Content uploaded by J. Peter Clinch
Author content
All content in this area was uploaded by J. Peter Clinch on Jul 09, 2019
Content may be subject to copyright.
ANALYSIS
Happiness, geography and the environment
☆
Finbarr Brereton, J. Peter Clinch⁎, Susana Ferreira
University College Dublin, Richview, Clonskeagh, Dublin 14, Ireland
ARTICLE INFO ABSTRACT
Article history:
Received 2 March 2007
Received in revised form 6 June 2007
Accepted 6 July 2007
In recent years, economists have been using socio-economic and socio-demographic
characteristics to explain self-reported individual happiness or satisfaction with life. Using
Geographical Information Systems (GIS), we employ data disaggregated at the individual
and local level to show that while these variables are important, consideration of amenities
such as climate, environmental and urban conditions is critical when analyzing subjective
well-being. Location-specific factors are shown to have a direct impact on life satisfaction.
Most importantly, however, the explanatory power of our happiness function substantially
increases when the spatial variables are included, highlighting the importance of the role of
the spatial dimension in determining well-being.
© 2007 Elsevier B.V. All rights reserved.
Keywords:
Subjective well-being
Spatial amenities
Geography
Environment
Geographical Information Systems
1. Introduction
The economics of happiness literature developed in the early
nineteen seventies with the pioneering work of such research-
ers as Richard Easterlin. Easterlin and subsequent authors,
such as Daniel Kahneman, believe that individual utility, tra-
ditionally thought by economists to be immeasurable and
hence proxied by income, can be measured directly. One
method is to employ happiness data from surveys as empirical
approximations of individual utility. The specific question
asked varies throughout the literature in terms of subject
matter (questions on happiness and life satisfaction are fre-
quently employed) and range of scale (three-point to ten-point
scales have been employed in the literature). These questions
elicit happiness or life satisfaction from individuals and mea-
sures such as these have been found to have a high scientific
standard in terms of internal consistency, reliability and vali-
dity (Diener et al., 1999)
1
and have been used extensively in the
economics literature in recent decades (see, e.g., Easterlin,
1974; 1995; 2001,orFrey and Stutzer, 2000; 2002a,b; 2004).
This literature has examined the role of socio-economic
and socio-demographic variables on individual well-being.
Established findings within the field include that character-
istics of the individuals themselves, their socio-demographic
characteristics, such as their age, gender and marital status,
influence their happiness. Similarly for micro-economic
characteristics, such as income, household tenure and em-
ployment status, with unemployment having a profound
negative influence on well-being. At the macro-economic
level, contributions have focused on the impact of national
ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
☆
Financial support of the Environmental Protection Agency
ERTDI program is gratefully acknowledged.
⁎Corresponding author. Tel.: +353 1 716 2751; fax: +353 1 716 2788.
E-mail addresses: finbarr.brereton@ucd.ie (F. Brereton),
peter.clinch@ucd.ie ( J.P. Clinch), susana.ferreira@ucd.ie
(S. Ferreira).
1
Firstly, measures of life satisfaction show temporal reliability,
even over a period of several years; secondly, they covary with
ratings made by family and friends, with interviewer ratings and
with amount of smiling in an interview; and finally, when self-
reports of well-being are correlated with other methods of
measurement, they show adequate convergent validity (Diener
and Suh, 1999).
0921-8009/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ecolecon.2007.07.008
available at www.sciencedirect.com
www.elsevier.com/locate/ecolecon
ARTICLE IN PRESSECOLEC-02889; No of Pages 11
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
inflation (Di Tella et al., 2001) and unemployment (Clark and
Oswald, 1994) rates and also the type of governance present in
the person's area (Frey and Stutzer, 2000). Happiness is found
to be inversely related to the inflation and unemployment
rates, but to increase with the level of direct democracy.
Prior literature in the economics field has demonstrated
that the area or location where an individual lives affects
quality of life. This is especially evident in the hedonic pricing
literature where there is a long tradition of constructing qua-
lity of life indices as the weighted averages of amenities in a
particular area, usually a city or region (see Rosen, 1974;
Roback, 1982 or Blomquist et al., 1988, for seminal contribu-
tions, and Chay and Greenstone, 2005, for a recent state-of-
the-art valuation exercise).
2
However, it wasn't until the 1990s that researchers began to
examine this spatial aspect of well-being in the economic
psychology literature. These more recent papers found that
characteristics of people's immediate surroundings (their
locality) influenced their well-being, but also that the wider
environment had an important role to play in explaining what
makes us happy. Environmental variables such as aircraft
noise (van Praag and Baarsma, 2005), air pollution (Welsch,
2002; 2006) and the prevailing climate (Frijters and van Praag,
1998 and Rehdanz and Maddison, 2005) are found to influence
welfare, as are environmental attitudes (Ferrer-i-Carbonell and
Gowdy, 2007). Findings indicate that excess noise levels ad-
versely affect well-being, as does air pollution and the influ-
ence of climate depends on the variable in question, indicating
the potential importance of spatial factors in determining well-
being.
In terms of examining the geography of well-being, pre-
vious studies were hindered by a lack of adequately disaggre-
grated data (Welsch, 2006; Rehdanz and Maddison, 2005). By
the authors own admission, data constraints at the local and
regional levels restricted their analysis to aggregated data at
the national level, or to focusing on a particular localised area
where richer data was available. Hence, thus far, the current
literature has stopped short of carrying out a holistic study of
the spatial element of well-being, due in no small part to these
data constrains, but also to the lack of availability of appro-
priate tools to carry out such analysis. For example, Rehdanz
and Maddison (2005) examine the influence on well-being of
climatic conditions, but including too many of their climate
variables in the model at once leads to problems of multi-
collinearity as some of their climate variables did not vary at
the national level (i.e., one record per country). They state that
their analysis was restricted to the country level and that it
would be interesting to see how climate would affect people's
happiness in different regions of a country. Ferrer-i-Carbonell
and Gowdy (2007) include a set of dummy variables indicating
the region where the individual lives to capture the (natural)
environment, proxying, for example, London and Manchester
as polluted areas. However, in the case of major cities in de-
veloped countries, pollution is, generally, a localised phenom-
enon and categorising an entire cities population under one
pollution level may severely under or overestimate their expo-
sure. Welsch (2006) uses life satisfaction scores to value air
pollution in European countries, but includes no within coun-
try variation in his estimation. Due to a lack of data, Welsch's
study was concerned with countries as the cross-sectional
units and he states that “future research may address the
question how regional or local happiness profiles are affected
by the corresponding environmental conditions. It is conceiv-
able that at a more disaggregated level the linkage between
environment and happiness is even more articulate than it is
with respect to national data”.van Praag and Baarsma (2005)
examine a localised problem and use postcodes to link their
respondents to objective noise burden, but due to issues of
anonymity, this application may only be available at city level
where populations are aggregated.
In this paper, we explicitly endeavour to examine the
importance of space in the determination of well-being, using
a more holistic approach. Firstly, we measure amenities at the
level of disaggregation at which individuals actually experi-
ence their surroundings, i.e. local level. This is facilitated
through the use of Geographical Information Systems (GIS), a
system for the visual display of spatial data. Using GIS, 1) the
level of disaggregation at which individuals are linked to their
surroundings is greatly improved; 2) the vector of spatial
variables included in the happiness function is expanded to
include variables with a potential influence on well-being, but
which have not been examined to date; and 3) distance mea-
sures are introduced, as one could hypothesize that the inten-
sity at which individuals experience their surroundings is a
function of proximity (as in the case of air pollution and noise).
The findings in the paper highlight the critical importance of
the role of the spatial dimension in determining well-being,
i.e., spatial variables are found to be highly significant with
large coefficients. We also find that the impact of spatial
amenities on life satisfaction is a function of distance, with the
most notable example being that of proximity to coast. This
has a large positive effect, which diminishes as one moves
further from the coast. The results may have potentially im-
portant implications for the setting of public policy, e.g. deci-
sions affecting the location of amenities negatively impacting
on well-being, such as waste facilities. Most importantly, the
explanatory power of our happiness function significantly
increases when the spatial variables are included, resulting in
three-times the variation in well-being being explained than
has been achieved in any previous cross-sectional study. This
indicates that geography and the environment have a much
larger influence on well-being than previously thought.
The paper proceeds as follows. Section 2 describes the
methodology (data, GIS requirements and the estimation stra-
tegy) used in the paper, Section 3 presents the results and
Section 4 concludes.
2
Roback (1982) found that the average person in her sample
would be willing to pay $69.55 per year for an additional clear day,
$78.25 per year to avoid an additional cloudy day, and $5.55 per
year to avoid an increase of 1 μg per cubic meter in particulate
matter. Blomquist et al. (1988) found that the difference in
compensation between the most and least desirable U.S. counties
in terms of the same bundle of local amenities comprising
climate, urban conditions and environmental quality was $5,146.
More recently, Berger et al. (2003) have shown that one standard
deviation changes in climate attributes (heating degree days), air
quality and crime produce annual compensation in the Russian
housing and labor markets of 7,839, 8,050 and 8,602 rubles
respectively, compared to a mean monthly salary of 1928 rubles.
2ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
2. Methodology
In this paper, we assume that the level of well-being attained
by an individual iin location kcan be represented by the
following indirect utility function:
ui;k¼aþb0xi;kþ0ai;kþ"i;ki¼1::::I;k¼1; ::::; Kð1Þ
where udenotes utility of individual iin location k,ais a vector
of spatial factors, some of which (e.g., commuting time, pro-
ximity toa coast) may vary at anindividual leveland
x
is a vector
of socio-economic anddemographic characteristics (age, gender
etc.) that are typically included in the literature (see, e.g., Clark
and Oswald, 1994; Di Tella et al., 2001 or Stutzer, 2004). In the
micro-econometric function, the individual's true utility is
unobservable, hence we use self-reported well-being as a proxy.
The well-being indicator (or proxy for individual utility) used
in this paper is based on the answers to the following question
(which was preceded by a range of questions regarding various
aspects of the respondent's life): ‘Thinking about the good and
bad things in your life, which of these answers best describes
your life as a whole?’. Respondents could choose a category on a
scale of one to seven(‘As bad as c an be’;‘very bad’;‘bad’;‘alright’;
‘good’;‘very good’;‘as good as can be’).
3
The use of self-reported
well-being introduces measurement error as the respondents
may be unable to communicate accurately their underlying
utility level. However, as Blanchflower and Oswald (2004a) point
out, it is measurement error in the independent variables that
would be more problematic in the econometric estimation, and
there is a broad consensus among previous studies that self-
reported well-being is a satisfactory empirical proxy of individ-
ual utility (see, e.g., Stutzer, 2004; Blanchflower and Oswald,
2004b; Ferrer-i-Carbonell and Frijters, 2004). Additionally, the
use of a latent variable framework (i.e. Ordered Probit) controls
for measurement error in the dependent variable.
Data on well-being and on the socio-demographic and
socio-economic characteristics used in the analysis come
from a survey
4
of a representative sample of 1,500
5
men and
women, aged 18 and over and living in Ireland. The survey
found a high well-being, in general, in Ireland with an average
of 5.5 on the seven-point scale. What makes this data set
particularly well suited for this paper is that it can be merged
with detailed geographical information as we know the area in
which the respondent lives. This information allows us to
match the survey data spatially to a national map of Ireland
using GIS and hence it is possible to combine subjective data at
the individual level with a vector of spatial amenities (a).
6
These two datasets are combined at the local (electoral
division)
7
level. However, to assess properly the impact on
individual well-being from changes in spatial amenities,
ideally, one would want to be able to match climate and
environmental factors to a particular individual rather than a
particular area. At present, however, the data do not allow this
and anonymity may preclude this in any case. Descriptions of
the variables and descriptive statistics are outlined in
Appendix A.
The use of data collected in Ireland is interesting in its own
right. In the last decade, the ‘Celtic Tiger’economy grew at a
record rate for a developed country (this and other trends are
documented in, for example, Clinch et al., 2002). Meanwhile,
the Economist Intelligence Unit (2004) has ranked Ireland as
first in its quality of life league table for 2005. Nevertheless,
there has been much concern regarding the implications of the
pace of economic growth for localized environmental quality
and life satisfaction generally (EPA, 2004). This makes Ireland
an appropriate subject for the analysis of the influence of
spatial amenities on subjective well-being. Furthermore,
issues surrounding heterogeneity of preferences may not be
as problematic in a small (approximately 70,000 km
2
) and
relatively homogenous country like Ireland, compared to other
nations. Also, by examining one country, issues of translation
and cultural bias in the well-being question should not arise.
8
As elements of the vector of spatial factors, we include
those variables which previous literature has shown to be
important determinants of well-being. Roback (1982) includes
the crime rate, population density and climatic conditions to
construct Quality of Life indices for US cities using the hedonic
pricing method. Blomquist et al. (1988) include similar
variables in their analysis, also of US cities, but in addition
they include the presence of waste facilities and proximity to
coast. More recently, research in the economics of happiness
has shown proximity to transport routes to affect well-being
(van Praag and Baarsma, 2005).
The dataset contains climate (from Collins and Cummins,
1996), environmental (from EPA, 2005) and other spatial data
(Urbis Database, 2006). Several climate variables were consid-
ered but following the advice of a climatologist, mean annual
precipitation, January mean daily minimum air temperature,
July mean daily maximum air temperature, mean annual
duration of bright sunshine and mean annual wind speed
3
Some studies treat self-reported life satisfaction data and
happiness data interchangeably. Veenhoven (1997) states that
“the word life-satisfaction denotes the same meaning and is often
used interchangeably with happiness.”Di Tella et al. (2001) report
a correlation coefficient of 0.56. However, Peiro (2006) points to
happiness and satisfaction as two distinct spheres of well-being.
He concludes that the first would be relatively independent of
economic factors while the second would be strongly dependent.
4
Urban Institute Ireland National Survey on Quality of Life (2001).
5
Due to missing observations the final sample consists of ap-
proximately (depending on the model specification) 1,467 observa-
tions. The effective response rate is 66.6%. The margin of error
using the entire sample is ± 2.5% at a 95% confidence level. The
2000 Register of Electors was used as the sampling frame.
6
GIS works well when applied to static data, and less well when
applied to time series analysis (Goodchild and Haining, 2004) and
hence is well-suited to the cross-sectional data employed in this
paper.
7
There are around 3,440 electoral divisions in Ireland which
represent the smallest enumeration area used by the Irish Central
Statistics Office in the collection of Census data. These areas are
relatively small, particularly in the city regions and those
represented in our sample range in size from 18ha (in cities) to
6,189ha (open countryside) (mean = 1,767, standard deviation =
1,538), with total populations ranging from 47 individuals to 8,595
(mean = 2,040, standard deviation = 2,073).
8
However, the extent to which these biases are problematic is a
matter of debate (Diener and Suh, 1999).
3ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
were chosen (similar to those included in Frijters and van
Praag, 1998).
As in Blomquist et al. (1988), variables capturing whether the
respondent lives near the coast, theviolent crime rate and pres-
ence of waste facilities in the respondent's area were included.
There is evidence suggesting that noise, smell and other
negative externalities from waste facilities may impact nega-
tively on well-being or quality of life (DG Environment, 2000). Air
pollution and water quality were considered as indicators of
environmental quality but regional variation is minimal (EPA,
2004). Additionally, population density (total population divided
by total area in km
2
(CSO, 2003)), traffic congestion and average
commuting timein each area were included tocapture crowding
and congestion effects. Also, a variable capturing voter turnout
in the Irish general election in 2002 (Kavanagh et al., 2004)is
included as an indicator for social capital (as in Putnam, 2000).
Due to data constraints, traffic congestion (number of vehicles
(DELG, 2002a,b) divided by the total length of primary roads per
local authority
9
area (NRA, 2003)) and thehomicide rate (number
of homicides per 100,000 of population (Garda Siochana, 2002))
are measured at the local authority level.
As in van Praag and Baarsma (2005), we include proximity to
airports.
10
However, we also include more detailed transport
data consisting of proximity to: major roads (national primary
and national secondary) (NRA, 2003); international, national
and regional airports; railway stations, and seaports (Urbis
Database, 2006). Access to transport routes could potentially
enter the micro-econometric function in two ways, positively
through accessibility and negatively through pollution and
noise. The latter was shown to be the case by van Praag and
Baarsma (2005) in relation to airport noise in Amsterdam.
As for the socio-economic and demographic variables, the
dataset includes an employment-status variable divided into
ten separate categories which follow the International Labour
Organisation (ILO) classification: employed (self-employed,
full-time employed and part-time employed), inactive (stu-
dent, working on home duties, disabled, retired, those not
working and not seeking work, and those on a government
training scheme) or unemployed (CSO, 2006). Unemployment
is further divided into two categories of those unemployed
having lost or given up their job combined with those not
working but seeking work, and those seeking work for the first
time. Additional individual characteristics contained in the
dataset and typically employed in the literature are age,
gender, educational attainment (primary, lower secondary/
junior high school, upper secondary/senior high school and
university degree), marital status (single, married, cohabiting,
widowed and separated/divorced), log of gross household
income,
11
whether the respondent is caring for a disabled
member of the family and the number of dependent children
in the household (1, 2, 3+). As an indicator of individual health
we use the number of times the respondent has visited the
doctor in the past year (never or once, two to five times and
six or more times a year). We also include household tenure
(owned outright, mortgaged, renting, or in public housing).
2.1. Geographical Information Systems methodology
GIS is a powerful computing tool that allows the visual repre-
sentation of spatially referenced data. It has advanced the tech-
nical ability to handle such data as countable numbers of points,
lines and polygons
12
in two-dimensional space (Goodchild and
Haining, 2004) and link various datasets using spatial identifiers
(Bond and Devine, 1991). It represents a solid base for spatial data
analysis and provides a range of techniques for analysis and
visualisation of spatial data. It provides effective decision support
through its database management capabilities, graphical user
interfaces and cartographic visualisation (Wu et al., 2001).
2.1.1. GIS in the economics literature
Research using GIS in the economics field has tended to be in
the area of environmental valuation through hedonic pricing
and a new generation of hedonic studies is using GIS to create
larger databases and define new explanatory variables in
combination with spatial econometric methods (see Bateman
et al., 2002; Lake et al., 1998). These hedonic models use a GIS
programme to develop neighbourhood characteristics that are
unique to each of their included observations (i.e. house or
property). GIS has enhanced the ability of these hedonic
models to explain variation in sale prices by considering both
proximity to, and extent of, environmental attributes (Paterson
and Boyle, 2002).
Baranzini and Ramirez (2005) use GIS to value the impact of
noise in Geneva, while Lynch and Rasmussen (2001) use GIS to
estimate the impact of crime on house prices in Jacksonville,
Florida, USA. Paterson and Boyle (2002) use GIS data to develop
variables representing the physical extent and visibility of
surrounding land use in a hedonic model of a rural/suburban
residential housing market. Bastian et al. (2002) use GIS data to
measure recreational and scenic amenities associated with
rural land, while Geoghegan et al. (1997) developed GIS data for
two landscape indices and incorporated them in a hedonic
model for Washington D.C, USA.
2.1.2. Creating variables using GIS
To capture accurately the influence of environmental and
location specific variables on individual well-being requires
variables to be measured at a high level of disaggregation i.e. at
9
For governance purposes, Ireland is divided into 34 different
regions called Local Authority areas. These generally equate to
one body per county and one for the three major urban areas of
Galway City, Limerick City and Cork City. Dublin is divided into
four areas and Tipperary is divided into two local authority areas.
These areas are relatively large and range in size from 2,035ha to
746,797ha (mean = 229,060, standard deviation = 226,508), with
total populations ranging from 25,799 individuals to 495,781
(mean = 177,377, standard deviation = 135,990).
10
All the proximity criteria are based on guidelines in Irish
Government policy documents (see, DELG, 2002b).
11
Income is expressed in thousands of euro. Missing values,
23.7% of those interviewed, were imputed based on the respon-
dent's socio-demographic characteristics including age, gender,
marital status, education level, area inhabited and employment
status. The original income variable was divided in 10 categories,
so mid-points were used (as in Stutzer, 2004). The survey was
carried out when Ireland was still using the Irish Pound, so we
converted to euros using the fixed rate of IR£1 = €1.26974.
12
A polygon is the GIS term for any multi sided figure.
4ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
the level at which individuals experience their environment.
Therefore they must be captured in a manner that reflects
individuals' perceptions of the amenity or disamenity in ques-
tion. Many facets of an amenity, such as intensity, frequency,
duration, variability, time of occurrence during the day etc.
(Bateman et al., 2001) will affect how an individual perceives
the amenity. GIS allows variables to be related spatially and
hence individuals can be linked to the geographic character-
istics of their surroundings. Hence, GIS could, in principle,
provide a full quantitative description of overall area quality if
all relevant data layers, for example concerning road networks
and public services, were available and were transformed in a
convenient way into spatial attributes (Din et al., 2001).
However, when specific household or property GeoCodes
(X,Ycoordinates) are unknown, as in the case of the house-
hold survey data used in this paper, neighbourhood areas
must be used as the reference point when creating environ-
mental variables. The typical method of doing this is to use
the mathematically-created centre or ‘centroid’of the area in
question
13
(as was the case in Craglia et al., 2001,whostudy
high intensity crime areas in England) and in this paper
we use the centroid of the respondents' electoral division.
This introduces a maximum measurement error equal to the
greatest distance between the centroid and the border of the
electoral division in question which will be greatest in rural
areas and smallest in the city regions.
The GIS requirements for this paper included the collection,
assimilation and pre-processing of digital, spatial datasets,
development of methods for spatio-temporal analysis and
production of summary statistics and cartographic representa-
tions. Thisprocess producedlayers of data which were ‘mapped’
into ArcView GIS. The data were entered into GIS as points (e.g.
the location of waste facilities), lines (e.g. roads), or polygons
(e.g. airports) within the categories of: meteorological; environ-
mental; transport; and administrative boundary data layers.
Different variables were entered in different ways. Some were
entered directly as the spatial coordinates for this data were
known, such as the airport co-ordinates. Others, such as the
climate layers were entered as raster maps and these were
converted to polygons for analysis purposes, as it was then
possible to link individuals to characteristics of their areas. All
data were converted to Irish National Grid co-ordinates.
Once the data layers were entered into the ArcView system,
variables were created to allow statistical analysis to take
place. For example, proximity to coast is measured as three
dummy variables; less than two kilometres from the coast,
between two and five kilometres and more than five kilo-
metres. This allows us to examine if the amenity/disamenity
values of the variables are functions of distance. We can also
disaggregate between different types of similar amenities e.g.
landfill and hazardous waste sites (EPA, 2005). Using proximity
tools within ArcMap, distance ‘buffers’were created from the
centroid (as in Craglia et al., 2001) of each specific electoral
division to a specified distance. Buffer analysis allows the
researcher to take a point or line feature and generate a
polygon containing all the area within a certain distance of the
feature (Bond and Devine, 1991). A tool called ‘select by
location’, was then used to identify the area where a particular
environmental condition is satisfied. The variables created
were either entered as columns of 0s and 1s, i.e. where the
dummy equaled 1 for a particular electoral division if the
condition was satisfied and 0 otherwise (e.g., 1 if an electoral
division was within a 50 km radius of an airport and 0
otherwise) or as continuous variables (as in the case of the
climate variables). These variables were then exported to the
statistical software package STATA so econometric analysis
could be carried out.
2.2. Estimation strategy
The stated aim of this paper is to examine the influence of
space and place on individual well-being. As a first step
towards capturing this influence, a micro-econometric happi-
ness function is specified (Model 1) in which we distinguish
between two distinct geographical areas of Ireland, i.e.,
between those respondents living in Dublin and those living
in the rest of the country. This split was considered appropriate
in a small (approximately 70,000 km
2
) and relatively homog-
enous country like Ireland where the Dublin area comprises
28% of the population in only 1.3% of the land area, accounts for
39% of the national total of Gross Value Added and, with a
population of 1.122 million, is the only urban area with a
population in excess of 150,000. In Model 1, which also controls
for a broad range of socio-economic and socio-demographic
characteristics of the individuals in question (age, age-
squared, gender, employment status, educational attainment,
health, marital status, income, number of dependent children
and household tenure), a dummy for Dublin might be seen as a
rough summary measure of the amenities in that area.
However, it does not provide much information regarding
which specific amenities are most valued by the individuals.
Therefore, in order to determine which site-specific factors are
most relevant to well-being, a subsequent model is estimated
(Model 2), corresponding to the estimation of Eq. (1), where the
spatial variables equate to the amenities contained in vector a.
This model contains the spatial amenities created using GIS
and other data at the electoral division level.
Finally, because the regressions combine data at different
levels of disaggregation (individual, electoral division and
local authority levels), the standard errors in all the regres-
sions are corrected for clustering (Moulton, 1990).
3. Results —assessing the importance of location
3.1. Model 1
Table 1 shows the results from the estimation of our models.
Following the recent literature (e.g., Ferrer-i-Carbonell and
Gowdy, 2007) and given the ordered nature of our dependent
13
A‘centroid' is the mathematical term for the centre of an area,
region, or polygon, calculated from points on its perimeter. In the
case of irregularly shaped polygons, the centroid is derived
mathematically and is weighted to approximate a ‘centre of
gravity.' These discrete X–Ylocations are often used to index or
reference the polygon within which they are located and some-
times attribute information is ‘attached,' ‘hung,' or ‘hooked' to the
centroid location.
5ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
Table 1 –Ordered probit regressions/dependent variable
‘life satisfaction’
Variable name Model 1 Model 2
Age Age 0.0188 0.0100
(1.45) (0.74)
Age-squared −0.0002 −0.0001
(1.46) (0.76)
Gender (female) Male −0.1665⁎⁎ −0.1719⁎⁎
(2.43) (2.20)
Employment status
(self employed)
Retired 0.0871
(0.54)
0.0350
(0.25)
Engaged in
home duties
−0.3941⁎⁎⁎
(3.22)
−0.2817⁎⁎
(2.19)
Student −0.1990 0.0251
(1.13) (0.10)
Seeking
work for 1st
time
−0.2090
(0.59)
−0.2061
(0.55)
Unemployed −0.9182⁎⁎⁎ −0.8674⁎⁎⁎
(4.26) (3.94)
Not working,
not seeking
work
−1.4317⁎⁎⁎
(3.95)
−1.4863⁎⁎⁎
(3.96)
Working
full-time
−0.1280
(1.26)
−0.0460
(0.50)
Working
part-time
−0.3695⁎⁎⁎
(2.80)
−0.2597⁎
(1.82)
Government
scheme
−0.6624⁎⁎⁎
(2.61)
−0.9309⁎⁎
(2.54)
Permanently
unable to work
−0.4888
(1.61)
−0.6247⁎
(1.94)
Education (primary) Lower
secondary/
junior high
school
0.4210⁎⁎⁎
(3.68)
0.3023⁎⁎
(2.24)
Upper
secondary/
senior high
school
0.1764⁎
(1.69)
0.1940⁎
(1.75)
Degree 0.0617 0.1590
(0.52) (1.15)
Health (visited the doctor
0 or 1 in the last year)
2–5 doctor visits −0.1555⁎⁎
(2.46)
−0.2224⁎⁎⁎
(2.61)
6 or more
doctor visits
−0.3851⁎⁎⁎
(3.10)
−0.4252⁎⁎⁎
(3.04)
Marital status (single) Married −0.0138 0.0720
(0.15) (0.74)
Co-habiting −0.1239 −0.2596
(0.83) (1.18)
Widowed 0.0880 0.1124
(0.57) (0.69)
Separated and
divorced
−0.3762⁎⁎
(2.04)
−0.1981
(1.00)
Log income Income (1000 s) 0.2103⁎⁎⁎ 0.2649⁎⁎⁎
(2.90) (2.95)
Number of children in the
household (no children)
1 Child 0.0215
(0.20)
−0.1197
(0.93)
2 Children −0.0829 −0.1111
(0.87) (1.09)
3 or more
children
−0.1772⁎
(1.89)
−0.1838⁎
(1.94)
Household tenure
(own outright)
Own with a
mortgage
−0.0194
(0.27)
0.0156
(0.20)
Rent privately 0.0342 −0.0033
(0.27) (0.02)
Public housing −0.5125⁎⁎⁎ −0.4781⁎⁎⁎
(4.69) (3.61)
Table 1 (continued)
Variable name Model 1 Model 2
Respondent is a carer 0.3314⁎0.2313
(1.70) (1.24)
Dublin dummy variable −0.7527⁎⁎⁎ −0.4430
(11.79) (1.12)
Spatial variables No Yes
Climate variables Precipitation 0.0005
(1.28)
Wind speed −0.3815⁎⁎
(2.36)
January
minimum
temperature
0.8082⁎⁎⁎
(3.33)
July maximum
temperature
0.0806⁎⁎⁎
(3.85)
Average
annual
sunshine
(hours)
−0.0011
(1.22)
Average commuting time 0.0057
(0.48)
Population density 0.0061⁎
(1.92)
Congestion −0.0001
(1.17)
Homicide rate 0.0570
(0.97)
Voter turnout 0.0160⁎
(1.84)
Proximity to landfill
(more than 10 km)
Contains
a landfill
−0.5145⁎
(1.87)
Within 3 km 0.4332
(1.55)
Between 3
and 5 km
0.2998
(0.95)
Between 5 and
10 km
−0.2359
(1.40)
Proximity to hazardous
waste facility (more than
10 km)
Contains a
hazardous
waste facility
−0.4190
(0.71)
Within 3 km −0.1993
(0.54)
Between 3
and 5 km
−0.3983
(1.01)
Between 5
and 10 km
−0.2888
(0.89)
Proximity to coast (more
than 5 km)
Within 2 km 1.1299⁎⁎⁎
(4.25)
2 to 5 km 0.2761
(1.34)
Proximity to beach
(more than
10 km)
Within 5 km −0.2248
(0.73)
Between 5 and
10 km
−0.1910
(0.62)
Proximity to rail station
(morethan10km)
Within 2 km −0.2868
(1.28)
Between 2
and 5 km
−0.3531
(1.37)
Between 5
and 10 km
−0.0391
(0.14)
Proximity to airport (more than 60 km)
Regional Within 30 km 1.2726⁎⁎⁎
(2.63)
Between 30 and
60 km
0.0543
(0.27)
6ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
variable, it contains results from ordered-probit regressions.
14
The reference groups for the independent variables are in
parentheses.
The results on the socio-economic and socio-demographic
characteristics in Model 1 are, broadly speaking, in line with
previous findings in the economic psychology literature. For
example, the coefficient on being unemployed is negative and
significant and, everything else being equal, reduces life
satisfaction substantially (see e.g., Blanchflower and Oswald,
2004a for similar results). Gender is significant and negative,
indicating that males are less satisfied with their lives than
females. Except for the study of Alesina et al. (2004) that finds
gender to be significantly related to life satisfaction in the USA,
in previous studies gender tends to emerge as insignificant in
life satisfaction regressions (Stutzer, 2004; Frey and Stutzer,
2000; Di Tella et al., 2001). We find that those with lower (junior
high school) or higher (senior high school) education are more
satisfied with life than those with a primary education level
(similar to Frey and Stutzer, 2000). As in Clark and Oswald
(1994) and Blanchflower and Oswald (2004b), being separated
or divorced is negative and significant. However, we find no
difference between married and single respondents. Having
three or more children is negative and significant at the 5%
level (similar to Clark and Oswald, 1994). Respondents visiting
their doctor two or more times a year are found to be less
satisfied with life than those not attending or attending only
once. Living in public housing is significant and negatively
related to life satisfaction at the 1% level with a large
coefficient. Perhaps surprisingly, being the carer of a disabled
family member emerges as positive and significant in the
regression. In line with the standard textbook prediction of
utility as an increasing function of income, our proxy for utility
(life satisfaction) is an increasing function of (log) income,
which emerges significant at the 1% level. Age emerges
insignificant in the regression. This is in contrast to the inter-
national literature which, generally, finds a U-shaped associ-
ation between life satisfaction and age.
Examining the influence of location on well-being, we find
the coefficient on the dummy variable for Dublin to be highly
significant and large; only the coefficients for being unem-
ployed and a discouraged worker are larger in magnitude.
Everything else being equal, those living in all areas outside
Dublin have a higher life satisfaction. This result is similar to
that in Ferrer-i-Carbonell and Gowdy (2007), who find indivi-
duals living in Inner London to be less happy, everything else
equal.
Having controlled for a large number of socio-economic and
socio-demographic characteristics, a reasonable hypothesis is
that factors related to the size of the settlement and other
location-specific factors may be responsible for lower life-
satisfaction levels in Dublin. For example, compared to any
other area in the country, unparalleled growth rates have
resulted in the capital having a much higher population
density than other areas and a significant traffic congestion
problem (DELG, 2002b). To test this hypothesis, Model 2
examines the importance of spatial amenities.
3.2. Model 2
Model 2, the results of which are reported in the last column of
Table 1, corresponds to Eq. (1). It builds on Model 1 by including
the variables with a spatial influence on well-being. These
include population density, congestion, commuting time and
the climatic and environmental variables. In this model, the
dummy for Dublin loses its significance. This result suggests
that the spatial variables explain an important part of the
difference between living in Dublin and other regions of
Ireland in terms of well-being.
15
The pseudo-R
2
of Model 2, at 0.16, exceeds all those obtained to
date in the international literature usinga cross-sectional dataset.
For example, Ferrer-i-Carbonell and Gowdy (2007) in their study of
subjective well-being and environmental attitudes, obtain a
pseudo-R
2
of 0.088. Another barometer of the explanatory power
of the model is the adjusted-R
2
in Table 2 reporting the OLS
results. The adjusted-R
2
increasesfrom0.21inModel1to0.33in
Model 2, which compares very favourably with those obtained in
other studies; Stutzer (2004) for example, in his analysis of Swiss
cantons, obtains an R
2
of 0.11. Since we control for similar socio-
economic and demographic characteristics of the individual as in
other studies of this nature, we believe this high adjusted-R
2
highlights the substantial influence of spatial amenities as
determinants of well-being.
16
Table 1 (continued)
Variable name Model 1 Model 2
National Within 30 km 0.1404
(0.40)
Between 30
and 60 km
0.5408
(1.55)
International Within 30 km 0.4294
(1.56)
Between 30
and 60 km
0.5371⁎⁎
(2.16)
Proximity to major road
(more than 5 km)
Contains a
major road
−0.6040⁎⁎
(1.97)
Within 5 km −0.5816⁎
(1.79)
Proximity to sea ports
(more than 10 km)
Within 3 km −0.5826
(1.63)
Between 3
and 5 km
0.0023
(0.01)
Between 5 and
10 km
0.2877
(0.85)
Number of observations 1467 1464
Likelihood Ratio −1845.59 −1692.85
Pseudo R
2
0.09 0.16
Note 1: ⁎significant at 10% level; ⁎⁎significant at 5% level; ⁎⁎⁎signif-
icant at 1% level.
Note 2: t-statistics in parentheses computed using White's Hetero-
skedasticity-corrected standard errors.
14
We also estimate OLS regressions (Table 2) and the results are
comparable.
15
We also estimate Model 2 without the Dublin dummy variable
and the results are almost identical (results available on request
from the authors).
16
Additional R
2
obtained in the literature include Blachflower and
Oswald (2004b) at 0.10, Di Tella et al. (2001) at 0.17 and Blanchflower
and Oswald (2004a) at 0.084. However, these papers use pooled data
over a number of years and hence, may not be directly comparable.
Proximity to airport (more than 60 km)
7ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
Table 2 –OLS regressions/dependent variable ‘life
satisfaction’
Variable name Model 1 Model 2
Age Age 0.0155 0.0070
(1.45) (0.70)
Age-squared −0.0002 −0.0001
(1.46) (0.75)
Gender (female) Male −0.1369⁎⁎ −0.1232⁎⁎
(2.45) (2.08)
Employment status
(self employed)
Retired 0.0677
(0.52)
0.0407
(0.39)
Engaged in
home duties
−0.3142⁎⁎⁎
(3.17)
−0.1940⁎⁎
(2.00)
Student −0.1520 0.0149
(1.06) (0.08)
Seeking work
for 1st time
−0.1706
(0.59)
0.0491
(0.21)
Unemployed −0.7810⁎⁎⁎ −0.6746⁎⁎⁎
(4.17) (3.56)
Not working,
not seeking
work
−1.1726⁎⁎⁎
(3.85)
−1.0821⁎⁎⁎
(3.60)
Working
full-time
−0.0994
(1.23)
−0.0306
(0.44)
Working
part-time
−0.3026⁎⁎⁎
(2.77)
−0.1628
(1.53)
Government
scheme
−0.5330⁎⁎
(2.48)
−0.6725⁎⁎
(2.39)
Permanently
unable to
work
−0.4181
(1.61)
−0.4681⁎
(1.82)
Education (primary) Lower
secondary/
junior high
school
0.3331⁎⁎⁎
(3.50)
0.2038⁎⁎
(2.05)
Upper
secondary/
senior high
school
0.1502⁎
(1.68)
0.1284
(1.47)
Degree 0.0582 0.0914
(0.58) (0.87)
Health (visited the doctor
0or1inthelastyear)
2–5 doctor
visits
−0.1348⁎⁎
(2.53)
−0.1720⁎⁎⁎
(2.63)
6 or more
doctor visits
−0.3149⁎⁎⁎
(3.04)
−0.3249⁎⁎⁎
(3.07)
Marital status (single) Married 0.0030 0.0671
(0.04) (0.91)
Co-habiting −0.0696 −0.1435
(0.55) (0.87)
Widowed 0.0773 0.0945
(0.61) (0.78)
Separated
and divorced
−0.3193⁎
(1.93)
−0.1537
(0.93)
Log income Income (1000 s) 0.1722⁎⁎⁎
(2.93)
0.2164⁎⁎⁎
(3.49)
Number of children in the
household (no children)
1 Child 0.0160
(0.18)
−0.0920
(0.96)
2 Children −0.0828 −0.1204
(1.02) (1.49)
3 or more
children
−0.1620⁎⁎
(2.06)
−0.1653⁎⁎
(2.24)
Household tenure
(own outright)
Own with a
mortgage
−0.0061
(0.10)
0.0149
(0.25)
Rent privately 0.0309 0.0100
(0.30) (0.08)
Public housing −0.4379⁎⁎⁎
(4.81)
−0.3594⁎⁎⁎
(3.52)
Table 2 (continued)
Variable name Model 1 Model 2
Respondent is a carer 0.2632⁎0.2371⁎
(1.73) (1.91)
Dublin dummy variable −0.6222⁎⁎⁎ −0.2434
(11.43) (0.78)
Spatial variables No Yes
Climate variables Precipitation −0.0003
(1.05)
Wind speed −0.2459⁎⁎
(2.13)
January
minimum
temperature
0.5558⁎⁎⁎
(3.19)
July
maximum
temperature
0.0543⁎⁎⁎
(3.77)
Average
annual
sunshine
(hours)
−0.0011⁎
(1.74)
Average commuting time 0.0034
(0.41)
Population density 0.0038
(1.63)
Congestion −0.0001
(1.36)
Homicide rate 0.0501
(1.06)
Voter turnout 0.0124⁎⁎
(2.12)
Proximity to landfill
(more than 10 km)
Contains
a landfill
−0.3736⁎
(1.90)
Within 3 km 0.2646
(1.32)
Between 3
and 5 km
0.2564
(1.05)
Between 5
and 10 km
−0.1346
(1.07)
Proximity to hazardous
waste facility (more
than 10 km)
Contains a
hazardous
waste facility
−0.2068
(0.47)
Within 3 km −0.1715
(0.64)
Between 3
and 5 km
−0.2998
(1.03)
Between 5
and 10 km
−0.1560
(0.66)
Proximity to coast (more
than 5 km)
Within 2 km 0.8351⁎⁎⁎
(4.32)
2 and 5 km 0.2271
(1.51)
Proximity to beach
(more than 10 km)
Within 5 km −0.1607
(0.73)
Between 5
and 10 km
−0.0923
(0.42)
Proximity to rail station
(more than 10 km)
Within 2 km −0.1705
(1.07)
Between 2
and 5 km
−0.2271
(1.22)
Between 5
and 10 km
−0.0142
(0.07)
Proximity to airport (more than 60 km)
Regional Within 30 km 0.8329⁎⁎⁎
(2.78)
Between 30 and
60 km
0.0284
(0.21)
8ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
Of the climate variables, the coefficient on mean annual
precipitation is positive indicating that, for Irish people,
increased rainfall slightly increases life satisfaction. This result
may, however, be driven by a positive correlation between rain
and scenic beauty.
17
The most spectacular landscapes in
Ireland are found in the wettest counties in the West of
Ireland. Rehdanz and Maddison (2005) find very scarce pre-
cipitation reduces happiness, which they hypothesize might
reflect the fact that climate could have an indirect effect on
happiness through landscape effects. However, in our case the
coefficient emerges insignificant at conventional levels.
Increases in the January minimum and July maximum tem-
peratures emerge as amenities and increase life satisfaction.
Wind speed emerges negative and significant in our regres-
sion, while surprisingly, we find that total annual sunshine is
negatively related to life satisfaction. However, it may be that
this result is driven by the correlation between elements of
rainfall not captured in our variable (e.g., intensity and
frequency) and sunshine.
As in the hedonic literature (e.g., Blomquist et al., 1988), we
find the presence of waste facilities in an individual's area to
be a disamenity. However, the type of, and distance from, the
waste facility in question matters. The coefficient on the
variable capturing if a landfill site is in operation in the
respondent's electoral division emerges negative and signif-
icant compared to those who live in electoral divisions more
than ten kilometres away. There is evidence suggesting that
noise, smell and other negative externalities from waste
facilities of this kind may impact negatively on well-being or
quality of life (DG Environment, 2000). Proximity to landfill
sites has been the subject of many hedonic analyses, such as
Blomquist et al. (1988) who use the number of landfill sites per
capita as a variable to construct Quality of Life indices, also
Nelson et al. (1992) and Havlicek (1985) who examine the price
effects of landfills on house values in US cities. Proximity to a
hazardous waste facility however, does not seem to have an
influenceintermsoflifesatisfaction.Itmaybethat
individuals are less aware of the presence of these facilities
in their areas. The coefficient on population density is positive
and significant at the 10% level. This result is similar to that of
Roback (1982), who finds population density to be an amenity.
Average commuting time and congestion emerge insignificant
in the regression as does the crime rate.
Proximity to coast emerges positive and significant with a
large coefficient, indicating that individuals living near the
coast enjoy higher life satisfaction, other things being equal.
Additionally there is evidence that the utility value of coast is a
function of distance with respondents living two kilometres or
less from the coast more satisfied with their lives, compared to
those living more than five kilometres from the coast. Those
living between two and five kilometres from the coast are also
more satisfied, if insignificantly so, but the coefficient is
reduced. Interestingly, proximity to beach emerges insignifi-
cant in the regression. It may be that, given Ireland's climate,
the amenity value of coastal areas is not a function of the
availability of a beach.
We find access to transport emerges as both an amenity and
disamenity, depending on the type of, and distance from, the
amenity in question. Life satisfaction is highest for those living
between thirty and sixty kilometres from an international
airport. It may be that those less than thirty kilometres away
are affected by the noise disamenity. In relation to regional
airports, the amenity value lies at less than thirty kilometres.
This result is not unexpected as these are smallairports and only
deal with smaller, less noisy aircraft and would have signifi-
cantly fewer arrivals and departures than do the larger airports.
Close proximity to a major road (less than five kilometres)
emerges as a disamenity,again with distance decay. This may be
capturing the noise affects of this transport route. Close
proximity to a seaport emerges insignificant in the regression.
4. Conclusion
In this paper we adopt a holistic approach to the examination
of the influence of geography and the environment on
happiness. Using GIS we are able to overcome many of the
difficulties that have prevented previous researchers addres-
sing this issue comprehensively. This is achieved by matching
individuals to their surroundings at a higher level of disag-
gregration and by expanding the vector of spatial variables
included in the happiness function. We also use proximity
measures to examine if the influence of spatial amenities on
life satisfaction is a function of distance.
The findings show that climate has a significant influence on
well-being, with wind speed negative and significant, but
increases in both January minimum temperature and July
maximum temperature are positive and significant. Access to
major transport routes and proximity to coast and to waste
facilities all influence well-being. However, the manner in which
Table 2 (continued)
Variable name Model 1 Model 2
National Within 30 km 0.0721
(0.28)
Between 30 and
60 km
0.3383
(1.44)
International Within 30 km 0.2603
(1.30)
Between 30 and
60 km
0.3851⁎⁎
(2.18)
Proximity to major road
(more than 5 km)
Contains a
major road
−0.3703⁎
(1.83)
Within 5 km −0.3543
(1.62)
Proximity to sea ports
(more than 10 km)
Within 3 km −0.3887
(1.42)
Between 3 and 5
km
0.0019
(0.01)
Between 5 and
10 km
0.2054
(0.75)
Number of observations 1467 1451
Adjusted R
2
0.21 0.33
Note 1: ⁎significant at 10% level; ⁎⁎ significant at 5% level;
⁎⁎⁎ significant at 1% level.
Note 2: t-statistics in parentheses computed using White's Hetero-
skedasticity-corrected standard errors.
17
A high correlation coefficient is observed between precipita-
tion and presence of Natural Heritage Areas (0.5874), the latter
being EU-designated areas of outstanding natural beauty.
9ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
they enter the happiness equation differs depending on the
amenity in question. Proximity to landfill is found to have a
negative affect on well-being. Proximity to coast has a large
positive effect, but its influence is a diminishing function of
distance. Additionally, the impact of proximity to major
transport routes has different effects depending on the type of,
and distance to, the amenity in question, e.g., while reasonable
proximity to international airports increases well-being, close
proximity to major roads decreases it. It may be that, in the
former case, the positive effect of access outweighs the negative
effect of noise, while the opposite may be true in thelatter case.
These results may have potentially important implications for
the setting of public policy, such as the location of waste
facilities, the routing of major roads, location of airports etc., so
as to have as minimal negative impact as possible on well-being.
Our findings highlight the critical importance of the role of the
spatial dimension in determining well-being, i.e., spatial vari-
ables are found to be highly significant with large coefficients. In
fact, the explanatory power of our happiness function substan-
tially increases when the spatial variables are included, resulting
in three-times the variation in well-being being explained than
has been achieved in any previous cross-sectional study. This
indicates that geography and the environment have a much
larger influence on well-being than previously thought, as im-
portant as the most critical socio-economic and socio-demo-
graphic factors, such as unemployment and marital status. This
finding has potentially important implications for setting
priorities for public policy as, in essence, improving well-being
could be considered to be the ultimate goal of public policy.
Acknowledgements
We thank Andrew Oswald, Michael Hanemann, Mirko Moro,
Alan Carr and two anonymous referees for helpful comments.
We thank Daniel McInerney and Sean Morrish for technical
assistance. The usual disclaimer applies.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.ecolecon.2007.07.008.
REFERENCES
Alesina, A., Di Tella, R., MacCulloch, R., 2004. Inequality and
happiness: are Europeans and Americans different? Journal of
Public Economics 88, 2009–2042.
Baranzini, A., Ramirez, J.V., 2005. Paying for quietness: the impact
of noise on Geneva rents. Urban Studies 42 (4), 633–646.
Bastian, C.T., McLeod, D.M., Germino, M.J., Reiners, W.M., Blasko,
B.J., 2002. Environmental amenities and agricultural land
values: a hedonic model using geographic information systems
data. Ecological Economics 40, 337–349.
Bateman, I.J., Jones, A.P., Lovett, I.R., Day, B.H., 2002. Applying
Geographical Information Systems (GIS) to environmental and
resource economics. Environmental and Resource Economics
22, 219–269.
Bateman, I., Day, B., Lake, I., Lovett, A., 2001. The effect of road
traffic on residential property values: a literature review and
hedonic price study. Scottish Executive Transport Research
Series. The Stationary Office, Edinburgh.
Berger, M.C., Blomquist, G.C., Sabirianova, P.K., 2003. Compensat-
ing Differentials in Emerging Labour and Housing Markets:
Estimates of Quality of Life in Russian Cities. IZA Discussion
Paper No, 900. October.
Blanchflower, D.G., Oswald, A.J., 2004a. Well-being over Time
in Britain and the USA. Journal of Public Economics 88,
1359–1386.
Blanchflower, D.G., Oswald, A.J., 2004b. Money, sex and happiness:
an empirical study. Scandinavian Journal of Economics 106 (3),
393–415.
Blomquist, G.C., Berger, M.C., Hoehn, J.P., 1988. New estimates of
quality of life in urban areas. American Economic Review 78 (1),
89–107.
Bond, D., Devine, P., 1991. The role of geographic information
systems in survey analysis. The Statistician 40 (2), 209–216.
Chay, K.Y., Greenstone, M., 2005. Does air quality matter: evidence
from the housing market. Journal of Political Economy 113 (2),
376–424.
Clark, A.E., Oswald, A.J., 1994. Unhappiness and unemployment.
Economic Journal 104, 648–659.
Clinch, J.P., Convery, F., Walsh, B.M., 2002. After the Celtic Tiger:
Challenges Ahead. O’Brien Press, Dublin.
Collins, J.F., Cummins, T. (Eds.), 1996. Agroclimatic Atlas of Ireland:
Working Group on Applied Meteorology. University College
Dublin, Ireland.
Craglia, M., Haining, R., Signoretta, P., 2001. Modelling high-
intensity crime areas in english cities. Urban Studies 38 (11),
1921–1941.
CSO, 2003. Census 2002: Population Classified by Area, volume 1.
Central Statistics Office, Dublin.
CSO, 2006. Quarterly National Household Survey. Central Statistics
Office, Dublin. Feb.
DELG, 2002a. Irish Bulletin of Vehicle and Driver Statistics.
Stationary Office, Dublin. Dec.
DELG, 2002b. The National Spatial Strategy 2002–2020. Stationary
Office, Dublin. Dec.
Diener, Ed., Suh, M.E., 1999. National differences in subjective
well-being. In: Kahneman, D., Diener, E., Schwarz, N. (Eds.),
Well-being: The Foundations of Hedonic Psychology. Russell
Sage Foundation, New York.
Diener, E.D., Suh, E.M., Lucas, R.E., Smith, H.L., 1999. Subjective
well-being: three decades of progress. Psychological Bulletin
125, 276–302.
Din, A., Hoesli, M., Bender, A., 2001. Environmental variables and
real estate prices. Urban Studies 28 (11), 1989–2000.
Di Tella, R., MacCulloch, R.J., Oswald, A.J., 2001. Preferences over
inflation and unemployment: evidence from surveys of
happiness. American Economic Review 91, 335–341.
DG Environment, 2000. A Study on the Economic Valuation of
Environmental Externalities from Landfill Disposal and Incin-
eration of Waste, Final Main Report, European Commission,
Brussels October.
Easterlin, R.A., 1974. Does economic growth improve the human
lot? In: David, Paul A., Reder, Melvin.W. (Eds.), Nations and
Households in Economic Growth: Essays in Honor of Moses
Abramovitz. Academic Press, New York.
Easterlin, R.A., 1995. Will raising the incomes of all increase the
happiness of all? Journal of Economic Behaviour and
Organisation 27, 35–47.
Easterlin, R.A., 2001. Income and happiness: towards a unified
theory. Economic Journal 111, 465–484.
Economist Intelligence Unit's Quality of Life Index, 2004. The
Economist, London.
EPA, 2004. Ireland's Environment. Environmental Protection
Agency, Wexford.
EPA, 2005. National Waste Database. Environmental Protection
Agency, Wexford.
10 ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
Ferrer-i-Carbonell, A., Frijters, P., 2004. How important is
methodology for the estimates of the determinants of
happiness? Economic Journal 114 (497), 641–659.
Ferrer-i-Carbonell, A., Gowdy, J.M., 2007. Environmental degrada-
tion and happiness? Ecological Economics 60 (3), 509–516.
Frey, B.S., Stutzer, A., 2000. Happiness, economy and institutions.
Economic Journal 110, 918–938.
Frey, B.S., Stutzer, A., 2002a. Happiness and Economics. Princeton
University Press, Princeton.
Frey, B.S., Stutzer, A., 2002b. What can economists learn from
happiness research? Journal of Economic Literature 40 (2),
402–435.
Frey, B.S., Stutzer, A., 2004. Stress that Doesn't Pay: The Com-
muting Paradox. IZA Discussion Paper No, 1278.
Frijters, P., van Praag, B.M.S., 1998. The effects of climate on
welfare and well-being in Russia? Climate Change 39, 61–81.
Garda, S., 2002. Annual Report. Garda Siochana, Dublin.
Geoghegan, J., Wainger, L.A., Bockstaell, N.E., 1997. Spatial
landscape indices in a hedonic framework: an ecological
economics analysis using GIS. Ecological Economics 23,
251–264.
Goodchild, M.F., Haining, R.P., 2004. GIS and spatial data analysis:
converging perspectives. Papers in Regional Science 83, 363–385.
Havlicek Jr., J., 1985. Impacts of solid waste disposal sites on
property values. In: Tolley, G.S., Havlicek Jr., J., Favian, R. (Eds.),
Environmental Policy: Solid Waste, vol IV. Ballinger, Cam-
bridge. MA.
Kavanagh, A., Mills, G., Sinnot, R., 2004. The geography of Irish
voter turnout: a case study of the 2002 general election. Irish
Geography 37 (2), 177–186.
Lake, I.R., Lovett, A.A., Bateman, I.J., Langford, I.H., 1998. Modeling
environmental influences on property prices in an urban
environment. Computer,Environment and Urban Systems 22 (2),
121–136.
Lynch, A.K., Rasmussen, D.W., 2001. Proximity, neighbour and the
efficiency of exclusion. Urban Studies 41 (2), 285–298.
Moulton, B.R., 1990. An illustration of a pitfall in estimating the
effects of aggregate variables on micro units. Review of
Economics and Statistics 72 (2), 334–338.
Nelson, A.C., Genereux, J., Genereux, M., 1992. Price effects of
landfills on house values. Land Economics 68 (4), 365–395.
NRA, 2003. National Route Lengths. National Road Authority,
Dublin.
Paterson, R.P., Boyle, K.J., 2002,. Out of sight, out of mind? using GIS
to incorporate visibility in hedonic property value models.
Land Economics 78 (3) , 417–425.
Peiro, A., 2006. Happiness, satisfaction and socio-economic
conditions: some international evidence. Journal of Socio-
Economics 35, 348–365.
Putnam, R.D., 2000. Bowling Alone: The Collapse and Revival of
American Community. Touchstone, New York.
Rehdanz, K., Maddison, D., 2005. Climate and happiness.
Ecological Economics 52, 111–125.
Roback, J., 1982. Wages, rents and the quality of life. Journal of
Political Economy 90 (6), 1257–1278.
Rosen, S., 1974. Hedonic prices and implicit markets: product
differentiation in pure competition. Journal of Political
Economy 82, 34–55.
Stutzer,A.,2004.Theroleofincomeaspirationsinindividual
happiness. Journal of Economic Behaviour and Organization
54 (1), 89–109.
Urban Institute Ireland National Survey on Quality of Life, 2001.
Urban Institute Ireland, Dublin.
Urbis Database, 2006. Urban Institute Ireland, Dublin.
van Praag, B.M.S., Baarsma, B.E., 2005. Using happiness surveys to
value intangibles: the case of airport noise. Economic Journal
115, 224–246.
Veenhoven, R., 1997. Advances in the understanding of happiness.
Revue Québécoise de Psychologie 18 (2), 29–74.
Welsch, H., 2002. Preferences over prosperity and pollution:
environmental valuation based on happiness surveys. Kyklos
55 (4), 473–494.
Welsch, H., 2006. Environment and happiness: valuation of air
pollution using life satisfaction data. Ecological Economics 58,
801–813.
Wu, Y.-H., Miller, H.J., Hung, M.-C., 2001. A GIS-based decision
support system for analysis of route choice in congested urban
road networks. Journal of Geographical Systems 3, 3–24.
11ECOLOGICAL ECONOMICS XX (2007) XXX–XXX
ARTICLE IN PRESS
Please cite this article as: Brereton, F. et al. Happiness, geography and the environment. Ecological Economics (2007),
doi:10.1016/j.ecolecon.2007.07.008
