# RECORD 7de41ea5f998e11b3cc730af7771c0a4b73baa2f
# TITLE: Approach to hedonic model and analyzing factors affecting real estate prices in Hanoi
# YEAR: 2022.0  VENUE: International Journal of Statistics and Applied Mathematics  DOI: 10.22271/maths.2022.v7.i4b.849
# AUTHORS: Nhat Hong Nguyen; Quang Duy Phung
# PAGES: 4  CANDIDATE TABLE PAGES (1-based): [3]
# ABSTRACT: In real estate valuation techniques and housing market research, real estate prices are often analyzed using The Hedonic model based on factors that affect real estate prices. The Hedonic model shows the factors that affect real estate prices. The determinants of real estate prices in Hanoi in this article use 240 observations. The variables that have the greatest influence on real estate prices are the location of the property, the area of the land, the area of the house, the distance from the property to the city center, and the distance from the property to the street frontage.


===== PAGE 1 (score 14, numeric-paren 0, table-words 1, var-words 12) =====

 
~101~ 
International Journal of Statistics and Applied Mathematics 2022; 7(4): 101-104 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ISSN: 2456-1452 
Maths 2022; 7(4): 101-104 
© 2022 Stats & Maths 
www.mathsjournal.com 
Received: 01-04-2022 
Accepted: 04-05-2022 
 
Nguyen Hong Nhat 
National Economic University, 
Hanoi, Vietnam 
 
Phung Duy Quang 
Foreign Trade University, 
Hanoi, Vietnam 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Corresponding Author: 
Phung Duy Quang 
Foreign Trade University, 
Hanoi, Vietnam 
 
 
 
 
 
 
 
 
 
 
Approach to hedonic model and analyzing factors 
affecting real estate prices in Hanoi 
 
Nguyen Hong Nhat and Phung Duy Quang 
 
DOI: https://doi.org/10.22271/maths.2022.v7.i4b.849 
 
Abstract 
In real estate valuation techniques and housing market research, real estate prices are often analyzed 
using The Hedonic model based on factors that affect real estate prices. The Hedonic model shows the 
factors that affect real estate prices. The determinants of real estate prices in Hanoi in this article use 240 
observations. The variables that have the greatest influence on real estate prices are the location of the 
property, the area of the land, the area of the house, the distance from the property to the city center, and 
the distance from the property to the street frontage. 
 
Keywords: Real estate price, Hedonic model, land area, house area, distance 
 
1. Introduction 
For many households, housing not only provides them shelter, but it is also a major asset. Real 
estate is one of the most important assets in the investment and buying and selling portfolio of 
households. Indeed, in industrialized countries, real estate is the asset that makes up the largest 
share of affluent households. As a result, the value of their apartments has a great impact on 
consumption, household savings and the supply-demand mechanism. Thus, home prices are of 
great concern to property developers, banks, policy makers or the general public as well as 
existing and potential owners (Ustaoilu, E. 2003) [10].  
The real estate market can be influenced by macroeconomic variables, spatial differences, 
community structural features and properties of real estate. Real estate valuation is necessary 
to provide a quantitative measure of the benefits and liabilities accrued from ownership of real 
estate. Valuations are performed by real estate agents, appraisal and inspection specialists, 
mortgage lenders, brokers, property developers, investors and fund managers, lending 
companies, market researchers, analysts, consultants and other professionals.  
 
2. The Hedonic model and experimental studies 
The Hedonic model is one of the oldest techniques of economic valuation developed by 
Lancaster (1966), Ridker (1967) [7], Grilihces (1971), Rosen (1974) and others. Originally used 
to study the relationship between air pollution reductions and property values, the Hedonic 
model became an important study in the second half of the 1970s and throughout the 1980s. 
During this period, the Hedonic pricing method has become a very important tool for academic 
research, used in theoretical and experimental studies to determine the monetary value of 
goods in relation to the environmental and physical characteristics of the property. 
The first basic Hedonic model was presented by Ridker (1967) [7]: 
 
𝑃𝑖= 𝑓(𝑆1𝑖, . . . , 𝑆𝑘𝑖; 𝑁1𝑖, . . . , 𝑁𝑚𝑖, 𝑍1𝑖, 𝑍2𝑖, . . . . , 𝑍𝑛𝑖)  
 
In which: Pi is the house price; S is the structural characteristics of the house (1 2 … k) such as 
the area of the house, the number of rooms, the type of construction and some other factors. 
N is neighborhood characteristics (1….m), such as distance to workplace, quality of school, 
local crime rate and other factors. 


===== PAGE 3 (score 96, numeric-paren 63, table-words 6, var-words 5) =====

 
~103~ 
International Journal of Statistics and Applied Mathematics 
https://www.mathsjournal.com 
 
Table 1: Description of variables 
 
Variable 
Description 
Expectation of Sign 
LogPRICE 
The logarithm of real estate prices 
 
AREA 
Land area (m2) 
+ 
HAREA 
House area (m2) 
+ 
DIS1: 
Distance from the property to the center (km) 
- 
DIS2: 
Distance from the property to street frontage (km) 
- 
WIDTH 
Front pavement width (m) 
+ 
SHAPE 
Real estate shape: dummy variable (gets the value 1 if the back is wider or equal to the front of the land; 
value 0 if the back is narrower than the front) 
+ 
NOF 
Number of floors: dummy variable (gets the value 1 if the number of floors <= 4; value 0 if > 4) 
- 
TERRACE 
Terrace: dummy variable (gets value 1 if there is a terrace; value 0 if there is no terrace) 
+ 
LOCATION 
Location: dummy variable (gets value 1 if property is in frontage; value 0 if property is in alleys) 
+ 
USE 
Usage: dummy variable (gets value 1 if the property is for business, office, or residential use; gets value 0 
for residential use only) 
+ 
 
Table 2: Descriptive statistics of variables 
 
Variable 
Average value 
Maximum value 
Minimum value 
Standard deviation 
LogPRICE 
8.20 
12.65 
7.24 
0.84 
AREA 
56.50 
260.60 
52.10 
24.3 
HAREA 
2.780 
4.30 
2.12 
1.23 
DIS1: 
24.60 
30.30 
22.40 
4.62 
DIS2: 
4.60 
6.60 
3.20 
1.26 
WIDTH 
6.30 
8.65 
5.40 
1.68 
SHAPE 
0.42 
1.00 
0.00 
0.12 
NOF 
4.20 
1.00 
0.00 
0.36 
TERRACE 
0.65 
1.00 
0.00 
0.21 
LOCATION 
0.43 
1.00 
0.00 
0.36 
USE 
0.54 
1.00 
0.00 
0.42 
 
Table 3: Results of 6 regression models 
 
Variable 
Model 1 
Model 2 
Model 3 
Model 4 
Model 5 
Model 6 
AREA 
0.0142*** 
(9.426) 
0.0143*** 
(9.426) 
0.0145*** 
(9.864) 
0.0141*** 
(10.124) 
0,.0148*** 
(10.012) 
0,.0146*** 
(9.942) 
HAREA 
0.0008 
(1.432) 
0.0008 
(1.420) 
0.0008 
(1.274) 
0.0009* 
(2.524) 
0,.0012* 
(2.788) 
0.0012* 
(2.982) 
DIS1: 
-0.045* 
(-1.984) 
-0.065* 
(-1.928) 
-0.0665* 
(-1.948) 
-0.0648* 
(-2.114) 
-0.0684* 
(-2.682) 
-0.0840* 
(-2.683) 
DIS2: 
-0.0016* 
(-1.914) 
-0.0016* 
(-1.976) 
-0.0016* 
(-1.978) 
-0.0016* 
(-1.928) 
-0.0017* 
(-1.926) 
-0.0015* 
(-1.968) 
LOCATION 
0.9284*** 
(7.192) 
0.8294*** 
(7.899) 
0.8368*** 
(9.827) 
0.8426*** 
(9.248) 
0.8424*** 
(9.842) 
0.8468*** 
(9.468) 
WIDTH 
-0.0085* 
(-1.826) 
-0.0088* 
(-1.824) 
-0.0086* 
(-1.946) 
-0.0084* 
(-1.964) 
-0.0094* 
(-1.848) 
 
SHAPE 
0.1283 
(0.984) 
0.1078 
(0.872) 
0.1268 
(0.924) 
0.1046 
(0.924) 
 
 
NOF 
-0.1193 
(-0.742) 
-0.1428 
(-0.846) 
-0.1464 
(-0.846) 
 
 
 
TERRACE 
-0.0426 
(-0.492) 
-0.0428 
(-0.548) 
 
 
 
 
USE 
0.0602 
(0.620) 
 
 
 
 
 
R2 
0.7824 
0.7918 
0.7824 
0.7784 
0.7689 
0.7846 
R2 correction 
0.7739 
0.7846 
0.7758 
0.7727 
0.7640 
0.7809 
Statistical value F 
91.8873*** 
109.8138*** 
119.1681*** 
136.4073*** 
155.7097*** 
213.9984*** 
Robust standard errors in parentheses 
*** p<0.001, ** p<0.05, * p<0.1 
Source: Estimated results from the model 
 
4. Result 
Consider model 1 
LogPRICE = 12.45 + 0.0142AREA + 0.0008HAREA −0.045DIS1 −0.0016DIS2 
 
+0.9284LOCATION −0.0085WIDTH + 0.1283SHAP −0.1193NOF- 0.0426TERRACE + 0.0602USE + ε  
(1) 
 
Model 1 has multicollinearity between variables and autocorrelation between USE and LOCATION, because the correlation 
coefficient between these two variables is 0.78. At the same time, the USE variable was not statistically significant, so the USE 
variable was excluded from model 1. 


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

 
~104~ 
International Journal of Statistics and Applied Mathematics 
https://www.mathsjournal.com 
 
Consider model 2 
LogPRICE = 12.45 + 0.0143AREA + 0.0008HAREA −0.0605DIS1 −0.0016DIS2 
 
+0.8294LOCATION −0.0088WIDTH + 0.1078SHAP −0.1428NOF- 0.0428TERRACE + ε  
(2) 
 
This model has eliminated autocorrelation but the test still shows signs of collinearity because 𝑅𝑖
2 is high. Hence model 2’s 
variables will be omitted. The TERRACE variable has low statistical significance and the sign is opposite to our expectation, so 
we remove the variable TERRACE from model 2. 
Similarly, when considering model 3, 4, we also remove variables NOF, SHAPE from the model. 
 
Consider model 3 
LogPRICE = 12.45 + 0.0148AREA + 0.0012HAREA −0.0684DIS1 −0.0017DIS2 
 
+0.8424LOCATION −0.0094WIDTH + ε  
(3) 
 
This model has eliminated autocorrelation but the test still shows collinearity because 𝑅𝑖
2 is high. Hence model 5’s variables will 
be omitted. The WIDTH variable has low statistical significance and the opposite sign is desired, so we remove the TERRACE 
variable from model 5. 
 
Consider model 4 
LogPRICE = 9.45 + 0.0146AREA + 0.0012HAREA −0.0840DIS1 −0.0015DIS2 
 
+0.8468LOCATION + ε  
(4) 
 
The variables of model 6 are all statistically significant, the signs of the coefficients are consistent with the expectations. The 
analysis results show that the price of real estate in the frontage is usually 84.68% higher than in alleys. For real estate in the alley, 
1m further from the frontage causes the average price to decrease by 0.12%. Similarly, if real estate is 1km away from the center, 
the price will decrease by 8.4% on average. If the land area increases by 1m2, the price will increase by 1.46%. 
We illustrate the following two cases: 
*Case 1: Real estate in the frontage LOCATION =1; DIS1= 0; AREA =60m2; HAREA = 186.2m2; DIS2=3.2 receives lnPRICE = 
11.39144. Then PRICE = 88.56 (equivalent to 86 billion 560 million VND). With a frontage house 3.2 km from the center, with 
an area of 60m2, the house area is 186.2m2 (4 floors), the price is 86 billion 560 million VND. 
*Case 2: real estate in alleys. LOCATION =0; DIS1= 0; AREA =60m2; HAREA = 186.2m2; DIS2=3.2 receives lnPRICE = 
10.54464. Then PRICE = 37.973 (equivalent to 37 billion 973 million). With houses in alleys 3.2 km from the center, with an area 
of 60m2, the house area is 186.2m2 (4 floors), the price has decreased significantly, at only 37 billion 973 million VND. 
 
5. Conclusion 
This article applies Hedonic model and the Status - Quality Trade-off Theory to determine real estate prices in Hanoi area. The 
results show that location variables, distance from property to street frontage, distance from property to center, land area and 
house area are important variables affecting property prices. 
 
6. References 
1. Gabriel Kayode Babawale. The Impact of Neighbourhood Churches on House Prices, Journal of Sustainable Development. 
3/2011;4(1):246-253. 
2. Hasan Selim. Determinants of house prices in Turkey: Hedonic regression versus artificial neural network, Expert Systems 
with Applications. 2009;36:2843-2852.  
3. Hoàng Hữu Phê, Patrick Wakely. Vị thế, chất lượng và sự lựa chọn khác: Tiến tới một lý thuyết mới về vị trí dân cư đô thị, 
tạp chí Đô thị học, xuất bản tại Vương quốc Anh. 1/2000;37:1.  
4. Kim. North Versus South: The Impact of Social Norms in the Market Pricing of Private Property Rights in Vietnam, World 
Development. 2007;35(12):2079-2095.  
5. Nguyễn Mạnh Hùng, Trần Văn Trọng, Lý Hưng Thành, Trần Thanh Hùng, Hoàng Hữu Phê, Phương pháp định giá BĐS ứng 
dụng thuyết vị thế - chất lượng, trang 1-8.  
6. Nguyễn Trọng Hoài, Phùng Thanh Bình, Nguyễn Khánh Duy. Dự báo và phân tích dữ liệu trong kinh tế và tài chính, NXB 
Thống kê, 2009. 
7. Ridker Ronald G, John Henning A. The Determinants of Residential Property Values with Special Reference to Air Pollution, 
The Review of Economics and Statistics. 1967;49(2):246-57. 
8. Sérgio AB, et al. Economics of Air Pollution: Hedonic Price Model and Smell Consequences of Sewage Treatment Plants in 
Urban Areas, Economics of Air Pollution. 8/2002;234:1-25.  
9. Sibel Selim. Determinants of house prices in Turkey: Hedonic regression model, Doðuþ niversitesi Dergisi, 2008, 65-76. 
10. Ustaoðlu E. Hedonic Price Analysis of Office Rents: A Case Study of the Office Market in Ankara, Middle East Technical 
University, Unpublished MSc, Thesis, 2003.