Home Office Research Study 217
The economic and social costs of crime
Sam Brand and Richard Price
Economics and Resource Analysis
Research, Development and Statistics Directorate
Home Office
The economic and social costs of crime
Home Office Research Studies
The Home Office Research Studies are reports on research undertaken by or on behalf of
the Home Office. They cover the range of subjects for which the Home Secretary has
responsibility. Titles in the series are listed at the back of this report (copies are available
from the address on the back cover). Other publications produced by the Research,
Development and Statistics Directorate include Research Findings, the Research Bulletin,
Statistical Bulletins and Statistical Papers.
The Research, Development and Statistics Directorate
RDS is part of the Home Office. The Home Office’s purpose is to build a safe, just and tolerant
society in which the rights and responsibilities of individuals, families and communities are
properly balanced and the protection and security of the public are maintained.
RDS is also a part of the Government Statistical Service (GSS). One of the GSS aims is to
inform Parliament and the citizen about the state of the nation and provide a window on the
work and performance of government, allowing the impact of government policies and
actions to be assessed.
Therefore –
Research Development and Statistics Directorate exists to improve policy making, decision
taking and practice in support of the Home Office purpose and aims, to provide the public
and Parliament with information necessary for informed debate and to publish information
for future use.
“The views expressed in this report are those of the authors, not necessarily those of the
Home Office (nor do they reflect Government policy).”
First published 2000
Application for reproduction should be made to the Communications and Development Unit,
Room 201, Home Office, 50 Queen Anne’s Gate, London SW1H 9AT.
© Crown copyright 2000 ISBN 1 84082 572 3
ISSN 0072 6435
i
Foreword
The costs of crime have become an increasingly important tool for decision-makers concerned
with crime and its impact on society. They help make explicit judgements about the relative
merits of alternative policies and programmes which are already implicit in decision-making
about how to allocate resources to tackling crime – both overall and between different types
of crime. However, the supply of good quality information on costs has not kept pace with the
demand for it. This study takes the first steps to addressing this problem.
Crime reduction and criminal policy is making progress but still a fair way behind some
areas of government in using evidence of effectiveness and cost effectiveness as the basis
for setting priorities and allocating resources. Many other departments routinely carry out
detailed cost-benefit appraisals and evaluations of new social policies. The Government’s
Crime Reduction Programme, and challenging new Public Service Agreements for the Home
Office, Criminal Justice System and other government bodies, are contributing to an
increased awareness of the role that cost of crime estimates can play in comparing the costs
of initiatives with the likely benefits that they can achieve.
Although they break new ground in this country, the cost estimates in this study are far from
perfect. Further work is necessary, and will be carried out, to ensure that the estimates are
robust, based on the best available evidence and capable of bringing a real change to the
way in which decision-makers at all levels view the problem of crime and how to tackle it.
Paul Wiles
Director of Research, Development and Statistics
Home Office
If you would like to comment on this paper please contact:
Philip Witcherley
Home Office Economics and Resource Analysis Unit
Room 271
50 Queen Anne’s Gate, London SW1H 9AT.
United Kingdom
Tel: +44 (0)20 7273 3284
Fax: +44 (0)20 7273 4013
E-mail: [email protected]
Acknowledgements
Many members of the Research, Development and Statistics Directorate commented and
assisted in the development of this study. We would particularly like to thank Carole Willis
and Sanjay Dhiri in ERA for their support and advice throughout. Thanks also go to Mark
Weiner and Phil Witcherley for taking on the project at an important stage and steering it in
the right direction.
We would like to thank the many people inside and outside the Home Office who
commented on various drafts of the study. Particular thanks go to Pat Mayhew for her incisive
and constructive comments and to Tracey Budd and Joanna Mattinson in the British Crime
Survey team for their help and guidance in extracting relevant information from the survey.
The CJS project team and consultative group on the development of the cost of crime
performance measure for the CJS have both generated a lot of debate and constructive
criticism, and have helped to ensure the real-world relevance of a potentially arcane study.
Particular thanks go to Simon Fish and Geoff Lewis (HM Treasury), Rob Culligan
(Association of Chief Police Officers), Paul Rayner (Serious Fraud Office) and Cecilia French
(Home Office), who all provided substantial contributions. Andrew Healey, Susanna
Mourato, Giles Atkinson and Ann Netten, of the Personal and Social Services Research
Units at the London School of Economics, University College London and the University of
Kent at Canterbury also kindly agreed to allow us to use some of their material in this study.
Finally, this work builds on the efforts of other cost of crime researchers around the world. In
particular, the pioneering work of Ted Miller, Mark Cohen and Brian Wiersema at the
National Institute of Justice in the US has helped accentuate the importance of this type of
work and to further its acceptance into the policy debate.
Sam Brand
Richard Price
Economics and Resource Analysis Unit
Home Office
October 2000
ii
The economic and social costs of crime
iii
Contents
Foreword i
Acknowledgements ii
Executive summary v
Section I: Introduction 1
Why measure the costs of crime? 1
Reasons for publishing this paper 2
Why “the economic cost of crime” as a performance measure? 3
Previous research and estimates for other countries 5
Total costs and average costs 6
Structure of the paper 6
Section II: Incidence of crime 9
Categorising types of crime 9
Measuring the incidence of crimes 11
A multiplier approach to counting crimes? 14
Confidence in the estimates 14
Section III: Methodological issues and principles 17
Key principles 17
Economic cost 17
Opportunity cost 17
Transfer payments 18
Categories of cost 19
Who bears the costs of crime? 19
Costs in anticipation of crime 20
Costs as a consequence of crime 23
Costs in response to crime 24
Measurement techniques 26
Costs in anticipation of crime 26
Costs as a consequence of crime 27
Costs of the response to crime 30
Section IV: Estimates and analysis 31
The costs of crime against individuals and households 34
Property crimes 35
Personal crimes 39
Commercial and public sector victimisation 43
iv
The economic and social costs of crime
Overview 43
Burglary not in a dwelling 46
Theft from a shop 46
Theft of and from commercial vehicles 46
Criminal damage 46
Robbery or till snatch 46
Fraud and forgery 47
Drug crime 50
Traffic and motoring offences and other non-notifiable offences 50
Wider economic distortions 52
Total cost of crime 56
Overview 56
Costs of crime by cost category 56
Costs of crime by offence type 57
Section V: Using and developing the estimates 59
Why the estimates are useful 59
Pitfalls to avoid in using the estimates 60
Development of cost of crime work programme 62
Emotional and physical impact on victims 62
Sexual offences 63
Quality of life and fear of crime 63
Police costs 64
Health services 64
Section VI: Appendices 65
Appendix 1: Best, low and high average cost estimates for selected
offence types 65
Appendix 2: Construction of a CJS performance measure 78
Appendix 3: Data sources and workings 79
Crimes against individuals and households 79
Commercial and public sector crime 80
Fraud and forgery 81
Drug crime 81
Traffic and other non-notifiable offences 81
Wider economic distortions 81
Section VII: Bibliography and related material 83
Executive summary
Every day decisions are made by policy makers and managers in the Criminal Justice System
which reflect implicit judgements about the relative seriousness of different crimes, or about
the benefits of pursuing one approach to reducing crime rather than another. This study
represents a first step towards making such judgements more explicit and in making sure they
better reflect the available evidence on the impacts on society of different types of crime.
Cost of crime estimates can play an important role in helping the government to achieve the
greatest impact on crime for the money spent. They can be used in both appraisal and
evaluation of crime reduction policies, such as those in the Government’s evidence-based
Crime Reduction Programme. They can help us to prioritise, focusing scarce resources on
policies that have the biggest impact on harm caused by crime, rather than simply the
number of crimes. Moreover, one of the two aims of the Criminal Justice System (CJS) is “to
reduce crime and the fear of crime and their social and economic costs”. This study reports
on progress towards a cost of crime measure that can be used to assess performance
against this aim. Figures used here represent the best available evidence, but nevertheless
needs to be much improved. The aim of this report is to stimulate debate and improvements
in the evidence.
The study concentrates largely on offences falling under notifiable offence categories.1 Not
all crimes are included in the study. The costs of drug trafficking and possession, handling
stolen goods, public order offences, other low-level disorder, fare evasion, summary and
non-summary motoring offences and other summary offences are not estimated.2 The study
does not, therefore, attempt to estimate the costs of all crime, but rather a subset of crime
where reliable information is available on the cost and the number of offences committed.
In order to get a true picture of the total impacts of crimes in notifiable offence categories,
we need to estimate the actual number of crimes in these categories, rather than the number
of crimes recorded. An approach has been devised which, as far as possible, links the total
estimated number of offences in a given year to changes in the number of offences recorded
by the police in that year. The British Crime Survey has been used to estimate actual
numbers of offences where possible. Table 1 gives details of the offence types for which
average cost estimates are presented.
v
1 Offences that police forces record and are required to report to the Home Office.
2 The cost of the criminal justice response is included in total for drug offences, motoring offences and other
summary offences, as is the cost of accidents involving illegal speed. These are, however, acknowledged to be
only partial estimates of the full cost of these offences.
Table 1: Notifiable offence categories in this study
Crime category Sub-categories included in this study Notifiable
offence codes3
Violence against Homicide 1-9; 11-15; 37.1
the person More serious offences
(excluding Homicide)
Less serious offences
Common assault 104; 105
Sexual offences 16-27; 74
Robbery Robbery of personal property 34A; 34B
Robbery of business property
Burglary Burglary and aggravated burglary 28-31
in a dwelling
Burglary and aggravated burglary
not in a dwelling
Theft and handling Theft of a vehicle 37.2; 39-49; 126
stolen goods Theft from a vehicle
Attempted theft of/from a vehicle
Theft from a shop (including
theft by an employee/other)
Theft of commercial vehicle
Theft from commercial vehicle
Other theft (including theft
of pedal cycle,theft from person,
other theft,but not handling
stolen goods)
Fraud and forgery4 51-53; 55; 60; 61;
814
Criminal damage Criminal damage against 56-59
individuals/households
Criminal damage against
commercial/public sector
vi
The economic and social costs of crime
3 As given in Appendix 3 of Criminal Statistics 1998 (Home Office, 1998a).
4 Only total costs are estimated for fraud and forgery.
‘Costs of crime’ in this paper refer to the full range of impacts of crime, approved where
possible in monetary terms – though this does not suggest that it is either straightforward or
always right to reduce the consequences of any crime into purely financial terms. Costs are
incurred in anticipation of crimes occurring (such as security expenditure and insurance
administration costs), as a consequence of criminal events (such as property stolen and
damaged, emotional and physical impacts and health services), and responding to crime
and tackling criminals (costs to the criminal justice system).
Costs have been measured using surveys of victims, such as the British Crime Survey and
Commercial Victimisation Survey, and estimates of industry turnover and costs, such as the
security and insurance industries. Resource cost estimates for the criminal justice system have
been derived from a model developed by the Home Office to track flows and costs through
the criminal justice process. Emotional and physical impacts of crime are, for the time being,
estimated using figures for people’s willingness to pay to avoid road traffic accidents, but
work is underway to derive better estimates reflecting more accurately of the impacts of
crime on victims.
Average costs of crime vary widely between offence categories. The most costly property
crimes are theft of vehicles, costing around £4,700 per incident.5 Burglaries cost an
average of £2,300, and criminal damage around £500. Personal crimes are far more
costly on average than property crimes. Homicides have been estimated to cost at least £1
million, with other violence against the person costing on average £19,000 per incident.
Robberies incur costs of almost £5,000 on average. Common assault is the least costly
personal crime, with an average cost of around £500 per offence.
The total cost of crime to England and Wales in 1999/2000 is estimated at around £60
billion, although this figure is still far from comprehensive, as it does not include important
costs such as fear of crime or quality of life impacts. Table 2 shows how this £60 billion is
split, by type of cost (such as property stolen, security expenditure and criminal justice system
resources) and by offence category (such as violence, robbery or burglary). Around £19
billion of the total cost of crime is the cost of property stolen or damaged. Nearly £18 billion
of the total is the direct emotional and physical impact on victims of crime, with a little over
£14 billion of this incurred as a result of violent crime. The response to crime by the CJS
constitutes around 20 per cent of the total cost of crime, at around £12 billion. Identifiable
costs in anticipation of crime – security expenditure and insurance administration costs –
came to over £5 billion, the bulk of this being security expenditure.
vii
Executive summary
5 All figures are given in 1999 prices.
viii
The economic and social costs of crime
Table 2: Summary of average and total cost estimates, by crime type and cost category
In response
In anticipation of crime (£) As a consequence of crime (£) to crime (£)
Property Emotional and Criminal Number of TOTAL
Security Insurance stolen and physical impact Lost Victim Health Justice System Average incidents COST
Offence category expenditure administration damaged on victims output services services (incl. Police) cost (£) (000s) (£ billion)
Crime against individuals and households
Violence against the person 2 – – 13,000 2,500 10 1,200 2700 19,000 880 16.8
Homicide – – – 700,000 370,000 4,700 630 22,000 1,100,000 1.1 1.2
Wounding (serious and slight) 2 – – 12,000 2,000 6 1,200 2,700 18,000 880 15.6
Serious wounding 10 – – 97,000 14,000 6 8,500 13,000 130,000 110 14.1
Other wounding 0 – – 120 400 6 200 1,300 2,000 780 1.5
Common assault 0 – – 240 20 6 – 270 540 3,200 1.7
Sexual offences 2 – – 12,000 2,000 20 1,200 3,900 19,000 130 2.5
Robbery/Mugging 0 40 310 2,400 420 6 190 1,400 4,700 420 2.0
Burglary in a dwelling 330 100 830 550 40 4 – 490 2,300 1,400 2.7
Theft 40 30 310 160 10 0 – 60 600 7300 4.4
Theft (not vehicle) – 20 130 100 4 0 – 90 340 3,800 1.3
Vehicle theft 70 50 500 220 20 0 – 30 890 3,500 3.1
Criminal Damage 10 20 190 200 30 0 – 60 510 3,000 1.5
All crime against individuals
and households (£ billion) 0.7 0.5 4.1 17.0 2.9 0.0 1.3 5.7 2,000 16,400 32.2
Commercial and public sector victimisation
Burglary not in a dwelling 900 50 1,200 – 40 – – 490 2,700 960 2.6
Theft from a shop 30 – 50 – – – – 20 100 31,000 3.1
Theft of commercial vehicle 3,400 1,500 4,600 – 60 – – 70 9,700 40 0.3
ix
Executive summary
Theft from commercial vehicle 240 110 320 – 10 – – 30 700 60 0.0
Robbery or till snatch 1,200 100 1,500 590 120 – 50 1,400 5,000 70 0.4
Criminal damage 340 20 440 – 30 – – 60 890 3,000 2.6
All commercial and public
sector victimisation (£ billion) 3.2 0.2 4.2 0.0 0.1 0.0 0.0 1.4 260 9.1
Fraud and forgery
All fraud and forgery (£ billion) 1.1 – 10.3 – – – – 0.6 – 9200 13.8
Traffic and motoring/other
non-notifiable offences
Illegal speed – – – – – – – – – – 0.9
Drug offences – – – – – – – – – – 1.2
Other indictable
non-motoring offences – – – – – – – – – – 1.0
Indictable motoring offences – – – – – – – – – – 0.5
Summary non-motoring offences – – – – – – – – – – 0.4
Summary motoring offences – – – – – – – – – – 0.8
All traffic and motoring/other
non-notifiable offences (£ billion) – – – 0.7 0.2 – 0.0 3.9 – – 4.8
TOTAL COST OF CRIME (£ billion) 4.9 0.6 18.6 17.7 3.3 0.0 1.3 11.6 – – 59.9
Notes:
1. Figures may not sum to total due to rounding errors
2. – indicates that no figure has been estimated
The average cost estimates given in this study are best estimates of costs given the
information available, but are inevitably imprecise. The quality of the available evidence on
the costs of crime is good in some cases, patchy in many, and poor in several. Some costs,
such as the fear of crime, or the impacts of crime on victims’ families, have not been
estimated, due to lack of data or lack of appropriate techniques through which to gather
data. Some costs are based on estimates from other fields of research. The cost estimates
are therefore sensitive to changes in assumptions made or to improvements in the quality of
the supporting data.
Throughout the study we attempt to highlight the problems with, and gaps in, the evidence,
and to identify the priorities for further work to ensure that these estimates can be used with
greater confidence. New methods need to be developed to estimate the costs of the fear of
crime and precautionary behaviour undertaken to reduce the risk of becoming a victim of
crime. Better estimates are needed for the emotional and physical impact on victims of
crime, health service costs, central and local government resources devoted to crime
prevention, and police resources. The Home Office has commissioned new research on the
emotional and physical impact of violent crime on victims in order to fill what is possibly the
most uncertain and important gap in our knowledge.
x
The economic and social costs of crime
Section I Introduction
Why measure the costs of crime?
Crime imposes a huge cost on society. Estimates from a number of recent studies range
widely from £35 billion to £60 billion per year.6 The potential savings to individuals and
households, businesses and the public sector from effective crime reduction measures are
therefore extremely large. Cost of crime estimates in this study show, for example, that
achieving the Government’s target of a 30% reduction in thefts of and from vehicles by
2004 could lead to savings to society of around £1 billion. A cost of crime measure
therefore provides a justification for resources spent on reducing crime, and provides an
indication of how successful the Government is at reducing the impacts of crime.
Estimates of the social and economic costs of crime7 can have an important role in
achieving the greatest impact on crime for the money spent. They can increase the
awareness of both policy-makers and the public in general of the full impact of crime on
society and the potential gains that could result from reductions in crime. Estimates of the
costs of individual crimes enable us to make better-informed decisions about which policy
measures are the most effective, by allowing meaningful comparisons to be made of the
costs and benefits offered by alternative crime reduction measures. They can also help us to
prioritise, focusing scarce resources on policies that have the biggest impact on harm
caused by crime, in addition to the number of crimes.
The estimates can be used both for policy appraisal – to value the likely benefits from
implementing alternative policy proposals, and so weigh these up against the likely costs of
implementation – and policy evaluation – identifying the size and value of the benefits that
have accrued from a policy. As in other policy areas, cost-benefit analysis cannot fully
encompass political or equity dimensions of appraisal and evaluation, and it is only one of
a number of complementary techniques. It does, however, provide a good basis for
answering many key questions about crime and crime prevention, such as:
1
6 For England and Wales
7 Throughout this paper we use the concept of “ social cost” in its economic sense – that is, the full impact on
society. The terms ‘social cost’, ‘economic cost’ and ‘social and economic cost’ are therefore used
interchangeably in this study. This includes costs imposed on individuals, households, businesses or institutions
by crimes they suffer directly (private costs) and wider impacts on society as a whole through, for example,
responses to the perceived risk of crime (external costs). The social cost of crime therefore includes both financial
costs reflected in expenditure, and ‘notional’ costs reflecting best assessments of the less tangible impacts of
crime, such as the emotional and physical impact on victims.
how can we use our existing resources in the most effective way?
how can we reduce the total cost of crime to society?
what is the correct level of resourcing for crime reduction activity?
should we concentrate only on preventing crime or should we do more to
mitigate its consequences?
Reasons for publishing this paper
This research paper serves a number of purposes:
To make public and open to debate research that the Economics and Resource
Analysis Unit of the Home Office has been engaged in over the last two years,
to share information and highlight major findings. The report will ensure that
the figures are open to scrutiny, so that they can be improved and gaps in the
data can be filled.
To provide information for the Crime Reduction Programme (CRP), a
comprehensive range of initiatives building on an evidence base of ‘what
works’ in reducing crime, and aimed at achieving the greatest impact on crime
for the money spent. An analysis of the costs and benefits of all CRP projects
will be a key part of the evaluation and future development of the programme.
Estimates of the cost of crime will allow us to estimate the savings generated
through CRP initiatives. These savings can then be compared with the costs to
show how cost-effective the initiatives have been.
To enable Crime and Disorder partnerships, local government officials,
criminologists, police, prison and probation service managers and those in
other operational agencies to carry out cost-benefit analyses that are
comprehensive and consistent. The paper aims to be accessible to anyone
working in the field of crime reduction.
To provide a basis for the development of a performance measure for the
Criminal Justice System (CJS). In the CJS Strategic Plan 1999-2002 the
Government has set the CJS the objective of reducing the economic cost of
2
The economic and social costs of crime
crime by 31 March 2002. The CJS Business Plan 2000-2001 notes that “the
costs to be tracked have been determined. A programme will be published in
Summer 2000 on the data available for those costs and detailing the work
continuing during 2000-2001 to improve data which is currently sketchy.” A
target for the reduction in the cost of crime is to be set by 31 March 2001.
This publication identifies the coverage of the costs that will be tracked, the
data that is currently available, and further work that is or will be happening in
2000-2001 to improve the accuracy of the estimates.
There are strong links between these different aims. A high-level understanding of the main
impacts of crime and the relative seriousness of different types of crime is vital in
highlighting areas where criminal policy needs to focus. Cost-benefit analysis of alternative
measures can help to inform the Criminal Justice System and other agencies about the most
effective mix of policies to bring down the cost of crime.
The estimates given in this study are far from comprehensive – rather, they represent a first step
towards a comprehensive set of estimates. Both the methodology and the estimates will be revised
on the basis of new information and research. The study does not cover the many issues involved
in cost-effectiveness and cost-benefit analysis for appraisal or evaluation. More details on
appraisal and evaluation of crime reduction initiatives can be found in Dhiri and Brand (1999).
Why “the economic cost of crime” as a performance measure?
One of the two key aims of the Criminal Justice System (CJS) is “to reduce crime and the
fear of crime and their social and economic costs”8 . In support of this aim, objective three
commits the criminal justice system to “a reduction in the economic cost of crime by 31
March 2002”.
The economic, or social, cost of crime is essentially a measure of the impact of crime on
society. It gives us a way of measuring the impact of policies aimed at reducing crime and its
consequences. Some crimes clearly have greater consequences than others. For example, a
murder has a greater impact on society than a shoplifting offence. A cost of crime performance
measure is designed to focus criminal justice system policy-makers and practitioners on the
most cost-effective solutions to crime, by ensuring due account is taken of both the effectiveness
of crime prevention measures and the relative seriousness of different offences, rather than
3
Introduction
8 CJS Strategic Plan 1999-2002 (Criminal Justice System, 1999). The distinction made in the Strategic Plan
between ‘social’ and ‘economic’ costs is not made in this document – the definition of economic costs
encompasses all possible social impacts.
simply focusing on the aggregate volume of crime. Figure 1.1 shows the striking difference in
the relative importance of different crimes against individuals and households9 when a) the
volume of offences is considered, and b) the cost to society is considered.
Figure 1.1: Volume and cost of offences
Volume:
Cost:
Figure 1.1 refers only to crimes against individuals and households – over 16 million crimes were estimated
to be committed in this category each year, at a total cost of around £32 billion. The relative proportions
shown in this figure will therefore differ from those based on a more comprehensive set of offence categories.
4
The economic and social costs of crime
Violence 53%
Sexual offences 8%
Robbery 6%
Common assault 5%
Burglary 8%
Theft of and from vehicles 10%
Other theft/handling 4%
Criminal damage 5% Attempted vehicle theft 1%
Violence 5%
Sexual offences 1%
Common assault 19%
Burglary 8%
Theft of and from vehicles 20%
Other theft/handling 22%
Criminal damage 17%
Attempted vehicle theft 5%
Robbery 3%
9 Figure 1.1 does not include crimes against the commercial and public sector, fraud and forgery, drug offences
or other non-notifiable offences.
The relative importance of violent crime in comparison with other, property crimes is
marked. When we focus on the volume of offences violent crimes come to around a quarter
of the total. When we focus on the cost of those offences rather than the volume, violent
crimes constitute nearly three-quarters of the total cost. This finding is one example of the
way in which cost of crime estimates can help illuminate potential areas where gains may
be made by new policies or the switching of resources from one area to another.
Comparisons of the relative CJS resources devoted to preventing or mitigating different
offences can also be made. Estimates of the average social cost of different offence types
can help decision-makers to assess whether the allocation of resources between
programmes in the CJS is suitably related to the overall impact of the crimes each
programme seeks to address.10 For example, CJS costs are estimated to represent over half
the total cost of common assault, but only around one-tenth of the total cost of theft and
handling offences.
Appendix 2 gives information on how a cost of crime performance measure could be
constructed from the estimates in this study.
Previous research and estimates for other countries
The total cost of crime has received attention in the past. Reports have been published by
various organisations on the total cost of crime, using varying degrees of sophistication in
their calculations. In 1998, the Association of British Insurers (Association of British Insurers,
1998b) calculated that the total cost of crime exceeded £35 billion. In 1999 an Audit
Commission Report, Safety in Numbers (Audit Commission, 1999), estimated the cost at
£50 billion a year. In 2000 a report in the Observer newspaper (Observer, 2000) adapted
figures from a paper published in the US (Anderson, 1999) to calculate that the annual cost
of crime in Britain was £60 billion. These estimates are based on different assumptions,
cover different crimes, costs and years. They do not imply that the cost of crime has
increased from £35 billion to £60 billion between 1998 and 2000. Neither is it clear that
the similarity between some of these estimates and the total cost estimate in this study is
more than coincidental, since the methodologies used differ, at least at the margin.
A number of international papers have attempted to cost crime in other industrialised
countries. Miller, Cohen and Wiersema (1996) investigated the cost to victims in the US of
5
Methodological issues and principles
10 It should be noted that other considerations such as deterrence or confidence in the CJS may also affect levels of
resourcing.
violent and property crime, including “pain and suffering”, and found the total cost to be
around $450 billion per year. Cohen (1998) attempted to estimate the monetary value of
saving a high-risk youth from a lifetime of delinquency and criminal activity. Aos, Phipps,
Barnoski and Lieb (1999) have created a cost-benefit model to evaluate crime prevention
activities in Washington State, USA, which compares the costs of crime prevention activity
with savings to the criminal justice system and to victims. Palle and Godefroy (1998) have
described plausible estimates for the monetary value of offending in 1996 to France, though
this study does not provide estimates of the pain and suffering of victims. The total cost of
crime to Australia in 1996, incorporating most of the cost categories in this paper, was
estimated by Walker (1997).
Total costs and average costs
The total cost of crime (which has received the most attention in recent years) and average
(or unit) costs of crime are both useful. The total costs of crime is important in assessing the
scale of the impact of crime. The total cost can also be broken down to get a good idea of
the magnitude of different types of cost, or of the contribution of particular types of crime to
the total impact on society. Average costs are vital in conducting cost-benefit analyses to
assess the value for money of individual policy initiatives. Average cost of crime estimates
focus on individual incidents, and allow us to get an idea of the relative impacts on average
of, for example, one theft of a vehicle in comparison with one robbery. Both are important
in bringing down the cost of crime in the most effective ways. This paper presents
information on both total costs and average costs.
Structure of the paper
Section I deals with the rationale for estimating the cost of crime, and for a cost of crime
performance measure.
Section II considers how to define and count criminal activity for the purposes of this exercise.
A method of measuring the incidence of actual victimisation and of tracking this through
time is developed, and its advantages and disadvantages discussed.
Section III explains some key economic concepts, and identifies and defines the different cost
categories and the components of each cost category that will be used in the exercise. A
methodology for the measurement of each cost component is considered. Alternative
6
The economic and social costs of crime
measurement techniques including stated preference or contingent valuation, surveys and
valuation using market prices are highlighted.
Section IV provides estimates of the average costs of crime for a range of different offence
types, and total cost estimates by crime type and cost category. It gives comparisons with
other estimates, and considers the implications of the estimates for crime reduction and
crime mitigation, for policy development and for CJS practitioners.
Section V discusses how the cost estimates should and should not be used. It attempts to
identify areas where our estimates need improvement and highlights areas where no
estimates are currently available. In the light of this discussion, some recommendations for
further work are made.
Section VI contains appendices outlining the data sources used and how estimates were
derived, and Section VII contains a bibliography and references.
7
Methodological issues and principles
8
The economic and social costs of crime
Section II Incidence of crime
Categorising types of crime
Defining what constitutes a crime often involves applying a rigid set of rules to complex
social interactions. Criminal activity ranges widely in scope, including, for example, murder,
damage to people or property, intimidation, appropriation of property, taking proscribed
substances and forging banknotes. Various methods have been devised to try to categorise
these activities, but for consistency, the categories used in this study are notifiable offence
categories (the types of offences that police forces record and are required to report to the
Home Office). This captures the majority of crimes that are likely to have the most severe
impacts, and makes the process of updating the figures and comparing them with the
volume of offences much simpler.
In addition to the notifiable offence categories, some non-notifiable offences which tend to
be relatively less serious in nature but sometimes have grave consequences are also
included. Driving above the speed limit, for example, would usually not cause direct harm
either to people or property, and if detected, would probably involve only a fixed penalty.
However, sometimes, driving over the speed limit causes or contributes to accidents
involving serious injury or loss of life.
The crimes covered by this study, and the sub-categories that have been used to divide these
categories into meaningful blocks for analysis, are listed below in Table 2.1.
9
Table 2.1: Notifiable offence categories in this study
Crime category Sub-categories included in this study Notifiable
offence codes11
Violence against Homicide 1-9; 11-15; 37.1
the person More serious offences
(excluding Homicide)
Less serious offences
Common assault 104; 105
Sexual offences 16-27; 74
Robbery Robbery of personal property 34A; 34B
Robbery of business property
Burglary Burglary and aggravated burglary 28-31
in a dwelling
Burglary and aggravated burglary
not in a dwelling
Theft and handling Theft of a vehicle 37.2; 39-49; 126
stolen goods Theft from a vehicle
Attempted theft of/from a vehicle
Theft from a shop (including
theft by an employee/other)
Theft of commercial vehicle
Theft from commercial vehicle
Other theft (including theft
of pedal cycle,theft from person,
other theft,but not handling
stolen goods)
Fraud and forgery12 51-53; 55; 60, 61;
814
Criminal damage Criminal damage against 56-59
individuals/households
Criminal damage against
commercial/public sector
10
The economic and social costs of crime
11 As defined in Appendix 3 of Criminal Statistics 1998 (Home Office, 1998a).
12 Only total costs are estimated for fraud and forgery.
There are many crimes which are not included in this list. Offences relating to the possession
or trafficking of drugs are not included, other than property crimes committed to fund drug
use, which are included under burglary, robbery and theft.13 The number of notifiable
offences falling under an “other notifiable offences” category, and the huge number of
other, non-notifiable, criminal activities, such as low-level disorder, fare evasion and
“breaches of the peace”, which could potentially have an impact on society, have not been
estimated. For some of these crimes, limited cost information is available. For others, neither
the number of offences nor cost information has been included. Table 3.1 in Section III gives
more details of the costs which are and are not estimated in this study.
Measuring the incidence of crimes
Whilst the notifiable offence categories have been used to determine the types of crime on
which this study will focus, the number of notifiable offences recorded by the police have not
been used as a measure of the incidence of crime. The number of recorded offences does
not reflect the actual number of offences committed. The police can record only those crimes
that come to their attention. Some incidents reported to the police are not recorded as a
notifiable offence, either because they may not fall into a notifiable offence category, or
because there may be insufficient evidence that a crime has actually taken place.
The British Crime Survey measures crimes against adults living in private households in
England and Wales. The 1998 survey estimated that, of the crimes that can be compared
with notifiable offence categories, “less than half were reported to the police, and only
about half of those that were reported were recorded” (Mirrlees-Black, Budd, Partridge and
Mayhew, 1998). In other words, the true number of offences against adults and households
was perhaps four times that recorded by the police. This conclusion is now well known and
widely recognised in the CJS. Offences that are not covered by the British Crime Survey,
such as shoplifting or fraud and forgery, are likely to have much lower reporting rates than
those that are covered. The total number of incidents in all notifiable offence categories is
therefore likely to be significantly more than four times the amount recorded by the police.
We need to know for each crime category the actual number of incidents occurring in England
and Wales each year so that we can estimate the actual impact of crime on society, not just the
11
Incidence of crime
13 Bennett (2000) notes that, of a sample of arrestees in the second developmental stage of the NEW-ADAM (New
English and Welsh Arrestee Drug Abuse Monitoring) programme, “ over two-thirds of the highest-rate offenders
(20 offences a month or more) reported using heroin or crack/cocaine” (p. ix). Nine per cent of all arrestees
were high-rate offenders, using heroin or crack/cocaine. This group was estimated to be responsible for over
half of all reported offences. Forty-two per cent of arrestees in the study thought that their drug use and crime
were connected.
Table 2.2: Estimated numbers of incidents, 1999-2000
Type of crime Recorded Crime Multiplier on Source of multiplier Estimated number
April 1999 to recorded estimate of actual incidents
March 2000 offences 1999/2000
(000s) (000s)
Crimes against individuals and households
Violence against the person 387 2.3 880
Homicide 1.1 1.0 none 1.1
Other violence against the person 386 2.3 BCS (1998) 880
of which: More serious offences 29 3.6 BCS (1998) 110
Less serious offences 357 2.2 BCS (1998) 780
Common assault 194 16.7 BCS (1998) 3,200
Sexual offences 38 3.5 BCS (1998) 130
Robbery
Robbery from individuals 72 5.8 BCS (1998) 420
Burglary
Burglary in a dwelling 443 3.2 BCS (1998) 1,400
Theft and handling
Theft from the person 76 9.9 BCS (1998) 760
Theft of a pedal cycle 131 3.5 BCS (1998) 460
Theft of vehicle 321 1.2 BCS (1998) 380
Theft from vehicle 566 3.9 BCS (1998) 2,200
Attempted vehicle theft 157 6.1 BCS (1998) 950
Other theft and handling 639 4.0 estimate 2,600
12
The economic and social costs of crime
Table 2.2 continued
Criminal Damage
Against individuals or households 473 6.3 BCS (1998) 3,000
Crimes against commercial and public sector
Robbery
Robbery of business property 12 5.8 CVS (1994) 70
Burglary
Burglary not in a dwelling 464 2.1 estimate 960
Theft and handling
Theft from a shop 292 100.0 estimate 29,000
Theft of commercial vehicle 0 N/A CVS (1994) 40
Theft from commercial vehicle 0 N/A CVS (1994) 60
Theft by employees (comm/public sector) 17 15.3 CVS (1994) 270
Theft by others (comm/public sector) 0 N/A CVS (1994) 1,400
Criminal Damage
Against commercial/public sector 473 6.3 estimate 3,000
Fraud and forgery
Fraud and forgery 335 42.6 NERA (2000) 9,200
Notes:
1. Source for recorded crime statistics: Table 6, Home Office Statistical Bulletin 12/00 (2000).
2. BCS = British Crime Survey. The number of crimes upon which the BCS multiplier estimates in Table 2.2 are based are adapted from Table 4.1 and Appendix C
of the 1998 British Crime Survey (Mirrlees-Black et al., 1998). The estimated number of actual incidents are different from those quoted in the British Crime
Survey because they have been adjusted to include crimes against under-16s and crimes recorded by the British Transport Police.
3. CVS = Commercial Victimisation Survey (Mirrlees-Black and Ross, 1995).
4. NERA = National Economic Research Associates. NERA (2000) estimated the total actual number of fraud and forgery offences, rather than a multiplier on
recorded offences. The multiplier estimate is therefore the total estimated number of offences divided by the number of recorded offences.
5. Figures may not sum to totals due to rounding.
6. Sources of unpublished multiplier estimates: Homicide – assumed that all offences are recorded. Other theft and handling – roughly equal to the multiplier for all
BCS crime, and for all comparable BCS property theft. Burglary not in a dwelling14 – half the BCS estimate for burglary in a dwelling. Theft from a shop – based
on survey of literature on nature and extent of shoplifting by Farrington (1999). Criminal damage against commercial and public sector – multiplier assumed
equal to multiplier for criminal damage against individuals and households.
13
Incidence of crime
14 2.1 = [(3.2-1)/2]+1
impact of those crimes that are reported or recorded. The British Crime Survey goes some way to
achieving this. However, crimes committed against commercial or public sector targets (e.g. theft
from a shop), or where there is no direct victim (e.g. handling stolen goods, some fraud and
forgery), are not covered by the survey, though it is believed that under-recording might be much
higher for some of these offences. For these crimes alternative data sources have been explored.
A multiplier approach to counting crimes?
The approach taken to counting crimes has been determined largely by the need to track
changes in the cost of crime over time, for the cost of crime performance measure. This
requires a system of counting crimes that can be readily updated and is not subject to wide
variations in data quality over time.
For this reason an approach has been devised that, as far as possible, ties the estimated
total number of incidents to changes in the number of recorded offences. For each crime, a
multiplier has been calculated equal to the ratio of the actual estimated number of crimes to
the number of crimes recorded. Multipliers have generally been calculated for the calendar
year 1997 in order to allow consistent comparison between the British Crime Survey and
recorded offences. Where no clear basis for calculating a multiplier exists, a figure has
been estimated. Where it is highly unlikely that the number of actual offences is related to
the number of recorded offences, and an estimate for the actual level of victimisation exists,
this estimate has been used for each year rather than a fixed multiplier.
Once multipliers have been calculated for all the offences of interest, they are applied to the
most recent recorded crime figures – April 1999 to March 2000 at the time of publication –
to construct a total number of incidents figure for each category. Table 2.2 gives details of
recorded crimes, multiplier estimates and their sources, and the estimated total number of
incidents for April 1999 to March 2000. (See pages 12 and 13).
Confidence in the estimates
Some of the estimates given in Table 2.2 are clearly more robust than others. Those
estimates derived from a comparison of British Crime Survey data and comparable
recorded crime figures are more robust than those estimates based on expert opinion but
little hard data. Even for these estimates, the relationship between the amount of crime
recorded by the police and the amount of crime estimated by the British Crime Survey may
14
The economic and social costs of crime
change over time. Mirrlees-Black et al. (1998) show that trends in recorded crime, reported
crime and BCS crime have differed somewhat between 1981 and 1997. This problem is
particularly acute where reporting rates have historically been low but may now be rising,
such as for domestic violence or racially-motivated offences. In April 1998 the police crime
recording rules changed in a number of ways. Although the estimated one-off effect of these
counting rule changes has been accounted for, differences in the types of offence now
recorded relative to the previous crime counting rules may affect the future relationship
between recorded and actual levels of crime.15
Even where the British Crime Survey offers an estimate, this may not be accurate. For
domestic violence and sexual offences in particular, there are factors at work that may
distort the true picture – for example, victims may be unwilling to report incidents to
interviewers where they have a close relationship with the offender, or where the offender
may be present when completing the survey. The British Crime Survey does not publish its
estimate of the level of sexual victimisation due to concerns over the accuracy of the results.
A self-completion module was introduced in the 1994 sweep of the survey (Percy and
Mayhew, 1997). This resulted in a much higher count of sexual victimisation than estimated
either by police recorded crime or British Crime Survey estimates of victimisation. However,
the estimate raised as many questions as it answered. In particular, the self-completion
responses magnified an issue already present in the main survey – that many victims did not
consider what happened to them to be a crime, but rather “just something that happens”,
even though what happened was legally a crime. This issue serves to highlight the tentative
nature of the multiplier estimate, and whilst the standard BCS estimate of the number of
sexual offences used in this study is likely to underestimate the true level of victimisation, no
reliable conclusions can be drawn about the extent of underestimation.
Fraud, theft from a shop and handling stolen goods are other areas where multiplier
estimates are particularly tentative. The estimate of just over 9 million fraud offences is
drawn from a report on the economic cost of fraud (NERA, 2000) commissioned by the
Home Office and the Serious Fraud Office as part of the development of a cost of crime
performance measure for the criminal justice system. The report acknowledges the partial
nature of this estimate and the fact that it is not suitable for tracking the total number of
offences each year.16
15
Incidence of crime
15 From 2000, the British Crime Survey will be run annually, on an increased sample size. This should allow more
regular monitoring and, if necessary, updating of the multiplier estimates.
16 The NERA estimate of the number of incidents of fraud each year is based on a summation of published
information from many different sources, including HM Customs and Excise, the Department of Health, the
Department of Social Security, the British Bankers Association and many others.
Theft from a shop is another hugely under-reported offence. Estimates of the number of
customer thefts are provided by the Commercial Victimisation Survey (CVS) (Mirrlees-Black and
Ross, 1995) and the Retail Crime Survey 1998 (British Retail Consortium, 1999). The CVS
counted nearly 6 million customer thefts in 1993, and the Retail Crime Survey nearly 4 million
in 1997. These estimates, however, require the retail outlet or head office to be aware that the
theft has taken place. Farrington (1999) brought together a number of studies on shoplifting.
He noted that police recorded crimes reflected only between 1 in 100 and 1 in 1000
shoplifting incidents in two department stores studied in 1984. Self-report data from various
studies also suggested that between 1 in 40 and 1 in 250 shoplifting offences led to a
conviction or caution.17 Given the uncertainties involved in these calculations, this study has
taken a fairly conservative approach and assumed 100 offences per recorded offence.
No estimates were made of the number of handling stolen goods offences, drugs offences,
other notifiable offences, traffic and motoring offences or other non-notifiable offences.
16
The economic and social costs of crime
17 There were nearly 120,000 offenders cautioned or convicted of theft from a shop in 1998 (Criminal Statistics,
1998). If each offender has been convicted or cautioned for 2 acts of shoplifting on average, and if we use
Farrington’s central assumption of 1 caution of conviction for every 150 offences, we find that there were
around 120,000 x 2 x 150 = 36 million offences. There were 281,000 recorded offences of theft from a shop
in 1998-99. Dividing the 36 million by 281,000 gives us a multiplier of around 128.
17
Section III Methodological issues and principles
Key principles
Economic cost
This study uses the terms “economic cost” and “social cost” to mean to full impact of crime on
society, to individuals, households, businesses and institutions, and encompassing both
“financial” impacts of crime and allowing a “notional” value for impacts which are not fully
or directly reflected in the financial consequences of crime – such as trauma and physical
injury. A distinction is sometimes made between the “economic” and “social” costs of crime.
Economic costs in this distinction are taken to mean financial costs – costs that can be readily
expressed in cash terms, such as stolen property or the cost of a prison place. Social costs
are taken to mean the impacts on society that cannot be readily expressed in cash terms.
This distinction, however, is a false one, reflecting practical difficulties with estimation rather than
any real differences. It would be misleading and incomplete to measure the economic cost of
crime in terms only of those costs that are already expressed in cash terms as this would omit
important impacts of crime and so would tell only part of the story. Crimes such as robbery or
violence against the person, which have significant ‘intangible’ costs, would appear much less
serious than they actually are, whilst other crimes would appear relatively more serious. It is
therefore important to try and quantify all the impacts of crime in common terms as far as
possible. Money can be used in this situation simply as a means of comparing one thing (e.g.
the physical impact of a broken leg) with another (e.g. the cost of a hospital bed).
This study treats the economic and the social costs of crime as one and the same, and holds
that, where at all possible, estimates should be made for all the impacts of crime. Simply
including costs which are easiest to measure often means excluding costs which have the
most severe impacts – such as the physical and emotional suffering of victims, to the
detriment of informed decision-making.
Opportunity cost
Opportunity cost is a central concept in economics. Measuring the opportunity costs of crime is a
key method of valuing the economic cost of crime to society. According to “The Green Book” (HM
Treasury, 1997), the opportunity cost of a resource is “the value of the resource in its most valuable
alternative use”. The concept of opportunity cost allows us to value the human, physical and
financial resources that will be ‘freed up’ for potential alternative uses when a crime is prevented.
Generally, the best measure of the opportunity cost of a resource is its market value, or price.
For example, the opportunity cost of a burglar alarm costing £100 is equal to the £100 that
cannot then be used to buy groceries. However, not all resources have a market value. The
emotional suffering of a person staying indoors at night because of the fear of crime is not
traded on the market, but still represents an opportunity cost to the extent that that person
values going out. Non-traded opportunity costs require different measurement approaches.
Transfer payments
The legal transfer of resources from one party to another occurs in many contexts within the
legal economy, for example through social security payments, subsidies or gambling. Such
transfers are not generally regarded in themselves as a loss to society. Crime too involves
some similar transfers; for example, property crimes involve a transfer of property from the
victim to the offender. The fundamental distinction between a transfer and a loss to society is
the distinction between a wanted and an unwanted transfer. A burglary, theft or robbery
involves an illegal transfer of property that is unwanted by one party, the victim, and the
transfer of the property out of the legal economy. This study treats transfers out of the legal
economy and into the illegal economy as costs of crime.
Insurance claims also involve a transfer of resources. Potential victims who take out
insurance policies in anticipation of crime pay premiums to an insurance company. Actual
victims of property crime who have taken out an insurance policy receive money from the
insurance company. Hence resources have been transferred as a result of a crime, from
potential victims with insurance to victims with insurance. Unlike property that is transferred
from victims to offenders, insurance has been entered into voluntarily by both parties, and
insurance claims are thus treated in this study as a transfer payment, not as a loss to society.
The only resources involved in insurance that represent a cost of crime to society rather than
a transfer are the resources used in insurance administration. Insurance companies require
staff, premises and equipment in order to provide, check and pay out on policies. The
resources used in insurance administration represent an opportunity cost to society, because
in the absence of crime these resources could be employed in a productive way elsewhere
in the economy.
18
The economic and social costs of crime
Categories of cost
Who bears the costs of crime?
There are a number of ways in which the costs of crime can be categorised. One way is by
who bears them – victims, those at risk of becoming victims (potential victims), the criminal
justice system and other services.
Victims18 face costs as a consequence of crime, through having property stolen, damaged or
destroyed, from the opportunity costs of time spent dealing with the crime and through the
emotional and physical impacts of crime.
Potential victims bear costs in anticipation of crime, through measures to reduce the risk of
victimisation (defensive expenditure, precautionary behaviour, and community initiatives),
measures to reduce the consequences of victimisation (i.e. insurance), and through reduced
quality of life and fear of crime.
Society bears the costs of resources devoted to bringing offenders to justice through the
criminal justice process, involving the Police Service, the Crown Prosecution Service,
Magistrates and Crown Courts, Legal Aid, and the Prison and Probation Services.
Crime involves wider economic distortions, such as the reduction in shops, services, facilities
and job opportunities in high-crime areas. These are considered in more detail under “wider
economic distortions” in Section IV. Other costs are also incurred as a consequence of crime
by employers of victims, victim support services, health and education services, and by the
offender and his or her family.
An alternative approach is to break down the costs of crime in relation to individual
incidents. This approach draws on the typology used by Davidson (1999) in Costing
Burglary Reduction, a paper presented at the British Criminology Conference. Costs are
incurred in anticipation of crimes occurring (mostly falling on potential victims). They are
incurred as a consequence of criminal events (falling mainly on victims, but also on services
dealing with the consequences, such as health services). There are also costs consequential
on the response to crime (falling mainly on the criminal justice system). This study uses the
anticipation – consequence – response categorisation. Table 3.1 summarises the types of
cost that are included, excluded (on theoretical grounds) and the main costs that could not
be estimated (generally through lack of adequate data).
19
Methodological issues and principles
18 Victims include individuals, households, businesses, organisations and institutions. This study defines victims as
the direct victim of a crime, and not, for example, family and friends. Victims of all crimes, whether recorded or
unrecorded, are included in this study.
Table 3.1: Costs estimated and not estimated in this study
Estimated Not estimated
In anticipation of crime
Security expenditure Precautionary behaviour
Insurance resources Fear of crime/Quality of life of
potential victims
Collective/community defensive
expenditure
Government crime prevention
activity
Insurance premiums
As a consequence of crime
Property stolen and damaged Insurance claims
Lost output Quality of life of victims
Emotional and physical impact
Health services
Victim support services
In response to crime
Police Criminal Injuries Compensation
payouts
Prosecution Witness costs
Legal aid and non legally-aided defence costs Miscarriages of justice
Magistrates and Crown Courts Offender and his/her family
Probation Service
Prison Service
Jury Service
Criminal Injuries Compensation resources
Costs in anticipation of crime
Measures to reduce the risk of victimisation:
There are a number of adverse consequences of becoming a victim of crime, which will be
dealt with in more detail in the consequences of crime section below.19 Such consequences
are perceived by potential victims, if only imperfectly. Potential victims will therefore be
20
The economic and social costs of crime
19 These comprise property stolen and damaged, costs of time spent dealing with or recovering from an incident,
and emotional or physical impacts of crime.
generally willing to take action to reduce the chance or risk of becoming a victim, where the
perceived benefits of doing so (in terms of reduced risk) outweigh the costs involved in the
action (in financial and opportunity cost terms).
These measures comprise defensive expenditure – expenditure on security measures such as
burglar alarms, fencing, lighting, security guards etc., and precautionary behaviour, such as
taking taxis instead of public transport, avoiding particular people or places, or staying at
home after dark. These measures are a cost of crime – they are based on the perception of
potential victims of the risk of crime, which (at least in the long run) is linked to the actual
rate of crime.
There is in fact little theoretical difference between defensive and precautionary measures,
since both are an attempt to reduce the risk of victimisation. In practice, there is a difference
in that most defensive expenditure centres on reducing the risk of property crimes such as
burglary, whereas most precautionary behaviour is centred on reducing the risk of personal
crimes such as robbery or sexual offences.
It is important to note that for some precautionary and defensive expenditures, a reduction
in the probability of victimisation is not the only consideration involved. For example,
reasons for driving children to school are likely to include convenience, speed, warmth and
road safety, as well as to reduce the risk of violent crimes against children. Care must be
taken to allow for this in any costing of crime, since attributing the entire cost of any action
or expenditure that indirectly reduces the risk of crime will overstate the cost of crime.
Defensive expenditure is affected by many things other than the perceived risk of
victimisation. It is affected by the ability of the potential victim to pay for security equipment.
Many wealthy individuals may have a low risk of victimisation but spend a great deal on
security, whilst many individuals of more limited means may have a high risk of victimisation
but are unable to afford security equipment. Technology is an important driver of changes in
defensive expenditure – if vehicle immobilisers become much more effective at reducing
crime (or much cheaper to install), for example, then (independently of changes in the risk of
crime) more people will buy them, because their expected value has increased.
The security choices of fellow potential victims will also affect expenditure – increased action
by others may displace crimes onto softer targets, or may have a wider benefit for adjacent
targets. Other determinants of defensive expenditure include the price of equipment and the
ability of criminals to circumvent such measures.
21
Methodological issues and principles
Another issue in expenditure on security is choice – many security features now come as
standard in cars (e.g. steering locks, alarms) and houses (e.g. window locks). To the extent
that any such features reduce crime (and the reduced quality of life through fear of crime
etc.) rather than yielding other, non-crime benefits, the extra cost of incorporating these builtin
features at the manufacturing stage should be counted as a cost of crime. Since
measurement of the cost of built-in security measures is difficult, the balance between addon,
after-sale security and built-in features will inevitably affect our cost estimates artificially.
The circumstances of individuals may also affect their ability or willingness to undertake
precautionary behaviour. Some may be unable to afford precautionary expenditure such as
taxis home at night, or a car to transport children to school, or may decide that the costs of
taking action outweigh the potential costs of crime they may face through not taking action.
Some potential victims may be unable to take precautionary action as a result of social
circumstance. In cases of domestic violence, for example, potential victims may be, or feel,
unable to remove themselves as a target of crime.
Groups of potential victims also undertake measures to reduce the collective risk of
victimisation. Whilst these measures