THE INTEGRATION OF ISLAMIC STOCK MARKETS: DOES A PROBLEM FOR INVESTORS?

This paper empirically investigates co-integration between Islamic stock market in Malaysia, Indonesia and the world by applying the Vector Auto Regression (VAR) method. This research used monthly data from January 2007 to May 2012 taken from authorized sources. The finding shows that there is no long-run or equilibrium relationship exists between FTSE Bursa Malaysia Emas Shariah (FBMES), Jakarta Islamic Index (JAKISL), and Dow Jones Islamic Market index (DJIM). Based on the result, hence, it can be concluded that the Islamic stock markets of Malaysia does not integrate with Indonesia’s, as well as with the world markets in the long run. This will create rooms for investors to diversify their investment portfolios, which puts Malaysia as one of their favoured investment destinations. From the Granger causality view, the return of JAKISL and DJIM is driven by the return of FBMES in the short run.


Introduction
The study of stock markets integration has received substantial interests from various stakeholders particularly after the 1987 stock market crash and the 1997-1998 Asian financial crises (Phuan et al., 2009;Marashdeh and Shrestha, 2010). There are several reasons that establish the importance of stock markets integration. First, it provides further opportunities in risk sharing among integrated markets at a cost of taking away the diversification benefit of international investment (Siskawati, 2011). Second, the information about stock market integration is then used by investors to set investment strategies based on the potential benefit that can be gained by diversifying in different stock markets (Kassim, 2010). Third, it promotes a vital condition for a country's financial sector to develop as a center for international financial sector (Reddy, 2003). Fourth, stock market integration could generate a higher productivity and economic growth across the economy by stimulating domestic savings and investments (Mohan, 2005). Fifth, it leads to financial stability by promoting effective operation of financial intermediaries and resource allocation (Trichet, 2005). Sixth, market liberalization reduces the cost of capital within the integrated markets due to risk sharing between the country's resident and foreign investors in the domestic economic activities (Tai, 2007).
Numerous researches have studied the conventional stock market integration among countries (see, for examples, Ratanapakorn and Sharma (2002), Yang et al. (2003), Narayan et al. (2004), Floros et al. (2005), and Majid et al. (2008). In comparison to the studies of market cointegration in conventional stock markets, there is little empirical research that examines the long run equilibrium among Islamic stock markets. Thus, this paper will contribute significantly to the literature by providing new evidence on Islamic stock market integration nexus in the specific context of FTSE Bursa Malaysia Emas Shariah Index (FBMES), Jakarta Islamic Index (JAKISL) and Dow Jones Islamic Market Index (DJIM). The structure of this paper is as followed: section 2 overview of Islamic stock Markets, section 3 presents literature review pertaining to previous researches on the stock market integration. Section 4 describes the data and variables used in this study. Section 5 discusses research methodology while Section 6 presents and analyses the empirical results. Section 7 concludes the findings.

Islamic Stock market
The Asian Financial Crisis 1997 is the turning point for the Islamic finance industry, as the investors have started to lose confidence in the existing conventional market systems and demand to switch to the Islamic markets. The growing demands for the Islamic investment products are evident in both developed countries, particularly in the UK, the US and Japan, and in developing countries such as Malaysia, Egypt and Sudan. Impressively, there are more than 600 Islamic financial institutions operating in more than 75 countries nowadays. Islamic banking and finance industry is no longer seen as competitor to its counterpart, instead it is recognized as a viable alternative to the conventional financial system as it has high potential to be developed further (Kassim, 2011).
Basically, the function of the Islamic stock market is similar to the conventional which is to provide a benchmark or performance indicator for a group of companies stocks' in a particular market.  (Rahim et al., 2009). While FBMES and JAKISL represent the Islamic market indexes for Malaysia and Indonesia Shariah-compliant investments respectively, the Dow Jones Islamic Market (DJIM) is referred to the first benchmark investment performance for the global Islamic index (Siskawati, 2011). Commencing in 1999, the DJIM include the Shariahcompliant stocks from 34 countries.

Literature Review
There have been few studies focusing on the integration of Islamic and conventional stock markets both regionally and globally, for examples, Roca et al. (1998), Wongbangpo (2000, Hee (2002), Azman-Saini et al. (2002), Dunis and Shannon (2005), Majid et al. (2007), Yusof and Majid (2007), Karim, Kassim and Arip (2010), Majid et al. (2008), and Siskawati (2011). Roca et al. (1998) investigate the price linkage among five ASEAN markets (i.e. Indonesia, Malaysia, Singapore, Thailand and Philippines) both in long-run and shortrun using cointegration technique based on Johansen procedure (1998), Granger Causality, Variance Decomposition and Impulse Response function for 1988-1995 period. They conclude that with an exception of Indonesia, ASEAN-5 markets were closely linked in the short-run but not in the long-run. Furthermore, the study shows that Malaysia is the most influential market while Singapore and Thailand are the markets with the most linkages with other markets. However, Indonesia has no linked at all with any other ASEAN markets. This indicates that there exist long run benefits for investors from diversifying their portfolio investments in these five countries, particularly in Indonesia.
Dunis and Shannon (2005) examine whether emerging markets still offer international investors a valuable diversification benefit. The study covers emerging markets from Southeast Asia (i.e. Indonesia, Malaysia and Philippine) and Central Asia (i.e. Korea, Taiwan, China and India) with the US and UK markets (as the reference of established markets). They found that the overall results indicate that international diversification into the emerging equity markets considered is beneficial for US investors during the period under study. In addition they explain the international portfolio diversification can contribute to a reduction in systematic risks since it reduces domestic market exposure that cannot be diversified. On the other hand, Middleton et al. (2008) showed that the beneficial opportunities from investing in Central and Eastern European emerging markets are still enormous, even during the financial crisis.
Another study about market integration in ASEAN region is conducted by Wongbangpo (2000). The study reveals that ASEAN stock markets, except for the Philippines, have long-run co-movement during the period 1985-1996. This means that an effective longterm diversification of an investor's portfolio among these stock markets cannot be achieved. By employing correlation and co-integration analysis, Hee (2002)  Nevertheless, these eight OIC stock markets are found to be co-integrated with the US, Japan and the UK stock markets, implying that investors can gain long-run risk reduction through portfolio diversifying in the MENA region but not in the Asian region.
Yusof and Majid (2007)  Similarly, Majid et al. (2008) study the integration among ASEAN-5 emerging stock markets and their interdependencies from the US and Japan. Employing a two-step estimation, cointegration and Generalized Method of Moments (GMM for period 1 January 1988 -31 December 2006, they found that ASEAN-5 (i.e. Malaysia, Thailand, Indonesia, The Philippines, and Singapore) stock markets are moving towards more integration among themselves or with the US and Japan, especially after financial crisis 1997. Furthermore, the results show that the Granger causal relations among the market in the region kept changing over the period. The degrees of short and long-run relationship have increased significantly. The study also reveals that each ASEAN market is found to be interrelated with the US and Japan in different ways.
In the more recent researches, Karim, Kassim and Arip (2010) examined the effects of the US Subprime Crisis on the integration of selected Islamic stock markets (i.e. Malaysia, Indonesia, US, UK and Japan). Employing Johansen-Juselius (JJ) cointegration approach, the study failed to prove the existence of cointegration among these Islamic stock market for both pre-crisis period (February 15, 2006July 25, 2007 and during crisis period (July 26, 2007-December, 2008. In contrast, Siskawati (2011)

Data Description
The monthly data for this study is retrieved from Bloomberg Database, spanning from January 2007 to May 2012. The logarithmic values of monthly market indices have been used as the proxy of selected market. The definitions of each variable and the time-series transformation are presented in Table 1.

Methodology
We adopted a Vector Autoregressive (VAR) model to examine the integration of Islamic stock market index between FBMES, JAKISL and DJIM, in which can be written as follow: (1) Our aim is to investigate the market co-integration among FBM Emas Syariah Index (FBMES), Jakarta Islamic Index (JAKISL) and Dow Jones Islamic Market index (DJIM). Deriving from equation 1, the estimation of expanded model of the integration of these Islamic stock market indexes is: where R is 3 x 3 matrix polynomial parameter estimators, (L) is lag length operators, A is an intercept, and et is a Gaussian error vector with mean zero and Ω is a Varian matrix.
To properly specify the VAR model, we follow the standard procedure of time series analyses. First, we apply the commonly used augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root tests to determine the variables' stationarity properties or integration order. Briefly stated, a variable is said to be integrated of order d, written 1(d), if it requires differencing d times to achieve stationarity. Thus, the variable is nonstationary if it is integrated of order 1 or higher. Classification of the variables into stationary and non-stationary variables is crucial since standard statistical procedures can handle only stationary series. Moreover, there also exists a possible long-run comovement, termed cointegration, among non-stationary variables having the same integration order. In this study, the lag is determined based on Akaike Information Criterion (AIC) which is commonly used for the VAR model. Second, we apply a VAR-based approach of the cointegration test suggested by Johansen (1988) and Johansen and Juselius (1990) to investigate the long run relationships among these variables. The cointegration test is based on the maximum likelihood estimation of the VAR model. If these variables were not cointegrated, the standard Granger causality test would be applied on the first difference of these variables. If these variables were cointegrated, a Granger causality test would be conducted based on a VAR with the introduction of an error correction term following the suggestion of Granger (1986). Granger causality tests are performed to identify the existence and nature of the causality relationship between the variables.
Finally, the Variance Decomposition (VDC) is performed to obtain the degree of exogeneity between variables outside of the sampling period. The VDC shows the percentage of forecast error variance for each variable that is attributed to its own shocks and to fluctuations in the other variables in the system.

Empirical Results
Table 2 provides the summary statistic of the Islamic stock returns (i.e. stock prices in first difference). FBMES has been the most active and profitable, showing the highest average daily returns of 9.071. This is followed by DJIM at 7.308 and JAKISL at 6.017. However, compared to other market indexes, JAKISL return (as reflected by the standard deviation) is very high and it also has the highest volatility 0.286 which is commensurate to its return. The smallest risk is experienced by FBMES 0.181. The negative value of skewness indicates that the series distributions are skewed to the left. All stock markets have positive kurtosis value with the highest value is DJIM at 3.840. This indicates that the distribution of stock market is leptokurtic than the normal distribution. The Jarque-Bera test rejects normality for all distribution.  Table 3 summarizes the results of both Augmented Dickey-Fuller (ADF) and Phillip Perron (PP) tests for all variables. As shown below, the null hypothesis of nonstationarity for the ADF and PP tests is accepted for all variables, indicating that all variables are non-stationary in level but become stationary after first differencing. Thus, they are integrated of order 1, or I(1).  , M.H., Yusni Anis. Y., Fidlizan, M., Azila, A.R., Emilda, H., & Nur Fakhzan, M. integration, we proceeded to test whether they were co-integrated. To achieve this, the Johansen Multivariate Co-integration test was employed. The results of the Johansen's Trace and Max Eigen-value tests are shown in Table 4, where it is found that there is no co integration among variables. This can be observed from the values of trace statistic and Max Eigen value, which are smaller than their critical values, respectively. Therefore, it can be concluded that there is no long-run or equilibrium relationship between FBMES, JAKISL and DJIM. Based on the result, hence, it can be concluded that the Islamic stock markets of Malaysia does not integrate with Indonesia's, as well as with the world markets in the long run. This will create rooms for investors to diversify their investment portfolios, which puts Malaysia as one of their favoured investment destinations.    The results of the VDC analysis are presented in Table 6

Conclusions
The objective of this paper is to investigate the integration among Islamic stock markets in Malaysia, Indonesia and World. Using VAR estimation technique, this paper found Indicator: uni-directional causality bi-directional causality

LNFBMES LNJAKISL LNDJIM
that there is no co-integration relationship (long term equilibrium) between these Islamic indices implying that the Malaysian Islamic stock markets does not integrated with Indonesian Islamic stock market as well as with the Islamic world stock markets in the long run. This will create rooms for both local and international investors to diversify their Islamic investment portfolios, which will put Malaysia as one of their investment destination. On the other hand, the results of the Granger causality test show a significant bi-directional causality between global Islamic stock market and Malaysian Islamic stock market (DJIM and FBMES) and a significant uni-directional causality relationship between FBMES and JAKISL. Therefore, these scenario suggesting that the return of Islamic stock market in Indonesia is significantly affected by the return of Islamic stock market in Malaysia. This result, hence leads to a conclusion that the return of JAKISL and DJIM is driven by the return of FBMES in the short run.