Content uploaded by Tuba Bircan
Author content
All content in this area was uploaded by Tuba Bircan on Feb 01, 2021
Content may be subject to copyright.
Article
Corresponding author:
Tuba Bircan, Centre for Political Research, Parkstraat 45 - bus 3602, 3000 Leuven, Belgium.
Email: Tuba.Bircan@soc.kuleuven.be
European Journal of Criminology
8(3) 198–212
© The Author(s) 2011
Reprints and permission: sagepub.
co.uk/journalsPermissions.nav
DOI: 10.1177/1477370811403443
euc.sagepub.com
Immigration, diversity
and crime: an analysis of
Belgian national crime
statistics, 2001–6
Tuba Bircan
Centre for Political Research, Belgium
Marc Hooghe
Centre for Political Research, Belgium
Abstract
In this article we investigate the assumed relationship between immigrant concentrations and
crime at the community level in Belgium for the period 2001–6. Multivariate analysis shows that
unemployment levels have a very strong effect on crime, thus rendering the ethnic composition
of the community non-significant. We distinguished various groups depending on ethnic origins.
Controls include the concentration of migrants and the number of incoming immigrants. Our
conclusion is that deprivation (and especially unemployment) is more important than direct
immigration effects in explaining community-level crime rates. We end with some theoretical and
policy considerations regarding the relation between ethnic diversity and crime.
Keywords
crime, ethnic diversity, ethnic groups, property crime, violent crime
Introduction
In many advanced industrial democracies, there is considerable concern about the costs
of immigration and the effect of immigration on crime. Some fear that a concentration of
ethnic minorities at the neighbourhood or community level will create higher levels of
crime. The study of neighbourhoods and crime has a long history in social science.
Stigler (1986) noted that social statisticians of the 18th and 19th centuries, particularly
Bircan and Hooghe 199
Quételet and Lexis, worked intensively on both criminological and demographic data.
For over a century, beginning with the influential Chicago School of social ecology,
generations of sociologists have investigated the link between population dynamics and
deviant behaviour (e.g. Short, 1971) and they have argued that neighbourhood transition
and heterogeneity contribute to crime levels in specific geographical areas. The notion
that immigration and crime are linked, therefore, is not a new one.
In order to explain the relationship between crime and community structure, crimi-
nologists often rely on social disorganization theory. According to this perspective, the
origins of crime are located not in the characteristics of individuals, but rather in the
social structural characteristics of neighbourhoods (Shaw and McKay, [1942] 1969). In
the wake of the immigration of Southern European and Mediterranean migrant popula-
tions into Central and Northern Europe, the ‘immigration and crime’ theme appeared in
Europe towards the end of the 1960s. From the mid-1960s on, a number of authors have
explored whether crime patterns of newcomers are associated with signs of cultural con-
flict and social disintegration. Criminal behaviour by Italians, Turks, Spaniards and
Yugoslavians has become the focus of criminological research (Ferracuti, 1968).
The alleged relationship between immigration and crime has been largely taken for
granted in right-wing and conservative media and policy circles (Butcher and Piehl,
1998; Hagan and Palloni, 1999; Martinez and Lee, 2000; Mears, 2001). Any association
between immigration and crime rates also has clear political consequences. Some politi-
cal rhetoric, in either implicit or explicit form, claims that crime will be reduced by
restricting levels of immigration. It is sometimes claimed that ethnic minorities threaten
safety levels. These claims are regularly made by right-wing politicians and extremists
(Walter and Kubink, 1993).
Although some right-wing political parties assume a link between ethnic heterogene-
ity and crime rates, a growing body of empirical research raises questions about the
causal assumption underlying this claim (Lee et al., 2001; Lee, 2003; Martinez, 2002;
Martinez and Lee, 2000). The main contention in this literature is that there is no signifi-
cant relationship between the presence of non-nationals and crime rates, once adequate
controls for deprivation, unemployment and other indicators of social and economic
exclusion are included in the analysis.
The relationship between ethnic diversity and crime occupies a substantial place in
both academic and policy research agendas. To date, however, most data regarding the
occurrence of crime have been collected in specific, mostly metropolitan, areas in the
United States. The evidence available from countries outside the US remains limited.
Moreover, the available research has mainly focused on urban crime and has often
excluded rural crime. Although crime rates are higher in urban than in rural areas
(Maguire and Pastore, 1995), it is clear that including only urban data is not sufficient if
one wants to arrive at a comprehensive explanation of crime in a country.
In this article we discuss new evidence based on national crime statistics in Belgium.
Since 2001, Belgian national police have collected these data in a uniform manner, and
they can be related to information about community characteristics. The statistics allow
us to test both economic and social theories on a national level, including rural and urban
areas. More specifically we investigate whether the concentration of immigrant groups
has a significant impact on crime rates. We also pay special attention to the effects of
200 European Journal of Criminology 8(3)
different ethnic groups in order to determine whether immigration has differential effects
on various types of crime.
The existing literature on the relationship between immigration and crime at the indi-
vidual level makes different claims about the causal relationship. Some scholars argue that
immigrants are more often involved in criminal acts than the native-born (McDonald,
1997; Olson et al., 2009; Tonry, 1997). Others argue the opposite (Butcher and Piehl,
2008; Sampson et al., 2005). However, research on how immigration affects rates of crime
at the macro level is largely absent (for exceptions see Hagan and Palloni, 1999; Lee et al.,
2001; Martinez, 2002). Immigrants enter a pre-existing socioeconomic social structure,
and therefore a macro-level analysis of the relationship between immigration and crime is
necessary. Research on the relationship between immigration and crime must take into
account the ecological impact of immigration, which may affect rates of criminal behav-
iour in that region. Moreover, Manfrass (1991) states that the concepts of ‘foreigner’ or
‘ethnic minority’ include a range of minorities that differ with respect to religion, race,
language, citizenship, material circumstances, cultural background, migration motives,
and histories of relations between the home country and the immigration country.
Similarly, Stowell and Martinez (2007) have demonstrated that the impact of immigration
on crime varies by ethnic group. Most criminological research on immigrant populations
and ethnic minorities has overlooked substantial cultural differences between and within
ethnic minorities (Albrecht, 1997). It could be argued, therefore, that an advancement of
our ‘understanding of the specific impacts of immigration on crime’ requires scholars to
consider the relationship of ‘each nationality separately’ (Lee, 2003: 132).
Moreover, the type of crime needs to be considered. Hagan and Palloni (1999) state that
immigration is more likely associated with instrumental crimes (i.e. property-based offences)
than with expressive crimes (i.e. violent offences not involving property). Therefore, if only
one type of crime is used to test the theory, relations may go undetected (Stowell and
Martinez, 2007). In this article we extend the previous research in two ways. First, we
explore the role of ethnic and immigrant geographical concentration on two types of crime:
property and violent crime. Second, various immigrant populations will be considered to
analyse whether they contribute similarly or differentially to community-level crime rates.
Diversity and crime in the context of social disorganization
theory
Exploring spatial dispersion patterns of crime is a continuing concern within the field of
criminology. Already in the first decades of the 20th century sociological theories
claimed a positive relationship between immigration and crime (Sellin, 1938; Shaw and
McKay, [1942] 1969). Contemporary criminological theories build on this assumption,
by suggesting that there is a causal relation between immigration and crime (Bankston,
1998; Chilton et al., 1995; Reid et al., 2005; Short, 1997). However, the results of macro-
level studies on the relationship between immigration and crime are mixed. Various
neighbourhood-level studies in areas with considerable immigrant populations have
shown that the concentration of recently arrived immigrants has no effect on, or actually
lessens the risk of, victimization (Butcher and Piehl, 2008; Lee et al., 2001, Sampson,
2008). Furthermore, Weijters et al. (2009) state that the percentage of ethnic minorities
Bircan and Hooghe 201
living in specific geographical contexts does not have a significant effect on delinquency.
In a recent study, Sampson (2008) claims that the relationship between immigration and
crime is largely a negative one.
Social disorganization theory, especially the earliest variant of the theory (e.g. Shaw and
McKay, [1942] 1969), suggests that large numbers of immigrants, with diverse cultural
backgrounds and limited economic resources, could weaken community cohesion
(Bankston, 1998). More recent variants of the theory provide a theoretical linkage between
features of populations and crime in contemporary macro-level research (Parker et al.,
1999; Sampson and Groves, 1989). In particular, it has been argued that economic depriva-
tion and inequality (Blau and Blau, 1982; Messner, 1982; O’Brien, 1983; Sampson, 2006)
are positively correlated with crime rates. Moreover, the social disorganization approach
has stressed the relevance of population heterogeneity, which impedes communication and
cooperation among distinct cultural groups (Bursik and Grasmick, 1993; Elliott et al., 1996;
Sampson and Groves, 1989; Shaw and McKay, [1942] 1969). Sampson (2006) argues that
vibrant communities with a high level of social cohesion have a preventive effect on crime.
Social disorganization theory predicts that increased rates of crime and other forms of
deviant conduct become enduring characteristics of some urban communities, especially in
poor, ethnically mixed and residentially unstable areas (Sampson and Raudenbush, 1999;
Sampson et al., 1997). Shaw and McKay ([1942] 1969) argued that the structural charac-
teristics of urban neighbourhoods make them particularly conducive to the occurrence of
delinquent and criminal acts. Indirectly, this theory leads to the assumption of a positive
relationship between immigration and crime. Furthermore, the literature suggests that
immigration, residential instability and economic disadvantage may be regarded as critical
factors that are responsible for the breakdown of informal social control. In addition, immi-
gration might also lead to concomitant increases in crime rates, because it undermines a
community’s ability to enforce effective social control over its residents (Bankston, 1998;
Lee et al., 2001; Lee and Martinez, 2002; Mears, 2002; Reid et al., 2005). One possible
explanation for the positive relationship between ethnic diversity and crime rates may be
that immigrants create a pool of potential victims with fewer resources to protect them-
selves and their property from crime (Butcher and Piehl, 1998). Reid et al. (2005) conclude
that immigration may increase the vulnerability of entire neighbourhoods by weakening
community ties that could protect neighbourhoods from criminal behaviour.
As noted, the heterogeneity of the community, poverty, a lack of resources and various
indicators for social disorganization have all been invoked to explain high criminality.
Social disorganization theory has a particular relevance for this study because it identifies
immigration as a neighbourhood characteristic associated with crime. In short, since ethnic
heterogeneity could have a negative effect on the social structure of a neighbourhood,
disorganization theory presupposes a positive relationship between immigration and crime.
Hypotheses
This overview of the literature allows us to develop a number of clear hypotheses. More
specifically, we will address the concern that a concentration of immigrants has a posi-
tive impact on crime levels. To test our hypotheses, we rely on new nation-wide crime
figures for Belgium. Since 2001 a uniform crime-recording protocol has been used by
202 European Journal of Criminology 8(3)
the Belgian police force, and this leads to a reliable measurement of registered criminal
acts in the entire country. Belgium is divided into 589 municipalities, with on average
17,900 inhabitants. The municipality can be considered an important political and admin-
istrative unit. Therefore we can make the claim that the average scale of a municipality
still allows for a feeling of ‘community’ among the inhabitants of that municipality.
The principal hypotheses tested in this article are:
H1. Crime rates will be higher in communities with a high concentration of non-Belgians.
H2. Different ethnic groups will have a different impact on crime rates in their community.
We also assume that this relationship will be valid for various forms of crime. We are
able to include a sufficient number of control variables in order to arrive at a fully speci-
fied model. Most studies on the spatial diffusion of crime focus on only metropolitan or
only urban areas. Since our data were collected across the entire territory of a country, we
include data from both rural and urban areas. Rural crime is generally under-studied in
criminology, but it is safe to assume that crime rates will be dramatically lower in rural
areas than in the urban regions of Belgium (Wells and Weisheit, 2004). Various control
variables therefore have to be included to be able to accommodate these differences.
Data and methods
Earlier studies investigating the effect of structural social features on crime rates have
used neighbourhoods as units of analysis (Krivo and Peterson, 1996; Logan and Stults,
1999; Miethe et al., 1991; Peterson et al., 2000; Warner, 2003). It is only recently that
uniform crime data have become available for the entire territory of Belgium. Before
2001, these data were assembled on a local or regional level, and were not always com-
parable (Lippens, 1995; Pattyn, 1994; Pauwels, 2007). New procedures were imple-
mented in 2001, and although self-evidently the ‘dark figures’ of unreported or
unregistered crime acts remain a problem, we can be confident that these data allow for
a reliable overview of registered crime in Belgium. Although we have to be extremely
cautious in comparing crime rates across countries, there is no indication that Belgium
would differ from other countries in Western Europe (Newman, 1999).
In line with the literature, we distinguish between violent crime and property crime
(Byrne, 1986; Wikström, 1991). In the Belgian police records, and consistent with the
Belgian criminal code, violent crime refers to the acts of ‘intentional assault and battery’,
‘vandalism’ (whether aimed at cars or other material goods) and ‘destruction and damag-
ing’. Property crime refers to theft from motor vehicles, stealing motor vehicles and
burglary. Although these two crime categories are strongly correlated, both theoretically
and empirically it makes sense to distinguish them in the analysis. Table 1 and Table 2
summarize the factor analysis of recorded property and violent crimes for 2006.
It is believed that these two different kinds of crimes will have different patterns of
geographical concentration (Kawachi et al., 1999; Kposowa et al., 1995). Whereas the
official police records contain information on only the absolute number of these offences,
for our analyses we use crime rates for every municipality, which are calculated as:
Bircan and Hooghe 203
CRt = [at/ut] * 1000,
where CR is the crime rate, t is the year, a is the number of crimes recorded and u is the
population of the municipality. Data are available for the years 2001 to 2006. Since we
have only six observations, we cannot conduct a reliable trend analysis. In order to take
account of temporary variations, we pooled the data for these six observations, so that the
crime rate for every municipality is composed of the average of six different observa-
tions. For every year the crime rate has been calculated using the population figure for
that year.
Data were analysed using linear regression, with the municipality (n = 589) as the unit
of analysis. Control variables (demography, absolute and relative poverty and structural
characteristics) will be included in the models. Measurement of immigration can be
made in terms of either the inflow (incoming immigrants) or the stock (immigrants
present at a given moment). The available statistical material defines ‘stock’ of immi-
grants as the stable presence of ethnic minority groups. Comparative models were con-
structed in order to explore the differentiation between several ethnic groups in Belgium.
In addition, it has to be noted that the data allow us to conduct analyses only at the level
of the community, as no individual-level data (whether about the victims or about the
perpetrators) are available for the Belgian context. As such, our conclusion is valid only
for the community level, and it does not allow any inference with regard to individual-
level causal relations.
Table 1. Characteristics of the factor scale violent crime (for 2006)
Items Item average Factor loading Cronbach’s alpha
if deleted
Intentional assault and battery 79.97 0.987 0.975
Vandalism aimed at cars 71.12 0.959 0.920
Vandalism aimed at other material goods 25.83 0.922 0.850
Destruction and damaging 118.87 0.990 0.979
Cronbach’s alpha 0.897
Notes: N = 589. Extraction method: principal component analysis.
Source: Belgian Federal Police, operational police information.
Table 2. Characteristics of the factor scale property crime (for 2006)
Items Item average Factor loading Cronbach’s alpha
if deleted
Theft from motor vehicles 28.40 0.968 0.936
Stealing motor vehicle 146.36 0.949 0.901
Burglary 113.69 0.953 0.909
Cronbach’s alpha 0.723
Notes: N = 589. Extraction method: principal component analysis.
Source: Belgian Federal Police, operational police information.
204 European Journal of Criminology 8(3)
The data indeed show significant differences across the time period (Figure 1). The
violent crime rate proved to be very stable, with 12–13 crimes registered per 1000 inhab-
itants. Since 2002 the number of property crimes has declined sharply. There is no scien-
tific evidence available to explain this drop, but the federal police claims that a targeted
approach toward specific forms of crime might be responsible for this decrease. Although
Belgium is a small country, it is crucial for our analysis to have sufficient variance at the
local level. If we plot the average property crime rate for the period 2001–6 on a map of
the Belgian municipalities (Figure 2), we can observe strong variations. The highest rates
are recorded in the central area of the capital, Brussels, with high rates also moving north
to the port of Antwerp, south to the city of Charleroi and east to the city of Liège. The
rural areas in the north-east and the north-west of the country, on the other hand, show
typically low rates of property crime.
For violent crime rates, the pattern of spatial distribution is less obvious (Figure 3).
Again, we find the highest rates in the main cities but, in contrast to property crime, these
high figures do not seem to spread out across the central urban area of Belgium. Violent
crime occurs more often in the southern, French-speaking part of the country than in the
northern, Dutch-speaking area. The municipalities in the upper north-western part of the
country are all seaside resorts. During the summer months the original population of
these municipalities is almost doubled by the influx of tourists, and this could explain the
very high figures recorded in these municipalities.
Building the model
After this initial exploration of the data, we can now explore a multivariate model to
explain these local differences. For the dependent variable we have the property crime
rate and the violent crime rate. Since this is a continuous variable, we can use ordinary
least squares (OLS) regression models. The level of analysis is the municipality.
Regression analysis requires that the observed values of the dependent variable follow a
8
9
10
11
12
13
14
15
2001 2002 2003 2004 2005 2006
Property crime rate Violent crime rate
Figure 1. Evolution of crime rates in Belgium, 2001–6.
Note: Reported and registered crimes in Belgium per 1000 inhabitants.
Source: Belgian Federal Police, operational police information.
Bircan and Hooghe 205
normal distribution. However, the original crime rates are not normally distributed. A
logarithmic transformation of the original data was therefore used, resulting in a suffi-
ciently normal distribution to be used in the analysis (Osgood, 2000).
For the independent variables, we use multiple measures of ethnic diversity. As a first
step, the stock of immigrants and the inflow of immigrants are chosen. We also used the
rates of specific subgroups within the ethnic minority population. More specifically, we
used rates of non-Belgian nationals, the inflow of immigrants, non-Belgian EU-nationals,
non-EU nationals, Turkish nationals, Moroccan nationals, African nationals, and asylum
seekers. African nationals are defined as the nationals of African countries other than
Morocco. All these rates are calculated per 1000 inhabitants for the years between 2001
and 2006.
We also need to include several control variables relating to economic structure and
urbanization. To be more specific, the social disorganization approach has stressed the
impact of poverty (Bailey, 1984; Lee, 2000; Messner and South, 1986; Parker, 1989;
Sampson, 2006), unemployment (Grant and Martinez, 1997; Miethe et al., 1991) and
income inequality (Blau and Blau, 1982; Farley, 1987; Shihadeh and Ousey, 1996) as
key determinants of urban crime rates (Crutchfield et al., 1982; Sampson and Groves,
1989; Shaw and McKay, [1942] 1969). Urbanization is generally considered to be a fac-
tor that is strongly related to crime, because it is assumed that crime occurs more often
Figure 2. Geographical distribution of property crime, 2001–6.
Note: Averages for 589 Belgian municipalities.
Source: Belgian Federal Police, operational police information.
206 European Journal of Criminology 8(3)
in urban areas. In the classic literature, urbanization has always been associated with
social disintegration (Christiansen, 1970; Wirth, 1938). One of the most straightforward
operationalizations of urbanization is the number of inhabitants. However, some of the
larger agglomerations in Belgium are for political reasons divided into small autono-
mous municipalities. For example, the Brussels agglomeration, with some 1 million
inhabitants, is divided into 19 separate and sometimes small municipalities. Therefore
we selected ‘population density’ as an indicator of urbanization, by dividing the popula-
tion by the size of the municipality.
In terms of economic structure, we used the mean income per inhabitant in the
municipality as an indicator of income. In other models, we tested poverty using the
median income, but this did not lead to substantially different results. For relative pov-
erty, we used the Gini coefficient as a measure of income inequality within the munici-
pality. We also included the unemployment rate as a control variable for social exclusion.
The underlying logic here is that poverty is not necessarily limited to a lack of financial
resources, but can also be considered as exclusion from meaningful social interaction,
most notably paid work. The unemployment rate is provided by the State Institute for
Labour, and is expressed as the unemployed as a percentage of the total labour force.
Finally, we considered the fact that Belgium is a federal country with strong
Figure 3. Geographical distribution of violent crime, 2001–6.
Note: Averages for 589 Belgian municipalities.
Source: Belgian Federal Police, operational police information.
Bircan and Hooghe 207
autonomous regions. Although crime and justice are still federal responsibilities, and
thus can be assumed to be homogeneous across the territory, various social, educational
and cultural competences are dealt with at the level of the regions. To monitor the pos-
sible effect of regional policy differences, we included a dummy variable to indicate
whether the municipality belongs to the Walloon, the Brussels or the Flemish autono-
mous region.
Results
When constructing our first model on property crime rates, we included the presence and
the inflow of immigrants in different models. The two models in Table 3 suggest that
neither the rate of non-Belgians nor the percentage of incoming immigrants has a signifi-
cant effect on property crime rates. Control variables mostly live up to theoretical expec-
tations, with significantly higher crime rates in urban areas and in communities with high
unemployment rates and higher levels of income inequality.
If we include the inflow of immigrants in the model (Model II in Table 3), there is no
significant change in the explained variance. Neither the presence of non-Belgians nor
the inflow of immigrants proves to be a significant determinant of property crime. The
control variables on population density, income inequality and unemployment rate retain
most of their original strength. This first analysis suggests that, in densely populated
urban areas with high unemployment rates, the property crime rate tends to be higher
regardless of the presence or inflow of non-Belgians.
Our next step in explaining property crime rates involves controlling for specific
migrant groups to ascertain whether the original model can be improved. In order to avoid
multicollinearity problems, each one of the variables was introduced separately in a full
model, as in Table 3. In Table 4, the results from five different OLS regressions are reported.
Each of the five models includes a new diversity indicator but all the control variables
(reported also in Table 3) used for fitting the models are the same in every model.
Table 3. Explaining property crime rates in Belgian municipalities (N = 589)
Model I Model II
Beta Beta
Non-nationals rate (log) 0.022 (0.024)
Incoming migrants rate (log) 0.063 (0.025)
Population density (log) 0.329*** (0.012) 0.343*** (0.013)
Mean income per capita (log) 0.130 (0.530) 0.084 (0.548)
Income inequality (Gini) 0.373*** (1.128) 0.325*** (1.136)
Unemployment rate (log) 0.632*** (0.096) 0.497*** (0.112)
Intercept –8.784 (5.235) –6.984 (5.417)
Adj. R2.65 .65
Notes: Entries are the result of an OLS regression analysis at the level of the municipality. Dependent
variable: average property crime rate, 2001–6. Standard errors in parentheses.
* p < .05; ** p < .01; *** p < .001
208 European Journal of Criminology 8(3)
Next we restrict the definition of non-Belgians to non-EU nationals. We also include
information on specific ethnic groups within the population. We observe that the relation
with non-EU-nationals is significant but there is no significant effect of the presence of
Moroccan or Turkish groups within the population. At first glance, there is a significant
relation with the concentration of African nationals, but it has to be remembered that
these are largely concentrated in the area of the capital, Brussels, which has high crime
rates. Self-evidently this correlation does not imply any causality. In any case, the number
of African nationals in Belgium is simply too small to be able to have any meaningful
impact on crime figures.
We follow a similar path to explain violent crime rates. First we determine whether
the percentage of non-nationals or the inflow of immigrants has an effect on violent
crime rates. The results of this analysis show that the unemployment rate is the main
determinant of violent crime (Table 5). Again, there is no significant effect of the pres-
ence of non-nationals or the inflow of immigrants. In both models, the unemployment
rate explains most of the violent crime rate. Subsequently, we again develop new mod-
els, with more refined measurements of ethnic diversity. In these models (Table 6) we
Table 4. Effect of different diversity variables on property crime (N = 589)
Beta Adj. R2
Non-EU nationals rate (log) 0.209*** (0.023) .66
EU nationals rate (log) 0.039 (0.019) .66
Turkish nationals rate (log) –0.050 (0.022) .66
Moroccan nationals rate (log) –0.105 (0.030) .67
African nationals rate (log) 0.205*** (0.024) .70
Notes: Entries are the result of an OLS regression analysis at the level of the municipality when all the
control variables are kept in the model as in Table 3. Dependent variable: average property crime rate,
2001–6. Standard errors in parentheses.
* p < .05; ** p < .01; *** p < .001
Table 5. Explaining violent crime rates in Belgian municipalities (N = 589)
Model I Model II
Beta Beta
Non-nationals rate (log) 0.064 (0.018)
Incoming migrants rate (log) –0.033 (0.019)
Population density (log) 0.062 (0.031) –0.030 (0.010)
Mean income per capita (log) –0.066 (0.359) –0.135 (0.423)
Income inequality (Gini) –0.024 (0.830) 0.014 (0.877)
Unemployment rate (log) 0.597*** (0.092) 0.584*** (0.086)
Intercept 3.747 (3.869) 6.716 (4.184)
Adj. R2.48 .41
Notes: Entries are the result of an OLS regression analysis at the level of the municipality. Dependent
variable: average violent crime rate, 2001–6. Standard errors in parentheses.
* p < .05; ** p < .01; *** p < .001
Bircan and Hooghe 209
do not observe any significant effect of non-EU nationals. For specific ethnic groups, we
observe the same results as for property crime, with no significant effect for Moroccan
and Turkish groups in the population. High unemployment rate (parameter = 0.706) still
remains the most substantial predictor of the violent crime rate in Belgian communities.
Discussion
Several studies have found a positive relationship between immigration and crime rates,
and this relationship has certainly affected crime policies (Martinez and Valenzuela,
2006). Our analysis of crime rates in Belgium, however, did not reveal any meaningful
significant relationship between the presence or the inflow of ethnic minorities and the
rates of property and violent crime. Unemployment rates clearly are the most important
determinant of property and violent crime. For the two largest ethnic minority groups in
the country (Turkish and Moroccan communities), we did not observe any significant
relationship with crime rates.
Because the relationship between ethnicity and crime is sensitive, it is crucial to men-
tion some limitations of the data. The research hypotheses were tested at the community
level including immigration figures, economic factors (unemployment rates, income),
population figures and crime rates for the period 2001–6. It is has to be noted that no
individual-level data (whether about the victims or about the perpetrators) are available
for the Belgian context. As such, the current analysis does not lead to any conclusions
regarding the individual level. However, given the absence of a fully representative sam-
ple of all perpetrators and/or victims of criminal acts, it is difficult to arrive at a reliable
estimate of these individual-level causal relations. Furthermore, in all models included in
this article the effect of the unemployment rate is so strong that this suggests that the
experience of unemployment is, in some way or another, related to the causal explana-
tion of criminal behaviour.
Given these caveats, we still believe that the current analysis is relevant, not only from
a theoretical perspective but also from a policy perspective. In Belgium too, a number of
political parties and policy advisers have suggested a causal relationship between diver-
sity and crime levels in urban areas. The current analysis suggests quite strongly that
economic indicators (most notably unemployment) have a strong significant relationship
Table 6. Effect of different diversity variables on violent crime (N = 589)
Beta Adj. R2
Non-EU nationals rate (log) 0.078 (0.019) 0.51
EU nationals rate (log) –0.013 (0.014) 0.50
Turkish nationals rate (log) –0.156 (0.025) 0.31
Moroccan nationals rate (log) –0.214 (0.025) 0.38
African nationals rate (log) 0.304*** (0.016) 0.64
Notes: Entries are result of an OLS regression analysis at the level of the municipality when all the control
variables are kept in the model as in Table 5. Dependent variable: average violent crime rate, 2001–6.
Standard errors in parentheses.
* p < .05; ** p < .01; *** p < .001
210 European Journal of Criminology 8(3)
with crime rates, and we did not find any comparable evidence for ethnic composition.
Despite the fact that we used a very comprehensive measurement of ethnic diversity,
relationships were almost never significant. We can therefore be quite confident in our
claim that any effort to explain the relationship between social disorganization and crimi-
nal behaviour should focus first of all on socioeconomic differences, and not on cultural
and ethnic diversity.
References
Albrecht HJ (1997) Ethnic minorities, crime and criminal justice in Germany. In: Tonry M (ed.)
Ethnicity, Crime and Immigration. Chicago: University of Chicago Press, 31–101.
Bailey WC (1984) Poverty, inequality, and city homicide rates: Some not so unexpected findings.
Criminology 22: 531–550.
Bankston CL (1998) Youth gangs and the new second generation: a review essay. Aggression and
Violent Behavior 3: 35–45.
Blau JR and Blau PM (1982) The cost of inequality: Metropolitan structure and violent crime.
American Sociological Review 47(1): 114–129.
Bursik RJ and Grasmick HG (1993) Neighborhoods and Crime: the Dimensions of Effective
Community Control. New York: Lexington Books.
Butcher KF and Piehl AM (1998) Cross-city evidence on the relationship between immigration
and crime. Journal of Policy Analysis and Management 17: 457–493.
Butcher KF and Piehl AM (2008) Crime, corrections, and California: What does immigration have
to do with it? Population Trends and Profile 9: 1–23.
Byrne JM (1986) Cities, citizens, and crime: the ecological/nonecological debate reconsidered. In:
Byrne JM and Sampson RJ (eds) The Social Ecology of Crime. New York: Springer, 77–101.
Chilton R, Teske R and Arnold H (1995) Ethnicity, race, and crime: German and non-German sus-
pects 1960–1990. In: Hawkins D (ed.) Ethnicity, Race, and Crime: Perspectives across Time
and Space. Albany, NY: SUNY Press, 323–340.
Christiansen KO (1970) Industrialization and urbanization in relation to crime and juvenile delin-
quency. In: Glaser D (ed.) Crime in the City. New York: Harper & Row, 47–55.
Crutchfield R, Geerken M and Gove W (1982) Crime rates and social integration: the impact of
metropolitan mobility. Criminology 20: 467–478.
Elliott DS, Wilson WJ, Huizinga D, Sampson RJ, Elliott A and Rankin B (1996) The effects of
neighborhood disadvantage on adolescent development. Journal of Research in Crime and
Delinquency 33: 389–426.
Farley JE (1987) Suburbanization and central-city crime rates: New evidence and a reinterpreta-
tion. American Journal of Sociology 93(3): 688–700.
Ferracuti F (1968) European migration and crime. In: Wolfgang ME (ed.) Crime and Culture:
Essays in Honor of Thorsten Sellin. New York: Wiley, 189–219.
Grant DS and Martinez R, Jr (1997) Crime and the restructuring of the U.S. economy: a reconsid-
eration of the class linkages. Social Forces 75: 769–799.
Hagan J and Palloni A (1999) Sociological criminology and the mythology of Hispanic immigra-
tion and crime. Social Problems 46: 617–632.
Kawachi I, Kennedy BP and Wilkinson RG (1999) Crime: Social disorganization and relative
deprivation. Social Science and Medicine 48(6): 719–731.
Kposowa AJ, Breault KD and Harrison BM (1995) Reassessing the structural covariates of violent and
property crimes in the USA – a county level analysis. British Journal of Sociology 46(1): 79–105.
Krivo LJ and Peterson RD (1996) Extremely disadvantaged neighborhoods and urban crime.
Social Forces 75: 619−650.
Bircan and Hooghe 211
Lee MT (2000) Concentrated poverty, race, and homicide. Sociological Quarterly 41: 189–206.
Lee MT (2003) Crime on the Border: Immigration and Homicide in Urban Communities. New
York: LFB Scholarly Publishing.
Lee MT and Martinez R (2002) Social disorganization revisited: Mapping the recent immigration
and black homicide relationship in northern Miami. Sociological Focus 35: 363–80.
Lee MT, Martinez R and Rosenfeld R (2001) Does immigration increase homicide? Negative
evidence from three border cities. Sociological Quarterly 42: 559−580.
Lippens R (1995) Contextualisering van politiële criminele statistiek. In: Hebberecht P, Bovyn M,
Colle P and Lippens R (eds) Politie en Criminaliteit in Context. Gent: Onderzoeksgroep
Criminologie.
Logan JR and Stults BJ (1999) Racial differences in exposure to crime: the city and suburbs of
Cleveland, 1990. Criminology 37: 251−276.
McDonald WF (1997) Crime and illegal immigration: Emerging local, state, and federal partner-
ships. National Institute of Justice Journal 232: 2–10.
Maguire K and Pastore AL (1995) Sourcebook of Criminal Justice Statistics 1994. Washington,
DC: US Department of Justice.
Manfrass K (1991) Türken in der Bundesrepublik. Nordafrikaner in Frankreich. Bonn: Bouvier.
Martinez RJ (2002) Latino Homicide: Immigration, Violence and Community. New York: Routledge.
Martinez RJ and Lee MT (2000) On immigration and crime. In: LaFree G and Bursik R (eds)
Criminal Justice 2000: The Changing Nature of Crime, Vol. I. Washington, DC: National
Institute of Justice.
Martinez RJ and Valenzuela A (eds) (2006) Immigration and Crime: Race, Ethnicity, and Violence.
New York: New York University Press.
Mears DP (2001) The immigration–crime nexus: toward an analytic framework for assessing and
guiding theory. Research and Policy 44: 1–19.
Mears DP (2002) Immigration and crime: What’s the connection? Federal Sentencing Reporter
14(5): 284–288.
Messner SF (1982) Poverty, inequality and the urban homicide rate. Criminology 20: 103–114.
Messner SF and South SJ (1986) Economic deprivation, opportunity structure, and robbery vic-
timization: Intra and interracial patterns. Social Forces 64: 975–991.
Miethe TD, Hughes M and McDowall D (1991) Social change and crime rates: an evaluation of
alternative theoretical approaches. Social Forces 70: 165–185.
Newman G (ed.) (1999) Global Report on Crime and Justice. New York: Oxford University Press.
O’Brien RM (1983) Metropolitan structure and violent crime: Which measure of crime? American
Sociological Review 48: 434–437.
Olson CP, Laurikkala MK, Huff-Corzine L and Corzine J (2009) Immigration and violent crime:
Citizenship status and social disorganization. Homicide Studies 13(3): 227–241.
Osgood WD (2000) Poisson-based regression analysis of aggregate crime rates. Journal of
Quantitative Criminology 16: 21–43.
Parker KF, McCall PL and Land KC (1999) Determining social-structural predictors of homicide:
Unit of analysis and related methodological concerns. In: Smith MA and Zahn MA (eds)
Homicide: a Source Book Social Research. Thousand Oaks, CA: Sage.
Parker RN (1989) Poverty, subculture of violence, and types of homicide. Social Forces 67: 983–1007.
Pattyn M (1994) Criminaliteit in de grote steden. Politeia 5: 12–16.
Pauwels L (2007) Buurtinvloeden en Jeugddelinquentie. The Hague: Boom Juridische Uitgevers.
Peterson RD, Krivo LJ and Harris MA (2000) Disadvantage and neighbourhood violent crime: Do
local institutions matter? Journal of Research in Crime and Delinquency 37: 31−63.
Reid LW, Weiss HE, Adelman RM and Jaret C (2005) The immigration–crime relationship:
Evidence across U.S. metropolitan areas. Social Science Research 34: 757–780.
212 European Journal of Criminology 8(3)
Rosenfeld R (1986) Urban crime rates: Effects of inequality, welfare dependency, region, and race.
In: Byrne JM and Sampson RJ (eds) The Social Ecology of Crime. New York: Springer, 116–130.
Sampson RJ (2006) How does community context matter? Social mechanisms and the explana-
tion of crime. In: Wikström P-O and Sampson RJ (eds) The Explanation of Crime: Context,
Mechanisms, and Development. New York: Cambridge University Press.
Sampson RJ (2008) Rethinking crime and immigration. Contexts 7: 28–33.
Sampson RJ and Groves WB (1989) Community structure and crime: Testing social-disorganization
theory. American Journal of Sociology 94(4): 774–802.
Sampson RJ and Raudenbush SW (1999) Systematic social observation of public spaces: a new
look at disorder in urban neighborhoods. American Journal of Sociology 105: 603−651.
Sampson RJ, Morenoff JD and Raudenbush S (2005) Social anatomy of racial and ethnic dispari-
ties in violence. American Journal of Public Health 95: 224–232.
Sampson RJ, Raudenbush S and Earls F (1997) Neighborhoods and violent crime: a multilevel
study of collective efficacy. Science 277: 918–924.
Sellin T (1938) Culture and Conflict. New York: Social Science Research Council.
Shaw C and McKay H ([1942] 1969) Juvenile Delinquency and Urban Areas. Chicago: University
of Chicago Press.
Shihadeh ES and Ousey GC (1996) Metropolitan expansion and black social dislocation: the link
between suburbanization and center-city crime. Social Forces 75: 649–666.
Short JF (1971) Introduction. In: Short JF (ed.) The Social Fabric of the Metropolis: Contributions
of the Chicago School to Urban Sociology. Chicago: University of Chicago Press, xi–xlvi.
Short JF (1997) Poverty, Ethnicity and Violent Crime. Boulder, CO: Westview Press.
Stigler SM (1986) The History of Statistics: the Measurement of Uncertainty before 1900.
Cambridge, MA: Harvard University Press.
Stowell, JI and Martinez RJ (2007) Displaced, dispossessed, or lawless? Examining the link between
ethnicity, immigration, and violence. Aggression and Violent Behavior 12(5): 564–581.
Tonry M (1997) Ethnicity, Crime, and Immigration: Comparative and Cross-National Perspectives.
Chicago: University of Chicago Press.
Walter M and Kubink M (1993) Ausländerkriminalität – Phänomen oder Phantom der (Kriminal-)
Politik? Monatsschrift für Kriminologie und Strafrechtsreform 76: 306–319.
Warner B (2003) The role of attenuated culture in social disorganization theory. Criminology 41:
73−97.
Weijters G, Scheepers P and Gerris J (2009) City and/or neighbourhood determinants? Studying
contextual effects on youth delinquency. European Journal of Criminology 6(5): 439–455.
Wells LE and Weisheit R (2004) Patterns of rural and urban crime. A country-level comparison.
Criminal Justice Review 29(1): 1–22.
Wikström P-O (1991) Urban Crime, Criminals and Victims. New York: Springer.
Wirth L (1938) Urbanism as a way of life. American Journal of Sociology 44(1): 1–24.
Author biographies
Tuba Bircan is a research fellow at the Department of Political Science of the Catholic
University of Leuven, Belgium, and working on her PhD. Her research interests are
social cohesion, migration, crime, economic inequality and anti-immigrant attitudes.
Marc Hooghe is a Professor of Political Science at the Department of Political Science
of the University of Leuven, Belgium. He holds PhDs in political science and in sociol-
ogy, and he has been president of the Belgian political science association. His main
research interests are social cohesion, social capital and subjective well-being.
