analysis of stock market linkages: chinese, indian and major markets ...

ANALYSIS OF STOCK MARKET LINKAGES:
CHINESE, INDIAN AND MAJOR MARKETS
MARLIANA ABAS
FACULTY OF BUSINESS AND ACCOUNTANCY
UNIVERSITY OF MALAYA
JULY 2009

Analysis of Stock Market Linkages:
Chinese, Indian and Major Markets
Marliana Abas
Bachelor of Business Administration (Hons) Finance
Mara University of Technology
2004
Submitted to the Graduate School of Business
Faculty Business and Accountancy
University of Malaya, in partial fulfillment
Of the requirements for the Degree of
Master of Business Administration
July 2009

ABSTRACT
This study examines the linkages of the two leading emerging markets i.e. Chinese
and Indian market with other developed markets. Using daily data from January 2000
to December 2008, we examine the stock market indices of China, India, United
States, United Kingdom, Japan and Hong Kong. We model the linkages among these
stock markets by using a simple correlation test, Granger causality test and co
integration test using error correction model. We found that Chinese and Indian
markets are correlated with all four developed markets under study. Both markets
have at least had unilateral causality with all four developed markets. The empirical
results suggest that the benefits of any short-term diversification, or speculative
activities, are limited between them.
ii

ACKNOWLEDGEMENTS
All praises to Almighty Allah, the Most Gracious and Most Merciful for giving me
the strength and patience in completing this project paper. I would like to express my
greatest appreciation to my most respected advisor, Dr. Gucharan Singh for his time,
moral support, ideas, suggestions and continued guidance throughout the period of
completing this project paper.
My deepest appreciation is also dedicated to my beloved family members especially
my parents Abas b. Md Yusof and Asiah bt. Mohd Ali, for their support and
understanding throughout the preparation of this project paper. Not forgotten to my
beloved person, Rahimy Bahaudin for his support, help, encouragement and
understanding.
I would also like to thank all my friends especially MBA students and everybody who
had directly or indirectly gives me the support, encouragement, constructive criticisms
and advises in preparing this research. All their kindness and willingness to help me
will be forever thankful and appreciated.
Thank you so much.
iii

TABLE OF CONTENT
ABSTRACT ………………………………………………………………………………... ii
ACKNOWLEDGEMENTS ………………………………………………………………...iii
LIST OF TABLES ………………………………………………………………………….vii
LIST OF FIGURES ………………………………………………………………………..viii
LIST OF ABBREVIATIONS …………………………………………………………….. ix
CHAPTER 1 INTRODUCTION
1.1 BACKGROUND OF STUDY .....................................................................................1
1.2 PROBLEM STATEMENT .........................................................................................3
1.3 OBJECTIVES OF THE STUDY................................................................................4
1.4 SIGNIFICANCE OF THE STUDY............................................................................5
1.5 SCOPE OF THE STUDY............................................................................................5
1.6 LIMITATIONS OF THE STUDY..............................................................................6
1.7 ORGANIZATION OF THE STUDY................................................................ 7
CHAPTER 2 OVERVIEW OF CHINA AND INDIA & LITERATURE
REVIEW
2.0 INTRODUCTION........................................................................................................8
2.1 CHINA’S AND INDIA’S ROLE TO WORLD ECONOMY...................................8
2.2 AN OVERVIEW OF CHINA AND INDIA STOCK MARKETS.........................13
2.2.1 CHINA STOCK MARKET ............................................................................ 13
2.2.2 INDIA STOCK MARKETS............................................................................ 15
2.3 EMPIRICAL EVIDENCE ON INTERNATIONAL STOCK MARKETS
LINKAGES, CO-MOVEMENT AND INTERDEPENDENCIES ........................16
2.4 STUDIES ON CHINESE AND INDIAN MARKETS ............................................23
iv

2.5 SUMMARY ................................................................................................................25
CHAPTER 3 RESEARH METHODOLOGY
3.0 INTRODUCTION......................................................................................................26
3.1 EMPIRICAL WORK ................................................................................................26
3.2 DATA ..........................................................................................................................29
3.3 INSTRUMENTATION AND SCALES ...................................................................31
3.4 METHODOLOGY ....................................................................................................32
3.4.1 THE CORRELATION TEST......................................................................... 32
3.4.2 GRANGER CAUSALITY TEST.................................................................... 34
3.4.3 UNIT ROOT TEST.......................................................................................... 35
3.4.4 COINTEGRATION TEST.............................................................................. 37
3.4.5 ERROR CORRECTION MODEL (ECM).................................................... 40
3.5 SUMMARY................................................................................................... 41
CHAPTER 4 RESEARH RESULTS
4.0 INTRODUCTION......................................................................................................42
4.1 DESCRIPTIVE STATISTICS..................................................................................42
4.2 ANALYSIS OF CORRELATIONS .........................................................................45
4.3 ANALYSIS OF GRANGER CAUSALITY TEST..................................................46
4.4 ANALYSIS OF UNIT ROOT TEST........................................................................50
4.5 ANALYSIS OF EAGLE GRANGER TEST ...........................................................51
4.6 ANALYSIS OF ERROR CORRECTION MODEL...............................................53
4.7 SUMMARY ................................................................................................................55
CHAPTER 5 CONCLUSION & RECOMMENDATIONS
5.1 CONCLUSION ..........................................................................................................57
5.2 RECOMMENDATION.............................................................................................60
v

REFERENCES.......................................................................................................................61
APPENDICES........................................................................................................................65
vi

LIST OF TABLES
Table 2.1: Top 10 Economies in 2007 9
Table 4.1: Descriptive Statistics 43
Table 4.2: Correlation of Stock Price Indices 45
Table 4.3: Results of the Granger Causality Test 47
Table 4.4: Augmented Dickey-Fuller 50
Table 4.5: Unit Root Test applied to the Residuals of
Cointegrating Regressions 52
Table 4.6: Bivariate ECM for Cointegrated Indices 54
vii

LIST OF FIGURES
Figure 2.1: Top 10 Countries by GDP in 2007 9
Figure 2.2: China’s and India’s GDP Growth 10
Figure 2.3: China’s Trade 11
Figure 2.4: India’s Trade 12
Figure 4.1: Movement of the Indices in the Observed Period 44
viii

LIST OF ABBREVIATION
DJIA Dow Jones Industrial Average
FTSE FTSE 100 Index
H0 Null hypothesis
H1 Alternative hypothesis
HSI Hang Seng Index
NIFTY National Stock Exchange S&P CNX Nifty Index
NIKKEI Nikkei 225 Stock Average
SENSEX Bombay Stock Exchange Sensitive Index
SHCOMP Shanghai Stock Exchange Composite Index
SHENZEN Shenzen Stock Exchange Composite Index
U.K. United Kingdom
U.S. United States
ix

1.1 BACKGROUND OF STUDY
CHAPTER 1
INTRODUCTION
International stocks markets have becoming more integrated in recent years. There are
several factors that contributed to the linkages or interdependencies of stock markets.
The progressive removal of restrictions and relaxation of controls on capital
movements, among others has increased the flow of funds across countries. Hence,
national stock exchanges are becoming more integrated and moving towards increasing
relationship with other international stock exchanges.
Risk reduction through the construction of diversifies portfolios is the foundation of
modern portfolio theory. The theory of portfolio selection developed by Harry
Markowitz and James Tobin (1952) states that diversification could eliminate risk if
returns are not correlated. If the correlation between the returns of the equity markets
under consideration increases, the risk exposure of the portfolio (all else being constant)
will start to increase and, at a certain point, international diversification will no longer
beneficial.
On the basis of portfolio selection theory, the increasing relationship among national
stock markets has created more opportunities for global international investment as
investors have begun including assets of foreign countries into their portfolio to reduce
risk and diversify effectively. Thus, knowing the correlations or relationship between
the returns of various national financial markets is important for the process of
allocating investments among these markets to reduce risk.
1

Recognizing the benefits of international diversification, numerous studies in finance
literature have concentrating on measuring the international linkages of national stock
markets across several developed and emerging markets. However, studies focusing on
the two major emerging markets namely China and India alone are rather limited.
The World Bank (1997) argues that the world’s financial markets are rapidly integrating
into a single global marketplace as investors are driven to developing countries in the
search for higher returns. However, history has shown that some once-emerging
economies, such as the United States and Japan have been successful. This has serve as
motivation for investors to allocate their investment in emerging market in order to
enhance portfolio return and reducing risk through diversification. The risk and return
in investing in emerging markets are significantly linked to the ability of these markets
to develop economically.
Two emerging markets, China and India have been called the Asian tigers due to the
unprecedented economic development experienced by both markets in recent years.
During the last decade, China’s economy as measured by GDP 1 has grown at the
average of 10 per cent per annum while India’s at 7 per cent per annum. During this
period, the trade volume, capital flows and mutual economic agreements with other
markets have also increased rapidly. Having large economic size, huge population and
dynamic economic growth, China and India emerge as two major prominent emerging
markets which contribute to the world economy. Huge potential returns from these
markets have attracted many investors to consider China and India in their investment
portfolio to maximize investment return and reduce risk.
1 GDP is calculated as the value of the total final output of all goods and services produced in a single year within a
country’s boundaries.
2

Therefore, this study aim at investigating the relationship and linkage of Chinese and
Indian markets with major developed markets namely United States (U.S), United
Kingdom (U.K), Japan and Hong Kong.
1.2 PROBLEM STATEMENT
In the light of globalization, the national stock markets are moving towards increasing
linkages to other international stock markets. From the perspective of international
investor, increasing stock markets linkage suggest that there is high correlation between
the markets where the separate markets move together resulting in less benefit of
diversification.
The question whether there is a relationship among stock markets is difficult to answer
without doing the research to investigate the problem. The relationships among
international stock markets have received a great deal of attention. Studies by Hillard
(1979), Eun and Shin (1989), Koch and Koch (1991) and Campbell and Hamao (1992),
among others have find the evidence of positive correlations in returns between stock
market around the world.
With the rapid growth of India and China, many investors would certainly consider to
invest in the two markets rather than in the advanced or developed markets. However,
the question of whether both markets are integrated with other stock markets so that
investing in India and China will provide the benefit of diversification is the major
concern raised by investors.
3

Thus, this study aims at providing some insight to the question above by investigating
the linkage of both Indian and Chinese market with other major developed markets. In
this study, we use four major markets namely United States (U.S.), United Kingdom
(U.K.), Japan and Hong Kong.
1.3 OBJECTIVES OF THE STUDY
The motivation behind this study is to investigate whether, despite the growing
importance and contribution of the two major emerging markets i.e., India and China to
the world economy, are their stock markets interdependent with other stock markets?
The aim is to present the linkages or relationship of Indian and Chinese with four other
major developed markets namely United States, United Kingdom, Japan and Hong
Kong using various econometrics techniques.
As such; the objectives of this study are as follows:
1. To examine the relationships of Indian and Chinese markets with four other
major developed markets namely United States, United Kingdom, Japan and
Hong Kong.
2. To analyse whether the two major emerging countries i.e. China and India are
interdependence with each other.
4

1.4 SIGNIFICANCE OF THE STUDY
This study will provide some useful information to investment practitioners, policy
makers, academicians and other researches where:
1. It provides knowledge on the relationship of Chinese and Indian market with
other major developed markets. Understanding the linkages and correlation
between these stock markets are important for policy makers and fund managers
in their investment decision and risk management.
2. It provides knowledge to investors who are considering investing in India and
China to take advantage of their rapid growth in order to enhance portfolio
return. By knowing the relationship between the markets, it can help investors to
hedge some risk in their international portfolio.
3. It contributes to the literature on the topic of stock market linkages focusing on
two major or leading emerging markets with other major markets.
1.5 SCOPE OF THE STUDY
This study focused on the linkages or relationship of the two major emerging markets
i.e. Chinese and Indian market with the other major developed markets over the period
of January 2000 to December 2008, a total of 1872 observations. Since China and India
have two main active stock markets in each country, we use will use both stock markets
in this study. Four stock markets is used to represent the major developed markets under
5

studies are United States (U.S.), United Kingdom (U.K.), Japan and Hong Kong.
For the purpose of this study, daily indices data from each stock market are intensively
used mainly to determine the relationship among stock markets since stock indices is
the widely used indicator to represent whole stock market of a country. The indices
used are as follows:
China : Shanghai Stock Exchange Composite Index
Shenzen Stock Exchange Composite Index
India : Bombay Stock Exchange Sensitive Index (Sensex)
National Stock Exchange S&P CNX Nifty Index (Nifty)
United States : Dow Jones Industrial Average (DJIA)
United Kingdom : FTSE 100 Index (FTSE)
Japan : Nikkei 225 Stock Average (Nikkei)
Hong Kong : Hang Seng Index (HSI)
1.6 LIMITATIONS OF THE STUDY
There are some limitations in this study. Since this study is exploratory in nature. It
attempts only to analyze the relationship of Chinese and Indian markets with four other
major developed markets namely United States, United Kingdom, Japan and Hong
Kong. It does not however, provide an in depth volatility measure of the stock markets.
6

1.7 ORGANIZATION OF THE STUDY
This study is structured into five chapters. Chapter 1 presents the background of the
study, problem statement, objectives, importance, scope, limitations of the study.
Chapter 1 also states an overview of Chinese and Indian stock markets and its roles to
the world economy.
Chapter 2 contains the literature review on international stock market linkages. This
chapter summarizes the relevance past literature on the subject matters.
Chapter 3 explains the data and methodology used in this study. It includes a
description about the data, data sources and the instrumentation used. Brief explanation
on the econometrics methodology adopted is also discussed in this chapter.
Chapter 4 discusses the findings of the analysis for this research. The data will be
interpreted and elaborated in order to achieve the objective of the study.
Lastly, Chapter 5 highlights the key findings and conclusion developed from the
research findings. Besides that, some recommendations will also be stated as the
guideline and suggestions to be considered for future study on related topic.
7

CHAPTER 2
OVERVIEW OF CHINA AND INDIA & LITERATURE REVIEW
2.0 INTRODUCTION
To start of this chapter, we will discuss on the economy of China and India as well as
their roles to the world economy. Understanding of each market is required before
analyzing their relationship with other major markets. Hence, we will briefly discuss the
background and development of both Chinese and Indian stock markets. We will also
look at some of the main literatures and past studies on stock market linkages or
interdependencies on international stock markets, Asia Pacific as well as Chinese and
Indian markets itself.
2.1 CHINA’S AND INDIA’S ROLE TO WORLD ECONOMY
The economic liberalization reforms undertaken by China and India are the key
elements to the success of their economy. Both China and India had highly restrictive
trade regimes until the late 1970s when they began to open to international trade. Their
openness to trade has led to greater capital flows trough imported inputs, new
technology and larger markets and spurs growth.
Both China and India have reaped handsome returns to opening up. They are the two
leading emerging economies that constitute unprecedented stories of economic
development. Measured on purchasing power parity (PPP), China is the second largest
economy in the world behind United States (U.S.) in 2007 while India ranks at fourth
behind Japan (see Table 2.1 and Figure 2.1).
8

Economy
Table 2.1
Top 10 Economies in 2007
GDP (PPP) USD
trillion
% Contribution to
world growth
United States 14.11 21.4%
China 7.104 10.8%
Japan 4.263 6.5%
India 3.065 4.6%
Germany 2.806 4.3%
United Kingdom 2.215 3.4%
Russia 2.089 3.2%
France 2.067 3.1%
Brazil 1.795 2.7%
Italy 1.787 2.7%
World 65.95 100.0%
Source: IMF World Economic Outlook 2008
Source: IMF World Economic Outlook 2008
Figure 2.1
Top 10 Countries by GDP in 2007
9

China and India have been growing rapidly since the early 1980s with the first cloaking
substantially higher growth rate. Specifically, average real GDP growth of China is at
approximately 10 per cent and 7 per cent for India (see Figure 2.2). The good
macroeconomic performance of both countries is expected to continue and real GDP is
expected to grow over 10 per cent in China and over 8 per cent in India in the medium
terms (IMF 2007).
Source: IMF World Economic Outlook 2008
Figure 2.2
China’s and India’s GDP Growth
China and India’s also has large and rapid expanding trade sector over the years. In
2008, China’s total value of import and export in 2007 reached USD 2.174 trillion (see
figure 2.3). China import has increased with an annual rate of 25 per cent since 2000.
This is due to the increase of the living standards which allow Chinese people to import
10

good from international market. In addition, the influence of price rises in the
international market plus further decrease of custom tax level contribute to the growth
of import in China.
On the other hand, China’s exports also increase rapidly with annual average growth
rate of 26 per cent since 2000. This is due to the continued reform on foreign trade
system since early 19 th centuries, resulted to the rose of partnership and private
enterprises which contribute to the exporting numbers.
Source: Bloomberg
Figure 2.3
China’s Trade
India also has undergone an impressive growth of international trade. Its import has
increased with an annual rate of 23 per cent since 2000. India’s exports also increase
rapidly with annual average growth rate of 21 per cent since 2000. This is in part due to
11

continued structural change and relaxation in industrial licensing and FDI restrictions.
Source: Bloomberg
Figure 2.4
India’s Trade
With the rapid economic expansion, China has emerged as a global manufacturing hub
and India as the service destination of the world. In summary, China and India had and
will continue to fuel the world economy and growth in the coming century making
investment in both countries attractive.
12

2.2 AN OVERVIEW OF CHINA AND INDIA STOCK MARKETS
China and India has experience an impressive economic development during the past
two decades and becoming the world fastest growing economy with substantial growth.
With its economic liberalization and deregulation, both countries have attracted the part
of the world’s policy makers and investors. Therefore, understanding the background of
each stock market may be prerequisite to the analysis of the relationship of these stock
markets with other major developed stock markets.
2.2.1 CHINA STOCK MARKET
In the early 1980s, China had adopted an open policy and widening foreign
economic relation and trade. This has led to the adoption of a legal framework to
facilitate foreign economic trade, foreign direct investment, the creation of
special economic zones, and involvement in international financial markets.
These changes has benefited the Chinese economy thus integrated China into the
world economy especially after the admission of China into the World Trade
Organization (WTO) in 2001.
China's primary trading partners included Japan, the EU, the U.S., South Korea,
Hong Kong, and Taiwan. China had a trade surplus with the U.S. of $83 billion
in 2000. With bilateral trade exceeding US$38.6 billion, China is India's largest
trading partner. China's global trade exceeded $2.4 trillion at the end of 2008.
China’s stock market has experienced tremendous growth and development
since the inceptions of the Shanghai Stock Exchange and the Shenzhen Stock
13

Exchange. The Shanghai Stock Exchange (SSE) was founded on November 26,
1990 and in operation on December 19, the same year. It is a membership
institution directly governed by the China Securities Regulatory Commission
(CSRC). After several years' operation, the SSE has become the most
preeminent stock market in Mainland China with over 71.30 million investors
and 860 listed companies as at December 2007. The total market capitalization
of SSE hit RMB 26.98 trillion (approximately USD3.7 trillion).
Established on 1990, the Shenzhen Stock Exchange (the SSE) is a mutualized
national stock exchange under the China Securities Regulatory Commission (the
CSRC), that provides a venue for securities trading. A broad spectrum of market
participants, including 540 listed companies, 35 million registered investors and
177 exchange members, create the market. As at December 2007, total market
capitalization of Shenzhen Stock Exchange stood at approximately USD800
billion. The market trades four hours a day and five days a week: 9:30 - 11:30
am; 1:00 - 3:00 pm. T+1 settlement is implemented.
In China, there are two types of shares namely class A and Class B shares. Class
A shares is the shares restricted to domestic investors, listed on either the
Shanghai or Shensen stock exchanges and are denominated in Renminbi (RMB).
These shares are not freely convertible to any international currencies. Chinese
investors are not allow to buy shares listed in Hong Kong and overseas, and
foreign investors are not allowed to trade class A shares. The class B shares
were created specifically for foreign investors and have been available to local
Chinese investors since 2001. However, their market is hardly compares with
14

that A shares which more numbers of shares has traded.
2.2.2 INDIA STOCK MARKETS
Much as in China, rising foreign exchange reserves and improved regulation in
stock markets have contributed to India's stock exchanges performing extremely
well in recent years. And because of increased foreign exchange reserves, India's
sovereign credit rating has improved to investment grade.
There are 23 recognized stock exchanges in India, including the Over the
Counter Exchange of India for providing trading access to small and new
companies. However, two major stock exchanges in India are the Bombay Stock
Exchange (BSE) and the National Stock Exchange (NSE). Established in 1875,
the Bombay Stock Exchange is the first stock exchange in the country. Over the
past 133 years, BSE has facilitated the growth of Indian corporate sector by
providing it with an efficient access to resources. Due to the generally high cost
of trading and obscure listing approvals, the Indian government made an attempt
to open up and liberalize BSE.
As an attempt to encourage competition, the Government of India set up the
National Stock Exchange of India in 1992 to serve as the second primary stock
market in India. The NSE was promoted by leading financial institutions with
the support of the Government of India. As a result of this, the cost of stock
trading in India has moved in line with global standard levels.
15

The BSE has about 4,800 Indian companies listed and significant trading
volume. As of December 2007, BSE's market capitalization stood at USD 1.79
trillion. Despite having only been established in 1992, the National Stock
Exchange also has market capitalization that is around US$ 1 trillion. In 2006,
Indian firms were able to raise US$19 billion from global markets.
2.3 EMPIRICAL EVIDENCE ON INTERNATIONAL STOCK MARKETS
LINKAGES, CO-MOVEMENT AND INTERDEPENDENCIES
Numerous studies have been done to investigate stock market linkages, integration or
interdependence. Stock market is said to be integrated when correlation exists between
markets. Although the results of these studies are mixed, inconsistent and sometimes
contradict with each other, the ultimate motivation behind the studies is the benefit of
diversification. If evidence of stock market linkage were found, it would imply that
there is a common force that brings these markets together. Hence, the benefit of
diversification would be limited.
Apart from analyzing only the interdependencies of stock markets, many researchers
have also focused on the impact of major events such as market crisis, market
liberalization, etc on the stock market linkages. Since the early 1990s, most of the
research on international stock market linkages has been concentrated on the mature and
emerging markets.
Jeon and Furstenburg (1990) perform a study on the inter-relationship among major
16

markets. They examine the relationships among stock prices in the major world stock
exchanges of Tokyo, Frankfurt, London and New York, using the vector autoregression
(VAR) approach to daily stock price indices of those markets for the period between
January 1986 and November 1988. The study shows a significant structural change with
regard to the correlation structure and leadership in the world stock markets crash of
October 1987. The degree of international co-movements in the stock price indices has
increased significantly since the crash.
Arshanapalli and Doukas (1993) focused their analysis on the capital markets of the
U.K., France, Germany and U.S. and they performed statistical tests for the existence of
bivariate cointegration between the U.S. stock market with each of the other major
markets. They found out that there exists cointegration for all potential pairs and
therefore there are small benefits from the international portfolio diversification of the
American investors.
Studies by Hilliard (1997) and Ioannis Asimakopouus, John Goddard and Costas
Siripoulos (2002) have investigated interrelationship between daily returns generated by
major stock exchanges. Evidence is found that strong interdependence exists between
the daily returns generated by United States and other selected major world indices.
Relationship between Latin American markets and the U.S market has been
documented in the study done by Fernández-Serrano and Sosvilla-Rivero (2002). They
investigate the long-run relationships between six major Latin American stock markets
and the U.S. during the period of 1995 to 2002. They use daily closing prices of six
major Latin American markets namely Argentina, Brazil, Chile, Mexico, Peru and
17

Venezuela and the U.S. market. Besides conducting normal cointegration test, they also
employ cointegration that allow for structural shifts in the cointegration vector. The
result suggests that without structural breaks, they only find cointegration in the cases of
Brazil and Mexico with U.S market. In contrast, allowing for structural breaks, they
find strong evidence in the relationship between the Argentine, Chilean and Venezuelan
indices and the Dow Jones index which represent U.S. market after the 1998 financial
turmoil and between the Brazilian and Mexican indices and the Dow Jones index before
such turbulence.
Aggarwal, Lucey and Muckley (2003) in their paper have examine time-varying
integration of European equity markets over the 1985 to 2002 period using daily data
for the main EU countries. They use estimates of traditional cointegration, the Haldane
and Hall Kalman filter technique, and dynamic eigenvalue analysis in their study. The
result shows the evidence of integration in European countries only after the
establishment of EMU and the ECB during 1997-98 periods. Result also indicates that
Frankfurt is the dominant market for equities in Europe.
In the paper by Canarella, Miller and Pollard (2008), they explore the dynamic linkages
between stocks market returns in NAFTA (i.e., Canada, Mexico and the U.S.). They
employ daily closing price of S&P TSX Composite Index in Canada, the IPC index in
Mexico and the S&P500 index in U.S which cover the period of January 1, 19992 to
December 31, 2007. They use cointegration techniques of Johansen and Juselius to
examine the long-run relationship between the three markets and impulse-response
analysis to evaluate the dynamic relationship between the three markets. In the study,
they fail to discover evidence that the NAFTA stock indices share long-run equilibrium
18

elationship since there is no evidence of a common long-term trend between the three
indices. They also find that the response response of each market to shocks in its own
market always exceeds the response to shocks in other markets with U.S. holds a
leading role.
Due to the substantial increase of capital flows from mature markets to emerging
markets of the Asian countries, considerable attention has been given to study the
possible linkages and interdependencies in major Asian countries. The general
consensus is that the correlations between emerging and developed stock markets are
generally on the increase.
Royfaizal, R. C, Lee, C and Mohamed, Azali (2007) analyse the stock markets
interdependencies between the Asean-5+3 and U.S. stock markets before, during and
after Asian financial crisis by using weekly stock indices expressed in local currencies
from July1997 to June 1998. They employ Granger-causality test based on VECM to
test the long run relationship among the stock markets. The study shows that the long-
run relationships between ASEAN 5+3 stock markets occur only for during- and post-
crisis period. They also found that US become dominant compare to other countries
after the crisis.
Choudhry et. al. (2007) studies the changes in the long run relationship between eight
Far East countries namely Thailand, Malaysia, Indonesia, Hong Kong, Singapore, the
Philippines, South Korea and Taiwan around the Asian financial crisis of 1997-98. They
also examine the change in the influence of the U.S. and Japanese stock markets in the
Far East region before, during and after the Asian financial crisis using daily stock price
19

indices from January 1, 1998 to January 1, 2003. Correlation coefficients, multivariate
cointegration, causality test and regression are conducted and results shows significant
long-run relationship and linkages between the Far East stock markets before, during
and after the crisis. They also find larger U.S. influence in all periods and some
evidence of increasing Japanese influence to the eight Far East countries.
Abbas Valadkhani and Surachai Chancharat (2008) has investigate the existence of
cointegration and causality between the stock market price indices of Thailand and its
major trading partners (Australia, Hong Kong, Indonesia, Japan, Korea, Malaysia, the
Philippines, Singapore, Taiwan, the UK and the USA), using monthly data spanning
from December 1987 to December 2005. They used both the Engle-Granger two-step
procedure (assuming no structural breaks) and the Gregory and Hansen test (allowing
for one structural break). From the result, they found evidence of potential long run
benefits from diversifying the investment portfolios internationally. They also found
that the stock returns of Thailand and three of its neighbouring countries (Malaysia,
Singapore and Taiwan) are interrelated.
Lim L.K. (2007) conducts a study to examine the dynamic interdependencies of five
ASEAN stock markets i.e. Indonesia, Malaysia, Philippines, Singapore and Thailand
with US stock market over the period of April 1990 to July 1997 using daily total
market-return indices for each stock market. The result indicates higher average returns
and correlations over the post crisis period. The result also indicates an increase in the
integration between the ASEAN-5 markets after the financial crisis and US market
returns have significant influence on the returns of all ASEAN-5 markets.
20

Another studies done by Click and Plummer (2005) who examine whether the ASEAN-
5 stock markets are integrated or segmented using cointegration technique using daily
and weekly stock index quotes in local currency data from July 1998 to December
2002. The empirical result suggests that the ASEAN-5 stock markets are cointegrated.
However, only one cointegrating vector is found, leaving four common trends among
the five variables. Hence, the ASEAN-5 stock markets are integrated, but the
integration is still far from complete.
Tan and Tse (2002) use daily data in local currencies over 1988-2000 to examine the
linkages among U.S., Japan, and seven Asian stock markets including Malaysia,
Philippines, Singapore, and Thailand. By truncating the data at the end of 1996 and
restarting the data in mid-1998 to create a pre-crisis and post-crisis comparison, they
find that markets appear to be more integrated after the crisis than before. They also
find that Asian markets are most heavily influenced by the U.S. but that the influence of
Japan is increasing. Another interesting result is that Malaysia is less affected by the
U.S. and Japan after the crisis, which can be attributed to the success of its capital and
currency controls, but Singapore and Malaysia still affect each other strongly, which
can be attributed to geographic proximity, economic linkages, and structural symmetry.
Siklos, P.L, and Ng, P. (2001) examines whether there is a common stochastic trends in
the U.S., Japan, Hong Kong, Korea, Singapore, Taiwan and Thailand using stock
market indices from each countries spanning from January 1976 to August 1995. They
employ Dicker Fuller and Philip-Perron to test the stationary property of the series and
using Vector Error Correction Model (VECM) and cointegration test to examine the
relationship between the indices. The result reveals that stock market integration is
21

largely features post-1987 U.S. stock market crash, and intensified during 1990s. Before
the 1987 crash, all seven stock market under study did not share the same common
trend, which means that investor in Asia Pacific market did not exploit diversification
opportunities. However, result shows that Asia Pacific stock markets appear to behave
as if trend in the U.S. and Japanese market did influence their stocks markets.
Cheung and Ho (1989) conduct a study on the causal relationship between the US
market and four Asian–Pacific markets, i.e., Australia, Hong Kong, Singapore and
Malaysia and find that a bi-directional relationship exist between the US and Singapore.
However, a unidirectional relationship running from the US market to the Hong Kong
market and to the Malaysian market is found.
Muhammad Naeem (2002) conducted a study on the interdependence of the major stock
markets in South Asia and the linkages between the markets with U.S and U.K stock
markets using monthly data from January 1994 to December 1999. Using both bivariate
and multivariate co integration analysis, he found no co integration between the South
Asian stock markets indices for the entire period but found co integration for the pre-
nuclear test period i.e from January 1994 to April 1998.
Golaka C Nath and Sunil Verma (2003) examine the interdependence of the three major
stock markets in South Asia. Using daily stock market indices of India (NSE-Nifty),
Singapore (STI) and Taiwan (Taiex) from January 1994 to November 2002, they
employ bivariate and multivariate co-integration test. The result shows that no
cointegration was found for the entire period thus conclude that there is no long run
equilibrium between India, Singapore and Taiwan.
22

2.4 STUDIES ON CHINESE AND INDIAN MARKETS
Although there are numerous literatures on financial or stock market integration, most
focus were on the developed markets. Until the late 1980s that Asian-Pacific stock
markets, which is also called “emerging market”, had grown in importance, a
considerable amount of work has been done to investigate the relationship and linkages
among the markets. However, only few studies that concentrating on the relationship of
two major emerging markets i.e. China and India with other major developed stock
markets in the world.
For example, Sharma and Kennedy (1977) examine the price behavior of Indian market
with the US and UK markets and conclude that the behavior of the Indian market is
statistically indistinguishable from that of the US and UK markets and find no evidence
of systematic cyclical component or periodicity for these markets. Other study by
Wong, Agarwal and Du (2005) using the BSE 200 data have found that the Indian stock
market is integrated with the matured markets of the World.
Chen, H., Lobo, B. J., and Wong, W. K. (2006) examines the bilateral relations between
three pairs of stock markets, namely India-U.S., India-China and China-U.S. They use
weekly stock index of Bombay Stocks Exchange National Index for India, All Shares
Index from Shanghai Stock Exchange for China, and the S&P 500 index for U.S.
market from January 2, 1991 to December 29, 2004. They apply Augmented Dickey-
Fuller and Phillip-Perron unit root test to test the stationary of the series and
subsequently employ a Fractionally Integrated Vector Error Correction Model
(FIVECM) to detect the co-movement of the pairs of stock markets. The result shows
23

that the three markets are fractionally co-integrated with each other. They also find that
U.S. market leads the Indian market in information spillover and leads the Chinese
market in return transmission. The result also suggests that the two emerging markets
appear to be more closely linked to each other relative to the U.S.
Study done by Chattopadhyay and Behera (2006) to examine whether reform in Indian
stock market has led to integration with the developed stock markets in the world
suggest that Indian stock market is not co-integrated with the developed market as yet
although some short-term impact does exist. The study also does not find any causality
between the Japanese stock market and Indian stock market.
More recently, Janak Raj and Sarat Dhal (2008) investigated the financial integration of
India’s stock market with global and major regional markets. They use six stock price
indices: the 200-scrip index of BSE of India to represent domestic market, stock price
indices of Singapore and Hong Kong to represent the regional markets and three stock
price indices of U.S., U.K., and Japan to represent the global markets. They use both
daily and weekly data from end-March 2003 to end-January 2008. To gauge the
integration of India’s stock market with global markets and major regional markets,
they employ correlation and the vector error correction and cointegration model
(VECM). The results suggests that Indian market’s dependence on global markets, such
as U.S. and U.K., is substantially higher than on regional markets such as Singapore and
Hong Kong while Japanese market has weak influence on Indian market.
Due to the limited studies on the linkages between China and India with major
developed markets, this study will focus on analysing whether the two major emerging
24

markets decouple from major developed markets namely U.S., U.K., Japan and Hong
Kong.
2.4 SUMMARY
With the massive growth of China and India in terms of economy and population, they
becoming among the major contributors to the world economy. Hence, both markets are
appealing to investors due to potential returns as their economies growing. Previous
studies have proven that stock market linkages or interdependencies are worth knowing
for portfolio diversification and return enhancement.
25

3.0 INTRODUCTION
CHAPTER 3
RESEARCH METHODOLOGY
Before discussing further on the methodology adopted in this study, we will briefly run
through few methodologies adopted in previous studies by other researchers. This
chapter will further elaborate on the description about the data i.e. all the indices used to
represent each stock markets. Data sources and the instrumentation used will also be
explained in this chapter.
3.1 EMPIRICAL WORK
Various studies discussed the issue of the relationship among stock markets in the last
decades. The majority of those studies used one or more correlation or regression
models to examine the price, returns or indices for different periods. Apart from
appropriate econometric technique used, there are several other aspects being
considered such as the time period to examine, whether there has been changes over
time or any structural breaks in relationships such as those events or crisis. Another
aspect to consider is whether the study should be conducted in the local currencies, U.S.
dollars or some other unit.
There are various methods used in measuring the relationship among stock markets.
The estimation of correlation coefficient is the most commonly method used in
estimating the relationship among stock returns. An increase in correlation is used as an
26

evidence of an increase in the linkage across stock markets. Ng (2000), Hashmi and Liu
(2001) and Sabri (2002) conducted the analysis of correlation to reveal the link between
the rates of changes in the Asian stock indices. Other study by King and Wadhwani
(1990) used cross-market correlation coefficient to identify whether there is any stock
market correlation between the United States, the United Kingdom, and Japan.
However, the use of correlation coefficient in the previous empirical studies on co-
movement of stock market price indices is questionable because the correlation
coefficients do not provide information on causal relationships between variables in the
model.
Cointegration analysis has been the most popular approach employed by academicians
and researchers in establishing linkages or relationship between financial markets.
Cointegration analysis was initially introduced through influential contributions by
Granger (1981), Eagle and Granger (1987) and Granger and Hallman (1991).
Granger (1969) proposed a test to address the question whether some variables causes
others. The test namely Granger causality consist of running regressions of one stock
return on its lagged values and on other stock returns. Hence, if the lagged values of one
stock return do not yield a statistically significant relationship, then it can be stated that
the stock return does not Granger-cause the other stock return. Mathur and
Subramanyam (1990), Arshanapalli and Doukas (1993), Malliars and Urrutia (1994),
Gerrits and Yuce (1999) and Yang (2002) used the concept of Granger causality to
analyze the linkages among stock prices. Recent study done by Sadhan Kumar
Chattopadhyay and Samir Ranjan Behera (2007) also used Granger causality to
investigate whether reform in Indian stock market has led to integration with the
27

developed stock markets in the world.
A recent feature of literature of stock market linkages is the use of tests for co
integration either bivariate or multivariate technique. Taylor and Tonks (1989) were the
first to apply bivariate co-integration in the UK and U.S markets to test the importance
of the latter after the abolition of foreign exchange controls in 1979. They found that
stock price index of the U.K. was cointegrated with the stock price index of the U.S.,
Germany and Netherland. Arshanapalli and Doukas (1993) also used bivariate co-
integration to explore changes patterns of dynamics interactions among national stock
market indices following the October 1987 crash.
Another group of studies has concentrated on examining stock market linkages by using
either multivariate co-integration methodology. A well-cited paper is Kasa (1992), one
of the first researchers to employ the Johansen co integration method to examine stock
prices integration. Other study, Phylaktis and Ravazzolo (2000) also used multivariate
cointegration analysis of Johansen to investigate the linkages among groups of stock
price levels by looking for the existence of potential linear combinations among them.
Hung, Sing, Cheung and Leung (2000) also apply the Johansen multivariate co-
integration test to examine the interdependence of the Asian emerging equity markets.
Many studies of stock markets have used the popular time series technique of vector
autoregressive model (VAR) to examine the relations among stock returns between
stock markets. The advantages of VAR analysis are that the VAR model is not subject
to any priori restrictions on the structural relationships among the variables, and the
analysis of the pattern of innovations and responses in different markets can be
28

precisely performed by the impulse response function analysis. A number of studies
have applied VAR models, for example, Eun and Shim (1989) have used VAR model to
identify main channels of interactions among stocks markets and to trace out the
dynamic response of one market to innovation in another. Other studies include Ammer
and Mei (1996), Yuce and Mugan (2000), Wu and Su (2001) and Yang, Kolari and Min
(2002).
3.2 DATA
Data used in this study are the stock market indices for every stock market. For Indian
market, we used Bombay Stock Exchange (BSE) Sensex 30 Index and National Stock
Exchange (NSE) S&P CNX Nifty. Meanwhile, for Chinese market we use Shanghai
Stock Exchange Composite Index and Shenzhen Composite Index.
The data for four major markets are Dow Jones Industrial Average (DJIA) for the U.S.,
FTSE-100 for U.K., Nikkei-225 Stock Average for Japan and Hang Seng Index for
Hong Kong. These indices are the most representative indices that reflect each market’s
performance.
Below are the brief explanations of each index.
Shanghai Stock Exchange Composite Index (Shanghai): Shanghai Index is a
capitalization-weighted index. The index tracks the daily price performance of
all A-shares and B-shares listed on the Shanghai Stock Exchange.
29

Shenzen Stock Exchange Composite Index (Shenzen): Shenzen Composite
Index is an actual market-capilatization weighted index with no free float, that
tracks the strong performance of all the A-share and B-share lists on Shenzen
Stock Exchange.
Bombay Stock Exchange Sensitive Index (Sensex): BSE Sensex 30 index is a
"market capitalisation-weighted" index of 30 stocks representing a sample of
large, well-established and financially sound companies. The Sensex is
calculated using a market capitalisation-weighted methodology 2 .
National Stock Exchange S&P CNX Nifty Index (Nifty): The S&P CNX
Nifty, a weighted average index is the leading index for large companies on the
National Stock Exchange of India. It consists of 50 companies representing 24
sectors of the economy,
Dow Jones Industrial Average (DJIA): DJIA is a price-weighted average
index of 30 blue-chip stocks that are generally the leaders in their industry. A
stock is typically added only if it has an excellent reputation, demonstrates
sustained growth, is of interest to a large number of investors and accurately
represents the sector(s) covered by the average. The index has been a widely
followed indicator of the stock market since October 1, 1928.
FTSE 100 Index (FTSE): This is the most widely quoted and popular index for
tracking the London stock exchange. FTSE-100 is a market capitalisation-
2 Market capitalization of a company is determined by multiplying the price of its stock by the number of shares
issued by the company.
30

weighted index, re-weighted every day. The index comprises of the 100 most
highly capitalized companies traded in London Stock Exchange.
Nikkei 225 Stock Average (Nikkei): The Nikkei is a price-weighted 3 average
of 225 top-rated Japanese companies listed in the First Section of the Tokyo
Stock Exchange.
Hang Seng Index (HSI): Hang Seng is a market capitalisation weighted index
based on 33 major stocks of Stock Exchange of Hong Kong (SEHK). The Hang
Seng Index (HSI) is a well-known benchmark for the performance of the Hong
Kong stock market.
3.3 INSTRUMENTATION AND SCALES
This study uses daily closing price for each index from January 2000 to December
2008, a total of 1,872 observations. The daily closing price data of the eight indices is
obtained from Bloomberg. The data considered only for those days where markets were
open in all the market.
Daily data has been used in order to capture potential interactions, for example, impulse
responses, because a month or even a week may be long enough to obscure interactions
that may last only a few days (Cotter, 2004). However, Yang, Kolari and Min (2002)
state that the frequency of data likely has only limited effects on the cointegration
analysis, as Hakkio and Rush (1991) have shown that, given a fixed sample period,
frequency of data did not affect cointegration results.
3 The value of the index is generated by adding the prices of each of the stocks in the index and dividing them by the
total number of stocks.
31

All of the indices are express in terms of local currencies to avoid problems associated
with transformation due to fluctuations in exchange rates and also to avoid the
restrictive assumption the relative purchasing power parity holds. In addition, the
preference for local currencies is focus on the domestic causes of stock market
interdependence. According to Leong and Felmingham (2001), by converting these
indices to a common currency there is a possibility that the impact of local economic
conditions and domestic economic policy maybe distorted. In addition, earlier studies
have found similar results if the price indices were measured in local currencies or were
converted into a common currency, usually into dollars. The data will be analysed
using E-View 6.1 which provides sophisticated data analysis, regression and forecasting
tools.
3.4 METHODOLOGY
In order to analyse the linkages of Indian and Chinese market with other major markets,
the data need to be statistically analyzed. This study adopts three methods available in
the literature as follows:
3.4.1 THE CORRELATION TEST
The first step of the analysis of stock market integration in this study involves a
simple correlation test to measure the strength and direction of the association
between the stock indices. The correlation value can range between -1.00 to
+1.00. The positive and negative sign indicate the direction of the relationship.
A positive sign indicates that the two indices covary in the same direction. On
32

the other hand, a negative value suggests that the two indices covary in opposite
directions.
The strength of the relationship is indicated by the magnitude of the correlation
coefficient. If the value of the coefficient is 0, this means that there is no linear
relationship between the two indices. If, on the other hand the absolute value of
the coefficient is 1 then there is a perfect linear relationship between the two
indices.
The significance of the correlation for each index provides a preliminary
indication about the strength of association between the indices of different
stock markets under study. According to Su Chan Leong and Bruce Felmingham
(2001), correlation coefficients are known to be biased upward if the share price
indices are heteroskedastic, thus it does not provide a sound basis for studies of
interdependence.
However, the analysis of correlation provides a commonly used preliminary
technique as adopted by few researchers in the earlier studies. Ng (2000),
Hashmi and Liu (2001) and Sabri (2002) conducted the analysis of correlation to
reveal the link between the rates of changes in the Asian stock indices. Other
study by King and Wadhwani (1990) used cross-market correlation to identify
whether there is any stock market correlation between the United States, the
United Kingdom, and Japan.
Correlation analysis only measures the degree of linear association between two
33

variables hence provides little insight on the dynamic linkages and causality
between stock markets. Therefore, we extend the analysis of stock market
integration by employing Granger Causality test.
3.4.2 GRANGER CAUSALITY TEST
Granger Causality test are conducted to further analyse the significance and
direction of causality between Chinese and Indian market with other major stock
markets. According to Granger (1969), this test will answer the question of
whether X causes Y. Y is said to be Granger-caused by X if X helps in the
prediction of Y, or equivalently if the coefficient on the lagged X are statistically
significant.
To show that X Granger cause Y, the first step is to consider an autoregression
for Y. Next, lagged values of X are added as extra independent variables.
Granger Causality test result is very sensitive to the number of lags used in the
analysis. Davidson and MacKinnon suggest using more rather than fewer lags.
There are four different criteria for specifying the lag length. This study will
adopt Akaike information criterion (AIC) suggested by Akaike (1970, 1973 and
1974).
34

.
The equation for the pairwise Granger causality tests are as follow:
Where,
Yt = αo + α1Yt-1 +...+ αiYt-1 + β1Xt-1+...+ βiXt-1+ µt
Xt and Yt = daily stock market index for country X and Y respectively
µt = error term at time t
The F test is used to test the hypotheses of the Granger Causality as follow:
H0: β1 = β2 = 0 (X does not Granger cause Y)
H1: At least one of the β1 ≠ 0
The null hypothesis is rejected if the computed F-value exceeds the critical F
value at the chosen level of significance. This implies that X does Granger cause
Y. The test will be performed in pair form between China and U.S, U.K, Japan
and Hong Kong. Similarly, the same method is repeated between India and U.S,
U.K, Japan and Hong Kong. We will also examine the causality between Indian
and Chinese market.
3.4.3 UNIT ROOT TEST
Before performing the co-integration analysis, the univariate properties of the
data series needs to be examined whether the series are non-stationary or contain
unit root. Other researchers also did this prior to the cointegration test.
35

A series is said to be stationary if the mean and variance of the series do not
systematically differ over the time period. In other word, they have constant
mean, constant variance and constant autocovariances for each given lag.
Regression in which the variables are non-stationary can lead to spurious result
where variables may share the same time trend even though they are not really
related.
Unit root test involves examining whether the series are stationary or not at level
and subsequently finding the order in which they are integrated if the series is
non-stationary. This study employs Augmented Dickey-Fuller (ADF) test to
determine the unit root property of the stock market indices.
This requires regressing ∆Yi on a constant, a time trend ∆Yt-1 and several lags of
dependent terms as follows:
Where,
∆Yt = γo + γ1Yt-1 + βi∑Yt-1+ εt
∆ = first difference operator
γo, γ1 and βi = coefficients to be estimated
Yt = non-stationary time series
εt = error term at time t
The following hypotheses are tested:
Ho:γ=0 (series contain a unit root)
H1:γ≠0 (series is stationary)
36

The test statistics known as the tau statistics are checked against the critical
values whose has been tabulated by Dickey and Fuller on the basis of Monte
Carlo simulations 4 . The null hypothesis of series contain a unit root is rejected if
t-statistics is smaller (more negative) than the critical value respectively.
Another indicator to look at is the Durbin-Watson (DW) test statistics. DW test
statistics has to be 2 or very close to 2 to verify that the test result is reliable i.e.
no autocorrelation problem.
After getting the order of integration using ADF test, the presence of co
movement between the stock market indices is testing using co-integration test.
3.4.4 COINTEGRATION TEST
Co-integration test is among the most widely used method in examining the
relationship or integration in financial market. The concept of co-integration was
introduced by Granger (1981, 1986) and further developed by Engle and
Granger (1987) which incorporates the presence of non-stationary, long-term
relationship and short-run dynamics in the modelling process.
The fundamental objective of co-integration analysis is to detect any common
stochastic trends in the price data and subsequently to use these common trends
for a dynamic analysis of the correlation in stock index. Eagle and Granger
(1987) propose a two step estimation method, where the first step consists of
estimating a long term equilibrium relationship and the second is the estimation
4
D. A. Dickey and W. A. Fuller, “Distribution of the Estimators for Autoregressive Time Series with a Unit Root,” 37
Journal of the American Statistical Association, vol. 74, 1979, pp. 427-431.

of the dynamic error correction relationship using lagged residuals. We will use
this method to test the presence of co-integration among the stock indices.
Holden and Thompson (1992) state that this two step method has the advantage
that the estimation of the two steps is quite separate, so that changes in the
dynamic model do not enforce re-estimation of the static model obtained in the
first step. As such, it offers a tractable modelling procedure.
A series is said to be integrated of order one i.e., I (1) if it becomes stationary
after the first differencing. If there exists a linear combination of two or more I
(1) series that it itself stationary, then the series are co integrated. Co-
integration necessitates that the variables be integrated of the same order. Thus,
this study employs Augmented Dickey Fuller (ADF) tests to determine the order
of integration for every stock return.
If both the variables are integrated of same order, we then estimate the co-
integrating regression using OLS.
Yt = β0 + β1 Xt+ µt
Where Y and X are non-stationary series. The residuals, denotes as µt are then
tested to ensure that they are I (0) by running Augmented Dickey-Fuller (ADF)
test. The time series is said to be co-integrated if the residual is itself stationary,
I (0). The residual will still be non-stationary if the time series are not co-
integrated. In effect the non-stationary I (1) series have cancelled each other out
38

to produce a stationary I (0) residual.
The following hypotheses are tested:
Ho: µt ~ I (1) error term or residuals contain unit root
H1: µt ~ I (0) error term or residuals is stationary
The test statistics against the critical values are checked. If t-statistics is smaller
(more negative) than the critical value, null hypothesis of residuals contain unit
root is rejected and conclude that the residuals or the error term is stationary.
This would imply that the two stock indices are co-integrated.
Another test statistics that can be used to test for non-stationarity of the µt is
using Durbin-Watson (DW). Here, DW is known as the Cointegrating
Regression Durbin Watson (CRDW) since it is applied to test the residuals of
potentially co-integrating regression. The critical values for CRWD were first
provided by Sargan and Bhargava (1983) 5 . The null hypothesis, d= 0 is rejected
if the computed d value is smaller than the critical value which means that the
two stock indices are not co-integrated. On the other hand, if the computed d
value is larger than the critical value, then it can be said that the two indices are
co-integrated.
5 . D. Sargan and A. S. Bhargava (1983). “Testing Residuals from Least Squares Regression for
Being Generated by the Gaussian Random Walk,” Econometrica, vol. 51, pp. 153-174
39

3.4.5 ERROR CORRECTION MODEL (ECM)
If the two stock indices are co-integrated, thus it also implies that there is a long-
term equilibrium relationship between the two. To test if there may be
disequilibrium in the short-run, we can use the residual of the co-integrating
regression to tie its short-run behaviour to its long-run value. According to the
Granger Representation theorem, if two variables are co-integrated, then the
relationship between the two can be expressed as an error correction model
(ECM). ECM was first used by Sargan and later popularized by Eagle and
Granger corrects for disequilibrium.
In this model, the error term from the OLS regression, lagged once acts as the
error correction term. The ECM allows the introduction of past disequilibrium as
explanatory variables in the dynamic behaviour of current variable thus enables
to capture both the short-run dynamic and long-run relationships between the
stock indices.
The basic ECM is as follows:
Yt-1 = α0 + α1 Xt-1 + α2 µt-1 + εt
Where µt-1, is the lagged value of error correction term derived from the co-
integration regression and εt is the residual or error term with the usual
properties. The model relates the changes in the dependent variable i.e., stock
index Y to the change in independent variables i.e., stock index X and the
“equilibrating” error in the previous period.
40

In this regression, Xt-1 captures the short-run disturbance in stock index X. The
F-tests of the differenced independent stock index Xt-1, give a sign of the short-
term causal effects. Meanwhile, µt-1 captures the adjustment toward the long-run
equilibrium. The long-run relationship is indicated through the significance of
the t-test of the lagged error correction term µt-1. However, the coefficient of the
error correction term α2 is a short-term adjustment coefficient. It will explain the
speed of adjustment back to equilibrium if it is statistically significant. In other
words, it represents the proportion by which the long-run disequilibrium in stock
index Y is being corrected in each short period.
3.5 SUMMARY
To sum up, four econometric methodologies are adopted to analyse the linkages of
Chinese and Indian markets with four other major developed markets namely
United States (U.S), United Kingdom (U.K), Japan and Hong Kong.
41

4.0 INTRODUCTION
CHAPTER 4
RESEARCH RESULTS
We will discuss the findings of all four econometric methodologies adopted in this
chapter. The findings will be analysed, interpreted and elaborated in order to
statistically analyse whether there is any linkages between Chinese and Indian markets
with four major developed markets.
4.1 DESCRIPTIVE STATISTICS
Some descriptive statistics for all the eight stock indices under study are given in Table
4.1. These include the distribution of mean, standard deviation, skewness and kurtosis.
For the purpose of comparison, the daily closing price index for each stock market is
transformed to the return form.
42

Table 4.1
Descriptive Statistics
SHCOMP SZCOMP SENSEX NIFTY
Mean 0.218 0.076 2.247 0.713
Median 0.480 0.380 8.120 2.750
Maximum 351.400 100.810 1222.660 349.900
Minimum -620.76 -181.03 -2283.76 -806
Std. Dev. 52.822 15.808 199.773 60.531
Skewness -1.225 -1.307 -1.129 -1.736
Kurtosis 21.293 20.376 21.475 29.526
Jarque-Bera 26469.88 23992.48 26919.85 55616.11
DJIA FTSE NKY HSI
Mean -1.116 -1.207 -5.65 -1.614
Median 4.200 0.900 -0.69 4.180
Maximum 889.350 462.150 1171.140 2332.540
Minimum -1187.63 -504.3 -1517.78 -3444.24
Std. Dev. 139.441 72.793 216.160 313.460
Skewness -0.535 -0.307 -0.648 -0.8
Kurtosis 10.413 9.229 8.154 18.526
Jarque-Bera 4358.92 3045.49 2194.624 18932.48
In term of absolute value, the mean of Sensex is greater than other indices. Meanwhile,
in the same observation period, Hang Seng index is the most volatile and the Shenzen
index is less volatile as the standard deviation of returns show. Table 4.1 also shows
that the indices’ skewness values are negative, and that all the indices have Kurtosis
values larger than 3, which indicate fat-tails. Therefore, the Jarque-Bera (JB) values of
the indices imply that none of the indices is normally distributed which is consistent
with prior literatures.
43

Figure 4.1
Movement of the Indices in the Observed Period
Figure 4.1 presents the movement of all the eight indices in the observed period from
January 2000 to December 2008. As can be seen in Figure 4.1, Hang Seng and Nikkei
record market capitalizations that are much higher than those of the other observed
indices. Interesting observation from Figure 4.1 is that Sensex’s market capitalization
started to increase in the year 2005. All stock indices started to decline or having
negative growth in the February 2008 onwards. This is contributed by the credit crisis
that took on a full head stem from August 2007 where most world stock markets are
falling in tandem with each other.
44

4.2 ANALYSIS OF CORRELATIONS
The correlations between the stock indices for the period of January 2000 to December
2008 are computed to measure the strength of the association between the stock indices.
Table 4.2 presents the simple correlation coefficients among the 8 stock indices under
study.
SHCOMP 1.000
Table 4.2
Correlation of stock price indices
SHCOMP SZCOMP SENSEX NIFTY DJIA FTSE NKY HSI
SZCOMP 0.985 1.000
SENSEX 0.749 0.686 1.000
NIFTY 0.750 0.688 0.999 1.000
DJIA 0.731 0.683 0.817 0.813 1.000
FTSE 0.530 0.529 0.532 0.522 0.814 1.000
NKY 0.417 0.397 0.514 0.500 0.762 0.912 1.000
HSI 0.828 0.791 0.918 0.918 0.895 0.733 0.671 1.000
While the numerical values of correlation coefficients may range from 1.0 to -1.0, the
majority of the correlation in Table 4.2 exceeds 0.3 and often exceeds 0.5. From the
result, we can see that both Shanghai and Shenzen are highly correlated with Hang Seng
with correlation coefficient of 0.82 and 0.79 respectively. This may be due to the close
proximity between the markets resulting in increase money flows as investors can easily
switch investments between the two markets. Thus stock price indices between China
and Hong Kong tend to correlate.
45

Correlation between Shanghai and Shenzen stock indices with both Sensex and Nifty
are also higher than correlation between them and DJIA. This suggests that both
Shanghai and Shenzen are more correlated with India as compared to U.S. in the period
under study. The increasing economic interdependence between the China and India
has contributed to the high correlation or co-movement between the two markets.
China has relatively lower correlation with U.K. and Japan as the correlation between
Shanghai and Shenzen indices with FTSE and Nikkei are lower. The result also
indicates that the correlations between Sensex and Nifty with Hang Seng index are
relatively high. This implies that India is highly correlated with Hong Kong. The
correlation between Sensex and Nifty with DJIA are also quite high suggesting that
there is a positive correlation between India and U.S.
Nevertheless, there is no negative correlation found between the eight markets under
study for the period. Overall, the results from the correlation coefficients suggest some
insight on the short-term relations between China and India with other major markets.
4.3 ANALYSIS OF GRANGER CAUSALITY TEST
The Granger Causality test is conducted to investigate direction of causality between
Chinese and Indian market with other major stock markets. The F-statistics from the
Granger Causality test results is presented in Table 4.3.
46

Table 4.3
Results of the Granger Causality Test
PANEL A
Causality F-statistics P-value
SHCOMP → DJIA 2.9592 0.0044
DJIA → SHCOMP 5.0911* 0.0000
SHCOMP → FTSE 1.8985 0.0660
FTSE → SHCOMP 5.6487* 0.0000
SHCOMP → NKY 2.0277* 0.0484
NKY → SHCOMP 1.1125 0.3522
SHCOMP → HSI 5.8967* 0.0000
HSI → SHCOMP 4.2704* 0.0001
SZCOMP → DJIA 2.7042* 0.0086
DJIA → SZCOMP 2.9075* 0.0050
SZCOMP → FTSE 1.2501 0.2717
FTSE → SZCOMP 3.6839* 0.0006
SZCOMP → NKY 2.1465* 0.0362
NKY → SZCOMP 1.0607 0.3865
SZCOMP → HSI 7.3174* 0.0000
HSI → SZCOMP 2.6358* 0.0104
PANEL B
Causality F-statistics P-value
Sensex → DJIA 1.3915 0.2045
DJIA → Sensex 13.7304* 0.0000
Sensex → FTSE 2.2727 0.0264
FTSE → Sensex 7.7012* 0.0000
Sensex → NKY 4.2241* 0.0001
NKY → Sensex 0.6580 0.7079
Sensex → HSI 2.2102* 0.0309
HSI → Sensex 3.4602* 0.0011
Nifty → DJIA 1.7164 0.1008
DJIA → Nifty 13.4686* 0.0000
Nifty → FTSE 2.3895* 0.0196
FTSE → Nifty 7.5921* 0.0000
Nifty → NKY 3.4333* 0.0012
NKY → Nifty 0.6698 0.6979
Nifty → HSI 1.9739 0.0551
HSI → Nifty 4.0323* 0.0002
47

PANEL C
Causality F-statistics P-value
SHCOMP → Sensex 2.6823* 0.0092
Sensex → SHCOMP 4.8723* 0.0000
SHCOMP → Nifty 2.5449* 0.0131
Nifty→ SHCOMP 4.1352* 0.0002
SZCOMP → Sensex 3.2893* 0.0018
Sensex → SZCOMP 3.3279* 0.0016
SZCOMP → Nifty 3.5449* 0.0009
Nifty→ SZCOMP 2.7291* 0.0081
* indicates significance at the 5% level
Panel A show the result of Granger Causality test between Chinese market and other
major markets. The test results suggest that there is a bi-directional causality between
both Shanghai and Hang Seng. A unidirectional causality exists between Shanghai and
DJIA where DJIA Granger causes Shanghai. Similarly, FTSE also Granger causes
Shanghai and hence they have a unidirectional causality. The results also show that
Shanghai Granger causes Nikkei which implies that they also have unidirectional
causality.
Similar result with Shanghai, Shenzen also has bidirectional causality with Hang Seng
and a unidirectional causality with FTSE and Nikkei. However, an interesting result to
note is Shenzen has bidirectional causality with DJIA which is not the case for
Shanghai.
Panel B show the result of Granger Causality test between Indian market and other
major markets. The results show that Sensex has bidirectional causality with HangSeng.
This result supports the high correlation found between them in the analysis of
correlation earlier.
48

Meanwhile, there is also evidence unidirectional causality from DJIA to Sensex and
FTSE to Sensex. In other words, Sensex is Granger caused by the DJIA and also FTSE.
Results also suggest that Sensex does Granger causes Nikkei.
A different causality is found in another Indian stock index i.e. Nifty. The results show
that Nifty has bidirectional causality with FTSE. There is also unidirectional causality
from DJIA to Nifty and Hang Seng to Nifty. Similar with Sensex, Nifty also does
Granger causes Nikkei.
The direction of causality between the two major emerging markets i.e. China and India
are also examined here. The results of Granger causality test between the two markets is
presented in Panel C. It is found that there is bidirectional causality between China and
India. This can be seen in the increased amount of bilateral trade between the two
countries over the years.
To sum up, overall results of Granger causality test suggest that Chinese market has
bidirectional causality with Hong Kong market. This support the evidence of high
correlation between the two markets in the analysis of simple correlation conducted
earlier. Another interesting result is the evidence of bidirectional causality between
China and India.
49

4.4 ANALYSIS OF UNIT ROOT TEST
Augmented Dickey-Fuller (ADF) test is used to test the stationarity of the stock price
indices. The result of ADF unit root test in Table 4.4 shows that the null hypothesis of a
unit root cannot be rejected, which indicates that all stock indices are non-stationary
series. However, since the t-statistics is smaller (more negative) than the critical value
in the first difference form, there is no evidence to support that the presence of unit in
the series. Hence, all the stock price indices are stationary, and integrated of order one, I
(1) which is consistent with results in the finance literature.
Stock Index
Table 4.4
Augmented Dickey-Fuller
Level
(with intercept and trend)
First Difference
(Constant)
SHCOMP -0.9696 -45.2038
SZCOMP -1.1197 -42.5494
Sensex -1.6728 -42.7364
Nifty -1.7862 -43.8788
DJIA -2.0061 -45.6734
FTSE -2.0221 -33.9525
NKY -1.8336 -42.6997
HangSeng -1.9343 -45.3721
1% Critical Value -3.9630 -2.5662
5% Critical Value -3.4122 -1.9410
10% Critical Value -3.1280 -1.6166
50

4.5 ANALYSIS OF EAGLE GRANGER TEST
After identifying that all stock indices are integrated of same order, we then estimate the
co-integrating regression using OLS. The residuals, denotes as µt are then tested to
ensure that they are stationary by running Augmented Dickey-Fuller (ADF) test. The
result of the ADF test on the residuals of the regression is presented in Table 4.5.
Schwarz Info Criterion is used to determine the appropriate number of lags. The DW
test statistics from each regression equations are also examined to ensure that there is no
autocorrelation problem.
Since the test statistics are smaller (more negative) than the critical value, null
hypothesis of residuals contain unit root is rejected and conclude that the residuals or
the error term is stationary. This would imply that the two stock indices are co-
integrated. Hence the results indicate that Shanghai and Shenzen Composite Index are
cointegrated with DJIA, FTSE, Nikkei, Hang Seng, Sensex and Nifty. In other words,
China is cointegrated with U.S., U.K., Japan, Hong Kong and India stock market.
Similarly, the results also indicate that Sensex and Nifty Composite Index are
cointegrated with DJIA, FTSE, Nikkei, HangSeng, Shanghai and Shenzen composite
index. In other words, India is also cointegrated with U.S., U.K., Japan, Hong Kong and
China stock market. The finding that stock market indices are cointegrated means that
there is one linear combination between the indices that forces these indices to have a
long-term equilibrium relationship even though the indices may wander away from each
other in the short-run.
51

Table 4.5
Unit Root Test applied to the Residuals of Cointegrating Regressions
Indices Test statistics for residuals
SHCOMP / DJIA -45.3205
SHCOMP / FTSE -45.6862
SHCOMP / NKY -44.6919
SHCOMP / HangSeng -44.3255
SHCOMP / Sensex -45.9762
SHCOMP / Nifty -45.7357
SZCOMP / DJIA -42.5724
SZCOMP / FTSE -42.6642
SZCOMP / NKY -41.7810
SZCOMP / HangSeng -40.8034
SZCOMP / Sensex -42.4048
SZCOMP / Nifty -42.0861
Sensex / DJIA -44.3596
Sensex / FTSE -45.1657
Sensex / NKY -45.0863
Sensex / HangSeng -43.4152
Sensex / SHCOMP -43.4695
Sensex / SZCOMP -42.5959
Nifty / DJIA -45.4221
Nifty / FTSE -46.2349
Nifty / NKY -45.9832
Nifty / HangSeng -43.8789
Nifty / SHCOMP -44.3998
Nifty / SZCOMP -43.4066
1% Critical Value -2.5662
5% Critical Value -1.9410
10% Critical Value -1.6166
52

4.6 ANALYSIS OF ERROR CORRECTION MODEL
As stated in Granger Representation theorem, the relationship between two cointegrated
variables can be expressed as error correction model (ECM) which is useful to to
capture both the short-run dynamic and long-run relationships between the stock
indices. Table 4.6 summarized the F-statistics, coefficient of the lagged value of error
correction term (ECT) and also t-ratio between pairs of stock market indices of China
and India with four major markets. This bivariate cointegration, if exist will reveal the
existence of a long run equilibrium.
The significance of the F-statistics on the lagged value of ECT, suggests that short-run
causality exists between the two cointegrating indices. On the other hand, the long-run
relationship is captured through the significance of the t test of the lagged error
correction term.
Panel A presents the result tested between Shanghai and Shenzen index with other
indices. The significance of the F stat on lagged value of independent stock index Xt-1,
suggests that short-run causality exist between Shanghai and all indices at the 5 per cent
level. There is also evidence of short-run causality between Shenzen and other indices
except DJIA, which means that there is no causal link from DJIA to Shenzen.
53

Table 4.6
Bivariate ECM for Cointegrated Indices
PANEL A
Stock Indices α2 t-stat for µt-1 F
SHCOMP / DJIA -0.047 -2.043* 3.068*
SHCOMP / FTSE -0.054 -2.334* 12.353*
SHCOMP / NKY -0.036 -1.535 51.614*
SHCOMP / HangSeng -1.024 -1.031 189.966*
SHCOMP / Sensex -0.062 -2.667* 93.646*
SHCOMP / Nifty -0.057 -2.440* 105.705*
SZCOMP / DJIA 0.015 0.664 0.431
SZCOMP / FTSE 0.013 0.571 4.593*
SHCOMP / NKY 0.032 1.374 36.134*
SHCOMP / HangSeng 0.060 2.573* 125.535*
SHCOMP / Sensex 0.019 0.835 59.706*
SHCOMP / Nifty 0.027 1.163 68.056*
PANEL B
Stock Indices α2 t-stat for µt-1 F
Sensex / DJIA -0.026 -1.108 48.293*
Sensex / FTSE -0.044 -1.895* 125.877*
Sensex / NKY -0.045 -1.935** 217.661*
Sensex / HangSeng -0.004 -0.154 611.053*
Sensex / SHCOMP -0.006 -0.245 89.781*
Sensex / SZCOMP 0.015 0.636 59.551*
Nifty / DJIA -0.049 -2.132* 44.562*
Nifty / FTSE -0.067 -2.909* 118.953*
Nifty / NKY -0.065 -2.787* 212.886*
Nifty / HangSeng -0.014 -0.611 628.000*
Nifty / SHCOMP -0.027 -1.163 103.151*
Nifty / SZCOMP -0.004 -0.181 67.349*
* indicates significance at the 5% level
** indicates significance at the 10% level
54

However, the significant of t stat on the lagged error correction term µt-1 suggests that
there is long-run relationship between Shanghai and DJIA, FTSE and both Indian stock
indices i.e. Sensex and Nifty at 5 per cent level. However for Shenzen, long-run
relationship is only found with Hang Seng.
The results of F stat and t ratio for Indian market is summarized in Panel B. Reviewing
the results in Panel B, we see that both Sensex and Nifty have short-run causality with
all the four major markets.
However, Sensex is found to have a long run relationship only with FTSE as evidence
by the significance of t stat on lagged error correction term at 5 per cent. Result also
suggests that Sensex has a long run relationship with Nikkei at 10 per cent level. As can
be seen from the result, Nifty is found to have a long run relationship with DJIA, FTSE
and NKY.
Although the ECT is only statistically significant in few equations, we cannot assume
that all other markets are non-causal to both Sensex and Nifty since the short run
channel still active indicated by the significance F stat in all equations of Sensex and
Nifty with other markets.
4.7 SUMMARY
In summary, the correlation analysis suggests some insight on the short-term relations
between China and India with other major markets. In addition, overall results of
Granger causality test suggest that Chinese market has bidirectional causality with
55

Hong Kong market. Another interesting result is the evidence of bidirectional causality
between China and India. We also found that China is cointegrated with U.S., U.K.,
Japan, Hong Kong and India stock market. Similarly, India is also cointegrated with
U.S., U.K., Japan, Hong Kong and China stock market.
56

5.1 CONCLUSION
CHAPTER 5
CONCLUSION AND RECOMMENDATION
This study has tried to investigate the linkage of both Indian and Chinese market with
other major developed markets namely United States (U.S.), United Kingdom (U.K.),
Japan and Hong Kong. Extending related empirical studies, we used correlation test as
preliminary phase of examining the linkages between the markets. We further use co
integration test to comprehensively investigate the direction of relationship.
Based on the analysis of correlation, we found that both China and India are highly
correlated with U.S. This is not surprising as U.S is the world’s foremost stock market
and has large influence on other stock markets. U.S. also is the main trading partner for
both China and India.
We also found that China is highly correlated with Hong Kong. This may be due to the
close proximity between the markets resulting in increase money flows as investors can
easily switch investments between the two markets. Thus stock price indices between
China and Hong Kong tend to correlate. India is also found to have high correlation
with Hong Kong.
Both Chinese and Indian markets are found to be highly correlated. The increasing
economic interdependence between the China and India has contributed to the high
correlation or co-movement between the two markets. Overall, the results from the
57

correlation coefficients suggest some insight on the short-term co-movement between
China and India with other major markets which suggests that the benefits of any short-
term diversification, or speculative activities, are limited between them.
Several interesting observations emerge from the Granger Causality analysis. First, the
findings show that there is bi-directional causality between China and Hong Kong.
Again, this result support the high correlation found between the two markets. China is
also found to have bidirectional causality with U.K. China also found to have
unidirectional causality with Japan where China Granger causes Japan but not vice
versa. Interestingly, different results found between Shanghai and Shenzen with U.S
market. We found that U.S does Granger causes Shanghai but not vice versa. However,
Shenzen found to have bi-directional causality with U.S market.
Second, U.S market is found to Granger causes Indian market but not vice versa.
Whereas, Indian markets is found to Granger cause Japanese market but not vice versa.
There is also short run causality between India and U.K and India and Hong Kong. This
result contradicts with Chattopadhyay and Behera (2006) where they found that US,
UK and HK stock markets Granger cause the India stock market but do not vice versa.
Thirdly, bidirectional causality is found between the Chinese and Indian markets. This
is also further support the high correlation found between the two markets earlier.
However, this finding is slightly differs from Chen, Lobo and Wong (2006) which
found that the Chinese market Granger causes the Indian market but not vice versa.
It was seen that there is evidence of co integration relationship across the markets under
58

study. China is found to be cointegrated with U.S., U.K., Japan, Hong Kong and India
stock market. Similarly, India is also cointegrated with U.S., U.K., Japan, Hong Kong
and China stock market. We also found that the Chinese market is cointegrated with
Indian market.
A subsequent error correction model (ECM) analysis shows that there is short-run
causality between the Chinese markets with all the other indices. However, long run
relationship is only found between China and U.S, U.K, Hong Kong and Indian market.
Similarly, the analysis of ECM suggests that the Indian market have short run causality
with all the four major developed markets. However, long run causality only found
between the Indian market with U.S., U.K. and Japanese market.
In conclusion, both Chinese and Indian market is correlated with all four developed
markets under study namely U.S., U.K., Japan and Hong Kong. This result is further
confirmed with the analysis of Granger causality where the both Chinese and Indian
markets have at least had unilateral causality with all four developed markets.
Another finding is the existence of linkages between Chinese and Indian market. The
relationship can be seen in the increasing bilateral trades between both countries. India
is the largest trading partner of China in South Asia.
This is not to say, however, that Chinese and Indian markets are no longer be beneficial
to investors, as they could switch investments into different emerging markets that has
sufficiently low correlation to developed markets.
59

5.2 RECOMMENDATION
This study has certainly raised several issues for further investigation. In order to get
substantial and better result, there are several recommendations that should be
considered in the future research.
It would be of interest to researcher to extend the study to generate further analysis on
the topics using more comprehensive methodology. The use of GARCH technique in
uncovering the short-run relationship between the stock markets will generate more
meaningful analysis.
Diversification benefit has not been eroded away if closer integration i.e. long term
relationship between two stock markets can be matched by reduction in volatility. Thus,
it would be interesting if this study can be extended to look into the volatility of each
stock market.
Considering structural break in further study would be interesting to address the issue of
how the crisis altered stock markets linkages or interdependencies. Further study may
analyse the stock market linkages before, during and after the recent global financial
crisis that affect almost all markets.
Finally, not that we examine the linkages among stock markets using only index data,
we also bear in mind the importance of macroeconomic variables that may affect the
linkages between markets such as foreign exchange, economic policy, etc. Thus,
incorporating that into the study of market linkages would give insightful results.
60

REFERENCES
Abbas Valadkhani and Surachai Chancharat (2008), “Dynamic linkages between Thai
and international stock markets”, Journal of Economics Studies, 35, pp. 425-441.
Arshanapalli, B. and Doukas, J. (1993), “International Stock Market Linkages:
Evidence from the pre- and post-October 1987 Period”, Journal of Banking and
Finance, 17 (1): 193-208.
Arshanapalli, B., Doukas, J. and Lang, L. (1995), “Pre and post-October1987 Stock
Market Linkages between U.S and Asian markets”, Pacific-Basin Finance Journal, 3
(1): 57-73.
Canarella. G, Miller, S. M, and Pollard, S. K. (2008), “Dynamic Stock Market
Interactions between the Canadian, Mexican, and the United States Markets: The
NAFTA Experience”, Dept. of Economics, Working Paper 2008-49.
Chan, K.C., Gup, B.E. and Pan, M. (1992), “An Empirical Analysis of Stock Prices in
Major Asian Markets and the United States”, Financial Review, 27 (2): 289-307.
Choudhry T, Lu L and Peng Ke (2007), “Common Stochastic Trends among Far East
Stock Prices: Effects of the Asian Financial Crisis”, International Review of Financial
Analysis, 16, 242-261.
Click RW and Plummer MG (2005), “Stock Market Integration in ASEAN after the
Asian Financial Crisis”, Journal of Asian Economics, 16, 5-28.
Cotter, John (2004),”International Equity Market Integration in a Small Open Economy:
Ireland January 1990 – December 2000”, International Review of Financial Analysis,
13 (2004) 669-685.
Dimaranan, B., Ianchovichina, E and Martin, W. (2007), “Competing with Giants: Who
61

Wins, Who Loses?” Chapter 3 in Winters, L. Alan and Yusuf, S. (eds.), Dancing with
Giants: China, India, and the Global Economy, World Bank, Washington DC.
Engle, R.F., and C.W.J. Granger (1987), “Co-Integration, error correction:
Representation, estimation and testing”, Econometrica 55:1251-1276.
Eun, C.S. and S. Shim (1989), “International Transmission of Stock Market
Movements”, Journal of Financial and Quantitative Analysis, vol. 24, pp. 41- 56.
Hamao, Y., R. Masulis, and V. Ng, (1990), “Correlation price changes and volatility
across international stock markets”, Review of Financial Studies, 3, pp. 281-307.
Hilliard, J.E. (1979), “Relationship between Equity Indices on World Exchanges”,
Journal of Finance, 34, pp. 103-104.
Ianchovichina, E. and Martin, W. (2004), “Economic impacts of China’s accession to
the World Trade Organization” World Bank Economic Review 18(1):3-28.
Koch, T.W. and P.D Koch (1993), “Dynamic Relationship among the Daily level of
National Stock Indexes” in International Financial Market Integration. Stansell, S.R.
(ed.), Blackwells, Mass.
Lee, I., R. Pettit and M, Swankoski (1990), “Daily Return Relationships Among Asian
Stock Markets”, Journal of Business Finance and Accounting, pp. 256-284.
Masih, A.M.M. and Masih, R. (1999), “Are Asian Stock Markets Fluctuations Due to
Mainly to Intra-regional Contagion Effects? Evidence Based on Asian Emerging
Markets”, Pacific-Basin Finance Journal, 7 (3): 251-82.
Masih, R. and Masih, A.M.M. (2001), “Long and Short-term Dynamic Causal
Transmission amongst International Stock Markets”, Journal of International Money
and Finance, 20: 563-87.
62

Nath, Golaka C and Sunil Verma. (2003), “Study of Common Stochastic Trend and Co-
Integration in the Emerging Markets: A case Study of India, Singapore and Taiwan”,
Working Paper 72, National Stock Exchange of India. Physica A 377 (2007) 173-180.
Rashid Sabri, Nidal (2002), “Increasing linkages of Stock Markets and Price Volatility”,
Journal of Financial Risk and Financial Risk Management, 16, pp. 349-373.
Roca, E.D. (1999), “Short-term and long-term price linkages between the equity
markets of Australia and its major trading partners”, Applied Financial Economics, 9:
501 -511.
Sharma, J.L and R.E. Kennedy (1977), “Comparative Analysis of Stock Price
Behaviour on the Bombay, London & New York Stock Exchanges”, JFQA, September
1977, pp. 391-403.
Siklos, P.L, and Ng, P. (2001), “Integration among Asia Pacific and International Stock
Markets: Common Stochastic Trends and Regime Shifts”, Pacific Economic Review, 6:
89-110.
Solnik, B, C Boucrelle and Y Le Fur (1996), “International market correlation and
volatility”, Financial Analysts Journal, vol 52, no 5, pp 17–34.
Tan, K.B., and Y.K. Tse. (2002), “The Integration of the East and South-East Asian
Equity Markets”, International Center for the Study of East Asian Development
Working Paper No. 2002-11.
Taylor, M.P. and Tonks, I. (1989), “The internationalization of stock markets and
abolition of UK exchange control”, The Review of Economics and Statistics 71: 332 -
336.
Toda, H.Y. and P.C.B. Phillips (1994), “Vector Autoregressions and Causality: A
Theoretical Overview and Simulation Study", Econometric Reviews 13, 259-285
Wong. W, A. Agarwal and J. Du (2005), “Financial Integration for Indian Stock
63

Market, A Fractional Cointegration Approach”, National University of Singapore, Dept.
of Economics , Working Paper No. 050.
64

APPENDICES
65

Appendix 1: Result of Granger Causality Test
Pairwise Granger Causality Tests
Date: 04/13/09 Time: 13:55
Sample: 1 1872
Lags: 7
Null Hypothesis: Obs F-Statistic Prob.
DDJIA does not Granger Cause DSHCOMP 1864 5.09111 1.E-05
DSHCOMP does not Granger Cause DDJIA 2.95922 0.0044
Null Hypothesis: Obs F-Statistic Prob.
DFTSE does not Granger Cause DSHCOMP 1864 5.64867 2.E-06
DSHCOMP does not Granger Cause DFTSE 1.89854 0.0660
Null Hypothesis: Obs F-Statistic Prob.
DNKY does not Granger Cause DSHCOMP 1841 1.11254 0.3522
DSHCOMP does not Granger Cause DNKY 2.02773 0.0484
Null Hypothesis: Obs F-Statistic Prob.
DHSI does not Granger Cause DSHCOMP 1864 4.27039 0.0001
DSHCOMP does not Granger Cause DHSI 5.89672 9.E-07
Null Hypothesis: Obs F-Statistic Prob.
DSENSEX does not Granger Cause DSHCOMP 1864 4.87225 2.E-05
DSHCOMP does not Granger Cause DSENSEX 2.68233 0.0092
Null Hypothesis: Obs F-Statistic Prob.
DNIFTY does not Granger Cause DSHCOMP 1864 4.13525 0.0002
DSHCOMP does not Granger Cause DNIFTY 2.54492 0.0131
Null Hypothesis: Obs F-Statistic Prob.
DDJIA does not Granger Cause DSZCOMP 1864 2.90751 0.0050
DSZCOMP does not Granger Cause DDJIA 2.70423 0.0086
Null Hypothesis: Obs F-Statistic Prob.
DFTSE does not Granger Cause DSZCOMP 1864 3.68395 0.0006
DSZCOMP does not Granger Cause DFTSE 1.25009 0.2717
66

Pairwise Granger Causality Tests
Date: 04/13/09 Time: 14:59
Sample: 1 1872
Lags: 7
Null Hypothesis: Obs F-Statistic Prob.
DNKY does not Granger Cause DSZCOMP 1841 1.06075 0.3865
DSZCOMP does not Granger Cause DNKY 2.14653 0.0362
Null Hypothesis: Obs F-Statistic Prob.
DHSI does not Granger Cause DSZCOMP 1864 2.63579 0.0104
DSZCOMP does not Granger Cause DHSI 7.31740 1.E-08
Null Hypothesis: Obs F-Statistic Prob.
DSENSEX does not Granger Cause DSZCOMP 1864 3.32791 0.0016
DSZCOMP does not Granger Cause DSENSEX 3.28934 0.0018
Null Hypothesis: Obs F-Statistic Prob.
DNIFTY does not Granger Cause DSZCOMP 1864 2.72914 0.0081
DSZCOMP does not Granger Cause DNIFTY 3.54487 0.0009
Null Hypothesis: Obs F-Statistic Prob.
DDJIA does not Granger Cause DSENSEX 1864 13.7304 2.E-17
DSENSEX does not Granger Cause DDJIA 1.39154 0.2045
Null Hypothesis: Obs F-Statistic Prob.
DFTSE does not Granger Cause DSENSEX 1864 7.70119 3.E-09
DSENSEX does not Granger Cause DFTSE 2.27267 0.0264
Null Hypothesis: Obs F-Statistic Prob.
DNKY does not Granger Cause DSENSEX 1841 0.65799 0.7079
DSENSEX does not Granger Cause DNKY 4.22408 0.0001
Null Hypothesis: Obs F-Statistic Prob.
DHSI does not Granger Cause DSENSEX 1864 3.46018 0.0011
DSENSEX does not Granger Cause DHSI 2.21024 0.0309
67

Pairwise Granger Causality Tests
Date: 04/13/09 Time: 15:13
Sample: 1 1872
Lags: 7
Null Hypothesis: Obs F-Statistic Prob.
DSHCOMP does not Granger Cause DSENSEX 1864 2.68233 0.0092
DSENSEX does not Granger Cause DSHCOMP 4.87225 2.E-05
Null Hypothesis: Obs F-Statistic Prob.
DSZCOMP does not Granger Cause DSENSEX 1864 3.28934 0.0018
DSENSEX does not Granger Cause DSZCOMP 3.32791 0.0016
Null Hypothesis: Obs F-Statistic Prob.
DDJIA does not Granger Cause DNIFTY 1864 13.4686 5.E-17
DNIFTY does not Granger Cause DDJIA 1.71643 0.1008
Null Hypothesis: Obs F-Statistic Prob.
DFTSE does not Granger Cause DNIFTY 1864 7.59207 5.E-09
DNIFTY does not Granger Cause DFTSE 2.38952 0.0196
Null Hypothesis: Obs F-Statistic Prob.
DNKY does not Granger Cause DNIFTY 1841 0.66979 0.6979
DNIFTY does not Granger Cause DNKY 3.43328 0.0012
Null Hypothesis: Obs F-Statistic Prob.
DHSI does not Granger Cause DNIFTY 1864 4.03233 0.0002
DNIFTY does not Granger Cause DHSI 1.97390 0.0551
Null Hypothesis: Obs F-Statistic Prob.
DSHCOMP does not Granger Cause DNIFTY 1864 2.54492 0.0131
DNIFTY does not Granger Cause DSHCOMP 4.13525 0.0002
Null Hypothesis: Obs F-Statistic Prob.
DSZCOMP does not Granger Cause DNIFTY 1864 3.54487 0.0009
DNIFTY does not Granger Cause DSZCOMP 2.72914 0.0081
68

Appendix 2: Result of Unit Root Test
Null Hypothesis: D(DJIA) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -45.67341 0.0001
Test critical values: 1% level -2.566199
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(DJIA,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(DJIA(-1)) -1.055136 0.023102 -45.67341 0.0000
R-squared 0.527441 Mean dependent var 0.031947
Adjusted R-squared 0.527441 S.D. dependent var 202.4986
S.E. of regression 139.2035 Akaike info criterion 12.71029
Sum squared resid 36216759 Schwarz criterion 12.71324
Log likelihood -11883.12 Hannan-Quinn criter. 12.71138
Durbin-Watson stat 2.002287
Null Hypothesis: D(FTSE) has a unit root
Exogenous: None
Lag Length: 1 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -33.95246 0.0000
Test critical values: 1% level -2.566200
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(FTSE,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(FTSE(-1)) -1.139396 0.033559 -33.95246 0.0000
D(FTSE(-1),2) 0.076408 0.023065 3.312728 0.0009
R-squared 0.532086 Mean dependent var 0.086693
Adjusted R-squared 0.531835 S.D. dependent var 105.9158
S.E. of regression 72.47030 Akaike info criterion 11.40530
Sum squared resid 9805380. Schwarz criterion 11.41122
Log likelihood -10656.25 Hannan-Quinn criter. 11.40748
Durbin-Watson stat 2.001974
69

Null Hypothesis: D(NKY) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -42.69974 0.0001
Test critical values: 1% level -2.566205
5% level -1.940994
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(NKY,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(NKY(-1)) -0.989099 0.023164 -42.69974 0.0000
R-squared 0.495013 Mean dependent var 0.269387
Adjusted R-squared 0.495013 S.D. dependent var 304.1730
S.E. of regression 216.1527 Akaike info criterion 13.59038
Sum squared resid 86902870 Schwarz criterion 13.59336
Log likelihood -12644.85 Hannan-Quinn criter. 13.59148
Durbin-Watson stat 2.001538
Null Hypothesis: D(HSI) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -45.37206 0.0001
Test critical values: 1% level -2.566199
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(HSI,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(HSI(-1)) -1.044187 0.023014 -45.37206 0.0000
R-squared 0.524139 Mean dependent var 0.605947
Adjusted R-squared 0.524139 S.D. dependent var 451.7115
S.E. of regression 311.6028 Akaike info criterion 14.32187
Sum squared resid 1.81E+08 Schwarz criterion 14.32483
Log likelihood -13389.95 Hannan-Quinn criter. 14.32296
Durbin-Watson stat 2.009205
70

Null Hypothesis: D(SHCOMP) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -45.20383 0.0001
Test critical values: 1% level -2.566199
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(SHCOMP,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(SHCOMP(-1)) -1.044601 0.023109 -45.20383 0.0000
R-squared 0.522286 Mean dependent var -0.011166
Adjusted R-squared 0.522286 S.D. dependent var 76.25882
S.E. of regression 52.70767 Akaike info criterion 10.76793
Sum squared resid 5192267. Schwarz criterion 10.77089
Log likelihood -10067.02 Hannan-Quinn criter. 10.76902
Durbin-Watson stat 1.996243
Null Hypothesis: D(SZCOMP) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -42.54944 0.0001
Test critical values: 1% level -2.566199
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(SZCOMP,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(SZCOMP(-1)) -0.984111 0.023129 -42.54944 0.0000
R-squared 0.492044 Mean dependent var -0.003214
Adjusted R-squared 0.492044 S.D. dependent var 22.15170
S.E. of regression 15.78774 Akaike info criterion 8.356879
Sum squared resid 465853.5 Schwarz criterion 8.359839
Log likelihood -7812.682 Hannan-Quinn criter. 8.357970
Durbin-Watson stat 2.000190
71

Null Hypothesis: D(SENSEX) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -42.73641 0.0001
Test critical values: 1% level -2.566199
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(SENSEX,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(SENSEX(-1)) -0.988576 0.023132 -42.73641 0.0000
R-squared 0.494236 Mean dependent var 0.169332
Adjusted R-squared 0.494236 S.D. dependent var 280.5015
S.E. of regression 199.4845 Akaike info criterion 13.42988
Sum squared resid 74375104 Schwarz criterion 13.43284
Log likelihood -12555.94 Hannan-Quinn criter. 13.43097
Durbin-Watson stat 1.999641
Null Hypothesis: D(NIFTY) has a unit root
Exogenous: None
Lag Length: 0 (Automatic based on SIC, MAXLAG=24)
t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -43.87881 0.0001
Test critical values: 1% level -2.566199
5% level -1.940993
10% level -1.616585
*MacKinnon (1996) one-sided p-values.
Augmented Dickey-Fuller Test Equation
Dependent Variable: D(NIFTY,2)
Method: Least Squares
Variable Coefficient Std. Error t-Statistic Prob.
D(NIFTY(-1)) -1.014957 0.023131 -43.87881 0.0000
R-squared 0.507425 Mean dependent var 0.053583
Adjusted R-squared 0.507425 S.D. dependent var 86.11787
S.E. of regression 60.44067 Akaike info criterion 11.04174
Sum squared resid 6827597. Schwarz criterion 11.04470
Log likelihood -10323.02 Hannan-Quinn criter. 11.04283
Durbin-Watson stat 1.998322
72

Appendix 3: Result of Error Correction Model
Dependent Variable: DSHCOMP
Method: Least Squares
Date: 04/12/09 Time: 18:59
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.241929 1.218767 0.198503 0.8427
DDJIA 0.011501 0.008752 1.314139 0.1890
UHAT(-1) -0.047268 0.023139 -2.042797 0.0412
R-squared 0.003276 Mean dependent var 0.226326
Adjusted R-squared 0.002209 S.D. dependent var 52.75969
S.E. of regression 52.70139 Akaike info criterion 10.76876
Sum squared resid 5185474. Schwarz criterion 10.77764
Log likelihood -10065.79 Hannan-Quinn criter. 10.77203
F-statistic 3.068653 Durbin-Watson stat 1.995559
Prob(F-statistic) 0.046718
Dependent Variable: DSHCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:31
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.309976 1.212867 0.255573 0.7983
DFTSE 0.072401 0.016668 4.343651 0.0000
UHAT(-1) -0.053956 0.023115 -2.334258 0.0197
R-squared 0.013061 Mean dependent var 0.226326
Adjusted R-squared 0.012003 S.D. dependent var 52.75969
S.E. of regression 52.44209 Akaike info criterion 10.75890
Sum squared resid 5134572. Schwarz criterion 10.76778
Log likelihood -10056.57 Hannan-Quinn criter. 10.76217
F-statistic 12.35336 Durbin-Watson stat 1.994775
Prob(F-statistic) 0.000005
73

Dependent Variable: DSHCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:34
Sample (adjusted): 3 1872
Included observations: 1861 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.516469 1.194059 0.432532 0.6654
DNKY 0.055175 0.005529 9.978958 0.0000
UHAT(-1) -0.035615 0.023200 -1.535107 0.1249
R-squared 0.052634 Mean dependent var 0.224610
Adjusted R-squared 0.051614 S.D. dependent var 52.87817
S.E. of regression 51.49546 Akaike info criterion 10.72247
Sum squared resid 4927011. Schwarz criterion 10.73139
Log likelihood -9974.263 Hannan-Quinn criter. 10.72576
F-statistic 51.61362 Durbin-Watson stat 1.999762
Prob(F-statistic) 0.000000
Dependent Variable: DSHCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:38
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.285966 1.112738 0.256993 0.7972
DHSI 0.069048 0.003591 19.23011 0.0000
UHAT(-1) -0.023981 0.023260 -1.030995 0.3027
R-squared 0.169089 Mean dependent var 0.226326
Adjusted R-squared 0.168199 S.D. dependent var 52.75969
S.E. of regression 48.11848 Akaike info criterion 10.58681
Sum squared resid 4322830. Schwarz criterion 10.59569
Log likelihood -9895.670 Hannan-Quinn criter. 10.59008
F-statistic 189.9660 Durbin-Watson stat 2.000000
Prob(F-statistic) 0.000000
74

Dependent Variable: DSHCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:40
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.046381 1.163817 0.039853 0.9682
DSENSEX 0.077645 0.005841 13.29320 0.0000
UHAT(-1) -0.061669 0.023124 -2.666890 0.0077
R-squared 0.091171 Mean dependent var 0.226326
Adjusted R-squared 0.090198 S.D. dependent var 52.75969
S.E. of regression 50.32406 Akaike info criterion 10.67645
Sum squared resid 4728199. Schwarz criterion 10.68532
Log likelihood -9979.478 Hannan-Quinn criter. 10.67972
F-statistic 93.64648 Durbin-Watson stat 1.995804
Prob(F-statistic) 0.000000
Dependent Variable: DSHCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:43
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.025892 1.157053 0.022377 0.9821
DNIFTY 0.272236 0.019168 14.20272 0.0000
UHAT(-1) -0.056455 0.023135 -2.440277 0.0148
R-squared 0.101717 Mean dependent var 0.226326
Adjusted R-squared 0.100755 S.D. dependent var 52.75969
S.E. of regression 50.03125 Akaike info criterion 10.66478
Sum squared resid 4673335. Schwarz criterion 10.67365
Log likelihood -9968.565 Hannan-Quinn criter. 10.66805
F-statistic 105.7048 Durbin-Watson stat 1.996519
Prob(F-statistic) 0.000000
75

Dependent Variable: DSZCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:48
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.079564 0.365258 0.217831 0.8276
DDJIA 0.001791 0.002624 0.682385 0.4951
UHAT(-1) 0.015380 0.023171 0.663771 0.5069
R-squared 0.000462 Mean dependent var 0.077257
Adjusted R-squared -0.000609 S.D. dependent var 15.78955
S.E. of regression 15.79435 Akaike info criterion 8.358785
Sum squared resid 465744.8 Schwarz criterion 8.367663
Log likelihood -7812.464 Hannan-Quinn criter. 8.362056
F-statistic 0.431400 Durbin-Watson stat 2.000286
Prob(F-statistic) 0.649664
Dependent Variable: DSZCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:52
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.094441 0.364476 0.259113 0.7956
DFTSE 0.015030 0.005013 2.998394 0.0027
UHAT(-1) 0.013224 0.023164 0.570891 0.5681
R-squared 0.004896 Mean dependent var 0.077257
Adjusted R-squared 0.003830 S.D. dependent var 15.78955
S.E. of regression 15.75928 Akaike info criterion 8.354339
Sum squared resid 463678.6 Schwarz criterion 8.363217
Log likelihood -7808.307 Hannan-Quinn criter. 8.357610
F-statistic 4.592924 Durbin-Watson stat 2.000309
Prob(F-statistic) 0.010238
76

Dependent Variable: DSZCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:54
Sample (adjusted): 3 1872
Included observations: 1861 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.153599 0.360198 0.426428 0.6698
DNKY 0.014137 0.001671 8.462610 0.0000
UHAT(-1) 0.031920 0.023239 1.373521 0.1698
R-squared 0.037439 Mean dependent var 0.079049
Adjusted R-squared 0.036403 S.D. dependent var 15.82475
S.E. of regression 15.53405 Akaike info criterion 8.325556
Sum squared resid 448347.7 Schwarz criterion 8.334469
Log likelihood -7743.930 Hannan-Quinn criter. 8.328841
F-statistic 36.13383 Durbin-Watson stat 2.003984
Prob(F-statistic) 0.000000
Dependent Variable: DSZCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:56
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.092327 0.342993 0.269180 0.7878
DHSI 0.017612 0.001112 15.84274 0.0000
UHAT(-1) 0.060041 0.023335 2.572969 0.0102
R-squared 0.118538 Mean dependent var 0.077257
Adjusted R-squared 0.117593 S.D. dependent var 15.78955
S.E. of regression 14.83215 Akaike info criterion 8.233074
Sum squared resid 410726.2 Schwarz criterion 8.241952
Log likelihood -7694.924 Hannan-Quinn criter. 8.236345
F-statistic 125.5355 Durbin-Watson stat 2.004895
Prob(F-statistic) 0.000000
77

Dependent Variable: DSZCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:58
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.031799 0.354200 0.089777 0.9285
DSENSEX 0.019464 0.001781 10.92764 0.0000
UHAT(-1) 0.019373 0.023208 0.834724 0.4040
R-squared 0.060115 Mean dependent var 0.077257
Adjusted R-squared 0.059108 S.D. dependent var 15.78955
S.E. of regression 15.31579 Akaike info criterion 8.297249
Sum squared resid 437948.8 Schwarz criterion 8.306127
Log likelihood -7754.928 Hannan-Quinn criter. 8.300520
F-statistic 59.70668 Durbin-Watson stat 2.000400
Prob(F-statistic) 0.000000
Dependent Variable: DSZCOMP
Method: Least Squares
Date: 04/12/09 Time: 20:59
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.026558 0.352723 0.075295 0.9400
DNIFTY 0.068347 0.005859 11.66574 0.0000
UHAT(-1) 0.027001 0.023224 1.162646 0.2451
R-squared 0.067951 Mean dependent var 0.077257
Adjusted R-squared 0.066952 S.D. dependent var 15.78955
S.E. of regression 15.25182 Akaike info criterion 8.288878
Sum squared resid 434297.8 Schwarz criterion 8.297755
Log likelihood -7747.101 Hannan-Quinn criter. 8.292148
F-statistic 68.05628 Durbin-Watson stat 2.000651
Prob(F-statistic) 0.000000
78

Dependent Variable: DSENSEX
Method: Least Squares
Date: 04/12/09 Time: 21:01
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 2.742296 4.500761 0.609296 0.5424
DDJIA 0.314579 0.032299 9.739623 0.0000
UHAT(-1) -0.025633 0.023138 -1.107806 0.2681
R-squared 0.049189 Mean dependent var 2.331102
Adjusted R-squared 0.048170 S.D. dependent var 199.4839
S.E. of regression 194.6200 Akaike info criterion 13.38158
Sum squared resid 70716237 Schwarz criterion 13.39046
Log likelihood -12508.78 Hannan-Quinn criter. 13.38485
F-statistic 48.29339 Durbin-Watson stat 2.000027
Prob(F-statistic) 0.000000
Dependent Variable: DSENSEX
Method: Least Squares
Date: 04/12/09 Time: 21:02
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 3.406072 4.333164 0.786047 0.4319
DFTSE 0.940191 0.059551 15.78797 0.0000
UHAT(-1) -0.043838 0.023128 -1.895469 0.0582
R-squared 0.118822 Mean dependent var 2.331102
Adjusted R-squared 0.117878 S.D. dependent var 199.4839
S.E. of regression 187.3579 Akaike info criterion 13.30552
Sum squared resid 65537286 Schwarz criterion 13.31440
Log likelihood -12437.66 Hannan-Quinn criter. 13.30879
F-statistic 125.8776 Durbin-Watson stat 2.000157
Prob(F-statistic) 0.000000
79

Dependent Variable: DSENSEX
Method: Least Squares
Date: 04/12/09 Time: 21:04
Sample (adjusted): 3 1872
Included observations: 1861 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 4.484725 4.175667 1.074014 0.2830
DNKY 0.405624 0.019445 20.86040 0.0000
UHAT(-1) -0.045133 0.023326 -1.934865 0.0532
R-squared 0.189821 Mean dependent var 2.346539
Adjusted R-squared 0.188949 S.D. dependent var 199.9605
S.E. of regression 180.0811 Akaike info criterion 13.22630
Sum squared resid 60253481 Schwarz criterion 13.23522
Log likelihood -12304.07 Hannan-Quinn criter. 13.22959
F-statistic 217.6606 Durbin-Watson stat 2.000942
Prob(F-statistic) 0.000000
Dependent Variable: DSENSEX
Method: Least Squares
Date: 04/12/09 Time: 21:05
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 2.677349 3.588203 0.746153 0.4557
DHSI 0.402338 0.011516 34.93595 0.0000
UHAT(-1) -0.003566 0.023142 -0.154082 0.8776
R-squared 0.395618 Mean dependent var 2.331102
Adjusted R-squared 0.394970 S.D. dependent var 199.4839
S.E. of regression 155.1658 Akaike info criterion 12.92847
Sum squared resid 44950709 Schwarz criterion 12.93735
Log likelihood -12085.12 Hannan-Quinn criter. 12.93174
F-statistic 611.0526 Durbin-Watson stat 2.002297
Prob(F-statistic) 0.000000
80

Dependent Variable: DSENSEX
Method: Least Squares
Date: 04/12/09 Time: 21:07
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 2.076601 4.408424 0.471053 0.6377
DSHCOMP 1.121792 0.083986 13.35682 0.0000
UHAT(-1) -0.005689 0.023260 -0.244586 0.8068
R-squared 0.087739 Mean dependent var 2.331102
Adjusted R-squared 0.086761 S.D. dependent var 199.4839
S.E. of regression 190.6338 Akaike info criterion 13.34019
Sum squared resid 67849124 Schwarz criterion 13.34907
Log likelihood -12470.08 Hannan-Quinn criter. 13.34346
F-statistic 89.78122 Durbin-Watson stat 1.999265
Prob(F-statistic) 0.000000
Dependent Variable: DSENSEX
Method: Least Squares
Date: 04/12/09 Time: 21:08
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 2.094906 4.475033 0.468132 0.6397
DSZCOMP 3.077211 0.284053 10.83322 0.0000
UHAT(-1) 0.014744 0.023190 0.635800 0.5250
R-squared 0.059968 Mean dependent var 2.331102
Adjusted R-squared 0.058961 S.D. dependent var 199.4839
S.E. of regression 193.5137 Akaike info criterion 13.37018
Sum squared resid 69914562 Schwarz criterion 13.37905
Log likelihood -12498.11 Hannan-Quinn criter. 13.37345
F-statistic 59.55112 Durbin-Watson stat 1.999819
Prob(F-statistic) 0.000000
81

Dependent Variable: DNIFTY
Method: Least Squares
Date: 04/12/09 Time: 21:10
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.856726 1.366313 0.627035 0.5307
DDJIA 0.089780 0.009804 9.157011 0.0000
UHAT(-1) -0.049278 0.023117 -2.131731 0.0332
R-squared 0.045562 Mean dependent var 0.739947
Adjusted R-squared 0.044539 S.D. dependent var 60.44290
S.E. of regression 59.08153 Akaike info criterion 10.99732
Sum squared resid 6517001. Schwarz criterion 11.00619
Log likelihood -10279.49 Hannan-Quinn criter. 11.00059
F-statistic 44.56213 Durbin-Watson stat 1.998830
Prob(F-statistic) 0.000000
Dependent Variable: DNIFTY
Method: Least Squares
Date: 04/12/09 Time: 21:12
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 1.054125 1.317246 0.800249 0.4237
DFTSE 0.275259 0.018103 15.20487 0.0000
UHAT(-1) -0.067185 0.023099 -2.908612 0.0037
R-squared 0.113024 Mean dependent var 0.739947
Adjusted R-squared 0.112074 S.D. dependent var 60.44290
S.E. of regression 56.95524 Akaike info criterion 10.92401
Sum squared resid 6056361. Schwarz criterion 10.93289
Log likelihood -10210.95 Hannan-Quinn criter. 10.92728
F-statistic 118.9525 Durbin-Watson stat 1.998875
Prob(F-statistic) 0.000000
82

Dependent Variable: DNIFTY
Method: Least Squares
Date: 04/12/09 Time: 21:13
Sample (adjusted): 3 1872
Included observations: 1861 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 1.383163 1.267851 1.090951 0.2754
DNKY 0.121598 0.005897 20.62171 0.0000
UHAT(-1) -0.064854 0.023272 -2.786778 0.0054
R-squared 0.186434 Mean dependent var 0.742826
Adjusted R-squared 0.185558 S.D. dependent var 60.58718
S.E. of regression 54.67777 Akaike info criterion 10.84240
Sum squared resid 5554786. Schwarz criterion 10.85132
Log likelihood -10085.86 Hannan-Quinn criter. 10.84569
F-statistic 212.8859 Durbin-Watson stat 1.999403
Prob(F-statistic) 0.000000
Dependent Variable: DNIFTY
Method: Least Squares
Date: 04/12/09 Time: 21:15
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.845340 1.081297 0.781783 0.4344
DHSI 0.122915 0.003468 35.44006 0.0000
UHAT(-1) -0.014138 0.023126 -0.611345 0.5410
R-squared 0.402178 Mean dependent var 0.739947
Adjusted R-squared 0.401537 S.D. dependent var 60.44290
S.E. of regression 46.75885 Akaike info criterion 10.52949
Sum squared resid 4081991. Schwarz criterion 10.53836
Log likelihood -9842.071 Hannan-Quinn criter. 10.53276
F-statistic 628.0008 Durbin-Watson stat 2.002731
Prob(F-statistic) 0.000000
83

Dependent Variable: DNIFTY
Method: Least Squares
Date: 04/12/09 Time: 21:17
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.656968 1.327095 0.495042 0.6206
DSHCOMP 0.362615 0.025246 14.36319 0.0000
UHAT(-1) -0.026996 0.023218 -1.162722 0.2451
R-squared 0.099504 Mean dependent var 0.739947
Adjusted R-squared 0.098539 S.D. dependent var 60.44290
S.E. of regression 57.38769 Akaike info criterion 10.93914
Sum squared resid 6148678. Schwarz criterion 10.94802
Log likelihood -10225.10 Hannan-Quinn criter. 10.94241
F-statistic 103.1508 Durbin-Watson stat 1.998251
Prob(F-statistic) 0.000000
Dependent Variable: DNIFTY
Method: Least Squares
Date: 04/12/09 Time: 21:19
Sample (adjusted): 3 1872
Included observations: 1870 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.663046 1.350626 0.490918 0.6235
DSZCOMP 0.993610 0.085652 11.60057 0.0000
UHAT(-1) -0.004182 0.023171 -0.180502 0.8568
R-squared 0.067292 Mean dependent var 0.739947
Adjusted R-squared 0.066293 S.D. dependent var 60.44290
S.E. of regression 58.40508 Akaike info criterion 10.97429
Sum squared resid 6368624. Schwarz criterion 10.98316
Log likelihood -10257.96 Hannan-Quinn criter. 10.97756
F-statistic 67.34913 Durbin-Watson stat 1.999138
Prob(F-statistic) 0.000000
84