Transcript Challenging the idyll: Does crime affect property prices

Challenging the idyll: Does crime affect
property prices in small towns?
Vania Ceccato and Mats Wilhelmsson
Introduction
• There is no novelty in saying that crime
concentrates in urban environments.
• Rural municipalities are often regarded as idyllic
safe places; a retreat from the problems of big
cities, including crime.
• The problem is that far too often low crime rates
in rural areas are taken as a sign of there being
‘no problem’, or that, just because fewer offences
occur, crime does not affect people living there.
Literature review
Source
Case study
Effect of crime on prices or rent
Kain and Quigley (1970)
St. Louis, USA
No effect
Thaler (1978)
Rochester, New York, USA
Negative
Hellman and Naroff (1979)
Boston, UK
Negative
Rizzo (1979)
Chicago and Boston, USA
Negative
Dubin and Goodman (1982)
Baltimore metropolitan area, USA
Negative
Tita et al. (2006)
Columbus, USA
Munroe (2007)
Charlotte, NC, USA
Lynch and Rasmussen (2001)
Jacksonville, Florida, USA
Bowes and Ihlanfeldt (2001)
Atlanta, USA
Negative
Gibbons (2004)
London, UK
Negative
Ceccato and Wilhelmsson (2011)
Stockholm, Sweden
Negative
Ceccato and Wilhelmsson (2012)
Stockholm, Sweden
Negative
Inconclusive
Negative
No effect/
Positive effect
Aim
• The aim of this study is to assess whether
crime, particularly burglary, affects property
prices in a rural municipality.
Case study: Jönköping
• The municipality has a housing market that is sufficiently
large to allow a hedonic analysis of the impact of safety on
property prices.
• Although Jönköping can be classified as a middle large
municipality in terms of total population in Sweden, with
an important university, it is located geographically isolated
from the three main urban Swedish centers: Stockholm,
Gothenburg and Malmö.
• The municipality has excellent transport communications
but it is relatively distant from the main national urban
centers.
• In terms of safety, the number of reported crimes per 1000
inhabitants in Jönköping is lower than the national average,
a typical characteristic of rural municipalities in Sweden.
16000
14000
Offence by 100 000 inhabitants
12000
10000
Sweden
Jönköping municipality
8000
Theft - Jönköping
Violence - Jönköping
6000
4000
2000
0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
Figure 2 – Offences per 100,000 inhabitants, Sweden total, Jönköping total, theft and violence rates. Data source, BRÅ (2013).
Hedonic modelling
• Hedonic price modelling is traditionally used to
assess property values and one’s willingness to
pay for the property.
• The price of a property reflects attributes
associated with it, which can be of two types:
– those related to the property itself and
– those related to the environment in which the
property is located.
• Controlling: spatial dependency, endogeneity,
outliers
Hypotheses
1. Residential burglary negatively impacts apartment
prices after controlling for attributes of the property
and neighborhood characteristics.
2. Residential burglary affects different market segments
differently. Residential burglary will have a stronger
negative effect on high-priced apartments regardless
of year.
3. The effect of residential burglary on property prices
varies over time. Property prices will be more
negatively affected in areas that show higher
increases in crime rates (changes).
Data
• The estimation of the hedonic equation in this article is
based on two cross-sectional data sets that include
arm’s-length transactions of apartment sales in cooperative housing societies.
• Using Geographical Information Systems (GIS), the
apartment sales data have been merged together with
land use, demographic and socio-economic data from
Jönköping’s City planning office.
• Crime data for 2005 and 2011 were provided by the
Jönköping Police and contained the coordinates of
each address.
Results (2011)
Area
Fee
Room
Top floor
Road50m
Water50m
DistJönköping
DistHuskvar
Age40_64
Age65_84
AgeOlder_85
RBurgrate11
Constant
R2
Moran’s I
Coefficient
.8525
-.235
.161
.027
.008
.406
-.251
-.051
-.203
-.340
-.247
-.230
14.675
.697
20.000
t-value
11.38
-5.03
.022
1.24
0.34
10.50
-18.20
.030
-6.00
-8.33
-.5.23
-3.44
29.73
Robust results (2011)
Burglary
OLS
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
-1.8
-2
IV
Spatial lag
Spatial Error
Conclusion
1. Residential burglary negatively impacts apartment
prices after controlling for attributes of the property
and neighborhood characteristics.
• Findings show that residential burglary has a
significant negative effect on property prices in
Jönköping in 2011
Conclusion
2. Residential burglary affects different market
segments differently. Residential burglary will have
a stronger negative effect on high-priced
apartments regardless of year.
• The variable ‘burglary rates’ is slightly more
significant in for upper quantile prices than for lower
or mid quantiles.
Conclusion
3. The effect of residential burglary on property prices
varies over time. Property prices will be more
negatively affected in areas that show higher increases
in crime rates.
• The models based on change in crime rate as an
explanatory variable produce a poor goodness of fit, the
hypothesis of effect of residential burglary on property
prices over time was only partially tested.