# RECORD 5242a8699dbfd95d3c698daaa18f7bab00d59c97
# TITLE: THE IMPACT OF LUXURY HOUSING ON NEIGHBORHOOD HOUSING PRICES: AN APPLICATION OF THE SPATIAL DIFFERENCE-IN-DIFFERENCES METHOD
# YEAR: 2020.0  VENUE: nan  DOI: 10.3846/ijspm.2020.13649
# AUTHORS: Chih-Min Liang; Chun-Chang Lee; You-Hsin Lin; Zheng Yu; Wen-Chih Yeh
# PAGES: 18  CANDIDATE TABLE PAGES (1-based): [9, 10, 11, 12]
# ABSTRACT: This study investigated the spatial spillover effects of luxury housing during and after construction, in regards to increases in housing prices in neighboring areas as well as the spatial dependence of neighboring housing. This study focused on already completed luxury housing in Taipei, Taiwan. First, the nearest-neighbor matching approach of propensity score matching was used to overcome the problem of data heterogeneity. The difference-in-differences (DD) method and spatial econometrics were used for analysis. The empirical results indicated that the spatial error model had the best goodness of fit. This indicated that housing prices increased by 13.0% during construction of luxury housing nearby. This indicated that housing prices increased by 5.8% after the construction of luxury housing nearby. The empirical results showed that the ongoing and completed construction of luxury housing had spillover effects on housing prices. The effect of ongoing construction of luxury housing was particularly large in scope, indicating its role as a predictor of psychological reaction in the market.


===== PAGE 1 (score 12, numeric-paren 1, table-words 0, var-words 11) =====

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unre-
stricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Copyright © 2020 The Author(s). Published by Vilnius Gediminas Technical University
THE IMPACT OF LUXURY HOUSING ON NEIGHBORHOOD
HOUSING PRICES: AN APPLICATION OF THE SPATIAL
DIFFERENCE-IN-DIFFERENCES METHOD
Chih-Min LIANG1, Chun-Chang LEE2,*, You-Hsin LIN2, Zheng YU2, Wen-Chih YEH2,**
1 Department of Public Finance and Tax Administration, National Taipei University of Business, Taipei, Taiwan
2 Department of Real Estate Management, National Pingtung University, Pingtung, Taiwan
Received 09 September 2019; accepted 07 June 2020
Abstract. This study investigated the spatial spillover effects of luxury housing during and after construction, in regards
to increases in housing prices in neighboring areas as well as the spatial dependence of neighboring housing. This study
focused on already completed luxury housing in Taipei, Taiwan. First, the nearest-neighbor matching approach of propen-
sity score matching was used to overcome the problem of data heterogeneity. The difference-in-differences (DD) method
and spatial econometrics were used for analysis. The empirical results indicated that the spatial error model had the best
goodness of fit. This indicated that housing prices increased by 13.0% during construction of luxury housing nearby. This
indicated that housing prices increased by 5.8% after the construction of luxury housing nearby. The empirical results
showed that the ongoing and completed construction of luxury housing had spillover effects on housing prices. The effect
of ongoing construction of luxury housing was particularly large in scope, indicating its role as a predictor of psychological
reaction in the market.
Keywords: luxury housing, propensity score matching, difference-in-differences, spatial dependence, spatial lag model,
spatial error model.
Introduction
With the improvement of living standards in Taiwan, 
housing now constitutes the highest expenditure with re-
spect to the basic necessities of life. In addition to consid-
ering affordability, transportation convenience, and loca-
tion when choosing housing, people have recently started 
pursuing high-standard, high-quality, high-profile, and 
high-priced housing.
In recent years, citizens of Taiwan have paid more 
attention to quality of life. This has led to the develop-
ment of a new type of housing, referred to in the market 
as luxury housing, derived from existing types of hous-
ing such as condominiums, studios, multi-story houses, 
and high-rise suites. Lin and Jou (2005) described typical 
luxury housing as housing with rare spatial resources such 
as beautiful landscapes and good locations, and character-
ized by a high total price, a large living space, good facili-
ties, and high prestige. According to the 15th provision of 
the “Property Price and Present Value Appraisal Criteria” 
as established by the Taipei City Revenue Service, build-
ings that are made from reinforced concrete or belong to 
a higher construction class are used for residential pur-
poses. Properties with a total value of over NT$80 mil-
lion, and a total registered area of ownership over 80 ping 
(1 ping is equal to 3.306 square meter) or a per apartment 
price of over NT$1 million per ping (excluding parking 
spaces), are defined as high-level housing or luxury hous-
ing. According to Wang (2008), luxury housing is a small 
market segment targeting upper class consumers; Wang 
suggested that such housing is showy by nature, and its 
target consumers are relatively homogeneous, which pro-
vides a cluster effect. Lin (2004) maintained that luxury 
housing has shaped a particular lifestyle for upper class 
society, and that it is not only local landmark, but also an 
indicator of wealth and ostentation.
Past studies on luxury housing have discussed histor-
ical development, purchase decision-making, integrated 
marketing, and positioning design of luxury housing 
(e.g. Lin & Jou, 2005; Ho, 2011; Wu, 2008; Chen, 2011). 
However, there is a lack of research regarding the effect 
of luxury housing on neighborhood housing prices.
International Journal of Strategic Property Management
ISSN: 1648-715X / eISSN: 1648-9179
2020 Volume 24 Issue 6: 456–473
https://doi.org/10.3846/ijspm.2020.13649
*Corresponding author. E-mail: lcc@mail.nptu.edu.tw
**Department of Real Estate Management, HungKuo Delin University of Technology, New Taipei, Taiwan


===== PAGE 9 (score 54, numeric-paren 20, table-words 9, var-words 5) =====

464
C.-M. Liang et al. The impact of luxury housing on neighborhood housing prices: an application...
groups of samples, and check if self-selection bias was 
effectively reduced. As shown in Table 2, before match-
ing, there were 352 samples in the experimental group 
and 6,940 samples in the control group, with significant 
differences being observed in terms of the number of 
rooms, number of living rooms, housing type, residence 
floor, distance to the nearest MRT station, distance to the 
nearest senior high school, and distance to the nearest 
junior high school. After matching, there were 352 sam-
ples in the experimental group and 4,772 samples in the 
control group, and in this case, significant differences in 
the variables were not observed, with the exception of 
the number of rooms and housing type. As shown in Ta-
ble 3, before matching, there were 542 samples in the ex-
perimental group and 9,953 samples in the control group, 
with significant differences being observed in terms of the 
number of rooms, number of living rooms, housing type, 
residence floor, distance to the nearest MRT station, dis-
tance to the nearest senior high school, and distance to the 
nearest junior high school. After matching, there were 541 
samples in the experimental group and 7,106 samples in 
the control group, and in this case, significant differences 
in the variables were not observed, with the exception of 
housing type.
Table 2. Differences in mean values between the experimental group and the control group during construction
Variable
Before matching
After matching
Treated group
(N = 352)
Control group
(N= 6,940)
t-value
Treated group
(N = 352)
Control group
(N= 4,772)
t-value
AREA
30.501
31.614
−1.280
30.501
30.265
0.180
AGE
20.903
20.091
1.280
20.903
21.322
−0.460
ROOM
2.293
2.492
−3.940***
2.293
2.452
−2.060**
LIVROOM
1.591
1.773
−7.010***
1.591
1.625
−0.880
BATH
1.446
1.451
−0.150
1.446
1.477
−0.680
MRT
290.630
554.350
−11.160***
290.630
275.740
1.320
HISC
402.650
676.610
−12.160***
402.650
389.060
0.910
JUNI
475.090
529.020
−3.400***
475.090
472.080
0.180
PRIM
390.430
377.690
1.160
390.430
387.770
0.220
Type
Frequency
Frequency
χ2
Frequency
Frequency
χ2
MULTISTORIED 
BUILDINGS
274
4,674
16.91***
274
3,431
5.78**
APARTMENT HOUSES
78
2,266
78
1,341
FLOOR
Frequency
Frequency
χ2
Frequency
Frequency
χ2
FLOOR1
10
392
5.07**
10
200
1.52
NONFLOOR1
342
6,548
342
4,572
Note: ***, **, and * indicate that the coefficient reaches the 1%, 5%, and 10% level of significance, respectively; significantly different from 0.
Table 3. Differences in mean values between the experimental group and the control group after construction
Variable
Before matching
After matching
Treated group
(N = 542)
Control group
(N = 9,953)
t-value
Treated group
(N = 541)
Control group
(N = 7,106)
t-value
AREA
34.824
33.692
1.730*
34.810
35.007
−0.220
AGE
26.329
24.157
4.300***
26.286
26.618
−0.520
ROOM
2.602
2.668
−1.690*
2.604
2.649
−0.740
LIVROOM
1.840
1.799
1.990**
1.839
1.847
−0.240
BATH
1.630
1.611
0.740
1.629
1.647
−0.450
MRT
302.180
628.310
−15.890***
302.340
302.040
0.030
HISC
507.980
725.530
−11.090***
507.920
514.520
−0.340
JUNI
415.650
551.230
−9.770***
416.190
408.900
0.580
PRIM
437.370
390.470
4.890***
437.040
427.340
0.900
Type
Frequency
Frequency
χ2
Frequency
Frequency
χ2
MULTISTORIED 
BUILDINGS
456
6,067
117.37***
455
4,880
56.73***
APARTMENT HOUSES
86
3,886
86
2,226
FLOOR
Frequency
Frequency
χ2
Frequency
Frequency
χ2
FLOOR1
32
789
2.92*
32
478
0.53
NONFLOOR1
510
9,164
509
6,628
Note: ***, **, and * indicate that the coefficient reaches the 1%, 5%, and 10% level of significance, respectively; significantly different from 0.


===== PAGE 10 (score 25, numeric-paren 0, table-words 4, var-words 7) =====

International Journal of Strategic Property Management, 24(6): 456–473
465
3.4. Descriptive statistics of the data
The acquisition of a construction license was set as the event 
(Table 4). The average total price of housing (PRICE) was 
NT$1,439.89, with the standard deviation of NT$1,113.18. 
The average housing area (AREA) was 30.93 ping, with the 
standard deviation of 16.42 ping. The average house age 
(AGE) was 20.13 years, with the standard deviation of 11.82 
years. The mean value of the residence floor (FLOOR) was 
0.041, indicating that4.1% of house residents lived on the 1st 
floor and 95.9% lived on other floors. The mean value of the 
housing type (TYPE) was 0.723, indicating that multisto-
ried buildings and apartment houses accounted for 72.3% 
and 27.7% of the sample, respectively. The average number 
of rooms (ROOM) was 2.40, with the standard deviation 
of 0.95. The average number of living rooms (LIVROOM) 
was 1.70, with the standard deviation of 0.49. The average 
number of bathrooms (BATH) was 1.44, with the standard 
deviation of 0.58. The average distance to the nearest MRT 
station (MRT) was 357 meter, with the standard deviation 
of 200.04 meter. The average distance to nearest senior high 
school (HISC) was 495.42 meters, with the standard devia-
tion of 243.69 meters. The average distance to nearest junior 
high school (JUNI) was 486.83 meters, with the standard 
deviation of 232.19 meters. The average distance to nearest 
primary school (PRIM) was 373.09 meters, with the stand-
ard deviation of 173.44 meters.
The acquisition of a usage license was set as the event 
(Table 5). The average total price of housing (PRICE) was 
NT$1,892.83, with the standard deviation of NT$1,153.54. 
The average housing area (AREA) was 33.73 ping, with 
Table 4. Descriptive statistics: During construction – Construction license acquisition
Variable
Mean
Standard deviation
Min.
Max.
PRICE (ten thousand NT$)
1,439.89
1,113.18
44
23,100
AREA (ping)
30.93
16.42
2.14
174.43
AGE (year)
20.13
11.82
0.10
53.80
ROOM
2.40
0.95
1
5
LIVROOM
1.70
0.49
1
5
BATH
1.44
0.58
1
5
TMRT
357.00
200.04
3.40
1397.44
HISC
495.42
243.69
13.11
1561.37
JUNI
486.83
232.19
17.56
1613.65
PRIM
373.09
173.44
16.42
1148.42
Frequency
Percent
Cumulative percent
TYPE
MULTISTORIED BUILDINGS
3,705
72.31
72.31
APARTMENT HOUSES
1,419
27.69
100.00
FLOOR
FLOOR1
210
4.10
4.10
NONFLOOR1
4,914
95.90
100.00
Table 5. Descriptive statistics: After construction – Usage license acquisition
Variable
Mean
Standard deviation
Min.
Max.
PRICE (ten thousand NT$)
1,892.83
1,153.54
217
33300
AREA (ping)
33.73
15.06
7.85
312
AGE (year)
24.61
11.51
0.1
55
ROOM
2.62
0.90
1
9
LIVROOM
1.80
0.47
1
8
BATH
1.60
0.59
1
8
MRT
415.77
229.40
2.46
1572.23
HISC
607.72
395.44
14.36
3617.10
JUNI
469.33
239.28
15.79
1561.37
PRIM
388.40
184.45
19.158
1334.04
Frequency
Percent
Cumulative percent
TYPE
MULTISTORIED BUILDINGS
5,335
69.77
69.77
APARTMENT HOUSES
2,312
30.23
100.00
FLOOR
FLOOR1
510
6.67
6.67
NONFLOOR1
7,137
93.33
100.00


===== PAGE 11 (score 51, numeric-paren 10, table-words 13, var-words 10) =====

466
C.-M. Liang et al. The impact of luxury housing on neighborhood housing prices: an application...
the standard deviation of 15.06 ping. The average house 
age (AREA) was 24.61 years, with the standard devia-
tion of 11.51 years. The mean value of the residence floor 
(FLOOR) was 0.068, indicating that 6.8% of house resi-
dents lived on the 1st floor and 93.2% lived on other floors. 
The mean value of the housing type (TYPE) was 0.698, in-
dicating that multistoried buildings and apartment houses 
accounted for 69.8% and 30.2% of the sample, respectively. 
The average number of rooms (ROOM) was 2.62, with the 
standard deviation of 0.90. The average number of living 
rooms (LIVROOM) was 1.80, with the standard devia-
tion of 0.47. The average number of bathrooms (BATH) 
was 1.60, with the standard deviation of 0.59. The average 
distance to the nearest MRT station (MRT) was 415.77 
meter, with the standard deviation of 229.40 meter. The 
average distance to nearest senior high school (HISC) was 
607.72 meters, with the standard deviation of 395.44 me-
ters. The average distance to nearest junior high school 
(JUNI) was 469.33 meters, with the standard deviation of 
239.28 meters. The average distance to nearest primary 
school (PRIM) was 388.40 meters, with the standard de-
viation of 184.45 meters.
4. Empirical results
This study investigated the effect of luxury housing on 
housing prices in neighboring areas during (indicated 
by a construction license) and after (indicated by a us-
age license) its construction. The spatial concepts were 
integrated into the regression model in order to solve the 
issues of spatial heterogeneity and spatial autocorrelation 
in the traditional OLS model. As shown in Table 6, Mo-
ran’s I reached the 1% level of significance both during 
and after construction, indicating the presence of positive 
autocorrelation and spatial clustering in housing prices in 
areas neighboring with luxury housing. Housing prices in 
administrative districts showed positive autocorrelation. 
The spatial dependence test was further used to determine 
which spatial model had the best explanatory power. Both 
LM-lag and LM-error reached the 1% level of significance. 
Both Robust LM-lag and Robust LM-error reached the 1% 
level of significance. Thus, the residuals differed in terms 
of heterogeneity. The results of the spatial lag model and 
spatial error model were further compared.
(1) During construction
Table 7 shows the results of evaluating the period after 
construction using the three models (OLS, SLM, SEM). 
The Breusch-Pagan test was used to determine heterosce-
dasticity of residuals. Its values in the three models were 
1483.72, 1436.74, and 1523.74, reaching the 1% level of 
significance. The value in the spatial error model was 
the highest, indicating heteroscedasticity of its residuals. 
R2 was equal to 0.797, 0.807, and 0.813; the spatial er-
ror model showed the highest explanatory power. Smaller 
AIC and SC values indicate better goodness of fit. AIC was 
equal to 1948.61, 1708.89, and 1642.35 and SC was equal 
to 2125.23, 1892.06, and 1818.97. The larger log likelihood 
values indicate better goodness of fit. The log likelihood 
values of the models were −947.30, −826.45, and −794.17, 
indicating that the spatial lag model and spatial error 
model had better goodness of fit. The results described 
above showed that the spatial error model was better than 
the traditional OLS regression model and spatial lag mod-
el. The evaluation results show that the spatial error model 
had the best goodness of fit. The results of the spatial error 
model are discussed further.
With regard to property attributes, the building trans-
fer area coefficient was 0.022, reaching the 1% level of sig-
nificance, which indicated that the housing price increased 
by 2.2% with each additional ping of the building transfer 
area. The house age coefficient was −0.013, reaching the 
1% level of significance. The coefficient of the square of 
house age was 0.001, reaching the 1% level of significance, 
which indicated prices of older housing were lower and 
such decrease in prices was decreased. The residence floor 
coefficient was 0.129, reaching the 1% level of significance, 
which indicated that the prices of housing located on the 
first floor were 12.9% higher than that on other floors. The 
coefficient of the number of rooms was 0.086, reaching the 
1% level of significance, which indicated that the hous-
ing price increased by 8.6% with each additional room. 
The coefficient of the number of living rooms was 0.173, 
reaching the 1% level of significance, which indicated that 
the housing price increased by 17.3% with each additional 
living room. The coefficient of the number of bathrooms 
was 0.083, reaching the 1% level of significance, which 
indicated that the housing price increased by 8.3% with 
each additional bathroom. The housing type coefficient 
Table 6. Spatial dependence during and after construction of luxury housing
Test
During construction
After construction
MI/DF
Z(I) value
MI/DF
Z(I) value
Moran’s I
0.1381***
21.626
0.2078***
34.996
LM-lag
1
289.411***
1
848.676***
LM-error
1
57.076***
1
95.030***
Robust LM-lag
1
435.721***
1
1183.005***
Robust LM-error
1
203.386***
1
429.358***
Note: MI/DF indicate Moran’s I and degrees of freedom. ***, **, and * indicate that the coefficient reaches the 1%, 5%, and 10% level of significance, 
respectively; significantly different from 0.


===== PAGE 12 (score 231, numeric-paren 186, table-words 10, var-words 2) =====

International Journal of Strategic Property Management, 24(6): 456–473
467
Table 7. Empirical results for the period during construction and after construction of luxury housing
During construction (N = 5124)
After construction (N = 7647)
OLS
Spatial lag 
model
Spatial error 
model
OLS
Spatial lag 
model
Spatial error 
model
Independent variable
coefficient
(S.E.)
coefficient
(S.E.)
coefficient
(S.E.)
coefficient
(S.E.)
coefficient
(S.E.)
coefficient
(S.E.)
Intercept
5.260***
(0.035)
3.914***
(0.091)
5.267***
(0.047)
6.010***
(0.029)
3.958***
(0.082)
6.055***
(0.042)
AREA
0.023***
(0.001)
0.022***
(0.001)
0.022***
(0.001)
0.018***
(0.001)
0.017***
(0.001)
0.017***
(0.001)
AGE
−0.013***
(0.001)
−0.013***
(0.001)
−0.013***
(0.001)
−0.004***
(0.001)
−0.004***
(0.001)
−0.004***
(0.001)
AGES
0.001***
(0.001)
0.001***
(0.001)
0.001***
(0.001)
0.001***
(0.001)
0.001***
(0.001)
0.001***
(0.001)
FLOOR
0.124***
(0.021)
0.126***
(0.020)
0.129***
(0.020)
0.221***
(0.013)
0.219***
(0.012)
0.221***
(0.012)
ROOM
0.085***
(0.007)
0.085***
(0.007)
0.086***
(0.007)
0.082***
(0.005)
0.078***
(0.005)
0.081***
(0.005)
LIVROOM
0.167***
(0.012)
0.166***
(0.011)
0.173***
(0.011)
0.097***
(0.008)
0.088***
(0.008)
0.092***
(0.008)
BATH
0.087***
(0.010)
0.085***
(0.010)
0.083***
(0.010)
0.018**
(0.007)
0.017***
(0.007)
0.019***
(0.007)
TYPE
0.068***
(0.012)
0.073***
(0.012)
0.069***
(0.012)
0.272***
(0.008)
0.255***
(0.008)
0.289***
(0.008)
Location fixed effects
yes
yes
yes
yes
yes
yes
TIME
0.427***
(0.009)
0.427***
(0.009)
0.430***
(0.009)
0.218***
(0.008)
0.211***
(0.007)
0.228***
(0.007)
TREAT
0.021
(0.024)
0.012
(0.023)
0.003
(0.029)
0.050**
(0.021)
0.014
(0.020)
0.027
(0.026)
DD
0.146***
(0.033)
0.132***
(0.032)
0.130***
(0.032)
0.050*
(0.026)
0.053**
(0.024)
0.058**
(0.025)
MRT
−0.001***
(0.001)
−0.001***
(0.001)
−0.001***
(0.001)
−0.001***
(0.001)
−0.001***
(0.001)
−0.001***
(0.001)
HISC
0.001
(0.001)
0.001
(0.001)
0.001
(0.001)
0.001
(0.001)
−0.001
(0.001)
−0.001
(0.001)
JUNI
0.001
(0.001)
−0.001
(0.001)
−0.001
(0.001)
0.001
(0.001)
−0.001
(0.001)
0.001
(0.001)
PRIM
−0.001
(0.001)
v0.001
(0.001)
−0.001
(0.001)
−0.001***
(0.001)
−0.001***
(0.001)
−0.001***
(0.001)
Spatial Lag(ρ)
−
0.205***
(0.0130)
−
−
0.298***
(0.011)
−
Lambda(λ)
−
−
0.406***
(0.022)
−
−
0.472***
(0.016)
Breusch-Pagan
1483.72***
1436.74***
1523.74***
8491.93***
7766.88***
8929.17***
Log Likelihood
−947.30
−826.45
−794.17
−1140.42
−803.25
−729.85
SC
2125.23
1892.06
1818.97
2522.28
1856.87
1701.13
AIC
1948.61
1708.89
1642.35
2334.85
1662.49
1513.69
R2
0.797
0.807
0.813
0.719
0.746
0.757
Note: The estimated coefficient value are non-standardized. ***, **, and * indicate that the coefficient reaches the 1%, 5%, and 10% level of significance, 
respectively; significantly different from 0.


===== PAGE 13 (score 29, numeric-paren 0, table-words 7, var-words 11) =====

468
C.-M. Liang et al. The impact of luxury housing on neighborhood housing prices: an application...
evaluation of data related to the period during construc-
tion. Thus, these results are not further described here. 
With regard to neighborhood characteristics, the coef-
ficient of the distance to the nearest MRT station was 
−0.001, indicating a 1% level of significance, and the coef-
ficient of the distance to the nearest primary school was 
–0.001, indicating a 1% level of significance.
With regard to DD properties, the estimated coef-
ficients of the construction license acquisition time for 
the three models were 0.218, 0.211, and 0.228, reaching 
the 1% level of significance, which indicated that housing 
prices were 21.8%, 21.1%, and 22.8% higher after acqui-
sition of a construction license by luxury housing than 
before. The estimated coefficients of location in the neigh-
boring area were 0.050, 0.014, and 0.027; the coefficient in 
the OLS model reached the 5% level of significance and 
the coefficients in the spatial lag model and spatial error 
model have not reached the 10% level of significance. 
The estimated coefficients of the interaction variable of 
construction license acquisition time and location in the 
neighboring area were 0.050, 0.053, and 0.058, reaching 
the level of significance. The DD estimated coefficients 
were 5.0%, 5.3%, and 5.8%.
In the spatial error model, the spatial error coefficient 
(λ) was 0.472, reaching the 1% level of significance, which 
indicated that interference factors in error terms caused 
spatial autocorrelation. In the spatial lag model, the spatial 
lag coefficient (ρ) was 0.298, reaching the 1% level of sig-
nificance, which indicated that there was spatial depend-
ence between neighboring areas and housing prices were 
affected by neighboring housing prices.
5. Discussion
The empirical results regarding basic property attributes 
indicated that the building transfer area had a significant 
positive effect on housing prices, which corresponded to 
the findings of Frew and Judd (2003), Martins-Filho and 
Bin (2005), and Abbott and Klaiber (2011). This indi-
cated that prices of housing with larger area are higher. 
House age was found to have a significant and negative 
effect on housing prices, which corresponded to the re-
sults reported by Glasock et al. (2000), Osland (2010), and 
Bin and Landry (2013). With time, housing material and 
facilities become older, which reduces the housing price. 
The square of house age was found to significantly and 
positively affect housing prices, which corresponded to the 
results reported by Malpezzi et al. (1987), Geltner et al. 
(2010), and Bin and Landry (2013). The house age variable 
can have non-linear changes and, thus, the square of house 
age was added that was found to have a significant positive 
effect on housing prices. The number of rooms was found 
to significantly and positively affect housing prices, which 
corresponded to the results reported by McMillen and 
Redfearn (2010) and Paredes and Iturra (2013). A larger 
number of rooms was found to result in higher housing 
prices. The numbers of living rooms and bathrooms were 
found to significantly and positively affect housing prices, 
was 0.069, reaching the 1% level of significance, which 
indicated that the housing prices in multistoried buildings 
and mansions were 6.9% higher than those in apartment 
houses. With regard to neighborhood characteristics, the 
coefficient of the distance to the nearest MRT station was 
−0.001, indicating a 1% level of significance. The three 
coefficients of the distance to the nearest school did not 
attain the level of significance.
With regard to DD properties, the estimated coef-
ficients of the construction license acquisition time for 
the three models were 0.427, 0.427, and 0.430, reaching 
the 1% level of significance, which indicated that hous-
ing prices were 42.7%, 42.7%, and 43.0% higher after the 
acquisition of a construction license by luxury housing 
than before. The estimated coefficients of location in the 
neighboring area were 0.021, 0.012, and 0.003, and did 
not reach the 1% level of significance, indicating that there 
was no significant difference between housing prices in 
areas neighboring and not neighboring with luxury hous-
ing. The estimated coefficients of the interaction variable 
of construction license acquisition time and location in 
the neighboring area were 0.146, 0.132, and 0.130, reach-
ing the 1% level of significance. The DD estimated coef-
ficients were 14.6%, 13.2%, and 13.0%. This indicated that 
the average difference in housing prices in areas neigh-
boring and not neighboring luxury housing was 14.6%, 
13.2%, and 13.0% during construction of luxury housing.
In the spatial error model, the spatial error coefficient 
(λ) was 0.406, reaching the 1% level of significance, which 
indicated that interference factors in error terms caused 
spatial autocorrelation. In the spatial lag model, the spa-
tial lag coefficient (ρ) was 0.205, reaching the 1% level 
of significance, which indicated that housing prices were 
affected by neighboring housing prices.
(2) After construction
The results of three models regarding the period after the 
completion of luxury housing construction are discussed 
further. The Breusch-Pagan test was used to determine 
heteroscedasticity of residuals. Its values in three models 
were 8491.93, 7766.88, and 8929.17, reaching the 1% level 
of significance. The value in the spatial error model was 
the highest, indicating heteroscedasticity of its residuals. 
R2 was equal to 0.719, 0.746, and 0.757; the spatial error 
model showed the highest explanatory power. AIC was 
equal to 2334.85, 1662.49, and 1513.69 and SC was equal 
to 2522.28, 1856.87, and 1701.13. The log likelihood val-
ues of the models were −1140.42, −803.25, and −729.85, 
indicating that the bias could not be solved in the OLS 
regression model but could be solved in the spatial lag 
model and spatial error model. The results showed that 
the spatial error model was better than the traditional 
OLS regression model and spatial lag model. The evalu-
ation results show that the spatial error model had the 
best goodness of fit. The results of the spatial error model 
are discussed further. With regard to property attributes, 
the results showed that the estimated coefficients and their 
significance levels were similar to those observed in the