# RECORD 3e2f420e36b96c93f3636cc35296dc7c69414687
# TITLE: Effect of Air Quality Improvements on Apartment Prices in Seoul: Application of Spatial Hedonic Price Model and Interpolation Method
# YEAR: 2013.0  VENUE: nan  DOI: 10.2457/srs.43.163
# AUTHORS: K. Nam; Jiao Son; Brian H. S. Kim
# PAGES: 13  CANDIDATE TABLE PAGES (1-based): [5, 8, 9, 11]
# ABSTRACT: nan


===== PAGE 1 (score 15, numeric-paren 0, table-words 0, var-words 15) =====

Studies in Regional Science, Vol. 43, No. 2, 2013, 163-175.
163
Effect of Air Quality Improvements on Apartment Prices in Seoul :  
Application of Spatial Hedonic Price Model and Interpolation Method
Kichan Nam*, Ji-Wan Son** and Brian H.S. Kim***
Abstract
    Using a spatial hedonic price model, this study estimated the elasticity and the implicit marginal costs of 
variables related to air pollutants that may affect apartment prices in Seoul.  Spatial econometric and interpo-
lation methods were employed to estimate the benefits of changes in the concentrations of air pollutants.   
Previous studies showed that dependence, spatial heterogeneity, and air pollution concentrations can be 
addressed with the use of these methods.  The average marginal willingness to pay with a 10% improvement 
in PM10 concentration is approximately 233,800 won (US$187.0) to 303,800 won (US$243.0) per month.   
With a healthy concentration (18.15% reduction), the average marginal willingness to pay is approximately 
424,300 won (US$339.4) to 551,400 won (US$441.1) per month.  Therefore, environmental factors can be 
important determinants of housing prices in Seoul.
JEL Classification : R00, R15, Q53
Keywords : Spatial Hedonic Price Model, Spatial Interpolation Model, Criteria Air Pollution, Marginal Willing-
ness to Pay
1.  Introduction
Cities are major political, economic, social, and cultural centers.  However, people have become 
increasingly concerned with the pace of growth and the development of cities.  Heavy population con-
centrations and the increase in the number of conurbations have resulted in undesirable environmental 
and health consequences.  Rapid development has also transformed many urban areas into metropoli-
tan areas with diverse cultures, ethics, and infrastructures.  Natural habitats and green spaces are dis-
appearing, along with problems in air and water pollution, diseases, and heat island effect.  Therefore, 
the concept of environmentally sound and sustainable development has received growing interest.
With a population of more than 10 million people, Seoul confronts many environmental and public 
health problems.  Air quality is one of Seoul’s main environmental problems.  The city suffers from 
poor air quality and poor visibility because of the smog formed from suspended particulate matter 
throughout the year.  Ozone (O3) levels, which have also been increasingly high, are way above those 
allowed according to Seoul’s ambient air standards.  The rapid increase in motor vehicles since the 
late 1980s has gradually changed the major pollutants in Seoul from sulfur dioxide (SO2) to nitrogen 
  　* Korea Research Institute for Human Settlement, 254 Simin-daero, Dongan-gu, Anyang-si, Gyeonggi-do, 431-
712, Korea
  ** Kyeryong Construction Industrial Co., LTD, Wolpyeongjung-ro 21 Beon-gil, Seo-gu, Daejeon, 302-717, Korea
*** Department of Agricultural Economics and Rural Development, Research Institute for Agriculture and Life Sci-
ences Seoul National University, 1 Gwanangno, Gwanak-gu, Seoul 151-921, Korea
　  E-mail : briankim66@snu.ac.kr  Phone : 82-2-880-4717  Fax :  82-2-873-5080
Received : 23rd  August, 2012  Accepted : 8th April, 2013
Ⓒ日本地域学会(JSRSAI) 2013


===== PAGE 5 (score 23, numeric-paren 0, table-words 4, var-words 10) =====

167
Effect of Air Quality Improvements on Apartment Prices in Seoul
Table 1.  All Variables
Variables
Name
Description
Unit
Dependent Variable
PART
Price of apartment
million won
Independent
Variables
Housing
M2
Area of apartment
m2
HEAT_D
Heat type
(1=individual heating, 0=central heating)
Dummy
CORRI_D
Corridor type
(1=stairway type, 0=corridor type)
Dummy
BATH
The number of bath
#
Complex
YEARS
Apartment age
Year
YEARS2
Re-development index
Year2
SIZE
The size of complex — total number of apartment 
units in the complex
Household
BRAND_D
Brand of apartment
(1=well known brand, 0=the others)
Dummy
Neighborhood
SCHOOL
College entrance rate
%
SUBWAY
Distance to nearest station
meter
GRDPC
Gross regional domestic product per capita
million won
Environment
SO2
Sulfur Dioxide
ppm
NO2
Nitrogen Dioxide
ppm
O3
Ozone
ppm
PM10
Particulate Matter
μm/m3
CO
Carbon Monoxide
ppm
Table 2.  Descriptive Statistic
Variables
Name
Mean
Min,
Max.
Stand.
Deviation
Dependent Variable
PART
599.57
150.00
3,200.00
371.37
Independent
Variables
Housing
M2
113.27
53.00
212.00
23.37
HEAT_D
0.69
–
–
0.46
CORRI_D
0.88
–
–
0.32
BATH
1.81
1
3
0.40
Complex
YEARS
9.46
0
33
7.07
YEARS2
139.36
0
1089
205.92
SIZE
750.57
63
5540
822.61
BRAND_D
0.35
–
–
0.48
Neighborhood
SCHOOL
0.73
0.60
0.81
0.05
SUBWAY
800.92
70
6027.00
886.97
GRDPC
2.59
2.18
3.56
0.30
Environment
SO2
0.01
0.01
0.01
0.001
NO2
0.02
0.02
0.02
0.002
O3
0.04
0.03
0.04
0.003
PM10
61.08
55.48
68.96
3.34
CO
0.68
0.46
0.82
0.06


===== PAGE 6 (score 5, numeric-paren 0, table-words 0, var-words 5) =====

Kichan Nam, Ji-Wan Son and Brian H.S. Kim
168
m2).  With regard to the apartment’s number of years, the sample (9.46 years) had only a year differ-
ence with Seoul’s average (10.53 years).  The state of the education system also strongly affects 
apartment price in Korea.  One of the indicators that reflect this is the college entrance rate of the 
region.  The selected sample household’s college entrance rate is about 73%.  Other important inde-
pendent variables are distance to public transportation and per capita GRDP (about 800 meters and 2.6 
million won, respectively, for sampled households).
Criteria air pollutants are used to reflect the regional environmental variable.  Figure 2 shows 
that each criteria air pollutant is diversely distributed in SMA.  The average concentration for these 
air pollutants is 0.007 ppm for sulfur dioxide (SO2), 0.02 ppm for nitrogen oxide (NO2), 0.038 for ozone 
(O3), 0.684 ppm for carbon monoxide (CO), and 61.084 μm/m3 for particulate matter (PM10).  These 
 
 
(a) SO2 
(b) O3 
 
 
(c) NO2 
(d) CO 
 
 
SO2 
 
O3 
 
NO2 
 
CO 
 
PM10 
 
(e) PM10 
 
Figure 2.  Distribution of criteria air pollutants in Seoul (Spatial Interpolation Result)


===== PAGE 8 (score 118, numeric-paren 90, table-words 4, var-words 2) =====

Kichan Nam, Ji-Wan Son and Brian H.S. Kim
170
Table 4.  Regression Results
Model
Base Model
Spatial Error Model (Distance=5 km)
OLS Robust
Estimation
OLS
ML
GM
(2-step)
GM (Iterated)
White Variance
λ
0.970
(0.018)
**
0.983
−
**
1.120
−
**
Constant
9.998
(0.322)
**
9.811
(0.322)
**
9.588
(0.623)
**
9.926
(0.340)
**
9.998
(0.335)
**
M2
0.013
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
HEAT_D
−−0.148
(0.029)
**
−0.129
(0.024)
**
−0.131
(0.024)
**
−0.123
(0.023)
**
−0.163
(0.034)
*
CORRI_D
0.178
(0.041)
**
0.139
(0.033)
**
0.146
(0.033)
**
0.136
(0.033)
**
0.167
(0.049)
*
BATH
−0.026
(0.034)
0.001
(0.029)
0.004
(0.029)
0.005
(0.028)
−0.034
(0.035)
YEARS
−0.031
(0.006)
**
−0.029
(0.005)
**
−0.031
(0.005)
**
−0.030
(0.005)
**
−0.033
(0.006
**
YEARS2
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
SIZE
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
BRAND_D
0.122
(0.029)
**
0.116
(0.024)
**
0.111
(0.024)
**
0.112
(0.024)
**
0.113
(0.027)
**
SCHOOL
0.782
(0.215)
**
0.550
(0.202)
**
0.513
(0.202)
*
0.520
(0.203)
*
0.697
(0.205)
SUBWAY
−0.001
(0.001)
**
−0.001
(0.001)
*
−0.001
(0.001)
*
−0.001
(0.001)
*
−0.001
(0.001)
**
GRDP
0.005
(0.001)
**
0.002
(0.001)
**
0.002
(0.001)
**
0.002
(0.001)
**
0.005
(0.001)
**
SO2
−54.235
(12.619)
**
−36.103
(12.490)
**
−36.375
(12.493)
**
−35.100
(12.514)
**
−52.734
(12.636
**
NO2
−41.089
(6.793)
**
−8.081
(8.254)
−4.468
(8.157)
−4.761
(8.338)
−42.995
(6.658)
**
O3
−4.916
(4.919)
−1.486
(5.683)
−1.301
(5.706)
−0.173
(5.784)
−3.981
(4.808)
*
PM10
−0.014
(0.003)
**
−0.012
(0.003)
**
−0.012
(0.003)
**
−0.010
(0.004)
**
−0.013
(0.003)
**
CO
0.473
(0.212)
0.191
(0.275)
0.152
(0.275)
0.142
(0.278)
0.413
(0.218)
Adj.R2/Sq. Corr.
0.786
0.854
0.737
0.734
0.786
AIC
−55.414
−197.433
−42.790
Breusch-Pagan
67.751**
59.679**
LM-LAG
(Robust-LAG)
165.338
(52.525)
**
**
LM-ERR
(Robust-ERR)
436.450
(323.637)
**
**
* : 5% significant level ; ** : 1% significant level ; ( ) : Standard error


===== PAGE 9 (score 142, numeric-paren 109, table-words 4, var-words 3) =====

171
Effect of Air Quality Improvements on Apartment Prices in Seoul
Table 5.  Sensitivity Analysis : Distance Effect
Distance
Weight Matrix=1.7 km
Weight Matrix=10 km
Weight Matrix=20 km
Model
Spatial Error Model
Spatial Error Model
Spatial Lag Model
Estimation
ML
GM
(2-step)
ML
GM
(2-step)
ML
IV
(2SLS)
ρ
0.174
(0.290)
−2.285
(0.399)
**
λ
0.834
(0.024)
**
0.631
−
**
0.972
(0.024)
**
1.031
−
**
Constant
9.222
(0.404) 
**
9.221
(0.382)
**
10.147
(0.476)*
10.965
(0.413)
**
8.153
(3.100)
**
34.214
(4.241)
**
M2
0.010
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
0.011
(0.001)
**
HEAT_D
−0.039
(0.022)
**
−0.063
(0.023)
**
−0.171
(0.026)
**
−0.173
(0.026)
**
−0.162
(0.041)
**
−0.178
(0.028)
**
CORRI_D
0.098
(0.029)
**
0.100
(0.031)
**
0.152
(0.037)
**
0.161
(0.037)
**
0.167
(0.041)
**
0.178
(0.040)
**
BATH
−0.034
(0.023)
0.024
(0.026)
0.003
(0.032)
0.005
(0.032)
−0.036
(0.034)
−0.018
(0.033)
YEARS
−0.024
(0.004)
**
−0.026
(0.005)
**
−0.032
(0.006)
**
−0.033
(0.006)
**
−0.033
(0.006)
**
−0.034
(0.006)
**
YEARS2
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
SIZE
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
0.001
(0.001)
**
BRAND_D
0.101
(0.021)
**
0.098
(0.022)
**
0.129
(0.026)
**
0.124
(0.026)
**
0.114
(0.029)
**
0.101
(0.028)
**
SCHOOL
0.527
(0.244)
*
0.580
(0.240)
*
0.822
(0.190)
**
0.758
(0.186)
**
0.681
(0.213)
**
0.917
(0.211)
**
SUBWAY
0.001
(0.001)
−0.001
(0.001)
*
−0.001
(0.001)
**
−0.001
(0.001)
**
−0.001
(0.001)
**
−0.001
(0.001)
**
GRDP
0.003
(0.001)
**
0.004
(0.001)
**
0.004
(0.001)
**
0.003
(0.001)
**
0.005
(0.001)
**
0.005
(0.001)
**
SO2
6.352
(13.978)
−15.283
(13.802)
−51.402
(11.155)
**
−55.429
(11.014)
**
−53.455
(12.472)
**
−43.373
(12.306)
**
NO2
−8.522
(9.304)
−21.450
(8.510)
*
−24.520
(6.394)
**
−22.118
(6.338)
**
−44.275
(6.963)
**
−26.219
(7.190)
**
O3
−3.978
(6.379)
1.040
(5.813)
−4.310
(4.749)
−4.758
(4.731)
−4.596
(4.951)
4.066
(4.958)
PM10
−0.006
(0.003)
−0.007
(0.003)
*
−0.021
(0.003)
**
−0.022
(0.003)
**
−0.013
(0.003)
**
−0.011
(0.003)
**
CO
0.245
(0.294)
0.089
(0.271)
0.144
(0.217)
0.178
(0.217)
0.454
(0.213)
*
−0.114
(0.224)
Adj.R2/Sq. Corr.
0.893
0.765
0.822
0.768
0.790
0.806
AIC
−300.176
−111.892
−63.135
Breusch-Pagan
51.830**
73.349**
73.127**
* : 5% significant level ; ** : 1% significant level ; ( ) : Standard error


===== PAGE 10 (score 9, numeric-paren 0, table-words 3, var-words 3) =====

Kichan Nam, Ji-Wan Son and Brian H.S. Kim
172
　　　　　　　　　　　　　　　　 　　
ln
dy
d
y
dx
dy
b
= 
(7)
　　　　　　　　　　　　　　　　　 　　dx
dy
y
b
=

(8)
　　　　　　　　　　　　　　　　　　
y y
x
x
h
b
b
=
=

(9)
The elasticity of spatial-lag model can be estimated as follows4 :
　　　　　　　　　　　　　　　ln PART
I
W
x
v
1
t
b
=
-
+
-
^
h
6
@

(10)
　　　　　　　　　　　　　　
ln
dx
d
PART
I
W
dx
dx
1
t
b
=
-
-
^
h
6
@

(11)
　　　　　　　　　　　　 　
dx
d PART
PART
I
W
1
1
$
t
b
=
-
-
^
h
6
@

(12)
　　　　　　　　　　　　
dx x
d PART
PART
I
W
x
1
h
t
b
=
=
-
-
^
^
e
h
h o
6
@

(13)
　　　　　　　　　　　　　　　　　　
I
x
1
h
t
b
=
-
-
6
@

(14)
The elasticity of spatial-error model can be estimated as follows :
　　　　　　　　　　　　　　　　ln PART
X
W
e
b
m
=
+
+
^
h

(15)
　　　　　　　　　　　　　　 　　
ln
dx
d
PART
dx
dx b
=
^
h

(16)
　　　　　　　　　　　　　　 　
ln
dx
d
PART
PART
$
b
=
^
^
h
h
(17)
　　　　　　　　　　　　　　　　　　　　
X
h
b
=

(18)
Spatial dependence generally affects the independent variables in SAR.  Considering the effect of spa-
tial dependence is important, but this paper uses Equation (14) for SAR and Equation (18) for SEM to 
estimate elasticity.
The value of air quality reflects the marginal willingness to pay (MWTP) of the improvements in 
air pollution concentration.  MWTP is calculated with the estimated elasticity of each air pollution 
source (Table 6) and the improvement level based on air quality standards of SMA.
PM10 is one of the most highly concentrated pollutants among criteria air pollutants (Table 
3).  The concentration level of PM10 needs to decrease by 11.084 μg/m3 (18.15%) to meet Seoul’s air 
quality standards.  Therefore, MWTP is estimated with the use of the elasticity of PM10 of each model 
in Tables 4 and 5 with the following cases : (1) reduction in PM10 concentration by 10% and 18.15%, (2) 
6.5% discount rate, and (3) apartment lifespan of 30 years5.  Table 7 shows the estimated average 
4　The functions of  I
W
i
I
i
1
1
t
t
-
=
-
-
^
c
h
m   is used for the conversion from Equation (13) to Equation (14).
5	 Since this study evaluates the value of environmental improvements on housing price, it is reasonable to use 
social discount rates that are generally used at the national level.  The Korea Development Institute (KDI), 
which is the main economic research institution in Korea, used the social discount rate of 6.5% for the years of 
2004 to 2008.  Also, Seoul city government’s current regulation to permit renewal and/or reconstruction for the 
apartment after 20 to 40 years of lifespan depending on the year it was built.  Therefore, the median of 30 years 
is used for lifespan of the apartments.


===== PAGE 11 (score 11, numeric-paren 0, table-words 3, var-words 3) =====

173
Effect of Air Quality Improvements on Apartment Prices in Seoul
monthly and annual MWTP6.  The estimated average MWTP for a 10% reduction in PM10 concentra-
tion is about 233,800 won (US$187.0) to 303,800 won (US$243.0) per month.  This range of values is 
relatively high compared with those found in previous studies that use OLS, but it remains reasonable 
in consideration of the spatial effects of air pollutants.
5.  Conclusions and Implications
Growing urban environment, public health, and housing problems have made people more sensi-
tive with issues on air pollution.  People’s choice of housing location and price could be based on the 
concentration of criteria air pollutants.  This study focuses on the evaluation of criteria air pollutant 
6	 The capital recovery factor, which is defined below, is used in the estimation of MWTP :
　　　　　　　　　　　　　　　　　CRF
P
R
r
r
r
1
1
1
N
N
=
=
+
-
+
^
^
h
h
	 where r is the discount rate, N is the life span period, P is the present value, and R is the equivalent price per 
year.
Table 6.  Elasticity Estimation of PM10
Mean
Coefficient
elasticity
OLS
61.084
−0.014
−0.855
W=5 km
ML
−0.012
−0.733
SEM GM
(2 Step)
−0.012
−0.733
SEM GM
(iterated)
−0.010
−0.611
OLS Robust
(White variance)
−0.013
−0.794
W=1.7 km
SEM GM
(2 step)
−0.007
−0.428
W=10 km
SEM GM
(2 step)
−0.022
−1.344
W=20 km
IV (2SLS)
−0.011
−0.205
Table 7.  Marginal Willingness to Pay
(Unit : 1,000 won)
Present Value
Ave. per year
Ave. per month
10%
18.15%
10%
18.15%
10%
18.15%
W=5 km
SEM GM
(2 step)
43,948.3
79,766.2
3,365.5
6,108.3
280.5
509.0
SEM GM
(iterated)
36,633.6
66,490.0
2,805.3
5,091.6
233.8
424.3
OLS Robust
47,605.7
86,404.3
3,645.5
6,616.6
303.8
551.4


===== PAGE 12 (score 9, numeric-paren 0, table-words 0, var-words 9) =====

Kichan Nam, Ji-Wan Son and Brian H.S. Kim
174
concentration and how this factor affects the price of apartments in SMA.  The marginal value of WTP 
for changes in criteria air pollutants is estimated with spatial hedonic price model.
The most highly concentrated criteria pollutant in Seoul is PM10.  A 10% decrease in PM10 con-
centration could be worth about 233,800 won (US$187.0) to 303,800 won (US$243.0) per month.  This 
range of values is based on estimated MWTP value for 30 years of apartment life span.  For a healthy 
level of PM10 concentration (reduction of 18.15%), the expected MWTP should be about 424,300 won 
(US$339.4) to 551,400 won (US$441.1) per month7.
Previous studies assert that if the concentration of sulfur dioxide (SO2) is improved by 10%, the 
marginal value of the benefit is about 36,000 won (US$32.7) (Choe and Shim, [4]) and 87,000 won 
(US$79.1) (Lee, [8]) per month.  A direct comparison of results is difficult because of the different 
methods and periods used by previous studies.  The normal lifespan of apartments in Korea before 
these are considered for renovation is 30 years.  Previous studies used the average apartment price of 
200 million (US$181,818) to 300 million (US$272,727) won.  However, the price of apartments signifi-
cantly increased since 2004, especially in Seoul, and interest toward the environment has 
grown.  Therefore, the estimated MWTP in this study is robust and meaningful.
Using spatial econometrics and spatial interpolation method, this study contributed to the litera-
ture not only by promoting the interests of SMA residents but also by addressing some of the problems 
in previous models and data.  However, econometric problems of inconsistent and biased estimators 
when spatial dependence is present that, in future studies, a more flexible approach is needed with 
functional form as well as spatial hedonic model.  Also, the effects of criteria pollutants other than 
PM10 are needed to directly compare with the results in other studies.
References
［ 1 ］	 Anselin, L. and Le Gallo, J., “Interpolation of Air Quality Measures in Hedonic House Price Models : Spatial 
Aspects,” Spatial Economic Analysis, vol. 1(1), 2006, pp. 31-52.
［ 2 ］	 Brasington, D. and Hite, D., “Demand for Environmental Quality : A Spatial Hedonic Analysis,” Regional Sci-
ence and Urban Economics, Elsevier, vol. 35(1), 2005, pp. 57-82.
［ 3 ］	 Chay, K. and Greenstone, M., “Does Air Quality Matter ?  Evidence from the Housing Market,” Journal of 
Political Economy, University of Chicago Press, vol. 113(2), 2005, pp. 376-424.
［ 4 ］	 Choe, J. and Shim, S.H., “An Effect of Air Quality on the Apartment Prices in Seoul : Using Hedonic Price 
Method,” Environmental and Resource Economics Review, vol. 11(2), 2002, pp. 261-279. (in Korean)
［ 5 ］	 Griliches, Z., “Hedonic Price Indexes Revisited : Some Notes on the State of the Art, Proceedings of the 
Business and Economic Statistics Section,” American Statistical Association, 1967, 1967, pp. 324-332.
［ 6 ］	 Griliches, Z., “Hedonic Price Indexes of Automobiles : An Econometric Analysis of Quality Change,” in Zvi 
Griliches (ed.), Price Indexes and Quality Change, Cambridge : Cambridge University Press, 1971.
［ 7 ］	 Kim, C.W., Phipps, T., and Anselin, L., “Measuring the Benefits of Air Quality Improvement : A Spatial 
Hedonic Approach,” Journal of Environmental Economics and Management, vol. 45, 2003, pp. 24-39.
［ 8 ］	 Lee, J.S., “Estimation of Implicit Marginal Price and Benefit of SO2 Density Decrease in Seoul Apartment 
7	 Previous studies assert that if the concentration of sulfur dioxide (SO2) is improved by 10%, the marginal value 
of the benefit is about 36,000 won (US$32.7) (Choe and Shim, [4]) and 87,000 won (US$79.1) (Lee, [8]) per 
month.  A direct comparison of results is difficult because of the different methods and periods used by previous 
studies.  The normal lifespan of apartments in Korea before these are considered for renovation is 30 
years.  Previous studies used the average apartment price of 200 million (US$181,818) to 300 million 
(US$272,727) won.  However, the price of apartments significantly increased since 2004, especially in Seoul, 
and interest toward the environment has grown.  Therefore, the estimated MWTP in this study is robust and 
meaningful.