Phone Theft Index

Abstract

This study presents what we believe to be the first Phone Theft Index, ranking the most stolen makes and models of mobile cell phones. It is based on over 200,000 recorded crimes involving such theft during 2004 and 2005. The theft “careers” of key phone models are examined. The market for mobile cell phones changes rapidly, as shown in theft career trajectories. The next steps in the development of the index would be for it to be risk-based, and routinely publicized on a timely and frequent basis. Complementary publications of information on “safe” handsets and practices would also stimulate the market in consumer safety.

Full text

Security Journal, 2008, 21, (212 – 227)
© 2008 Palgrave Macmillan Ltd 0955–1622/08 $30.00
www.palgrave-journals.com/sj
Phone Theft Index
Jen Mailley
a
, Roni Garcia
b
, Shaun Whitehead
a
and Graham Farrell
a
a
Midlands Centre for Criminology and Criminal Justice, Loughborough University , Loughborough LE11 3TU ,
U.K .
E-mail: G.Farrell@lboro.ac.uk
b
The National Mobile Phone Crime Unit, c / o Territorial Policing HQ, Victoria Embankment, Westminster,
London, SW1A 25L, UK.
This study presents what we believe to be the fi rst Phone Theft Index, ranking the most stolen
makes and models of mobile cell phones. It is based on over 200,000 recorded crimes involving
such theft during 2004 and 2005. The theft “ careers ” of key phone models are examined. The
market for mobile cell phones changes rapidly, as shown in theft career trajectories. The next
steps in the development of the index would be for it to be risk-based, and routinely publicized
on a timely and frequent basis. Complementary publications of information on “ safe ” handsets
and practices would also stimulate the market in consumer safety.
Security Journal (2008) 21, 212 – 227. doi: 10.1057/palgrave.sj.8350055
Keywords: mobile phone theft ; theft index ; risk and crime ; National Mobile Phone Crime Unit ;
cell phone theft
Introduction
Mobile phone theft stands out like a thumb sore from texting. Whereas most types of crime
declined in the last decade, mobile phone theft increased dramatically. According to the
British Crime Survey, violent crime fell 43 per cent ( Jansson et al ., 2006, p. 65 ) and house-
hold acquisitive crime 55 per cent (Hoare and Cotton, 2006 ) between 1995 and 2005.
In contrast, mobile phone theft rose sharply in the second-half of the 1990s (Harrington
and Mayhew, 2001 ). Victimized mobile owners rose from two per cent in 2000 – 2001
to seven per cent in 2002 – 2003 ( Allen et al ., 2005 ) during which time mobile phone
ownership surged ( Mintel, 2004 ). By 2006, there were more mobile phone connections
than people in the U.K. population ( Ofcom, 2006, p. 138 ). Knives, guns and extreme
violence have been used in mobile phone robberies ( BBC, 2002a, b, 2003a, b, 2006a, b ;
Mobile, 2006 ). Consequently, it is arguably not surprising that headlines have denounced
“ the mobile phone crime wave ” ( BBC, 2002c ) and the fact that mobiles “ fuel robbery rise ”
( BBC, 2006c ).
Harrington and Mayhew ’ s (2001) study of the nature and extent of mobile phone theft
remains the most comprehensive work to date. They estimated that there were 710,000
such thefts in a year, and that one in fi ve stolen mobiles were taken during either thefts from
the person (15 per cent) or robbery (4 per cent; p. 14). They found that one-third of mobiles
were taken during “ other thefts ” , often when a mobile was left unattended on public transport,
in offi ces or leisure facilities. Offenders were predominantly males aged under 24 years with
around one-third aged between 15 and 16 years (p. 36). Victims were also predominantly

Jen Mailley et al.
Phone Theft Index
213
male and over three-quarters of offences involved male offender(s) on male victims, with 39
per cent of victims aged between 18 and 29 years (p. 47).
There has been some progress in the development of prevention efforts to tackle mobile
phone crime. There has been an increased attention from the police, some increase in the
blacklisting of stolen mobiles and some effort to tackle the reprogramming of stolen phones
(see Mailley et al ., 2006 ). However, the extent of the problem means there is a need for
further work to design out the problem. It is increasingly recognized that designing out
crime can be effective. Efforts to design out car crime, described further below, may be the
fl agship for successful efforts relating to frequently stolen products. Yet it is often diffi cult
to develop anti-theft designs, and even more diffi cult to implement them. Consequently, the
role of incentives, particularly market-based incentives, have been identifi ed as critical to
stimulating crime prevention efforts (see Pease, 1998 ; Ledbury et al ., 2006 ). Among the
more prominent market-based incentives are theft indices. The role of theft indices in stimu-
lating design against crime is explored below, and the mobile Phone Theft Index detailed,
after the cost of mobile phone theft is briefl y examined.
The cost of mobile phone theft
Many readers might be surprised to learn that the average robbery imposes a cost upon so-
ciety which is greater than the average car theft. This is despite the fact that mobile phones
are generally less valuable than cars. The reason is that the cost of replacing the property is
only one component of the overall cost. Robbery victims can incur considerable costs in
terms of physical, emotional and psychological damage. In addition, costs related to the
criminal justice system, including the police, can be extensive, as can reduced productivity
if victims require time away from work. When such costs are included, the Home Offi ce
estimates a robbery costs £ 7,282 on average, while a car theft costs £ 4,138 ( Dubourg and
Hamed, 2005, p. 7 ). These cost estimates also put a quantifi able explanation on the fact that
robbery is often viewed as one of the more serious crimes. Clearly, not all mobile phone
thefts are robberies, but it is arguably the rise in robberies that, due to their seriousness, have
generated much policy interest.
Combining estimates of the average cost of crimes with the number of mobile phone
thefts gives an estimate of the total social cost of mobile phone theft. Harrington and
Mayhew ’ s estimate of the number of mobile phone thefts in the U.K., while now somewhat
dated, remains the most comprehensive. The total estimated cost of mobile phone theft in
the U.K. comes out at over £ 1 billion per annum ( Table 1 ). Clearly, this estimate could be
refi ned if more recent data were available, and this should be the aim of future research. It
is also worth noting that it is probably a conservative estimate. That is because Cohen et al .
(2004) , using a different methodology, estimated the average cost of crime to be several
times higher than the Home Offi ce estimates used here.
The role of theft indices
Theft indices are increasingly important in the repertoire of levers which can motivate
design against crime efforts. The earliest such index is, to our knowledge, that of the Highway

Phone Theft Index
214
Loss Data Institute in the U.S., which has produced a Car Theft Index since 1982 (see
Hazelbaker, 1997 ).
In 1992, the U.K. Home Offi ce produced its fi rst Car Theft Index ( Houghton, 1992 ). It
was preceded by two classic studies. The fi rst was a case study of steering column locks
forming part of the “ Crime as Opportunity ” study by Mayhew et al . (1976) ; see also
Mayhew et al. (1992) . The second was Southall and Ekblom ’ s exploration of the potential
for a “ crime free car ” ( Southall and Ekblom, 1985 ). Houghton ’ s index appeared in the Crime
Prevention Unit papers publication series of the Home Offi ce. Chronologically, the index
was sandwiched between a study of the nature and extent of the car crime problem ( Webb
and Laycock, 1992 ) and a study of car crime prevention in car parks ( Webb et al ., 1992 ), in
the same publication series.
While pioneering, the 1992 Car Theft Index was fairly crude. It offered only three cate-
gories of risk classifi cation: high, medium and low (see Houghton, 1992 , Tables 1 and 2 ,
pp. 16 – 20) and some limited analysis. It took 5 years for another iteration of the index to
emerge, but these have now been published annually since 1997 (see e.g. Home Offi ce,
2005a ). The Car Theft Index appears to have been closely followed by immensely success-
ful efforts on the part of car manufacturers to design out crime. In a recent review of the
impact of the Car Theft Index, Gloria Laycock wrote:
After several failed attempts to persuade the car manufacturers to reconsider the
design of vehicles and build in better security, the U.K. government developed an
index that ranked models of cars by their vulnerability to theft. This was a lever
intended to press the manufacturers into changing their behaviour. It raised the
profi le of car theft with the public and showed which types of vehicles were most
vulnerable … . ( Laycock, 2004, p. 25 ).
Since the publication of the Car Theft Index, there has been a dramatic decline in car
thefts in the U.K. It has not, to date, proved possible to develop an experimental evaluation
that excludes alternative possible explanations, yet it is clear that the decline in car theft
began not long after the publication of the Car Theft Index, and gained momentum thereaf-
ter. It is also the case that few other explanations have been offered to explain the surge in
Table 1 Annual cost of mobile phone theft in the U.K.
Crime type Per cent of crimes Total crimes Cost per crime Total cost ( £ )
Others thefts 32% 227,200 844 191,756,800
Theft from vehicle 29% 205,900 858 176,662,200
Burglary dwelling 20% 142,000 3,268 464,056,000
Theft from person 15% 106,500 634 67,521,000
Robbery 4% 28,400 7,282 206,808,800
Total 100 710,000 1,106,804,800
Notes : “ Per cent of crimes ” from Harrington and Mayhew (2001, p. 13, Figure 3.1). The percentages relate
to crimes against adults but are applied to all crimes here. Cost per crime is from Dubourg and Hamed (2005, p.7,
Table 2.1) .

Jen Mailley et al.
Phone Theft Index
215
anti-theft design efforts and the drop in crime. Figure 1 reproduces the key chart from Lay-
cock’s (2004) review, with the permission of the author.
A range of design innovations have emerged to tackle car theft. Rick Brown showed com-
pelling evidence of the impact of immobilizers ( Brown, 2004 ). Barry Webb et al . (2004)
scrutinized progress in the design of systems to monitor vehicle licensing and registration.
Brown and Clarke (2004) argued that, while there is evidence of increasing international traf-
fi cking in stolen cars, there is little substantive research on the issue. There are many parallels
between designing out car theft and designing out mobile phone theft. However, to date, anti-
theft design efforts relating to mobile phones are in their infancy. The more recent sibling of
the Car Theft Index is the Bike Theft Index ( Home Offi ce, 2002, 2005b ). This ranks stolen
mopeds, scooters and motorcycles in order of risk, and uses a similar methodology.
Inspired by the Car Theft Index, this study presents what we believe to be the fi rst mobile
Phone Theft Index. We acknowledge at the outset that further research should seek to build
upon this work, as outlined below. However, a mobile Phone Theft Index may assist in
Table 2 Makes of stolen mobile phones in 2004 ( N =129,186)
Rank Make Per cent Cumulative %
1 Nokia 55.3 55.3
2 Sony Ericsson 12.4 67.7
3 Samsung 11.4 79.2
4 Motorola 7.4 86.6
5 Siemens 3.9 90.5
6 LG 1.6 92.1
7 Sharp 1.7 93.8
8 Panasonic 1.7 95.5
9 Sagem 1.3 96.8
10 NEC 1.2 98.0
11 Hutchinson 0.3 98.3
12 O2 0.3 98.6
13 Vodafone 0.3 98.9
14 Phillips 0.3 99.2
15 Orange 0.2 99.4
16 Alcatel 0.2 99.6
17 T-Mobile 0.1 99.7
18 Virgin 0.1 99.8
19 Blackberry 0.1 99.9
20 BT 0.0 99.9
21 Sanyo 0.0 100.0
22 Palm 0.0 100.0
23 Toshiba 0.0 100.0
24 i-mate 0.0 100.0
25 Bosch 0.0 100.0
26 Qtek 0.0 100.0
27 Nikon 0.0 100.0
28 Sendo 0.0 100.0
Notes : Missing=12,753. All makes shown registered some thefts but those with too few to register to one decimal
place are shown as 0.0 per cent.

Phone Theft Index
216
providing customers with information that motivates the mobile phone industry into further
anti-crime efforts. The key difference between this index and its predecessors is that absolute
counts of crime rather than counts relative to sales or “ the population in circulation ” (wheth-
er cars, bikes or mobile phones) is the key measure. The only reason for this is that the data
on mobile phone sales by model are currently, to our knowledge, confi dential industry data.
Here, we keep with the defi nition used by Harrington and Mayhew (2001) for mobile
phone theft. This includes not only theft of various forms such as snatch-theft, theft from the
person, and theft from various public and private places including bars, shops, offi ces and
other workplaces. It also includes robberies and crimes where mobiles were taken, including
burglaries and theft from motor vehicles.
Data
The data analysed here were crimes recorded by police in Greater London and which in-
volved the theft of a mobile phone. The bulk of the preparatory work for the analysis con-
sisted of many weeks of data cleaning. Although the data were routinely collected by the
police, it was not in a format that was immediately conducive to analysis.
Police information technology typically records many items of information in free-text
fi elds. This was the case for the names of the makes and models of stolen mobile phones.
Free-text fi elds meant that anything could be typed into the available space. This allows for
spelling mistakes, typos, entries in the wrong fi eld, accidentally combining text that is meant
to be in two fi elds, blank fi elds, and other possibilities such as text saying as “ Awaiting in-
formation ” or similar. The main outcome was that the same manufacturer names and hand-
set model names were often recorded in many different ways. However, in order to count the
number of occurrences of each make and model in the database, it was necessary that
the information was in the same format for each name for every case in the data set. The
0
100000
200000
300000
400000
500000
600000
700000
1934
1938
1942
1946
195
0
19
54
19
58
1962
19
66
197
0
1974
19
78
198
2
1986
19
90
1994
1
998
200
2
Years
Numbers of Vehicles
First Car Theft
Index published
Figure 1 . Number of vehicles stolen by year.

Jen Mailley et al.
Phone Theft Index
217
analytic computer software counted only identically formatted names as the same name.
Hence, it would classify the following as nine different makes:
Motorola; Motor-rola; Motarola; Motorolas; Motorollo; Motorrolla; MOTOROLA;
Moto; Motoralo.
As well as differences in spelling, even differences in the use of upper or lower case charac-
ters, differences in syntax and the use of spaces, result in seemingly different manufacturers.
Even the relatively simple manufacturer name “ Nokia ” was recorded in over 400 different
ways, and others such as Sony Ericsson had far more.
The problem with the names of manufacturers was compounded by the issue of model
numbers. When model numbers, including their various prefi xes and suffi xes (is it the 6630
or the 6630i?) were introduced, the number of error permutations increased exponentially.
Although some analytic software can identify “ similar ” names within parameters, this type
Table 3 Makes of stolen mobile phones in 2005 ( N =125,840)
Make Per cent Cumulative %
Nokia 49.1 49.1
Samsung 13.7 62.7
Sony Ericsson 13.3 76.0
Motorola 11.4 87.3
LG 3.3 90.6
Siemens 2.6 93.2
NEC 1.2 94.4
Sagem 1.1 95.5
Sharp 1.1 96.6
O2 0.6 97.2
Hutchinson 0.5 97.8
Panasonic 0.5 98.3
Blackberry 0.3 98.6
Orange 0.3 98.9
Vodafone 0.3 99.2
Phillips 0.2 99.4
T-Mobile 0.2 99.6
Alcatel 0.1 99.7
Virgin 0.1 99.8
Sanyo 0.0 99.9
Palm 0.0 99.9
BT 0.0 100.0
i-mate 0.0 100.0
Toshiba 0.0 100.0
Qtek 0.0 100.0
Sendo 0.0 100.0
Fujitsu 0.0 100.0
Bosch 0.0 100.0
Nikon 0.0 100.0
Total 100.0
Notes : Missing=12,198. All makes shown registered some thefts but those with too few to register to one decimal
place are shown as 0.0 per cent.

Phone Theft Index
218
of analysis was not appropriate because it would have grouped together too many model
types that needed to be separately identifi ed (such as the D500 and D600, the 6610 and the
6310, the 3310 and the 3610, or the K700i and the K750i). Having said this, some handset
model names were grouped where this was appropriate. For example, the black and hot pink
Motorola Razr V3 and Razr V3i were grouped together as Motorola Razr, and the Nokia 3510
and 3510i were combined, as they were effectively the same handsets for present purposes.
This type of data cleaning is not uncommon when police records are used for research. It
has been widely used in relation to the study of crime hot spots and repeat victimization (see
e.g. Sherman et al ., 1989 ; Johnson et al ., 1997 ). In those analyses, it was the street names
and addresses of locations that needed cleaning so that same-location crimes could be iden-
tifi ed. Newer police IT systems now commonly use “ justifi cation ” for addresses, that is,
they are cross-referenced against a list of known street addresses so that errors are reduced.
In the future, perhaps, police IT systems should include such justifi ed fi elds to record more
accurately the names and models of frequently stolen goods.
To clean the many thousands of cases of data, a line of code (specifi cally, it was SPSS
syntax) was written for each instance where data cleaning was necessary. A single line of
code would clean-up the same error if it occurred elsewhere in the data. In total, over 13,000
lines of code were needed to clean the data; around a quarter of these ( n = 3,433) related to
manufacturer names and three-quarters to model names ( n = 10,000 + ). Names and models
were verifi ed against lists from manufacturer ’ s websites or from more extensive internet
searches as required. Where the make or model of the phone was an empty fi eld (or a version
of this such as “ awaiting information ” ) unrecognizable or appeared to be incorrect, the data
was classifi ed as missing. Around 70 per cent of cases required some form of data cleaning.
The end result was that, of the original 255,353 cases, over 90 per cent of cases were in-
cluded in the analysis relating to makes, and around 70 per cent relating to models. The
sample sizes are detailed in what follows.
The total number of recorded monthly mobile phone thefts remained fairly stable across
the 2 years in question. There was a fall in the number of mobile phone thefts recorded
during late 2004 and early 2005 but the monthly total climbed to its previous levels by
mid-2005. An explanation for this variation is not offered here although the reduction was
most evident in Nokia handsets, as detailed further below, and the resurgence represented by
increases in thefts of handsets from other manufacturers.
The most stolen makes
Tables 2 and 3 show the makes (manufacturers) of stolen handsets for 2004 and 2005, re-
spectively, and the percentage of thefts accounted for by their handsets. Nearly half of all
handsets stolen in 2005 were from Nokia. Samsung, Sony Ericsson and Motorola account
for most of the remainder, and this “ big four ” account for almost nine in ten thefts each year.
Around two dozen other makes account for the remainder, with LG overtaking Siemens for
fi fth place in 2005. Hence, stolen handsets are primarily from four manufacturers with
Nokia dominating the overall theft market. The extent to which this refl ects the market share
of these companies is discussed further below.
The number of thefts per month for the four makes most frequently stolen plus a group
for all other makes is shown as Figure 2 . There has been a steady decline in thefts of Nokia

Jen Mailley et al.
Phone Theft Index
219
models over the twenty-four months. Samsung thefts increased overall, because after expe-
riencing a dip in thefts in late 2004 and early 2005, there was a strong resurgence in the
second-half of 2005. Although many models of Motorola were stolen throughout the period,
the rise in overall prominence of Motorola in 2005 was accounted for almost solely by in-
creased thefts of the Razr models, described further below.
This brief glimpse at the rankings of manufacturers clarifi es why it is primarily models
from Nokia, Samsung, Motorola and Sony Ericsson that feature in the top-twenty stolen
handset models discussed in what follows.
A risk-based index by make
It was possible to develop a preliminary risk-based index by make of mobile phone. Sales
data by manufacturer was publicly available from Mintel (2004) . The distribution of sales
by manufacturer is shown in Table 4 alongside the distribution of thefts. The top row of
data relates to Nokia which accounted for 36 per cent of handset sales in 2004 and for 55
per cent of thefts. There are therefore a disproportionate proportion of thefts of Nokias rela-
tive to sales for that year, at the rate of 1.5 stolen for every one sold. This “ theft ratio ” is
shown in the fi nal column of the table for each manufacturer. According to this measure,
Nokia has the highest theft index, followed by Sony Ericsson and Samsung, each of which
have a theft ratio greater than one. Motorolas, Siemens, LG and “ Others ” have theft ratios
less than one, indicating that they accounted for a smaller proportion of thefts than sales in
2004. As with most indicators, this one is not perfect because the proportion of sales in a
given year may differ from the proportion of handsets in circulation or regular use. How-
ever, as a partial indicator, it suggests that, relative to sales that year, Nokia was the riskiest
phone to own in 2004.
0
1000
2000
3000
4000
5000
6000
Jan
04
Feb
04
Mar
04
Apr
04
May
04
Jun
04
Jul
04
Aug
04
Sep
04
Oct
04
Nov
04
Dec
04
Jan
05
Feb
05
Mar
05
Apr
05
May
05
Jun
05
Jul
05
Aug
05
Sep
05
Oct
05
Nov
05
Dec
05
Date
Number Stolen
Nokia
Sony E
Other
Samsung
Motorola
Figure 2 . Thefts by make, 2004 – 2005.

Phone Theft Index
220
The theft ratios can be compared between manufacturers. The ratio of 1.5 for Nokia can
be compared to the ratio of 0.8 for Motorola to suggest that, in 2004, a Nokia had a risk al-
most twice that of a Motorola (strictly speaking, 1.5 / 0.8 = 1.875 times riskier). Such com-
parisons suggest Nokia was 1.25 times riskier than Sony Ericsson, and 1.36 times riskier than
Samsung. Handsets from “ Other ” manufacturers had only 0.4 of the risk of Nokia according
to this measure. It is worth noting that this measure does not refl ect longer-term changes in
the market and the possible lag between sales and thefts. For example, if Motorola sales and
market share were increasing rapidly in 2004 then it could feasibly obtain a higher proportion
of sales relative to thefts due to this change rather than due to actual risk levels.
The most stolen models
The top-20 stolen handset models of 2004 and 2005 are listed in Tables 5 and 6 . Each table
shows the data by quarter for the year, with annual totals in the fi nal column of each table. The
data columns show the cumulative percentage of handsets stolen. The percentage stolen that
relates to each model can be derived by subtracting the cumulative percentage in any given
row from the value in the row above it. The exception to this rule is the fi rst data row in each
table, which is also the percentage of total stolen handsets accounted for by that model.
Only two models making the top-20 lists were not from the “ big four ” manufacturers.
The Sharp GX10 ranked 20th for the fi rst quarter of 2004, dropped to 26th in quarter two
then fell further down the charts ( Table 5 ). The LG U8120 entered the top-20 in 13th at the
end of 2004 and rose to eighth in 2005 before dropping out of the top-20 by the third quarter
( Table 6 ). These two models represented a high percentage of all thefts of handsets manu-
factured by LG and Sharp, although again this could refl ect sales and use patterns.
Theft careers
A well-known feature of the mobile phone market is that it is quickly evolving. New models
frequently enter the market. Owing to miniaturization they are often smaller. They have
new and interesting shapes, facia and screens. They have improved software and increased
memory. They have various new bells and whistles or improvements on the old models.
Table 4 Risk-based mobile phone theft index for manufacturers, 2004 ( N =116,433 stolen)
Make Sold (% of total) Stolen (% of total) Theft ratio
Nokia 36 55.3 1.5
Sony Ericsson 10 12.4 1.2
Samsung 10 11.4 1.1
Motorola 9 7.4 0.8
Siemens 10 3.9 0.4
LG 7 1.6 0.2
Others 18 7.9 0.4
Total 100 100.0
Note : Data on per cent share of sales is from Ofcom (2004) .

Jen Mailley et al.
Phone Theft Index
221
Table 5 Stolen mobile phones rankings, 2004
Rank Quarter 1 (n=22,131) Quarter 2 (n=22,237) Quarter 3 (n=20,538) Quarter 4 (n=19,090) Total for 2004 (n=83,996)
Make &
model
Cum % Make and
model
Cum % Make and
model
Cum % Make and
model
Cum % Make and
model
Cum %
1 Nokia 7250 9.1 Nokia 7250 7.7 Nokia 3310 5.3 Nokia 6230 9.7 Nokia 7250 6.6
2 Nokia 3310 16.2 Nokia 3310 13.4 Nokia 7250 10.5 Nokia 3310 14.0 Nokia 3310 12.2
3 Nokia 8310 22.8 Sony E T610 18.7 Samsung E700 15.7 Samsung E700 18.2 Sony E T610 16.9
4 Sony E T610 28.1 Samsung E700 23.9 Nokia 6230 20.8 Nokia 7250 21.9 Samsung E700 21.3
5 Nokia 6100 33.1 Nokia 8310 28.2 Sony E T610 25.2 Sony E T610 25.3 Nokia 8310 25.5
6 Nokia 7210 37.9 Nokia 6100 32.4 Nokia 6100 29.0 Nokia 6100 28.2 Nokia 6100 29.5
7 Samsung E700 41.1 Nokia 7210 36.1 Nokia 8310 32.3 Sony E K700i 30.9 Nokia 6230 33.2
8 Nokia 3510 44.3 Nokia 6310 39.2 Nokia 6600 35.4 Nokia 6610 33.4 Nokia 7210 36.5
9 Nokia 6310 47.4 Nokia 6600 42.2 Nokia 6310 38.2 Nokia 6600 35.9 Nokia 6310 39.4
10 Nokia 3410 50.5 Nokia 3510 44.9 Nokia 7210 40.9 Nokia 6310 38.3 Nokia 6600 41.9
11 Samsung V200 52.8 Nokia 3410 47.6 Nokia 3410 43.0 Nokia 8310 40.6 Nokia 3510 44.4
12 Nokia 8210 54.9 Nokia 3200 49.3 Nokia 3510 45.1 Nokia 7610 42.6 Nokia 3410 46.7
13 Nokia 3210 56.8 Samsung V200 51.0 Nokia 6610 46.9 LG U8120 44.5 Nokia 6610 48.6
14 Nokia 6610 58.5 Nokia 3210 52.6 Sony E T630 48.7 Nokia 3510 46.3 Nokia 3210 50.3
15 Sony E T68 60.0 Nokia 6610 54.1 Nokia 3210 50.4 Nokia 7210 48.0 Nokia 8210 51.8
16 Samsung A800 61.5 Nokia 3100 55.6 Motorola V600 52.1 Nokia 3410 49.6 Samsung V200 53.2
17 Nokia 3330 63.1 Nokia 7600 57.0 Nokia 3200 53.7 Motorola V600 51.2 Nokia 3200 54.6
18 Nokia 5210 64.5 Nokia 8210 58.4 Motorola V300 55.0 Sony E T630 52.8 Nokia 6210 55.8
19 Nokia 6600 65.9 Motorola V300 59.8 Sony E P900 56.4 Samsung E800 54.3 Nokia 3100 57.0
20 Sharp GX10 67.3 Samsung A800 61.1 Nokia 6210 57.7 Nokia 7600 55.8 Sony E T630 58.2
Other
a
Others=32.7% 100.0 Others=38.9% 100.0 Others=42.3% 100.0 Others=44.2% 100.0 Others=41.8% 100.0
a
Number of ranks of other models varied by Quarter: Q1=297, Q2=289, Q3=328, Q4=345.

Phone Theft Index
222
Table 6 Stolen mobile phone rankings, 2005
Rank Quarter 1 (n=22,131) Quarter 2 (n=22,237) Quarter 3 (n=20,538) Quarter 4 (n=19,090) Total for 2004 (n=83,996)
Make &
model
Cum % Make and
model
Cum % Make and
model
Cum % Make and
model
Cum % Make and
model
Cum %
1 Nokia 6230 12.8 Nokia 6230 13.9 Nokia 6230 13.9 Nokia 6230 13.3 Nokia 6230 13.5
2 Sony E K700i 16.7 Samsung D500 21.5 Samsung D500 23.1 Samsung D500 22.1 Samsung D500 20.6
3 Nokia 3310 20.1 Sony E K700i 25.3 Motorola Razr 28.5 Motorola Razr 29.9 Motorola Razr 25.4
4 Sony E T610 22.7 Motorola Razr 28.7 Sony E K700i 31.8 Sony E K750i 33.5 Sony E K700i 28.8
5 Nokia 6610 25.3 Nokia 3310 31.2 Nokia 6630 34.7 Nokia 6680 36.7 Nokia 3310 31.2
6 Samsung E700 27.8 Nokia 6610 33.5 Nokia 6680 37.1 Sony E K700i 39.6 Nokia 6610 33.3
7 Samsung D500 30.3 Nokia 7610 35.5 Nokia 3310 39.4 Nokia 6630 42.2 Nokia 6630 35.3
8 LG U8120 32.6 Sony E T610 37.5 Sony E K750i 41.4 Sony E W800i 43.9 Sony E T610 37.0
9 Nokia 7610 34.9 Nokia 6630 39.4 Nokia 6610 43.4 Samsung D600 45.5 Nokia 7610 38.7
10 Nokia 7250 37.1 Nokia 6310 41.2 Nokia 7610 45.0 Nokia 3310 47.2 Nokia 6310 40.3
11 Nokia 6100 39.1 Nokia 6100 42.8 Nokia 6310 46.6 Nokia 6610 48.6 Nokia 6100 41.8
12 Nokia 7600 41.0 LG U8120 44.4 Nokia 6100 48.0 Samsung E720 49.9 Nokia 6680 43.4
13 Motorola Razr 42.9 Samsung E700 46.0 Sony E T610 49.3 Nokia 6310 51.1 Sony E K750i 44.8
14 Nokia 6310 44.8 Nokia 7250 47.5 Nokia 6210 50.6 Nokia 6100 52.3 Samsung E700 46.2
15 Nokia 6600 46.6 Nokia 7600 48.9 Samsung E720 51.9 Nokia 8800 53.4 LG U8120 47.6
16 Sony E T630 48.1 Nokia 6600 50.2 Nokia 7250 53.0 Sony E K608i 54.4 Nokia 7250 48.9
17 Samsung E800 49.5 Nokia 6210 51.6 Samsung E700 54.1 Nokia 6210 55.3 Nokia 6210 50.1
18 Nokia 8310 51.0 Nokia 8310 52.8 Nokia 8310 55.1 Nokia 7610 56.2 Nokia 7600 51.3
19 Nokia 7210 52.4 Sony E T630 54.0 Nokia 7600 56.0 Nokia 1100 57.1 Nokia 6600 52.5
20 Nokia 3510 53.7 Samsung E800 55.1 Sony E T630 57.0 Sony E T610 58.0 Nokia 8310 53.6
Other
a
Others=46.3% 100.0 Others=44.9% 100.0 Others=43.0% 100.0 Others=42.0% 100.0 Others=46.3% 100.0
a
Number of ranks of other models varied by Quarter: Q5=317, Q6=332, Q7=344, Q8=347.

Jen Mailley et al.
Phone Theft Index
223
Technological convergence means a mobile phone can now include an MP3 player, a Global
Positioning System, Internet and email access, a digital camera, TV and video options, various
games, word processing and spreadsheet software, Bluetooth interaction with other electronic
goods and computers, plus various other PDA (Personal Digital Assistant) options. The result
is that new and better handsets are attractive targets for theft as soon as they enter the legal
market.
The result of the dynamic market for mobile phones is that particular models come
and go relatively quickly from the sale and resale (including theft) markets. While many,
perhaps most, handset models barely register a signifi cant presence in the index
presented here, others have different patterns of theft. These are presented here as
theft “ careers ” . The terminology is that of crime-related career research elsewhere in-
cluding offender criminal careers (see e.g. Blumstein et al ., 1986a, b ), the criminal
careers of places ( Sherman, 1994 ) and victim careers ( Farrell et al ., 2001 ). In the present
context, the onset of a career is when a product is fi rst stolen. The theft careers of mobile
phones have varying trajectories. Theft careers vary in frequency and duration and can
enter into decline at varying rates. The remainder of this section illustrates some of the
diversity of theft career trajectories shown primarily through a sample of models that are
more prominent in the theft rankings. The graphic representation of theft careers for fi ve
prominent mobile phone models is shown in Figure 3 . Figure 3 portrays monthly crimes
rather than the quarterly counts that are used in Tables 5 and 6 .
We thank an anonymous reviewer for noting that, while the terminology of criminal ca-
reer research is used here, the lesser known terminology of product lifecycles would also
apply (see e.g. Pease, 1997 ; Clarke, 1999, pp. 31 – 32 ).
In the fi rst quarter of 2004, the Nokia 7250 was the most stolen handset. It accounted for
9.1 per cent of thefts that quarter ( Table 5 ). In Figure 3 , the Nokia 7250 was at the peak of
its career to the lefthand side of the chart in early 2004, and experienced a steady decline in
theft frequency thereafter. Its decline was counterbalanced by the rise in thefts of other mod-
els, suggesting that some models may be successors to others in their theft careers, just as
some models succeed others in the legal sales market. The other theft career that is gener-
ally in decline in Figure 3 is that of the Sony Ericsson T610. This model was never stolen as
much as the Nokia 6230, but its lower number of thefts appear to peak slightly later in April
and May 2004 and decline steadily thereafter.
The Nokia 6230 was the most stolen mobile phone model in 2004 and 2005 combined.
This may well refl ect its sales and the volume of models in circulation, and further research
should explore this possibility as detailed below. Yet it was not always ranked fi rst in thefts.
Figure 3 shows the steep rise in theft frequency of the Nokia 6230. In the fi rst quarter of
2005, over three times as many Nokia 6230s were stolen as the second ranked handset, the
Sony Ericksson K700i. Figure 3 appears to show a peak in thefts in July 2005 and the begin-
ning of a gradual decline in thefts after that date. Whether that decline continued steadily or
gathered pace in 2006 is a question to answer with more recent data.
At the time of writing in mid-2006, the Motorola Razr, in its various forms, was arguably
one of the U.K. ’ s most iconic handsets. Its theft career shows a dramatic rise. Hardly any
Razrs were recorded stolen in the bulk of 2004. In the fourth quarter of 2004, over a hundred
Razrs were reported stolen and it was the 36th most stolen handset. The number stolen more
than trebled by the fi rst quarter of 2005 when it ranked 13th, accounting for close to two per

Phone Theft Index
224
cent of stolen handsets. The “ Rise of the Razr ” is shown in Figure 3 . By the end of 2005, the
Razr overtook the Samsung D500 as the second most-stolen phone and accounted for near-
ly eight per cent of handset thefts in the last quarter. If its theft career maintained its upward
trajectory and that of the Nokia 6230 continued its decline, we would anticipate that the
Razr overtook the 6230 in absolute number of thefts in the fi rst-half of 2006.
The Samsung D500 shows a somewhat different theft career trajectory to the other mod-
els in Figure 3 . Thefts of the D500 rose enormously in the fi rst part of 2005 so that it
quickly became the second most-stolen handset model. Yet the career trajectory is somewhat
more truncated than the Nokia 6230. Thefts of the Samsung D500 remained relatively stable
for the 6 months of May to October 2005, but declined thereafter, appearing to be in fairly
steep decline in December 2005.
Next steps
The Car Theft Index was the intellectual precursor of the mobile Phone Theft Index. As with
that index, the fi rst effort is not anticipated as the wholly completed product. The authors
would view the current work as successful if it stimulates further development of the index.
There are two key steps to the production of a useful mobile Phone Theft Index:
1. A risk-based index would be preferable to the counts used here. The traditional meth-
od, as used for car theft, would be to compare thefts to the number of phones in cir-
culation. However, we hypothesize that for phones, a more incisive indicator of risk
could be developed by comparing thefts during crimes that involve target selection
(particularly snatch-theft where only a mobile is taken) to those that do not (such as
burglary or car theft). A preliminary analysis along the lines of this second indicator
suggests that decisions are made to steal some models, such as the Motorola Razr, far
more than others, and this fi nding is supported by our interviews with mobile phone
thieves.
0
100
200
300
400
500
600
700
800
900
1000
Number Stolen
Nokia 6230
Samsung D500
Motorola
Razr
Sony E. T610
Nokia 7250
Jan-
04
Feb-
04
Mar-
04
Apr-
04
May-
04
Jun-
04
Jul-
04
Aug-
04
Sep-
04
Oct-
04
Nov-
04
Dec-
04
Jan-
05
Feb-
05
Mar-
05
Apr-
05
May-
05
Jun-
05
Jul-
05
Aug-
05
Sep-
05
Oct-
05
Nov-
05
Dec-
05
Date
Figure 3 . Mobile phone theft “ careers ” .

Jen Mailley et al.
Phone Theft Index
225
2. The key to producing a useful index with a chance of stimulating anti-theft action
by the mobile phone industry, is that its production should be timely and frequent. The
index would preferably be produced and publicized by the Home Offi ce or the National
Mobile Phone Crime Unit, or perhaps by Ofcom or the Trading Standards Institute.
Production on a quarterly, or perhaps biannual, basis would allow it to track the mar-
ket better than an annual index. To speed up the production process, there needs to be
analysis to determine the optimal level of data cleaning required, because this is the
main task involved. It may be that the rankings required to classify handsets as either
high, medium or low risk (the categories used in the Car Theft Index) could be pro-
duced from the original police records. This would eliminate the need for the exten-
sive data cleaning that was undertaken for the present study. An appropriate next step
for research would be to determine the correlation between the theft rankings using
cleaned and un-cleaned data. The aim should be to determine the minimum amount of
data cleaning necessary to produce an accurate index. It may be that some are always
required, but the present study may have already completed the bulk of the task or at
least demonstrated that it is feasible: the bulk of data cleaning takes place on the fi rst
occasion of writing the clean-up computer code, so for each subsequent update it is a
far smaller task.
There is scope for further analysis of handset thefts with a view to informing crime preven-
tion practice. Do robbers target different models to thieves, and under what circumstances?
Which phone models are stolen during lifting (from bars and tables) and which during other
types of theft? Which models are stolen from persons of different ages? Are more pricy
phones disproportionately targeted or is it the best-sellers that can be quickly fenced? There
is a further need to explore the who, what, where, why, when and how of mobile phone
thefts linked to specifi c models. Qualitative interviews with offenders could produce useful
complementary insight into their decision-making. Anti-theft designs relating to mobile
phones have recently been reviewed elsewhere but there is a clear need for further industry
efforts to design out crime ( Whitehead et al ., 2008 ). Data-driven recommendations based on
further research would be a means to this end.
Concluding note
The motor car has been around since shortly after the internal combustion engine was
invented in 1885. Car theft emerged shortly thereafter, but it took close to a century until
the fi rst car theft indices emerged. There are reasonable grounds for believing that those
indices have stimulated anti-theft design efforts and a major decline in the car theft. The fi rst
mobile phone was used in the U.K. in 1985 by comedian Ernie Wise ( BBC News, 2005 ).
Mobile phone theft has only really been with us since the mid-1990s. Yet it is already
a headline crime topic, bucking the trends of decreases elsewhere, and showing little sign
of abating as the mobile phone continues to progress and integrate other technologies.
They say history repeats itself. If so, further iterations of a theft index that stimulate
design against crime efforts should come sooner rather than later to avert the next century of
crime.

Phone Theft Index
226
Acknowledgements
The funding from the Engineering and Physical Sciences Research Council under grant
EP / C52036X / 1 is acknowledged. We thank Gloria Laycock for her permission to reproduce
her chart of car theft and the Car Theft Index, and the anonymous reviewers of this journal.
References
Allen , J . , Forest , S . , Levi , M . , Roy , H . , Sutton , M . and Wilson , D . ( 2005 ) Fraud and Technology Crimes: Find-
ings from the 2002/03 BCS and 2003 Offending, Crime and Justice Survey . Home Offi ce Online Report 34/05.
London: Home Offi ce . Available at http://www.crimereduction.gov.uk/fraud20.htm (accessed July 2006) .
BBC ( 2002a ) Phone Theft at Gun Point . BBC News online . 23 August 2002. Available at http://www.bbc.co.uk/
threecounties/news/2002/08/23/phone_theft.shtml (accessed August 2006) .
BBC ( 2002b ) Trauma of Phone Theft Victim . BBC News online . 30 January 2002. Available at http://news.bbc.
co.uk/1/hi/uk/1791577.stm (accessed July 2006) .
BBC ( 2002c ) Under Phone Lock and Key . BBC News online . 30 January 2002. Available at http://news.bbc.
co.uk/1/hi/uk/1791216.stm (accessed August 2006) .
BBC ( 2003a ) Man Robbed with Replica Gun . BBC News online . 29 April 2003. Available at http://news.bbc.
co.uk/1/hi/england/kent/2985555.stm (accessed August 2006) .
BBC ( 2003b ) Good Samaritan Knifed by Teenager . BBC News online . 14 July 2003. Available at http://news.bbc.
co.uk/1/hi/england/bristol/3065829.stm (accessed August 2006) .
BBC ( 2005 ) Mobiles Rack Up 20 Years of Use. BBC News Online . 1 January 2005. Available at http://news.bbc.
co.uk/1/hi/technology/4138449.stm (accessed March 2006) .
BBC ( 2006a ) Knife Used in Mobile Phone Theft . BBC News online . 30 May 2006. Available at http://news.bbc.
co.uk/1/hi/england/wear/5030942.stm (accessed August 2006) .
BBC ( 2006b ) Teenager Jailed for CCTV Mugging . BBC News online . 29 June 2006. Available at http://news.bbc.
co.uk/1/hi/england/lancashire/5128634.stm (accessed August 2006) .
BBC ( 2006c ) Mobiles and MP3s Fuel Robbery Rise . BBC News online . 20 July 2006. Available at http://news.bbc.
co.uk/1/hi/uk_politics/5195910.stm (accessed August 2006) .
Brown , R . ( 2004 ) The Effectiveness of Electronic Immobilisation: Changing Patterns of Temporary and Permanent
Vehicle Theft . In Maxfi eld, M.G. and Clarke, R.V. (eds) Understanding and Preventing Car Theft . Vol. 17 of
Crime Prevention Studies. Monsey, NY: Criminal Justice Press .
Brown , R . and Clarke , R . V . ( 2004 ) Police Intelligence and Theft of vehicles for Export: Recent U.K. Experience.
In Maxfi eld, M. G. and Clarke, R. V. (eds) Understanding and Preventing Car Theft , Volume 17 of Crime
Prevention Studies , Monsey, NY: Criminal Justice Press , pp. 173 – 192 .
Blumstein , A . , Cohen , J . , Roth , J . A . and Visher , C . A . (eds) ( 1986a ) Criminal Careers and “ Career Criminals ” .
Vol. I. Washington, DC: National Academy Press .
Blumstein , A . , Cohen , J . , Roth , J . A . and Visher , C . A . (eds) ( 1986b ) Criminal Careers and “ Career Criminals ” .
Vol. II Washington, DC: National Academy Press .
Clarke , R . V . ( 1999 ) Hot Products: Understanding, Anticipating and Reducing Demand for Stolen Goods . Police
Research series paper 112. London: Home Offi ce . Available at www.homeoffi ce.gov.uk/rds/prgpdfs/fprs112.pdf
(accessed July 2006) .
Cohen , M . , Rust , R . T . , Steen , S . and Tidd , S . T . ( 2004 ) Willingness-to-Pay for Crime Control Programs . Criminology .
Vol. 42 , No. 1 , pp 89 – 109 .
Dubourg , R . and Hamed , J . ( 2005 ) Estimates of the Economic and Social Cost of Crime in England and Wales:
Costs of Crime Against Individuals and Households 2003/04. In The Economic and Social Costs of Crime against
Individuals and Households 2003/04 . Home Offi ce online report 30/05 London: Home Offi ce , pp. 3 – 15 .
Farrell , G . , Tseloni , A . , Wiersema , B . and Pease , K . ( 2001 ) Victim Careers and ‘ Career Victims?: Towards a
Research Agenda ’ . In Farrell, G. and Pease, K. (eds) Repeat Victimization . Vol. 12 of Crime Prevention Studies.
Monsey, NY: Criminal Justice Press , pp 241 – 254 .
Hazelbaker , K . ( 1997 ) Insurance Industry Analyses and the Prevention of Motor Vehicle Theft . In Felson, M. and
Clarke, R.V. (eds) Business and Crime Prevention . Monsey, NY: Willow Tree Press .

Jen Mailley et al.
Phone Theft Index
227
Harrington , V . and Mayhew , P . ( 2001 ) Mobile Phone Theft . Home Offi ce Research Study 235. London: Home
Offi ce Development and Statistics Directorate .
Hoare , J . and Cotton , J . (2006) Property Crime. In Walker, A., Kershaw, C., and Nicholas, S. (eds) Crime in Eng-
land and Wales 2005/06. Home Offi ce Statistical Bulletin 12/06 . London: Home Offi ce , pp. 85 – 108 .
Home Offi ce ( 2002 ) The Bike Theft Index 2002 . London: Home Offi ce . Available at http://www.crimereduction.
gov.uk/bti2002.pdf (accessed July 2006) .
Home Offi ce ( 2005a ) The Car Theft Index 2004 . London: Home Offi ce Communications Directorate . Available at
http://www.crimereduction.gov.uk/vehiclecrime47.htm (accessed July 2006) .
Home Offi ce ( 2005b ) The Bike Theft Index . London: Home Offi ce . Available at http://www.crimereduction.gov.
uk/vehiclecrime54.htm (accessed July 2006) .
Houghton , G . ( 1992 ) Car Theft in England and Wales: The Home Offi ce Car Theft Index . Police Research Group
Crime Prevention Unit series paper 33. London: Home Offi ce .
Jansson , K . , Coleman , K . and Kaiza , P . ( 2006 ) Violent Crime . In Walker, A., Kershaw, C. and Nicholas, S. (eds)
Crime in England and Wales 2005/06 . Home Offi ce Statistical Bulletin 12/06, London: Home Offi ce .
Johnson , S . , Bowers , K . and Hirschfi eld , A . ( 1997 ) New Insights into the Spatial and Temporal Distribution of
Repeat Victimisation . British Journal of Criminology . Vol. 37 , No. 2 , pp 224 – 241 .
Laycock , G . ( 2004 ) The UK Car Theft Index: An Example of Government Leverage . In Maxfi eld, M.G. and Clarke,
R.V. (eds) Understanding and Preventing Car Theft . Vol. 17 of Crime Prevention Studies Monsey, NY: Criminal
Justice Press , pp 25 – 44 .
Ledbury , M . , Miller , N . , Lee , A . , Fairman , T . and Clifton , C . ( 2006 ) Understanding Policy Options . Home Offi ce
Online Report 06/06. London: Home Offi ce .
Mailley , J . , Whitehead , S . and Farrell , G . ( 2006 ) Progress and Prospects in the Prevention of Mobile Phone Theft .
Justice of the Peace . Vol. 170 , pp 404 – 407 .
Mayhew , P . , Clarke , R . V . and Hough , M . ( 1992 ) Steering Column Locks and Car Theft . In Clarke, R.V. (ed.)
Situational Crime Prevention: Successful Case Studies . 1st Edition Albany, NY: Harrow and Heston , pp 52 – 65 .
Mayhew , P . , Clarke , R . V . and Hough , J . M . ( 1976 ) Crime as Opportunity . London: HMSO .
Mintel ( 2004 ) Telecommunications Retailing – May 2004 . Available at www.mintel.com (accessed July 2006) .
Mobile ( 2006 ) Time to Crack Down on Crime . Mobile , Friday 23rd June 2006. Available at http://www.mobileto-
day.co.uk/content/2996.asp?men=5 & sub=9 (accessed August 2006) .
Ofcom ( 2004 ) The Communications Market 2004 – Telecommunications Appendices . August 2004, Offi ce of
Communications. Available at http://www.ofcom.org.uk/research/cm/cmpdf/telecom_apndx.pdf (accessed
August 2006) .
Ofcom ( 2006 ) The Communications Market 2006, 3 Telecommunications . August 2006, Offi ce of Communica-
tions. Available at http://www.ofcom.org.uk/research/cm/cm06/telec.pdf (accessed December 2006) .
Pease , K . ( 1997 ) The Role of Routine Activities and Rational Choice Theory in Predicting the Future . In Newman,
G. Clarke, R.V. and Shoham, S. (eds) Situational Crime Prevention and Rational Choice . Dartmouth Press:
Aldershot, UK , pp 231 – 247 .
Pease , K . ( 1998 ) Crime, Labour, and the Wisdom of Solomon .
Policy Studies . Vol. 19 , No. 3/4 , pp 255 – 265 .
Sherman , L . W . ( 1994 ) Hot Spots of Crime and Criminal Careers of Places . In Eck, J. and Weisburd, D. (eds) Crime
and Place . Vol. 4 of Crime Prevention Studies. Monsey, NY: Criminal Justice Press , pp 35 – 52 .
Sherman , L . W . , Gartin , P . R . and Buerger , M . E . ( 1989 ) Hot Spots of Predatory Crime: Routine Activities and the
Criminology of Place . Criminology . Vol. 27 , No. 1 , pp 27 – 55 .
Southall , D . and Ekblom , P . ( 1985 ) Designing for Car Security: Towards a Crime Free Car . Crime Prevention Unit
paper 4. London: Home Offi ce .
Webb , B . and Laycock , G . ( 1992 ) Tackling Car Crime: The Nature and Extent of the Problem . Crime Prevention
Unit paper 32. London: Home Offi ce .
Webb , B . , Brown , B . and Bennett , K . ( 1992 ) Preventing Car Crime in Car Parks . Crime Prevention Unit paper 32.
London: Home Offi ce .
Webb , B . , Smith , M . and Laycock , G . ( 2004 ) Designing-Out Crime Through Vehicle Licensing and Registration
Systems . In Maxfi eld, M.G. and Clarke, R.V. (eds) Understanding and Preventing Car Theft . Vol. 17 of Crime
Prevention Studies. Monsey, NY: Criminal Justice Press , pp 67 – 84 .
Whitehead , S . , Mailley , J . , Storer , I . , McCardle , J . and Farrell , G . ( 2008 ) In Safe Hands: A Review of Mobile Phone
Anti-Theft Designs . European Journal on Criminal Policy and Research . Vol. 14 , No. 1 , pp 39 – 60 .