AN ANALYSIS OF STOCK MARKET INTEGRATION IN THE ASIAN DEVELOPED AND EMERGING MARKETS

Purpose of the study: This study investigates Short-run, Long-run, and Casual relationships in the Asian Developed and Emerging stock market indices for the period of 19 years weekly data of stock market indices of Asian Developed and Emerging Markets which are Japan (Nikkei 225), South Korea (KOSPI), Pakistan (KSE 100), China (SSE Composite), Sri Lanka (ASPI), India (BSE 200) and Malaysia (KLSE composite) from January 2001 to December 2019. Methodology: To analyze long-run and short-run relationships among the Asian developed and emerging stock markets, this study practices Descriptive Statistics, Correlation Matrix, Unit Root Test, Johansen Co-Integration Test, Vector Error Correction Model, Granger Causality test, Variance Decomposition and Impulse Response Function (IRF). Main findings: By employing the ADF and P.P. tests, the results specify that the entire variables' data are non-stationary and stationary in exact order, which is 1 st difference. The Johnson Co-integration test found one cointegration relationship, where the results are consistent with Granger causality, Variance Decomposition, and Impulse Response Function (IRF). Application of the study: As the current research has focused on finding out the comovements in the Asian developed and emerging markets. So, the applications are that the survey found short-run and long-run relationships in these countries' stock markets. The study's originality: The current study has selected seven Asian developed and emerging stock markets and weekly updated time series data to investigate short-term and long-term linkages. So, this study found long-run comovements in these stock indices, which contributes to the literature. In addition, these stock markets have limited diversification benefits for international investors, while short-term diversification benefits may exist.


INTRODUCTION
The liberalization globally leads the international markets to integration among them, Rodriguez and Rodrik (2011). Researchers defined the market integration that the linkages or comovements existing in all over the world markets economies are integration in those economies, Forbes and Rigobon (2012). Further research studies indicated that when different markets are integrated, how much risk these markets will have there would have the same level of return (Errunza and Losq (1985). Researchers examined that international financial integration encourages economic growth, which concludes that the research does not support the logic that international financial integration stimulated economic growth (Edison et al. (, 2012). Due to increased integration in world economies, it is perceived rapid increase in global capital mobility from direct or indirect investment (Chen et al. (2016). Financial integration is the cause of abolishing capital flows, foreign exchange transaction, technological advancement, flows of information, and the financial transaction became more accessible among the international countries (Arshanapalli and Doukas, 1993; Khan, Khan, Ullah, Usman, Farhat, 2020). Chan et al. (1997) scrutinized the stock markets integration internationally, examined relationships among eighteen nations, and found no long-run cointegration among the countries, showing diversification benefits. Chen et al. (2012) investigated the behaviour of stock prices in six major Latin American stock exchanges 1 . The researchers found a long-run relationship, which concluded that diversification was of fewer benefits.
This study is also based on this approach, which examines the Long-run, Short-run, and Casual relationship in the Asian developed and emerging stock markets. Asian developed markets are Japan and South Korea, and Asian emerging markets are Pakistan, China, Sri Lanka, India, and Malaysia under World Bank criteria 2  The study is based on time series data of Asian Developed and Emerging stock markets. The sample size is based on 19 years of weekly data of Asian developed and emerging markets. The needs investigated are China (SSE Composite), India (BSE 200), Japan (Nikkei 225), Malaysia (KLSE composite), Pakistan (KSE 100), South Korea (KOSPI), and Sri Lanka (ASPI) from 2001 to 2019. The study is based on secondary data. The sample consists of the Asian developed and emerging countries for which sufficiently long series of weekly data is available. The information has been abstracted from various sources 3 .

Variables of the Study
The relationship among the developed and emerging stock markets have been analyzed by different researchers such as Chan et al. (1997), Chen et al. (2012), Gilmore and McManus (2002), and Click and Plummer (2015), etc. These countries selected variables both from emerging and developed markets. All of them have used different models and found different results. Due to the particular emphasis on the Asian region, the current study investigates the long-run, short-run, and causal relationship among the Asian developed and emerging stock markets. The present study investigates the association among two Asian developed stock markets of Japan and South Korea and five emerging stock markets of Pakistan, China, Sri Lanka, India, and Malaysia. The study analyzes the long-run, short-run, and causal relationship among the stock as mentioned above market indices of Japan (Nikkei 225), South Korea (KOSPI), Malaysia (KLSE), Pakistan (KSE 100), China (SSE), Sri Lanka (CSE), and India (BSE).

Estimation Tools
To investigate the relationships in the Asian developed and emerging stock markets, the current study employed descriptive statistics, correlation matrix, Unit Root Tests, Johansen Co-Integration Test, Vector Error Correction Model (VECM), Granger Causality test, Variance Decomposition, and Impulse Response Function (IRF). This methodology is followed by various existing studies such as Canarella et al. (2018), Gurcharan and Pritam (2010).

Descriptive Statistics
The Descriptive Statistics brought to scrutinize the behaviour of the data using for further analysis. The descriptive Statistics comprises mean, median, maximum, minimum, standard deviation, skewness, and kurtosis. The average return of each stock market and standard deviation measure how much risk is in each stock market of the study. Skewness measures that the data is either positively or negatively skewed, which tells the symmetry of the data. Kurtosis shows that the information is either normally distributed or not.

Correlation Matrix
The Correlation Matrix is used to examine the association in the variables. This is also used to identify that the variables are positively or negatively correlated (Canarella et al. (, 2018). The drawback of this tool is it is only used to examine the relationship between the variables but cannot investigate that the variables have short-run or long-run relationships. For example, the correlation between the two stock markets may be positive or negative. When the two variables are highly correlated, the co-efficient value maybe +1 or -1. The positive relationship shows, when one variable return increases, the other return will also increase, and the antagonistic relationship shows that one variable return increases, the other will decrease. The co-efficient value near zero implies no ties.

Unit Root Test
The Unit Root Tests are used to investigate that the data is either stationary or non-stationary. This is important because if the information is static in the same order, cointegration tests will be employed. The cointegration assumption is that the data must be stationary at the same level. This study practices Augmented Dicky Fuller (ADF) and Philips Peron Test (P.P.) to identify the fact that the data is stationary in the same order. The ADF assumption is that the error term variance is constant, and the error term is the independent variable. P.P. is a generalized form of ADF. The models for ADF and P.P. are as under, The series is non-stationary is the null hypothesis of both the ADF test and P.P. test.

Johansen Co-Integration Test
To examine the long-run relationship among the Asian developed and emerging stock markets, this study practices the Johansen cointegration test (1988  Is the Metrix used for the long-run relationship and Is the matrix of the long-run coefficients that represents up to g-1 co-integrating relationship.

Granger Causality Test
The Granger Causality Test is employed to examine the causal relationships between the variables of the research study. Where = stationary variables = uncorrelated white noises series

Variance Decomposition
The Variance Decomposition is employed to examine how many variances in a stock market are due to its variation and how many disagreements are due to the innovation of other developed and emerging stock markets counterparts (Moon (2001). It demonstrates the Variance Decomposition for the individual stock market of advanced and emerging economies of the research study. The research study used Variance Decomposition analysis to examine the dynamic changes in the stock markets of the study, which identify that either the dynamic changes in the stock markets are due to their innovations or due to other stock markets.

Impulse Response Function
The impulse response function is applied to check out the response of the dependent variable to the dynamic changes of each variable under study. Thus, there is an equation separately for each variable. The effects over time can watch out under the impulse response function. If there are "s" variables, so it could be generated." " impulse responses. To analyze the effects of shocks, we first estimate VAR. So VAR can be written as:
When one standard deviation shock is bringing in the variable on the period "t", then identifies the consequences of this innovation. Table 1 summarizes several descriptive statistics for the countries' stock markets under investigation. In the table, the Mean comprises the average weekly return, maximum and minimum indicates maximum and minimum weekly return, standard deviation specifies the risk, and the skewness suggest that the data is either positively or negatively skewed. Therefore, the countries stock markets having higher average weekly return and the low standard deviation is fruitful for investments. Here in this table, Sri Lanka (CSE) displays the highest maximum return of 17.9% and a low risk of 2.8%. Secondly, South Korea (KOSPI) has a high maximum return of 17%, and risk is 3.3 %. Malaysia (KLSE) has the lowest maximum return of 6.7%, with the lowest risk of 1.9%. The countries India (BSE) have the maximum weekly return of 13.2%, China (SSE) 11.7%, remaining and Japan (N225) 11.4% and Pakistan 10.9% respectively. The stock indices understudy all the stock indices are negatively skewed except Japan (Nikkei 225) stock index and China (SSE), which indicates that the countries' weekly returns under study were negative in more weeks than positive except Japan and China stock markets. The Jarque-Bera suggests that the data is not normally distributed. The kurtoses measure shows that the data is leptokurtic, where the values of all the sample variables are more significant than 3, which concluded that the information is non-normally distributed. Skewness indicates that either the data is negatively or positively skewed. (*) suggests that the Jarque-Bera stats are significant at a 5 percent level, and (N) is the numbers of observations.

Correlation Matrix
The research study employed a correlation matrix to identify the relationship in the Asian Developed and Emerging Stock market indices. The correlation between the two stock markets may be positive or negative. When the two variables are highly correlated, the co-efficient value maybe +1 or -1. The positive relationship shows, when one variable return increases, the other return will also increase, and the antagonistic relationship shows that one variable return increases, the other will decrease. The co-efficient value near zero implies no ties. Table 2

Unit Root Test
The research study employed a correlation matrix to investigate the association among the variables. Still, to explore the long-run and short-run relationship, the task is using the cointegration test. The cointegration test assumes that the data must be non-stationary and integrated of the same order, preferably I(1). To clarify the fact that either the data is stationary in the same order or not, this research employs two tests Augmented Ducky Fuller (ADF) and Philips Parron  Table 3 summarizes the results of the Unit Root Tests. By employing the ADF and P.P. tests, the results specify that all the variable's data is non-stationary and stationary in the first difference. The (*) in the table shows significance at a 5 percent level. So it is clear that we can investigate the long-run relationship by employing the cointegration test.

Lag Structure
This study is examining the long-run and short-run relationship among the variables under study. To identify long-run and short-run relationships, this study employs a cointegration test. To run the Co-integration test lag value is needed. Different researchers used different lag length criteria. By finding the appropriate lag vale, Vector Autoregressive Statistics are using in this study. Table 4 summarizes the lag values, which indicates that this research study will use lag 1 for running the cointegration test which is an appropriate lag value by Schwarz criteria, Shezad et al (2014). Most of the researchers used Akaike Information Criteria (AIC). This study uses S.C. because S.C. is significant at lag 1 and AIC is significant at lag 4. The results from lag 1 will be accurate instead of lag 4.

Vector Error Correction Model (VECM)
As the Johnsen Co-integration test found one cointegration relationship among the Asian developed and emerging stock markets of Asia. To investigate the long-run and short-run association among markets Vector Error Correction Model (VECM) will employ. The VECM identifies that either the stock market indices are positively or negatively inter-linked.  (2015) found longrun equilibrium associations in the Asian region.   Table 7 summarizes the casual relationship among the time series variables under study. The casual relationship indicates

Variance Decomposition
The research study employed Variance Decomposition to examine that how many variances in any stock market are happening in percentage by its advancement and how many clashes are occurring by other stock markets of the research study, Moon (2018)

Impulse Response Function (IRF)
The Impulse Response function (IRF) is employed to forecast that how much the shocks in one country are affecting the other country. The figure 1 to 7 summarizes the IRF of the Asian developed and emerging stock markets. The x-axis in the figures indicates the data period up to ten periods, so this research study is analyzing weekly data. Here the Impulse Response Function can forecast the shocks occurring in one country up to ten weeks in the other country. The y-axis indicates the percentage change that occurred from shocks in one country to the other country. Figure 1 charts the response of KSE to the stock markets indices of the study. The response of KSE to KSE indicates that is highly positive during the first two weeks then gradually decreased and comes to level in the third week then consistent up to ten weeks. Figure 2 shows the response of BSE to the other seven stock indices of the study. The response of BSE to BSE indicates that in the first week the response is highly positive, which is gradually decreased and in the third week, it is positive but consistent up to ten weeks. Figure 3 shows the response of CSE to the other stock market indices of the study. The response of CSE to CSE is highly positive in the first weeks then gradually decreased up to the fourth week and then consistent up to ten weeks. Figure 4 graphs the response of KLSE to the stock markets indices of the study. The response of KLSE to KLSE indicates that is highly positive during the first two weeks then gradually decreased and comes to a level in the fourth week which is consistent up to ten weeks. Figure 5 charts the response of SSE to the stock markets indices of the study. The response of SSE-to-SSE indicates that is highly positive during the first week then gradually decreased and becomes negative in the second week, again come to a level in the third week then consistent up to ten weeks. The response of SSE to N225 is positively the least increase in first two weeks then decreased and come to a level in the third week, which is consistent up to ten weeks. The response of SSE to CSE is positive in the first week, then negative in the second week again come to level in the third week and consistent up to ten weeks. Figure 6 charts the response of KOSPI to the stock markets indices of the study. The response of KOSPI to KOSPI indicates that is highly positive during the first week then gradually decreased and becomes negative in the second week, again come to a level in the third week then consistent up to ten weeks. Figure 7 graphs the response of N225 to the stock markets indices of the study. The response of N225 to N225 indicates that is highly positive during the first week, then becomes negative in the second week and in the third week again positive and consistent up to ten weeks. -.005 .000 .005 .010 .015 .020 .025 .   Response of N225 to Cholesky One S.D. Innovations