International Journal of Housing Policy

ISSN: 1949-1247 (Print) 1949-1255 (Online) Journal homepage: https://www.tandfonline.com/loi/reuj20

Housing finance reform in Mexico: the impact of

housing vacancy on property crime

César M. Fuentes & Vladimir Hernandez

To cite this article: César M. Fuentes & Vladimir Hernandez (2014) Housing finance reform in
Mexico: the impact of housing vacancy on property crime, International Journal of Housing Policy,
14:4, 368-388, DOI: 10.1080/14616718.2014.955332

To link to this article: https://doi.org/10.1080/14616718.2014.955332

Published online: 17 Sep 2014.

Submit your article to this journal

Article views: 482

View related articles

View Crossmark data

Citing articles: 12 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=reuj20
International Journal of Housing Policy, 2014
Vol. 14, No. 4, 368
388, http://dx.doi.org/10.1080/14616718.2014.955332

Housing finance reform in Mexico: the impact of housing vacancy

on property crime
C
esar M. Fuentesa* and Vladimir Hernandezb
a
Department of Urban and Environmental Studies, El Colegio de la Frontera Norte, Ciudad
Ju
exico; bDepartment of Architecture, Universidad Aut

arez, M
onoma de Ciudad Ju
arez,
Ciudad Ju

arez, Mexico

Since 2000, the numbers of vacant housing units have increased dramatically in
many parts of Mexico. Much of this has been tied to the liberalisation of the
housing market where private homebuilders play an important role in the
production of housing and the role of public sector is limited to the support of
housing finance. Public debates concerning the regulation of private sector
homebuilders as well as the practices and policies of government lending
agencies have hinged on housing overproduction and vacancy rates in recent
years, with the latter associated with social problems in many cities. This paper
explicitly examines the relationship between vacancy rate and levels of
property crime (burglaries) in Ciudad Ju
arez, Chihuahua, Mexico. Along with
data from the Population and Housing Census, the State Public Security Office
data for the 2008
2009 period on property crimes (burglary), were geocoded,
mapped and aggregated to the census-tract level. The ordinary least square
regression results show that there are significant local variations in the
relationships between the risk of property crimes and the percentage of vacant
housing units, commercial and service land use and residents who are male
aged 15
24. The findings strongly link housing overproduction with vacancy
rates and crime, and suggest that private sector homebuilders and the mortgage
allocation system should be more strongly regulated in Mexico.
Keywords: vacant housing units; burglaries; ordinary least regression; Ciudad
Ju
arez, Mexico

Introduction
Since 1990, housing finance reform has transformed how housing is acquired and
produced in Mexico. On the one hand, government agencies, which were already
an important source of housing finance, expanded their lending volume consider-
ably as agencies such as Institute of National Worker’s Housing Fund (INFONA-
VIT),1 Housing Fund of Institute of Social Security of State Workers
(FOVISSSTE),2 Central Bank of Mexico (FOVI),3 National Fund for Popular

*Corresponding author. Email: [email protected]

Ó 2014 Taylor & Francis
 International Journal of Housing Policy 369

Housing (FONHAPO),4 Federal Mortgage Society (SHF)5 and Limited Financial

Societies (SOFOLES)6 began to provide more loans. On the other hand, housing
built by private homebuilders also advanced (Monkkonen, 2011).
During the period 2005
2010 more than 12.4 million houses were built. The
national housing stock increased from 23.2 million units in 2005 to 35.6 million in
2010. However, the percentage of vacant units also rose precipitously in Mexico,
increasing from 3 million (11.6%) in 2005 to 5 million (14.0%) in 2010 (Banco
Bilbao Vizcaya Argentaria (BBVA), 2011; S
anchez & Salazar, 2011). Moreover,
the cities where more housing units were built also developed higher vacancy rates.
Ciudad Ju
arez (23%) and Tijuana (20%) received national attention for the number
of housing units built and the high percentage of vacant units (BBVA, 2011). In
Ciudad Ju
arez, the vacancy rate rose from 6.52% (22,121 units) in 2005 to 30%
(110,087 units) in 2010 (Instituto Nacional de Estad
ısticas y Geograf
ıa [Inegi],
2010).
The new policy assigned the leading role to private sector homebuilders and
identified a limited role for the public sector in financing the production of housing.
The change has led to the overproduction of housing developments containing thou-
sands of identical small tract houses (around 45 m2) mostly along the urban periph-
ery (Monkkonen, 2008). Moreover, the government loan allocation system has
disproportionately directed investment towards places with larger numbers of sala-
ried employees (Monkkonen, 2011). The combination of overproduction of ‘social
interest housing’7 and misallocation of loans subsequently shaped a foreclosure cri-
sis that produced hundreds of vacant units.
At the same time, since 2007 the city of Ciudad Ju
arez, Chihuahua, has experi-
enced a high rate of homicides as a consequence of the war between drug cartels
fighting each other over control of both local market distribution and the most
important routes for transporting drugs into the United States. As a consequence of
the high level of impunity and the diversification of the drug cartels activities,
many other crimes increased such as robbery, burglary, carjacking, simple assault,
etc., mainly in social interest housing developments (Fuentes & Hern
andez, 2013).
These processes have contributed to the abandonment of additional units by their
residents. As a result, most vacant housing units had been vandalised; copper wiring
and appliances have been stolen; and some vacant units have been utilised in drug-
related crimes, including homicides, assaults, etc. In this context, multiple actors
have speculated that the high rate of housing vacancy has stimulated high levels of
delinquency, especially in recent social interest housing developments.
Most of the studies that have empirically tested the relationship between hous-
ing vacancy rate and property crime have been conducted in the United States. A
few mapping and/or spatial studies on criminal activity in Mexico also exist
(Vilalta, 2009; 2010; Vilalta & Muggah, 2012). However, none of them have con-
centrated in the impact of socio-economic variables such as vacant housing and its
link with different crimes such as burglaries in developing country such as Mexico.
370 C.M. Fuentes and V. Hernandez

A central goal of this paper is, therefore, to address this oversight and provide evi-
dence for international comparison. This article, furthermore, examines an issue
which has been a focus of attention in housing and urban studies in recent years.
The impact of vacant housing units on property (burglaries) in the city of Ciudad
Ju
arez, Chihuahua, Mexico has also been a particularly well-debated local issue.
The remainder of this article is divided into four sections. The first part provides
a brief history of the most relevant policy changes to the housing finance system in
Mexico. The second section explores the process through which Ciudad Juarez
experienced both high vacancy and delinquency rates. The third section presents
the results of the OLS multiple regression model, which assesses the relationship
between the vacancy rate and property crime at the census tract level. The fourth
section presents the conclusions of this paper that outline the implications of this
analysis for the regulation of the mortgage lending allocation system.

Housing finance reform in Mexico

The liberalisation of the housing system that began in the early 1990s has trans-
formed the way that housing is produced and acquired in Mexico (Monkkonen,
2011). At the end of 1995 an expansion in lending for housing by government agen-
cies began following the devaluation of the Mexican peso (Edwards, 1998). The
economic crisis generated by the currency failure affected the housing sector as
banks ceased lending for housing. Although private banks withdrew from the hous-
ing finance market, government agencies increased the numbers of loans that they
provided. Government agencies such as INFONAVIT, FOVISSSTE, FOVI, FON-
HAPO, SHF and SOFOLES are the most important source of lending (Monkkonen,
2011). Commercial banks have subsequently resumed mortgage lending, but their
share of the market remains small.
Since 2000, the reform included an initiative to foster the expansion of the pri-
vate construction industry in Mexico. INFONAVIT convened a series of meetings
with homebuilders to convince them that low-cost housing could be profitable.
INFONAVIT facilitated access to loans for participating developers, offered con-
struction loans connected to mortgages and assisted builders with state and local
permitting as well as land acquisition. According to Coulomb, Le
on, Puebla, and
Castro (2009), the construction companies focused on the following four elements
in determining the location of housing developments: (1) the existence of demand
with high purchasing power; (2) the availability of accessible financial aid for
buyers; (3) the ability to build the development at a competitive cost; and (4) the
existence of cheap land on which to build.
The new system led to the production of housing developments that are com-
posed mainly of thousands of identical, small tract houses8 located in the urban
periphery, which are dramatically altering the urban structure of Mexican cities
(Monkkonen, 2008). The system has been criticised for its lack of planning and
 International Journal of Housing Policy 371

Table 1. Ratio of new dwellings according to type and size in Mexico (2001
2006).

Basic dwelling Social dwelling Economic dwelling Medium dwelling Large dwelling

<30 m 2
30
45 m 2
45
55 m 2
55
100 m 2
>100 m2
0.2% 27.3% 32.2% 40.0% 0.3%
























59:7



























!

Source: Coulomb et al. (2009).

limited access to basic amenities, such as supermarkets, public spaces and public
transportation (Maya & Cervantes, 1999). During the 2000
2006 period, almost
60% of the new housing units in the country have a size of between 30 and 45 m2
(see Table 1).
The design, size, cost and localisation of housing developments have thus been
increasingly determined by the economic interests of private homebuilders (Coulomb
et al., 2009), with the new role of the public sector limited to financing the produc-
tion of housing. Additional effects of these changes include increasing urban mobil-
ity, isolation of many housing developments from the rest of the city, and insufficient
urban infrastructure, such as public transportation, street lighting, police stations, etc.
This situation has contributed to a social
spatial context in which opportunities to
commit crimes in such housing developments abound. According to Williams,
Galster, and Verma (2014, p. 382) ‘several characteristics of the neighbourhood

the location of properties vis-a-vis the street, access to mass transit, and the presence
and design of public spaces and facilities
will independently affect both how
many foreclosures and how much crime will be observed there’.

From housing overproduction to vacancy

In recent years, thousands of small tract houses have been left vacant, especially
along urban peripheries. Over the last 10 years, the percentage of vacant units
increased from 11.6% (3 million units) in 2000 to 14.0% (5 million units) in 2010
(BBVA, 2011; S
anchez & Salazar, 2011). Higher numbers of vacant housing units
are observed in two types of locations: (1) cities with high rates of out migration
(33%) and (2) northern border cities (23%) (BBVA, 2011).
At the national level, two border cities have received national attention for their
high percentage of vacant housing units: Ciudad Ju
arez9 (23%) and Tijuana (20%)
(BBVA, 2011). The vacancy rate in Ciudad Ju
arez rose from 6.52% (22,121 units)
in 2005 to 30% (110,087 units) in 2010 (Inegi, 2010). There are three main factors
driving this phenomenon: (1) the oversupply of local housing markets by private
developers; (2) the local impacts of the global economic recession that has left
many production workers without jobs, with many of them unable to pay-off their
372 C.M. Fuentes and V. Hernandez

loans and leading them to abandon their homes; and (3) many residents have fled
the cities as a result of the high level of violence (Fuentes & Hern
andez, 2013).
In Mexico, the new housing policy emphasised the connection between salaried
employment and mortgage lending, which means that industries with large numbers
of salaried workers receive an implicit subsidy (Monkkonen, 2011). Ciudad Ju
arez
is an export-processing zone, or maquiladora10 hub, which has created thousands
of employment opportunities for production workers since the mid-1960s. Employ-
ment peaked in 2000 with 249,509 positions. Ciudad Ju
arez was the recipient of
thousands of loans as a product of the high percentage of the economically active
population employed in the formal sector of the economy. During the 2001
2006
period, more than 112,678 new houses were built in Ciudad Ju
arez, and as a result
the city became the national leader in the production of housing (Maycotte &
S
anchez, 2009). However, 86.78% of these new housing units constructed were in
the segment of social housing it means that most of them have a size of 30
45 m2
of construction (Table 2).
Forecasts for housing demand did not foresee the reduction in population
growth that would result from the economic crisis in 2000. Moreover, since 2000,
Ciudad Ju
arez has been the only city along the United States
Mexico border where
the average increase in housing units exceeded the annual average increase in
households, which created an excess capacity of 12% (Fuentes & Pe~na, 2006). This
situation worsened in 2010 when the annual increase in housing units exceeded the
annual increase in households by 126% (Romo, C
ordova, Fuentes, & Brugu
es,
2012). Many of these units were not sold, and a landscape of consolidated neigh-
bourhoods that emerged appeared abandoned because of the number of vacant units
(Maycotte & S
anchez, 2009). Essentially, overcapacity in the housing market left
many units vacant, especially in the low-income segment.

Table 2. Housing production by type in Ciudad Ju
arez, Chihuahua, Mexico.

Social interest Middle-income Upper-income

Year housing dwellings dwellings Total

2001 10,413 3037 539 13,989

2002 15,012 577 240 15,829
2003 20,149 2823 415 23,387
2004 20,664 1062 406 22,132
2005 16,258 3188 942 19,796
2006 15,290 1255 1313 17,545
Total 97,786 11,942 2951 112,678
Percentage 86.78 10.60 2.62 100.00

Source: Maycotte and S
anchez (2009).

 International Journal of Housing Policy 373

From foreclosure to vacancy and crime

In the early 2000s, the city suffered substantial job losses in manufacturing (from
2000 to 2003, approximately 83,074 jobs were lost) because of the brief downturn
in the US economy. Furthermore, during the 2007
2009 global economic recession
the city lost an additional 34,000 job positions Instituto Municipal de Investigaci
on
y Planeaci
on [IMIP], 2010). As a result, many workers who were fired and could no
longer maintain repayments on their mortgages, abandoned their homes. According
to INFONAVIT officials, more than 4300 houses for which the institute had pro-
vided loans were abandoned in the city during the period 2005
2010 (Vel
azquez
& Mart
ınez, 2011).
Foreclosures harm neighbourhoods by triggering off extended vacancies and
stimulating the appearance of abandoned and blighted buildings, and in turn attract-
ing deviant activities (Immergluck & Smith, 2006). In Ciudad Ju
arez, police reports
indicate that more than 8228 vacant housing units were vandalised, and theft of wir-
ing, copper or appliances were frequently reported (Carrasco, 2012). Vacant proper-
ties are perceived as attractive unguarded spaces, which yield elevated rates of a
variety of deviant activities (e.g., substance abuse) that fuel other crimes, including
violent crimes such as robbery, assault and homicide (Carrasco, 2012). This out-
come is discussed by the broken windows thesis that encompasses ideas elaborated
more fully within the classic and contemporary formulations of social disorganisa-
tion theory and routine activities perspectives (Arnio, Baumer, & Wolff, 2012).
The broken windows thesis (Corman & Mocan, 2005; Jang & Johnson, 2001;
Silverman & Della-Giustina, 2001; Willson & Kelling, 1982) states that visual
signs of abandonment, such as high vacancy rates, can fuel additional incivility and
‘physical and social disorder that may translate into higher levels of crime by dis-
couraging community involvement and providing unregulated private and public
spaces in which deviant activities can flourish’ (Arnio et al., 2012, p. 1599).
Vacant and abandoned buildings are often considered a component of neigh-
bourhood physical disorder (as opposed to social disorder). Physical disorder refers
more to sustained conditions rather than to particular events. However, these condi-
tions are, of course, created by events and actions (Skogan, 1990). Social disorder
includes specific undesirable events and behaviour (e.g., aggressive panhandling
or public drinking) that may in turn cause problems that are more serious.
However, whether this pattern reflects causation is disputed (Sampson &
Raudenbush, 1999). The broken windows thesis encompasses ideas that are elabo-
rated more fully within social disorganisation and routine activities theories (Arnio
et al., 2012).
Social disorganisation theory (Haining & Ceccato, 2005; Haining, Ceccato, and
Kahn 2007; Sampson, 1993; Weisburd & Mazerolle, 2000) and routine activities the-
ory (Cohen & Felson, 1979; Kennedy & Forde, 1990; Koening & Linder, 2004) also
explain criminal spatial variation as responses to environmental conditions (Vilalta,
374 C.M. Fuentes and V. Hernandez

Figure 1. Spatial distribution of housing vacancy in Ciudad Ju
arez, Chihuahua, 2010.

Source: Author’s elaboration based on the Population and Housing Census (Inegi, 2010).

2010). The spatial distribution of vacancy rates is displayed in Figure 1. Data from
2005 and 2010 indicate that the number of census tracts with high vacancy rates
increased mainly in the south-east region of the city of Ciudad Ju
arez where most of
the social interest housing developments are located. The map of vacancy rates high-
lights key elements of the notion of a dual city, which refers to the polarisation of
the city into two spatial units according to socio-economic status. Several main high-
ways and the railroad system divide the city into four sectors (north-west, south-
west, north-east and south-east) that vary by socio-economic status.
 International Journal of Housing Policy 375

The city’s downtown is located on the western bank of the Grande River/Bravo
River and is surrounded by old residential areas. The railroad system follows inter-
state 45 along a west
east axis across the centre of the city. Most of the western
part of the city was developed as households built their own houses incrementally.
In contrast, private sector homebuilders developed most of the eastern part of the
city. The north-east section contains housing developments for high-income resi-
dents, while the south-east section contains social interest housing for low-income
residents. Two features of this spatial pattern deserve attention. First, the census
tracts with the highest vacancy rates are concentrated in two recent social housing
developments
Riveras del Bravo and Sederos de San Isidro
in the south-eastern
part of the city. Second, pockets of high vacancy rates are observed in the old part
of the city surrounding the downtown area and in the high-income developments
located in the north-eastern part of the city.11
Since 2007, the city has simultaneously experienced unprecedented and
extreme levels of criminality mainly related to fighting among the drug cartels
that desire to control both local market distribution and the most important
routes for transporting drugs into the United States. Some housing develop-
ments with high rates of vacancy became retail distribution centres and spaces
for the consumption of drugs. As a result, more than 9626 men, women and
children have been killed between 2008 and 2012 (Mon
arrez & Cervera,
2013). The high level of impunity and diversification of the drug cartels’ activ-
ities have also led to an increase in many other crimes, such as robbery, bur-
glary, carjacking, assault, etc. (Fuentes & Hern
andez, 2013). In this context,
the housing developments with high vacancy rates became the target of many
property-related and violent crimes. According to the Municipal Secretary of
Public Safety, 977 property crimes (burglaries) were reported in 2009.
The spatial pattern of property crimes (burglaries) in Ciudad Ju
arez is character-
ised by a south
north division (see Figure 2). Three features of the spatial pattern
of residential burglaries deserve attention. First, property crimes tend to be concen-
trated in areas adjacent to the centre of the city with commercial and service land
use and along the main roads. This finding is supported by previous research in
other countries, which suggests that the intensity of crime tends to decrease with
increasing distance from the city centre (Bowers & Hirschfield, 1999; Brantingham
& Brantingham, 1984; Kohfeld & Sprague, 1988). Second, property crime in the
city is concentrated in high-density housing developments. Third, pockets of ele-
vated rates of property crime exist in the older part of the city. The presence of
many abandoned houses may contribute to the spatial pattern of property crimes.
Routine activity theory holds that residential burglary is largely a crime of opportu-
nity and supports the routine activities hypothesis (Cohen & Felson, 1979; Kennedy
& Forde, 1990; Koening & Linder, 2004). The routine activity argument also high-
lights the tendency of burglars to commit offenses in their own neighbourhoods.
This view suggests that vulnerable communities are located close to areas where
376 C.M. Fuentes and V. Hernandez

Figure 2. Rate of burglaries in Ciudad Ju
arez, Chihuahua by Census Tract (2009). Source:
Author’s elaboration based on data from the State Secretary of Public Safety (2010).

offenders live (Arnio et al., 2012; Hebert & Hyde, 1985; Smith, 1976; Wiles &
Costello, 2000; Wright, Logie, & Decker, 1995).

Methods and data

The literature review has provided theoretical elements necessary to now continue
with the analysis of the relationship between housing vacancy and property crime
(burglaries) in the context of a Mexican city such as Ciudad Ju
arez, Chihuahua.
 International Journal of Housing Policy 377

This study draws on two primary data sources. Property crime (burglary) was
collected from the Chihuahua State Public Security Office (SPSO) for 2008 and
part of 2009. The location and date of each reported crime was collected, geocoded
and aggregated at the census-tract-level utilising ArcMap geographic information
systems software. The numbers of geographic units (census tracts) were 615.
To identify the effect of housing vacancy rate on neighbourhood property crime
(burglaries), we estimate a reduced form equation relating housing vacancy rate
and others socio-economic variables to neighbourhood property crime (Immergluck
& Smith, 2006). The multiple ordinary least square (OLS) regression has the fol-
lowing form:
ln Ci ¼ a þ b1Pi þ b2 Bi þ b3 Zi þ b4 Fi þ ei: (1)
While Ci represents the natural log of property crime rate in census tract, Pi is
the population density of the tract and Bi is the density of commercial and service
employment density in the tract. Zi is a vector of resident characteristics that might
be expected to affect neighbourhood crime rates based on the literature. Fi is the
percentage of vacant housing units, measured by the number of vacant housing
divided by the number of owner occupiable housing units in tract i. The ei’s are
independent and normally distributed unobservable error terms (or residuals) with
mean zero and constant variance. The OLS method is typically employed to esti-
mate the parameters (see Miller, 1990; Selvin, 1998). The method is based on a set
of assumptions including normality, homogeneity of variance and independence of
residuals. Spatial autocorrelation (or spatial dependency) and spatial nonstationarity
(spatial heterogeneity) are two properties of spatial data that may undermine the
assumptions of traditional regression models (Bailey & Gatrell, 1995).
Based on the 2010 census and the theoretical framework, 11 structural measures
were examined in the regression model, estimated at the census tract level. Ecologi-
cal theories, especially social disorganisation and routine activity, have demon-
strated the tendency of criminal events to cluster in space. Ecological theory holds
that understanding the characteristics of places
including their physical and social
measures
that affect the number of targets and offenders in an area is necessary in
understanding of the causes of crime (Cahill & Mulligan, 2007).
Routine activities theory focuses on the presence of opportunities for crime in
an area shaped by residents’ daily activities and determined by the spatial and tem-
poral intersection of three key elements: suitable targets, motivated offenders and
the lack of capable guardians (Cohen & Felson, 1979; Miethe & Meier, 1994).
These three factors must be present simultaneously for a burglary to occur. The
structural measures suggested by the opportunity framework indirectly measure the
availability of targets and offenders.
The ‘offender’ measure includes variables that may explain why an individual
would commit a crime. Specifically, previous empirical research suggests that the
contextual characteristics of neighbourhoods in which offenders live include high
378 C.M. Fuentes and V. Hernandez

unemployment rates (Hartnagel, 2004; Kohfeld & Sprague, 1988; Neustrom &
Norton, 1995; Rountree & Land, 2000), high proportions of low-income households
(Bursik & Grasmick, 1993; Kennedy & Forde, 1990; Pratt, 2001), low levels of
education (Ehrlich, 1975; Pratt, 2001) and lone-parent families (Bottoms & Wiles,
1988; Bowers & Hirschfield, 1999).
The ‘target’ variables associated with the risk of residential burglary include the
value of dwellings (Bursik & Grasmick, 1993; Kenedy & Forde, 1990; Kohfeld &
Sparague, 1988; Paternoster & Bushway, 2001), tenure and type of housing
(Ceccato, Haining, & Signoretta, 2002; Neustrom & Norton, 1995), household
income (Bursik & Grasmick, 1993; Roundtree & Land, 2000), foreclosure (Arnio
et al., 2012; Ellen, Lacoe, & Sharygin, 2013; Immergluck & Smith, 2006) and resi-
dential mobility (Ceccato et al., 2002; Hartnagel, 2004; Pettiway, 1982; Pratt,
2001; Sampson and Groves, 1989). Vacant properties may facilitate certain types
of property crimes, such as the theft of wiring and appliances (Ellen et al., 2013).
Vacant properties also provide easy targets for vandalism and trespassing and may
provide safe havens for prostitution and drug-related crimes that can potentially
lead to more serious violent crimes (Spelman, 1993).
Measures of social control or guardianship in an area include residential stabil-
ity, racial heterogeneity, proportion of single-person households and level of family
disruption (Wilcox, Land, & Hunt, 2003). The dependent variable was transformed
to a natural log because the crime figures are substantially skewed. In the model,
the natural log of property crime (burglaries) is utilised as the dependent variable.
Of the 11 independent variables included in this study, 5 offender measures
were included: population without employment, population density, residents with-
out social security coverage, residents who are males 15
24 years old and percent-
age of land utilised for commercial and service functions. The influence of
demographic characteristics on the incidence of crimes has been recognised.
Five variables measured the target: vacant housing, dwelling size, population that
was born in another state, population that has moved within five years. To measure
guardianship, two variables were included: female-headed households and population
density. Some variables, such as population without employment and population
without income, may serve as both offender and target variables (Malczewski &
Poetz, 2005). All variables employed were extracted from the Population and Hous-
ing Census (Inegi, 2010) at census tract level (see descriptive statistics in Table 3).
Pearson’s correlation co-efficients were also utilised to test for multicollinearity
in the independent variables; bivariate correlations between the candidate predic-
tors were calculated (see Table 4). The highest positive correlation co-efficients
were observed between property crime rate and commercial density employment,
proportion of residents who were 15
24 year old males, and female-headed house-
holds. Negative correlation co-efficients were observed for the proportion of the
population born in another state. All correlations between independent variables
were below the danger level of 0.7 (Clark & Hosking, 1986).
 International Journal of Housing Policy 379

Table 3. Descriptive statistics for Ciudad Ju
arez, Chihuahua, per census tracts (n D 565).

Variables Mean Minimum Maximum S.D.1

Ln of property crime rate 0.64 0.00 45.22 2.38

Percentage of vacant housing units 22.37 2.88 76.11 9.99
Population without employment 6.88 0.9 18.79 2.86
Proportion of female-headed households 25.66 9.22 46.22 6.32
Proportion of population without social 26.43 10.74 53.48 7.63
security coverage
Proportion of residents who are male aged 15
24 50.12 29.27 66.00 3.30
Population density 55.20 0.05 187.44 35.17
Commercial and service employment density 4.67 0.04 106.09 8.24
Percentage of population moved within five years 5.2677 0.15 21.50 2.75
Percentage of population born in another state 0.2703 0.07 0.99 0.11
Social disadvantaged index ¡0.19 ¡1.50 14.07 1.11
1
Standard deviation.
Source: Own elaboration based on XII Population and Housing Census (Inegi, 2010).

The impact of housing vacancy rate on property crime in Ciudad Ju
arez,

Chihuahua (2010)
To examine relationships between property crimes (burglary) and the socio-
economic characteristics of neighbourhoods, especially vacant housing units, we
developed a multiple-linear regression model. Model selection was based on the

Table 4. Bivariate correlation (Pearson’s r) between independent variables and crime

levels.

Variables Ln of property crime rate

Percentage of vacant housing units .140

Population without employment .059
Proportion of female-headed households .216
Proportion of population without social security coverage .124
Proportion of residents who are male aged 15
24 .283
Population moved within five years .130
Proportion of population born in another estate ¡.111
Population density .036
Commercial and service employment density .425
Social disadvantaged index ¡.235

Note: Significance levels are as follows: p < 0.05; p < 0.01.
380 C.M. Fuentes and V. Hernandez

Akaike Information Criterion (AIC) to select the ‘best’ subset of independent varia-
bles to include in the multiple regression analysis (Akaike, 1981; Millar, 1990). The
final model included the natural log of burglaries rate as the dependent variable and
seven independent variables: percentage of housing units vacant, tertiary sector
employment (commercial and services) density, proportion of population without
social security coverage, proportion of female-headed households, percentage of resi-
dents who have moved within 5 years, proportion of residents who are males aged
15
24 years and social disadvantage index. This specification yielded an AIC of
2001.34. The objective here is not to identify all root causes of neighbourhood crime
but to provide a measure of the impact of vacancies on property crime.
The estimated regression co-efficients of the variables that control for offenders
(i.e., proportion of residents who are males aged 15
24 years, percentage of resi-
dents who have moved within 5 years) and targets (i.e., percentage of vacant hous-
ing and tertiary employment density) were statistically significant at the 0.05 level
(see Table 5).
This model explained 22.9% of the variability in the relative risk of property
crimes. The co-efficient for the percentage of vacant housing units was positive,
indicating that areas with high vacancy rate have higher levels of property crime.
As the proportion of vacant housing units increases, there is an increase in the prop-
erty crime (burglaries) rate. The nature of the semilog functional form facilitates
the interpretation of the effect of the rate of vacancies on property crime. For exam-
ple, a 10% increase in the proportion of vacant housing units is associated with
8.4% increase in the expected number of burglaries. This finding is associated with
social disorganisation and routine activity theories, which state that visual signs of
abandonment in communities, such as high vacancy rates, can fuel additional

Table 5. OLS regression model for relative risk of property crime in Ciudad Ju
arez.

Independent variables Model 1 (Ln of burglary rate)

Percentage of vacant housing units 0.850400 (0.429698)

Population moved within five years 0.628306 (0.463545)
Proportion of residents who are male aged 15
24 20.03997 (5.604756)
Proportion of female-headed households 0.200654 (0.683636)
Proportion of population without social security coverage 0.451826 (0.589162)
Commercial and service employment density 0.056038 (0.006587)
Social disadvantaged index ¡0.04219 (0.052580)
N 615
Adjusted R 2 0.220323
R2 0.229108

Note: Significance levels are as follows: p < 0.05; p < 0.01.
 International Journal of Housing Policy 381

incivility as well as physical and social disorder that may translate into higher levels
of crime (Skogan, 1990). Immergluck and Smith (2006) observe similar statistically
significant results for violent crimes and foreclosure but not for property crimes.
But Williams et al. (2014) found that completed foreclosures temporally lead to
property crime, though such is not the case of for violent crime.
The commercial and service employment density is a proxy variable for land
use. This variable tests the land use hypothesis for tertiary employment density and
is statistically significant. The co-efficient for tertiary employment density has a
positive effect on property crimes, indicating that areas with high commercial and
service employment density have higher levels of property crime. For example, a
10% increase in the proportion of commercial and service employment density is
associated with a 0.59% increase the expected number of burglaries.
The proportion of residents who are male aged 15
24 years has a positive asso-
ciation with property crimes (burglaries). A 10% increase in the proportion of resi-
dents who are young males is associated with 200% increase the expected number
of burglaries. This finding is consistent with the routine activities hypothesis that
neighbourhoods with high proportions of young resident provide the quintessential
environment for the convergence of three elements: motivated offenders, suitable
targets and the absence of capable guardianship (Henderson & Stone, 1999). This
finding is similar to the patterns that Malaczewski and Poetz (2005) observe in Port-
land, Oregon where neighbourhoods with high proportions of university students
are at higher relative risk for residential burglaries.
The co-efficient associated with the percentage of residents who have moved
within five years is positive and indicates that census tracts with higher percentages
of new residents have higher levels of property crime. A 10% increase in the pro-
portion of residents who have moved within five years is associated with 10.2%
increase in the expected number of burglaries.
A ranking of independent variables by their beta weights indicates that tertiary
employment density has the largest impact on property crime. The impact of this
variable was followed by the proportion of vacant housing units, proportion of resi-
dents who are males aged 15
24 years and percentage of residents who have
moved within 5 years.
The only regression co-efficient with an unexpected sign was the social disad-
vantage index, which is negative but not statistically significant. The co-efficient
for guardianship (i.e., the proportion of female-headed households) on property
crimes was not statistically significant. In addition, the co-efficients for offenders
(i.e., the proportion of population without social security coverage and social disad-
vantage index) were not statistically significant predictors of property crime.
In the model described in Equation (1), there may be a concern about reverse
causality. That is, there may be a concern that property crime, in fact, causes
vacancy as well the other way around, a problem referred to as simultaneity. Crime,
for example, may encourage home-owners to flee neighbourhoods without paying
382 C.M. Fuentes and V. Hernandez

Table 6. Summary of Moran’s I for vacant housing and homicide,

assault, burglaries in Ciudad Ju
arez, Chihuahua, Mexico
(2009
2010).

Burglaries

Moran’s I ¡.009755
Moran´s I expected ¡0.005882
z-score ¡0.193898
p-value 0.846256

Source: Own elaboration based on data from the State Secretary of Public
Safety 2010 and Population and Housing Census (Inegi, 2010).

off their loans, increasing vacant levels. If simultaneity occurs between crime and
vacancy rate, OLSs would not accurately measure the effect of vacancy on crime.
However, with a cross-sectional data it is almost impossible to make a strong casual
claim about vacancy and crime no matter what statistical tests are used.
Since the multiple regression model explains only 22.9% of the variance of the
relative risks of residential burglary, it is clear that there are other factors influenc-
ing the risk unaccounted for in the model. This can be attributed to the fact that the
estimated parameter represents global averages of processes that might exhibit a
substantial degree of spatial variation12 (Fotheringham, Charlton, & Brunsdon,
2002). Thus, some of the unexplained variance may be associated with the assump-
tion of spatial stationarity behind the multiple regression model.
One way of detecting the problem of misspecification of the relationships
described by the multiple regression model is to use the Moran’s I statistic (see,
e.g., Bailey & Gatrell, 1995; Fotheringham et al., 2002; Haining, 2003). This statis-
tic can be used to verify that the residuals from the global model are randomly dis-
tributed. The results from the regression model do not exhibit statistically
significant positive autocorrelation (Moran’s I statistic D
.009755) (see Table 6).
The spatial distribution of the results suggests that there are not local variations
(spatial nonstationarity) in the relationship between the relative risk rates of bur-
glaries and the contextual characteristics (Malczewski & Poetz, 2005).

Conclusions
Since housing reform in Mexico, the government has allowed the private sector of
homebuilders to take the leading role, with government agencies expanding lending
considerably, supporting private demand. This new policy has had some negative
outcomes such as the overproduction of housing developments and misallocation in
the distribution of government loans towards places with larger number of salaried
 International Journal of Housing Policy 383

employees (Monkkonen, 2011). The combination of both policies has produced

hundreds of vacant units most of them in the social interest housing developments.
Our study finds that higher housing vacancy rates lead to higher levels of prop-
erty crime at appreciated levels. A one point (0.01) increase in the percentage of
vacant housing is expected to increase the property crime in a tract by 0.84%, all
other things being equal. This finding suggests that housing vacancy rates may have
important social and economic consequences for neighbourhoods. An increase in
property crime is an important social and economic cost that must be incorporated
into policy making concerning real estate and mortgage lending policies and regula-
tion (Immergluck & Smith, 2006). Other variables also seem to have a positive
effect on property crime. For example, tertiary employment density, the proportion
of residents who are male aged 15
24 years and the percentage of residents who
have moved within 5 years.
Most of these findings concur with the US literature on foreclosure, that identify
a positive relationship between crime and foreclosure (Arnio, Baumr, & Wolff,
2012; Ellen et al., 2013; Immergluck & Smith, 2006; Williams et al., 2014). Our
study also provides evidence that the growth of housing vacancy and their concen-
tration in places with larger numbers of salaried employees, such as the manufactur-
ing centres in the north of the country (see Monkkonen, 2011), in recent years has
been specifically driven by the government loan system. Private homebuilders iden-
tify the existence of demand, and thus overbuild in certain cities, according to the
availability of accessible financial aid for buyers. The findings here suggest that
some reduction of the credit flow in places with large salaried employees may be
merited, ‘since it appears that they may be leading to social ills that impact those
beyond the lender or borrower’ (Immergluck & Smith, 2006, p. 866).
Arguably, housing finance reform needs to better align the roles of both private
homebuilders and public sector financing. On the one hand, given the fact that high
vacancy rates lead to property crime and not vice versa, it may be necessary to regu-
late the production of housing units in the private sector: that is, do not start to build
new units until all excess housing has been taken up. The areas of high vacancy/
oversupply coincide with areas of low population growth that suggests ongoing
issues of oversupply for many years. Areas of low vacancy will start to correct over
the coming few years as long as demand and supply are allowed to harmonise.
In addition, our findings provide some guidance for the design of crime preven-
tion measures and social housing development programmes. Indeed, despite data
limitations, the results have implications for urban planning and social development
policies. One suggestion would be that the government implements programmes to
rebuild and re-use vandalised housing units. These policies might help rebuild
social cohesion between residents and reduce the rate of vacancy. However, it may
also be necessary to regulate the private construction industry, including the urban
infrastructure near new housing developments to avoid negative outcomes, such as
384 C.M. Fuentes and V. Hernandez

the isolation of low-income housing developments from the rest of the city, which
create opportunities to commit different types of crimes.
The comparison in terms of public policy between US literature on foreclosure
and this paper is that in the US case, the problem of foreclosure has been tied to the
rise of higher risk subprime mortgage lending (Immergluck & Smith, 2006) pro-
vided by the private banks. In the Mexican case, the main issue is the (mis)alloca-
tion of government loans that disproportionately direct investments towards places
with larger number of salaried workers where the private homebuilders choose to
focus housing developments (Monkkonen, 2011). In both cases, the policy options
suggested concern improved regulatory regimes. However, there remains a debate
about whether increased regulation might possibly restrict access to credit for some
potential borrowers (Immergluck & Smith, 2006).

Acknowledgements
We would like to thank the two anonymous referees and the editor for helpful suggestions on
an earlier draft. The opinions, findings and conclusions or recommendations expressed in
this paper are those of the authors and do not necessarily reflect those of our institutions.

Notes
1. The INFONAVIT is the most important institution in the Mexican housing finance sys-
tem, it is a tripartite organization governed by worker, business and government
representatives.
2. FOVISSSTE is other large government lender which provides loans to government
employees, it operates in a manner similar to INFONAVIT, using funds from a 5% con-
tribution for each salaried employee of the government (Monkkonen, 2011).
3. Fovi is a trust fund of the Central Bank of Mexico.
4. Fund for Popular Housing (FONHAPO) provides subsidised loans for low-income
people.
5. FOVI was taken over by a newly created agency, the Federal Mortgage Society (SHF)
was a newly created agency a. Although the SHF manages the housing fund of the cen-
tral bank, it was created with the primary goal of developing a secondary mortgage mar-
ket by extending the guarantee programme and standardising initiation and the
administration of mortgages across agencies that originate loans (Babatz, 2004).
6. Limited Financial Societies (SOFOLES) is a recently created institution, a group of
non-depository lenders known as Limited Financial Societies, or SOFOLES. The
SOFOLES were created as part of NAFTA with funds from the Central Bank of Mex-
ico, the World Bank, and the United States, with the goal of enabling U.S. non-bank
financial institutions and Canadian investment societies to participate in the Mexican
mortgage market (Pickering, 2000).
7. In Mexico the Federal Housing Law (2010) defines ‘social interest housing’ as those
which the value of the construction do not exceed fifteen times the minimum wage at
year. Social housing can also corresponds to housing types defined by floor area: basic
housing (<30 m2) and social housing (up to 45 m2). Some authors such as Coulomb
 International Journal of Housing Policy 385

et al. (2009) define social interest housing as those homes financed by different govern-
ment lending agencies through subsidised mortgage loans.
8. In Mexico, small houses are defined by their surface area: basic housing (<30 m2) and
social housing (31 to 45 m2).
9. Ciudad Ju
arez is the largest city in the northern Mexican state of Chihuahua. It is a met-
ropolitan area located along the U.S.-Mexico border and is the twin city of El Paso,
Texas. It was the home of 1,332,131 residents in 2010 (39% of Chihuahua’s popula-
tion). The population quadrupled between 1960 and 2010, increasing from 276,995 to
1332,131 inhabitants (Fuentes & Pe~na, 2010).
10. A maquiladora is a labour-intensive assembly operation. In its simplest organizational
form, a Mexican maquiladora plant imports inputs from a foreign country—typically,
the United States—processes these inputs and ships them back to the country of origin
for finishing and sale.
11. Many high-income families fled to El Paso, Texas because of the high level of violence
and abandoned their houses. According to the El Paso Police Department, more than
13,000 residents of Ju
arez moved to El Paso.
12. The consequences of spatial autocorrelation are the same as those of time series auto-
correlation; the OLS estimators are unbiased but inefficient, and the estimates of the
variance of the estimators are biased (Dubin, 1998).

References
Akaike, H. (1981). Likelihood of a model and information criteria. Journal of Econometrics,
16, 3
14.
Arnio, A., Baumer, E., & Wolff, K. (2012). The contemporary foreclosure crisis and US
crime rates. Social Science Research, 41, 1598
1614.
Babatz, G. (2004). The role of the Sociedad Hipotecaria Federal in the development of the
Mexican mortgage market. Housing Finance International, 18(3), 42
51.
Bailey, T.C., & Gatrell, A.C. (1995). Interactive spatial data analysis. New York, NY: Wiley.
Banco Bilbao Vizcaya Argentaria. (2011). Situaci
on inmobiliaria en M
exico. M
exico:
Author.
Bottoms, A.E., & Wiles, P. (1988). Crime and housing policy: A framework for crime pre-
vention analysis. In T. Hope & M. Shaw (Eds.), Communities and crime reduction (pp.
84
98). London: HMSO.
Bowers, K., & Hirschfield, A. (1999). Exploring links between crime and disadvantage in
north-west England: An analysis using geographical information systems. International
Journal of Geographical Information Science, 13, 59
84.
Brantingham, P., & Brantingham, P. (1984). Patterns in crime. New York, NY: Macmillan.
Bursik, R.J. Jr., & Grasmick, H.G. (1993). Economic deprivation and neighborhood crime
rates, 1960
1980. Law and Society Review, 27, 263
268.
Cahill, M., & Mulligan, G. (2007). Using geographically weighted regression to explore local
crime patterns. Social Science Computer Review, 25(2), 174
193.
Carrasco, H. (2012). Viviendas abandonadas, una historia de crisis e inseguridad. El Diario
de Ju
arez, Ciudad Ju
arez [Vacant housing a story of crisis and insecurity]. Retrieved
from http://diario.mx/Print.2012-12-15_b43d8890/viviendas-abandonadas-una-historia
de crisis.pdf. Accessed in August 2013.
Ceccato, V., Haining, R., & Signoretta, P. (2002). Exploring offence statistics in Stockholm
City using spatial analysis tools. Annals of the Association of American Geographers,
92, 29
51.
386 C.M. Fuentes and V. Hernandez

Clark, W.A., & Hosking, P.L. (1986). Statistical methods for geographers. New York, NY:
John Wiley and Sons.
Cohen, L., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 44, 588
608.
Corman, H., & Mocan, N. (2005). Carrots, sticks and broken windows. The Journal of Law
and Economics, 48(1), 235
266.
Coulomb, R., Le
on, P., Puebla, C., & Castro, J. (2009). Efectos de la producci
on privada de la
vivienda de inter
es social en M
exico [Effects of the private social housing production in
Mexico]. Revista Tecnolog
ıa y Construcci
on, 25, 33
51.
Dubin, R. (1998). Spatial autocorrelation; a primer. Journal of Housing Economics, 7,
304
327.
Edwards, S. (1998). The Mexican peso crisis: How much did we know? When did we know
it? World Economy, 21(1), 1
30.
Ehrlich, I. (1975). On the relation between education and crime. In F.T. Juster (Ed.), Educa-
tion, youth and human behavior (pp. 313
337). New York: McGraw-Hill.
Ellen, I., Lacoe, J., & Sharygin, C. (2013). Do foreclosure cause crime? Journal of Urban
Economics, 74, 59
70.
Fotheringham, A.S., Brunsdon, C., & Charlton, M. (2002). Geographically weighted regres-
sion: The analysis of spatially varying relationships. Chichester: John Wiley &Sons.
Fuentes, C., & Hern
andez, V. (2013). Assessing spatial pattern of crime in Ciudad Ju
arez,
Chihuahua, Mexico (2009): The macrolevel, mesolevel and microlevel approaches.
International Journal of Criminology and Sociological Theory, 6(4), 242
259.
Fuentes, C., & Pe~ na, S. (2006). La pol
ıtica de vivienda en la frontera norte de M
exico [The
housing policy in the Northern border of Mexico]. In G. Ordo~ nez & M. Reyes (Eds.),
Los retos de la pol
ıtica social en la frontera norte [The challenges of the social policy in
the northern border] (pp. 45
63). Tijuana: Plaza y Valdez and Colef.
Fuentes, C., & Pe~na, S. (2010). Globalization and its effects on the urban socio-spatial struc-
ture of a transfrontier metropolis: El Paso, Tx-Ciudad Ju
arez, Chih.-Sunland Park, NM.
In K. Staudt, C. Fuentes, & J. Mon
arrez (Eds.), Cities and citizenship at the U.S.-Mexico
border: The Paso del Norte Metropolitan Region (pp. 32
45). New York, NY: Palgrave,
Macmillan.
Haining, R. (2003). Spatial Data Analysis: Theory and Practice. Cambridge, UK: Cambridge
University Press.
Haining, R., & Ceccato, V. (2005). Assessing the geography of vandalism: Evidence from a
Swedish city. Urban Studies, 42, 1637
1656.
Haining, R., Ceccato, V., & Kahn, P. (2007). The geography of homicide in Sao Paolo Brazil.
Environmental and Planning A, 39(7), 1632
1653.
Hartnagel, T.F. (2004). Correlates of criminal behaviour. In R. Linden (Ed.), Criminology: A
Canadian perspective (pp. 120
163). Toronto, ON: Thomson-Nelson.
Herbert, D.T., & Hyde, S.W. (1985). Environmental criminology: Testing some area hypoth-
eses. Transactions of the Institute of British Geographers, 10, 259
274.
Immergluck, D., & Smith, G. (2006). The impact of singles family mortgage foreclosure on
neighborhood crime. Housing Studies, 21(6), 851
866.
Instituto Municipal de Investigaci
on y Planeaci
on. (2010). Plan de desarrollo urbano [Urban
development plan]. Ciudad Ju
arez: Instituto Municipal de Investigaci
on y Planeaci

on.
Instituto Nacional de Estad
ısticas y Geograf
ıa. (2010). XII Censo de Poblaci
on y Vivienda
[Housing and population census], Aguascalientes: Instituto Nacional de Geograf
ıa y
Estad
ıstica.
Jang, S., & Johnson, B. (2001). Neighborhood disorder, individual religiosity and adolescent
use of illicit drugs: A test of multilevel hypothesis. Criminology, 39(1), 109
144.
 International Journal of Housing Policy 387

Kennedy, L.W., & Forde, D.R. (1990). Routine activities and crime: An analysis of victimi-
zation in Canada. Criminology, 28, 137
151.
Koening, D.J., & Linden, R. (2004). Conventional or “street” crime. In R. Linden (Ed.),
Criminology: A Canadian perspective (pp. 408
443). Toronto, ON: Thomson-Nelson.
Kohfeld, C.W., & Sprague, J. (1988). Urban unemployment drives urban crime. Urban
Affairs Quarterly, 24, 215
241.
Malczewski, J., & Poetz, A. (2005). Residential burglaries and neighborhood socioeconomic
context in London, Ontario: Global and local regression analysis. Professional Geogra-
pher, 57(4), 516
529.
Maya, E., & Cervantes, J. (1999). La producci
on de vivienda del sector privado y su prob-
lem
atica en el municipio de Ixtapaluca, Estado de M
exico [The production of housing
by the private sector and problems in the municipality of Ixtapaluca, State of Mexico].
Mexico City: Plaza, Vald
es y Facultad de Arquitectura, UNAM.
Maycotte, E., & S
anchez, E. (2009). Ciudades dispersas, viviendas abandonadas: La pol
ıtica
de vivienda y su impacto territorial y social en las ciudades mexicanas [Disperse cities,
vacant housing: The housing policy and its territorial and social impact in the Mexican
cities]. Paper presented in European Association of Urban Studies, Barcelona, Spain.
Retrieved from http://upcommons.upc.edu/revistes/bitstream/2099/1/04 PROCCIDINGS
M4 09 0025.pdf. Accessed in November 2012).
Miethe, T.D., & Meier, R.F. (1994) Crime and its social context. Albany, NY: State Univer-
sity of New York Press.
Miller, A.J. (1990). Subset selection in regression. New York, NY: Chapman and Hall.
Mon
arrez, J., & Cervera, L. (2013). Spatial and temporal behavior of three paradigmatic
cases of violence in Ciudad Ju
arez, Chihuahua M
exico: Feminicide, homicide and invol-
untary disappearances of girls and women (1993
2013) (Report presented to United
Nations Human Rights, No. 34). Ciudad Ju
arez: El Colegio de la Frontera Norte.
Monkkonen, P. (2008). Using online satellite imagery as a research tool: Mapping changing pat-
terns of urbanization in Mexico. Journal of Planning Education and Research, 28, 225
236.
Monkkonen, P. (2011). The housing transition in Mexico: Expanding access to housing
finance. Urban Affairs Review, 47(5), 672
695.
Neustrom, M.W., & Norton, W.M. (1995). Economic dislocation and property crime. Jour-
nal of Criminal Justice, 23(1), 29
39.
Paternoster, R., & Bushway, S.D. (2001). Theoretical and empirical work on the relationship
between unemployment and crime. Journal of Quantitative Criminology, 17, 391
407.
Pettiway, L. (1982). Mobility of robbery and burglary offenders: Ghetto and nonghetto
spaces. Urban Affairs Quarterly, 18, 255
270.
Pickering, N. (2000). The SOFOLES: Niche lending or new leaders in the Mexican mortgage
market. Joint Center Center for Housing Housing StudiesStudies. Cambridge, MA: Har-
vard University.
Pratt, T.C. (2001). Assessing the relative effects of macro-level predictors of crime: A meta-
analysis (PhD dissertation). The Division of Criminal Justice of the College of Educa-
tion, University of Cincinnati, OH.
Romo, L., C
ordova, G., Fuentes, C., & Brugu
es, A. (2012). La vivienda nueva en la ciudad de
Chihuahua: Oferta y demanda [The new housing in the city of Chihuahua: Supply and
demand]. Econom
ıa, Sociedad y Territorio, XII(40), 657
688.
Rountree, P.W., & Land, K.C. (2000). The generalizability of multilevel models of burglary
victimization: A cross-city comparison. Social Science Research, 29, 284
305.
Sampson, R. (1993). The community context of violent crime. In W. Wilson (Ed.), Sociology
and the public agenda (pp. 259
286). Newbury Park, CA: Sage.
388 C.M. Fuentes and V. Hernandez

Sampson, R., & Groves, B.W. (1989). Community structure and crime: testing social-disor-
ganization theory. American Journal of Sociology, 94, 774
802.
Sampson, R., & Raudenbush, S. (1999). Systematic social observation of public spaces; a
new look at disorder urban neighborhood. The American Journal of Sociology, 106,
603
651.
S
anchez, L., & Salazar, C. (2011). Lo que dicen las viviendas deshabitadas sobre el censo
de poblaci
on 2010 [The vacant housing units: The population and housing census].
Coyuntura demogr
afica, 1, 66
72.
Selvin, S. (1998). Modern applied biostatistical methods using S-plus. New York, NY:
Oxford University Press.
Silverman, E., & Della-Giustina, J. (2001). Urban policing and the fear of crime. Urban Stud-
ies, 38(5
6), 941
957.
Skogan, W. (1990). Disorder and decline: Crime and the spiral of decay in American neigh-
borhoods. Berkeley, CA: University of California Press.
Smith, T.S. (1976). Inverse distance variations for the flow of crime in urban areas. Social
Forces, 54, 802
815.
Spelman, W. (1993). Abandoned buildings: Magnets of crime? Journal of Criminal Justice,
21, 481
495.
Vel
azquez, S., & Mart
ınez, W. (2011). Viviendas deshabitadas efecto de la crisis econ

omica,
violencia e inseguridad en Ciudad Ju
arez [Vacant housing effect of the economic crisis,
violence and insecurity in Ciudad Ju
arez]. Retrieved from http://promepca.sep.gob.mx/
archivospdf/proyectos/proyecto211423.pdf. Accessed in August 2013
Vilalta, C. (2009). La geograf
ıa local del naromenudeo: Patrones, proceso y recomenda-
ciones de pol
ıtica urbana [The local geography of narcotics: Patterns, process and reco-
mendations of urban policy]. Estudios Demograficos y Urbanos, 24(1), 49
77.
Vilalta, C. (2010). The spatial dynamics and socioeconomic correlates of drug arrests in
Mexico City. Applied Geography, 30, 263
270.
Vilalta, C., & Muggah, R. (2012). Violent disorder in Ciudad Juarez: A spatial analysis of
homicide, discussion paper 1. Rio de Janeiro: HASOW.
Weisburd, D., & Mazerolle, L. (2000). Crime and disorder in drug hot spots: Implications for
theory and practice in policing. Police Quarterly, 3(3), 331
349.
Wilcox, P., Land, K.C., & Hunt, S.A. (2003). Criminal circumstance: A dynamic multicon-
textual criminal opportunity theory. New York, NY: Aldine de Gruyter.
Wiles, P., & Costello, A. (2000). The road to nowhere: The evidence for traveling criminals.
Home Office Research Study. London: Home Office.
Williams, S., Galster, G., & Verma, N. (2014). Home foreclosures and neighborhood crime
dynamics. Housing Studies, 29(3), 380
406.
Willson, J., & Kelling, G. (1982). Broken windows: The police and neighborhood safety.
Atlantic Monthly, 249(3), 29
39.
Wright, R., Logie, R., & Decker, S. (1995). Criminal expertise and offender decision-
making: An experimental study of the target selection process in residential burglary.
Journal of Research in Crime and Delinquency, 32, 39
53.