Estimating the true scale of shop theft in Nottingham

Estimating the true scale of
shop theft in Core Cities in England
Dr. James Hunter, Laura Garius,
Professor Andromachi Tseloni and Dr. Paul Hamilton
Follow us on Twitter @ShopTheftRisk
The Research Problem:
• Shop theft increased in England and Wales
by 6% between 2012-13 (ONS, 2014) unlike the fall in many other offence
categories since 2003;
• Cost of shop theft to Core City A economy
is estimated at £24million per annum;
• Between 2009 and 2014 there were 18,725
officially recorded incidents of shop theft in
Core City A;
• Community safety partnership estimates
that this only accounts for around 12% of
all shop theft occurrences in the City
(based upon CVS).
Project aims:
• Estimate true scale of shop theft in Core
Cities across England;
• Develop neighbourhood level risk profiles
designed to inform retailers and crime
reduction agencies;
• Neighbourhood profiles informed by (a) true
scale of shop theft; (b) offender profiles; (c)
effectiveness of security measures; and (d)
neighbourhood context;
• Knowledge generated by project
disseminated to retailers and crime
reduction agencies via social enterprise
company.
Estimating true scale of shop theft: two new approaches
• Approach One: Evaluation of shop theft trends
using Commercial Victimisation Survey (CVS)
across different retail sectors;
• Identification of hidden level of shop theft
through application of CVS trends onto different
retail sectors at local authority level;
• Approach Two: Modelling of spatial variations in
official recorded shop theft incidence at LSOA
level;
• Identification of significant socio-demographic,
economic and neighbourhood factors;
• Use predicted levels of shop theft arising from
multivariate model.
Commercial Victimisation Survey [CVS]
CVS
• Home Office Victim
Survey;
• Crimes Against
Businesses;
• Experiences in
previous12 months;
• Four Sectors;
• G] Retail/ Wholesale;
• Telephone Survey;
• Representative sample
4,000 businesses;
• Sweeps: 1994, 2002,
2012, 2013.
Strengths
• Evidence gap (National
Statistician Review,
2011);
• ‘Dark figure of
unrecorded crime’;
• Investigates police
reporting behaviour;
• Explores nature of
business crime.
Issues
• Premises name;
• Limited internal
information;
• Premises residency <12
months;
• Multiple response
vertical retail sectors;
• ‘Most recent
incident…’;
• SIC definition of ‘retail’.
Approach One Methodology
Nomis UK Business Counts [UBC] Data: Counts businesses
in each vertical retail sector for every core city
Scale down the Rates per 1,000 sector businesses
(nationwide) to the number of sector businesses present in
each of the core cities in 2013.
Drop the CVS estimates (by vertical sector) onto the
existing Police [PRC] Data for the same time period (2013)
City-specific comparisons between the number
of recorded shop thefts occurring in a city and
the number of shop thefts predicted by the
CVS to have actually occurred
An overall percentage of police reporting
for each core city & unique reporting
patterns for the different vertical retail
sectors
TRUE ESTIMATE OF SHOP THEFT IN CORE CITY A (2013) BY VERTICAL RETAIL
SECTOR
Police Recorded Crime [PRC] Estimate
Commerical Victimisation Survey [CVS] Estimate
18000
NUMBER OF SHOP THEFTS
16000
14000
12000
10000
8000
6000
4000
2000
0
7%
VERTICAL RETAIL SECTOR
TRUE ESTIMATE OF SHOP THEFT IN CORE CITY B (2013) BY
VERTICAL RETAIL SECTOR
Police Recored Crime [PRC] Estimate
Commercial Victimisation Survey [CVS] Estimate
18000
NUMBER OF SHOP THEFTS
16000
14000
12000
10000
8000
6000
4000
2000
0
6%
VERTICAL RETAIL SECTOR
TRUE ESTIMATE OF SHOP THEFT IN CORE CITY C (2013) BY VERTICAL RE TAIL
SECTOR
Police Recored Crime [PRC] Estimate
Commercial Victimisation Survey [CVS] Estimate
NUMBER OF SHOP THEFTS
35000
30000
25000
20000
15000
10000
5000
0
6%
VERTICAL RETAIL SECTOR
Approach Two: Explaining spatial variations in shop theft at
the LSOA level
Categories of shop theft
offenders:
1. teenagers;
2. homeless;
3. hard-up households;
4. offenders with alcohol and
substance abuse problems;
5. ‘cry for help’ individuals;
6. ‘bored’ affluent individuals.
Retail profile of neighbourhood:
• Number of retail outlets;
• Type of retail outlets;
• Type of goods on sale.
Level and type of security
Function of neighbourhood:
• Central business district;
• Mixed use neighbourhood;
• Residential;
• Transport hub.
Location of retail outlets:
• Shopping centres;
• High street;
• Out of town shopping centres;
• Street corners.
Response of crime reduction
agencies
Who is living – and who is shopping in the neighbourhood
LEVEL OF SHOP THEFT
Modelling spatial variations in shop theft at the LSOA level
Empirical evidence: Three types of shopping areas and
shop theft neighbourhoods in Core City A 2009-2014
Central business district:
Local high streets:
Shopping precincts and
corner shops:
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Supermarkets (30.7%);
Pharmacists (11.8%);
IT/computing (5.1%);
Clothing (4.4%);
Furniture (4.2%);
Department stores (3.8%).
Supermarkets (77.9%);
Food and alcohol (5.8%)
Pharmacists (4%);
Department stores (1.6%);
Furniture (1%)
Food only (0.6%);
Off License (0.6%);
Petrol stations (6.3%).
Supermarkets (88%);
Petrol stations (7%)
Newsagents (4%);
Food and alcohol (2%).
What explains spatial variations in shop theft in Core City A at the
LSOA level? Correlation coefficients (n=182) (* significant at .05)
Positive relationship:
Negative relationship:
• Change in daytime population (.63***);
• Primary Age (-.20**), Retired (-.19*);
• Drug offences (.64***)
• Deprivation (.15*);
• Owner occupation (-.21**);
• Lack of employment (.16*)
• Semi-detached (-.23**);
• Barriers to services and housing (.27**);
• Job seekers allowance (-.17*), income
support (-.35***);
• Aged 20-30 (.27**), Single (.23**);
• Lone parents not in employment (.25**);
• Students (.18*);
• Private renting (.20**);
• Mixed White Asian (.24**), Chinese (.38***);
• EU Original Expansion Countries (.29***);
• Arrived in UK after 2010 (.35***);
• Working in hotels (.19**), finance (.23**) or IT (.28***);
• Presence of retail outlets (.48***);
• Size of retail outlets (.35***);
• Bakers (.43***), Food (.35***), Mobile phones (.47***), Newsagents (.41***),
Clothing (.40***), Cosmetics (..42***).
• Type of neighbourhood (-.27***).
Multivariate modelling of spatial variations in shop theft
rate in Core City A: empirical results (n=182) (* significant at .05)
• Explanatory power of model (Adjusted R2)=0.56 ***
Independent variable:
Drug offences per 1000
population
Difference between
daytime and resident
population
Retail units per 1000
population
Standardised
coefficient
.45
Significance
Tolerance
VIF
.000***
.766
1.306
.37
.000***
.541
1.849
.16
.011*
.589
1.697
Police recorded and new estimates of level of shop
theft 2009-2014
Core City:
A
B
C
Police Recorded
Estimate
Approach One: Approach Two:
CVS Trends
Modelling LSOA
estimate
spatial variations
18,725
287,634
34,766
17,867
297,783
31,367
27,326
445,775
50,416
Future developments of our approach:
• Work in progress – project only been
operational for six months;
• Inclusion of data from other Core Cities;
• Improve retail data at neighbourhood
level – switch from NOMIS estimates to
ONS Points of interest data;
• Develop indicators of risk by retail type of
move from simple modelling of number of
retail units to take account of type of
retail presence within different
neighbourhoods;
• Thank you for listening!