Socio‐Economic Determinants of Crime Rates: Modelling Local Area Police‐Recorded Crime
| Published date | 01 May 2016 |
| Date | 01 May 2016 |
| DOI | http://doi.org/10.1111/hojo.12157 |
The Howard Journal Vol55 No 1–2. May 2016 DOI: 10.1111/hojo.12157
ISSN 2059-1098, pp. 207–225
Socio-Economic Determinants of
Crime Rates: Modelling Local Area
Police-Recorded Crime
ROGER TARLING and RICHARD DENNIS
Roger Tarling is Professor of Social Research and Criminology, University of
Surrey; Richard Dennis is Statistical Analyst, 1st Central Insurance
Abstract: Social disorganisation theory and economic theory posit relationships between
crime and socio-economic conditions of an area. This article examines those relationships
by analysing, for the first time, data for 322 local authorities in England. Significant
determinants of property crime rates are found to be deprivation, the percentage of young
people in the population, density of population/urbanisation, and movement in and out of
the area. The variables found to be associated with violent crime are deprivation, density
of population, population movements, and ethnic heterogeneity. Our resultsconfirm much
previous research but add fresh insights, especially regarding demographic distributions
and population movements.
Keywords: police-recorded crime; area correlates of crime; spatial distribu-
tion of crime; social disorganisation; economic theory; rational choice theory;
multilevel modelling
Discourse regarding the spatial distribution of crime and its associated
patterning with other socio-economic variables has a long history and
the subject of much empirical research stemming, in the main, from two
theoretical perspectives: social disorganisation theory and economic the-
ory/rational choice theory of crime causation. In this article we draw on
previous research to identify correlates of crime rates and extend that re-
search evidence by creating some new variables and by analysing, for the
first time, local authority level data.
In the first half of the last century the ‘Chicago School’ analysed crime
in different areas of Chicago and found that crime rates were inversely re-
lated to economic disadvantage. The most economically deprived areas (as
defined by a concentration of poverty) and areas of social disorganisation
(as characterised by densely populated urban inner-city neighbourhoods,
diverse ethnic and racial composition, and movement in and out of the
area) experienced the highest levels of crime (Shaw and McKay 1942).
207
C
2016 The Howard League and John Wiley & Sons Ltd
Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK
The Howard Journal Vol55 No 1–2. May 2016
ISSN 2059-1098, pp. 207–225
A British study drawing on this theoretical approach was undertaken by
Baldwin, Bottoms and Walker (1976) and more recently by Sampson and
Groves (1989).
Economists have also been prominent in this field of research in the
latter half of the 20th Century. Impetus came from Becker (1968) and
subsequent development from Ehrlich (1973) who formulated what has
become known as the Becker-Ehrlich economic model of crime.1Their
approach:
follows the economists’ usual analysis of choice and assumes that a person commits
an offense if the expected utility to him exceeds the utility he could get by using
his time and other resources at other activities. Some persons become ‘criminals’,
therefore, not because their basic motivation differs from that of other persons, but
because their benefits and costs differ. (Becker 1968, p.176)
Thus, individuals are rational and engage in either legal or illegal be-
haviour according to the returns from each activity. Rational choice the-
ory, which came later, is based on exactly the same underlying premises
regarding offender motivation and behaviour and can be considered as a
restatement of the Becker-Ehrlich economic theory.
Although formulated as a micro theory of individual behaviour, the
Becker-Ehrlich economic model has been tested mainly at the macro level
on aggregate data by comparing crime rates and economic indicators, in
particular income, unemployment and wealth, at the area level.2In the
UK, the unit of analysis has invariably been the 43 police force areas of
England and Wales (Carr-Hill and Stern 1979; Willis 1983; Timbrell 1990;
Reilly and Witt 1996; Carmichael and Ward 2000; Machin and Meghir
2004; Tarling 2009; Bandyopadhyay 2012; Han, Bandyopadhyay and
Bhattacharya 2013). Three studies compared data for the standard re-
gions of England plus Wales (Witt, Clarke and Fielding 1999; Carmichael
and Ward 2001; Wu and Wu 2012). There have also been two studies
comparing the six regions of Scotland (Reilly and Witt 1992; Pyle and
Deadman 1994). In the USA, the unit of analysis has often been the 50
states (Pratt and Cullen 2005; Cook and Winfield 2013) and in Germany,
the eleven West German Laender (Entorf and Spengler 2000).
A limitation of all these economic studies is the level of aggregation
(police force area, state, l¨
ander) and several authors lament the inability to
disaggregate the data further,which would permit more precise inferences
about the socio-economic and spatial dynamics of crime (see, for example,
Carr-Hill and Stern 1979; Witt, Clarke and Fielding 1999). Also, Willis
(1983) makes the point that many, if not most, police force areas, com-
prise a mix of heterogeneous smaller urban and rural areas. One study
did, however, analyse data for small geographical areas: a recent study by
Hooghe et al. (2011) compared data for 589 municipalities of Belgium.
Furthermore, all studies analysed police-recorded crime data, so any in-
ferences drawn are only valid to the extent that recorded crime represents
the level of crime in an area – an issue we return to in the ‘Methodology’
section.
208
C
2016 The Howard League and John Wiley & Sons Ltd
To continue reading
Request your trial