Study for Factors Influencing Pre-Owned Housing Prices in China

Muge Zhao

Stony Brook Institute at Anhui University, Anhui University, Hefei, Anhui, 230031, China

Keywords: Pre-Owned Housing Prices, Influencing Factors, Multiple Linear Regression Model, Geographically

Weighted Regression Model, Random Forest.

Abstract: Against the backdrop of the sustained development of China's real estate market and active pre-owned housing

transactions, it is crucial to explore the factors influencing pre-owned housing prices. This article focuses on

this and uses a review method to integrate literature from multiple fields. Research has found that a

combination of multiple factors influences pre-owned housing prices. In the inherent properties of a house,

factors such as age, decoration, layout, and floor level directly affect its price. In terms of the external

environment, location factors have a significant impact, and the completeness of supporting facilities such as

education, healthcare, commerce, and transportation in the surrounding area greatly affects its market value.

The supporting services, such as property services and greening rates within the community, will also have

an impact on housing prices. In addition, the article summarizes existing research methods, such as multiple

linear regression, geographically weighted regression, and random forest models. A deep understanding of

these factors and mechanisms can not only help homebuyers and investors make scientific decisions but also

provide theoretical support for the government to formulate precise and effective real estate regulation policies

and help stabilize the market.

1 INTRODUCTION

In current China, the real estate market has undergone

years of development and has become a significant

mainstay industry of the national economy. In recent

years, with the continuous advancement of

urbanization and the continuous growth of the urban

population, the demand for housing has become

increasingly strong. However, due to limited land

resources, the growth rate of the new housing supply

has gradually slowed down, and the pre-owned

housing market has become increasingly active. Its

transaction scale continues to expand, and its

proportion in the housing market continues to

increase. From January to November 2023, the

volume of business of pre-owned houses and the sales

area of newly built commercial houses in China

increased by 6.9% compared to the same period in

2022. Compared to the same period last year, the

volume of business of pre-owned houses accounted

for around 40% of the whole transaction volume of

new and pre-owned houses, a growth of about 10

percentage points. In some major cities, the

https://orcid.org/0009-0004-2330-4418

proportion of pre-owned house transaction volume

even exceeded 50% (source from The Paper). At the

same time, the government's regulation policies on

the real estate market are becoming increasingly

refined, and the positioning of "housing for living, not

for speculation" is deeply implemented, aiming to

stabilize housing prices and ensure people's

livelihoods. In this context, studying the influencing

factors of resold housing prices in China is of great

significance. For homebuyers, it can help them

accurately grasp market dynamics, weigh various

factors, and make rational purchasing decisions,

avoiding blindly following trends and causing

economic losses.

Previous literature has studied the factors

influencing resold housing prices from various

research perspectives. Yang, Shao & Peng (2023)

selected factors such as the presence or absence of

elevators and decoration as independent variables,

constructed a multiple linear regression model, and

found a significant correlation between pre-owned

housing prices in Beijing and their independent

variables. Guo, Yu & Ke (2022) first used a multi-

factor covariance model to determine the main factors

Zhao, M.

Study for Factors Inﬂuencing Pre-Owned Housing Prices in China.

DOI: 10.5220/0013814000004708

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Innovations in Applied Mathematics, Physics, and Astronomy (IAMPA 2025), pages 59-64

ISBN: 978-989-758-774-0

affecting pre-owned housing prices in Yantai City

and established a valuation model and Lasso

regression model to ultimately select three important

factors: property fees, administrative districts, and

main floors. Zhao (2021) constructed a hedonic price

model and a geographically weighted regression

(GWR) model, and used interpolation analysis to

discover that different factors have varying degrees of

impact on pre-owned housing prices in different

regions of Guangzhou, indicating spatial

heterogeneity. However, some literature selects a

relatively single research perspective, and the

selected research models are not perfect enough. For

example, the interaction between influencing factors

may affect the research model. Mi (2018) studied the

factors affecting pre-owned housing prices in

Guangzhou but did not take into account the

interaction between the insignificant influencing

factor of whether the decoration is fine and other

factors in the regression model.

This article aims to analyze the more common

influencing factors of pre-owned housing prices and

explore more effective research models, providing

pricing references for real estate transactions and

facilitating consumers to evaluate the cost-

effectiveness when purchasing pre-owned houses.

2 RESEARCH METHODS FOR

INFLUENCING FACTORS OF

PRE-OWNED HOUSING

PRICES

2.1 Multiple Linear Regression Model

This article found through reviewing existing

literature that multiple studies have used multiple

linear regression models and their related models to

investigate the factors affecting pre-owned housing

prices. The multiple linear regression model is a

commonly used statistical analysis method that has

advantages for studying this problem. Firstly, it can

consider multiple influencing factors; Secondly, it

can make the variable relationship clear and intuitive.

This model assumes a linear relationship between the

dependent variable (pre-owned housing prices) and

the independent variable (various influencing factors),

01122 pp

yxx x

ββ β β

= + + +⋅⋅⋅+ +

(1)

Among them

are pre-owned housing prices,

are various influencing factors,

corresponding regression coefficients, and

random error terms. By estimating the regression

coefficients, it is possible to clarify the direction and

degree of influence of each independent variable on

the dependent variable. A positive regression

coefficient indicates a positive correlation between

the factor and housing prices, while a negative

coefficient indicates a negative correlation, and the

magnitude of the coefficient reflects the strength of

the impact. For example, if the regression coefficient

of the housing area is positive and the value is large,

it indicates that the larger housing area can contribute

to the higher housing price. Liu, Jin & Wang (2017)

and others studied the influencing factors of pre-

owned housing prices in Nanjing. Using a multiple

linear regression model, they selected factors such as

house area and number of bedrooms as independent

variables, and calculated that the area and whether

there is an elevator have a greater impact on unit area

housing prices, while area and number of bedrooms

have a smaller impact on unit area housing prices;

Shen (2017) also used regression models to study the

key factors impacting resold housing prices in Beijing,

selecting factors such as greening rate and housing

area, and established regression equations to verify

the significance of the effects of each influencing

factor. Finally, the conclusion was drawn that the

interaction between urban areas, school district

houses, and subway houses was significant; Chen

(2019) also used a multiple linear model to study the

important factors influencing resold housing prices in

Shenyang, and found that the price of pre-owned

housing purchased by buyers is mainly influenced by

the building area, elevator equipment, and location.

However, multiple linear regression models also have

limitations, as they may lack the ability to capture

complex interactions between variables.

2.2 Geographically Weighted

Regression Model

Some articles also utilize more optimized models,

such as the geographically weighted regression model,

which is a regression analysis method that considers

spatial nonstationarity and extends the traditional

linear regression model. The traditional linear

regression model assumes that the regression

coefficients remain fixed throughout the entire study

area, but in the real world, many phenomena exhibit

spatial nonstationarity, meaning that the relationships

between variables change with spatial location. The

GWR model allows regression coefficients to vary

spatially by assigning a local weight to each sample

point, establishing a local regression model that more

accurately describes the characteristics of spatial data

and overcomes this problem. Its basic form is:

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy

(,) (,)

iii kiiiki

yuv uvx

ββ

=+ +



(2)

is the observed value of the dependent

variable at the position;

),(

is the coordinates of

the location;

),(

0 ii

and

),(

iik

are the

intercept at the position

),(

and the regression

coefficient of the independent variable, respectively;

is the value of the kth independent variable of the

ith sample point;

is the number of independent

variables;

is a random error term; The key to the

GWR model is to determine the local weight of each

sample point, and commonly used weight functions

include Gaussian kernel function, double square

kernel function, etc. For example, Gao (2020) studied

the influencing factors of pre-owned housing prices

within the third ring road of Wuhan city, and used this

model to regress the unit price of houses and eight

characteristic variables; Liu (2016) also used this

model to study the impact of residential

characteristics on the listing prices of pre-owned

houses in Xuzhou City. The geographically weighted

regression model he constructed is mainly used to

study the spatial relationship between commodity

residential prices and residential characteristics, that

is, influencing factors; Li, Wang & Qi (2024) used a

geographic regression weighted method to explore

the influencing factors of residential costs in

Shanghai.

2.3 Random Forest Regression Model

Compared to traditional multiple linear regression

models, the random forest model has a stronger

ability to handle nonlinear relationships and

robustness. Random forest is a model that constructs

multiple decision trees and synthesizes their predicted

results to make the final decision. The core formula

may vary slightly depending on the type of task

(regression or classification), for example, for

regression problems,

() ()

yx f x

∧



, (3)

Among them,

)(xy

∧

is the predicted value of the

random forest on the input

;

is the number of

decision trees in a random forest;

)(xf

is the

predicted value of the input

for the kth decision

tree. This formula indicates that the prediction result

of a random forest for regression problems is the

average of the predicted values of all decision trees

for the same input. Qin Yanjiao conducted a

comparative study on multiple linear regression

models and random forest algorithms for housing

price prediction models, and found that for the

random forest model, the average absolute error and

root mean square error are smaller, making it more

advantageous in the housing price research sector;

Wei (2021) studied the important factors influencing

resold housing prices in Nanning and found that

random forests are suitable for feature selection on

high-dimensional data; Sun (2019) simultaneously

used multiple regression models and random forest

regression models to explore the impact of regional

factors on the unit area price of pre-owned houses in

Shijiazhuang; Li, Wang & Tong (2023) also

conducted a study on the nonlinear impact of street

quality on housing prices based on the random forest

model, and found that the random forest model has a

higher goodness of fit compared to traditional linear

models.

2.4 Other Methods

In addition to the research methods mentioned above,

there are also research methods such as the median

regression model, covariance regression model, and

stratification model, which will not be introduced one

by one in this article.

3 THE INFLUENCING FACTORS

OF CURRENT PRE-OWNED

HOUSING PRICES IN CHINA

3.1 Location Factors

Location factors are significant factors affecting pre-

owned housing prices. Generally speaking, under

other equal conditions, the more convenient the

transportation, the more complete the supporting

facilities (such as shopping malls, hospitals, schools),

and the more concentrated the resources in the core

urban area, the higher the pre-owned housing prices.

In Beijing, Xicheng District and Haidian District are

the core areas, while Daxing District and Tongzhou

District are slightly more remote noncore areas.

Studies have shown that the resold housing costs of

the first two are higher than those of the latter two,

indicating that location significantly affects housing

prices (Yang et al., 2023). In the process of forming

the costs of pre-owned houses in Baotou City, the

Study for Factors Inﬂuencing Pre-Owned Housing Prices in China

location factor of the houses also has a significant

impact. The three districts of Baotou City, Kundulun

District, Qingshan District, and Donghe District, have

different pre-owned housing prices, indicating that

the prices of pre-owned housing in Baotou City are

greatly influenced by regional development (Ma,

2023). The resold housing costs in Zone A of Hefei

City are greatly influenced by whether it is equipped

with a subway, as the transportation facilities,

hospitals, shopping malls, and other supporting

facilities in Zone A are relatively complete

(Jin&Wang, 2023). Location factors play a major role

in the important factors influencing resold housing

prices in Changchun, including supporting facilities,

transportation conditions, and future development

prospects (Yan & Zhao, 2021). Exploring the reasons

behind this, this article analyzes that location factors

significantly affect housing prices for the following

reasons. Location with convenient transportation, low

commuting costs, and high demand for real estate,

such as houses near subway stations with higher

prices. The well-equipped location, surrounded by

schools, hospitals, and shopping malls, provides

convenient living and enhances the value of the

property. Places with beautiful environments, such as

near parks, lakes, etc., have high living comfort and

correspondingly increase housing prices. In addition,

the concentration of resources in urban core areas and

strong economic vitality often result in high housing

prices.

3.2 Community Supporting Facilities

and Services

Community-supporting facilities and services are

also important factors affecting housing prices. The

level of community property services and the green

environment of the community can all affect pre-

owned housing prices. Generally speaking, under

other equal conditions, pre-owned houses with higher

levels of community property services and better

green environment in the community have higher

housing prices. In the study on the influencing factors

of pre-owned housing prices in Hefei City, regarding

the influence factors of sample characteristic prices in

the overall samples of the primary city area and

nonmain urban area, residential areas with

standardized property management and high greening

rates are more favored by homebuyers. Jin and Wang

(2023) used a multiple linear regression model to

analyze the influencing factors of pre-owned housing

prices in Hefei City. The non standardized

coefficients of property prices and greening rates

reached 0180 and 0.063, respectively, indicating that

pre-owned housing buyers are increasingly

concerned about the supporting services and facilities

within the community. This inspires for real estate

developers to strengthen their green environment and

improve their property service management level.

3.3 Self Conditions of Pre-Owned

Houses

The quality of pre-owned houses significantly affects

their prices. This includes factors such as decoration

level, number of rooms, availability of elevators, and

floor height. Yang et al. (2023) analyzed the

transaction data of pre-owned housing prices in

Beijing and used a regression method to calculate the

regression coefficients of the independent variable

bedroom number as 2030.524, the independent

variable living room number as 2694.756, the

independent variable decoration method (divided into

simple decoration and fine decoration) as 4097.853,

the independent variable low floor as 546.939, the

high floor as -402.045, and the independent variable

elevator as 4812.397. It can be seen that the self-

equipped conditions of pre-owned houses

significantly affect the pre-owned housing prices. In

Hefei City, the three factors of building area, number

of bedrooms, and decoration situation rank among the

top in terms of their impact on resold housing costs.

Jin and Wang (2023) calculated the regression

coefficients using a regression model, where the non

standardized coefficients of the independent variables

of number of bedrooms, number of living rooms, and

equipped elevators were 0.139, 0.020, and 0.709,

respectively. From this, it can be seen that the

condition of a pre-owned house itself is a factor that

buyers highly value.

3.4 Age and Property Rights Issues

This article studies the key factors influencing resold

housing prices, so age and property rights cannot be

ignored. In the real estate market, there are significant

differences in age and property rights between pre-

owned and non pre-owned houses. Second-hand

houses have a wide range of ages, ranging from a few

years to several decades. Increasing the age of a house

can lead to significant depreciation, affecting its

structural safety and facility performance, thereby

reducing its price. The age of the new house is zero,

and the quality is relatively more guaranteed. In terms

of property rights, pre-owned houses have diverse

properties, including commercial housing, affordable

housing, etc., with different transaction restrictions

and taxes, and may also have hidden risks of property

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy

disputes. Non pre-owned houses have clear property

rights and are mostly commercial properties. There

are relatively few property disputes, and because

there is no loss of use, the remaining years of property

rights are relatively longer, making them more

attractive in the market. In the study conducted by Jin

and Wang (2023) on the key factors impacting resold

housing prices in Hefei, the non standardized

coefficients for property ownership and housing age

were -0.296 and 0.004, respectively, using a

regression model.

3.5 The Impact of Special Factors on

Pre-Owned Housing Prices

For different regions and periods, there will be

different special factors that affect pre-owned

housing prices; that is, the impact of special factors

on pre-owned housing prices has temporal and spatial

differences. Li (2022) studied the spatiotemporal

characteristics and influencing factors of pre-owned

housing prices in Chengdu. Among the many factors

studied, the determining factors of pre-owned

housing prices were land price grade and school

district attributes. The impact on housing prices

became increasingly strong from 2012 to 2021, but

the popularity of school district housing decreased

after 2018. Zhao (2023) found that when studying the

important factors influencing resold housing prices in

Hangzhou, Other factors being equal, pre-owned

houses with high waterlogging risk will suffer a 3%

price discount, while high waterlogging risk in

surrounding areas will suppress the costs of resold

houses in that area. Li (2022) studied the influences

of the pandemic on resold housing costs in Wuhan

and learned that from January to September 2020, the

transaction costs of resold houses in Wuhan generally

decreased, with April being the highest peak. Due to

proper control of the pandemic, housing prices began

to steadily rise in October 2020. It can be seen that

special and unpredictable factors such as natural

disasters will inevitably have an influence on resold

housing costs.

4 SUGGESTIONS

In the existing literature on pre-owned housing prices,

this article finds that they generally lack discussion

on some government policy factors, and government

policies also have an undeniable influence on resold

housing costs. The restriction policy extends the time

for housing to be traded again, and a large number of

properties are locked up. When the supply of pre-

owned housing in the market decreases and demand

is relatively stable, prices are difficult to rise

significantly. And the tightening of credit policies,

such as raising loan interest rates, increasing the

repayment pressure on homebuyers, suppressing the

demand for home purchases, and thus bringing

downward adjustment momentum to pre-owned

housing prices. Therefore, in the field of research on

the influencing factors of pre-owned housing prices,

this article suggests that researchers pay more

attention to national or regional policies on pre-

owned housing, especially in the event of major

events such as the epidemic, what important measures

and policies have been taken by the government to

regulate pre-owned housing prices.

5 CONCLUSION

In summary, this study systematically analyzed and

summarized the important factors influencing pre-

owned housing prices in China. Factors such as

location, community-supporting facilities, and micro

individual conditions of pre-owned houses all have

varying degrees of influence on resold housing costs.

Previous researchers have mostly used multiple linear

regression models, while geographically weighted

regression models and random forest models also

have advantages in the study of pre-owned housing

prices.

These research findings not only provide clear

price judgment criteria for homebuyers, enabling

them to comprehensively consider various factors and

make more reasonable choices when purchasing a

house; It also provides data support for investment

decisions of real estate enterprises, helping them

optimize housing allocation and enhance market

competitiveness. At the same time, this has

significant reference value for the government to

formulate real estate regulation policies, which helps

the government to implement precise policies,

stabilize housing prices, and promote the real estate

market to develop continuously. However, the estate

market is complex and ever-changing, and we still

need to continue to pay attention to emerging

influencing factors in the future, such as the course of

city renewal and the popularization of green building

standards, which may have a potential impact on pre-

owned housing prices. We look forward to further

deepening and expanding future research, providing

more forward-looking theoretical and practical

guidance for the development of the real estate

market, and jointly promoting the Chinese estate

Study for Factors Inﬂuencing Pre-Owned Housing Prices in China

market to move towards a high-quality development

stage in stability.

REFERENCES

Chen, M. 2019. Analysis of factors influencing the second

hand housing prices of residential buildings in

Shenyang. Northern Economic and Trade, (06), 54-56.

Gao, H. 2020. Research on spatial differentiation and

influencing factors of pre-owned housing prices within

the third ring road of Wuhan City. Master's Thesis,

Wuhan University.

Guo, Y., Yu, H. & Ke, J. 2022. Research on the influencing

factors of pre-owned housing prices in Yantai City

based on covariance regression. Journal of Ludong

University (Natural Science Edition), 8(03), 232-240.

Jin, C. & Wang, X. 2023. Analysis of factors influencing

the prices of second hand houses in Hefei City based on

feature price model. Journal of Hubei Second Normal

University, (09), 53-61.

Li, C. 2022. Analysis of spatiotemporal characteristics and

influencing factors of pre-owned housing prices in

Chengdu based on multi-source data. Master's Thesis,

Fuzhou University.

Li, G., Wang, J. & Qi, J. 2024. The spatial and temporal

differentiation pattern of residential prices in Shanghai

and its influencing factors: an analysis based on a

geographically weighted regression model. Regional

Research and Development, 43(06), 74-81.

Li, Y. 2022. Analysis of the impact of the COVID-19 on

the price of pre-owned housing in Wuhan. Master's

Thesis, Huazhong Agricultural University.

Li, Y., Wang, N. & Tong, Z. 2023. Research on the

nonlinear impact of street quality on housing prices

based on random forests: a case study of Guangzhou

City. Tropical Geography, 43(08), 1547-1562.

Liu, B., Jin, Y. & Wang, S. 2017. Multiple linear regression

analysis of factors affecting pre-owned housing prices

in Nanjing. Journal of Nanjing Industrial Vocational

and Technical College, 17(01), 17-19.

Liu, J. 2016. Research on the impact of residential

characteristics on the listing price of urban second hand

houses. Master's Thesis, China University of Mining

and Technology.

Ma, X. 2023. Research on the prediction method of pre-

owned housing prices in Baotou City based on

composite models. Master's Thesis, Inner Mongolia

University of Science and Technology.

Mi, W. 2018. Exploring the influencing factors of pre-

owned housing prices. Master's Thesis, Central China

Normal University.

Shen, X. 2017. Statistical analysis of the pre-owned

housing market in Beijing. Master's Thesis, Beijing

Institute of Technology.

Sun, H. 2019. Analysis of factors influencing the prices of

second hand houses in Shijiazhuang City Hebei

Province. Master's Thesis, Tianjin University of

Finance and Economics.

Wei, X. 2021. Analysis of factors influencing the prices of

second hand houses in Nanning City based on random

forests. Light Industry Technology, 37(10), 96-98.

Yan, M. & Zhao, L. 2021. Research on the influencing

factors of pre-owned housing prices in Changchun City.

Journal of Shenyang Institute of Technology (Social

Sciences Edition), (02), 35-39.

Yang, J., Shao, X. & Peng, Z. 2023. Research on the

influencing factors of pre-owned housing prices in

Beijing. Real Estate World, (04), 25-28.

Zhao, L. 2023. The impact of waterlogging risk on real

estate prices. Master's Thesis, Zhejiang University.

Zhao, X. 2021. Research on spatial differentiation and

influencing factors of pre-owned housing prices in

Guangzhou based on multi-source data. Master's Thesis,

South China University of Technology.

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy