex work in the pre and post 1984 periods. To these
can be added a variety of regulations related to advertising and those directed at the
clients of prostitutes.
Figure 18.1 presents a timeline of how these classes of offences have been applied
between 1977 and 2013. We have compiled this data from a variety of sources:
18 A New Victorian Era: Getting Crime Out of Commercial Sex 297

1977–1984 (Neave, 1985, vol. II, 38); 1984/1985–1988/1989 (Perkins, 1991: 125;
Pinto et al., 1990); 1984–2013 (VicPol, 1985, 1986, 1987, 1988, 1989, 1990, 1991,
1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013 inclusive).1 What is immedi-
ately apparent is the sharp decline in charges laid by police for activities related to
sex work in the past 35 years. The number of charges laid in 2012/2013 (n = 149) is
more than 90 % fewer than those laid in the peak year of 1980 (=1728), while the
lowest year, 2010/11 (n = 81), is less than 5 % of the 1980 peak.

Decriminalisation as a Prevention Strategy: Sex Work


in Victoria Since 1984

A disease, public agitation, two inquiries and two pieces of legislation framed the
first steps toward decriminalising the sex industry in Victoria. The early 1980s saw
the spread of the human immunodeficiency virus, otherwise known as the acquired
immunodeficiency syndrome (HIV/AIDS). What had largely been ignored as a
problem for male gay community and intravenous drug users was now a public
health issue—and commercial sex work represented a potential high risk (Frances,
2007; Perkins, 1991). At the same time, both street prostitution and brothels had
spread into residential areas, which led to the formation of a resident lobby group
Westaction (Perkins, 1991: 111). As a result of the lobbying, a working party was
established by the Victorian Parliament to investigate the location of brothels. This
working party produced two significant outcomes. The first was the Planning
(Brothels) Act 1984 (Vic) that established a permit system and regulatory frame-
work for decriminalising brothels in certain areas. The act delegated the authority to
issue permits to local councils. The second was the establishment of the Inquiry into
Prostitution.
Chaired by Marcia Neave, the Inquiry Into Prostitution was given the following
terms of reference:
1. To inquire into and report upon the social, legal and health aspects of prostitution
in all its forms insofar as those matters are relevant to the powers and functions
of the Victorian Parliament and Government.
2. In reporting to the government the inquiry shall amongst other things:
(a) Examine the implications of the various alternative actions available to the
government with respect to all parties involved in prostitution with particular
reference to the implications of decriminalisation and legalisation thereof.
(b) Consider other inquiries relevant to prostitution.

1
Note: Perkins data for 1986/1987–1988/1989 contained a category ‘other’ these were zero, nine
and three charges laid for the respective years. These have been included in the ‘aiding and abet-
ting’ category in Fig. 18.1. Furthermore, Perkin’s categories ‘soliciting for immoral purposes;
soliciting/accost for prostitution; and loiter/soliciting by prostitutes’ have all been included under
our category of ‘street offences’ in Fig. 18.1.
298 A. Masters and A. Graycar

(c) Consider the relevant law and practice in relation to prostitution elsewhere.
(d) Have regard to the reports of the Interdepartmental Monitoring Committee
relating to the Planning (Brothels) Act 1984.
3. In consideration of the above matters to recommend to the Victorian Government
as to whether existing laws and town planning practices require amendment and
to recommend on other measures necessary and desirable with respect to prosti-
tution in Victoria (Neave, 1985: v).
The inquiry eventually made 91 recommendations and identified three policy
options. Essentially the options were to retain criminal penalties for prostitution-
related offences; repeal the criminal penalties (decriminalisation); or regulate pros-
titution (legalisation). The inquiry recommended a combination of repeal and
regulate—repeal penalties for most prostitution related offences (yet retaining pen-
alties for street prostitution); regulate brothels via town planning and licence opera-
tors to exclude criminals (Neave, 1985: 8–9). As a result the parliament eventually
passed the Prostitution Regulation Act 1986 (Vic).
For those in the sex work industry the new legislation was problematic to say
the least. The Prostitutes’ Collective of Victoria claimed at a forum arranged by the
Australian Institute of Criminology (AIC) the Prostitution Regulation Act ‘failed
to meet its stated objectives and should be viewed as a lesson in defective policy
making and how to avoid it’ (Dobinson, 1992: 117). Legal problems for sex work-
ers were both exacerbated and new problems created. Much of this related to the
limited number of licenced brothels under the new regime. Estimates made to the
Neave Inquiry placed the number of brothels in the state at around 120 with an
additional 37 escort agencies, employing between them 1800–2300 of the then
3000–4000 sex workers in the state (Neave, 1985: 47). By the time the Prostitution
Regulation Act gained royal assent, only 42 brothels were licenced under the
Planning (Brothels) Act. This left the majority of sex workers outside the ‘legal-
ised’ sphere of licenced brothels. Furthermore, discriminatory and punitive
employment practices by brothel licencees (e.g. a system of fines for minor rule
infractions; loss of choice in clients; inflexible working hours) meant illegal sex
work was more profitable (Dobinson, 1992: 118). Yet it was not only the sex work-
ers who believed the legislation failed.
The Victorian Brothel Association also made contributed to the AIC forum. By
this stage, the Association represented 64 legal brothel owners. For these owners,
the Planning (Brothels) Act resulted in higher cost pressures associated with com-
pliance, competition from the illegal industry, trading hour restrictions and an
employment cap. The last point meant that legal brothels only employed 2000 of the
then estimated 6000 sex workers in Melbourne (Richardson, 1992). Another result
of the two prostitution acts was a redeployment of police resources, with the
Melbourne Vice Squad reduced to three or four members, who had little time to visit
the licenced brothels, let alone act in relation to the unlicenced ones (Richardson,
1992). From these perspectives, the early years of decriminalisation in Victoria
amounted to little improvement.
18 A New Victorian Era: Getting Crime Out of Commercial Sex 299

Current Legislation to Regulate the Market


for Commercial Sex

The Prostitution Control Act (1994)(Vic) superseded the Prostitution Regulation


Act. The new legislation retained all the criminal and regulatory offences and added
regulations to control offensive behavior toward prostitutes and the consumption of
alcohol in brothels. Over the past two decades a number of amendments have been
made and in 2010 the Prostitution Control Act was renamed the Sex Work Act 1994.
The Sex Work Act 1994 now contains some 17 criminal offences with a penalty of
five or more years in prison, 22 offences carrying jail terms of less than five years and
77 regulatory offences carrying a fine of which 18 can be dealt with by an infringe-
ment notice as an on-the-spot fine (see Appendix). The criminal offences include
those associated with procurement—mainly regarding children, but also protecting
adults—and living off the earnings of prostitution. Keeping unlicenced brothels
(s.22), or holding more than two licences (s.75), are also criminalised with the clear
intent of keeping organised criminal enterprises out of the industry. Sex workers are
subjected to a small number of regulatory requirements prohibiting street work in
certain places (s.13(1)(a)(b)(c), s.13(2)(a)(b)(c), i.e. near schools, hospitals, places
of worship or where children congregate) and similar restrictions apply to clients
(s.12(1)(a)(b)(c)). Sex workers are also required to practice safe sex (s.18A, s.20), as
are their clients (also s.18A). When operating a brothel either singly or in pairs, sex
workers are not required to obtain a licence, but are required to register as a small
business (s.24) and apply annually for a licence exemption (s.24A).
Sex work service providers are defined in the Act as ‘a person carrying on a busi-
ness of a … brothel [or] escort agency’. They are subject to 52 offences or regula-
tions specific to sex work service providers and 42 general offences or regulations
that apply to all persons on premises. These are detailed in Appendix. Providers, sex
workers and other people all have restrictions on how they advertise the industry
placed upon them (s.17(1)(2)(3)(4)). Finally, the rights of those in the industry are
protected by a series of regulations upon the regulators. These require inspectors to
show their identification (s.61B) and respect confidentiality (s.61DA(8), s.61Z).
Police and other public servants are similarly bound by secrecy provisions (s.87)
and also to destroy fingerprints of those in the industry when no longer required by
them (s.88A(1)) or other authorities (s.88A(3)).
While data is scarce on the extent of illegal sex work in Victoria there are some
indicators of the effect of this legislative change. Figure 18.1 provides an overview
of police law enforcement activity in the industry between 1977 and 2013. While
there is an obvious decline in charges over the years, this likely reflects police
resources being allocated to other crimes, as claimed by the Victorian Brothel
Association above. Law enforcement only ever had a deterrent effect and was never
prosecuted in a manner to eliminate commercial sex. Even at the peak of police
activity in the late 1970s, there was less than a one-in-three chance that a person
involved in the sex industry would be arrested and charged in any given year.
Another consideration is the geographic impact of the decriminalisation process.
Local councils have retained a veto power on the establishment of brothels since the
300 A. Masters and A. Graycar

Fig. 18.2 Geographic distribution of legal brothels (boundary in red). dark blue, metropolitan,
shaded, rural. Source: prepared by authors

Planning (Brothels) Act was passed. The effect of this veto power is that legalised
brothels have remained principally within the metropolitan areas of Victoria. Of the
96 licenced brothels operating today, two are in Geelong and the remainder are in
Melbourne (RhED, 2014). While the total has fluctuated over the years, the model of
two in Geelong and the remainder in Melbourne is unchanged since the Neave
Inquiry (1985: 152). Rural community resistance was noted by Neave with visits to
Geelong, Shepparton and Morwell, all places known to have brothels operating at the
time. During town meetings locals opposed decriminalisation. Neave noted that
these meetings were heavily attended by religious groups (Neave, 1985: 35–36).
Figure 18.2 shows the current geographic distribution of the legal brothels in Victoria.

Discussion

The previous sections briefly detailed three decades of reform and decriminalisation
in the Victorian sex industry. This section analyses the legal industry created by
these reforms against the criteria developed by Levi and Dorn (2006) for the assess-
ment of regulations applicable to corporate managers and auditors (Table 18.2).
Where necessary, our analysis draws on some vignettes of cases and alternative
approaches from other jurisdictions.
18 A New Victorian Era: Getting Crime Out of Commercial Sex 301

Table 18.2 Summary of 18 draft criteria for legislative/regulatory crime risk assessment
Quality of the text Vulnerability of the market Attraction/repulsion
1. Low level of clarity of 7. Discretionary power or 13. The proposed changes will be
the text itself responsibility for com- widely publicised (considering
2. Complexity when con- pliance put on officials/ that knowledge of a change is a
sidered alongside exist- agents prerequisite to exploiting it)
ing regulations 8. Creation of or increase in 14. The proposed change is in a
3. Difficult or time con- conflicts of interest for field in which/adjacent to which
suming to apply in officials and/or agents criminals are already active
everyday practice (current and future) 15. The methods required to exploit
(trader opinion) 9. Lack of requirement the vulnerabilities listed in col-
4. Rule unlikely to be for rotation of official/ umn 2 are within the existing
robust in the face of on- agents competencies of many
the-spot challenge 10. Lack of requirements criminals
(enforcer opinion) for guidelines on how 16. Those crime methods do not
5. Uncertain justiciability officials/agents should require complicated crime proj-
(legal opinion) discharge their duties ect teams or technologies
6. There are easier ways 11. Lack of requirements for 17. Those methods are skills pos-
of attaining the same effective supervision sessed by at least some poten-
policy objectives 12. Lack of back-up by law tial entrants (as key informants
enforcement (criminal report)
or administrative) 18. Potential/actual perpetrators’
values are not seriously vio-
lated by either targets or meth-
ods of crime (as key informants
report)
Source: Levi & Dorn, 2006: 242

Quality of the Text

Quality of the text refers to the refers to the clarity, consistency, comprehensibility
and accessibility of legislation or regulatory text to the users, not the drafters (Levi
& Dorn, 2006). The decriminalisation of brothels required a significant increase in
the amount of legislation to differentiate between legal and illegal brothels and to
establish a system of licencing. Over the years, the original legislation—Prostitu-
tion Regulation Act 1986 (Vic) grew from 43 to 81 pages as the Prostitution Control
Act 1994 (Vic), and eventually settled at over 200 pages in the current form—the
Sex Work Act 1994 (Vic). The 80+ criminal and regulatory offences associated with
sex work in Victoria are listed in Appendix, which also indicates to whom these
offences are applicable. While the offences appear relatively straightforward there
remain numerous complicating factors. For example, the offence of living on the
earnings of a sex worker under s.10(1) excludes licence holders and those ‘that do
not hold a relevant position (whether in his or her own right or on behalf of any
other person) in the sex work service providing business or exercise a significant
influence over or with respect to the management or operation of that business’.
Applying Levi and Dorn’s (2006, 239) clarity test, the lay practitioner could easily
believe that the life partners and children of street sex workers and those in unli-
cenced brothels may still be guilty of the offence. Despite only 18 charges being
302 A. Masters and A. Graycar

laid in the past 2 years for earnings related offences (Vicpol, 2012, 2013), the text
remains unclear.
Levi and Dorn pose two questions to consider when comparing the complexity
of new regulations to the old. First are the new rules (i.e. the current version of the
Sex Work Act) consistent with their predecessors. Second are the number of regu-
lations small enough not to overwhelm practitioners and regulators. On the first,
the rules are reasonably consistent, reflecting an evolution of the law as problems
are identified and addressed. For example, while children generally are not per-
mitted on licenced premises or premises used for sex work (s.11A), an exception
has been made for those less than 18 months old. This represented a balanced
change to the law to accommodate the needs of self-employed sex workers who
are also parents of young infants2 and illustrated reasonable consistency between
the new and the old.
However, on the second question, the number of regulations can overwhelm
practitioners and regulators alike. For example, council officers charged with con-
trol of licenced and unlicenced brothels faced complex evidential requirements to
tackle the illegal side of the industry. Several councils found the only way to gather
evidence in a timely fashion was to hire private investigators to engage in commer-
cial sex at illegal brothels (Lucas, 2007a). Without this tactic, Councils were forced
to give 2 days’ notice before an inspection (Lucas, 2007b). This complexity flowed
into Levi and Dorn’s next criterion and demonstrated the difficult and time-
consuming measures required to prove the operation of an illegal brothel.
The difficulties of on-the-spot challenges and uncertainty in justiciability to con-
trolling illegal brothels from a legislative point of view is most aptly underscored by
the presence of a massage parlour that operated as an illegal brothel within view of
the headquarters of Consumer Affairs Victoria (CAV), the state government depart-
ment responsible for regulating sex work. First reported in the media in September
2007 (Moor, 2007), the illegal brothel was still operating 12 months later (Moor,
2008). In 2010, a legislative change prompted by this resulted in an increased num-
ber of brothel keeping charges being preferred (Table 18.3). The change clarified
exactly which agency was responsible for controlling illegal brothels. Victoria
Police were given new powers of entry and responsibility for investigating illegal
brothels; the Sex Work Act made CAV responsible for regulating licenced brothels;
and councils remained responsible for planning permits (Delacorn et al., 2011).
While Table 18.3 shows a jump in brothel keeping offences in 2011/2012 and
2012/2013, this jump is more likely associated with the Victoria Police forming a
Sex Industry Unit, which effectively re-created the Vice Squad, previously dis-
banded in 1999.
From the practitioners’ (i.e. sex worker) point of view, Frances (2007, 275)
explained that ‘although these laws have been in operation for over a decade, the

2
Care needs to be applied to terms such as ‘infant’ when researching sex work in Australia. In the
Northern Territory sex work-related legislation refers to persons less than 18 years of age as
infants.
Table 18.3 Sex Work Offence: Charges laid in Victoria 2003/2004–2012/2013
Offence category 03/04 04/05 05/06 06/07 07/08 08/09 09/10 10/11 11/12 12/13 Total
Street work 285 488 325 182 73 84 110 29 54 35 1665
Client 40 64 41 19 23 9 14 34 92 73 409
Brothel keeping 6 38 5 8 4 7 0 7 25 20 120
Procuring 6 11 15 21 17 10 11 4 6 2 103
Earnings 1 5 5 4 6 4 1 2 12 6 46
On premises 3 1 9 4 1 3 1 5 3 13 43
Advertising 0 0 3 1 0 2 0 0 0 0 6
Brothel work n/a since 1984
Total sex 341 607 403 239 124 119 137 81 192 149 2392
work-related
18 A New Victorian Era: Getting Crime Out of Commercial Sex

offences
Source: Compiled by authors from Victoria Police Crime Statistics
303
304 A. Masters and A. Graycar

majority of Victorian workers are still employed in illegal operations or on the


street, because of the expensive and cumbersome licensing system’. The Victorian
legislature again updated the law in 2012; yet the number of licensed brothels did
not increase leaving most sex workers on the illegal side of the industry.

Vulnerability of the Market

The devolution of controlling part of the sex industry to local government- and
state-based agencies shifted discretionary power and responsibility for compliance
from the police. By the mid-2000s, a multitude of agencies had overlapping respon-
sibility for criminal and regulatory control of sex work. Local government councils
ensured licencees complied with planning laws and responded to constituent com-
plaints. Conflict of interest existed where council officers inspected both legal and
illegal brothels. Furthermore, budgetary measures restricted council ability to rotate
these inspectors. At the state level, Victoria Police retained powers in relation to
criminalised activity associated with the sex industry, including state offences
related to sexual servitude. CAV was responsible for compliance with licencing
regulations. At the federal level, the Australian Federal Police investigated sexual
servitude offences and the Department of Immigration and Citizenship had respon-
sibility for ensuring visa compliance (Joint Committee on the Australian Crime
Commission, 2002; Segrave, 2009). As can be seen in Fig. 18.1 and Table 18.3,
there is a steady decline in police activity in relation to the state laws on prostitution.
Enforcement of brothel keeping offences was at a minimum when local council
inspectors policed both legal and illegal brothels while state and federal police have
focused their attention on sexual servitude, immigration and human trafficking
offences associated with the sex industry (see Larsen, et al., 2012; Putt, 2007). The
result of this multi-tiered approach meant that responsibility was often shuffled
between the various government agencies creating vulnerabilities in the market as
described by Levi and Dorn (Table 18.2).
The multi-tiered approach to regulating the sex industry has proven problematic.
Between 2002 and 2010, the Coordinator of Planning Enforcement for Yarra City
Council accepted $134,260 in bribes from the operators of three illegal brothels.
The payments into a gambling account were made in exchange for early warnings
about any impending raids by authorities. This council officer, had been involved in
the investigation of other illegal brothels, investigations which included local, state
and federal authorities—each with a different role in policing the legal and illegal
sex industry (McKenzie & Beck, 2011c; 2011b). Furthermore, the investigation
revealed federal agencies had intelligence about the corrupt officer as early as 2002
(McKenzie & Beck, 2011c). One of the bribers operated both legal and illegal
brothels, which placed the council officer in contact with them as part of their duties,
thus highlighting the vulnerability of certain high risk functions to corruption
18 A New Victorian Era: Getting Crime Out of Commercial Sex 305

(Gorta, 2006: 209-10).3 The bribed council officer had a high profile in fighting
illegal brothels, having testified before a Commonwealth Parliamentary Committee
on the issue (Joint Committee on the Australian Crime Commission, 2002). The
evidence furnished to the parliamentary committee by this officer and others out-
lined how the policy decision to decriminalise brothels in Victoria had led to unin-
tended consequences. The consequences included an increase in the number of
illegal brothels under the cover of the legal ones and confusion as to which level of
government was responsible for what under the new arrangements (Joint Committee
on the Australian Crime Commission, 2002). Ironically, the council officer was
directly asked whether he had heard of illegal brothels being tipped off before
raids—to which he answered ‘no…’ (Joint Committee on the Australian Crime
Commission, 2002: 10). The scale of the bribery for a single official is indicative of
the illegal profits being generated by the unregulated sex industry.
The case of the Yarra Council inspector highlighted a period of extreme market
vulnerability. Inspectors had a conflict of interest in trying to control both legal and
illegal brothels at the same time. Councils with limited budgets were not in a posi-
tion to rotate or effectively supervise their officials; thus this particular inspector
was left vulnerable to an environment whereby the same activity—sex work—was
legal in one place and illegal literally next door. Finally, as the Yarra Council inspec-
tor pointed out to the parliamentary committee, local officials were not backed up
by law enforcement at either state or federal level.

Attraction/Repulsion

Since 1984, changes to the laws on sex work have been widely publicised as the
Victorian legislature sought to balance conflicting social values. A search of the
Factiva ™ database of the two most popular Victorian newspaper—The Age and the
Herald Sun—shows that since 1992, there have been 180 reports related to illegal
brothels. Our analysis of these reports shows that 84 of these are predominantly
concerned with the policy debate. Naturally, press coverage tends to peak as changes
are debated, proposed and implemented.
Since aspects of the commercial sex were criminalised, organised crime groups
have been active in the field (Neave, 1985; Costigan, 1984; Perkins, 1991). Criminal
groups involved in the illegal side of the Victorian sex industry possess the methods
and competencies to exploit the vulnerabilities listed in the second column of

3
It is worth noting that this council officer falls into 8 of 15 high-risk functions identified by Gorta
(2006: 209-10). These are ‘(1) Inspecting, regulating or monitoring the standards of premises,
businesses, equipment or products; (2). Providng a service to new immigrants [in this case sex-
workers]; (3) Issuing qualifications or licences to indicate their proficiency or enable them to
undertake certain types of activities; (4) Providing a service to the community where demand fre-
quently exceeds supply; … (6) Issuing, or reviewing the issue of, fines or other sanctions; … (10)
Making determinations/handing down judgements about individuals or disputes; … (12) Having
discretion concerning land re-zoning or development applications; … (15) Having regular deal-
ings with the private sector other than for the routine purchasing of goods and services’.
306 A. Masters and A. Graycar

Table 18.2. These methods are often uncomplicated, requiring neither teams nor
technology to apply. Besides utilising bribery as a tool to subvert law enforcement,
at least one illegal operator used the cover of their licenced brothel to operate a
chain of unlicenced ones (McKenzie & Beck, 2011a). Where complexity exists—
i.e. the knowledge required to bypass or exploit federal migration laws in cases
where illegal brothels use human trafficking networks as a source of sex workers—
methodologies are already well known in the region and the groups are sufficiently
flexible to respond to legislative or changes in enforcement practices (Larsen et al.,
2012; Putt, 2007; Yea, 2012). These criminal groups will go to extremes to protect
their business—in one instance a young man attempting to ‘rescue’ a sex slave was
killed in front of a brothel (McKenzie & Beck, 2011a) and in another, a man and
wife involved in the industry were murdered in their home (Silvester et al., 2003).
Thus those wishing to enter Victoria’s illicit sex trade may find themselves chal-
lenged by the murderous lengths already employed by existing players.
On Levi and Dorn’s final point, for those involved in the illegal side of Victoria’s
sex industry their values are not seriously violated by the methods they choose to
meet the demand side of the market. In 1975, Victoria Police estimated there were
100 massage parlours operating as brothels in the state (VicPol, 1975, 12). Their
estimate increased to 150 the following year (VicPol, 1976, 55). Since decriminali-
sation of brothels was first introduced, the estimated number of illegal brothels has
always been significantly greater than the licenced variety. Neave (1985, 47) esti-
mated there were 3000–4000 sex workers, mostly working in 120 brothels in
Victoria. In 1992, there were 64 licenced brothels employing approximately 2000 of
the estimated 6000 sex workers in the state (Richardson, 1992, 136). In 2004, ‘the
sex-worker industry in Victoria is comprised of: about 2000 workers for 400 illegal
brothels; 500 street workers, 5000 sole traders; 1600 workers in about 90 legal
brothels; and 400 escorts working for 20 agencies’ (Lamperd, 2004). By 2010,
operators reported about 400 illegal brothels, a number conflicting with sex workers
who indicated there were only about 140, and prostitution enforcement officers esti-
mating there were 67 (Moor, 2010). In a submission to CAV in May 2014, the
Australian Adult Entertainment Industry Inc. repeated the estimate of 400 illegal
brothels in the state in a submission to CAV (Albon, 2014). Their calculation is
based on the number of advertisements for ‘massage therapy outlets’ in suburban
papers. While the estimates are not official figures they are the best available. The
reduction in overall sex work-related charges (Fig. 18.1) are in stark contrast to the
increased number of illegal brothels. It can reasonably be concluded that the illegal
industry has significantly expanded in the wake of decriminalisation.

Conclusion

The policy choice to decriminalise sex work in Victoria has clearly reduced some
crime related to commercial sex, but has had unintended effects elsewhere. These
effects include an apparent increase in instances of sexual servitude (slavery),
18 A New Victorian Era: Getting Crime Out of Commercial Sex 307

illegal immigration, exploitation and corruption. Victoria represents an ideal case


study on several grounds. The Prostitution Regulation Act 1986 (Vic) represented
the first significant shift in policy, followed a decade later by the Sex Work Act 1994
(Vic) and Sex Work Regulations (2004) (Vic). Regulation of the sex industry is
divided between three levels of government—local government approves planning
for brothels; the state government is responsible for preventing child exploitation
occurring in brothels and the federal government has jurisdiction over cases of sex-
ual servitude (i.e. slavery) and the illegal immigration of sex workers. As far back as
1991, the changes were highly criticised by those in the sex industry and reformers
who observed ‘that legislation in Victoria has failed to meet its stated objectives and
should be viewed as a lesson in seriously defective policy making and how to avoid
it’ (Dobinson, 1992). Finally, the Victorian experience can be compared to the cir-
cumstances other Australian state jurisdictions.
In terms of crime prevention, what decriminalisation has done is likely minimal.
Despite the large drop in the number of sex industry-related charges in the past
several decades, the increased number of illegal brothels and the growth in human
trafficking provide a counterpoint to this statistical success.
What decriminalisation has done successfully is shift the social perspective on
sex work. The presence of an illegal brothel across the road from the regulatory
authority for over a year illustrates how the tolerance has permeated both society
and the regulators. Despite media attempts at stirring a moral panic, the broader
community tends to accept the sex industry—legal and illegal—as part of the fabric
of society. Law enforcement attention and resources is rightly focused on the severe
end of the scale—the involvement of organised criminal groups and human traffick-
ing. It cannot be said decriminalisation has prevented crime in the sex industry, but
in this instance it has replaced one set of Victorian values with a new set of values
for Victoria.

Appendix: Offences Under the Sex Work Act 1994 (Vic)

Sex Work Act 1994 regulatory offences, application and penalties

Regulation SWa SPb Cc Od Maximum penalty


S.5(1) Causing or inducing child to take part ✓ ✓ ✓ 10 yrs
in sex work
S.6(1) Obtaining payment for sexual services ✓ ✓ 15 years
provided by a child
S.7(1) Agreement for provision of sexual ✓ ✓ ✓ 15 years
services by a child
S.8(1)(a) Forcing person into or to remain in ✓ ✓ ✓ 10 year
sex work (assault)
S.8(1)(b) Forcing person into or to remain in ✓ ✓ ✓ 10 year
sex work (intimidate)
S.8(1)(c) Forcing person into or to remain in ✓ ✓ ✓ 10 year
sex work (supply or offer drugs)
308 A. Masters and A. Graycar

Regulation SWa SPb Cc Od Maximum penalty


S.8(1)(d) Forcing person into or to remain in ✓ ✓ ✓ 10 year
sex work (through fraud)
S.9(1)(a) Forcing person to provide financial ✓ ✓ 10 year
support out of sex work (assault)
S.9(1)(b) Forcing person to provide financial ✓ ✓ 10 year
support out of sex work (intimidate)
S.9(1)(c) Forcing person to provide financial ✓ ✓ 10 year
support out of sex work (supply or offer
drugs)
S.10(1)Living on earnings of sex worker ✓ ✓ 5 years
S.11(1) Allowing child to take part in sex ✓ ✓ ✓ 10 year
work
S.11A(1) Child over 18 months not to be in ✓ 1 year 120PUf.
brothel
S.12(1)(a) Street sex work—offences by ✓ 30,60,90 PU
clients—inviting or soliciting near a place of 3,6,9 months
worship (1st, 2nd, 3rd and subsequent
offence)
S.12(1)(b) Street sex work—offences by ✓ 30,60,90 PU
clients—inviting or soliciting near a hospital 3,6,9 months
(1st, 2nd, 3rd and subsequent offence)
S.12(1)(c) Street sex work—offences by ✓ 30,60,90 PU
clients—inviting or soliciting near a school, 3,6,9 months
education and care service premises or
children’s services centre (1st, 2nd, 3rd and
subsequent offence)
S.12(1)(d) Street sex work—offences by ✓ 30,60,90 PU
clients—inviting or soliciting near a public 3,6,9 months
place regularly frequented by children and in
which children are present at the time of the
loitering (1st, 2nd, 3rd and subsequent
offence)
S.12(2)(a) Street sex work—offences by ✓ 10, 30, 60 PU
clients—A person must not loiter in or 1,3,6 months
frequent a public place for the purpose of, or
with the intention of, inviting or soliciting
any person to offer sex work services
himself or herself with him or her or another
person or of being accosted by or on behalf
of a sex worker (1st, 2nd, 3rd and
subsequent offence)
S.12(2)(b) Street sex work—offences by ✓ 10, 30, 60 PU
clients—a person must not in a public place 1,3,6 months
invite or solicit any person to offer sex work
services himself or herself with him or her or
another person (1st, 2nd, 3rd and subsequent
offence)
S.13(1)(a) Street sex work—offences by sex ✓ 10, 30, 60 PU
workers—inviting or soliciting near a place 1,3,6 months
of worship (1st, 2nd, 3rd and subsequent
offence)
18 A New Victorian Era: Getting Crime Out of Commercial Sex 309

Regulation SWa SPb Cc Od Maximum penalty


S.13(1)(b) Street sex work—offences by sex ✓ 10, 30, 60 PU
workers—inviting or soliciting near a 1,3,6 months
hospital (1st, 2nd, 3rd and subsequent
offence)
S.13(1)(c) Street sex work—offences by sex ✓ 10, 30, 60 PU
workers—inviting or soliciting near a 1,3,6 months
school, education and care service premises
or children’s services centre (1st, 2nd, 3rd
and subsequent offence)
S.13(1)(d) Street sex work—offences by sex ✓ 10, 30, 60 PU
workers—inviting or soliciting near a public 1,3,6 months
place regularly frequented by children and in
which children are present at the time of the
loitering (1st, 2nd, 3rd and subsequent
offence)
S.13(1)(2) Street sex work—offences by sex ✓ 5, 10, 15, 25, 30, 60
workers—place of worship, hospital, school, PU 1,3,6 months
etc., near children (1st, 2nd, 3rd and
subsequent offence)
S.15 Being in, entering or leaving unlicensed ✓ ✓ ✓ ✓ 10, 30, 60 PU
brothel (without reasonable excuse) 1,3,6 months
S.16 Offensive/intimidatory behaviour ✓ ✓ ✓ 3months 30PU
toward sex worker
S.17(1) Controls on advertising by sex work ✓ ✓ ✓ 40PU
service providers—not describe services
S.17(2) … not to be broadcast or televised ✓ ✓ ✓ 40PU
S.17(3) … not to induce a person to [work] ✓ ✓ ✓ 40PU
as sex worker or in brothel
S.17(4) … not to [advertise] using words ✓ ✓ ✓ 40PU
‘massage’, ‘masseuse’, ‘remedial’
S.18A(1) Sex workers and clients must ✓ ✓ 20PU
adopt safer sex practices use condom or
other appropriate barrier
S.18A(2) Sex workers and clients must ✓ ✓ 20PU
adopt safer sex practices minimise the risk of
acquiring or transmitting sexually
transmitted diseases while providing or
receiving those services
S.19(1) Permitting sex worker infected with ✓ ✓ 50PU
a disease to work in a brothel
S.20 Sex worker working while infected ✓ 20PU
with a disease
S.21(1)(a) Sell, supply or consuming liquor ✓ 100PU
in brothel
S.21(1)(b) Permit the sale, supply or ✓ 100PU
consumption of liquor in brothel
S.21A Operating brothel other than in a ✓ 3 years/360PU
building
S21E(3)(a) Requirement to give name and ✓ ✓ ✓ 5PU
address to police—refuse or fail
310 A. Masters and A. Graycar

Regulation SWa SPb Cc Od Maximum penalty


S21E(3)(b) Requirement to give name and ✓ ✓ ✓ 5PU
address to police—false name
S21E(3)(c) Requirement to give name and ✓ ✓ ✓ 5PU
address to police—incomplete address
S21E(3)(d) Requirement to give name and ✓ ✓ ✓ 5PU
address to police—fail to give in writing if
requested
S21E(7)(d) Requirement to give name and ✓ ✓ ✓ 5PU
address to police—must comply
S21G(1)—Offence to contravene banning ✓ ✓ ✓ 20PU, IN2PU
notice or fail to comply with police
directions—must not enter or attempt to
enter declared area
S21G(2)—Offence to contravene banning ✓ ✓ ✓ 20PU, IN2PU
notice or fail to comply with police
directions—must leave declared area when
directed
S.22(1)(a) A person must not knowingly or ✓ 5 years/1200PU
recklessly carry on a business as a sex work
service provider … without a licence
S.22(1)(b) … in breach of any condition of a ✓ 5 years/1200PU
licence
S.22(1)(c) … when a licence is suspended ✓ 5 years/1200PU
S.22(1A)(a) A person must not carry on a ✓ 240PU
business as a sex work service provider
without a licence
S.22(1)(b) … in breach of any condition of a ✓ 240PU
licence
S.22(1)(c) … when a licence is suspended ✓ 240PU
S22(3) A person must not assist in the ✓ 5 years /600PU
carrying on of a sex work service providing
business at a time when he or she knows that
subsection (1) is being contravened or is
reckless as to whether or not subsection (1)
is being contravened
S.24(1) Register for businesses to which ✓ 30PU
section 23 applies (1–2 worker brothels to be
registered (licence not required))
S.24A(4) Annual statement regarding ✓ 30PU
exemption (1–2 workers)
S.39A Licencee identity cards (to be carried) ✓ 10PU
S.40AA Endorsement of licence ✓ 10PU
S.40A Surrender of licence INe ✓ 10PU IN1PU
S.41 Cancelled or suspended licence must be ✓ 25PU IN3PU
returned IN
S.42(1) Requirement for licensee to be in ✓ 60PU
effective control of business—must be
regularly and usually in charge at the brothel
S.42(2) … must give regular and substantial ✓ 60PU
attendance at the brothel
18 A New Victorian Era: Getting Crime Out of Commercial Sex 311

Regulation SWa SPb Cc Od Maximum penalty


S.42(3) …must properly control and ✓ 60PU
supervise any approved manager …
S.42(4) … must … ensure that any … ✓ 60PU
person connected with the licensee’s
business complies with the provisions of this
act and any other laws …
S.42(5) … establish procedures designed to ✓ 60PU
ensure that the licensee’s business is
conducted in accordance with the law and in
a suitable manner
S.42(6) … monitor the conduct of the ✓ 60PU
licensee’s business in a manner that will
ensure, as far as is practicable, that those
procedures are complied with
S.42(7) If a business is run by more than one ✓ 60PU
licensee, each licensee must ensure that at
least one licensee is nominated as the
licensee in effective control of the business
at any one time and notify the authority in
writing of the nomination as soon as is
practicable
S.43 Updating of application ✓ 60PU
S.45 False or misleading information (re ✓ 60PU
licence application)
S.46 Requirement to notify changes in ✓ 60PU
information provided
S.46AA Authority may conduct checks upon ✓ 60PU
receipt of information relating to change of
ownership of premises
S.46AA Annual licence fee and statement ✓ 60PU
S.46d Production of information to authority ✓ 10PU
S.47(3) Licence cancellation (fail to notify ✓ 60PU
circumstances under s.47(1))
S.47A(6) Death, disability, etc. of licensed ✓ ✓ 50PU
sex work service provider (person carrying
on to comply with licence)
S.49(3) [a] brothel must at all times when ✓ 60PU / 6 months
open for business be personally supervised
by the licensee or an approved manager
S.52(4) Grant or refusal of approval or ✓ 60PU
renewal (approved manager to notify
changes)
S.52(5) Grant or refusal of approval or ✓ 60PU
renewal (approved manager to comply with
conditions, etc.)
S.52AAA Approved manager identity cards ✓ 10PU
(to be carried)
S.52AB Endorsement of certificate of ✓ 10PU
approval (approved manager to notify
variations)
312 A. Masters and A. Graycar

Regulation SWa SPb Cc Od Maximum penalty


S.52A(3) Surrender of approval (by ✓ 10PU IN1PU
approved manager) IN
S.53C(3) Person must imply with authority ✓ 25PU
imposed conditions
S.54B Cancelled or suspended approval ✓ 10PU
must be returned
S.57 Licensee not to carry on business with ✓ 120 PU / 1 year
unlicensed partner, etc.
S.58(a) Power to require licensee, etc. to ✓ 20PU IN2PU
state name and address (to police or
inspector—refuse or fail) IN
S.58(b) Power to require licensee, etc. to ✓ 20PU IN2PU
state name and address (to police or
inspector—false name or address) IN
S.58(c) Power to require licensee, etc. to ✓ 20PU IN2PU
state name and address (to police or
inspector—incomplete address) IN
S.59(3)(a) Power to require person to state ✓ ✓ ✓ ✓ 20PU IN2PU
age—refuse or fail IN
S.59(3)(b) Power to require person to state ✓ ✓ ✓ ✓ 20PU IN2PU
age—give false particulars IN
S.59(3)(c) Power to require person to state ✓ ✓ ✓ ✓ 20PU IN2PU
age—supply false evidence IN
S.60(1) Display of licence IN ✓ 10PU IN1PU
S.60(2) Display of licence (not false) IN ✓ 10PU IN1PU
S.60A(1) Display of prescribed signage ✓ 10PU
relating to sexual slavery
S.61(1) Licencee refusing to produce licence ✓ 10PU IN1PU
or certificate of approval IN
S.61(1) Licencee refusing to produce ✓ 10PU IN1PU
identity card IN
S.61B(1) Production of identity card (by ✓ 10PU
inspector)
S.61C(1) Accounts and other documents ✓ 10PU IN1PU
available for inspection IN
S.61C(2) Accounts and other documents ✓ 10PU IN1PU
available for inspection (by surrendered or
cancelled licencee) IN
S.61C(3) Accounts and other documents ✓ 10PU IN1PU
available for inspection (by suspended
licencee) IN
S.61DA(3) Persons to answer questions in ✓ ✓ ✓ ✓ 10PU
relation to suspected non-licenced sex work
service providing businesses
S.61DA(6) Persons to answer questions in ✓ ✓ ✓ ✓ 10PU
relation to suspected non-licenced sex work
service providing businesses (provide proof
of name and address)
18 A New Victorian Era: Getting Crime Out of Commercial Sex 313

Regulation SWa SPb Cc Od Maximum penalty


S.61DA(8) Persons to answer questions in ✓ 50PU
relation to suspected non-licensced sex work
service providing businesses (inspector not
to divulge information)
S.61P(2)(a)—Embargo Notices (person ✓ ✓ ✓ ✓ 10PU
knowingly sells embargoed thing)
S.61P(2)(b)—Embargo Notices (person ✓ ✓ ✓ ✓ 10PU
knowingly leases embargoed thing)
S.61P(2)(c)—Embargo Notices (person ✓ ✓ ✓ ✓ 10PU
knowingly moves embargoed thing)
S.61P(2)(d)—Embargo Notices (person ✓ ✓ ✓ ✓ 10PU
knowingly transfers embargoed thing)
S.61P(2)(e)—Embargo Notices (person ✓ ✓ ✓ ✓ 10PU
knowingly otherwise deals embargoed thing)
S.61U Refusal or failure to comply with ✓ ✓ ✓ ✓ 10PU
requirement
S.61 W Offence to give false or misleading ✓ ✓ ✓ ✓ 10PU
information
S.61Z Confidentiality (inspectors not ✓ 10PU
divulge)
S.62(3)(a) Powers of entry (person not to ✓ ✓ ✓ ✓ 1 year 120PU
prevent a member of the police)
S.62(3)(b) Powers of entry (person not to ✓ ✓ ✓ ✓ 1 year 120PU
obstruct/hinder a member of police)
S.75 Persons not to have interest in more ✓ 5 years 600PU
than one brothel licence or permit
S.77(1) Offences under Planning and ✓ 1200PU + 60PU daily
Environment Act 1987
S.81(3) Publication of declaration (of ✓ ✓ ✓ ✓ 6 months 60PU
proscribed brothel not to be removed, etc.)
S.82(1) Offences with respect to proscribed ✓ ✓ ✓ ✓ 12months 60PU
brothels (entering/leaving)
S.82(4) … premises to be a proscribed ✓ ✓ 12months 120PU
brothel are used as a brothel at any time after
service of notice …
S.87(2) Secrecy (unauthorised disclosure by ✓ 60PU
Advisory Committee, police, public service)
S.88A(1) Destruction of fingerprints, etc. (by ✓ 10PU
police when of no further use)
S.88A(3) Destruction of fingerprints, etc. (to ✓ 10PU
be returned to police [by other authorities]
when of no further use)
a
SW—sex worker
b
SP—sex work service provider—i.e. owners and managers of brothels (not small businesses with
one or two self-employed sex workers)
c
C—Sex work client
d
O—Other people—owners/occupiers of premises, police and inspectors, the general public, etc.
e
IN—Infringement Notice can be issued.
f
PU—Penalty Unit
314 A. Masters and A. Graycar

References

Albon, W. (2014). Public submission to consumer affairs Victoria: Response to regulatory impact
statement for the sex work (fees) regulations 2014. St Kilda: Australian Adult Entertainment
Industry Inc.
Braithwaite, J. (2005). Markets in vice markets in virtue. Annandale: The Federation Press.
Costigan, F. (1984). Royal commission on the activities of the federated ship painters and dockers
union, final report (Vol. 3). Canberra: The Government of the Commonwealth of Australia and
the Government of the State of Victoria.
Daniels, K. (Ed.). (1984). So much hard work: Women and prostitution in Australian history.
Sydney: Fontana/Collins.
Delacorn, A., Ross C., Merner, B., & Lesman, B. (2011). Sex work and other acts amendment bill
2011. Research papers. Melbourne: Victorian Parliamentary Library.
Dobinson, S. (1992). Victorian situation with legislation. In S. A. Gerull & B. Halstead (Eds.), AIC
conference: Sex industry and public policy (pp. 117–120). Canberra: Australian Institute of
Criminology.
Dunstan, K. (1968). Wowsers: Being an account of the prudery exhibited by certain outstanding
men and women in such matters as drinking, smoking, prostitution, censorship and gambling.
Melbourne, VIC: Cassell Australia.
Frances, R. (2007). Selling sex: A hidden history of prostitution. Sydney: UNSW Press.
Gorta, A. (2006). Corruption risk areas and corruption resistance. In C. Sampford, A. Shacklock,
C. Connors, & F. Galtung (Eds.), Measuring corruption (pp. 203–219). Burlington: Ashgate.
Haberfeld, M. R., & Sheehan, D. (Eds.). (2013). Match-fixing in international sports: Existing
processes, law enforcement, and prevention strategies. New York, NY: Springer.
Joint Committee on the Australian Crime Commission. (2002). Reference: Trafficking in women
for sexual servitude. Official committee hansard—18 November 2003. Melbourne: Parliament
of Australia.
Lamperd, R. (2004, June 5). The calling. Herald Sun, p. 1.
Larsen, J. J., Renshaw, L., Gray-Barry, S., Andrevski, H., & Corsbie, T. (2012). Trafficking in
persons monitoring report: January 2009-June 2011. In A. Tomison (Ed.), AIC Monitoring
Reports No. 19. Canberra: Australian Institute of Criminology.
Levi, M., & Dorn, N. (2006). Regulation and corporate crime: Managers and auditors. European
Journal on Criminal Policy and Research, 12, 229–255.
Lucas, C. (2007a, January 25). Councils pay private detectives to have sex. The Age.
Lucas, C. (2007b, January 26). Councils defend brothels tactic. The Age.
McKenzie, N., & Beck, M. (2011a, March 5). The secret world of Melbourne’s sex trade. The Age,
p. 1.
McKenzie, N., & Beck, M. (2011b, August 11). Guilty plea over bribes by brothels. The Age, p. 3.
McKenzie, N., & Beck, M. (2011c, November 17). Authorities took decade to act on brothel bribe
evidence. The Age, p. 1.
Moor, K. (2007, September 15). Buck passes on sex in the city. Herald Sun.
Moor, K. (2008, September 9). Blitz on illegal brothels. Herald Sun.
Moor, K. (2010, January 21). 40 illegal brothels offering full sex, 100 more offering a sexual ser-
vice and more than 450 illegal prostitutes working alone criminals infiltrating legal brothel
industry. Herald Sun.
Morgan, R., & Clarke, R. V. (2006). Legislation and unintended consequences for crime. European
Journal on Criminal Policy and Research, 12, 189–211.
Neave, M. (1985). Inquiry into Prostitution, final report, the Neave Report. Melbourne: Inquiry
into Prostitution.
Paoli, L., & Donati, A. (2013). The sports doping market: Understanding supply and demand, and
the challenges of their control. New York, NY: Springer.
Perkins, R. (1991). Working girls: Prostitutes, their life and social control. Canberra: Australian
Institute of Criminology.
18 A New Victorian Era: Getting Crime Out of Commercial Sex 315

Pinto, S., Scandia, A., & Wilson, P. (1990). Prostitution law in Australia. In D. Chappell (ed.),
Trends and issues in criminal justice No. 22. Canberra: Australian Institute of Criminology.
Putt, J. (2007). Human trafficking to Australia: A research challenge. In Toni Makkai (ed.), Trends
and issues in crime and criminal justice No. 338. Canberra: Australian Institute of Criminology.
RhED. (2014). Listings. http://sexworker.org.au/Listings/BrothelEscort/
ListofMetropolitanLegalBrothels.aspx. Retrieved 15 April 2014.
Richardson, P. (1992). The Victorian brothel owners’ perspective. In S. A. Gerull & B. Halstead
(Eds.), AIC conference: Sex industry and public policy (pp. 135–137). Canberra: Australian
Institute of Criminology.
Savona, E. U. (2006). Proofing EU legislation against crime. European Journal on Criminal Policy
and Research, 12(3-4), 177–178.
Segrave, M. (2009). Order at the border: The repatriation of victims of trafficking. Womens Studies
International Forum, 32(4), 251–260.
Silvester, J., Munro, I., & Jackson, A. (2003, October 25). The day security failed the couple with
love for sale. The Age.
VicPol. (1975). Annual report. Melbourne, VIC: Victoria Police.
VicPol. (1976). Annual report. Melbourne, VIC: Victoria Police.
VicPol. (1985). Statistical review of crime 1984. Melbourne, VIC: Victoria Police.
VicPol. (1986). Statistical review of crime 1985/86. Melbourne, VIC: Victoria Police.
VicPol. (1987). Statistical review 1986/87. Melbourne, VIC: Victoria Police.
VicPol. (1988). Statistical review 1987/88. Melbourne, VIC: Victoria Police.
VicPol. (1989). Statistical review 1988/89. Melbourne, VIC: Victoria Police.
VicPol. (1990). Statistical review 1989/90. Melbourne, VIC: Victoria Police.
VicPol. (1991). Statistical review 1990/91. Melbourne, VIC: Victoria Police.
VicPol. (1992). Statistical review 1991/92. Melbourne, VIC: Victoria Police.
VicPol. (1993). Crime statistics 1992/93. Melbourne, VIC: Victoria Police.
VicPol. (1994). Crime statistics 1993/94. Melbourne, VIC: Victoria Police.
VicPol. (1995). Crime statistics 1994/95. Melbourne, VIC: Victoria Police.
VicPol. (1996). Crime statistics 1995/96. Melbourne, VIC: Victoria Police.
VicPol. (1997). Crime statistics 1996/97. Melbourne, VIC: Victoria Police.
VicPol. (1998). Crime statistics 1997/98. Melbourne, VIC: Victoria Police.
VicPol. (1999). Crime statistics 1998/99. Melbourne, VIC: Victoria Police.
VicPol. (2000). Crime statistics 1999/2000. Melbourne, VIC: Victoria Police.
VicPol. (2001). Crime statistics 2000/01. Melbourne, VIC: Victoria Police.
VicPol. (2002). Crime statistics 2001/02. Melbourne, VIC: Victoria Police.
VicPol. (2003). Crime statistics 2002/03. Melbourne, VIC: Victoria Police.
VicPol. (2004). Crime statistics 2003/04. Melbourne, VIC: Victoria Police.
VicPol. (2005). Crime statistics 2004/05. Melbourne, VIC: Victoria Police.
VicPol. (2006). Crime statistics 2005/06. Melbourne, VIC: Victoria Police.
VicPol. (2007). Crime statistics 2006/07. Melbourne, VIC: Victoria Police.
VicPol. (2008). Crime statistics 2007/08. Melbourne, VIC: Victoria Police.
VicPol. (2009). Crime statistics 2008/09. Melbourne, VIC: Victoria Police.
VicPol. (2010). Crime statistics 2009/10. Melbourne, VIC: Victoria Police.
VicPol. (2011). Crime statistics 2010/11. Melbourne, VIC: Victoria Police.
VicPol. (2012). Crime statistics 2011/12. Melbourne, VIC: Victoria Police.
VicPol. (2013). Crime statistics 2012/13. Melbourne, VIC: Victoria Police.
Yea, S. (2012). ‘Shades of grey’: Spaces in and beyond trafficking for Thai women involved in
commercial sexual labour in Sydney and Singapore. Gender Place and Culture, 19(1), 42–60.
Part VII
Technologies of Crime Prevention
Chapter 19
Technology, Opportunity, Crime and Crime
Prevention: Current and Evolutionary
Perspectives

Paul Ekblom

Introduction: The Nature of Technology


and Technological Change

The world of crime science and situational crime prevention (SCP) has often drawn
on technology for practical purposes. This will increase as the world in which crime
is committed and prevented becomes ever more technologically based. However,
with some notable exceptions, there has been little attempt to explicitly theorise
about the role of technology in SCP. This chapter seeks to correct this deficiency—to
relate technology to key concepts in SCP, with the aim of giving us a more self-aware
and detached view of what technology is, and how it fits with crime science.
This chapter starts by briefly considering the nature of technology. It then looks
at the relationship between technology, opportunity, problems and solutions. But
opportunity is a more subtle concept than many in the field assume, needing further
development for present purposes. Opportunity is therefore discussed in relation to
both traditional frameworks of crime science and to a more integrated and detailed
counterpart, the Conjunction of Criminal Opportunity. But all the opportunity

I am grateful to Benoit Leclerc for his helpful editing, Richard Wortley and anonymous reviewers
for their inputs to this chapter.
P. Ekblom (*)
Design Against Crime Research Centre, Central Saint Martins, University of the Arts London,
London, UK
Department of Security and Crime Science, UCL, London, UK
Applied Criminology Centre, University of Huddersfield, Huddersfield, UK
e-mail: p.ekblom@csm.arts.ac.uk

© Springer International Publishing Switzerland 2017 319


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_19
320 P. Ekblom

frameworks need supplementing by an account of the dynamics of crime, especially


through the concepts of scripts and script clashes. A major section then examines
the relationship between crime and technological change, covering adaptations and
clashes over longer timescales, in the shape of co-evolutionary arms races between
offenders and preventers. This spans both biological and cultural evolution. Then
come sections on the practicalities of adopting a deliberately evolutionary approach
to prevention—gearing up against crime, innovation and design—and finally some
weaknesses of purely technological approaches to crime prevention. The conclu-
sion reviews the significance of understanding technology for crime science. This
enhanced understanding of how changing technology can both create and block
opportunity for crime, nuisance and terrorism (henceforth, crime) is needed to help
us anticipate, detect and respond to the many changes in the crime and security
world we can expect to encounter during the rest of the twenty-first century.
Throughout, low-tech and hi-tech, material and cyber technology are covered in
parallel. Co-evolutionary processes intertwine the crime and crime prevention
sides and introduce significant symmetry, so a proper understanding of the past,
present and future role of technology in the security field can only be achieved by
considering them together.

The Nature of Technology

Although technology is a rich and complex aspect of human life, somewhat surpris-
ingly, until recently the field has had few theorists of its own. But Mitcham (1979)
identifies four dimensions: artefact (tools, manufactured products etc.), knowledge
(scientific, engineering, technological know-how, plus insight from social and phys-
ical sciences), process (problem-solving, research and development, invention,
innovation), and volition (ethics, technology as social construction). Arthur’s recent
(2009) seminal attempt to provide a theory of technology characterises it on differ-
ent scales: as a means to fulfil a particular human purpose; an assemblage of prac-
tices and purposes; and the entire collection of devices and engineering practices
available to a culture. These levels interact with each other and the entire economy:
“As the collective technology builds, it creates a structure within which decisions
and activities and flows of goods and services takes place.” (p. 194)
For Arthur, technology starts with phenomena—natural effects (e.g. gravitation
or electricity) that exist independently in nature. Technology is organised around
central principles, which are the application of one or more phenomena for some
purpose; principles in turn are expressed in the form of physical or informational
components which are combined, often hierarchically, to meet that purpose.
Technological domains are toolboxes of potential, which are clustered around some
common set of phenomena or applied principles such as movement of mechanical
parts, or of electrons.
These frameworks equally apply to technology in the field of crime and crime
prevention, but need particularisation. Mitcham’s volitional dimension for example
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 321

could include the social institution of crime and the social forces of conflict between
individuals, or between individuals and wider social groups like the state. Arthur
refers to multiple purposes; extending these to the multiple stakeholders that hold
them is especially important in the case of criminal conflicts. As will be seen, these
extensions bestow special qualities on technology in crime and its prevention.

A Focus on Opportunity

Crime science centres on the immediate causes of criminal events and views oppor-
tunity in terms of two related perspectives. In the Rational Choice agenda (Cornish
& Clarke, 1986), an opportunity emerges when the offender perceives risk and
effort as low and reward high. Complementing this psychological approach is the
ecological Routine Activities perspective (Cohen & Felson, 1979) of likely offender
encountering a suitable target in the absence of capable guardians. At a higher eco-
logical level is the opportunity structure (Clarke & Newman, 2006)—the entire
pattern of available opportunities for crime.
At this level there are obvious affinities with Arthur’s “decisions and activities
and flows of goods and services” quoted above. But crime science’s familiar theo-
retical perspectives require some extension and integration to fully and efficiently
engage with technology. Here we discuss “static” aspects of these perspectives;
dynamic and change-related aspects are addressed in later sections.
Rational Choice does not cover prevention that works simply by defeating the
offender1 (e.g. with a massive concrete barrier), or completely designing out the
crime target (e.g. the disappearance of the vehicle tax disc in the UK in 2014 and its
replacement by electronic means of controlling revenue totally obliterated the crime
of forging these discs). Here, any choice has to be exercised at a more strategic level
(“How else can I make some money?”).
Tools and weapons are considered “crime facilitators”, and “control tools/weapons”
appears under “Increase the effort” in the 25 Techniques of SCP (e.g. Clarke & Eck,
2003). However, a direct and explicit theoretical treatment of technology is lacking.
In the “Hot Products” approach (Clarke, 1999) to identification of risk factors for
becoming targets of theft, some of these factors are directly related to technology—
in particular Concealable, Removable, Enjoyable and Disposable. (The same applies
to protective factors, for example those reducing the risk of theft of cell phones
(Whitehead et al., 2008).) Attempts to apply the Hot Products approach in a practi-
cal exercise in crime-proofing of new personal electronic items such as cameras and
phones (Armitage, 2012) revealed difficulties of linking up terms and concepts in
technology and crime science (Ekblom & Sidebottom, 2008).

1
Thanks to Richard Wortley for this point. But it is worth noting that the capabilities technology
brings mean that defeat may rarely be absolute. This is illustrated by the jewel thieves who drilled
through a metre of concrete to reach the safe deposit boxes of London’s Hatton Garden diamond
dealers in 2015—ironically using diamond-tipped cutters.
322 P. Ekblom

Within the core Routine Activities model, the capability of guardians is an obvi-
ous conceptual peg for preventive technology. This applies equally to the other
“crime preventer” roles that were introduced later (Clarke & Eck, 2003), namely
place managers (e.g. surveillance and access control technology) and handlers of
offenders (e.g. electronic tagging used to monitor curfew of convicted offenders).
At first sight, however, there is little that explicitly covers technology for offending.
Cohen and Felson (1979) did originally include offender capacity under “likely” but
this has got lost in recent years and unfortunately most writers now refer, too nar-
rowly, to the “motivated” offender.
The Conjunction of Criminal Opportunity (CCO—Ekblom, 2010, 2011 and
https://5isframework.wordpress.com/conjunction-of-criminal-opportunity/
accessed 17 June 2015) remedies some of these limitations. It seeks to integrate
the Rational Choice and Routine Activities approaches together with others on
both the situational and offender side, and to provide a consistent and all-encom-
passing conceptual framework and a unified terminology. CCO explicitly includes
offenders’ resources for committing crime (Ekblom & Tilley, 2000; Gill, 2005)
and can be readily extended to cover resources for the familiar preventer roles
discussed above (plus many more with its more flexible role concept, including
“engineer” or “designer”).
In brief, CCO offers twin perspectives: (1) on the proximal causes of criminal
events; and (2) on interventions in those causes to reduce the events’ likelihood and/
or harm. CCO comprises agents (offenders, crime preventers and crime promoters,
who inadvertently, carelessly or deliberately make crime more likely or harmful,
including those engineers who ignore crime risks); and entities (material or human
target; target enclosure; wider environment; and the criminal’s psychological, social
and material resources both for offending, and for avoiding offending). Under CCO,
a criminal event happens when an offender who is predisposed, ready and equipped
to offend (and lacking the resources to avoid offending) encounters, seeks or creates
a situation containing a target that is vulnerable, attractive or provocative, in an
enclosure and/or wider environment that is tactically insecure and perhaps motivat-
ing in some way, facilitated by the absence of ready and able preventers and perhaps
too by the presence of promoters. When these preconditions are met (and perceived
to be met) the offender decides to proceed, and acts. When they are blocked, weak-
ened or diverted by a security intervention, or by natural circumstances, the offender
either cannot so act; is not provoked or prompted to act (Wortley, 2008); or decides
on balance that the perceived reward is not worth the effort and risk.
This ability to map systematically and in detail offers a fuller view of proximal
criminogenic mechanisms, and the criminocclusive mechanisms that deliberately
block, weaken or divert them (Pawson & Tilley, 1997; Ekblom, 2011). Although
space is lacking here, each of the 11 elements of CCO can be focused on in turn (and
perhaps in combination) to examine the diverse roles technology, both material and
informational, can play in contributing to the causes of criminal events, and to inter-
ventions in these causes. For example, offender presence in an enclosure can be
controlled through automated access control (keeping them out) or electronic tag-
ging (keeping them in). Intelligent software can act in a preventer role (e.g. making
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 323

decisions and acting on suspicious financial transactions) or even, with automated


fraud, as an offender.
The explicit incorporation in CCO of resources for offending allows another
insight. In SCP, opportunity is typically considered an attribute of the situation.
But this is incomplete. An open window three floors up is only an opportunity to
an offender with the resources of agility, courage and/or climbing technology
such as a ladder. Opportunity is thus an ecological interaction between situation
and offender.
Additionally, the concept of opportunity makes no sense without specifying “to
do what”—i.e. there must be some purpose, which ultimately stems from an agent’s
predisposition, whether that agent is offender, preventer or promoter. Thus, we can
define opportunity as an ecological concept, relating to how agents encounter, seek
or create a set of circumstances in which their resources enable them to cope with
the hazards and exploit the possibilities in order to achieve their multiple goals. The
goals of course can be positive (getting the money …) or “hygiene” related (…
whilst avoiding arrest or injury).

Opportunities and Problems

Another conceptual element to consider is the relationship between opportunities


and problems. SCP methods are usually selected, implemented and evaluated
through a problem-oriented approach (e.g. Clarke & Eck, 2003; Goldstein, 1990).
But criminals have problems too, and there is a payoff for crime prevention in gen-
eral, and for discussing technology specifically, from highlighting a symmetry of
circumstance between offenders and preventers. With criminal conflict, problems
and opportunities are intimately entangled: one party’s opportunity will invariably
be another’s, or the state’s, problem.
From a neutral position, a problem is some set of environmental circumstances
that hinders an agent (or agents), equipped with a certain set of resources, from
immediately achieving a particular goal or goals. “Goals” is usually plural because
often the difficulty is in resolving some conflict between positive and hygiene
goals—burgle the house without getting caught; having tranquil enjoyment of a
house without expensive and unsightly fortification; tackling a burglary hotspot
without restricting pedestrian movement. In this sense, a problem is what stops
achieving an opportunity from being child’s play.
We can visualise the problem-opportunity relationship (Fig. 19.1a–e) as a circle
broken by an arc of “preconditions that need to be met before the opportunity can
be realised and the goal achieved”.2 The arc is of a greater or lesser angle representing
how demanding of effort and resources the problem is to solve with an acceptable
level of risk of harm, and of failure. If there are several problems we could envisage

2
Alternative imagery for problems/solutions includes that of ‘glass half-full versus glass half-
empty’; for multiple ones Arthur uses chains.
324 P. Ekblom

a b c

child’s play a harder problem a complex problem

d e

multiple conflicting goals a field of problems/ opportunities

Fig. 19.1 Opportunities and problems

multiple breaks; and we could represent multiple, conflicting goals as intersecting


circles with a common break. Beyond individual instances, we can envisage fields
of opportunity for crime and its prevention, the initial problem being where to dis-
cern a circle of opportunity; this is equivalent to (legitimate) entrepreneurs scanning
the field to discern just where they might be able to make some money.

Technology, Opportunities and Problems

How, then, does technology fit with opportunity, opportunity reduction and
problem-solving?
On the opportunity side, the resources and circumstances can obviously include
technological elements. But technology can be quite fundamental to defining oppor-
tunity. This is partly because it extends human capability to cope and exploit, and
partly because (following Arthur, 2009), it always has a purpose. Indeed, offenders
may have purposes for the technology other than those intended by the engineer/
designer—for example misusing or misappropriating it as discussed below.
On the problem side, technology can contribute to solutions by bridging gaps in
the circle of opportunity—for example, how to stop car alarms going off as intended
(problem for offenders), or inappropriately (problem for car owners and their neigh-
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 325

bours). It can even start an entirely fresh circle, opening the doors to new rounds of
problem and opportunity—as with the arrival of CCTV—monitoring misbehaviour,
or spying on changing rooms. Technology can help either party adapt to contextual
conditions necessary for offending or preventive mechanisms to be triggered (Tilley,
1993). Besides the social and architectural context these may include the prevailing
weather conditions such as sun, rain or snow. Likewise, the weather can cause tech-
nology to fail, as anyone whose car alarm has been set off by the wind will attest.
Besides prevention, technology can halt a criminal attack that has been launched
(for example personal attack alarms or “smokecloaks” to obscure vision), and miti-
gate the adverse consequences of crime—whether backing up the data on a stolen
cell phone, or providing for business continuity after a terrorist attack.3
And technology can resolve design contradictions (Ekblom, 2012a) or trade-offs
between security and, say, safety or profitability. An example is the millimetre-wave
airport body scanner that aids security but reduces privacy, where one resolution is
the substitution of a personal (point-and-laugh) body image on the operator’s screen
with a generic, computer-generated outline that still displays suspicious items. In
solving problems, technology thus enables rapid adaptation of the good or the bad
party to the challenges posed by its material and social habitat, to the pursuit of its
goals. But finding technological solutions that enable opportunities may not always
be straightforward. To quote Arthur, “At one end of the chain is the need or purpose
to be fulfilled; at the other is the base effect that will be harnessed to meet it. Linking
the two is the overall solution, the new principle, or concept of the effect used to
accomplish that purpose. But getting the principle to work properly raises chal-
lenges, and these call for their own means of solution, usually in the form of the
systems or assemblies that make the solution possible.” (2009:110) And one might
add that technology can also obstruct achievement of purposes when it fails the
user, engenders a vulnerability to attack, or is difficult to obtain without extra cost
and risk; and sometimes too when the opponent’s technology succeeds.
Cyberspace creates a new technological domain for opportunities and problems,
but whether recently emerged crimes in silico are merely always reconfigurations of
familiar clashes in vivo, with certain constraints removed (such as spatial conjunc-
tions, inertia, conservation of mass etc.), or entirely new ones is debatable. Certainly
the question of identity and trust in online transactions has become a huge issue.
Applying CCO to cyberspace, Collins and Mansell (2004:64) noted that trust fits
into the framework in several ways. “An Internet shopper who is too trusting may
act as a careless or negligent crime promoter, as may a system designer. Conversely,
being an effective crime preventer means being equipped with appropriate applica-
tions and systems. Offenders exploit misplaced trust, sometimes to an expert degree
and are aided by software and hardware based resources, for example, ‘skimming’
devices fitted into cash machines to clone cards.”
Taking the broader view, technological extensions of human agents’ capabilities,
and the technologically modified situations in which those capabilities are exer-
cised, combine to engender a range of opportunities for crime, and likewise for

3
Thanks to Benoit Leclerc for these ideas.
326 P. Ekblom

crime prevention. The combinations range from simple to complex; and from direct
routes to indirect, roundabout ones such as the mediating effect of technology on
people’s routine movements which lead them to crime-generating situations
(Brantingham & Brantingham, 2008). And there are always knock-on and interac-
tion effects between technologies and with other social and physical circumstances
generating unforeseen consequences (for example see Tenner, 1996).

The Dynamics of Crime: Technology, Scripts and Script


Clashes

Criminal events are unlike chemical reactions, where the ingredients are simply
brought together and in a single step reliably produce a brightly coloured precipitate
or nasty smell. The coming-together of proximal criminal ingredients may result
from influences at levels ranging from the individual offender creating the opportu-
nity—which is not a Routine Activity—to emergent societal influences such as mar-
ket forces. For its part, Rational Choice does not cover the actions linking successive
decisions. A complete picture of crime as opportunity therefore cannot be devel-
oped without incorporating a dynamic view.
Cornish (1994) developed the dynamic dimension with his seminal article on
crime scripts, boosting understanding of the procedure of crime commission, and
promising identification of particular pinch points in the script where interventions
might be targeted for maximum effect. But scripts can be better woven into the
opportunity/problem perspective presented here if they give more explicit attention
to goals, plans and resources.
Being instrumental, scripts can be influenced by technology in various ways as
problems are solved and opportunities realised. Problems come in hierarchies or
clusters—much as goals do in means-end relationships—e.g. a subsidiary problem
to breaking into the house could be the lack of a crowbar. Offenders must learn to
bypass alarms or pick locks. They must often take extra steps to obtain tools or
weapons; perhaps also to learn how they work, how they can be used, or even
hacked; and maybe to safely dispose of them or eliminate traces on them such as
from DNA or electronic usage data.
The tools themselves can shape or constrain criminal behaviour. Designers refer
to “persuasive technology” (Lockton, Harrison, & Stanton, 2008), and the idea that
devices (e.g. cash machines) have scripts “expected” of their users (Latour, 1992).
Certainly the properties of knives, locks or network routers influence the kinds of
action offenders can contemplate undertaking, and their performance during the
event itself. Material items are often misused, sometimes created, as props for
con-tricks or ambushes which may involve more or less elaborate scripting. The gay
hookup facility Grindr has for example been used to lure victims to robberies.
Applying the procedural dimension to risk factors reveals glossed-over subtle-
ties. For example, the Concealable factor in CRAVED hot products (Clarke, 1999)
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 327

may be criminogenic at the getaway stage when it is the thief who pockets the stolen
smartphone; but the same factor may protect the phone, safely in the owner’s pocket
or bag, at the target-seeking stage.
The procedural analysis of behaviour applies to preventers as well as offenders,
covering all the above considerations. The preventer’s script may be quite closely
interrelated to the offender’s script, for example in collecting money from a cash
machine, and stealing or robbing it. Especially significant for technology is the
concept of the script clash (Ekblom, 2012a). This is where the offender’s script
engages with the preventer’s in such issues as:
• Surveill v conceal
• Exclude v permit entry
• Wield force v resist
• Conceal v detect criminal intent
• Challenge suspect v give plausible response
• Surprise/ambush v warning
• Trap v elude
• Pursue v escape
• Foster trust v become suspicious
• Constrain v circumvent4
Technology can favour one side over the other, creating an opportunity either for
crime, or for prevention, variously relating to targets (e.g. resistant or vulnerable,
concealable or detectable), enclosures (hardened or vulnerable, excluding or per-
mitting entry), environments (e.g. illuminated evenly so as to minimise scope for
ambush, or with deep shadows) and resources for offending (e.g. tools with ambigu-
ous or clear-cut criminal purpose).
Clashes are the fulcrums on which prevention must be designed. The security
design task is to arrange the situation to favour preventers over offenders.
Discriminant technologies are often crucial—for example the swing-down fire
escapes enabling residents to flee a burning building, but hindering offenders from
entering; and the “what you know, what you have and what you are” discriminators
in ICT security (respectively, passwords, tokens and biometrics).

Technological Change

The original Routine Activities article (Cohen & Felson, 1979) theorised how
changes in the weight/bulk of items like TV sets over several decades made them
more suitable as targets for theft. But we can view such changes over greater times-
cales—indeed Felson and Eckert (2015) argue that technological change is of major
importance in understanding longer-term trends in the crime and crime prevention

4
Thanks to Benoit Leclerc for this clash.
328 P. Ekblom

field, illustrating the point with, for example, crime impacts over history of the
adoption of horse transport, and the move into towns and cities.
Although early technologies (such as Acheulian stone hand-axes introduced by
Homo erectus) endured for hundreds of thousands of years, under Homo sapiens
technological change has become the norm. Ogburn (1922), an early historian of
technology, saw technological change as the primary driver of social change. In
understanding change processes, subsequent debates have occurred over the relative
predominance of technological or resource push versus market or demand pull (see
Arthur, 2009; and Sten, 2014 for a thorough review of the development of thought
in this area). The current view seems to be that of a system of mutual influence and
feedback loops between supply and demand for technology, albeit with technology
taking a significant and often-leading role. Technologically induced changes diffuse
through the wider society at differential rates, leading to cultural lags in adjustment
(Ogburn, 1922). Many commentators (e.g. Arthur, 2009) have noted that the rate of
technological change has been accelerating, which can only amplify such lags and
any negative consequences, e.g. from crime prevention techniques trailing those of
crime commission.

Cultural and Biological Evolutionary Perspectives

Accounts of the change process have usually drawn on evolutionary themes, and
this is adopted here. The pattern of technological evolution has been variously seen
as slow and cumulative (as with Gilfillan’s (1935) account of the development of
ships through many individual inventions), or operating on a range of saltatory
scales. According to Arthur (2009) these range from small “standard engineering”
advances or tweaks, to more radical innovations (such as the leap from steam to
electrical propulsion of locomotives) and those that transform whole industries and
beyond (such as ICT). The pressures shaping technological evolution variously
relate to market forces, networking and both physical and social constraints.
It is self-evident that viewing technological evolution as a subset of cultural evo-
lution can give useful insights. But by combining this with conventional, biological
evolution we can go beyond the human and ecological factors than the “vanilla”
version of crime science supplies, gaining insight from fresh concepts and a certain
detachment from conventional viewpoints.
Early attempts to draw on biological evolution to understand technology include
Samuel Butler’s idea of “Darwin amongst the Machines” (1863) but these resem-
bled superficial analogies (with curious visions of coupling steam engines!). Later
attempts have been more penetrating. Biological and cultural evolution have previ-
ously been viewed as rivals for explaining human behaviour (e.g. see Roach &
Pease, 2013), but the scope for fundamental tie-ups between them has increased
(e.g. Godfrey-Smith, 2012). In fact, “Universal Darwinism” (Nelson, 2007) envis-
ages a common “evolutionary algorithm” (Dennett, 1995) comprising variation of
individual organisms, practices or products; selection on the basis of adaptation to
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 329

some natural, social or commercial environment; and replication or transmission


whether through genes, blueprints or imitating/copying the product.
Detailed differences are, however, instructive. For example, Arthur (2009) notes
the relative rarity of combinatorial mechanisms in biological evolution (such as the
symbiotic coming together of various bacteria and archaea to generate the great leap
forward of eukaryotic cells, supporting all advanced life forms). This contrasts with
its pervasiveness in technological evolution, where variety is commonly generated
by bringing together new assemblages of components or principles: the jet engine,
for example, does not result from a gradual modification of piston/propeller engines.
To focus on biology, most tools and weapons evolved by animals are simply
grown as part of their anatomy (e.g. shells) or physiology (venom). But spiders man-
ufacture webs to trap insects, and termites build protective mounds. Technology
therefore does not require a large brain to create and exploit it, but it does need a
research and development process. Humans apart, this is not undertaken by con-
scious engineers and designers over years but mediated over millennia by the gene-
based evolutionary process in countless generations of individuals. Those spiders
which, by virtue of their genes, reliably and blindly produce web variants better-
adapted to serve their own foraging agenda in the immediate environment (catch
prey, avoid being eaten in turn, economise on effort and materials) survive, repro-
duce and pass those genes on so they come to predominate in the spider population.
Dawkins (1982) refers to the reach of genes to reliably generate structures and
behaviours beyond the body of their owner as the “extended phenotype” (the phe-
notype being the mature body, the result of protracted interactions between genes
and environment during the course of individual development). Humans do have
inbuilt tools and weapons—arguably, our hands. (Morgan and Carrier (2013) con-
troversially claim that fists have anatomically evolved as a specialised fighting
structure.) But with the possible exception of implements for throwing (see below),
we have little in the way of fixed phenotypic extensions. The knuckle-duster
weapon, for example, is not produced by specific “build knuckle-duster” genes,
rather through our generic “ability to build tool/weapon” and “understand causal-
ity” genes conferring the manipulative rather than pugilistic aspects of our hands,
and associated brainpower to operate them purposively.
At the most general level our capacity to support cultural evolution, has itself
biologically evolved. Whiten and Erdal (2012) ponder the tripling of human
brain size in the last 2.5 million years, and explain it in terms of “the evolution
of a new socio-cognitive niche, the principal components of which include forms
of cooperation, egalitarianism, mindreading (also known as “theory of mind”),
language and cultural transmission, that go far beyond the most comparable phe-
nomena in other primates. This cognitive and behavioural complex allows a
human hunter-gatherer band to function as a unique and highly competitive pred-
atory organism.” [2119]
The evolution of technology has been a significant part of this process (although
the technological history literature seems to have paid scant attention to the wider
cultural evolution material). The anatomy and fine neurological control of the hand,
and the construction and use of tools, together constitute a powerful example of
330 P. Ekblom

genetic/cultural co-evolution (Tocheri, Orr, Jacofsky, & Marzke, 2008), a process


whereby cultural and biological changes amplify and channel one another. Elsewhere
on the body, paleontological and comparative studies (e.g. Roach, Venkadesan,
Rainbow, & Lieberman, 2013) suggest that the unique weapon-throwing capacity of
humans involved feedback between changes to arm, shoulder and back anatomy and
the development of projectiles, a process which may have begun with Homo erectus
some two million years ago.
From a broader perspective Godfrey-Smith (2012) identifies macro-level,
“cultural phylogenetic changes” such as the Neolithic Revolution’s shift from
hunting/gathering to farming. These are comparable to Arthur’s (2009) sugges-
tion that major, transformative leaps in technological ability occur when we
switch to a new domain, e.g. from mechanical to cyber. The cultural/technological
evolution of farming introduced a phase-change in human existence. It increased
population density, led to ownership of fixed parcels of land and other property,
plus the development of written recording of that ownership; fostered emergence
of hierarchies; and enabled the development of societal roles with specialist skills
unrelated to subsistence. In turn, these ultimately technologically induced changes
together with a cascade of others including the invention of mass transportation,
can be said to have drastically and progressively reshaped the routine activities of
society (Cohen & Felson, 1979; Felson & Eckert, 2015); the places the activities
occur in and the journeys between them (Brantingham & Brantingham, 2008);
and the nature and supply of targets, tools and weapons. Such changes have gen-
erated both readiness to offend (for example via more sources of conflict), and
made possible many more criminal opportunities leading to more frequent, and
more diverse, criminal events. And major phase changes continue to unfold—we
are in the midst of those stemming from the emergence of ICT, bioengineering
and climate change.
Farming, fire, and the axe, in enabling forest clearance in favour of grassland,
illustrate a further evolutionary concept: niche construction. This fundamental, but
only recently recognised process (Laland, Odling-Smee, Feldman, & Kendal,
2009), is where a given species pervasively shapes its own environment to its own
advantage and simultaneously adapts to survive within it, as with corals building
entire reefs out of their limestone skeletons. We humans have come to constitute and
determine our own environment in both social and technological terms, for better
and for worse.

Biological Lag

To Ogburn’s cultural lag, we may add biological lag. Evolutionary psychology


(e.g. see Roach & Pease, 2013; Ekblom, Sidebottom, & Wortley, 2015) explores
the possibility that human genes, including those influencing our perceptions
and behaviour, remain adapted to life in the Pleistocene epoch (which ended
some 11,000 years BP), when we were hunter-gatherers living in small mobile
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 331

bands with limited weapons and tools. The remarkable human capacity for
cooperation described above evolved in this period, creating the backdrop
against which crime must be understood. But unfortunately we are also pretty
accomplished at conflict between individuals, and between groups; and in
inventing, using and improving tools and weapons in both cooperation and con-
flict. Behavioural tendencies appropriate for the Pleistocene circumstances—
where time, space, materials and local population size provided natural
constraints on conflict, differential wealth, things to steal and violence—are
now inappropriate where weapons of easy, stand-off killing and mass destruc-
tion are available together with a cornucopia of portable high-value goods.
Inappropriate, too, when we are enclosed in vehicles where lack of natural
empathic signals between conflicting individuals can perhaps unleash road rage.
Cultural evolution has largely been able to compensate for inadequate psycho-
logical/ecological controls, including through the development of social con-
trols relating to reputation; and societal institutions such as law, criminalisation
and enforcement. But because these remain imperfect, direct interventions to
improve security remain necessary (Schneier, 2012).

Technology, Disruption, Co-evolution and Arms Races

Thwarted commercial burglars can simply return to attack a fence with a more
powerful bolt-cutter—an example of tactical displacement. But the offensive or
defensive tools themselves may change, and the balance of technological advan-
tage between offenders and preventers alters over time. Technological historians
have long identified perturbations and disruptions of a more general nature
(Christensen & Raynor, 2003; Ogburn, 1922). Disruptive trends like automation,
remote monitoring and operation, self-design and production, mass customisa-
tion, miniaturisation, portability including power supply, and the break between
appearance and functionality, will all keep perturbing the balance between offend-
ers and preventers.
Change can be exogenous (driven by external forces like the arrival of the motor
vehicle, acting as crime resource and target par excellence, or police patrol car);
endogenous, by the playing out of script clashes involving the adaptation and
counter-adaptation of criminals and preventers to one another’s tools, techniques
and weapons; or more usually, a combination in which exogenous changes perturb
the endogenous ones. However, the nature of that perturbation, and the interactions
between different perturbations, can only be resolved at the level of the fine detail
of particularities and perhaps only in retrospect. An example is the digital TV set-
top box, intended to allow existing analogue TV sets to receive digital stations. A
compact (and for a few years, widely available) object initially costing around
£100, we might have expected the box to become a hot product—until the TV
service providers decided to subsidise the cost and instead make their money from
the service charge.
332 P. Ekblom

Co-evolutionary Struggles

In the short term one can imagine the mutual adaptation of conflicting scripts—the
first bicycle-parking script might have been “cycle to cake shop, leave bike outside,
buy cake, return, mount bike and pedal off”; the first bike theft script “see unat-
tended bike, get on and pedal off”. Soon these would be followed by various elabo-
rations such as “lock bike”, “break bike lock” etc. In the medium term come “crime
harvests” Pease (2001), in which some product, say the mobile phone, is designed
and developed in a way that is naïvely vulnerable to crime and attractive to offend-
ers (a failure to “think thief”—Ekblom, 1997). Soon after coming on the market it
becomes both a popular purchase and a popular steal. This is usually followed by
desperate commercial or governmental measures to retrofit security which often
result in clunky, user-unfriendly or unreliable products.
In the longer term, such adaptations and counteradaptations can extend into pro-
longed co-evolutionary struggles (Ekblom, 1999; Sagarin & Taylor, 2008). These
are also known as arms races or “Red Queen’s games”, where you have to keep
running merely to stay in the same place (from Alice Through the Looking Glass—
see van Valen 1973; Dawkins and Krebs 1979). Classic examples are the evolution
of the safe (Shover, 1996), coins and banknotes, and more recent means of payment
such as online purchases. Once started, arms races may unfold at an irregular pace
(perhaps resembling the biological “punctuated equilibrium” of Eldredge and Gould
(1972)). At any point in such a criminal co-evolutionary sequence, we may encoun-
ter further harvests in the form of breakouts or “evolutionary surprise attacks”
(Tooby & DeVore, 1987), where a new tactic, tool or weapon becomes available or
is invented out of the blue and, for a while at least, overwhelms the opposition’s
defences. Imagine, for example, the devastating effect on cyber security if someone
discovered how to predict/identify the very large prime numbers relied on in most
security protocols.
Historical changes, and co-evolution especially, mean that knowledge of what works,
including technological solutions to crime problems, is a wasting asset that needs con-
tinual replenishment by new sources of variety. A contemporary example here is what
happens when the automotive industry rests on its laurels. A convincing case can be
made (see Chap. 22) that the “security hypothesis”—sustained technological and proce-
dural improvements in the security of homes, vehicles, shops etc.—accounts for the
striking crime drop seen over the last two decades. A significant contributor to these
improvements has been the inclusion (e.g. mandated by EU Directive) of immobilisers
in vehicles; Brown (2013) thoroughly reviews the evidence. Recently, however, and as
Brown anticipated, car thieves have managed to circumvent the security of keyless top-
end models such as the Land Rover Evoque (www.bbc.co.uk/news/technology-29786320
accessed 20 February 2015). These are currently disappearing into shipping containers
and heading abroad so fast that insurers are declining cover unless, say, cars are parked
off-street and primitive security devices like add-on steering wheel locks fitted. A similar
reversion to past technology is evidenced (www.theguardian.com/world/2013/jul/11/
russia-reverts-paper-nsa-leaks accessed 20 February 2015) by the Russian Federal
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 333

Guard Service. Following the Snowden affair, they have purchased old-fashioned type-
writers in order to deter leaks, by the traceability of individual machines’ key imprints.
In asymmetrical warfare, insurgents sometimes gain advantage by reversion to old-
fashioned weapons or means of communication.

Accelerants

Co-evolution through conflict, as just described, constitutes a powerful accelerant


of technological change in both criminal and military arenas since the two oppos-
ing sides focus sharply, consistently and persistently on countering one another’s
resources and capabilities. But co-evolution unfolds against a background of fur-
ther accelerants. Ogburn (1922) and later technological historians (see Sten, 2014)
have identified factors including increased population size enabling more people to
invent things; a greater stock of pre-existing technologies to combine; and com-
munications media enabling recording and dissemination of inventions and tech-
niques (including lock-picking sites on the Internet (Ekblom, 2014b)), and
capitalistic competition. Arthur (2009) additionally sees a qualitative change, with
the human economy becoming increasingly generative—shifting from optimising
fixed operations, towards creating new and flexible combinations and offerings for
the market.
The last relates to the biological concept of the evolution of evolvability (Dawkins,
2003). This refers to the fact that some organisms evolve sets of body-plan genes
that facilitate the orderly and efficient generation of variety. Moreover this is not just
random variety, where the chances are that such a spanner dropped into the works
of a complex machine would be more likely to harm than help, but plausible vari-
ants (Kirschner & Gerhart, 2005) with a good chance of survival and possibly of
conferring benefit.
The same process can be seen with cultural and especially technological evolu-
tion, and in fact we can see processes of combination, co-evolution, modularity and
evolution of evolvability coming together. In crime, facilities like script kiddies
enable less-accomplished programmers to generate computer viruses. 3D printers,
originally design prototyping tools, have been used to boost criminals’ own capacity
in, say, manufacturing accurately fitting and realistic-looking scanning mouthpieces
for ATMs to read/transmit customers’ card details; and in rapidly updating the
shapes as soon as the bank security team modify the ATM front panel (Krebs, 2011).
This aspect is more significant than the printers’ claimed ability to produce working
firearms (e.g. www.wired.com/2014/11/atlas-314-3-d-printed-guns-bullets/
accessed 3 March 2015). There is now also on the market an Internet-of-Things kit
and support service for connecting up and remotely activating whatever one wants
(www.bbc.co.uk/news/technology-31584546 accessed 3 March 2015). One imag-
ines this will interest terrorists and other criminals.
334 P. Ekblom

Gearing Up Against Crime

In the face of co-evolution, accelerants, and the background of dramatic


changes in technologies and their applications, the appropriate strategic
response for professional preventers is to try to out-innovate adaptive offend-
ers. Anything less risks winning battles, but losing campaigns. “Gearing up
against crime” approaches (Ekblom, 1997) suggest ways of doing this. Some
are lessons transferred from a range of other evolutionary struggles such as
arms races in the military domain, human versus nature (e.g. antibiotics versus
resistant bacteria, pests versus pesticides) or the purely natural (e.g. predator
versus prey, immune system versus pathogens) (Sagarin & Taylor, 2008).
Practically speaking, running arms races (more fully covered in Ekblom, 1997,
1999, 2015 and below on innovation) involves generating plausible variety of
responses by relying on theory and interchangeable practical elements; build-
ing in the capacity for security upgrades, especially “broadcastable” ones as
with Windows security patches; and developing security “pipelines”, as with
bank cards and satellite TV decoders, such that as soon as offenders crack one,
a new one is slotted into place.
It is prudent to establish systems for spotting and quickly reacting to new tech-
nologically enabled or targeted crime. But given the lead time to develop security
functionality, anticipation plays an important role in prevention. The traditional
anticipatory method of the problem-oriented approach to crime prevention—
induction of risk and protective factors from past patterns of hotspots etc.—has
been applied to “hot products” (those whose design and exposure render them at
greatest risk of theft: Clarke, 1999; Armitage, 2012). But this struggles with non-
linear changes in technology. However, alternative, foresight- or horizon-scan-
ning-based approaches are applicable: see, for example the UK Government’s
Foresight Programme activities covering crime in general (DTI, 2000) and cyber-
crime (www.gov.uk/government/publications/cyber-trust-and-crime-prevention
accessed 3 March 2015). Technology roadmapping, which seeks to connect future
requirements with upcoming trends in technology (e.g. www.technology-road-
maps.co.uk/secure_environment/ accessed 3 March 2015), could be applied to
both crime prevention and the misuse of emerging technologies and technological
combinations by offenders.
Horizon-scanning can be rendered more systematic and rigorous by various
crime science approaches. On causes, the Routine Activities perspective can be
used to identify changes in any of its three causal components that might make
crime events more likely (Pease, 1997). The CCO can prompt more detailed ques-
tions along similar lines (Ekblom, 2002): what future technological changes might
affect offender presence, target vulnerability, offender resources etc.? We might
also ask what changes might tip the balance of particular script clashes. On risk and
protective factors, a generic approach—the Misdeeds and Security framework
(Ekblom, 2005)—asks how scientific and technological innovations might generate
opportunities for crime through:
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 335

• Misappropriation (theft)
• Mistreatment (damage or injury)
• Mishandling (e.g. smuggling, data transfer)
• Misrepresentation (fraud)5
• Misbegetting (counterfeit)
• Misuse (as tool or weapon)
• Misbehaviour (for spraying graffiti for example)
• Mistakes (e.g. false alarms)
More specific risk factors for, say, misappropriation of products can be devel-
oped, as with CRAVED (Clarke, 1999). In turn, we can link these factors to generic
anticipated trends in technology, including greater miniaturisation and less use of
energy, to spot upcoming criminogenic changes.
Equivalent protective factors/opportunities for prevention are:
• Secured against Misappropriation, e.g. vehicles with built-in immobilisers
• Safeguarded against Mistreatment, e.g. street signs that avoid couching regula-
tions in provocative, confrontational terms
• Scam-proofed against Mishandling, Misbegetting and Misrepresentation, e.g.
fold-over airline baggage labels concealing holidaymakers’ addresses from bur-
glars’ touts; or anti-copying functions within DVDs
• Shielded against Misuse, e.g. one-time syringes
• “Sivilised” against Misbehaviour, e.g. metro station seating shaped to discourage
rough sleeping
These factors can be used descriptively; or as a technology requirement specifi-
cation e.g. by the police (Ekblom, 2005) intended to encourage the development of
the appropriate preventive capabilities. This last is an instance of market pull rather
than technology push.

Technology, Innovation and Design

Arthur (2009), building on Ogburn (1922), emphasises the importance of combina-


tion of prior, especially modular, elements of technology in generating new prod-
ucts. With crime prevention, systematically generating plausible variety at the level
of principles derives from the use of tested theory and what-works evidence (Ekblom
& Pease, 2014). This is the case whether that theory and research relates to techno-
logical or social domains (e.g. how to make wall-penetrating surveillance work, and
how to use it with what effect on offenders). That theory must be in a suitably ana-
lytic, integrated and accessible state to generate variety at a sufficient pace to out-
innovate adaptive offenders against the backdrop of social and technological change,
as noted. Arguably, CCO fits this requirement, being a suite of generative, analytic

5
Thanks to Benoit Leclerc for Misrepresentation, and Ken Pease for Misbegetting.
336 P. Ekblom

principles for preventive intervention, which channel more or less complex mecha-
nisms through practical intervention methods. (In this, in itself it resembles Arthur’s
(2009) account of a domain of technology.) Innovation in prevention can also draw
on recombination of modular practice elements (e.g. how to mobilise people as
preventers) (Ekblom, 2011).
However, concepts can only be realised by people. We need professional design-
ers closely working with practice-experts and users for the theory and practical
know-how to come together to generate intervention measures that both work in
principle and are workable in practice; and moreover which meet a range of other
requirements such as cost, aesthetics, durability, a small carbon footprint (Armitage
& Monchuk, 2011), business continuity and public safety. This takes design beyond
the pure homespun practicality of the police and hard-edged engineers. Recent
reviews of design and crime are in Ekblom, 2012a, 2014, and see also www.desig-
nagainstcrime.com and www.designingoutcrime.com/.
The design process needs hefty doses of intuition, inspiration and creativity,
helped by developing a “think criminal” mindset, and always being alert to the
need to “reframe” the presenting problem (Lulham, Camacho Duarte, Dorst, &
Kaldor, 2012). But it must also be systematic, constrained and supported by theo-
retical and methodological discipline. The use of frameworks like CCO, scripts
etc. as discussed, and the more specific situational perspectives behind them, can
contribute to this. One approach to feeding crime science into design is the Security
Function Framework (Ekblom, 2012b; Meyer & Ekblom, 2011) for specifying
criminocclusive products, i.e. those that reduce the possibility, probability and
harm from criminal events. This framework seeks to develop a rationale for secure
designs in terms of
• Purpose (what/who are the designs for, i.e. to reduce what crimes, and serve what
other goals, for which stakeholders?)
• Niche (how do they fit within the security ecosystem? Inherently secure prod-
ucts, dedicated security products, or securing products which confer protection
as a side-benefit to some main function like being a handbag)
• Mechanism (how do they work, causally, to serve security and other goals?)
• Technicality (how are they constructed, of what materials, and how are they
operated?)
A counterpart framework could be developed for criminals’ tools and weapons.
Clarke and Newman (2006) have developed the acronym MURDEROUS to describe
weapons of choice for terrorists: Multipurpose, Undetectable, Removable,
Destructive, Enjoyable, Reliable, Obtainable, Uncomplicated and Safe.
A final approach in the engineering innovation context is TRIZ, the theory of
inventive principles (Ekblom, 2012a; www.triz-journal.com/ accessed 3 March
2015). In brief, TRIZ facilitates the engineering design process by bulk scrutiny of
past inventions and inventive trends. The TRIZ approach has developed a generic
set of 40 inventive principles, abstracted widely across vast numbers of patents; 39
recurrent contradiction principles that routinely face designers seeking to meet con-
flicting user requirements, or technical tradeoffs such as strength versus weight.
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 337

Contradictions and tradeoffs are inherent in crime prevention: there is the central
contradiction between “offender wants the item” versus “owner wants the item”;
and in practice there are specific script clashes, as previously described. There are
also contradictions within security, such as “higher fence keeps intruders out” ver-
sus “blocks natural surveillance”; and “lighting means more security but greater
carbon footprint” (cf. Armitage & Monchuk, 2009). Each contradiction constitutes
a problem as previously defined, in terms of an obstacle to someone’s achievement
of their goals. A central procedure within TRIZ is to identify two principles in con-
flict, then look up to see what inventive principles were previously used to resolve
that contradiction. New technology has always been at the forefront of such resolu-
tions—for example, the arrival of the internal combustion engine relaxing the
strength/weight trade-off, so the military tank simultaneously has both armour and
manoeuvrability, unlike its predecessors the armed knight and the cavalryman, who
had one or the other.
A final aspect of TRIZ worth mentioning is its identification of a range of evolu-
tionary trends in various technological fields (e.g. a trend from fixed links between
components, to hinged links, to infinitely variable links like bicycle chains, to elec-
tromagnetic fields). Knowing such trends can help us anticipate where the next
improvement in some product, process or system might be expected to come from,
whether introduced by legitimate engineers or criminal ones. In this respect, TRIZ
provides yet another, more technologically particular, evolutionary perspective to
those already described.
Beyond the technical side of innovation, designers and those that employ them
must consider the business dimension of market forces in spurring development of
products and getting them sold. Space precludes further discussion but it is worth
briefly noting that opportunity relating not to criminal conflict but commercial com-
petition (e.g. McElwee & Smith, 2015; Trott, 2005), raises interesting theoretical
and practical issues. And Arthur (2009) considers demand too limited a concept to
depict processes at the level of the economy, preferring to refer to “opportunity
niches” (e.g. p. 174). There was no specific demand for penicillin, for example, until
it was discovered, but there was a broader need for something to be done about
bacterial infection. Working this concept into crime science may prove rewarding.

Weaknesses of Technology

Technology has its drawbacks for crime prevention as much as in any other domain
of application. Solution- or supply-driven approaches to problems can canalise
responses, constraining both current interventions and future adaptability. The rush
into public-space CCTV surveillance, despite indications of restricted utility (Welsh
& Farrington, 2008) reflects this. Investment in technology of the rigid, capital-
intensive kind can hinder adaptation to changes and enhance the lag behind adaptive
criminals—a case of “field obsolescence”. Techno-fixes can be superficial, of the
“bolt-on, drop-off” kind. However well-designed and constructed, they can also fail
338 P. Ekblom

at the interface with humans if that part of the preventive system is inadequately
integrated or updated. One example is the Grippa Clip (Ekblom et al., 2012), a care-
fully designed and trialled clip for preventing theft of customers’ bags by anchoring
them to pub/café tables. Despite plaudits from customers, police and bar staff, and
successful utilisation by customers in bars in Barcelona and a café at a London sta-
tion, in one UK bar chain they were ignored. Indications were that utilisation
depended on the crime climate (serious enough in Barcelona for bar staff to not be
shy about alerting customers to these security aids), and on staff motivation and
commitment (in the UK bars this seemed lacking, but was firmly present in the
London café, which “nurtured” its personnel and established mutual commitment
with the company).
Technology can make things worse, with false burglar or car alarms wasting
police time and annoying neighbours. It can also be self-defeating: clunky password-
based security systems overload the memory and exceed an employee’s “compli-
ance budget” (Beautement, Sasse, & Wonham, 2008)—that proportion of their
work time and effort they are willing to dedicate to security procedures (beyond
which they cut corners, like writing passwords down).
But none of these failings are inherent limitations of technology—only technol-
ogy that is over-relied upon in isolation from human/system considerations; poorly
designed (e.g. to be abuser-unfriendly without being simultaneously user-friendly;
and note the DAPPER acronym in Chap. 22); rigid and constraining in the face of
the messy complexity of real life; and incapable of being adapted to changing pat-
terns of risk during its lifetime of use, through material or software upgrades
(Ekblom, 1997).
A broader issue is complexity. Arms races, human system failures etc. show that
technology is often embedded in complex adaptive systems, where introducing
change at one point is followed by the various agents adjusting to the change, and to
each other’s new stance (a theme inaugurated by Ogburn (1922)). Add to this the
complexity of interactions between technologies (think how many technologies
came, or were brought, together to enable the attack on the Twin Towers) and we
can see how unpredictable the crime (or crime prevention) impact of individual
technological developments can be. Crime prevention faces a rich and challenging
future in which sense-making rather than watertight, orderly explanation and pre-
diction may well play a greater part in practice, as it is doing in the economy as a
whole (Arthur, 2009).

Conclusion

Technology forms a pervasive and routine part of the human ecosystem, and consti-
tutes the means of individuals and groups to extend their native capabilities to adapt
to and exploit that ecosystem, whether for legitimate or illegitimate purposes. It is
not an optional bolt-on component to crime science, but is fundamental to concepts
of opportunities, problems and solutions.
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 339

Technology both creates and solves problems for offenders and crime preven-
ters, and—as the other side of the coin—helps to block or generate opportunities for
these parties. It plays many causal roles in the clashing scripts of offenders and
preventers: in the language of CCO it can produce or modify targets, enclosures and
wider built environments; enable or restrict presence in the crime situation; and sup-
ply resources for offending, avoiding offending and preventing offending. All of
these apply to both material and cybercrime, and to both hi-tech and low-tech prod-
ucts and systems. Since every criminal or preventive action has a potential techno-
logical dimension, those seeking to understand and intervene in crime must be
technically aware. Since it is so varied—and variable—crime scientists must under-
stand technology’s fundamental nature, as shown for example by Arthur, and be
able to grasp the specifics of technologies in particular crime situations, scripts and
script clashes, both functionally and … technically.
One underlying theme of this chapter has been the symmetry between problems
and solutions facing offenders, and the solutions and problems facing preventers.
But the point having being made, it may be worth exploring what asymmetries
remain, and their significance for prevention. To a certain degree, professional
developers of preventive technologies work in an intensive way, with considerable
investment and application of know-how, time and other resources in coming up
with solutions. This may also happen with organised or networked crime and terror-
ism, but in many cases the development process may be extensive, with thousands
of offenders “trying it on” in modest ways to develop ways to circumvent particular
locks, or surveillance cameras, or find loopholes in financial systems. The extensive
process resembles all those bacteria mutating away in their billions, with a few com-
ing up with just the right change in their metabolic repertoire to overcome the latest
antibiotic.
Technologies interact with one another and with social and environmental con-
texts, generating challenging levels of complexity and unpredictability. And tech-
nology evolves, under the drivers and constraints of market forces, societal
requirements, the material laws of physics and chemistry, the logical rules and
conventions of ICT, the ingenuity and innovativeness of engineers and designers, a
never-ending succession of technologically induced opportunities and problems and
in some cases, the co-evolutionary arms race between offenders and preventers.
This evolution unfolds in biological, cultural and specifically technological domains.
(Interestingly, Arthur (2009) points out how technology itself is developing ever
increasing resemblance to biology in its complexity, adaptability and integration/
communication between parts.)
For the foreseeable future, technology will be a significant influence in both gen-
erating and reducing crime. However, successful crime prevention through technol-
ogy cannot be based on some narrow and linear technological determinism but must
be undertaken through the understanding of and influence upon complex social/
physical/informational systems. The design of technology must thread its way
between diverse requirements and in particular to intelligently discriminate in
favour of rightful owners/users over criminal abusers and misusers. While a purely
technological approach to crime prevention has notable weaknesses and limitations,
340 P. Ekblom

technology that is carefully designed in line with the tested theories of crime sci-
ence, developed, updateable and updated on an appropriate timescale, and that is
well-integrated with the human parts of a security system, can make a significant
contribution to the well-being of individuals, organisations and society.
But criminals are forever seeking opportunities to misappropriate, mistreat,
misuse and misbehave with technology as they adapt to the changing socio-
technological environment, attempting to cope with its hazards and exploit its vul-
nerabilities. The Red Queen’s game will doubtless continue to run throughout the
remainder of the twenty-first century and beyond. Only strategic, evolutionary and
innovative thinking based on plausible theory and empirical research and develop-
ment can stop the nice guys finishing last. But it seems (www.bbc.co.uk/news/
technology-31416838 accessed 5 March 2015) they will have to contend with
3D-printed knuckledusters!

References

Armitage, R. (2012). Making a brave transition from research to reality. In P. Ekblom (Ed.), Design
against crime: Crime proofing everyday objects (Crime prevention studies, Vol. 27). Boulder,
CO: Lynne Rienner.
Armitage, R., & Monchuk, L. (2009). Reconciling security with sustainability: The challenge for
eco-homes. Built Environment, 35(3), 308–327.
Armitage, R., & Monchuk, L. (2011). Sustaining the crime reduction impact of Secured by Design:
1999 to 2009. Security Journal, 24(4), 320–343.
Arthur, W. B. (2009). The nature of technology. What it is and how it evolves. London: Allen Lane.
Beautement, A., Sasse, M. A., & Wonham, M. (2008). The compliance budget: Managing security
behaviour in organisations. In NSPW'08: Proceedings of the 2008 Workshop on New Security
Paradigms Workshop (pp. 47–58). New York: Association for Computing Machinery.
Brantingham, P., & Brantingham, P. (2008). Crime pattern theory. In R. Wortley & L. Mazerolle
(Eds.), Environmental criminology and crime analysis. Cullompton: Willan.
Brown, R. (2013). Reviewing the effectiveness of electronic vehicle immobilisation: Evidence
from four countries. Security Journal. doi:10.1057/sj.2012.55.
Butler, S. (1863). To the editor of the press, Christchurch, New Zealand, 13 June. Wellington:
Victoria University.
Christensen, C., & Raynor, M. (2003). The innovator's solution. Harvard: Harvard Business Press.
Clarke, R. (1999). Hot products: Understanding, anticipating and reducing demand for stolen
goods (Police research series paper, Vol. 112). London: Home Office.
Clarke, R., & Eck, J. (2003). Become a problem solving crime analyst in 55 small steps. London:
Jill Dando Institute, University College London.
Clarke, R., & Newman, G. (2006). Outsmarting the terrorists. London: Praeger Security
International.
Cohen, L., & Felson, M. (1979). Social change and crime rate changes: A routine activities
approach. American Sociological Review, 44, 588–608.
Collins, B., & Mansell, R. (2004). Cyber trust and crime prevention: A synthesis of the state-of-
the-art science reviews. London: Department for Business, Innovation and Science. Retrieved
January 21, 2015, from www.foresight.gov.uk/Cyber/Synthesis of the science reviews.pdf.
Cornish, D. (1994). The procedural analysis of offending and its relevance for situational preven-
tion (Crime prevention studies, Vol. 3). Monsey, NY: Criminal Justice Press.
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 341

Cornish, D., & Clarke, R. (Eds.). (1986). The reasoning criminal: Rational choice perspectives on
offending. New York: Springer.
Dawkins, R. (1982). The extended phenotype. Oxford: Oxford University Press.
Dawkins, R. (2003). The evolution of evolvability. In S. Kumar & P. Bentley (Eds.), On growth,
form and computers. London: Academic Press.
Dawkins, R., & Krebs, J. (1979). Arms races between and within species. Proceedings of the Royal
Society of London B, 205(1161), 489–511. doi:10.1098/rspb.1979.0081.
Dennett, D. (1995). Darwin’s dangerous idea. London: Penguin.
DTI. (2000). Turning the corner. Report of Foresight Programme’s Crime Prevention Panel.
London: Department of Trade and Industry.
Ekblom, P. (1997). Gearing up against crime: A dynamic framework to help designers keep up
with the adaptive criminal in a changing world. International Journal of Risk, Security and
Crime Prevention, 2, 249–265.
Ekblom, P. (1999). Can we make crime prevention adaptive by learning from other evolutionary
struggles? Studies on Crime and Crime Prevention, 8, 27–51.
Ekblom, P. (2002). Future imperfect: Preparing for the crimes to come. Criminal Justice Matters,
46, 38–40. London: Centre for Crime and Justice Studies, Kings College.
Ekblom, P. (2005). How to police the future: Scanning for scientific and technological innovations
which generate potential threats and opportunities in crime, policing and crime reduction. In
M. Smith & N. Tilley (Eds.), Crime science: New approaches to preventing and detecting
crime. Cullompton: Willan.
Ekblom, P. (2010). The conjunction of criminal opportunity theory. Sage Encyclopedia of
Victimology and Crime Prevention, 1, 139–146.
Ekblom, P. (2011). Crime prevention, security and community safety using the 5Is framework.
Basingstoke: Palgrave Macmillan.
Ekblom, P. (2012a). Happy returns: Ideas brought back from situational crime prevention’s explo-
ration of design against crime. In G. Farrell & N. Tilley (Eds.), The reasoning criminologist:
Essays in honour of Ronald V. Clarke (Crime science series, pp. 163–198). Cullompton:
Willan.
Ekblom, P. (Ed.). (2012b). Design against crime: Crime proofing everyday objects (Crime preven-
tion studies, Vol. 27). Boulder, CO: Lynne Rienner.
Ekblom, P. (2014). Designing products against crime. In G. Bruinsma & D. Weisburd (Eds.),
Encyclopedia of criminology and criminal justice. New York: Springer Science + Business
Media.
Ekblom, P. (2015). Terrorism—Lessons from natural and human co-evolutionary arms races. In
M. Taylor, J. Roach, & K. Pease (Eds.), Evolutionary psychology and terrorism. London:
Routledge.
Ekblom, P., Bowers, K., Gamman, L., Sidebottom, A., Thomas, C., Thorpe, A., et al. (2012).
Reducing handbag theft in bars. In P. Ekblom (Ed.), Design against crime: Crime proofing
everyday objects (Crime prevention studies, Vol. 27). Boulder, CO: Lynne Rienner.
Ekblom, P., & Pease, K. (2014). Innovation and crime prevention. In G. Bruinsma & D. Weisburd
(Eds.), Encyclopedia of criminology and criminal justice. New York: Springer Science + Business
Media.
Ekblom, P., & Sidebottom, A. (2008). What do you mean, ‘Is it secure?’ Redesigning language to
be fit for the task of assessing the security of domestic and personal electronic goods. European
Journal on Criminal Policy and Research, 14, 61–87.
Ekblom, P., Sidebottom, A., & Wortley, R. (2015). Evolutionary psychological influences on the
contemporary causes of terrorist events. In M. Taylor, J. Roach, & K. Pease (Eds.), Evolutionary
psychology and terrorism. London: Routledge.
Ekblom, P., & Tilley, N. (2000). Going equipped: Criminology, situational crime prevention and
the resourceful offender. British Journal of Criminology, 40, 376–398.
Eldredge, N., & Gould, S. (1972). Punctuated equilibria: An alternative to phyletic gradualism. In
T. Schopf (Ed.), Models in paleobiology. San Francisco: Freeman Cooper.
342 P. Ekblom

Felson, M., & Eckert, M. (2015). Crime and everyday life (5th ed.). London: Sage.
Gilfillan, S. (1935). Inventing the ship. Chicago: Follett.
Gill, M. (2005). Reducing the capacity to offend: Restricting resources for offending. In N. Tilley
(Ed.), Handbook of crime prevention and community safety. Cullompton: Willan.
Godfrey-Smith, P. (2012). Darwinism and cultural change. Philosophical Transactions of the
Royal Society B, 367, 2160–2170.
Goldstein, H. (1990). Problem-oriented policing. Philadelphia: Temple University Press.
Kirschner, M., & Gerhart, J. (2005). The plausibility of life: Resolving Darwin’s dilemma. New
Haven, CT: Yale University Press.
Krebs, B. (2011). Retrieved January 22, 2015, from http://krebsonsecurity.com/2011/09/
gang-used-3d-printers-for-atm-skimmers/
Laland, K., Odling-Smee, J., Feldman, M., & Kendal, J. (2009). Conceptual barriers to progress
within evolutionary biology. Foundations of Science, 14(3), 195–216.
Latour, B. (1992). Where are the missing masses? The sociology of a few mundane artifacts. In
W. Beijker & J. Law (Eds.), Shaping technology (pp. 205–224). Cambridge, MA: MIT Press.
Lockton, D., Harrison, D., & Stanton, N. (2008). Design with intent: Persuasive technology in a wider
context. In H. Oinas-Kukkonen, P. Hasle, M. Harjumaa, K. Segerståhl, & P. Øhrstrøm (Eds.),
Persuasive technology: Third International Conference, PERSUASIVE 2008, Oulu, Finland, June
4–6, 2008, Proceedings (Lecture notes in computer science, Vol. 5033). Berlin: Springer.
Lulham, R., Camacho Duarte, O., Dorst, K., & Kaldor, L. (2012). Designing a counterterrorism
trash bin. In P. Ekblom (Ed.), Design against crime: Crime proofing everyday objects. Boulder,
CO: Lynne Rienner.
McElwee, G., & Smith, R. (Eds.). (2015). Exploring criminal and illegal enterprise: New perspec-
tives on research, policy & practice. London: Emerald Group.
Meyer, S., & Ekblom, P. (2011). Specifying the explosion-resistant railway carriage—A desktop
test of the Security Function Framework. Journal of Transportation Security, 5, 69–85.
Mitcham, C. (1979). Philosophy and the history of technology. In G. Bugliarello (Ed.), The history
and philosophy of technology (pp. 163–189). Champaign-Urbana, IL: University of Illinois Press.
Morgan, M., & Carrier, D. (2013). Protective buttressing of the human fist and the evolution of
hominin hands. Journal of Experimental Biology, 216, 236–244.
Nelson, R. (2007). Universal Darwinism and evolutionary social science. Biology and Philosophy,
22, 73–94.
Ogburn, W. (1922). Social change with respect to culture and original nature. New York:
B.W. Huebsch.
Pawson, R., & Tilley, N. (1997). Realistic Evaluation. London: Sage.
Pease, K. (1997). Predicting the future: The roles of routine activity and rational choice theory. In
G. Newman, R. V. Clarke, & S. Shoham (Eds.), Rational choice and situational crime preven-
tion: Theoretical foundations. Aldershot, UK: Dartmouth Press.
Pease, K. (2001). Cracking crime through design. London: Design Council.
Roach, J., & Pease, K. (2013). Evolution and crime. London: Routledge.
Roach, N., Venkadesan, M., Rainbow, M., & Lieberman, D. (2013). Elastic energy storage in the
shoulder and the evolution of high-speed throwing in Homo. Nature, 498, 483–486.
Sagarin, R., & Taylor, T. (Eds.). (2008). Natural security: A Darwinian approach to a dangerous
world. Berkeley: University of California Press.
Schneier, B. (2012). Liars and outliers: Enabling the trust that society needs to thrive. New York:
Wiley.
Shover, N. (1996). Great pretenders: Pursuits and careers of persistent thieves. London: Westview
Press.
Sten, K. (2014). The emerging dynamics of innovation: The case of IT Industry in India. Master’s
thesis, Copenhagen Business School. Retrieved March 25, 2015, from http://studenttheses.cbs.
dk/bitstream/handle/10417/4797/katri_sten.pdf?sequence = 1
Tenner, E. (1996). Why things bite back: Technology and the revenge of unintended consequences.
New York: Knopf.
19 Technology, Opportunity, Crime and Crime Prevention: Current and Evolutionary… 343

Tilley, N. (1993). Understanding car parks, crime and CCTV: Evaluation lessons from safer cities
(Crime prevention unit paper, Vol. 42). London: Home Office.
Tocheri, M., Orr, C., Jacofsky, M., & Marzke, M. (2008). The evolutionary history of the hominin
hand since the last common ancestor of Pan and Homo. Journal of Anatomy, 212(4),
544–562.
Tooby, J., & DeVore, I. (1987). The reconstruction of hominid behavioral evolution through stra-
tegic modelling. In W. Kinzey (Ed.), The evolution of human behavior: Primate models
(pp. 183–227). New York: SUNY Press.
Trott, P. (2005). Innovation management and new product development, 2005. Upper Saddle River,
NJ: Prentice Hall.
van Valen, L. (1973). A new evolutionary law. Evolutionary Theory, 1, 1–30.
Welsh, B., & Farrington, D. (2008). Effects of closed circuit television surveillance on crime. A
systematic review. Oslo: Campbell Collaboration.
Whitehead, S., Mailley, J., Storer, I., McCardle, J., Torrens, G., & Farrell, G. (2008). IN SAFE
HANDS: A review of mobile phone anti-theft designs’. European Journal on Criminal Policy
and Research, 14, 39–60.
Whiten, A., & Erdal, D. (2012). The human socio-cognitive niche and its evolutionary origins.
Philosophical Transactions of the Royal Society B, 367, 2119–2129.
Wortley, R. (2008). Situational precipitators of crime. In R. Wortley & L. Mazerolle (Eds.),
Environmental criminology and crime analysis. Cullompton: Willan.
Chapter 20
Passive WiFi Radar: A New Technology
for Urban Area Surveillance

Kevin Chetty and Karl Woodbridge

Introduction

Wireless Fidelity (WiFi) technology has experienced a major and rapid global roll-
out since its introduction in the late 1990s and is now the standard communication
modality for accessing broadband internet worldwide. It is a pervasive technology
and ubiquitous in our everyday lives, operating in our homes, places of work and
study, right through to transport hubs and the coffee shops and shopping malls we
visit in our leisure time. WiFi is based on the IEEE 802.11 communications proto-
cols (IEEE, 1999) and defines the standards for wireless local area networks
(WLANs) which allow electronic devices to connect to the internet or exchange
data. By the end of 2011 it was estimated that 439 million, or 25 % of all households
worldwide had a WiFi network set-up (Wu, 2011). Additionally WiFi provider
iPass state that at the end of 2014 the number of global public WiFi hotspots reached
approximately 47.7 million.

Concept of Operation

Passive radar is sometimes referred to as parasitic or piggyback radar as unlike


traditional radar system systems which transmit their own radio-frequency (RF)
signals to interrogate targets of interest, passive radar systems make use of RF

K. Chetty (*)
Department of Security and Crime Science, University College London,
35 Tavistock Square, London WC1h 9EZ, UK
e-mail: k.chetty@ucl.ac.uk
K. Woodbridge
Department of Electronic and Electrical Engineering, University College London,
London, UK

© Springer International Publishing Switzerland 2017 345


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_20
346 K. Chetty and K. Woodbridge

signals which already exist in the atmosphere. It dates back to the earliest days of
radar (Griffiths, 1993) and interest in the field has varied cyclically over the years.
Griffiths (Griffiths, 2008) explains that the most recent resurgence is due to the
increased availability and affordability of both computing power and hardware, as
well as the transition from the more traditional illuminators of opportunity origi-
nating from television (Griffiths & Long, 1986) and FM radio (Lauri et al., 2007)
broadcast stations, to the escalating numbers of new digital communications pro-
tocols such as GSM, digital audio broadcasting, digital video broadcasting ter-
restrial (Poullin, 2005), and wireless networks based on the IEEE 802 standards
such as WiFi (802.11) and WiMAX (802.16). The result of this resurgence has
new wave of passive radar systems (Poullin, 2005; Tan et al., 2005; Zemmari
et al., 2014) which utilise these transmissions primarily for maritime and air sur-
veillance applications.
Passive radar systems essentially consist of a series of synchronised radio receiv-
ers that are spatially separated from the transmitter which they are using. This spa-
tial separation between the transmitter and receivers puts them into a class of radar
system known as bistatic radar (Willis, 2007).
The basic concept of WiFi passive radar is illustrated in Fig. 20.1: In many
typical urban environments there exist WiFi access points which allow people to
connect their electronic devices (e.g. laptops and smart-phones) to the internet.
The WiFi access points flood their surrounding areas with RF communication
signals (shown in blue) so as people move through these signal fields they scatter

Fig. 20.1 WiFi passive radar concept of operation. A person moving within the coverage area of
a WiFi access point scatters the signal field in all directions, which can then be detected by sur-
rounding passive receivers
20 Passive WiFi Radar: A New Technology for Urban Area Surveillance 347

the signals in all directions, and induce a frequency shift within the signal through
a phenomenon known as the Doppler effect (Gill, 1965). By placing a single or set
of surveillance receivers around the environment it is then possible to detect these
scattered echoes, and thus detect the presence of any moving targets of interest.
However, most urban environments are generally already cluttered with other
WiFi and interfering signals in the immediate vicinity so passive radar systems
must also employ a reference receiver to monitor the output from the WiFi access
point. The reference channel is then used to inform the surveillance receiver of
exactly which scattered signals to identify during the cross-correlation signal pro-
cessing. Signal processing can reveal whether or not moving targets such as peo-
ple have been detected in the surveillance receivers’ field-of-view, and if so,
information relating to the range of the target from the receiver, direction of
movement and velocity can be obtained.

Theoretical Feasibility Assessment

WiFi Passive radar is a new concept so in order to determine the feasibility of the
passive WiFi radar concept described above, two types of theoretical analyses were
conducted to predict and quantify its operational characteristics. These were (1) An
ambiguity function waveform analysis to gauge the accuracy to which WiFi signals
can detect moving targets and (2) A signal-to-noise ratio (SNR) performance simu-
lation to identify the distances over which such systems can operate.

Ambiguity Function Waveform Analysis

The ambiguity function is a mathematical tool used for analysing or designing


waveform signals in a radar system. It gives the radar designer an insight into the
achievable range resolution of the waveform being assessed, i.e. the minimum dis-
tance between two targets which can be distinguished by the system. It represents
the response of a matched filter to the signal for which it is matched, as well as to
mismatched Doppler shifted signals (Woodward, 1953). Table 20.1 shows the typi-
cal range resolutions of various communications waveforms which have previously
been used in passive radar:

Table 20.1 Range Source of signals Range resolution


resolutions of various
FM radio transmitter Few kilometres
opportunistic signals which
have been used in passive Digital video broadcasting Kilometre (approx)
radar (DVB) station
GSM cell phone Few 100 m
basestation
348 K. Chetty and K. Woodbridge

It can be seen from Table 20.1 that the range resolutions which can be achieved
by transmission signals previously used as the basis of passive radar systems are
suitable for maritime and air surveillance where a ranging error of between a few
100 m to a kilometre is acceptable. However, these range resolutions are far too
coarse for monitoring urban areas because a target position uncertainty of 100 m
cannot be used effectively. An ambiguity function analysis was applied to various
WiFi waveforms and the results revealed that a range resolution of approximately
12 m can be obtained using the current WiFi standards (Chetty et al., 2009), while
increased bandwidths in future versions of WiFi will allow range resolutions of less
than a metre. The range resolution of the current WiFi standards could be have uses
for identifying the positions of targets of interest in typical urban settings such as
within different rooms of a building or along a section of a street. Overall these
analytical outcomes suggest that the detection accuracy of passive WiFi radar is fine
enough to find surveillance applications in urban environments.

Signal-to-Noise Ratio Performance Simulation

WiFi operates over approximately 46 m indoors and 92 m outdoors. However, WiFi


passive radar relies on detecting signal reflections from personnel which are orders
of magnitude less than the power transmitted by WiFi access points (approximately
100 mW). It was therefore also necessary to conduct an analysis to gauge the dis-
tances over which WiFi passive radar can operate. To do this a simulation platform
was developed to assess the Signal-to-Noise Ratio (SNR). The SNR compares the
level of a reflected target signal in the surveillance receiver to the level of back-
ground noise and decreases as the range of the target increases due to propagation
losses. It is defined as the ratio of signal power to the noise power and is often
expressed in decibels (dB). For radar systems an SNR of greater than 10 dB is often
assumed to be the minimum level required to confirm that a target has been success-
fully detected (Skolnik, 1980).
Our simulation was built around the Radar Equation (Skolnik, 1980) which is a
fundamental and well-known mathematical construct in the field of radar that allows
the radar engineer to predict the SNR for a given set of operational and target param-
eters. The simulation result shown in Fig. 20.2 illustrates the area within which
target can be detected with SNR’s greater than 10 dB. In this case the WiFi transmit-
ter (Tx) and passive surveillance receiver (Rx) are separated by a distance of 30 m
and the target has a radar cross section of 1 m2 to represent a person. The surveil-
lance area (i.e. the area enclosed by the red 10 dB contour line) covers approxi-
mately a 100 m radius around the surveillance receiver which again suggests that
passive WiFi radar could be utilised as a surveillance technology in urban areas.
Note that the simulation is an idealised scenario and does not take into account cer-
tain other factors such as direct signal interference and multipath reflections which
can degrade the performance of the passive WiFi radar system.
20 Passive WiFi Radar: A New Technology for Urban Area Surveillance 349

Fig. 20.2 Predicted SNR performance of passive WiFi radar

Proof-of-Principle Experiments

Following our theoretical analysis of passive WiFi radar which indicated that the
current technology standards was suitable for urban area surveillance, the next
objective was to run a series of proof-of-principle experiments and demonstrate its
ability to detect people in a range of scenarios. Two experimental campaigns were
carried out; the first campaign investigated the ability of a passive WiFi radar sys-
tem to detect people within indoor environments under realistic conditions, i.e.
where the system has to contend with clutter and multipath components which can
mask target responses, cause fading of the signal or generate ghost targets. The
second examined the possibility of detecting people through walls which presents a
challenging barrier for the WiFi signal to penetrate whilst maintaining a minimum
detectable signal level.

Operation within an Indoor Environment

The detection experiments were set up within a typical office environment at


University College London (UCL) in order to generate the clutter and multipath
reflections one would expect in any real-world indoor urban scenario. A schematic
diagram of the test area is shown in Fig. 20.3. Two synchronised receivers which
worked on the WiFi frequency of 2.4 GHz were used as the passive reference and
350 K. Chetty and K. Woodbridge

Fig. 20.3 Experimental


set-up to demonstrate the
not to scale
concept of passive WiFi
radar Corridor

Offices

Target

Surveillance 7m
Receiver 46°

m
Office

10.5
7.6
m

6m

Reference
AP
Receiver
WiFi Access
Point

Offices

surveillance receivers (Derham et al., 2007). The reference receiver (shown in


green) was located in a conference room and used to monitor the signals being
transmitted by a WiFi access point which allowed the office workers on the whole
floor to access the internet. For any given experiment, one or more participants were
asked to walk or run along the corridor in order to create Doppler shifted reflections
of the original transmitted WiFi signal which would then be detected by a surveil-
lance receiver located 7 m away. The distances from WiFi access point to the target
position and surveillance receiver were 10.5 m and 7.6 m respectively.
The velocity-range graph plotted in Fig. 20.4 is the first demonstration of person-
nel detection in an indoor environment using passive WiFi radar. The result relates
to an experiment in which a person ran along the corridor (shown in Fig. 20.3).
Within experimental error (approximately ± 6 m in range and ± 3 m per second in
velocity) and achievable system resolution, the target response appeared within the
expected region of the graph. Note that on the line of zero velocity we also see the
responses from the stationary clutter in the office such as walls, filing cabinets, and
desks. Above and below the target response there are also low-power unwanted
artefacts which correspond to correlation sidelobes that are generated in the signal
processing stages.
20 Passive WiFi Radar: A New Technology for Urban Area Surveillance 351

20 0

–2
15
–4
10
–6

5 Target
Velocity (m/s)

–8

0 –10

–12
–5
Stationary Objects –14
–10 (walls, filing cabinets etc)
–16
–15
–18

–20 –20
0 10 20 30 40 Signal Power
Range (m) [dB]

Fig. 20.4 Detection of personnel within a high clutter indoor environment using passive WiFi
radar

Detecting People Through-Walls

Another series of experiments were conducted to investigate the ability of passive


WiFi radar for detecting targets through walls (Chetty, Bo, & Woodbridge, 2014),
and were carried out within the bricked pavilion structure shown in Fig. 20.5.
Similar to the indoor detection experiments, measurements were taken as a person
moved around within the coverage area of the WiFi access point. However, in these
experiments both the reference and surveillance receivers were located on the other
side of the wall. The result shown in Fig. 20.6 represents the detection of a person
walking in the surveillance cell. Note that in this case the target velocity (plotted in
terms of a Doppler shift) has a negative value as the person was walking away from
the surveillance receiver.

Key Benefits

The list below outlines some of the main features of passive WiFi radar technology
and describes the benefits it can offer over other types of surveillance and monitor-
ing systems:
352 K. Chetty and K. Woodbridge

Fig. 20.5 Experimental test site for examining the through-wall detection capability of passive
WiFi radar

25
20
15
10
5
Doppler (Hz)

0
–5
–10
–15
–20
–25
0 12 24
Range (m)

Fig. 20.6 Detection of a person through a wall in the experimental test scenario. Note that a cell
averaging technique and interference suppression algorithm have been applied to this result

• Uncooperative Detection:
Unlike some monitoring technologies which require a person to carry a coopera-
tive device such as an RFID tag or mobile phone, passive WiFi radar relies only
on detecting reflected signals from targets of interest and therefore operates in a
completely uncooperative manner.
• Motion Sensitivity:
20 Passive WiFi Radar: A New Technology for Urban Area Surveillance 353

The motion sensitivity, or Doppler resolution, of the system is determined by the


length of time which the system is able to record target signal reflections. As pas-
sive WiFi radar does not employ a dedicated transmitter it can allocate the major-
ity of its time budget to “listening” for target reflections and therefore offers
excellent motion sensitivity compared to active radars. Moreover this high mea-
surement precision aids Doppler-based tracking (Nadjiasngar, Middleton, &
Inggs, 2012) for monitoring moving targets over period of time. Our existing
passive WiFi radar prototype system is now able to detect small swaying move-
ments (Tan et al., 2014) and we are currently investigating its ability to sense the
small chest-wall movements of a person as they breathe.
• Cost:
Passive WiFi radar is a receiver-only system of a significantly lower cost than a
comparable active radar system which employs an expensive transmitter section
(Wu, 2011). Furthermore the system is not subject to spectrum licencing, which
itself can be very costly, because it does not transmit any of its own RF signals.
• Operational Environments:
Unlike closed circuit television (CCTV), passive WiFi radar can operate unaf-
fected in all weather conditions such as heavy rain, dense smoke and fog, and
when it is completely dark.
• Through-Wall Detection:
As demonstrated in our previous research, passive WiFi radar is able to detect
people through walls at standoff distances (Chetty, Smith, & Woodbridge, 2012;
Chetty et al., 2014). The common location of WiFi access points within many
building interiors gives a significant advantage to through-wall passive WiFi
radar detection over their active radar counterparts. Active through-wall radar
systems are required to transmit an RF signal through the wall to interrogate a
target of interest prior to detecting the faint echo, again through the wall. This
double-pass of the RF signal through the wall can significantly degrade the SNR.
• Covert:
The system is entirely inert in the sense that it has no effect on the WLAN which
it is exploiting, e.g. it does not use any of the networks bandwidth and thus affect
the data transfer rates, or even communicate with the WiFi access point itself.
This property makes passive WiFi radar a covert system and therefore difficult to
jam with other RF signals.

Crime and Security Applications

Based on the capabilities and functions described above and in addition to the ubiq-
uitous and pervasive nature of WiFi communication networks, we believe that pas-
sive WiFi radar has the potential to be a valuable crime prevention and detection
tool with widespread applicability, either as a standalone surveillance system or in
conjunction with other technologies such as CCTV. In this section we describe the
354 K. Chetty and K. Woodbridge

potential applications of passive WiFi radar for crime prevention and detection, and
outline some of the system requirements associated with the various types of
deployment:
• General Surveillance in Public Urban Areas:
Networks of distributed passive WiFi radar sensors could be deployed relatively
easily and cheaply over large urban areas where, in addition to WLANs, CCTV
usually already exists. One important crime prevention application here could be
to enhance the existing surveillance capability by addressing the problem of
large volumes of redundant video images which are fed into CCTV control
rooms: CCTV operators are typically required to observe the recordings from
many cameras simultaneously, and though these cameras are equipped with pan,
zoom and tilt capabilities; it is a near impossible task for the operator to ensure
the cameras are always directed toward areas where people are located and activ-
ities of interest are occurring.
Passive WiFi could work alongside CCTV systems to automatically slave
them towards areas where movement is detected, allowing a more useful stream
of video images to the operators. Moreover, the potential of passive WiFi radar
to identify aggressive or anomalous behaviour (see section on Technology
Evolution) could be used to further enhance this function and slave cameras in
real-time to areas which are required to be assessed with high-priority. Finally
the tracking capability of passive WiFi radar (Falcone et al., 2012) could be com-
bined with video analytics to add a level of autonomy to the surveillance net-
work. For example if the CCTV operator wanted to monitor a specific person
over a period of time, the radar could track that person and facilitate automatic
handover between different cameras as they move around.
• Though-the-Wall Surveillance:
For law enforcement agencies being able to determine the numbers and locations
of people in an enclosed area would be useful in scenarios such as hostage situ-
ations or prior to forcing entry in a house raid, or when storming a building. The
through-wall capability of passive WiFi radar could possibly also be extended to
through-floor and through-ceiling detection in structures such as high rise build-
ings for covertly monitoring terrorist activities.
For through-wall surveillance applications there is a requirement for the sen-
sor to be mobile so the architecture of the passive WiFi radar system would have
to be designed to meet this criteria, i.e. the reference and surveillance antennas
would not be fixed and spatially distributed around a large area, but would be
co-located and allowed to have degrees of freedom in their pointing directions.
The signal processing and user interface would also need to be integrated on a
single unit for control by an individual operator.
• Emergency Evacuation:
Within the domain of Intelligent environments, agent based models are being
used in real-time to analyse information relating to the locations of fires, explo-
sions, dense smoke etc from smart sensors to generate optimised emergency
20 Passive WiFi Radar: A New Technology for Urban Area Surveillance 355

evacuation routes (Chooramun, Lawrence, & Galea, 2011; Gwynne et al., 1999)
using intelligent active dynamic signage (Xie et al., 2012). These evacuation
routes are then updated in real-time as the fire or attack progresses. However, a
major issue for the models is that there is a limited availability of data relating to
the positions and numbers of people at any given instant. Passive WiFi radar
could therefore be combined into the intelligent sensor network to aid the agent
models by seeding the simulations with information about the locations and
movements of people. The result would be a tailored and dynamic evacuation
plans for individuals based on their location and current direction of travel.
• Perimeter Monitoring Around Buildings:
Passive WiFi radarcould be used as an alarm system when a predefined perimeter
around a building has been breached. To keep costs low for the home security
market, the sensor could take the form of a single receiver which bolts-on to an
existing WiFi access point. For public, government and industrial buildings the
technology could also be integrated with the existing CCTV infrastructure for
enhanced surveillance capabilities described above.

Technology Evolution

Extension to 4th Generation Wireless Networks

The development in recent years of 4th Generation (4G) wireless networks such as
Worldwide Interoperability for Microwave Access (WiMAX) and Long Term
Evolution (Lte) have allowed internet access over long distances of many tens of
kilometres within wireless metropolitan area networks (WMANs). Passive WiFi
radar could therefore evolve toward passive WiMAX or Lte radar in order to pro-
vide surveillance over much larger areas.
We have investigated the potential application of passive WiMAX radar for
maritime surveillance (Chetty et al., 2010) and the outcomes suggest that due to
its low-cost, ease of large-scale deployment and no requirement to operate with
a spectrum licence, it could be used to extend existing surveillance coverage
areas, or “fill-in” blind spots. Potential applications include monitoring long
stretches of coastline to counter drug smuggling and illegal immigration, ship
based detection of approaching small boats in tackling piracy, and surveilling
harbour areas and shipping channels for non-cooperating unidentified vessels
and suspicious movements associated with terrorist activities. Additionally in
developing countries where it is unlikely that wired communications infrastruc-
tures will be established on a large scale, 4G and 5G wireless communications
are expected to undergo a significant rollout. These countries typically have
little or no protection of coastlines, harbour and port areas, so utilisation of the
wireless transmissions through passive radar technology could prove to be a
useful surveillance tool.
356 K. Chetty and K. Woodbridge

Motion Characterisation Using Micro-Doppler Analysis

WiFi signal reflections from moving targets that are detected by the surveillance
receiver experience a shift in their frequency due to the well-known Doppler effect
(Gill, 1965). Micro-Doppler is an additional effect whereby any other motions on
the main body of movement, such as vibrations or rotations induce further Doppler
perturbations around the main Doppler shift. These perturbations, or micro-Doppler
signatures, are characteristic to particular types of movement and can therefore be
used to provide extra information about the targets of interest using classification
techniques (Chen, Tahmoush, & Miceli, 2014).
For passive WiFi radar, micro-Doppler signatures associated with the way in
which a person’s torso and limbs move could be used indicate the type of activity
they are carrying out. For example a person who is walking whilst holding a ruck-
sack or concealing an object under one arm while the other arm swings naturally
would generate a different micro-Doppler signature to someone who is walking
absent with any belongings on their person, and similarly for motions such as
punching, kicking, and limping. Potential application of micro-Doppler analysis
include; recognising aggressive behaviour in the night-time economy, determining
if an alarm has been triggered by an animal or human (e.g. in border surveillance)
or identifying erratic behaviour in the bulk movements of large crowds. Moreover
there may also be a possibility that micro-Doppler signatures for the same types of
motions (e.g. walking or running) but between different people can be differentiated
due to subtle differences in body motions, and thus used for gait-based identifica-
tion. Initial results from our real-time passive WiFi radar system prototype (Chetty
et al., 2014) have shown that difference body movements such as swaying and
squatting to pick up an object give distinct Doppler-time characteristics with many
micro-Doppler components (Tan, Woodbridge, & Chetty, 2014).

Summary

This chapter describes a novel approach for surveillance in urban areas using passive
WiFi radar technology. The main features of technology and its advantages over other
surveillance technologies are explained. The application of passive radar for crime
prevention and detection applications within Crime and Security are significant and
some examples of where and how the technology could be deployed are discussed.
The expansion of wireless networks in towns and cities across the world has pro-
vided a basis for the technology to be deployed ubiquitously wherever there is a
requirement for area monitoring and surveillance. The future demand for wireless
networks to provide higher power communication signals and larger bandwidths to
facilitate broader connection coverage faster data rates will result in performance
improvements of wireless based passive radar systems in terms of both surveillance
20 Passive WiFi Radar: A New Technology for Urban Area Surveillance 357

coverage and detection accuracy. Passive WiFi radar is a new technology for monitor-
ing the movements of people, and we envisage that it will become integrated within
the urban wireless landscape with widespread applications in the growth of cities.

References

Chen, V. C., Tahmoush, D., William J. (2014) Radar Micro-Doppler Signatures: Processing and
Applications. ISBN: 9781849197168. doi:10.1049/PBRA034E.
Chetty, K., Smith, G., Guo, H., & Woodbridge, K. (2014). Target detection in high clutter using
passive bistatic WiFi radar, Radar Conference, 2009 IEEE. Pasadena, CA, 2009, 1–5.
doi:10.1109/RADAR.2009.4976964.
Chetty, K., Smith, G. E., & Woodbridge, K. (2012). Through-the-wall sensing of personnel using
passive bistatic WiFi radar at standoff distances. Geoscience and remote sensing. IEEE
Transactions, 50(4), 1218–1226.
Chetty, K., Tan, B., & Woodbridge, K. (2009). Data processing for real-time wireless passive radar,
Radar Conference, 2014 IEEE. Cincinnati, OH, 2014, 0455–0459. doi:10.1109/
RADAR.2014.6875634.
Chetty, K., Woodbridge, K., Guo, H., & Smith, G. E. (2010). Passive bistatic WiMAX radar for
marine surveillance, Radar Conference, 2010 IEEE. Washington, DC, 2010, 188–193.
doi:10.1109/RADAR.2010.5494627.
Chooramun, N., Lawrence, P. J., & Galea, E. R. (2011). Implementing a hybrid space discretisation
within an agent based evacuation model. In R. D. Peacock, E. D. Kuligowski, & J. D. Averill
(Eds.), Pedestrian and evacuation dynamics (pp. 449–458). Boston, MA: Springer US.
Derham, T. E., et al. (2007). Design and evaluation of a low-cost multistatic netted radar system.
Radar, Sonar & Navigation, IET, 1(5), 362–368.
Falcone, P., Colone, F., Macera, A., & Lombardo, P. (2012). Localization and tracking of moving
targets with WiFi-based passive radar, Radar Conference (RADAR), 2012 IEEE. Atlanta, GA,
2012, 0705–0709. doi:10.1109/RADAR.2012.6212229.
Gill, T. P. (1965). The Doppler effect: An introduction to the theory of the effect. London: Logos
Press.
Griffiths, H. D. (1993) Bistatic Radar - Principles And Practice, Microwave Conference/Brazil,
1993, SBMO International, pp. 519–526. doi:10.1109/SBMO.1993.587208.
Griffiths, H. D. (2008). New directions in bistatic radar, Radar Conference, 2008. RADAR ‘08.
IEEE, Rome, 2008, pp. 1–6. doi:10.1109/RADAR.2008.4720719.
Griffiths, H. D., & Long, N. R. W. (1986). Television-based bistatic radar. Communications, radar
and signal processing. IEE Proceedings F, 133(7), 649–657.
Gwynne, S., et al. (1999). A review of the methodologies used in the computer simulation of
evacuation from the built environment. Building and Environment, 34(6), 741–749.
Howland, P. (2005). Editorial: Passive radar systems. Radar, sonar and navigation. IEE Proceedings,
152(3), 105–106.
IEEE. (1999). IEEE Std 802.11-1999, part 11: Wireless LAN medium access control and (MAC)
and physical layer (PHY) specifications.
Lauri, A., Colone, F., Cardinali, R., Bongioanni, C., & Lombardo, P. (2007) Analysis and Emulation
of FM Radio Signals for Passive Radar, Aerospace Conference, 2007 IEEE, Big Sky, MT,
2007, pp. 1–10. doi:10.1109/AERO.2007.353068.
Nadjiasngar, R., Middleton S., & Inggs, M. (2012). Doppler-only tracking with the recursive
Gauss-Newton filter, Radar Systems (Radar 2012), IET International Conference on, Glasgow,
UK, 2012, pp. 1–5. doi:10.1049/cp.2012.1646.
Poullin, D. (2005). Passive detection using digital broadcasters (DAB, DVB) with COFDM modu-
lation. Radar, sonar and navigation. IEE Proceedings, 152(3), 143–152.
358 K. Chetty and K. Woodbridge

Skolnik, M. I. (1980). Introduction to radar systems (2nd ed.). New York: McGraw-Hill.
Tan, B., Woodbridge, K., & Chetty, K. (2014) A real-time high resolution passive WiFi Doppler
radar and its applications. In International radar conference. France.
Tan, D. K. P., et al. (2005). Passive radar using global system for mobile communication signal:
Theory, implementation and measurements. Radar, sonar and navigation. IEE Proceedings,
152(3), 116–123.
Willis, N. J. (2005) Bistatic Radar, Second Edition, Electromagnetics and Radar Series Edition:
illustrated, reprint, SciTech Publishing, ISBN:1891121456, 9781891121456
Woodward, P. M. (1953). Probability and information theory with applications to radar. New York:
Pergamon.
Wu, J. (2011) a quarter of households worldwide now have wireless home networks. Apr 04, 2012.
Market report by Strategy Analytics.
Xie, H., et al. (2012). Experimental analysis of the effectiveness of emergency signage and its
implementation in evacuation simulation. Fire and Materials, 36(5–6), 367–382.
Zemmari, R., et al. (2014). GSM passive coherent location system: Performance prediction and
measurement evaluation. Radar, Sonar & Navigation, IET, 8(2), 94–105.
Chapter 21
Micro UAV Crime Prevention: Can We Help
Princess Leia?

Matthew Ritchie, Francesco Fioranelli, and Hervé Borrion

Introduction

Unmanned Aerial Vehicles (UAVs) are becoming increasingly popular due to the
impressive capabilities, broad range of applications and cost-saving opportunities
they offer. Also known as Unmanned Aerial Systems (UAS) or drones, these plat-
forms, in particular micro UAVs under 3 kg, are increasingly found in the hands of
untrained operators and probably malicious ones in a near future. As such, these
systems are likely to represent a source of danger and criminal threats, introducing
new operational challenges to police forces.
Prevention of drone crime relies on the ability to pre-empt potential risks, and
implement appropriate measures to affect offenders’ motivation to commit crime,
disrupt their attempts, and minimise the consequences. As seen in the other chapters
of this book, prevention measures are highly specific, and effective implementation
requires relevant tactical information. In particular, prior knowledge of the presence
and location of problematic UAVs is needed to decide when and where prevention
measures should be used. Therefore, achieving UAV detection could soon become
a pressing challenge for law enforcement.
At the time of writing, a representative of DroneShield, a US company manufac-
turing UAV detection systems, indicated in the magazine Motherboard that they were
unable to honour several orders as exportation of UAV detection systems has not yet
been authorised by the US State Department (Koebler, 2014). Amongst the companies
waiting for their products to arrive feature Pinewood studios. Known for blockbusters
including Harry Potter, James Bond, Batman and Alien, the studios was planning to

M. Ritchie • F. Fioranelli
UCL Department of Electronic and Electrical Engineering, London, UK
H. Borrion (*)
UCL Department of Security and Crime Science, London, UK
e-mail: h.borrion@ucl.ac.uk

© Springer International Publishing Switzerland 2017 359


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_21
360 M. Ritchie et al.

use it to prevent unauthorised pictures being taken of the Star Wars Episode VII. Several
UAVs have allegedly been flying over the movie set with the hope to capture some
images of this very closely guarded secret. In order to shed a light on the issue of UAV
detection, we investigated empirically the possibility to detect civilian micro UAVs
that would threaten the (commercial) interests of Star Wars.
The chapter is divided into five sections. In the first, we provide an overview of
current UAV uses and list their domains of application. In the second, we introduce
the less glamorous side of UAVs, and identify various ways in which UAVs may
cause or facilitate the occurrence of crime. In the third section, we discuss the main
challenges for law enforcement. As part of this discussion, we present current capabil-
ity gaps that ought to be addressed in order to achieve micro UAV crime prevention.
In the fourth section, we focus on one of those capabilities, and review some of the
principles that have been proposed to detect micro UAVs. Finally, the chapter presents
the physical principles behind radar detection, and concludes with an experimental
study to answer our initial question: Could a conventional radar system be used to
detect civilian micro UAVs that would fly toward Pinewood studios’ movie set?

The UAV Society

UAVs can now be found in a range of sizes, shapes and contexts. Their size classification
range are commonly noted as nano, micro, tactical, MALE (Medium Altitude Long
Endurance) and HALE (High Altitude Long Endurance) as well as others. Musial (2008)
provided a comprehensive description of various system architectures for airships, fixed-
wing systems, helicopters and quadrotor UAVs. Over the years, large systems have dem-
onstrated their capacity to produce a real game-change in war theatres. Systems such as
General Atomics’ Predator have shown that military missions can now be technically
carried out from centres situated thousands of kilometres away from targets (Mayer,
2009). However, their impact is mostly perceived through TV documentaries and news
reports. In contrast, smaller civilian UAVs have been operating for several years in our
urban and rural environments, and already delivered local benefits to the population.
Micro UAVs weighing just a few hundred grams, for example, have been used by police
forces to support law enforcement operations (Pilkington, 2014) and, in particular, to
improve situation awareness. The “UK’s first drone arrest” occurred in 2010, according
to a newspaper report that detailed the contribution of an airborne thermal imaging cam-
era to the arrest of a car thief (Hull, 2010). Whilst police UAVs have so far been restricted
to an information gathering role (Constantinescu & Nedelcut, 2011, Lega, Ferrara,
Persechino, & Bishop, 2014), platforms carrying more active payloads are now adver-
tised by several manufacturers. Amongst them, a “riot control” UAV that can make use
of “pepper spray, plastic bullets, paintballs, strobe lights and […] lasers” has recently
appeared on the market (Kelion, 2014). Beside these ethically questionable applications,
UAVs are also employed in more mundane activities such as crop mapping, sheep herd-
ing, delivery of medical equipment, environmental monitoring, search and rescue, air-
craft inspection and filming. Beyond commercial uses, UAVs are also employed in a
range of recreational activities including drone racing (Bloom, 2014).
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 361

In the coming years, the number of small-UAVs, micro UAVs and commer-
cial licences are expected to increase exponentially (TGC, 2014). UAVs are
likely to be used for more applications including, for example, pollination.
Equally, the number of households having their own UAV will probably increase
significantly. This is supported by the fact that these systems already outper-
form a number of popular products (e.g. handheld digital cameras, kites and
remote control model aircraft) and that new applications will become possible
as UAVs are designed to perform more advanced functions (e.g. collision
avoidance).

The Dark Side of UAVs

Through their surveillance capability, UAVs already play an important role in law
enforcement, for example to provide situation awareness during protests. In the
future, it is possible that more widely and systematic use of UAV swarms will allow
police forces to further reduce crime rates. For example, advanced UAVs deployed
near luxury stores could be employed to track so-called “smash-and-grab gangs” as
they leave a crime scene on motorcycles. However, dual-use issues are common to
most advanced technologies (Nissen, 2014), and the impact of UAVs on society will
not be limited to controlling crime. Greater use of such systems could also bring
about various risks (Lucas, 2015).
As an analogy, the introduction of mobile phone (as an example of recent techno-
logical innovation) has been accompanied by various crime issues. Mailley, Garcia,
Whitehead, and Farrell (2008) pointed out that mobile phones handsets are frequently
stolen by criminals; they give criminals enhanced communication capability to plan
future crime; and they can be used to remotely trigger explosive devices. Finally they
make their users more vulnerable to certain types of attacks, e.g. vishing (IC3, 2008).
In the case of UAVs, coarse examination of potential threats and vulnerabilities
reveals various risks that those systems could cause or exacerbate. These risks can
be classified into three categories. These are described below with illustrating
examples of previous incidents.
Criminal activities targeting UAVs, payloads or operators
• In 2010, three UAVs used for environmental research work were stolen at the
University of Bergen (Reuder, 2010).
Misuses and Criminal activities perpetrated using UAVs
• In April 2014, “a man […] has become the first person in the UK to be success-
fully prosecuted for the dangerous and illegal flying of an unmanned aircraft”
after flying over a nuclear submarine facility (CAA, 2014).
Legal UAV-related activities that facilitate crime
• UAV operators may be more vulnerable to theft of properties due to the distrac-
tion caused by operating a UAV.
362 M. Ritchie et al.

UAV Crime Prevention

The rise of mobile phones has been accompanied by an increase in mobile phone
thefts. To address this phenomenon dedicated police teams were created such as the
National Mobile Phone Crime Unit in the UK. Given the similarity between these
two domains, one may expect that, sooner or later, specialist police units with dedi-
cated techniques and equipment will be established to focus exclusively on crimes
involving UAVs.
In terms of operational capabilities, such police units may seek to have the same
control over UAVs as with other vehicles. Their powers may even be extended fur-
ther to compare with those employed in air traffic control.
• Verification and Evidence collection
– Verifying the administrative status of UAVs, e.g. licence and insurance.
– Verifying compliance to legal requirements, e.g. model, weight, activity,
flight path, altitude, and distance to people.
– Verifying compliance to the code of practice for surveillance technology and
personal information (e.g. ICO, 2014).
• Control
– Operating UAVs
– Grounding UAVs
– Requisitioning UAVs
One of the main challenges for law enforcement is that UAVs can help reduce the
risk of being caught for criminals. First, those systems can be used to carry out cer-
tain actions remotely and anonymously; second, they can evade capture attempts by
moving to higher altitudes; and third, they can disappear relatively rapidly and leave
very little trace at the crime scene.
Another challenge is that a wider range of existing prevention measures have
been developed and implemented in the built environment with the assumptions
that certain events are unlikely to occur because they are too “dirty, difficult or
dangerous”. However, the new capability of UAVs, and their suitability to so-
called 3D tasks place new constraints on police forces, change the risks, and poten-
tially undermine the effectiveness of existing crime prevention and crime detection
measures.
In order for the police to collect tactical information about, or take control of,
UAVs, some basic capabilities would be required. These are described below with a
technical definition for each of them.
• Detection: ascertaining the presence or absence of an entity in given geographic
and time domains
– Detecting UAVs
– Detecting individuals operating a UAV
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 363

• Localisation: determining the geospatial position of an entity and overlaying


these grid points onto a map an operational area.
– Localising UAVs
– Localising stolen or unauthorised UAVs
– Localising individuals operating a UAV
• Tracking: determining the geospatial position of an entity across a time window
– Tracking UAVs
– Tracking UAV operators
• Identification/Recognition: determining an entity’s identity, and being able to
recognise it
– Identifying UAVs
– Identifying UAVs on a watch list (e.g. stolen ones)
– Identifying UAV owners or UAV operators
Several of these capabilities such as detection of vehicles on a watch list could be
achieved for cars and lorries (Coaffee, 2004; Rhead, Gurney, Ramalingam, &
Cohen, 2012) but are poorly applicable to highly versatile vehicles such as UAVs.
At the time of writing, the French Authorities have been investigating several
cases of drones that have penetrated the 2.5 km radius no-fly zone around nuclear
power plants (Tamblyn, 2014). In a context where many countries are concerned by
European jihadists returning from Syria, inability to rapidly identify the perpetra-
tors and to effectively control them illustrates the societal risks caused by such a
disruptive technology.

UAV Detection

At this stage, UAV crime prevention relies extensively on the ability to detect and
localise platforms or their operators. Without precise information about the presence
and location of UAVs, control measures would be implemented with limited efficacy.
Multi-model netted sensor fence have been proposed to detect and classify rela-
tively large airborne vehicles such as light aircraft. Shi, Fante, Yoder, and Crawford
(2005), for example, provide a comprehensive description of such a system: “The
primary component in the system is a radar fence capable of detecting approaching
targets and providing a cue to two other sensors that perform the discrimination task.
[…] Acoustic microphone arrays [detecting the sound originating from the aircraft’s
engine(s)] are used as the second sensor modality in the system […] (1) to provide
target direction of arrival (DOA) estimates that will then be fused with radar measure-
ments to form a target track; (2) to provide a means for target identification and clas-
sification; and (3) to mitigate false alarms.” Finally “the third sensor modality in the
fence is an optical system which is cued by the radar and/or acoustic sensors and slews
in angle to acquire, track and identify the potential airborne threat. A boresighted vis-
ible camera is also used for improved target resolution during the daytime.”
364 M. Ritchie et al.

Attempts to develop a UAV detection system have been made during a recent
military competition where one team developed a multi-sensor system whose archi-
tecture resembles the one described above. “The team used radar for the long-range
detection outside the city canyons, an acoustic sweep for close range detection,
video cameras for a visual on the target and finally a RF Geolocation, a sensor that
detects video downlink signals from the UAVs” (Azimi, 2012).
Other attempts to develop acoustic sensing system have been published by Case,
Zelino, and Rigling (2008). Such technology is becoming more mature and can now
be found in the market. For example, the company Droneshield (2014) advertises an
acoustic detection system that compares the ambient audio signal recorded by a
microphone with common drone acoustic signatures in a database.
Detection can also be performed based on the data transmitted between UAVs
and their ground station controllers, as this data link is always present within com-
mercial micro-UAVs it can be exploited for detection. Peacock & Johnstone (2013)
developed and tested a detection method “capable of detecting and identifying a
Parrot AR Drone v2 within Wi-Fi operational range”. In the private sector, a US
company called Domestic Drone Countermeasures currently advertises a prototype
drone detection system (Mlot, 2014). Their system is designed to detect data signals
propagating in a given area, and applies a standard triangulation technique to deter-
mine the location of the transmitters. The presence of drones is then inferred by
monitoring the presence of unknown transmitters suddenly appearing on the spec-
trum. According to their website, “the Primary Command and Control Module can
typically communicate with nodes up to 200 ft away. Each RF Sensor Node can
typically detect drones within 50 ft in all directions. Frequency detection range
1 MHz to 6.8 GHz” (DDC, 2014).

Radar Detection

Aircraft detection is currently performed using either co-operative transponder sys-


tems or active Radar systems which do not require the co-operation of the target. In
principle those systems could be adapted to detect much smaller targets such as
micro UAVs (Moses, Rutherford, & Valavanis, 2011).
Radar is an acronym for “RAdio Detection And Ranging”, and defines a sensor
able to measure the range (i.e. distance) and in most cases the velocity of an object.
Radars are able to extract this information on targets of interest by using a method
of transmitting electromagnetic waves, and analysing the returned scattered waves.
The architecture of a radar system typically consists of a transmitter generating an
electronic signal, one or multiple antennas successively converting this signal into
electromagnetic waves and capturing some of the scattered waves, and a receiver
converting the captured energy into electronic signals that could be analysed using
some form of computational device (Skolnik, 2008). A radar system also includes
a visualisation module to display the extracted information for the user. A simple
diagram of a radar system is shown in Fig. 21.1. The returned electromagnetic
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 365

Fig. 21.1 Principle of radar detection

signal contains information on the targets range and Doppler as well as more subtle
information such as its resonance frequency, micro-Doppler and range profile.
Micro-Doppler is the phenomenon of the observed micro-motions on top of the
bulk main Doppler component of target’s motion, such as the blade rotations of a
helicopter on top of its forwards main Doppler velocity (Chen, Tahmoush, &
Miceli, 2014).
Depending on their complexity, radar systems can simply provide an indication
on the absence or presence of targets in the area of interest, as well as estimate addi-
tional information on the number, distance, spatial location, speed, as well as other
characteristics of the targets. Compared with optical sensors, radar systems are less
sensitive to weather conditions, generally operate over longer ranges and can provide
detection capabilities during both day and night, regardless of fog, haze, or clouds.
Radar systems have been used for many decades in a variety of civilian and military
applications. Practical uses include air traffic control (Nolan, 2010), geological surveys
(Oyan et al., 2012), terrain mapping (Dubois et al., 1995), prediction of potentially
dangerous weather phenomena such as severe storms and tornadoes (Alberty & Crum,
1990), vital signs monitoring (Sachs et al., 2014), collision avoidance for aircrafts and
cars (Rohling, 2010), surveillance of areas in war and security enforcement contexts
(Charvat et al., 2012).
The probability of successful target detection and the range at which this can be
achieved depends on numerous parameters, such as transmit power, the fraction of
time when the energy is transmitted (“duty cycle”), the capacity of concentrating the
electromagnetic energy towards a particular direction (“antenna gain”), the fre-
quency of the transmitted waves, the power of the thermal noise, the power of the
background of undesired echoes (called “clutter”), and the power of other signals at
the same frequency generated by external sources (called “interference”), which can
mask the reflections from the targets of interest. An important effect on the detec-
tion performance is also played by the reflective properties of the target, which
366 M. Ritchie et al.

depend on its shape and physical composition, as well as on the frequency and
direction of arrival of the transmitted waveforms. These properties are generally
summarised in the concept of target Radar Cross Section (RCS) which has the unit
of an area (square metres) and is a measure of the power backscattered by the target
towards the direction of the radar (Stimson et al., 2014). All the aforementioned
parameters can be combined in the “radar equation”, which defines the detection
performance of a radar system. A more detailed description of the fundamental
principles of radar can be found in (Stimson et al., 2014).

Experiment

Micro UAVs are challenging targets for radar detection because of their size and the
fact that they are made mostly of low conductivity materials such as plastic, rubber,
and foam, which reduces their RCS in comparison with metallic targets.
An experiment was performed using a radar system in order to assess the detect-
ability of UAVs and signal properties seen by a radar.

Aim

To assess the suitability of conventional radar systems to be used for UAV detection,
we set out to empirically answer three questions:
1. Can a conventional radar system be used to detect micro UAVs hovering or fly-
ing over an area of interest?
2. What do the reflected radar signals from a typical micro UAV look like? Both in
the time domain (RCS) and Doppler domain (micro-Doppler).
3. What tactical information can a conventional radar system provide in the context
of crime prevention? For example can intension of the drone movement be char-
acterised. If the drone is only hovering in place it may not be a great threat, but
if it is moving directly towards an important aspect such as a building / protected
airspace it may be a greater threat.

Method

Design

The quantity measured during this study is the electromagnetic power received at
the radar system after the transmission of a radar pulse. When a small UAV is pres-
ent in the area under test, the electromagnetic signal is reflected back towards the
radar, and the level of received power increases compared with the situation of no
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 367

targets present in the area under test, when only noise return is received at the
radar. This increase in received power provides an indication that the target is pres-
ent. The time for the electromagnetic power to propagate from the radar to the
UAV and back to the radar is related to the physical distance of the target. Each
single pulse is regularly transmitted according to a specific frequency (Pulse
Repetition Frequency, PRF), and the electromagnetic power after each transmis-
sion is recorded as a function of time.
The characteristics of the returned signals were analysed with the aim of defining
how a radar sees a micro drone and what information on the target can be gained
using a radar. These points are important when considering using a radar system to
detect, track and classify micro drones.

Instruments

Radar: NetRAD

The radar system used in this study is the netted radar system NetRAD which has
been developed over a number of years by the Radar Group at University College
London (Derham et al., 2007). NetRAD has been used primarily to gather data for
sea clutter analysis and maritime targets detection in an experimental campaign in
South Africa and in the UK, providing novel results in this research domain
(Al-Ashwal et al., 2011, 2014a, 2014b).
NetRAD consists of three different nodes, each including both a radar trans-
mitter and a receiver. The nodes operate in a networked mode (hence NetRAD)
where they are online on a local network and can be controlled via remote access
through a laptop during the measurements. Networked radars have many advan-
tages over individual radar systems, such as enhanced target signatures, advanta-
geous properties of the clutter, multi-perspective views on targets (Inggs et al.,
2014). Hence, such a system would likely be a good candidate for this applica-
tion of detecting small drones which may be difficult to detect and track using a
single radar.
In order to preserve the coherency (i.e. synchronisation) of the data collected at
each NetRAD node, the same clock signal has to be shared in all the nodes. In its
basic configuration, NetRAD is used in “wired” mode where one of the nodes works
as “master node” generating and providing the same clock signal to the other nodes.
This is achieved via Ethernet cables that physically connect the nodes. NetRAD can
also operate in “wireless” mode where each node is equipped with a GPS Disciplined
Oscillator (GPSDO) which provides the clock signal thanks to the synchronisation
to the same constellation of GPS satellites. This configuration avoids having cables
that run between the nodes, and therefore allows longer physical separation between
the nodes during the measurements
NetRAD transmits electromagnetic pulses at a carrier frequency of 2.4 GHz—the
same frequency as the IEEE 802.11 standard Wi-Fi—with 45 MHz bandwidth.
Parameters such as pulse length, pulse repetition frequency, and number of recorded
368 M. Ritchie et al.

pulses can be selected by the user and flexibly changed at each recording. For this
study the following values have been used: 0.6 μs pulse length, PRF between 1 and
5 kHz, and a duration of between 1 and 10 s of recorded data for each measurement.
The transmitted pulse was modulated with a linear chirp. Each node is equipped with
a directional antenna with 24 dBi gain and 10° by 10° beam-width. The power of the
transmitted signal was 200 mW (23 dBm). NetRAD can also use high power amplifier
with over 400 W peak transmitted power, which allows the radar to operate at ranges
up to approximately 2 km. For this study of UAV detection the high power amplifier
of NetRAD was not used, as the UAV was flying relatively close to the radar, but the
results shown would also translate to longer ranges when using more power.
Figure 21.2 shows one of the NetRAD nodes being deployed in the measurement
area, with the transmitter and the receiver antenna placed onto two tripods.
Figure 21.3 shows a close-up view of the NetRAD node with cables connected at
the front and a monitor to check the computer built in within the node.

UAV: DJI Phantom 2 Vision+

The UAV used in this study is the DJI Phantom 2 Vision+ represented in Fig. 21.4
(DJI, 2015). This is a popular commercial micro UAV which weighs about 1.3 kg
and houses a gyroscope, accelerometer, magnetometer and pressure sensor, as
well as ultrasound sensors for altitude. It also carries a 1080p HD (GoPro) camera
that can record a video footage on-board and transmit downlink video data to a
tablet or phone.

Fig. 21.2 View of a NetRAD node with transmitter and receiver antennas
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 369

Fig. 21.3 Close-up view


of a NetRAD node with
connected cables and a
monitor

Fig. 21.4 DJI Phantom 2 vision+


370 M. Ritchie et al.

The DJI Phantom can be piloted through a control unit which operates at
5.8 GHz Wi-Fi (the IEEE 802.11 ac/n standards). It is therefore particularly suit-
able to be used in conjunction with NetRAD, as there is no reciprocal interference
between the 2.4 GHz waveforms transmitted and received by the radar, and the
5.8 GHz signals for the UAV flight control. The DJI phantom is also a very stable
platform, capable of flying in outdoor scenarios even in windy conditions, and easy
to control and move along a desired trajectory. It has a maximum flight time of
approximately 25 min.

Protocol and Setting

The experiments for this study have been conducted in a large open football
field at the UCL sports ground to the North of London. The three NetRAD
nodes were placed along a straight line and spaced 40 m. Node 3 was used as
transmitter and receiver (noted as Tx and Rx, respectively), whereas Node 1
and Node 2 were only used as receivers. The UAV performed different types of
motion in the area under test, approximately at 70 m from the middle node, as
shown in Fig. 21.5.
In the experiment described in this chapter, both the UAV and the operator were
present in the area under test. At the beginning of the measurement, the person
started walking towards Node 3 on a linear trajectory, while the UAV started flying
in the opposite direction. The geometry of the measurement set-up and the trajecto-
ries for the person and the UAV are shown in Fig. 21.5. During a measurement the

Fig. 21.5 Set-up of the UAV detection experiment


21 Micro UAV Crime Prevention: Can We Help Princess Leia? 371

antennas were not moved or steered; they were kept in the same position and pointed
as indicated by the dashed lines in Fig. 21.5.

Results

As indicated in the description of the radar system used in this study, each dataset
consists of a selectable amount of recorded pulses, which corresponds to a certain
time interval. The results can be presented to the user either in the form of a single
recording, hence a two-dimensional profile of received power as a function of range,
or in the form of a three-dimensional Range-Time-Intensity (RTI) image, obtained
by stacking multiple profiles next to each other.
Figure 21.6 shows the resulting data plotted as a Range Time Interval (RTI)
image for the aforementioned experiment using 10 s of data recorded at Node 3.
The colour scale in the image is logarithmic and the dynamic range normalised
to 30 dB below the strongest reflection related to the person walking close to the
radar. Both targets, the UAV and the person, are clearly visible within the data
and they intersect at approximately 5 s, when both person and UAV are at the
same distance from the radar. The UAV target is shown to be moving away from
the radar, while the individual is shown to be moving towards the radar. The
signal scattered from the UAV is weaker than that scattered by the person
because of the different RCS of the two targets. This difference is approximately
8–10 dB at a distance of about 70 m, which means in practice that the signal
scattered from the UAV is approximately ten times (an order of magnitude)
weaker than that scattered by the person. The RCS of the average person can be

0 0
1
–5
2
3 Operator
–10
4
Time [s]

5 –15

6
UAV –20
7

8
–25
9
10 –30
50 60 70 80 90 100 110 120 130 140 150
Range [m]

Fig. 21.6 RTI image for UAV and person moving in the area under test
372 M. Ritchie et al.

assumed to be approximately 1 m2 at 2.4 GHz [Skolnik, 2008] hence the RCS of


this UAV is estimated to be about 0.1 m2. This indicates that a radar system that
is only capable of detecting humans will require additional sensitivity of
8–10 dB in order to also successfully detect micro drones, i.e. a system able to
correctly detect signals which are an order of magnitude weaker. The UAV trace
becomes weaker when the UAV is at more than 90 m away from the radar (this
is empirically visible from the colour changing from white/light grey colour to
darker grey colour). However, the UAV is still detectable even at those dis-
tances and its signal is approximately 10 dB above the noise floor. This result
confirms that conventional radar systems can detect micro UAVs hovering and
flying in the test environment.
In the case of UAVs micro-Doppler is related to the rotation of the blades, and
any vibration of the main drone fuselage, and can be exploited to distinguish the
UAV from other potential targets (e.g. birds), or to identify a particular model of
UAV (De Wit et al., 2012, Molchanov et al., 2013).
The micro-Doppler of the data within Fig. 21.6 was evaluated and is shown in
Fig. 21.7. The operator is shown to have a positive Doppler signature with lines
extending above and below from their main movement created by the motion of
their arms and legs. The micro UAV is shown to have a negative main Doppler, but
no micro-Doppler components were detected on top of this motion. The authors
conducted further research on this aspect; the results can be found in (Fioranelli
et al., 2015).

50 0

–5

0 –10
Doppler [Hz]

–15
Operator

–50 –20

–25
UAV

–100 –30
0 1 2 3 4 5 6 7 8 9 10
Time [s]

Fig. 21.7 Doppler-time plot of operator and UAV micro Doppler signatures
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 373

Limitations

This study presents some limitations that need to be discussed.


A coarse qualitative assessment of the tracks suggests that the radar used pro-
vided accurate information, showing the direction, range and speed of both the
operator and the UAV. However, the UAV was neither filmed nor equipped with a
GPS receiver. This would have allowed tracking of its trajectory during the motion,
thus providing the precise spatial coordinates of the UAV location at different time
instants during the measurement as a “ground truth” for comparison with the
response from the radar.
The radar system used in this study, NetRAD, is not purposely designed for UAV
detection. The antennas were arranged to stare at a fixed location. For full tracking
of UAV targets a radar would have to gain both range and azimuth information to
accurately locate the target. Real world deployed systems can complete this by
using rotating antennas or electrically steered beams to gain azimuth information
(Stimson et al., 2014).
The results suggest that the wavelength of the transmitted signal (12.5 cm corre-
sponding to 2.4 GHz operational frequency) is too short for effective micro-Doppler
analysis of the UAV. This was shown in Fig. 21.7 where micro-Doppler signatures
were seen for the operator but not for the micro UAV. Higher radar frequencies and
therefore shorter wavelengths would be more beneficial for these techniques. This is
because as a rule of thumb half wavelength should be comparable with the length of
the moving parts of the objects whose micro-Doppler has to be characterised, in this
case the blades of the UAV which are only a few centimetre long.
The experiments were performed in an open field, without the complications that
may be encountered in an actual urban environment making UAV detection more
challenging. Complications may include obstructions from building, structures, or
other moving objects that can hide the UAV from the radar, as well as sources of
interference that can disrupt the normal operations of the radar (for instance devices
working with 2.4 GHz Wi-Fi would cause severe interference to NetRAD). The
choice of the operational frequencies to avoid interference and the most suitable
locations to maximise the area seen by the radar should be carefully assessed to
deploy effectively a radar system for UAV detection in an urban environment.

Summary and Conclusions

This chapter investigates some of the issues that would be faced by society as a
result of the growing use (and potentially misuse) of UAVs. We discuss some of the
tasks law enforcement would need to achieve in order to prevent UAV crime, and
highlight the need for more effective UAV detection capability. Drawing from the
example of Pinewood studio trying to keep Star Wars’ movie sets confidential, we
ask whether a radar system could be used as a non-co-operative system to determine
the presence and location of micro UAVs threatening Princess Leia’s privacy.
374 M. Ritchie et al.

In the second half of this chapter this question is examined empirically. Using a
conventional radar (NetRAD), we demonstrate that such a system can detect micro
UAVs hovering/flying over an area of interest, showed the relative signal levels in
comparison to a human target and finally provide practical information in the con-
text of crime prevention. We also point out that with limited information about the
structure of the target (e.g. presence of propellers), a high false alarm rate would be
expected as UAVs could not be differentiated from birds.
Considering the limitations of the radar system used in this study, it could be hypothe-
sised that UAV detection performance can be improved in actual systems developed specifi-
cally for this purpose. Choosing higher operational frequency, compared to that used in the
experiments, and introducing steerable antennas during measurements to track the UAVs
motion would provide substantial benefits, including allowing discrimination between birds
and UAVs. However, the effect of urban structures (e.g. buildings) on the radar signal would
need to be determined in order to evaluate the effectiveness of radar systems for UAV
detection.

Acknowledgements This work was supported by the Engineering and Physical Sciences
Research Council (EPSRC) [grant number EP/G037264/1]. The authors would like to thank Prof.
Hugh Griffiths for his support, which has been helped greatly by the IET A.F. Harvey Engineering
Prize 2013.

References

Al-Ashwal, W. A., Baker, C. J., Balleri, A., Griffiths, H. D., Harmanny, R., Inggs, M., et al. (2011).
Statistical analysis of simultaneous monostatic and bistatic sea clutter at low grazing angles.
Electronics Letters, 47(10), 621–622. doi:10.1049/el.2011.0557.
Al-Ashwal, W. A., Woodbridge, K., & Griffiths, H. D. (2014a). Analysis of bistatic sea clutter -
part I: Average reflectivity. IEEE Transactions on Aerospace and Electronic Systems, 50(2),
1283–1292.
Al-Ashwal, W. A., Woodbridge, K., & Griffiths, H. D. (2014b). Analysis of bistatic sea clutter -
part II: Amplitude statistics. IEEE Transactions on Aerospace and Electronic Systems, 50(2),
1293–1303.
Alberty, R.L., & Crum, T.D. (1990). Nexrad and the Operational Doppler Weather Radar for The
1990s and beyond. Geoscience and Remote Sensing Symposium. IGARSS '90. 'Remote Sensing
Science for the Nineties', 10th Annual International (pp. 725–728).
Azimi, A. (2012, September 21). Competition offers solutions to detecting UAVs. Marines.
Retrieved from http://www.marines.mil/
Bloom, D. (2014, October 16). Attack of the drones! Star Wars inspires 70mph mid-air races by
French drone fans - who say it’s the new Formula 1. MailOnline. Retrieved from http://www.
dailymail.co.uk/
CAA (2014, April 2). First conviction for illegal use of an unmanned aircraft. Civil Aviation
Authority. Retrieved from http://www.caa.co.uk
Case, E. E., Zelino, A. M., & Rigling, B. D. (2008, July). Low-cost acoustic array for small uav
detection and tracking. Aerospace and Electronics Conference, 2008. NAECON 2008. IEEE
National.
Charvat, G. L., et al. (2012). A through-dielectric ultrawideband (UWB) switched antenna-array
radar imaging system. IEEE Transactions on Antennas and Propagation, 60, 5495–5500.
21 Micro UAV Crime Prevention: Can We Help Princess Leia? 375

Chen, V. C., Tahmoush, D., & Miceli, W. J. (2014). Radar micro-Doppler signatures: Processing
and applications. London: The Institution of Engineering and Technology.
Coaffee, J. (2004). Rings of Steel, Rings of Concrete and Rings of Confidence: Designing out
Terrorism in Central London pre and post September 11th. International Journal of Urban and
Regional Research, 28, 201–211. doi:10.1111/j.0309-1317.2004.00511.x.
Constantinescu, S. G., & Nedelcut, F. (2011). UAV systems in support of Law Enforcement forces.
International Conference of Scientific Paper AFASES 2011 (pp. 1211–1219).
DDC. (2014). Retrieved from http://www.ddcountermeasures.com/
Derham, T. E., Doughty, S., Woodbridge, K., & Baker, C. J. (2007). Design and evaluation of a
low-cost multistatic Netted Radar System. IET Radar Sonar & Navigation, 1(5), 362–368.
De Wit, J. J. M., Harmanny, R. I., & Premel-Cabic, G. (2012). Micro-Doppler analysis of small
UAVs. Paper presented at the Radar Conference (EuRAD), 9th European.
DJI. (2015). Retrieved from https://www.dji.com/product/phantom-2-vision-plus
Droneshield. (2014). Retrieved from http://www.droneshield.org/
Dubois, P. C., Van Zyl, J., & Engman, T. (1995). Measuring soil moisture with imaging radars.
IEEE Transactions on Geoscience and Remote Sensing, 33(4), 915–926.
Fioranelli, F., Ritchie, M., Griffiths, H., & Borrion, H. (2015). Classification of loaded/unloaded
micro-drones using multistatic radar. Electronics Letters, 51(22), 1813–1815.
Hull, L. (2010, February 12). Drone makes first UK ‘arrest’ as police catch car thief hiding under
bushes. MailOnline. Retrieved from http://www.dailymail.co.uk/
IC3. (2008, January 17). Vishing attacks increase. Internet Crime Complaint Centre. Retrieved
from http://www.ic3.gov/
ICO. (2014). In the picture: A data protection code of practice for surveillance cameras and per-
sonal information. Information Commissioner’s Office. Retrieved from http://ico.org.uk/
Inggs, M., Griffiths, H., Fioranelli, F., Ritchie, M., & Woodbridge, K. (2014). Multistatic radar:
System requirements and experimental validation. 2014 International Radar Conference,
13–17 October 2014, Lille, France.
Kelion, L. (2014, June 18). African firm is selling pepper-spray bullet firing drones. BBC News.
Retrieved from http://www.bbc.co.uk/news
Koebler, J. (2014, September 19). Star Wars Ordered a ‘Droneshield’ to prevent leaks on set.
Motherboard. Retrieved from http://motherboard.vice.com/en_uk
Lega, M., Ferrara, C., Persechino, G., & Bishop, P. (2014). Remote sensing in environmental
police investigations: Aerial platforms and an innovative application of thermography to detect
several illegal activities. Environmental Monitoring and Assessment, 186(12), 8291–8301.
Lucas, G. R. (2015). In P. K. Valavanis & J. G. Vachtsevanos (Eds.), Handbook of unmanned aerial
vehicles (pp. 2865–2878). Dordrecht: Springer.
Mailley, J., Garcia, R., Whitehead, S., & Farrell, G. (2008). Phone Theft Index. Security Journal,
21(3), 212–227.
Mayer, J. (2009, October 26). The predator war. The New Yorker. Retrieved from http://www.
newyorker.com
Mlot, S. (2014, June 19). Test a personal Drone Detection System for $500. PCMagazine.
Retrieved from http://uk.pcmag.com/
Molchanov, P., Egiazarian, K., Astola, J., Harmanny, R. I., & de Wit, J. J. M. (2013). Classification
of small UAVs and birds by micro-Doppler signatures. Paper presented at the Radar Conference
(EuRAD).
Moses, A., Rutherford, M. J., & Valavanis, K. P. (2011, September). Radar-based detection and
identification for miniature air vehicles. In Control Applications (CCA), 2011 IEEE
International Conference on (pp. 933–940). IEEE.
Musial, M. (2008). System architecture of small autonomous UAVs. Saarbrücken: VDM Verlag.
Nissen, T. G. (2014). Designing for socially acceptable security technologies. Doctoral disserta-
tion, University College London, London.
Nolan, M. (2010). Fundamentals of air traffic control. Boston, MA: Cengage Learning.
Oyan, M. J., Hamran, S.-E., Hanssen, L., Berger, T., & Plettemeier, D. (2012). Ultrawideband
gated step frequency ground-penetrating radar. IEEE Transactions on Geoscience and Remote
Sensing, 50(1), 212–220.
376 M. Ritchie et al.

Peacock, M., & Johnstone, M. N. (2013). Towards detection and control of Civilian Unmanned
Aerial Vehicles. Proceedings of the 14th Australian Information Warfare Conference, Edith
Cowan University, Perth, Western Australia, December, 2013. Perth, Western Australia: Edith
Cowan University.
Pilkington, E. (2014, October 1). ‘We see ourselves as the vanguard’: The police force using
drones to fight crime. The Guardian. Retrieved from http://www.theguardian.com/
Rhead, M., Gurney, R., Ramalingam, S., & Cohen, N. (2012). Accuracy of automatic number plate
recognition (anpr) and real world UK number plate problems. In Proceedings of the 46th IEEE
International Carnahan Conference on Security Technology. IEEE.
Reuder, J. (2010, November 29). SUMOutstyr stjålet fra UiB. Radiostyrk. Retrieved from http://
www.radiostyrt.no/content.php/249-SUMO-utstyr-stjålet-fra-UiB?
Rohling, H. (2010). Milestones in radar and the success story of automotive radar systems. Radar
Symposium (IRS), 11th International (pp. 1–6, 16–18).
Sachs, J., Helbig, M., Herrmann, R., Kmec, M., Schilling, K., & Zaikov, E. (2014). Remote vital
sign detection for rescue, security, and medical care by ultra-wideband pseudo-noise radar. Ad
Hoc Networks, 13 Part A, 42–53.
Shi, W., Fante, R., Yoder, J., & Crawford, G. (2005, June 2). Multi-modal netted sensor fence for
homeland security. Proc. SPIE 5796, Unattended Ground Sensor Technologies and Applications
VII (vol. 416). doi:10.1117/12.607139.
Skolnik, M. (2008). Radar handbook (3rd ed.). New York: McGraw-Hill Education.
Stimson, W., Griffiths, H., Baker, C., & Adamy, D. (2014). Stimson’s introduction to airborne
radar (3rd ed.). Mendham, NJ: SciTech Publishing.
Tamblyn, T. (2014, October 31). Someone is flying drones over France’s nuclear power stations.
The Huffington Post. Retrieved from http://www.huffingtonpost.co.uk
TGC. (2014, July 14). Teal Group Predicts Worldwide UAV Market Will Total $91 Billion in Its
2014 UAV Market Profile and Forecast. Teal Group Corporation. Retrieved from http://www.
tealgroup.com
Chapter 22
Technology for Crime and Crime Prevention:
A Supply Side Analysis

Graham Farrell and Nick Tilley

Introduction

Security technology is increasingly elegant. Consider the modern car. The driver
walks up to the vehicle with the car’s RFID-fob about them and the door automati-
cally unlocks. The presence of the RFID-fob allows the car to be started—there is
no need to insert it into a slot. The wing mirrors open. The alarm, tracker and immo-
biliser are deactivated. As the car pulls away the central door locks engage, reducing
the opportunity for offenders to open them and steal, for example, a handbag, at
traffic lights. The inbuilt distributed satnav device (its distribution making it diffi-
cult to steal) warns the driver when they exceed the speed limit (deterring or avoid-
ing inadvertent speeding). The top-notch sound system plays, but its parts are also
distributed around the car so it is at no risk of theft. A warning bleeps when the car
veers too far towards the other side of the road, reducing the risk of negligent driv-
ing and collisions. A light warns the driver of a car in their blind spot (ditto). The
driver stops to buy fuel. They press an internal button to open the fuel tank cover
(siphoning off fuel is more difficult since the button is only accessed from inside the
car, and this is easier to use and more secure than a key). At-pump credit card pay-
ment eliminates the possibility that the driver exposes themselves and the car to risk
if they go inside to pay. But if the driver walks to pay for fuel, the car’s security
measures are automatically reactivated, reducing risk of opportunistic thefts.

G. Farrell, Ph.D. (*)


Center for Criminal Justice Studies, School of Law, University of Leeds,
Liberty Building, Leeds LS2 9JT, UK
e-mail: g.farrell@leeds.ac.uk
N. Tilley (*)
Department of Security and Crime Science, University College London,
35 Tavistock Square, London WC1H 9EZ, UK

© Springer International Publishing Switzerland 2017 377


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6_22
378 G. Farrell and N. Tilley

When the driver returns, security is once again automatically deactivated. On arrival
at the destination, the car automatically parks safely avoiding unintended damage to
other vehicles. As the driver walks away the locks automatically engage, the win-
dows shut, the wing mirrors fold back, the alarm is activated, the immobiliser is set,
and the tracker is armed (all making theft of and from cars more difficult or risky).
The driver has to do nothing. The car secures itself automatically. The car’s security
is its default state, the security is not obtrusive or ugly, and the effects of the
increased security have dramatically cut car crime in recent years (as detailed fur-
ther below). No-one’s rights or liberties are infringed and the monetary costs are
low whilst the convenience as well as crime prevention benefits are substantial.
Good security goes hand-in-hand with improved quality of life.
This chapter explores the relationship between crime and technological develop-
ments, focussing on car theft as an example. It shows that technology can both
increase and reduce opportunities for crime. It argues that security has to be “ele-
gant” if it is to be effective in preventing crime and introduces “DAPPER” as an
acronym to capture the qualities required. Those trying to design crime prevention
need to think in DAPPER terms if devices are to be widely adopted and used. Finally
the chapter argues that DAPPER technologies may need a governmental or regula-
tory push if they are to be introduced.
If more crime opportunities are supplied, more will be used. If fewer are sup-
plied, fewer will be used. So says supply-side criminology as we conceive it here.
Technology can be both a source of crime opportunities and a means of curtailing
opportunities. It cuts both ways. Producers of technology often generate opportuni-
ties unintentionally. They sometimes curtail them intentionally. Producers of tech-
nology also sometimes unintentionally reduce the supply of crime opportunities.
They may very occasionally create crime opportunities intentionally. Supply-side
criminology proposes that reducing crime opportunities is the key to crime control.
For the purpose of this chapter we are using a very broad definition of technology:
the use of science to invent useful things or to solve problems. Hence, we do not con-
fine ourselves to the various branches of engineering. We also refer to some simple
(and also old) technologies as well as more sophisticated (and contemporary) ones.
Table 22.1 presents examples of technological developments that have either
deliberately or inadvertently marginally increased or reduced the supply of crime
opportunities.
Those technologies that create crime opportunities intentionally are orientated to
solving problems that would otherwise be faced by prospective offenders. Once the
pro-crime technologies are developed and the erstwhile problem for the offender
has been solved, they increase the supply of opportunities that would otherwise be
closed and hence increase levels of crime. Table 22.1 shows a range of examples of
technological developments that are intended to be useful to offenders (albeit that
there are some nifty, if disingenuous, rationalisations for developing the technolo-
gies, for example that it is being done to reveal the weaknesses in security as a kind
of public service!). Some crime-enabling crime-focused technologies are old and
simple, such as knuckle dusters and flick knives; others are recent and comprise
information technology orientated to damaging the network or to enabling offend-
ers to steal identities and to profit from doing so.
22 Technology for Crime and Crime Prevention: A Supply Side Analysis 379

Table 22.1 Examples of technology and its intended and unintended effects on the supply of
crime opportunities
Create opportunity Reduce opportunity
Intentionally Speed limit detectors; flick knives; Security technology, for example
knuckle dusters; skimmers for credit centrally locked cars, milled coins,
cards; innovations in mood-changing safes, alarms, chastity belts, dye
drugs; RFID jammers; phone tags, RFID tags, DNA profiling,
hacking tools; computer viruses, CCTV, niche-free nappy changing
worms, trojans and bots tables, blue lights in public
lavatories, smoke cloaks, mosquito
alarms, airport scanners
Inadvertently Mobile phones; satnav devices, Double glazing; street lighting;
laptop computers; cordless public lavatories; supplying less
screwdrivers/drills; online banking; toxic gas to households; online
tap and pay cards; public lavatories; shopping; central air conditioning.
online shopping; rohypnol;
automatic weapons

Technologies may also be useful to offenders, even if that was not the intention
of their developers: first they can produce goods that comprise suitable targets for
theft, and second they can provide resources facilitating criminal acts. Ron Clarke
(1999, 2012) describes the attributes of goods suitable for theft with his acronym,
“CRAVED”. This refers to concealable, removable, accessible, valuable, enjoyable
and disposable. Other things being equal an increase in the supply of goods with
these characteristics will increase opportunities for theft. Technology has helped
create them. Mobile phones comprise an extreme example, cars another and the
internet a third. Cars and mobile phones have been stolen in very large numbers.
Mobile phones and cars also contribute to increases in crime opportunities because
of their obvious usefulness to offenders in the planning and conduct of crimes.
Other postwar technological developments that have inadvertently increased the
supply of suitable targets and thereby yielded a substantial crime harvest include car
radios, lightweight televisions, satnavs, credit cards, laptop computers, digital cam-
eras, and internet chat rooms.
Some technological developments may reduce the supply of crime opportunities
as a by-product of their intended use. Take technology-driven efforts to deal with
uncomfortably hot or cold weather. In the Northern hemisphere this has led to wide-
spread installation of double-glazed (double-paned) windows. In the South it has
led to the widespread installation of air conditioning. Double-glazing makes break-
ing into houses more difficult. Moreover it often includes inbuilt window and door
locks that are automatically activated when the door or window is shut. On the other
hand inbuilt air conditioning makes keeping doors and windows open less likely in
the hot weather, and hence reduces the supply of opportunities for walk-in burglar-
ies. Many households, particularly in areas of North America with significant sea-
sonal variations in weather, may have central air conditioning as well as
double-glazed windows and doors.
380 G. Farrell and N. Tilley

Finally, technologies may be devised specifically to reduce crime opportunities.


Parts of the security industry are devoted to this. Shops, banks, dwellings, airports
and cars all make quite extensive use of security technologies to reduce crime
opportunities. Shops, for example, use CCTV, dye and RFID tags, automated release
of smoke-like vapours that disorientate offenders, property marking that is invisible
to the naked eye, safes, alarms, self-locking trolleys, and chip and pin credit card
checks, to reduce the supply of crime opportunities to would-be shoplifters, bur-
glars, commercial robbers and fraudsters.
Readers will notice that the same technology can both increase and reduce the
supply of crime opportunities. Public lavatories, for example, reduce the need for
public urination at the same time as they create secluded private spaces for drug
dealing and sexual abuse. Internet shopping reduces the use of stores and hence
exposure to crime opportunities, but it also creates opportunities to steal delivered
goods left on doorsteps.
Moreover, as Ekblom (1999) has noted, there can be an “arms race” between
those technological developments attempting to curtail the supply of crime oppor-
tunities and those attempting to widen them by sidestepping the preventive tech-
nologies. The history of credit card fraud furnishes an example. The card first
reduced the supply of cash, the thieves’ erstwhile favourite crime target and reduced
opportunity. The card then became a focus for a series of technological develop-
ments aimed at releasing its potential as a crime target. These efforts have been met
by further efforts to thwart the offender. There has been toing and froing between
developments aimed at reducing the supply and developments aimed at increasing
the supply of crime opportunities furnished by credit and debit cards. The recently
introduced contactless card promises to provide a rich new supply of crime oppor-
tunity despite some pre-emptive efforts to ensure better security (Whitehead &
Farrell, 2008).

Technology and Situational Crime Prevention

Situational crime prevention was defined in 1997 as comprising, “opportunity-


reducing measures that (1) are directed at highly specific forms of crime, (2)
involve the management, design or manipulation of the immediate environment is
as systematic and permanent as way as possible, (3) make crime more difficult and
risky or less rewarding and excusable as judged by a wide range of offenders”
(Clarke, 1997). (4) More recently, changing the situation to make it less crime pro-
voking would likely be added to the definition (Cornish & Clarke, 2003; Wortley,
2001, 2008).
Technology is clearly important in situational crime prevention. It helps shape
the immediate environment facing the potential offender and may make crime more
or less risky, more or less difficult, more or less rewarding, more or less excusable,
and more or less provocative. Moreover, it tends to do so in a relatively permanent
22 Technology for Crime and Crime Prevention: A Supply Side Analysis 381

Table 22.2 Examples of technology in situational crime prevention


Reduce
Increase effort Increase risk Reduce reward Remove excuses provocation
Door locks CCTV Dye tags Flashing speed Queue
limit reminders management lanes
Window locks RFID tags Property Vehicle Queue
marking headlight management with
warning sound called numbers
Safes Airport Cash safes in Vehicle seatbelt Card operated exits
scanners buses notification light at stations
Double glazing Caller-ID Prepayment Vehicle satnav/ Perimeter-limited
cards for gas and GPS speeding shopping trolleys
electricity alert sounds
Air Breathalyser Milled edges to Vehicle built-in
conditioning silver and gold breathalyser
coins
Immobilisers Alarms
Trackers e-Passport

Table 22.3 Examples of technology in situational crime production


Increase
Reduce effort Reduce risk reward Adding excuses Add provocation
Battery Multistorey Paint Internet sites advocating/ Detachable car
operated car parks spray legitimating violence badges
drills and cans
screwdrivers
Lightweight Speed limit Guns Paint sprayed walls Malfunctioning
TVs detectors alarms
Mobile Public Unmilled Loud music/car
phones lavatories coins radios
Laptop Rohypnol $500 Online pornography
computers Euro note
Satnavs Violent movies
Pistols Mechanical looms
(Luddites)

and cost effective manner. Whereas, say, police patrol requires significant recurrent
costs, inbuilt technology is often a more cost efficient one-off cost. Moreover, tech-
nology purchasing costs often fall over time due to economies of scale and
competition.
Tables 22.2 and 22.3 give examples respectively of ways in which technological
developments have reduced and increased the supply of crime opportunities by
increasing or decreasing effort, increasing or decreasing risk, reducing or increasing
rewards, removing or adding excuses, or removing or adding provocation.
382 G. Farrell and N. Tilley

Technology and Car Crime

Car crime accounted for some 25 % of all recorded crime in England and Wales in
the early 1990s. Since then the number of incidents has fallen substantially. Figure
22.1 shows the trend.
Technology has played a large part in making cars more secure from crime. Figure
22.2 shows the percentage of cars on the road in England and Wales with various
security devices installed, as found in the Crime Survey of England and Wales (previ-
ously called the British Crime Survey). It shows that between 1991 and 2006–2007
the proportion of cars with central locking rose from around one in three to almost
nine in 10. Likewise there was a substantial growth in the proportion of cars with
alarms and with electronic immobilisers. At the same time there was a fall in the pro-
portion of cars with window etching and mechanical immobilisers. More sophisti-
cated electronic technology was becoming more widespread at the expense of simpler
mechanical technology. The survey only distinguished electronic and mechanical
immobilisers from 1999 but it is clear that electronic immobilisers spread rapidly
from the early 1990s and that this fits well with the onset of the decline in car theft.
Farrell, Tilley, and Tseloni (2014) calculated the trajectory of the spread of electronic
immobilisers to estimate that the rise was underway by the early 1990s. Similarly they
discuss how the timing of the spread of immobilisers and declines in vehicle theft
have also coincided in other countries, and conclude that this is not a coincidence!
The effectiveness of the security devices fitted to cars can be shown in two ways.
First, as indicated in Fig. 22.3 some modes of entry have fallen dramatically since

Fig. 22.1 Rate of vehicle-related theft in England and Wales 1981–2013 (Crime Survey for
England and Wales)
22 Technology for Crime and Crime Prevention: A Supply Side Analysis 383

100
Central locking
Percent of vehicle-owning households 90
Electronic immobiliser
80 Car alarm
Window etching
70 Mechanical immobiliser

60

50

40

30

20

10

0
91

92

93

94

95

96

97

98

99

00

-2

-3

-4

-5

-6

-7
01

02

03

04

05

06
19

19

19

19

19

19

19

19

19

20
Year

20

20

20

20

20

20
Fig. 22.2 Changes in the percentage of cars with various security devices (Crime Survey for
England and Wales)

Fig. 22.3 Method of entry for theft of motor vehicles 1995–2010 (Crime Survey for England and
Wales)

1995, whilst others have remained the same or increased (Farrell et al., 2014). The
falls have occurred most dramatically for forced entry, where improvements in
security might have some effect. Where entry is not forced, for example where the
car is left open or where the offender has a key, there has not been so steep a fall.
Indeed the number of thefts where the offender has a key has been increasing
recently, maybe as a result in part of burglaries where the thief has acquired a car
key and then stolen the car.
384 G. Farrell and N. Tilley

Table 22.4 Security protection factors in England and Wales, as calculated from data included in
the 2001–2007 sweeps of the CSEW
Security protection factor
Security configuration Theft of motor vehicle Theft from motor vehicle
ACET 25.4 6.6
CEM 25.2 2.5
ACEM 19.1 5.7
ACE 15.3 5.5
ACM 12.8 6.4
CE 11.8 2.8
ACEMT 10.3 5.4
AEM 8.4 3.0
EM 7.7 1.7
AE 5.6 2.4
CM 5.4 2.6
AC 5.0 3.9
E 4.0 1.7
AM 3.7 2.3
M 2.8 1.9
C 2.7 2.0
A 1.2 1.5
No security 1.0 1.0
Key: A alarm, C central locking, E electronic immobiliser, M mechanical immobiliser, T tracking
device

The car had furnished a rich supply of crime opportunities until the early 1990s,
both because car theft was straightforward and because there were growing num-
bers of cars for the offender to choose from. Numbers of cars on the road have
continued to grow whilst numbers of car thefts have fallen, especially where secu-
rity devices have made offending more difficult or more risky.
A variety of security devices are often found in a car. A second indicator of the
effectiveness of security devices comes from a comparison of the protection fur-
nished by individual security devices and suites of devices, in comparison to cars
with no security. Table 22.4 shows the “Security Protection Factor” (SPFs) for the
more common combinations of security device fitted to cars, as estimated from
British Crime Survey sweeps from 2001 to 2007 (Farrell, Tilley, Tseloni, & Mailley,
2011, Farrell, Tseloni, & Tilley, 2011). The SPFs show the likelihood of a theft
where there is no security to that for a car with a given security combination. Hence,
the first line shows that a car fitted with a combination of Alarm, Central locking,
Electronic immobiliser and Tracking device (ACET) is 25 times less at risk of theft
of a motor vehicle than a vehicle with no security. The list is ranked for theft of a
motor vehicle. The overall finding is clear from simply scanning the findings for
different combinations. Larger combinations tend to be more effective than smaller
ones; inclusion of electronic immobilisers tends to help prevent thefts of motor
vehicles; and inclusion of alarms tends to help prevent theft from motor vehicles.
22 Technology for Crime and Crime Prevention: A Supply Side Analysis 385

This accords with expectations. Immobilisers make it more difficult to drive cars
away but not to break in and steal things from them (for an overview of the extensive
evidence on the importance of immobilisers in preventing theft of cars, see Brown,
2013a). Alarms may deter people from breaking into cars, but do nothing to stop
thieves driving them away if they do get in. They reduce the supply of opportunities
for different types of crime.
A similar approach has been used to tease out variation in the effectiveness of
security devices for preventing household burglary (Tseloni & Thompson, 2015;
Tseloni, Thompson, Grove, Tilley, & Farrell, 2014). Here, door and window locks
were found to be particularly effective as well as external security lighting and inter-
nal lights on timers. Perhaps surprisingly, household burglar alarms were found to
convey little or no additional reduction in burglary risk (Tilley, Farrell, & Clarke,
2015; Tilley, Farrell, Grove, Thompson, & Tseloni, 2015). It is also telling that, over
the period when locks and lights have become more sophisticated (as well as frames
and panes of glass due to the use of reinforced or stronger materials), more inele-
gant types of security declined. That is, window bars and grilles, which are unsightly,
have been in long-term decline. Alongside the steep declines in household burglary
that have been experienced, these patterns, we suggest, are attributable to the spread
of elegant security and the decline of inelegant security. This is important because it
means possibilities of a “fortress society” are, perhaps counter-intuitively, waning
in parallel with crime’s decline, and that some common fears of security are often
misplaced. Hence, our notion of “elegant security” is discussed next.

Elegant Technology for Effectively Reducing the Supply


of Crime Opportunities

Some technologies can be clunky, ugly, inconvenient, costly and intrusive. That
which is convenient, cheap, and unobtrusive is more liable to be used and less liable
to provoke opposition. In cars ugly, heavy, iron mechanical immobilisers are applied
manually to lock the gearshift or steering wheel in place. This comprises a clunky,
inconvenient and intrusive technology to make it more difficult to steal cars, and
moreover they are not particularly effective—clubs and crook-locks are often easily
overcome by thieves.
We use the acronym, DAPPER, to capture the attributes of “elegant security”.
Dapper refers to
Default (the default position is secure rather than insecure);
Aesthetic (the form of the security is either pleasing or neutral);
Powerful (the preventive effect is strong and not easily circumvented);
Principled (the security measure is acceptable to all, ideally increasing rather than
reducing liberties);
Effortless (the security requires little or no time or thought to engage), and
Rewarding (the benefits exceed the costs).
386 G. Farrell and N. Tilley

As detailed in the opening paragraph of this chapter, the elegance of the technology
associated with car crime reduction is not incidental to its efficacy. We regard it as
essential. The security devices have the potential to reduce the supply of crime
opportunities. Their practical acceptance, adoption, and routine application how-
ever, are also important if that potential is to be realised. It is to this that elegance is
orientated. In the case of car crime, security has indeed become more elegant.
Opportunities have declined and in spite of the increase in numbers of cars on the
road, numbers of car crimes have dropped precipitously.

The Road to Security Elegance

The improvements in vehicle security are not confined to technological develop-


ments. They include, for example, registration arrangements, making it difficult to
alter a car’s identity when sold on to a third party and provisions for car parking that
provide for more natural surveillance increasing the risk to offenders (see Webb,
Smith, & Laycock, 2004). According to Newman (2004), the first compulsory reg-
istration for cars was in Connecticut in 1901, with licence plates introduced in
Massachusetts in 1903.
Ingenious physical security devices for motor vehicles have a long history, a
quite extensive list having been described as long ago as 1910 (Perry, 1910).
Newman (2004: 243–246) lists a series of innovations in security technology from
push button door locks used in Cadillacs from 1915, key operated starter switches
in Chryslers from 1949, vehicle identification numbers from 1969 (when the stan-
dards for lock complexity were also raised), central locking from the 1980s, elec-
tronic immobilisers invented in 1985, trackers in place from 1989, and “electronic
security devices, including transponders, becoming standard in most models” from
2002. This is a tale of increasing technological elegance as well as increasing tech-
nological capacity to reduce the supply of opportunities for crime furnished by cars.
It shows an evolution towards elegance. And elegance means a reduction in depen-
dency on the frailties of human behaviour.
Persuading manufacturers to improve security technology has not always been
easy. Some persuasion has been needed. This can come from various sources,
including consumer demand, police pressure, insurance company carrots, govern-
mental exhortation, or governmental regulation (or the prospect of it). In the event,
governmental exhortation and regulation has often proven necessary to put pressure
on manufacturers to turn their attention to security and attend to its effectiveness.
Laycock (2004) shows how the Car Theft Index was needed in the UK to persuade
car manufacturers to improve the security of their products. In Europe more gener-
ally and in Australia minimum standard immobilisers have become a requirement
for all new cars. Eck and Eck (2012) have recently argued that regulatory measures
be taken in relation to those inadvertently supplying crime opportunities as a side-
effect of their commercial activities to make them accountable for the consequential
crimes produced. This may turn out to be a means of stimulating the development
of elegant security technologies.
22 Technology for Crime and Crime Prevention: A Supply Side Analysis 387

Conclusion

This chapter has argued that technology is important in the supply of crime oppor-
tunities. It can increase or decrease them intentially or unintentionally. The chapter
has shown that technology affects the supply of criminal opportunities in five distin-
guishable ways—by increasing or decreasing the effort required to commit a crime;
by increasing or decreasing the risks involved in committing a crime; by increasing
or decreasing the rewards from committing a crime; by adding or removing an
excuse for a crime; and by increasing or decreasing the provocation of a crime. The
use of technology in the prevention of car crime comprises an instructive case study.
There is strong evidence that technology has been effective in producing substantial
drops in the theft of and from cars since the early 1990s. Moreover, the technology
has become increasingly “elegant”, by which we mean that it is operated by default,
is unobtrusive, is pleasing (or at least non-displeasing) aesthetically, brings benefits
that exceed its costs and does not threaten civil liberties.
We believe that the success of car security devices including, in particular, their
individual and collective “elegance” has important implications for the develop-
ment of future security technologies. Consumers have embraced changes in retail
methods, the development of credit and debit cards, multipurpose mobile phones
and lightweight laptop computers. These developments have also increased the sup-
ply of crime opportunities. Responding with or, better still, incorporating powerful,
elegant security technologies to curtail those opportunities seems to us to be impor-
tant. It would be better that the producers of these crime opportunities try to contain
them. Failing that it may be necessary for governments to step in using regulatory
powers to require that security be added to reduce crime opportunities that are,
sometimes unknowingly, created (see Brown, 2013b; Clarke & Newman, 2005 on
forms of persuasion). But events of recent decades suggest there are grounds for
optimism that security technology can dramatically reduce crime while increas-
ing the liberty of all.

References

Brown, R. (2013a). Reviewing the effectiveness of electronic vehicle immobilisation: Evidence


from four countries. Security Journal, 1–25. DOI: 10.1057/sj.2012.55.
Brown, R. (2013b). Regulating crime prevention design into consumer products: Learning the les-
sons from electronic vehicle immobilisation. Trends and Issues in Crime and Criminal Justice
No 453. Canberra: Australian Institute of Criminology.
Clarke, R. V. (1997). Situational crime prevention: Successful case studies (2nd ed.). Monsey, NY:
Criminal Justice Press.
Clarke, R. V. (1999). Hot products: Understanding, anticipating and reducing demand for stolen
goods. Police research series paper 112. London: Home Office.
Clarke, R. V. (2012). Opportunity makes the thief. Really? And so what? Crime Science, 1(3), 1–9.
Clarke, R. V., & Newman, G. (2005). Modifying criminogenic products: What role for govern-
ment? In R. Clarke & G. Newman (Eds.), Designing out crime from products and systems
(Crime prevention studies, Vol. 18, pp. 7–83). Monsey, NY: Criminal Justice Press.
388 G. Farrell and N. Tilley

Cornish, D. B., & Clarke, R. V. (2003). Opportunities, precipitators and criminal decisions: A reply
to Wortley’s critique of situational crime prevention. Crime Prevention Studies, 16, 41–96.
Eck, J. E., & Eck, E. B. (2012). Crime place and pollution: Expanding crime reduction options
through a regulatory approach. Criminology and Public Policy, 11(2), 281–316.
Ekblom, P. (1999). Can we make crime prevention adaptive by learning from other evolutionary
struggles? Studies on Crime and Crime Prevention, 8, 27–51.
Farrell, G., Tilley, N., Tseloni, A., & Mailley, J. (2011). The crime drop and the security hypothesis.
Journal of Research in Crime and Delinquency, 48(2), 147–175.
Farrell, G., Tilley, N., & Tseloni, A. (2014). Why the crime drop? In M. Tonry (Ed.), Why crime
rates fall and why they don’t, volume 43 of crime and justice (pp. 221–290). Chicago: University
of Chicago Press.
Farrell, G., Tseloni, A., & Tilley, N. (2011). The effectiveness of car security devices and their role
in the crime drop. Criminology and Criminal Justice, 11(1), 21–35.
Laycock, G. (2004). The U.K. car theft index: An example of government leverage. In M. G.
Maxfield & R. V. Clarke (Eds.), Understanding and preventing car theft, crime prevention
studies (Vol. 17). Monsey, NY: Criminal Justice Press.
Newman, G. (2004). Car safety and car security: An historical comparison. In M. G. Maxfield &
R. V. Clarke (Eds.), Understanding and preventing car theft (Crime prevention studies, Vol. 17,
pp. 217–248). Monsey, NY: Criminal Justice Press.
Perry, H. W. (1910). Anti joy ride devices. Scientific American. 49, 58, 70.
Tilley, N., Farrell, G., Grove, L. E., Thompson, R., & Tseloni, A. (2015). Do burglar alarms
increase burglary risk? A counterintuitive finding and possible explanation. Crime Prevention
and Community Safety: An International Journal, 17(1), 1–19.
Tilley, N., Farrell, G., & Clarke, R. V. (2015). Target suitability and the crime drop. In M. A.
Andresen & G. Farrell (Eds.), The criminal act: The role and influence of routine activities
theory (pp. 59–76). London: Palgrave Macmillan.
Tseloni, A., & Thompson, R. (2015). Securing the premises (pp. 32–35). February: Significance.
Tseloni, A., Thompson, R., Grove, L. E., Tilley, N., & Farrell, G. (2014). The effectiveness of
burglary security devices. Security Journal. doi:10.1057/sj.2014.30.
Webb, B., Smith, M., & Laycock, G. (2004). Designing out crime through vehicle licensing and
registration systems. In M. G. Maxfield & R. V. Clarke (Eds.), Understanding and preventing
car theft (Crime prevention studies, Vol. 17, pp. 67–84). Monsey, NY: Criminal Justice Press.
Whitehead, S., & Farrell, G. (2008). Policing and corporate social responsibility: Anticipating
mobile phone “smart wallet” crime. Policing: A Journal of Policy and Practice, 2(2),
210–217.
Wortley, R. (2001). A classification of techniques for controlling situational precipitators of crime.
Security Journal, 14, 63–82.
Wortley, R. (2008). Situational precipitators of crime. In R. Wortley & L. Mazerolle (Eds.),
Environmental criminology and crime analysis (pp. 48–69). Cullompton, UK: Willan.
Index

A predictive, 218
ABMs. See Agent-based models (ABMs) properties, 220
Aborted robberies strengths, 219–220
bystanders, 29 temporal dynamics, 219
characteristics, 24, 25 Ambiguity function waveform analysis, 347, 348
co-offenders, 30 Anticipatory prevention, 235
doubt, 30 ABM-simulations, 236, 237
fear of victims, 30 artificial society, 237
home robberies attraction values, 238
gloves, 26 built-in prevention elements, 237
plan, 25 commercial/entertainment businesses, 239
scared, 25 computational methods, 240
unexpected events, 25 guardian agents, 237
victims, 25, 26 guardian movement, 240
moral concerns, 29 offenders and targets, 237
physical, 29–30 police patrol strategies, 239
planning robberies, 27–28 route organisation, 239
police, 28–29 simulation and practical application, 236
robbery events, 22 stylised facts, 236
victims, 30 surveillance strategies, 238
Accompanying Administrative Document tax evasion simulation, 238
(AAD), 257 victimised targets, 239
Acquired immunodeficiency syndrome vulnerable point, 239
(HIV/AIDS), 297
Agent-based models (ABMs), 235
advantages, 219 B
applications, 216, 217 Banca Popolare Italiana (BPI) cases, 261
characteristics, 219 British Crime Survey, 382, 384
complementary approach, 216 Broker, 160, 195
computational criminology, 220, 221 Brothels
computational model, 215 alcohol, 299
decision-making process, 219 legal and illegal, 306
explanatory (see Explanatory ABMs) geographic distribution, 300
mechanisms, 216 licenced, 298
observation and manipulation, 219 unlicenced, 299

© Springer International Publishing Switzerland 2016 389


B. Leclerc, E.U. Savona (eds.), Crime Prevention in the 21st Century,
DOI 10.1007/978-3-319-27793-6
390 Index

C failures, 37
Car crime, 382–385 gang-related violence, 134, 137
Carousel fraud, 257 global level, 1
CAV. See Consumer Affairs Victoria (CAV) logistic regression model, 135
CCO. See Conjunction of criminal opportunity longitudinal modeling techniques, 136
(CCO) network approach, 135
Certification of Destruction (COD), 257 SCP (see Situational crime prevention (SCP))
City-to-city network, 172 social network analysis, 4
Colombian cocaine, 189 street gangs, 4
Commercial sex. See Victorian commercial technical process, 9
sex market terrorism, 1
Computational criminology, 220, 221 victimization, 135
Conjunction of criminal opportunity (CCO), 322 Crime prevention intervention
Consumer Affairs Victoria (CAV), 302, agents, 226
304, 306 characteristics, 229
CPL. See Crime proofing of legislation (CPL) crime events, 222, 223, 227
Crime drop, 332, 386 crime pattern theory, 226
Crime mapping depict criminological theory, 228
criminal events, 114, 115 environmental criminology, 222
ecological/atomistic fallacy, 115 implementation and evaluation design, 221
environmental design, 113 mechanisms, 224, 228
kernel density calculations, 114 optimal test, 223
kernel density mapping, 116–118 requirements, 227
local spatial analysis, 118–121 routine activity approach, 225
MAUP, 116 simulation study, 224, 225
sophistication without complexity, 121 social systems, 229
spatial crime analysis, 116 temporal and spatio-temporal variants, 225
temporal aggregation, 122–124 types, evaluation, 222
Crime mapping and spatial analysis validation, 228
crime analysis, 106 Crime proofing
crime control, 106 IA, 251
crime data/calls, 105 in European Union, 249, 250
crime theory, 103–105 risk assessment, 252
criminal opportunity, 101 Crime proofing of legislation (CPL), 276
criminology, 108–109 ex ante evaluation, 275
environmental criminology, 101 ex post evaluation, 275
hotspots, 101 ITTP, 276
mobile targets, 105 TPD (see Tobacco Products Directive
policing, 101 (TPD))
rational choice theory, 102 Crime Risk Assessment (CRA) Procedure, 269
routine activity theory, 102–103 Crime script analysis (CSA), 144, 147, 148
strengths and weaknesses, 107–108 actions, 52
Crime prevention, 307 behavior and experience, 49
adolescents, 135 child sexual offenders, 61
attacks, 2 compatible, 53
crime events, 4 crime events, 53
crime mapping, 4 criminology, 55
crime problems, 3, 4 future behaviour, 61
crime script analysis, 2, 3 human trafficking, 55
criminal networks, 2 insightful, 54
criminological role, 9 limitations, 52
detecting and tracking, 5 method/investigation, 52
development, 136 police data, 52
economic and social costs, 2 resilient, 54
Index 391

robbery, 51 structural properties


sexual offenses, 50, 53 communications, 194
social situations, 50 core and peripheral actors, 193, 194
systematic, 53 JJTE, 192, 193
Crime-proofing legislation, 294 SNRM/JPPM, 192
Criminal networks types, 192
classification, 188 triangulating information, 207
empirical approaches, 188 women, 207
graphic representation, 206 Dynamic analysis (DNA)
neutralise and dismantle, 189 illicit markets/financial supervision, 199
static and dynamic analyses, 191 investigation, phases, 200–205
target and dismantle, 196 SIENA, 200
CSA. See Crime script analysis (CSA) types, 199

D E
Decriminalisation ECRA. See Extended crime risk assessment
brothels, 297 (ECRA)
regulate prostitution (legalisation), 298 EU Member States (MSs), 257
sex work, 293 Explanatory ABMs, 221
DG Home, 250 characteristics, 217
Dividends-received deduction system, 253 cognitive functioning, 217
DJI Phantom 2 Vision+, 368–370 crime prevention (see Crime prevention
DNA. See Dynamic analysis (DNA) intervention)
Drone formal computational model, 217
micro, 367 mechanisms, 218
microphone, 364 simulation, 217
prevention, 359 Extended crime risk assessment (ECRA)
UAS, 359 limitations, 290
Drug trafficking network, 187, 192–194, menthol and slim cigarettes
201–206 costs/harm, 289
attributes and motivations, 196 criminological assumptions, 284
broker, 195 ITTP, 284, 285, 287, 288
centrality and intermediation perpetrators, 288
(betweenness), 195 victims, 288, 289
criminal networks, 188, 189
dynamic networks analysis, 199–201
ethnographic research, 190 F
hierarchical vision, 198 Foot and Mouth Disease (FMD), 254
investigation
first phase, 201–202
second phase, 201–204 G
third phase, 202–206 Gang rivalries
legal and illicit activities, 190 bloods and crips, 166, 167
mafia organisations, 187 hyperdyadic contagion, 167
non-mafia-type organisations, 187 intra-group conflict, 168, 169
police investigations, 188 partial transitivity, triads, 168
quantitative and qualitative reciprocity, dyads, 168
information, 189 spatial structure, 169, 170
scale-free networks, 195 types, 168
sex, 197 Gang violence, 166, 174–180
SNA (see Social network analysis (SNA)) asserting dominance, 170
Spanish Guardia Civil, 189 blood/crip, 171
static network analysis, 191–192 city to city, 181
392 Index

Gang violence (cont.) J


city-level violence, 170 JPPM, 192, 195, 198, 201, 206
eligible court cases, 171
injunctions, 182, 183
limitations, 184, 185 K
Los Angeles County, 166 Kernel density mapping
network generation, 172 crime prevention, 116
rivalries (see Gang rivalries) criminal events, 117
social networks, 166 spatial resolution, 117
strengthening intervention strategies, surface generation technique, 117
183, 184 violent crime, 117
structural variables, 172–174
subset violence
bloods and crips, 176, 177 L
properties, 180 Legislation
reciprocity, 174, 175 anti-money laundering, 253
transitive patterns, 177, 178 bearer share, 252
visualization, 178, 179 prohibition strategies, 254
Geometry crime theory turnover of officials, 253
awareness space, 103 Legislative/regulatory crime risk assessment,
crime attractors, 103 294, 301
crime events, 105 Local spatial analysis
crime generators, 103 crime displacement, 118, 119
crime pattern/problem, 104 criminal events, 119, 120
criminal opportunities, 103 global evaluation, 119, 120
social and physical environment, 104
4th Generation (4G) wireless networks, 355
M
MAUP. See Modifiable areal unit problem
H (MAUP)
High-stress traffic environments, 77 Mechanisms for Assessing the Risk of Crime
HIV/AIDS. See Acquired immunodeficiency (MARC) project, 249
syndrome (HIV/AIDS) Micro-Doppler, 365, 372, 373
Mitigating prevention, 235
co-occurrence, 238
I de-escalating aggressive behaviour, 241
IAB. See Impact Assessment Board (IAB) de-escalation of (consequences of)
Illicit trade of tobacco products (ITTP) offending, 240
activities, 276 execution phase, 237
costs/harm, 289, 290 heart rate, 241
crime, 284, 286 real perpetrators, 238
crime risk probability, 280 Modifiable areal unit problem (MAUP), 116
criminal opportunities, 276 Multi-model netted sensor fence, 363
menthol and slim cigarette, 282
perpetrators, 288
victims, 288, 289 N
Impact assessment (IA), 251 NetRAD, 367, 368
Impact Assessment Board (IAB), 290
Initial screening (IS), 277
Intra-group conflict, 168, 169 O
IS. See Initial screening (IS) OBR. See Offender-based research (OBR)
Italian National Authority, 248 Offender-based research (OBR)
ITTP. See Illicit trade of tobacco products crime, 11
(ITTP) crime control, 13
Index 393

law, 12 manufacture, 157


motivation, 11 roles, 157–158
SCP technique, 11 criminal network
sexual contact, 11 associations, 159
street vendors, 13 brokers, 160
street-level crime, 12 legitimate and illegitimate Actors, 160
Offenders’ perspectives, 37 network structure, 161
business environment, 35 strategic position, 160–161
crime, 35 risks, 159
crime science, 42 Private cost, 268, 269
failures Proofing the legislation, 270–272
characteristics, 38 Prostitution
criminal opportunity, 39 criminal laws, 294
gambling habit, 38 legalisation, 298
losses, 39 morality, 296
opportunities, 41 offences, 296
opportunity, 39 Prostitution Control Act (1994)(Vic), 299, 301
fear, 38 Prostitution Regulation Act 1986 (Vic), 298,
iatrogenic effect, 37 301, 307
instrumental crimes, 36
limitations, learning, 35
routine activities, 37 R
script analysis, 43 Radar cross section (RCS), 366, 372
traditional crime prevention, 35 Radar detection
Organized crime and crime prevention aircraft detection, 364
crime commission process, 146 architecture, 364
criminal groups, 145 Micro-Doppler, 365
criminal networks, 146 numerous parameters, 365
drug trafficking, 145 uses, 365
social structures and links, 147 Range Time Interval (RTI), 371
street and violent crime, 145 Rational choice theory
crime prevention, 102
criminal opportunities, 102
P effort, 102
Parrot AR Drone v2, 364 risks, 102
PCRA. See Preliminary crime risk assessment RCS. See Radar cross section (RCS)
(PCRA) Robbery
Planning (Brothels) Act 1984 (Vic), 296 analysis, 23–24
Potential-victim and actual-victim attack, 88–89
scripts, 81 characteristics, 25, 31, 83
Predictive ABMs, 218 counterfactual strategy, 20
Preliminary crime risk assessment (PCRA) crime and daily activity, 84
banning menthol, 280, 287 crime incidents, 83
banning slim cigarettes, 280, 287 crime prevention, 82, 94–95
European tobacco market, 279, 281 crime victims, 77
herbal products, 286 crimes-in-action, 78, 92
ingredients and emissions, 279, 282 criminal and driving histories, 78
labelling and packaging, 279, 283–285 criminal behaviors, 97
nicotine-containing products, 286 daily activities, 77
novel tobacco products, 286 driver reactions, 78
Prevention strategies drug sales area, 90
crime commission process GPS tracking, 93
chemical precursors and equipment, high-crime areas, 77
156–157 hunting process, 32
394 Index

Robbery (cont.) bystanders, 72


individual track, 85, 96 crime script stage, 67
interpersonal script, 92 crime set up phase, 71
interpretation problems, 33 public awareness, 71
intimidating threats, 86–87 rape, 72
money and physical attack, 91 situational crime prevention, 73
police attention, 85 rape cases, 59
reactive tactics, 91 script analysis, 73
recruitment, 22 script and actions
sample and interviews, 22–23 continuation, 66
sexual abuse, 79 instrumental actualisation, 66
sexual assault, 97 interaction and crime completion, 66, 67
taxi driver, 77, 82 offender approach method, 66
taxi driving, 78–79 offender/s and victim prehistory, 63
victims, 31 post-actions, 67
Routine activity theory setting, 63
crime problems, 103 sexual assault, 66
motivated offender, 102 sexual violence, 61
place manager, 102 situational crime prevention, 60, 62
RTI. See Range Time Interval (RTI) victims, 59
violence, 60
SIENA. See Simulation Investigation for
S Empirical Network Analysis (SIENA)
SCP. See Situational crime prevention (SCP) Signal-to-noise ratio (SNR), 348
Script analysis. See also Crime script analysis Simulation Investigation for Empirical Network
(CSA) Analysis (SIENA), 200, 208
crime scripts, 79–80 Simulation-based Training of Resilience in
crime-in-action, 81 Emergencies and Stressful
reactive tactics, 80–81 Situations (STRESS), 240–241
victim reports, 79 Situational crime prevention (SCP), 380
victim scripts, 80 agents, 16
Semi-illegal market, 294 commit, 10
Sex work crime opportunities, 10
criminal and regulatory control, 304 law, 10, 14
criminal penalties (decriminalisation), 298 risk, 10
HIV/AIDS, 297 robbery, 10
legal problems, 298 time and place, 10
licenced brothels, 298 Situational prevention measurement
Planning (Brothels) Act 1984 (Vic), 297 EU Commission, 248
regulate prostitution (legalisation), 298 legislation, 247
Sex Work Act 1994, 299, 301, 307 loopholes, 248
Sexual offenses against women regulatory gaps, 247
acquaintance rape, 62 Smash-and-grab gangs, 361
attack, 60 SNA. See Social network analysis (SNA)
court transcripts, 64–65 SNR. See Signal-to-noise ratio (SNR)
crime event, 60, 73 SNRM, 192, 195, 202, 206, 207
crime script, 61, 62, 68–70 Social cost, 269, 289–290
criminal opportunity, 60 Social crime, 293
data and analytic strategy, 63 Social network analysis (SNA)
motivation, 60 adoption, 132
prevention applications, 130
acquaintance rape, 73 category, 130
acquaintances, 72 cocaine trafficking network, 189
attitudes, 72 crime and terrorist, 143
Index 395

crime enforcement T
advantages, 132 Technology
agency, 132 accelerants, 333
despite evidence, 133 arms races, 338
hierarchical organization, 133 biological lag, 330, 331
strategic analysis, 133 bridging gaps, 324
topological features, 134 CCO, 322
crime groups, 144 co-evolutionary struggles, 332, 333
crime prevention, 134–137 crime, 334, 335
crime scenes, 144, 152, 154 crime science, 321
crime scripts, 154 cultural and biological evolutionary
criminal network, 143, 150, 151, 156 perspectives, 328–330
criminologists and law enforcement design contradictions, 325
agencies, 129 disadvantages, 337, 338
dependent variable, 131 ecological interaction, 323
drug trafficking, 143 environmental circumstances, 323
effectiveness and applicability, 188 equivalent protective factors/
etiology and dynamics, 132 opportunities, 335
interdependence, 130 factors, 321
law enforcement, 151 goods and services, 321
limitations, 161 hot products approach, 321
mathematical graph theory, 130 human agents’ capabilities, 325
methamphetamine, 149 human life, 320
methodological techniques, 129 innovation and design, 335–337
neighborhood networks and influence, 131 opportunities and problems, 323, 324
network map, 150, 151 problem-oriented approach, 323, 334
organized crime, 155 psychological approach, 321
static analysis, 191–192 resources and circumstances, 324
surveillance targets, 144 scripts and clashes, 326, 327
Spanish Guardia Civil, 189, 209 social and architectural context, 325
Spatial analysis. See Crime mapping and social institution, 320
spatial analysis technological change, 319, 320
Static network analysis tools and weapons, 321
classification, 191 Testo Unico Bancario (TUB), 253
socio-demographic attributes, 191 Three-step checklist
vulnerabilities, 191 accessibility, market, 263
Step-based checklist burdensome obligations, 256
criminal implications, 254 corporate structure, 268
structure, 255–256 crime risk indicators, 256
unintended crime risks, 254–269 discretionary power, 260
Subset-to-subset network, 172 distribution/production, 259
Supply-side crime prevention economic and organised crime, 262
commercial activities, 386 economic impact, 268
crime opportunities, 379, 380 economic/legal requirements, 266
electronic security devices, 386 excise diversion funds, 257
elegant technology, 385, 386 expected effects, 265
intended and unintended effects, 378, 379 grants, subsidies and compensation
security technology, 377 schemes, 258
technology jurisdiction, 261, 262
car crime, 382, 385 law enforcement activity, 259
decline, 385 law enforcement agencies, 264
security protection factors, 384, 385 legal persons, 267
situational crime prevention, 380, 381 market/sector, 263
Swedish Tax Legislation, 249 number of criminals, 265
396 Index

Three-step checklist (cont.) corporate managers and auditors, 300, 301


professional requirements, 266 crime-proofing legislation, 294
regulated sector, 261 decriminalisation, 293
skills and knowledge, 266 difficult and time-consuming measures, 302
socio-demographic characteristics, 267 legal brothels, 300
systems/structures, 263 legislative/regulatory crime risk
tax allowances, 257 assessment, 294, 301
taxation, 258 licencing, 301
victims, 267 on-the-spot challenges, 302
Tobacco Products Directive (TPD), 281 police law enforcement activity, 299
ECRA (see Extended crime risk prostitution, 294, 296
assessment (ECRA)) Prostitution Control Act (1994)(Vic), 299
impact assessment, 276, 278–279 Rural community resistance, 300
IS, 277, 280 semi-illegal markets, 294
PCRA, 277, 280 sex work, 297–298
TPD. See Tobacco Products Directive (TPD) Sex Work Act 1994, 299
uncertainty in justiciability, 302
vulnerability, market
U corruption, 304
UAS. See Unmanned aerial systems (UAS) criminal and regulatory control, 304
Unmanned aerial systems (UAS), 359 legal and illegal brothels, 305
Unmanned aerial vehicles (UAVs) multi-tiered approach, 304
architectures, 360 VSPBs. See Victim self-protective behaviors
classification, 360 (VSPBs)
conventional radar system, 366
crime prevention
control, 362 W
detection, 362 Wireless fidelity (WiFi) technology, 349–352
identification/recognition, 363 ambiguity function waveform analysis,
operational capabilities, 362 347, 348
tracking, 363 cost, 353
verification and evidence collection, 362 covert, 353
criminal activities, 361 crime and security applications
design, 366–367 emergency evacuation, 354, 355
detection perimeter monitoring around
multi-model netted sensor fence, 363 buildings, 355
Parrot AR Drone v2, 364 public urban areas, 354
DJI Phantom 2 Vision+, 368–370 though-the-wall surveillance, 354
drone crime, 359 4th generation (4G) wireless networks, 355
limitations, 372, 373 IEEE 802.11 communications
mobile phone, 361 protocols, 345
NetRAD, 367–369 micro-Doppler analysis, 356
radar detection, 364–366 motion sensitivity, 353
RCS, 372 operational environments, 353
RTI, 371 passive radar, 345, 346
smash-and-grab gangs, 361 proof-of-principle experiments
indoor environment, 349–351
through-wall detection, 351, 352
V signals, 347
Victim self-protective behaviors (VSPBs), 80 SNR performance simulation, 348, 349
Victorian commercial sex market, 304, 305 surveillance receivers, 347
attraction/repulsion, 305–306 through-wall detection, 353
clarity test, 301 uncooperative detection, 352

You might also like