Available via license: CC BY-NC 4.0
Content may be subject to copyright.
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
399
The Impact of Education Level on Real Estate Asset Allocation
Wancheng Jiang *
School of Innovation and Development, Central University of Finance and Economics, Beijing,
China
* Corresponding Author Email: 2021312519@email.cufe.edu.cn
Abstract. As an important asset class, the allocation decision of real estate is of great significance
to the wealth accumulation of individuals and families. This paper builds a theoretical framework
based on the theoretical perspective of human capital, takes the CHFS database in 2015,2017 and
2019 as the benchmark, and constructs the Probit model, two-way fixed effect model and
intermediary effect model to carry out empirical testing, revealing the causal relationship and
mechanism between education level and real estate asset allocation. The results show that the
education level significantly increases the individual real estate asset allocation, and this causality is
still significantly established under multiple robustness tests. Financial information attention and
financial knowledge play an important intermediary role in this process. In addition, the impact of
educational level on real estate asset allocation was more obvious in the urban sample than in the
rural groups. This study not only enriches the theoretical basis of relevant fields, but also provides a
scientific basis for policy makers, which helps to optimize the allocation of resources in the real estate
market and promote the sustainable economic development.
Keywords: Real estate assets; education level; financial information attention; intermediary effect.
1. Introduction
In the context of global economic integration, real estate, as one of the important economic fields,
has a profound impact on the wealth accumulation and economic development of individuals, families
and even the country [1]. Especially in China, the real estate market has experienced rapid
development and has become one of the important pillars to support economic growth. However, as
the market increasingly matures and complexity, the allocation decision of real estate assets has
become more and more complex and difficult. In this context, it is particularly important to study
how individuals make real estate asset allocation decisions, and which factors influence these
decisions [2].
As an important indicator to measure the individual ability and social status, the education level
plays an important role in the real estate asset allocation. With the general improvement of education
level, people's knowledge structure and cognitive ability have been greatly expanded, which will
naturally affect their decisions on real estate asset allocation [3]. On the one hand, the improvement
of education level may enable individuals to analyze market dynamics more rationally and grasp
investment opportunities; on the other hand, education may also affect individuals' risk preference,
investment concept and other factors, thus further affecting their real estate asset allocation decisions.
Therefore, it is of great significance to deeply study the influence of education level on real estate
asset allocation to understand individual economic behavior and optimize the resource allocation in
the real estate market.
This paper will use the method of empirical research, using the CHFS database in 2015,2017, and
2019 as the benchmark to construct Probit model, two-way fixed effect model and intermediary effect
model to carry out empirical test, in order to reveal the causal relationship and the mechanism of
action between education level and real estate asset allocation. This study not only reveals the internal
relationship between education level and real estate asset allocation, but also provides a scientific
basis for policy makers and a reference for investors in decision-making.
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
400
2. Literature Review and Research Hypotheses
2.1. Education Level and Real Estate Asset Allocation
The theory of human capital is widely recognized in the academic field, and its core view is to
emphasize the decisive role of individual investment such as education, training and skills in
individual productivity and economic performance [4]. This theory regards education not only as the
accumulation of knowledge, but also as a long-term investment, through which individuals can
significantly improve their knowledge reserve, skill level and production capacity, so as to obtain
higher economic returns in the future [5]. Explore the influence of education on real estate asset
allocation. First, from the perspective of the income effect, an increase in education is usually
accompanied by an increased income. As income increases, individuals have more capital for asset
allocation without being limited to a single real estate sector [6]. They began to invest their capital in
more diverse areas, such as stocks and bonds, to diversify their assets and diversify their risks. Second,
the improvement of education is also accompanied by increased risk awareness. Higher educated
people tend to have a deeper understanding of the financial market and investment risks, which makes
them pay more attention to risk control and diversification in asset allocation [7]. This risk awareness
makes the highly educated people more cautious and rational in making real estate investment
decisions, and their real estate assets are better quality. In addition, the demand for educational
resources is also an important factor affecting the real estate asset allocation. Higher educated families
tend to pay more attention to their children's education and development, so they may be more willing
to buy houses in areas with quality educational resources [8]. This demand drives the rise of real
estate prices in specific areas, but also reflects the profound impact of educational resources on the
property market. Therefore, this paper proposes that the study hypothesis H1.
H1: Education level will increase the individual's real estate asset allocation.
2.2. Education Level, Financial Information Attention and Real Estate Asset Allocation
There is a significant positive correlation between the education level and the attention to financial
information. Specifically, with the improvement of individual education level, their attention to
financial information also shows an increasing trend. This phenomenon can be explained from
multiple perspectives. First, the deepening of education not only expands the knowledge field of
individuals, but also improves their information processing ability, enabling them to more effectively
analyze and evaluate the value of financial information. This ability makes it easier for them to
identify the financial dynamics that are closely related to their own investment decisions, which in
turn increases their attention to financial information [9]. Secondly, the improvement of education
level is also accompanied by the deepening of individual understanding of the importance of financial
knowledge, which further encourages them to actively seek and pay attention to financial information,
so as to make more sensible and reasonable decisions in financial activities such as real estate asset
allocation [10]. Therefore, it can be considered that the education level is one of the important factors
affecting the attention of financial information, which indirectly acts on the individual real estate asset
allocation behavior through the influence of information processing ability and financial knowledge
understanding. Financial information attention plays a vital role in the process of real estate asset
allocation. For those individuals who pay close attention to financial information, they are often able
to assess the risks and benefits more comprehensively when allocating real estate assets [11].
Specifically, they pay close attention to market dynamics, such as price fluctuations, supply and
demand relations, and the impact of policy changes on the real estate market. These detailed
observations and analysis enable them to predict market trends more accurately, so as to make more
rational investment decisions [12]. In addition, individuals with high attention to financial
information show significant advantages in identifying investment opportunities. They not only
deeply study market data, but also explore potential high-quality investment targets by analyzing
industry trends and interpreting economic data [13]. This keen market insight and judgment, enable
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
401
them to seize the fleeting investment opportunities in the complex real estate market. Therefore, the
study hypothesis H2.
H2: Financial information attention plays an intermediary role in the impact of educational level
on real estate asset allocation. That is, the education level will increase the individual's real estate
asset allocation by increasing the attention of individual financial information.
2.3. Education Level, Financial Knowledge and Real Estate Asset Allocation
There is a positive correlation between educational attainment and financial knowledge. With the
improvement of individual education level, their ability to obtain, process and apply financial
information is also increased accordingly [14]. On the one hand, highly educated individuals have a
deeper understanding of how the financial market operates. They can clearly grasp the supply and
demand relationship of the market, understand the internal mechanism of price formation, and have
an insight into the behavior patterns of market participants [15]. This deep understanding of the
market enables them to predict market trends more accurately and provide strong guidance for real
estate asset allocation. On the other hand, individuals with high educational levels have more accurate
cognition of different investment tools [16]. They understand the characteristics, risks and returns of
various financial instruments, and can choose the appropriate investment instruments for asset
allocation according to their own investment objectives and risk tolerance [17]. This precise cognition
enables them to flexibly allocate assets in the real estate market and realize the steady appreciation of
wealth [18]. Financial knowledge plays a crucial role in individual real estate asset allocation
decisions. Individuals with rich financial knowledge can assess the risks and benefits more
comprehensively and accurately when making investment decisions in the real estate market [19].
They not only pay attention to the price changes and rental income of the real estate market, but also
can conduct in-depth analysis from the macro economy, policy environment, market supply and
demand and other dimensions, so as to make more rational investment choices [20]. For example,
they may analyze the impact of macroeconomic policies on the real estate market and predict the
future trend of housing prices or focus on the changes in the market supply and demand relationship
and look for investment opportunities. At the same time, individuals with rich financial knowledge
are better at identifying investment opportunities in the real estate market [21]. They can find out
through market research, data analysis and other means, potential high-quality investment targets,
and grasp the investment opportunity. This keen market insight and judgment enable them to obtain
higher returns in the real estate market. Therefore, this paper proposes that the study hypothesis H3.
H3: Financial knowledge plays an intermediary role in the influence of educational level on real
estate asset allocation. That is, the education level will increase the individual's real estate asset
allocation by increasing the individual's financial knowledge.
3. Research Design
3.1. Variable Selection
The explained variable in this paper is the real estate asset allocation. This paper measures the
allocation of individual real estate assets from two aspects, 1 is whether to hold the property
(D_House), which is a virtual variable, when the individual holds the property, the variable equals to
1, otherwise the value is 0; the second is the property proportion (P_House), using the ratio of the
house assets to the total asset value as the proxy variable. Educational level is the explanatory variable
of this paper. Education level refers to the range and level of knowledge and skills acquired by a
person, which usually includes multiple fields such as culture, science, technology, society and
history. The level of education is often determined by many factors, such as school education, family
education and social environment. In the survey, the education level was divided into nine categories:
no school, primary school, primary school, middle school, senior high school, technical secondary
school (vocational high school), junior college (higher vocational college), university undergraduate,
master's degree, doctor's degree, and the value of 1-9 from small to large. This paper selects financial
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
402
information attention and financial knowledge as the intermediary variables. Financial information
attention refers to the degree of people pay attention to all kinds of financial information in economic
life, which reflects the investment willingness of individuals. In the process of variable construction,
the attention of financial information is directly defined based on issues related to CHFS. Financial
knowledge refers to an individual or organization's understanding of financial markets, financial
instruments and products, and the ability to use them to make informed financial decisions, reflecting
an individual's investment ability. Specifically, this paper analyzed the respondents' understanding
and cognition of interest rate calculation, inflation and investment risk, and used the main factor
analysis method to extract variables related to financial knowledge. Next, this paper weighted the
extracted factors according to the ratio of variance contribution rate to cumulative variance
contribution rate in the two-factor rotation matrix, to obtain a comprehensive financial knowledge
score index for comprehensively evaluate the financial knowledge level. In this paper, age, marriage,
gender, health status, family size, total family income, financial industry practitioners in the home,
risk preference, property Sharp rate and economic environment were selected to control the
interference of other factors as much as possible and ensure the unbiased nature of the regression
results. All the variables are shown in the following Table 1.
Table 1. Variable selection and definition
type
Variable name
Variable
abbreviation
Variable definition
explained
variable
Whether to own
the property
D_House
The value of the held property is 1, otherwise, it is 0
Share of property
P_House
The proportion of housing assets in the total assets
explanatory
variable
Education level
edu
Have not been to school, primary school, junior high
school, senior high school, technical secondary school
(vocational high school), junior college (higher
vocational college), university bachelor's degree,
master's degree, doctor's degree, from small to large,
assigned the value of 1-9
metavariable
Financial
information
attention
att
According to the interviewees, the high attention value is
assigned to 1, otherwise it is 0
Financial
knowledge
kno
Yin Zhichao (2014)
controlled
variable
age
age
Current year-year of birth
marriage
marriage
Married take 1, otherwise 0
sex
gender
And 1 for males and 0 for females
health condition
health
Health level 1-5 is good-bad
Family size
size
Total family population
Household
income
lnincome
Total income household log value
There are people
working in the
financial industry
fin
There is a financial industry practitioner in the home for
1, otherwise it is 0
Risk appetite
risk
Degree of risk preference 1 – 5 corresponds to no-
preference-preference
Real estate Sharp
rate
Sha
Property Sharpe rate index
economic
environment
lnPGDP
The log of per capita GDP value of the province
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
403
3.2. Data Sources
The individual-level data in this paper are all obtained from the CHFS database, namely the China
Household Finance Survey (China Household Finance Survey) database, which is a nationally
representative household-level financial database. The CHFS database is operated and managed by
the China Family Finance Survey and Research Center and aims to comprehensively collect the
financial information of Chinese families through scientific sampling survey methods. The CHFS
database survey covers household financial assets and non-financial assets, including housing assets,
liabilities and credit constraints, income, consumption, social security and insurance, demographic
characteristics, employment, and payment habits. This comprehensive survey content enables the
CHFS database to comprehensively and carefully depict the family economic and financial behavior.
The sample data of CHFS database is scientific and accurate. The stratification, the sampling design
method proportional to scale measurement (PPS) is adopted in data collection, and the sample data is
collected and updated in combination with field visits and quarterly telephone visits. Economic and
environmental data at the provincial level are derived from the China Statistical Yearbook.
The obtained raw data are processed as follows. (1) The absolute value data of a large magnitude
is log-treated to alleviate the possible heteroscedasticity problem caused by the large difference in
magnitude. Such as total family income, etc. (2) All continuous variables were bilateral tailed at the
1% level to alleviate the interference caused by extreme values as much as possible. In this paper,
using the combined CHFS databases from 2015,2017,2019 and 2019,17,254 valid samples were
obtained.
3.3. Descriptive Statistical Analysis
The basic descriptive statistics of the variables are shown below in Table 2.
Table 2. Descriptive Statistical Analysis
VarName
Obs
Mean
SD
Min
Median
Max
D_House
17254
0.95
0.22
0
1
1
P_House
17254
0.87
0.21
0
1
1
edu
17254
3.31
1.51
1
3
9
div
17254
0.16
0.15
0
0.162
1
att
17254
0.16
0.32
0
0
1
age
17254
54.25
12.61
26
54
83
marriage
17254
0.89
0.31
0
1
1
gender
17254
0.79
0.41
0
1
1
health
17254
2.68
0.97
1
3
5
size
17254
3.46
1.56
1
3
8
lnincome
16865
10.50
1.67
0
10.820
13.140
fin
17254
0.01
0.08
0
0
1
risk
17254
2.10
2.20
0
1
5
Sha
17254
0.10
0.03
0
0.107
0.152
lnPGDP
17254
10.97
0.43
10.172
10.929
12.009
3.4. Model Construction
Since the data sample used in this paper is family survey data across years, it is appropriate to
choose a panel effect model. There is dual heterogeneity between years and individuals, thus, in this
paper, the empirical test will be carried out on the basis of this model. Specifically, since whether to
hold the property (D_House) is a virtual variable, this paper selects Probit as the benchmark, as shown
in (1), (2) and (3) respectively.
(1)
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
404
(2)
(3)
To calculate the probability, the relationship with the satisfaction formula (2) is the conversion
function. i mean the ith survey user, and t is the t-th year. Is one of the cores explained variables,
indicating whether the ith respondent holds the property in the t year. Is the constant term in the model.
It is the core explanatory variable, the corresponding coefficient to be estimated, and the expected
result is positive, that is, the education level helps to increase the probability of holding a property.
For the many control variables selected in this paper, they are individual fixed effects, year fixed
effect, and random disturbance terms.
(4)
Property ratio (P_House) is a continuous variable, so OLS is selected as the benchmark model in
this paper, and the above is the model setting. The proportion of the property in the t-year of the i th
survey respondents. Is the constant term in the model. Still the core explanatory variable, the core
coefficient concerned here, and the expected result is positive. Other parameters remain the same.
(5)
(6)
The mediating effect model aims to identify whether there is a mediating variable in the causal
relationship connecting the independent variable and the dependent variable, the independent variable
has an influence on the dependent variable by influencing the mediating variable. This paper
introduces a three-step mediation effect to complete the analysis. The specific model setting is shown
in Equations (5) and (6). Equation (5) will test whether the individual education level is promoted
(investment attention and investment ability), reflected by the coefficient. Equation (6) will test
whether it promotes its real estate asset allocation, including two aspects, reflected by the coefficient.
If the coefficient and significant are both, it indicates that the intermediary effect of investment
attention and investment ability exists.
4. Empirical Results and Analysis
4.1. Benchmark Regression
Benchmark regression results are shown below in Table 3. The regression results for uncontrolled
variables were reported in columns (1) and (2) and in columns (3) and (4). In the absence of controlled
variables, the influence coefficient of education on D_House and P_House was 0.0770 and 0.0092,
respectively, which passed the significance test at the 1% level, indicating that education would
increase the probability and ratio of individual real estate asset allocation. After the addition of control
variables, the influence coefficient of education level on D_House and P_House was significantly
0.1056 and 0.0111, respectively, which once again proves the positive impact of education level on
real estate asset allocation. The study hypothesis that H1 was confirmed.
The improvement of education level significantly affects the individual real estate asset allocation
strategy. With the improvement of education level, the income level of individuals generally increases,
and financial literacy and risk awareness are also enhanced correspondingly enhanced, which makes
them more rational and comprehensive in asset allocation [6]. In addition, the demand for educational
resources from highly educated families also promotes their decision to buy houses in quality school
districts, further increasing the allocation of real estate assets [8]. Therefore, the improvement of the
education level not only improves the economic ability of individuals, but also changes their
investment philosophy, and ultimately increases the proportion and strategic choice of individuals in
the real estate asset allocation.
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
405
Table 3. Benchmark regression results
(1) D_House
(2) P_House
(3) D_House
(4) P_House
edu
0.0770***
0.0092***
0.1056***
0.0111***
(5.5367)
(8.1081)
(6.5716)
(8.7150)
age
0.0019
-0.0005***
(1.0873)
(-3.5000)
marriage
0.1412**
0.0016
(2.2933)
(0.2861)
gender
-0.0259
-0.0058
(-0.5352)
(-1.4062)
health
-0.0099
0.0039**
(-0.4792)
(2.2314)
size
0.0679***
0.0080***
(4.6973)
(6.7703)
lnincome
-0.0397***
-0.0051***
(-2.8456)
(-4.8751)
fin
-0.0162
-0.0303
(-0.0619)
(-1.4667)
risk
-0.0233
0.0037*
(-0.9721)
(1.9145)
Sha
-0.4475
0.1231**
(-0.7751)
(2.4438)
lnPGDP
-0.1796***
-0.0060
(-3.5828)
(-1.3377)
_cons
1.6633***
0.8433***
3.6559***
0.9187***
(25.5405)
(194.9184)
(6.5612)
(18.6605)
Note: * * * p <0.01, * * p <0.05, * p <0.1, with t statistics in parentheses
4.2. Robustness Test
Multiple forms form robustness tests of replacement of explained variables, fixed effects of control
province, and PSM. Specific instructions are as below in Table 4. Measurement error is one of the
important reasons for the model regression bias. Changing different variable measurement methods
is helpful to alleviate the estimation bias caused by measurement error, and further strengthen the
effectiveness of the empirical results. Specifically, this paper chooses the log value of housing assets
as the proxy variable of real estate asset allocation here and carries out the regression analysis again.
The results are as follows. The influence coefficient of education level on the log value of housing
assets is 0.3382, which is positively significant at the level of 1%, indicating that the education level
will significantly increase the real estate asset allocation of individuals, which is well consistent with
the previous one.
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
406
Table 4. Replacement was explained by explanatory variable robustness test
Coefficient
Coefficient
edu
0.3382***
lnincome
0.1236***
(27.6874)
(12.8552)
age
0.0014
fin
-0.0110
(0.9775)
(-0.0572)
marriage
0.2324***
risk
-0.0114
(4.3607)
(-0.6675)
gender
-0.2222***
Sha
3.3028***
(-5.7657)
(6.7974)
health
-0.0796***
lnPGDP
0.7161***
(-4.9132)
(16.3492)
size
0.0940***
_cons
1.3677***
(8.5193)
(2.8429)
Changing the regression model can provide multiple estimation results under different model
Settings, thus strengthening the robustness of the regression results. Therefore, this paper adds the
provincial fixed effect based on the original year and the individual bidirectional fixed effect,
conducts the regression analysis again. The results are shown in Table 5. The influence coefficients
of education level on D_House and P_House were significantly 0.1178 and 0.0116, respectively,
which once again proved the positive impact of education level on real estate asset allocation.
Table 5. Control the province fixed effect
(1)
(2)
D_House
P_House
edu
0.1178***
0.0116***
(7.1148)
(9.0251)
age
0.0024
-0.0005***
(1.3519)
(-3.1795)
marriage
0.1310**
0.0016
(2.1216)
(0.2792)
gender
-0.0362
-0.0064
(-0.7410)
(-1.5335)
health
-0.0073
0.0039**
(-0.3523)
(2.2217)
size
0.0664***
0.0075***
(4.4944)
(6.2478)
lnincome
-0.0403***
-0.0049***
(-2.8516)
(-4.6314)
fin
-0.0551
-0.0338
(-0.2099)
(-1.6383)
risk
-0.0220
0.0041**
(-0.9190)
(2.1215)
Sha
-2.1575
0.1876
(-1.4658)
(1.5893)
lnPGDP
0.2645
-0.0095
(1.3957)
(-0.6285)
_cons
-1.3021
0.9488***
(-0.6054)
(5.5040)
lnsig2u
-0.9712***
(-4.6780)
N
16864
16865
Note: * * * p <0.01, * * p <0.05, * p <0.1, with t statistics in parentheses
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
407
Self-selection bias is also one of the important factors that may interfere with the regression results
of this paper. Individuals with better education may naturally show differences in property asset
allocation with individuals with poorer education due to differences in other factors. The propensity
score matching method (Propensity Score Matching, PSM) is a widely used method in economics to
deal with self-selection bias. It is mainly used to compare the correlation between two groups of data,
detect the deviation between two groups of data, and compare the difference between two groups of
data. In this paper, PSM will be used to select the individuals with the most similar characteristics to
alleviate self-selection bias. Specifically, this paper chose the near-neighbor matching method,
calculated the propensity score of the sample with the probit model, and selected all control variables
as covariates of the PSM. First, the matching effect is tested in this paper, and the results are shown
in the Fig. 1. It can be seen that the black dot symbol in the figure reports the degree of deviation
between the covariates before matching, while the symbol reports the degree of deviation between
the covariates after matching. After completing the matching, the deviation degree of all covariates
decreased significantly, and was basically less than 10%, which indicates that the matching effect
was good and better alleviated the problem of self-selection bias.
Fig. 1 Matching effect analysis
In the regression analysis, the influence coefficient of education on D_House and P_House was
significantly 0.1056 and 0.0110, respectively, which again proved the positive impact of education
on real estate asset allocation (See Table 6).
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
408
Table 6. The PSM robustness test
(1)
(2)
D_House
P_House
edu
0.1056***
0.0110***
(6.5415)
(8.5947)
age
0.0016
-0.0005***
(0.9186)
(-3.3779)
marriage
0.1528**
0.0010
(2.4528)
(0.1797)
gender
-0.0192
-0.0060
(-0.3934)
(-1.4346)
health
-0.0139
0.0034*
(-0.6653)
(1.9340)
size
0.0671***
0.0082***
(4.5873)
(6.8148)
lnincome
-0.0379***
-0.0049***
(-2.6736)
(-4.5597)
fin
-0.0209
-0.0304
(-0.0799)
(-1.4719)
risk
-0.0238
0.0033*
(-0.9860)
(1.7120)
Sha
-0.3652
0.1281**
(-0.6254)
(2.5112)
lnPGDP
-0.1782***
-0.0065
(-3.5310)
(-1.4559)
_cons
3.6288***
0.9244***
(6.4717)
(18.6659)
lnsig2u
-0.9371***
(-4.5573)
N
16618
16619
Note: * * * p <0.01, * * p <0.05, * p <0.1, with t statistics in parentheses
4.3. Inspection of Financial Information Attention and Financial Knowledge Mechanism
First, this paper will test the possible intermediary role of financial information attention in the
real estate asset allocation affected by education level. The regression results are shown as follows.
In column (1), the influence coefficient of education level on the attention to financial information
was 0.0208, which passed the significance test of 1% level, indicating that the education level will
significantly increase individuals' attention to financial information. In columns (2) and (3) of
financial information attention, the influence coefficient of D_House and P_House was 0.0401 and
0.0729 respectively, both of which passed the significance test of 1% level, indicating that the
financial information attention will significantly increase the probability and proportion of individual
real estate asset allocation. The results together show that the educational level will significantly
increase the attention of individuals to financial information, thus increasing the probability and
proportion of individual real estate asset allocation, assuming that H2 is confirmed (See Table 7).
The education level affects the individual financial information attention, and then affects the
allocation of real estate assets. Specifically, individuals with high educational level pay more attention
to financial information because they have higher financial literacy and information processing ability.
This attention allows them to more comprehensively assess risks and returns, identify investment
opportunities and optimize their portfolios when allocating their real estate assets. Therefore,
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
409
financial information attention plays an important intermediary role in the process of real estate asset
allocation affected by education level.
Table 7. Intermediary effect of financial information attention
(1)
(2)
(3)
att
D_House
P_House
edu
0.0208***
0.0075***
0.0096***
(11.1909)
(5.5566)
(7.5756)
att
0.0401***
0.0729***
(7.2978)
(14.5308)
age
0.0011***
0.0001
-0.0006***
(4.7795)
(0.5420)
(-4.0165)
marriage
0.0124
0.0149**
0.0008
(1.4798)
(2.4476)
(0.1434)
gender
0.0076
-0.0017
-0.0064
(1.2218)
(-0.3815)
(-1.5375)
health
-0.0075***
-0.0005
0.0044**
(-2.8574)
(-0.2429)
(2.5420)
size
0.0010
0.0060***
0.0079***
(0.5619)
(4.6879)
(6.7560)
lnincome
0.0033**
-0.0034***
-0.0054***
(2.0856)
(-2.9803)
(-5.1461)
fin
0.1474***
-0.0079
-0.0404**
(4.7900)
(-0.3553)
(-1.9686)
risk
-0.0346***
-0.0006
0.0061***
(-11.6753)
(-0.2842)
(3.1736)
Sha
0.1155
-0.0496
0.1133**
(1.5747)
(-0.9275)
(2.2632)
lnPGDP
-0.0057
-0.0171***
-0.0056
(-0.8824)
(-3.6324)
(-1.2629)
_cons
0.2098***
1.1063***
0.9043***
(2.9494)
(21.2370)
(18.4790)
N
16865
16865
16865
Note: * * * p <0.01, * * p <0.05, * p <0.1, with t statistics in parentheses
Financial knowledge may also play an intermediary role in the real estate asset allocation
influenced by the education level, and the regression results are as follows. In column (1), the
influence coefficient of education level on financial knowledge was 0.0152, which passed the
significance test of 1% level, indicating that the education level will significantly increase the
individual's financial knowledge. In columns (2) and (3), the influence coefficient of financial
knowledge on D_House and P_House was 0.0944 and 0.0108 respectively, both of which passed the
significance test of 1% level, indicating that financial knowledge will significantly increase the
probability and proportion of individual real estate asset allocation (See Table 8). Together, the results
show that educational level significantly increases individual financial knowledge, thus increasing
the probability and proportion of individual real estate asset allocation, assuming that H3 is confirmed.
In the process of real estate asset allocation influenced by education level, financial knowledge
plays a key intermediary role. High educated individuals are better able to understand and cope with
the complexity and uncertainty of the real estate market through accumulating financial knowledge.
They can understand how the real estate market works and predict market trends to make more
accurate investment decisions. At the same time, they can also effectively manage investment risks
and ensure the safety and steady appreciation of assets. This influence mechanism shows that
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
410
improving the level of individuals' financial knowledge will not only help to optimize their real estate
asset allocation strategy, but also can improve their wealth accumulation level.
Table 8. Mediating effect of investment ability
(1)
(2)
(3)
div
D_House
P_House
edu
0.0152***
0.0944***
0.0108***
(17.0131)
(5.8062)
(8.4062)
kno
0.9107***
0.0200*
(6.9031)
(1.8913)
age
0.0006***
0.0014
-0.0005***
(5.1833)
(0.8205)
(-3.5656)
marriage
0.0134***
0.1276**
0.0013
(3.3041)
(2.0497)
(0.2376)
gender
0.0058*
-0.0328
-0.0059
(1.9218)
(-0.6716)
(-1.4239)
health
-0.0013
-0.0099
0.0039**
(-0.9891)
(-0.4765)
(2.2475)
size
-0.0045***
0.0722***
0.0081***
(-5.3035)
(4.9392)
(6.8353)
lnincome
0.0052***
-0.0453***
-0.0052***
(6.7393)
(-3.1932)
(-4.9592)
fin
0.0160
-0.0456
-0.0305
(1.0812)
(-0.1741)
(-1.4767)
risk
-0.0009
-0.0217
0.0037*
(-0.6319)
(-0.8998)
(1.9224)
Sha
-0.1686***
-0.3260
0.1264**
(-4.7523)
(-0.5584)
(2.5068)
lnPGDP
-0.0162***
-0.1678***
-0.0057
(-5.2093)
(-3.3187)
(-1.2660)
_cons
0.2259***
3.5077***
0.9142***
(6.5587)
(6.2482)
(18.5343)
lnsig2u
-0.9125***
(-4.5257)
N
16865
16864
16865
4.4. Urban-rural Heterogeneity Analysis
The difference between urban and rural environment may make the impact of education level on
real estate asset allocation show a difference. This difference not only reflects the difference of social
and economic environment between urban and rural areas, but also profoundly reveals the important
role of education in wealth accumulation and asset allocation in modern society. Therefore, according
to the individual permanent residence of town or village, the samples are divided into town and village,
and carry out the regression analysis again, and the results are as follows. In the town sample, the
influence coefficient of edu on D_House and P_House was 0.1513 and 0.0214, respectively, which
both passed the significance test at the 1% level. In the village sample, the influence coefficient of
edu on D_House and P_House were 0.1241 and 0.0095, respectively, which also passed the
significance test at the 1% level (See Table 9). The effect of edu on D_House and P_House is more
pronounced in towns than in the rural sample. That is, compared with the rural sample, the impact of
education level on real estate asset allocation is more prominent in the urban sample.
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
411
First of all, the influence of education degree on real estate asset allocation in the urban sample is
more prominent, which can be analyzed from multiple dimensions. In the urban environment,
educational resources are relatively concentrated, and the education level is generally high, which
enables individuals to have a more comprehensive understanding of the basic knowledge, investment
concepts and risk management strategies of the real estate market in the process of receiving
education. Therefore, urban residents with higher education often show more rationality and
professionalism when allocating real estate assets and can better make reasonable investment
decisions according to their own economic conditions and market trends. Second, the urban real estate
market is becoming more mature and stable, which provides a broader investment space for highly
educated individuals. The urban real estate market has a more perfect legal and regulatory system, a
more abundant investment varieties and a more active trading market, and these conditions provide
more choices for well-educated investors. At the same time, with the continuous advancement of
urbanization, the potential of the urban real estate market is gradually being released, which brings
more investment opportunities and higher investment returns to investors. In contrast, the impact of
education on real estate asset allocation was relatively weak. This is mainly because of the relative
lack of educational resources in rural areas, and the education level is generally low, which makes it
difficult for rural residents to understand and grasp the real estate market information. In addition,
the rural real estate market is relatively small, and the transaction activity is not high, which also
limits the choice space of rural residents in the real estate asset allocation.
Table 9. Analysis of heterogeneity
(1)
(2)
(3)
(4)
town
village
town
village
D_House
D_House
P_House
P_House
edu
0.1513***
0.1241***
0.0214***
0.0095***
(4.6456)
(5.9635)
(7.8304)
(6.0672)
age
0.0019
0.0025
-0.0005**
-0.0005***
(0.7352)
(1.0580)
(-1.9955)
(-2.6723)
marriage
-0.1507
0.2696***
-0.0098
0.0057
(-1.4763)
(3.3461)
(-1.0902)
(0.7777)
gender
-0.2214**
-0.0127
-0.0206***
-0.0028
(-2.4077)
(-0.2082)
(-2.5888)
(-0.5564)
health
-0.0320
0.0031
0.0090***
0.0008
(-1.1600)
(0.1047)
(3.5093)
(0.3356)
size
0.0172
0.1076***
0.0034**
0.0121***
(0.9812)
(4.7633)
(2.0479)
(7.1184)
lnincome
-0.0312*
-0.0317
-0.0043***
-0.0053***
(-1.6531)
(-1.6398)
(-2.7202)
(-3.7673)
fin
0.0000
-0.0039
0.0962
-0.0295
(.)
(-0.0134)
(0.9405)
(-1.3689)
risk
-0.0698*
0.0001
0.0001
0.0054**
(-1.9358)
(0.0027)
(0.0359)
(2.2161)
Sha
-0.1335
-0.4231
0.3298***
0.0499
(-0.1592)
(-0.5397)
(4.1355)
(0.7580)
lnPGDP
-0.0510
-0.2179***
0.0083
-0.0095*
(-0.6583)
(-3.2438)
(1.1202)
(-1.6525)
_cons
2.7398***
3.6161***
0.7497***
0.9534***
(3.0842)
(4.8721)
(8.9744)
(15.0892)
lnsig2u
-4.8592
-0.4732**
(-0.4585)
(-2.3170)
N
6726
10134
6730
10135
Note: * * * p <0.01, * * p <0.05, * p <0.1, with t statistics in parentheses
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
412
5. Conclusions and Suggestions
5.1. Conclusion
As an important asset class, real estate has a significant impact on the accumulation of wealth of
individuals and families. As an important indicator to measure individual ability and social status, its
role in real estate asset allocation is increasingly attracting attention. It is particularly important to
study the impact of education level on real estate asset allocation. In this context, this paper deeply
studies the impact of education level on real estate asset allocation, builds a theoretical framework
based on the theoretical perspective of human capital theory, and constructs the Probit model, two-
way fixed effect model and intermediary effect model based on the CHFS database in 2015,2017 and
2017 as the benchmark to carry out empirical tests. The main study conclusions are summarized as
follows:
(1) The education level significantly increases the individual real estate asset allocation. The
human capital theory provides some explanations from the aspects of income effect, risk awareness
and educational resource demand, and the empirical results of this paper correspond to them, and the
conclusion still holds significantly after replacing the various forms of robustness test of the explained
variables, the fixed effects of the control province and PSM.
(2) Financial information attention and financial knowledge play an intermediary role in the
influence of education level on real estate asset allocation. That is, the education level will increase
the individual real estate asset allocation by increasing the individual financial information attention
and financial knowledge.
(3) Compared with rural groups, the impact of education level on real estate asset allocation is
more obvious in the urban sample.
5.2. Recommendations
Attach importance to the role of education in improving the ability of real estate asset allocation.
The government should continue to increase investment in education and improve the overall level
of education, especially financial education, so as to enhance the public's understanding of the real
estate market and asset allocation ability. Educational institutions can add more courses related to the
real estate market, such as real estate investment, financial analysis, etc., to cultivate students'
investment awareness and risk identification ability. All sectors of society should encourage and
support individual lifelong learning, and constantly improve their ability of self-education and self-
investment.
To enhance the public's attention to financial information and financial knowledge level. Media
and financial institutions should strengthen the information disclosure and interpretation of the real
estate market and improve the public's attention to and understanding of financial information. The
government can set up a financial knowledge popularization platform or organize lectures on financial
knowledge to help the public improve their financial literacy and reduce investment blindness. When
providing financial services, financial institutions should strengthen the financial knowledge
education of customers to ensure that customers make decisions on the premise of fully understanding
the market risks and investment opportunities.
Pay attention to the differences between urban and rural areas and optimize the urban real estate
asset allocation environment. The government should strengthen the supervision of the urban real
estate market to ensure the standardized operation of the market and reduce the investment risks. For
rural areas, the government should guide and support the development of rural economy, improve the
income level of rural residents, and provide more investment channels and opportunities for rural
residents. At the same time, the government can promote the integrated development of urban and
rural areas, strengthen the economic ties and exchanges between urban and rural areas, and narrow
the gap in education and asset allocation between urban and rural areas.
Highlights in Business, Economics and Management
PGMEE 2024
Volume 41 (2024)
413
References
[1] Ullah F, Al-Turjman F. A conceptual framework for blockchain smart contract adoption to manage real
estate deals in smart cities. Neural Computing and Applications, 2023, 35(7): 5033-5054.
[2] Garriga C, Gete P, Tsouderou A. The economic effects of real estate investors. Real Estate Economics,
2023, 51(3): 655-685.
[3] Zhang Elegant. Study on the mechanism of education influence on family risk financial asset allocation.
Economic longitude, 2023,40 (02): 140-149.
[4] Hu Yaozong, Wang Feng, Yao Hao. Human capital theory Sinicization and its innovation - Mr.Jin Xibin
Human capital Thought and Value. Education Science in China (Chinese and English), 2023,6 (04): 143-
146.
[5] Lu Wenhui. Explain the problem of China's economic transformation from the comparison of human
capital theory and human feasible ability theory. Exploration of economic issues, 2011, (11): 20-24.
[6] Zhang Elegant. Study on the mechanism of education influence on family risk financial asset allocation.
Economic longitude, 2023,40(02):140-149.
[7] Ding T, Li Y, Zhu W. Can Digital Financial Inclusion (DFI) effectively alleviate residents’ poverty by
increasing household entrepreneurship? –an empirical study based on the China Household Finance
Survey (CHFS). Applied Economics, 2023, 55(59): 6965-6977.
[8] CAI Wenbo, Gan Xueyan. Analysis of uncertainty and stickiness information of investment behavior in
family shadow education - Empirical study based on CHFS survey data. Review of Educational
Economics, 2021,6 (06): 109-128.
[9] Peng D. Financial Market Participation under the Digital Divide: An Empirical Study Based on CHFS
Data. Frontiers in Business, Economics and Management, 2023, 8(1): 127-133.
[10] Yang Z, Gao J, Yu D. Cognition ability, financial advice seeking, and investment performance: New
evidence from China. International Review of Finance, 2024.
[11] Wang S. The Household Head's Educational Level and Financial Decision-Making.
[12] Wei Zhao, Song Quanyun. Household asset allocation under Internet finance. Finance and Economics
Science, 2016, (07): 52-60.
[13] Wu Zaihua, Xie Beichen, Guo Xinhua. Lending constraints, financial literacy and the mystery of "limited
participation" in the Chinese household stock market. Modern Finance and Economics (Journal of Tianjin
University of Finance and Economics), 2017,37 (12): 20-35.
[14] Pang Ming, Wang Hui, Zhang Zhihao. Digital Financial Inclusion, Financial Literacy, and Family
Financial Asset Selection. Friends of the Accounting, 2024, (10): 40-48.
[15] Chen F, Jiang G, Gu M. Financial knowledge and responsible credit card behavior: exploring mediators
and moderators. International Journal of Bank Marketing, 2024.
[16] Zhou C. The Impact of Financial Literacy on Family Venture Capital--Analysis Based on 2019 CHFS
Data. Highlights in Business, Economics and Management, 2023, 20: 730-741.
[17] Zheng Lu, Xu Minxia. Does the traditional family concept inhibit the participation of urban residents in
commercial endowment insurance? - An Empirical Analysis from the perspective of Financial Trust and
Financial Literacy. Financial Research, 2021, (06): 133-151.
[18] Chen Q, Kang X, Wang Y. Financial Literacy and Consumption Inequality of Chinese Households -
Evidence from CHFS Data.2023.
[19] Tan Yanzhi, Peng Qianrui. Financial competence, financial decision-making, and poverty. Economic
Theory and Economic Management, 2019, (02): 62-77.
[20] Aliya, Tingting Wang, Han Fuling. The perspective of regional differences in financial capacity, credit
constraint mitigation and household consumption. Financial Review, 2021,13 (04): 92-110 + 126.
[21] Jia Xianjun. How does financial knowledge affect families' participation in the financial management
market? - Empirical analysis based on the CHFS data. Economic longitude, 2020,37 (04): 159-167.
