University of Southern California Dissertations and Theses

An empirical analysis of the relationship between real estate investment and regional economy in China

(USC Thesis Other)

An empirical analysis of the relationship between real estate investment and regional economy in China

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

An Empirical Analysis of the Relationship
between Real Estate Investment and
Regional Economy in China

Xing Ming

Dissertation Submitted to
Dana and David Dornsife College of Letters, Arts and Sciences
Sol Price School of Public Policy
University of Southern California
in Fulfillment of the Requirements for the Degree of
Master of Arts in Economics/Master of Planning

Committee
Joel David, Assistant Professor, Chair
Richard K. Green, Professor
Harrison Cheng, Professor
Degree Conferral Date: August 11, 2015
1

Abstract

China has gone through dramatic development in the last decade. Since 2010, China Gross
Domestic Products (GDP) surpassed Japan as the world’s second-largest economy. That fact
showed a signal of the economy recovery from the recession, and leads a number of literatures
to work on the triggers to the Chinese economy. There is no doubt that real estate investment,
as a component of GDP, plays a fundamental role in economic growth.

Many literatures have discussed the topic of real estate investment and economic growth from
different perspectives. In this paper, the author will go over some researches and studies in the
related field. Some professionals started their research from testing the Granger causality
between investment and economic growth. Other professionals focused on figuring out the
mechanism between these two sectors and how these two sectors influence on each other. In
this paper, the author will also discuss the topic from above two ways.

In the first section, the author will develop an empirical analysis based on the provincial data in
China from 2000 to 2013. Firstly, the author will employ a cluster analysis by dividing the 31
provinces into East, Middle and West districts based on their locations. In the second section,
the author will discuss how to analyze the relationship and Granger causality between real
estate investment and economic growth theoretically. The author will reexamine the Granger
causality between real estate investment and economic growth in each district, which will be
regarded as the basis of further analysis.

On the basis of theoretical and empirical analysis in previous two sections, the author will
investigate the impact what real estate investment has on urban development and vice versa.
In conclusion, real estate investment can Granger cause the economic growth, while economic
growth can also Granger cause the real estate investment. Furthermore, during the study
period, real estate investment has observable influence on economic growth in China. Last but
not least, the author would like to discuss what factors could be potentially involved in the
relationship between real estate development and economic growth based on qualitative
analysis, and put forward some recommendation for policy makers.

Keywords: Real Estate Investment, Economic Growth, Panel Data, Granger Causality
2

Acknowledgements

I would like to acknowledge and express my sincere thanks and gratitude to all those people
who help, support and guide me toward the successful completion of my dissertation.

I thank every committee member who would like to offer me valuable ideas of the topic and
give me suggestions, comments and corrections towards the completion of my dissertation.

I would like to express my earnest thanks to the chair of my committee, Professor Joel David,
who would like to help me and serve as the advisor of my dissertation.

I would like to extend my heartfelt appreciation to Professor Richard Green for giving me
constructive and truly helpful instructions, encouragements and guidance.

I would like to express my sincere gratitude Professor Harrison Cheng for sparing his previous
time to be a member of my committee.

Special thanks to my family for their immense love, encouragement, and trust in me. I would
also like to thank to my friends for giving me spiritually support and encouragement towards
the completion to my dissertation.

3

Table of Contents
Abstract ................................................................................................................................ 1
List of Figures ........................................................................................................................ 4
List of Tables ......................................................................................................................... 4
Chapter 1 Introduction .......................................................................................................... 5
Chapter 2 Literature Review .................................................................................................. 8
Chapter 3 Theoretical Model and Methodology ................................................................... 10
3.1 Empirical Model...................................................................................................................... 10
3.2 Granger Causality ................................................................................................................... 11
Chapter 4 Empirical Analysis and Results ............................................................................. 13
4.1 Data Resources ....................................................................................................................... 13
4.2 Empirical Analysis ................................................................................................................... 13
4.2.1 Cluster Analysis ......................................................................................................................... 13
4.2.2 Regression Results in Nationwide ............................................................................................ 14
4.2.3 Regression Results in the East District ..................................................................................... 15
4.2.4 Regression Results in the Middle District ................................................................................. 17
4.2.5 Regression Results in the West District .................................................................................... 18
4.3 Granger Causality Results ........................................................................................................ 19
4.3.1 Unit Root Test ............................................................................................................................ 19
4.3.2 Granger Causality Results ......................................................................................................... 20
Chapter 5 Conclusions ......................................................................................................... 24
References .......................................................................................................................... 26

4

List of Figures

Figure 1.1 National Annual Growth Trends of GDP and Investment
Figure 1.2 Percentages of GDP and Real Estate Investment
Figure 1.3 National Annual Growth Rates of GDP and Investment

List of Tables

Table 3.1 Descriptions of Variables
Table 4.1 Descriptions of East, Middle, and West Districts
Table 4.2 Regression Results in Nationwide
Table 4.3 Regression Results in the East District
Table 4.4 Regression Results in the Middle District
Table 4.5 Regression Results in the West District
Table 4.6 Unit Root Test
Table 4.7 Unit Root Test in the East, Middle, and West Districts
Table 4.8 Granger Causality Test Results in Nationwide
Table 4.9 Granger Causality Test Results in the East District
Table 4.10 Granger Causality Test Results in the Middle District
Table 4.11 Granger Causality Test Results in the West District
5

Chapter 1 Introduction

Within the past decade, Chinese economy experienced a boom along with rapid urbanization
process. As a significant indicator of economic growth, Gross Domestic Product (GDP) follows
an upward trend in the past decade. Meanwhile, real estate development within the past
decade also experienced a rapid increase. A lot of real estate development projects are going
on in China, especially in major cities, like Beijing and Shanghai. As an important component of
the Gross Domestic Product (GDP) in China, it is necessary to study the relationship between
GDP growth and real estate development. Even though China is experiencing rapid increasing in
GDP with the real estate boom, it is evitable whether the real estate boom would make a
positive or negative contribution to the overall urban development in a short term or a long
term. There is no doubt that real estate is now a hot topic for many economists and literatures.

Figure 1.1 National Annual Growth Trends of GDP and Investment

As it is shown in Figure 1.1, GDP, total investment in fixed assets and real estate investment
have upward trend in the past decade. Both GDP and real estate investment are experiencing
an increase during the period of 2000 to 2013. The percentages of real estate investment of
GDP showed in Figure 1.2 also represent an upward trend in both nationwide and three
districts. It indicates the important role of real estate industry in the whole economic growth.
$-
$100,000
$200,000
$300,000
$400,000
$500,000
$600,000
$700,000
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
GDP Total Investment in Fixed Assets Real Estate Investment
6

Figure 1.2 Percentages of Real Estate Investment of GDP

Figure 1.3 National Annual Growth Rates of GDP and Investment

0.00%
5.00%
10.00%
15.00%
20.00%
25.00%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
East Middle West Nation
0.00%
5.00%
10.00%
15.00%
20.00%
25.00%
30.00%
35.00%
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Growth Rate of GDP Growth Rate of Real Estate Investment
7

Furthermore, taking a closer look of GDP growth rate and real estate investment growth rate
(Figure 1.3), the author find that GDP and real estate investment have been experiencing
similar path of growth from 2000 to 2013. That fact indicates the importance of researching on
the relationship between regional economy and real estate industry. In this paper, the author
discussed the relationship between real estate development and China’ economic growth on
both national and provincial basis. The whole paper is divided into following 4 sections.

First of all, the author went over the work and papers literature and professionals did in past
decades. Many economist and literatures studies the economic growth either in a city or in a
country. Some literatures also study created models to study what would be the indicators and
driving forces for the economic growth. For the literatures that study the real estate
development, they point out the important link between real estate development and
economic growth, which would be a foundation of this paper.

Secondly, the author discussed the empirical economic model based on previous researches
and the methodology of Granger causality test. In this section, the author focused on the
process of creating the empirical model and the variables involved. The author would propose
three hypotheses with regards to current situations in China and would discuss in more details
in the next section.

Thirdly, the author applied an empirical analysis, using the model created in the above section.
The data included panel data and time series data in both national degree and provincial
degree. The author retrieved data from National Bureau of Statistics of China. In the end, the
author drew a conclusion based on the results of regression and Granger causality results.

Last, the author provided some recommendations with reference of the previous results and
finding. The author further discussed how the newly released policies would influence on the
future real estate market and the economic growth trend in a macro and quantitative scale.

8

Chapter 2 Literature Review

Considering the endogenous variables, Mills created a simultaneous equations model to discuss
whether the United States has overinvested in housing (Mills, 1987). In the paper, Mills divided
the total output into housing investment and everything else. Based on data of national income
accounts and capital stock from 1929 to 1983, he introduced lagged value for investment
capital stock and run an empirical analysis. He made a conclusion that return of non-housing
capital is higher than that of housing capital.

Green came up with an idea to test the relationship between residential and non-residential
investment and GDP via Granger tests (Green, 1997). In the paper, Green used quarterly
National Income and Products Data from 1959 to 1992 in US to examine whether the
residential investment and non-residential investment could Granger cause GDP, and verse vice.
As a result, residential investment Granger causes GDP, while non-residential investment does
not Granger cause GDP. On the other hand, GDP can Granger cause non-residential investment,
but could not Granger cause residential investment.

Followed by the idea Green came up with, Coulson and Kim tested the Granger causality of
residential and non-residential investment with GDP (Coulson, N. Edward and Kim, Myeong-Soo,
2000). They believed that proving the Granger causal relationship is not enough, and further
study of determinates of GDP is needed. In order to do that, they divided the GDP into our
components, which are residential investment, non-residential investment, consumption and
government spending, and used a vector auto regression (VAR) model with equations of above
four components. Besides reiterating Green’s causality results, they also tested the response of
national income to the shocks in both residential and non-residential investment. As a result,
they drew the conclusion that the influence of residential investment shocks on GDP is greater
than that of non-residential investment shocks.

In addition to the neoclassical economic growth model, Madsen took consideration of the
demand of supply shocks in prices and quantities, and suggested two alternative tests to verify
the Granger causality between investment and economic growth (Madsen, 2002). The author
used the data of 18 OECD countries from 1950 to 1999, and divided investment into two
sectors, which are equipment and structures. In conclusion, investment in the former sector is
mainly driven by supply, while investment in the latter sector is mainly driven by demand.

Wen also investigate the relationship between economic growth and residential related capital
formation from the aspects of both equipment and structures as well (Wen, 2001). Wen
9

extended the problem in both capital sector and business sector. He introduced a block
exogeneity test and explained the mechanism of related Granger causality through different
sectors. It turns out that the residential sector in capital formation actually Granger causes
economic growth, and then further Granger cause capital formation in the business sector.

Brito and Pereira regarded housing as a consumption and an investment good for households
based on an assumption that housing and other assets cannot be perfectly substituted by each
other (Paulo, M. B. Brito, Pereira, M. Alfredo, 2002). The authors established an endogenous
growth model according to the housing price and other types of capital, including physical
capital and human capital, and discussed the balanced growth path in the long run.

Hong discussed the dynamic relationship between real estate investment and economic growth
by employing GMM (Hong, 2014). In his paper, Hong applied an empirical analysis based on the
panel data of 284 Chines prefecture cities from 1994 to 2010. He divided all these cities into
three regions – East, Middle and West - based on their locations. The influences of real estate
investment on economic growth show different tracks in three regions. Generally speaking, real
estate investment has positive influence on economic growth in a short run, while its influence
in a long run is weaker and negative. He also suggested that control of the size and growth rate
of the real estate investment should be taken consideration when making relative policies and
regulations.

Additionally, other literatures also did research in this field from other perspectives. Leung
discussed the relationship between economic growth and increasing housing price via an
endogenous growth model (Leung, 2003). He mentioned that the increasing rate of housing
price might be underestimated due to the inefficient variables selected. Zhang (Zhagn, 2014)
employed panel data into a simultaneous equations model and conclude that the government
should take measures to avoid the aggravation of the gap between housing price and income.

Some Chinese professionals also have dramatic debates on this topic. Zhao and Chen (Zhao,
Chun-ming; Chen, Hao, 2011) used residential sales price data of 1991-2009 to study the impact
of real estate price movements to import based on GMM analysis. Wang (Wang, 2009)
discussed the impacts of urban openness on real estate prices by examining the effectiveness of
Balassa-Samuelson hypothesis. As a result, real estate prices are not only related to property
market and urban factors, but also related to the openness of regional economy. Other
researchers focus on whether the policy could have impact on real estate industry. Li (Li, 2013)
took consideration of land supply, housing price and household consumption and drew a
conclusion of a positive wealth effect of the real estate market.

10

Chapter 3 Theoretical Model and Methodology

In this chapter, the author will talk about the models and methodology in two sections. In the
first section, the author will discuss an empirical model, which will be used in Chapter 4 for
empirical analysis. The author built the model with references of previous papers and
researches on China’s real estate development.

In the second section, the author will discuss the methodology of Granger causality. Many
professionals discussed the relationship between real estate investment and regional economic
development by testing the Granger causality between them. Even though Granger causality is
not the real causality, it could still be used to predict the trend of correlation. In Chapter 4, the
author will test the Granger causality between real estate investment and economic
development based on national data.

3.1 Empirical Model

As discussed earlier, the focus of this paper is to analyze the relationship between real estate
investment and economic development. So in this model, the author used real estate
investment as the dependent variables.

𝑅𝐸𝐼 𝑡 = 𝛼 0
+ 𝛼 1
(𝑑 _𝐺𝐷𝑃 )
𝑡 + 𝛼 2
𝐼𝑛𝑐 𝑡 + 𝛼 3
𝑆 𝑡 + 𝛼 4
𝐿𝑃
𝑡 + 𝛼 5
𝑈𝑛𝑑𝑣𝑙𝑝 𝑡 + 𝜇 𝑡

Specific descriptions of variables are explained in the following table.

Table 3.1 Descriptions of Variables
Variables Description
REI The amount of real estate investment per capita in each province.
d_GDP GDP pre capita is generally treated as an effective indicator of regional
economic growth. d_GDP is the first difference format of GDP per capita.
Inc In this model, income refers to disposable income, which could better
represent individual consumption ability.
S Saving is the balance in people’s banking accounts in the end of last year.
LP In the model, the author used the average of purchase price of a square
meter of land to represent the land vale.
Undvlp Undeveloped land (square meters).

11

Since real estate investment is a contributor to Gross Domestic Products (GDP), there is a
problem of endogenous variable in the model. In order to eliminate the endogenous problem
between real estate investment and GDP, the author took the first difference format of GDP as
one of the independent variables.

In addition, the author introduced disposable income and savings from last year as indicators of
consumption abilities. Other than that, the author also introduced land value and areas of
undeveloped land as indicators of real estate market.

In order to reduce the skew of variables, the author took the natural logarithms format of all
variables and got the formatted equation as followed,

𝑙𝑜𝑔𝑅𝐸𝐼 𝑡 = 𝛼 0
+ 𝛼 1
𝑙𝑜𝑔 (𝑑 _𝐺𝐷𝑃 )
𝑡 + 𝛼 2
𝑙𝑜𝑔𝐼𝑛𝑐 𝑡 + 𝛼 3
𝑙𝑜𝑔𝑆 𝑡 + 𝛼 4
𝑙𝑜𝑔𝐿𝑃 𝑡 + 𝛼 5
𝑙𝑜𝑔 𝑈𝑛𝑑𝑣𝑙𝑝 𝑡 + 𝜇 𝑡

Before importing data in the model, the author came up with some hypothesis.

Hypothesis 1: Regional GDP should have positive influence on real estate development.

Hypothesis 2: Provinces or provincial cities with higher consumption abilities should have
more real estate development.

Hypothesis 3: Provinces or provincial cities with more potentials should have more real estate
development.

The third hypothesis is a little vogue than the other two because the standard to value the
development potentials is difficult to define. In this paper, the author will only discuss this topic
in a relatively practical way on the basis of current issues.

In next chapter, the author will import the panel data in 31 provinces and provincial cities
during the year 2000 to 2013 into the model. The author will then analyze the relationship
between the real estate investment and economic growth and discuss these hypotheses based
on empirical analysis.

3.2 Granger Causality

There is no doubt that many professionals have applied Granger causality test in their
researches. Since Green examined the granger causality based on American data, it is necessary
12

to re-examine whether the granger causality is still … in China. On the other hand, with the
rapid development and increasingly hot debate of the real estate market in China in the past
decade, it is necessary to reiterate the granger causality between real estate investment and
economic growth. The author will explain the methodology of Granger causality test in this
chapter and will employ an empirical analysis in the next chapter based on nationwide
quarterly data from the first quarter in 2005 to the last quarter in 2014 in China in order to see
it there is any changes in the result of Granger causality.

In brief, the concept of Granger causality is that if lagged values of a stationary variable X1 can
improve the ability to predict another stationary variable X2 after controlling for lagged values
of X2, X1 is said to “Granger Cause” X2. Even though the Granger causality is not the real
causality, it is still widely used to testify the relationship between two variables by many
professionals.

The equation of Granger causality test used in this paper is simple as followed (Robert F. Eagle;
C.W.J Granger, 1987).
𝑌 𝑡 = ∑ 𝛼 𝑖 𝑋 𝑡 −𝑖 +
𝑚 𝑖 −1
∑ 𝛽 𝑗 𝑌 𝑡 −𝑖 𝑚 𝑗 −1
+ 𝜇 1𝑡
And
𝑋 𝑡 = ∑ 𝛾 𝑖 𝑋 𝑡 −𝑖 +
𝑛 𝑖 −1
∑ 𝛿 𝑗 𝑌 𝑡 −𝑗 𝑛 𝑗 −1
+ 𝜇 2𝑡

The hypotheses for above equations are:
𝐻 0
: 𝛼 𝑚 = 0, 𝑚 = 1, … , 𝑗
And
𝐻 0
: 𝛿 𝑛 = 0, 𝑛 = 1, … , 𝑗

Before employing Granger causality test, we need to run the unit root test first. The purpose of
the unit root test is to examine whether the variables are stationary. Only if the two variables
are stable, can we further employ the Granger causality test. It is the prerequisite of the
Granger causality test. If we just apply the Granger causality test without testing whether the
variables are stable, the regression results would be spurious. The widely used way to examine
the unit root is Fisher-ADF test.

Only if the variables are stable (in other words, there is no unit root for all variables), can the
author further run Granger causality test. The lag lengths chosen in above equations have
significant impact on rejecting or accepting the null hypothesis mentioned above. In this paper,
the author will follow the criteria of SC and AIC as the standard to decide the lag lengths.

13

Chapter 4 Empirical Analysis and Results

4.1 Data Resources

As mentioned in last chapter, the analysis has two main sections – empirical analysis and
Granger causality test. In the first analysis, the author will use panel data; while in the second
analysis, the author will use time series data instead. The author will explain the reasons later.
All data could be retrieved from National Bureau of Statistics of China.

Since urban development shows regional differences, operating an empirical analysis based on
time series has some limitations and weakness. On the contrary, panel data is an effective and
efficient solution to the above problem. Thus, in the empirical analysis of the relationship
between real estate investment and economic development, the author will apply panel data in
31 provinces and provincial cities in China from 2000 to 2013 into the economic model
mentioned in previous chapter.

On the other hand, the Granger causality test is generally used with time series data. Moreover,
the author would like to discuss the “causality” just between real estate investment and
economic development in a macro level in particular. So in this section, the author will retrieve
nationwide quarterly data ranging from the first quarter in 2005 to the last quarter in 2014 in
China and run Granger causality test accordingly.

4.2 Empirical Analysis

4.2.1 Cluster Analysis

Since each region has different path of economic development and different characteristics of
real estate market, it is necessary to apply a cluster analysis first. It is well known that real
estate is about “location, location and location”. It indicates that location is an inevitable
element when we talk about real estate market. When we search for comparable properties in
real world, the first step is to define the radius. Based on this theory, the author will divide the
31 provincial cities in China into 3 parts according to their locations. According to National
Bureau of Statistics of China, the total 31 provinces can be divided into East district, Middle
district and West district. Specific description of each district is as followed.

14

Table 4.1 Descriptions of East, Middle, and West Districts
District Province
East Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian,
Shandong, Guangdong, Hainan
Middle Shanxi, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei, Hunan
West Neimenggu, Guangxi, Chongqing, Sichuan, Guizhou, Yunan, Xizang, Shan’xi,
Gansu, Qinghai, Ningxia, Xinjiang
After grouping the 31 provinces and provincial cities in China, the author will import the panel
data into the empirical model to test the relationship between real estate investment and
economic growth in East, Middle and West districts respectively.

4.2.2 Regression Results in Nationwide

After clustering the panel data based on locations, the author took the log of each variable and
imported the panel data in each district into previous empirical model. The regression results
are as followed. The author will analyze the results and come up with reasonable explanation
with nationwide panel data first and then discussed the results in East district, Middle district
and West district respectively in more details.

Table 4.2 Regression Result in Nationwide
REI Coef. Std. Err. t P>|t| [95% Conf. Interval]
d_GDP 0.5791655 0.2068509 2.80 0.005 0.1721246 0.9862063
Inc 0.6313696 0.1795258 3.52 0.001 0.2780991 0.9846400
S 0.7908891 0.1510422 5.24 0.000 0.4936685 1.0881100
LP 0.0849524 0.0332817 2.55 0.011 0.0194607 0.1504442
Undvlp 0.1392993 0.0176651 7.89 0.000 0.1045378 0.1740607
_cons -3.906396 0.1800849 -21.69 0.000 -4.260767 -3.552026
Sample Size: 340
R
2
= 0.950074743
Adjusted R
2
= 0.94432677

As it is showed in the result based on nationwide panel data, the R
2
is 0.950074743 and the
adjusted R
2
is 0.94432677, which means that there is a linear correlation in the regression
model. The empirical equation can be further rewritten as

𝑅𝐸𝐼 𝑡 = −3.906396 + 0.5791655 ∗ (𝑑 _𝐺𝐷𝑃 )
𝑡 + 0.6313696 ∗ 𝐼𝑛𝑐 𝑡 + 0.7908891 ∗ 𝑆 𝑡 + 0.0849524 ∗ 𝐿𝑃
𝑡 + 0.1392993 ∗ 𝑈𝑛𝑑𝑣𝑙𝑝 𝑡 + 𝜇 𝑡
15

The p values of all independent variables show that their relationships with real estate
investment are significant to 99% level. In other words, the model is workable to analyze the
real estate investment market and the author will discuss the relationship between real estate
investment and each variable by analyzing the coefficients. To be more specific, the coefficient
of GDP growth is 0.5791655, which indicators its positive influence on real estate investment. It
testifies the first hypothesis mentioned above.

Speaking of people’s consumption behavior, disposable income and savings in the end of last
year could be useful indicators. As it is showed above, the coefficients of these two variables
are 0.6313696 and 0.7908891 respectively. It represents that the more money people earn or
save, the more they would put into the real estate investment market. It testifies the second
hypothesis mentioned above.

With regards to the real estate market, land value and areas of undeveloped land are regarded
as effective and efficient parameters. As the result showed, the coefficient of land value is
0.0849524, which indicates that the land value has positive influence on real estate investment.
And yet, the coefficient of land value is not very high, which means that even though land value
has positive influence on real estate investment, it is not the most fundamental determinant.
Similar to that of land value, the coefficient of the areas of undeveloped land is also only
0.1392993, which implies the similar role of this parameter. Thus, the third hypothesis the
author came up with is not wrong, but the influence of real estate market on investment is not
that observable.

Since each province and provincial city has a different picture of economy and a different path
of development, it is necessary to analyze what the relationship between real estate and
regional economy in district level. The author will run the regression of the empirical model
based on the panel data in East district, Middle district and West district respectively in the
following section.

4.2.3 Regression Results in the East District

As it is showed in the Table 4.3, the R
2
is 0.944820757 and the adjusted R
2
is 0.936938008,
which means that there is a linear correlation in the regression model.

16

Table 4.3 Regression Result in the East District
REI Coef. Std. Err. t P>|t| [95% Conf. Interval]
d_GDP 0.6899385 0.4124301 1.67 0.097 -0.1278342 1.5077110
Inc -0.0661174 0.3439235 -0.19 0.848 -0.7480541 0.6158193
S 1.3413040 0.3101535 4.32 0.000 0.7263265 1.9562810
LP 0.0638758 0.0515987 1.24 0.219 -0.0384348 0.1661865
Undvlp 0.2070521 0.0282119 7.34 0.000 0.1511131 0.2629911
_cons -3.7387880 0.2389521 -15.65 0.000 -4.212586 -3.2649900
Sample Size: 121
R
2
= 0.944820757
Adjusted R
2
= 0.936938008

The empirical equation can be further rewritten as
𝑅𝐸𝐼 𝑡 = −3.738788 + 0.6899385 ∗ (𝑑 _𝐺𝐷𝑃 )
𝑡 – 0.0661174 ∗ 𝐼𝑛𝑐 𝑡 + 1.341304 ∗ 𝑆 𝑡 + 0.0638758 ∗ 𝐿𝑃
𝑡 + 0.2070521 ∗ 𝑈𝑛𝑑𝑣𝑙𝑝 𝑡 + 𝜇 𝑡

Taking closer look of the above equation, the author came up with some findings in the east
district. Since the coefficient of GDP growth is positive, it is clear that GDP growth has positive
influence on real estate investment. It makes sense since regional economy should have
positive impacts on investment, and real estate is one of many kinds of investments. Thus, the
hypothesis 1 is proved to be correct in the east district. And yet, the p value of GDP growth is
only 0.097, which represents that GDP growth only have significant impacts on real estate
investment to 10% degree.

The p value of the undeveloped land variable is 0, which underline the strong relationship
between undeveloped land and real estate investment. The coefficient of the undeveloped land
variables is 0.2070521, which is positive to the dependent variable. It means that the more
undeveloped land, the more money will be invested into the real estate market in the east
district in China. This proves the validation of hypothesis 3. On the contrary, the influence of
land value on real estate is not obvious since the p value of land value is 0.219. From this
perspective, even the coefficient of land value to real estate investment (0.0638758) is positive,
it cannot be treated as a strong determinant for decision-making.

However, not all of the hypotheses the author made in Chapter 3 are correct. For instance, it is
hard to tell whether the hypothesis 2 is true or false in the east district. In theory, provinces or
provincial cities with higher consumption abilities should have more real estate investment. In
the empirical model, it can be translated into positive coefficients of the independent variables,
income and savings in the end of last year. As a matter of fact, the result shows that the p value
17

of savings in the end of last year is 0, while the p value of disposable income is 0.848. The result
shows an interesting fact. On one hand, how much people would invest in real estate market
does not depend on their disposable income. On the other hand, the coefficient of the savings
in the end of last year is positive. Since the coefficient of savings in the end of last year is as
high as 1.341304, a possible explanation of the fact is that people in the east district use their
savings to invest in the real estate development instead of their disposable income.

4.2.4 Regression Results in the Middle District

Table 4.4 Regression Result in the Middle District
REI Coef. Std. Err. t P>|t| [95% Conf. Interval]
d_GDP 0.6174963 0.2806007 2.20 0.031 0.0585111 1.1764810
Inc 1.3620520 0.3756952 3.63 0.001 0.6136284 2.1104750
S 0.3374857 0.314169 1.07 0.286 -0.2883711 0.9633425
LP 0.0827332 0.0814534 1.02 0.313 -0.0795303 0.2449968
Undvlp 0.0670503 0.0423081 1.58 0.117 -0.0172318 0.1513324
_cons -4. 632801 0. 2986298 -15.51 0. 000 -5. 2277020 -4. 0379000
Sample Size: 88
R
2
= 0.978414135
Adjusted R
2
= 0.974960397

As it is showed in the result, the R
2
is 0.978414135 and the adjusted R
2
is 0.974960397, which
means that there is a linear correlation in the regression model. The empirical equation can be
further rewritten as

𝑅𝐸𝐼 𝑡 = −4.632801 + 0.6174963 ∗ (𝑑 _𝐺𝐷𝑃 )
𝑡 + 0.1.362052 ∗ 𝐼𝑛𝑐 𝑡 + 0.3374857 ∗ 𝑆 𝑡 + 0.0827332 ∗ 𝐿𝑃
𝑡 + 0.0670503 ∗ 𝑈𝑛𝑑𝑣𝑙𝑝 𝑡 + 𝜇 𝑡

The p value of GDP growth is 0.031, which indicates that the significance of GDP growth on real
estate investment is to the 95% level. Furthermore, the coefficient of GDP growth is 0.6174963,
which shows its positive impacts on real estate investment. Although the coefficient of GDP
growth in the middle district is a little lower than that in the east district, the result still shows
strong evidence to prove the first hypothesis.

The p value of savings in the end of last year is 0.286, which indicates weak relationship
between savings in the end of last year and real estate investment. In contrast, disposable
income has a strong relationship with real estate investment with the p value of 0.001. What is
more, the coefficient of disposable income is 1.362052. In other words, people with higher
18

income would more like to invest in the real estate market. Compare with people in the east
district, people in the middle district show a different consumption habit and preference of real
estate investment. However, similar to that in the east district, it is also hard to provide strong
proof of the accuracy of the second hypothesis.

The p values of land value and undeveloped land are 0.313 and 0.117 respectively. It indicates
that the influence of these two variables on real estate investment is not significant to even 90%
level. The results could not prove the truth of the third hypothesis in the middle district. Thus,
in the middle district, land value and areas of undeveloped land are not sensitive parameters to
predict the trend of real estate investment.

4.2.5 Regression Results in the West District

Table 4.5 Regression Result in the West District
REI Coef. Std. Err. t P>|t| [95% Conf. Interval]
d_GDP 0.3853338 0.3465458 1.11 0.269 -0.3011708 1.0718380
Inc 0.9809703 0.3701259 2.65 0.009 0.2477537 1.7141870
S 0.5828580 0.2798729 2.08 0.040 0.0284318 1.1372840
LP 0.0434620 0.0549159 0.79 0.430 -0.0653260 0.1522500
Undvlp 0.0928807 0.0278807 3.33 0.001 0.0376492 0.1481122
_cons -4.0684140 0.4495186 -9.05 0.000 -4.9589070 -3.1779210
Sample Size: 131
R
2
= 0.943776599
Adjusted R
2
= 0.935885596

As it is showed in the result, the R
2
is 0.943776599 the adjusted R
2
is 0.935885596, which
means that there is a linear correlation in the regression model. The empirical equation can be
further rewritten as

𝑅𝐸𝐼 𝑡 = −4.068414 + 0.3853338 ∗ (𝑑 _𝐺𝐷𝑃 )
𝑡 + 0.1.9809703 ∗ 𝐼𝑛𝑐 𝑡 + 0.582858 ∗ 𝑆 𝑡 + 0.043462 ∗ 𝐿𝑃
𝑡 + 0.0928807 ∗ 𝑈𝑛𝑑𝑣𝑙𝑝 𝑡 + 𝜇 𝑡

Different from the results in the east district and middle district, GDP growth in the west district
did not show strong relationship with real estate investment since the p value of GDP growth is
0.269. In other words, the result did not show a persuasive clue of the relationship between
regional economy and real estate development. Thus the first hypothesis cannot be proved true
in the west district.

19

Speaking of people’s consumption abilities, which refer to disposable income and savings in the
end of last year in the model. On one hand, the p value of disposable income is 0.009, which
represents that the result is significant at the 99% level. The coefficient of disposable income is
0.9809703, which refers to its solid relation with real estate investment. On the other hand, the
p value of savings in the end of last year is 0.040, which represents that its positive impact on
real estate investment is significant at the 95% level. Therefore, the second hypothesis is
proved to be accurate in the west district.

However, the two variable used to represent the real estate market revealed different pictures
in the regression results. In particular, the p value of land value is 0.430, which is proved not to
be a convincing parameter, while the p value of areas of undeveloped land is as low as 0.001
(significant at 99% level). As far as the author knew, the land value in the west district is not as
high as that in the east district and middle district, so that it is reasonable that investors did not
make much consideration of land value when they decide to invest in the real estate market.
With regards to the areas of undeveloped land, its coefficient is 0.0928807. It means that the
more undeveloped land, the more likely people would like to invest.

In summary, even the relations between real estate investment and each variable are relatively
clear to the national degree, the three districts have different characteristics of the relationship
between regional economy and real estate investment. Generally speaking, the three
hypotheses mentioned in previous chapter have been proven to be true in most cases.

4.3 Granger Causality Results

As it is known that Granger causality is not real causality, yet Granger causality analysis is widely
used in business forecasting and policy making model anyway. The purpose of running the
Granger causality test in this paper is to further explore the relationship between economic
growth and real estate investment in particular. The author applied quarterly time series data
from the first quarter in 2005 to the last quarter in 2014 (40 periods) in nationwide. All data can
also be retrieved from China Bureau of Statistics.

4.3.1 Unit Root Test

The first step to test the Granger causality is to run the unit root test. The purpose of running
the unit root test is to test the stationary feature for the panel data. In this paper, the author
will use Fisher-ADF test, which is widely used to test whether there are unit roots.

20

At the 1% significance level, the Fisher-ADF test represents that nationwide GDP and real estate
investment have a unit root, while the first differences of GDP and real estate investment
become stationary at the 1% and 5% significance level respectively. The Fisher-ADF test result
strongly suggests that GDP and real estate investment should be taken as I (1) series, and the
author will continue to implement Granger causality test by using their first differences. Table
4.6 and Table 4.7 present the unit root tests for the whole country and in three districts
respectively.
Table 4.6 Unit Root Test
Variables Whole Country
lnGDP 0.3783
∆lnGDP 0.0073***
lnREI 0.0130
∆lnREI 0.0004**
∆ is the first order difference
***, rejects the null hypothesis at the 1% level
**, rejects the null hypothesis at the 5% level
*, rejects the null hypothesis at the 10% level

Table 4.7 Unit Root Test in the East, Middle, and West Districts
Variables East District Middle District West District
lnGDP 0.2277 0.3173 0.1862
∆lnGDP 0.0037*** 0.0041*** 0.0241**
lnREI 0.3291 0.2366 0.1392
∆lnREI 0.0047*** 0.2551* 0.1466*
∆ is the first order difference
***, rejects the null hypothesis at the 1% level
**, rejects the null hypothesis at the 5% level
*, rejects the null hypothesis at the 10% level

In a micro level, the Fisher-ADF test represents that GDP and real estate investment in three
districts have a unit root at the 1% significance level. In more details, the first differences of
GDP and real estate investment became stationary at the 1% significance level respectively in
the east district. The first differences of GDP and real estate investment became stationary at
the 1% and 10% significance level respectively in the middle district. The first differences of GDP
and real estate investment became stationary at the 5% and 10% significance level respectively
in the west district. Thus, the Fisher-ADF test results strongly suggest that GDP and real estate
investment should be taken as I (1) series, and the author will continue to implement panel
21

Granger test by using their first differences. The following table presents the unit root tests for
the three districts.

4.3.2 Granger Causality Results

After running the unit root test, the next step is to implement the Granger causality test for the
variables.
Table 4.8 Granger Causality Test Results in Nationwide
Null Hypothesis Obs Lags F-Statistics P Value
GDP causes real estate investment 38 1 0.00019 0.9892
Real estate investment causes GDP 0.00088 0.9765
GDP causes real estate investment 37 2 13.0806 7.00E-05***
Real estate investment causes GDP 12.5904 9.00E-05***
GDP causes real estate investment 36 3 12.8268 2.00E-05***
Real estate investment causes GDP 23.4293 7.00E-08***
GDP causes real estate investment 35 4 5.70457 0.0020***
Real estate investment causes GDP 5.34133 0.0028***
GDP causes real estate investment 34 5 0.51340 0.7633
Real estate investment causes GDP 0.81265 0.5528
***, rejects the null hypothesis at the 1% level
**, rejects the null hypothesis at the 5% level
*, rejects the null hypothesis at the 10% level

According the results, the null hypothesis does not reject at any level with the lag lengths of
one, while it rejects at the 1% level with the lag lengths of two to four. The results showed a
clear and strong two-way Granger causality relationship between GDP and real estate
investment in nationwide within four years. However, when the author picked up five years as
the lag length, the null hypothesis did not reject either. According to the criteria of SC and AIC
to decide the lagged period of the variables, the author picked up the lagged period as 5, which
means that in a long run, the Granger causality between regional economy and real estate
investment in nationwide becomes weaker.

As the results portrayed in Table 4.9, the null hypothesis rejects at the 1% level in both sides
with the lag lengths smaller than five on both sides. There is no doubt that there is a strong
bidirectional Granger causality between GDP and real estate investment. In other words, GDP
Granger causes real estate investment and real estate investment Granger causes GDP.
However, according to the criteria of SC and AIC to decide the lagged period of the variables,
22

the appropriate lagged period should be five. In that circumstance, the null hypothesis does not
reject at any level. As it is known, real estate industry is facing an extremely fast-speed growth
in recent decades. And yet, after the peak of real estate development growth, real estate
industry could not further cause the increasing regional growth.

Table 4.9 Granger Causality Test Results in the East District
Null Hypothesis Obs Lags F-Statistics P Value
GDP causes real estate investment 38 1 35.6379 8.00E-07***
Real estate investment causes GDP 35.5956 9.00E-07***
GDP causes real estate investment 37 2 22.2944 9.00E-07***
Real estate investment causes GDP 23.2290 6.00E-07***
GDP causes real estate investment 36 3 8.46759 3.00E-04***
Real estate investment causes GDP 15.6979 3.00E-06***
GDP causes real estate investment 35 4 5.14971 0.0034***
Real estate investment causes GDP 5.26944 0.0030***
GDP causes real estate investment 34 5 1.47757 0.2356
Real estate investment causes GDP 1.56464 0.2096
***, rejects the null hypothesis at the 1% level
**, rejects the null hypothesis at the 5% level
*, rejects the null hypothesis at the 10% level

Table 4.10 Granger Causality Test Results in the Middle District
Null Hypothesis Obs Lags F-Statistics P Value
GDP causes real estate investment 38 1 21.4039 5.00E-05***
Real estate investment causes GDP 21.8222 4.00E-05***
GDP causes real estate investment 37 2 12.9995 7.00E-05***
Real estate investment causes GDP 13.9113 4.00E-05***
GDP causes real estate investment 36 3 6.19128 2.20E-03***
Real estate investment causes GDP 15.0880 4.00E-06***
GDP causes real estate investment 35 4 4.79472 0.0050***
Real estate investment causes GDP 4.94889 0.0042***
GDP causes real estate investment 34 5 1.87079 0.1388
Real estate investment causes GDP 0.71618 0.6177
***, rejects the null hypothesis at the 1% level
**, rejects the null hypothesis at the 5% level
*, rejects the null hypothesis at the 10% level

23

Compared with the results in the east district, the Granger causality results in the middle
district is similar with the lag lengths less than five as it is showed in Table 4.10. Chinese
government proposed a policy of promoting the rise of the central area by developing new city
plans and real estate development. That would be a reasonable explanation of the relationship
between regional economy and real estate development in previous analysis.

Table 4.11 Granger Causality Test Results in the West District
Null Hypothesis Obs Lags F-Statistics P Value
GDP causes real estate investment 38 1 0.41103 5.26E-01***
Real estate investment causes GDP 0.45324 5.05E-01***
GDP causes real estate investment 37 2 14.3391 4.00E-05***
Real estate investment causes GDP 13.7467 5.00E-05***
GDP causes real estate investment 36 3 1.68909 1.91E-01***
Real estate investment causes GDP 6.91511 1.20E-03***
GDP causes real estate investment 35 4 3.63529 0.0176**
Real estate investment causes GDP 4.58941 0.0062***
GDP causes real estate investment 34 5 2.87942 0.0367**
Real estate investment causes GDP 2.60956 0.0520*
***, rejects the null hypothesis at the 1% level
**, rejects the null hypothesis at the 5% level
*, rejects the null hypothesis at the 10% level

According to the criteria of SC and AIC, the proper lag lengths should be five. As the results
portrayed above, the null hypothesis that GDP causes real estate investment rejects at the 5%
level, while the null hypothesis that real estate investment causes GDP rejects at the 10% level.
Looking backwards, there is a strong two-way Granger causality relationship between regional
economy and real estate investment on both sides with the lag lengths of one to four. In other
words, in the west district, the Granger causality between these two variables is stable in both
short term and long term. As a matter of fact, real estate industry in the west district has more
potentials as well as regional economy.

To conclude, the Granger causality between GDP and real estate investment is proven to be
strong in a short run in both nationwide and three districts. In a long run, only the west district
has an obvious Granger causality relationship between regional economy and real estate
investment in a long run. But overall, the Granger causality has been clarified.

24

Chapter 5 Conclusions

Chinese economy is experiencing tremendously rapid development and real estate industry is
said to be able to promote the whole economy in both macro and micro degree. The important
of real estate investment to regional economy is especially debated in China. Based on previous
analysis, real estate industry is definitely regard as an indispensable parameter and
determinant in the national economic system.

Based on the regression results of nationwide panel data, GDP growth, disposable income,
savings in the end of last year have apparently positive influences on real estate investment. On
the other hand, land value and areas of undeveloped land also have positive influences,
however the influences are slight on real estate investment.

From a micro perspective, GDP growth only has a positive impacts on real estate investment in
the east and middle district since the regional economy in the west district is not as prosperous
as that in the other two districts. With regards to people’s consumption abilities, the three
districts behaviors in three ways. In the east district, people prefer to invest real estate with
their savings; in the middle district, they would like to put their liquid asset into real estate
investment; and in the west district, disposable income and savings are both the resource of
real estate investment. Speaking of development potentials, the west district displays a more
promising picture of real estate development.

The author hypothesized that economic growth and real estate investment have bidirectional
impacts on each other. This paper has re-examined the relationship between real estate
investment and regional economy. Consequently, GDP per capita Granger causes real estate
investment and real estate investment Granger causes GDP per capita as well in nationwide and
in three districts individually as well. However, in a relatively long run, the Granger causality of
regional economy and real estate investment becomes weaker rather than that in a short run.

There is no doubt that the real estate market in China is making a fundamental contribution to
Gross Domestic Products. However, the outcome of current real estate development is more
like a double-edged sword to regional economy. Following a normal business cycle, the
development of an industry will follow a normal distribution in a long term. On the basis of
current situations, China’s real estate market is close to the peak, which means the whole
industry will face a slowdown in growth or decreasing in the next few years.

Many policies have been proposed to control the “unlimited” growth of real estate investment.
For example, the Chinese government came up with the policy of increasing the down payment
25

for purchasing a second house as a family. Generally speaking, Chinese authority aims to make
a higher threshold of real estate investment. It is a measure to cool down the extremely real
estate investment.

Another interesting fact that happens recently is that many Chinese investors started to put
their money into the real estate market overseas, such as US and Australia. Some Chinese big
companies, such as China Vanke Co. and Greenland Holding Group Co., have started several
projects in San Francisco, Los Angeles, New York and etc. (KPMG, 2014). It is too early to tell
whether it is the right decision to make considering current global economy.

There still are many topics related to Chinese real estate investment need to discuss. In this
paper, the author employed latest data into a simplified research model. In further research,
the author could take account of some political and environmental parameters into the model
to predict the development trend of real estate investment and how the real estate market and
regional economy influence on each other.

26

References
Coulson, N. Edward and Kim, Myeong-Soo. (2000). Residential Investment, Non-residential Investment
and GDP. Real Estate Economics, 233-247.
Green, K. R. (1997). Follow the Leader: How Changes in Residential and Non-residential Investment
Predict Changes in GDP. Real Estate Economics, 253-270.
Hong, L. (2014). The Dynamic Relationship between Real Estate Investment and Economic Growth:
Evidence from Prefecture City Panel Data in China. IERI Procedia, 2-7.
KPMG. (2014). China Inbound Investing in U.S. Real Estate. KPMG.
Leung, K. Y. (2003). Economic Growth and Increasing House Prices. Pacific Economic Review, 183-190.
Li, Y. (2013). Land Supply, Housing Price and Household Consumption: An Estimation Based on the Panel
Data Simultaneous Equations Model. Journal of Nanjing Agricultural University (Social Sciences
Edition), 54-63.
Madsen, B. J. (2002). The Causality between Investment and Economic Growth. Economics Letters, 157-
163.
Mills, S. E. (1987). Has the United States Overinvested in Housing? Journal of the American Real Estate
and Urban Economics Association, 212-217.
Paulo, M. B. Brito, Pereira, M. Alfredo. (2002). Housing and Endogenous Long-term Growth. Journal of
Urban Economics, 246-271.
Robert F. Eagle; C.W.J Granger. (1987). Co-Integration and Error Correction: Representation, Estimation,
and Testing. Econometrica, 251-276.
Wang, S. (2009). Urban Openness and Real Estate Prices: Empirical Evidence from Thirty-Five Large Scale
Chinese Cities. Nankai Economic Studies, 91-102.
Wen, Y. (2001). Residential Investment and Economic Growth. Annals of Economics and Finance, 437-
444.
Zhagn, C. (2014). The Endogeneity and Interactive Relationship between Housing Price and Income
Distribution. Statistical Research, 63-69.
Zhao, Chun-ming; Chen, Hao. (2011). Analysis on Relation between Real Estate Price Movements and
Import in China: Based on GMM and Principal Component Factor. Journal of International Trade,
28-34.

Abstract (if available)

Abstract China has gone through dramatic development in the last decade. Since 2010, China Gross Domestic Products (GDP) surpassed Japan as the world’s second-largest economy. That fact showed a signal of the economy recovery from the recession, and leads a number of literature to work on the triggers to the Chinese economy. There is no doubt that real estate investment, as a component of GDP, plays a fundamental role in economic growth. ❧ Many literature have discussed the topic of real estate investment and economic growth from different perspectives. In this paper, the author will go over some researches and studies in the related field. Some professionals started their research from testing the Granger causality between investment and economic growth. Other professionals focused on figuring out the mechanism between these two sectors and how these two sectors influence on each other. In this paper, the author will also discuss the topic from above two ways. ❧ In the first section, the author will develop an empirical analysis based on the provincial data in China from 2000 to 2013. Firstly, the author will employ a cluster analysis by dividing the 31 provinces into East, Middle and West districts based on their locations. In the second section, the author will discuss how to analyze the relationship and Granger causality between real estate investment and economic growth theoretically. The author will reexamine the Granger causality between real estate investment and economic growth in each district, which will be regarded as the basis of further analysis. ❧ On the basis of theoretical and empirical analysis in previous two sections, the author will investigate the impact what real estate investment has on urban development and vice versa. In conclusion, real estate investment can Granger cause the economic growth, while economic growth can also Granger cause the real estate investment. Furthermore, during the study period, real estate investment has observable influence on economic growth in China. Last but not least, the author would like to discuss what factors could be potentially involved in the relationship between real estate development and economic growth based on qualitative analysis, and put forward some recommendation for policy makers.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Growth, trade and structural change in low income industrializing economies

PDF

Three papers on dynamic externalities and regional growth

PDF

Essays on firm investment, innovation and productivity

PDF

An empirical analysis of the quality of primary education across countries and over time

PDF

Essays in panel data analysis

PDF

Essays in political economy and mechanism design

PDF

Pathways of local economic development: tales of cities in the United States and South Korea

PDF

Essays on urban and real estate economics

PDF

Three essays in United States real estate markets

PDF

Regulation and rights reallocation: how private car license plate auction policy in Shanghai affects travelers’ welfare

PDF

Increased access to capital: evaluation of the New Market Tax Credit Program in New York

PDF

The economy-environment trade-off in China's air pollution policies

PDF

Avoiding middle-class planning 2.0: media arts and the future of urban planning

PDF

Population disaggregation for trade area delineation in retail real estate site analysis

PDF

Financial crises and trade policy in developing countries

PDF

Essays in empirical asset pricing

PDF

Housing consumption-based asset pricing and residential mortgage default risk

PDF

Two essays on major macroeconomic shocks in the Japanese economy: demographic shocks and financial shocks

PDF

Essays on the econometrics of program evaluation

PDF

Three essays on supply chain networks and R&D investments

Asset Metadata

Creator Ming, Xing (author)

Core Title An empirical analysis of the relationship between real estate investment and regional economy in China

School College of Letters, Arts and Sciences

Degree Master of Planning / Master of Arts

Degree Program Planning / Economics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag economic growth,Granger causality,OAI-PMH Harvest,panel data,Real estate investment

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor David, Joel (committee chair), Cheng, Harrison (committee member), Green, Richard Kent (committee member)

Creator Email daisy.418.ming@gmail.com,xingming@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-588262

Unique identifier UC11301080

Identifier etd-MingXing-3568.pdf (filename),usctheses-c3-588262 (legacy record id)

Legacy Identifier etd-MingXing-3568.pdf

Dmrecord 588262

Document Type Thesis

Format application/pdf (imt)

Rights Ming, Xing

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA