On the Interdependence Structure of Market Sector Indices - Essay Example

Project Report On the Interdependence Structure of Market Sector Indices: the Case of Qatar Exchange This research paper investigates the interrelationships among the variety of specific indices of the Qatar Exchange, including: Banking and Financial Institutions, Insurance, Industrial indices, and Services. The author examines three key issues: the short-run causal relationships amongst the sectors, the long-run relationships amongst the sectors, and the relative degree of exogeneity/endogeneity of each sector. In order to perform this analysis there were employed several econometric models, including: Johansen’s multivariate cointegration, generalized forecast error variance decomposition, and Granger’s causality. For empirical analysis there were used both daily and weekly closing stock price indices for the four sectors of the Qatar Exchange for the period from January, 2008 to April, 2011. The results of the study based on the Johansen’s multivariate cointegration analysis have shown that all four sector indices had interrelationships in a long-term equilibrium. The findings based on the Granger’s causality analysis have shown that Banking and Financial Institutions sector had significant impact on the price predictability in the other sectors in the short-run, while Services sector had much lower impact. The analysis based on the findings of the generalized forecast error variance decomposition has shown that the Services and Insurance sectors appeared to be the most endogenous sectors while both the Industrial and the Banking and Financial Institutions sectors were the most exogenous sectors (Walid, 2012). Table of Contents 1. Introduction…………………………………………………………………4 2. Literature Review……………………………………………………………5 3. Research Objectives and Research Questions ……………………………..6 4. Research Design (Methods and Methodology)……………………………..7 5. Data Collection and Analysis……………………………………………….8 6. Empirical Analysis…………………………………………………………..8 7. Conclusion…………………………………………………………………12 8. References…………………………………………………………………14 Introduction Both individual and institutional investors usually try to diversify their portfolios in order to minimize the risks and to generate most efficient outcomes and/or expected returns. There are recognized two major strategies under portfolio diversification: cross-market investment to in different classes of assets or sectors and global diversification of portfolio by investing across both domestic and foreign assets. However, it is worth to mention that in order to create a well-diversified asset portfolio, prospective investors should take into consideration the fact of interdependence structure of financial markets (Walid, 2012). There were carried out many empirical researches, which provided evidence of interdependencies of international financial markets as a result of globalization and increased level of integration. However, it is also important to understand the nature of interdependence structure amongst different sectors in a specific country, because it can provide prospective and actual investors with some practical insights on potential gains derived from the diversified portfolio. Also, the knowledge of how much each index influences a price variation in other indices can be helpful for policymakers’ decision inferences. As Qatar Exchange market is relatively unexplored market, this paper aims to examine the internal dynamics of Qatar Exchange market with a focus made on the interdependence structure amongst the sector-specific indices of the Qatar Exchange (Walid, 2012). Literature Review Over the past thirty years there were carried out many empirical researchers aimed to explore integration amongst financial markets. Majority of these studies have illustrated interdependencies amongst the international financial markets. The number of studies focusing on the issue of interdependence structure in the industry or specific country is more limited. Arbela´ezet al. (2001) have carried out a research, examining both the long-run and short-term linkages among the stock indices of the Medellı´n Stock Exchange, Columbia. Having analyzed six sector indices (including financial, general, commerce, industrial, various and select) the authors found both long-run cointegration relationships and short-term linkages amongst the sector indices of the Columbian market in more than 50% of cases. Ewing (2002) also has carried out a study attempting to determine the relative importance of individual S&P stock indices’ shocks (financials, utilities, industrials, transportation, and capital goods) for the forecast error variance among other indices (Walid, 2012). Lafuente and Ruiz (2004) have carried out analysis of the Spanish stock market, and illustrated positive correlation between the new market index volatility (for technological corporations) and between the volatility and return of Spanish sector indices (industrial, financial, and utilities). Wang et al. (2005) have examined the Chinese Stock Exchange in Shenzhen and Shanghai for interdependence and unveiled a high degree of interdependence across various sectors. Another one study was carried out by Constantinou et al. (2008) who have explored both short-term and long-term dynamic relationships amongst the economic sectors of the Cyprus Stock Exchange and found no evidence of long-run relationships amongst the sectors (Walid, 2012). Moreover, their findings have shown that over time the analyzed sector indices were absolutely independent, suggesting great opportunities for domestic investors. Patra and Poshakwale (2008) also have explored the behavior of sector indices of the Athens Stock Exchange and found that banking sector had significant influence over variance in the insurance and construction sectors, and slightly lower influence over variance of the investment, industrial, and holding sectors. Al-Fayoumiet al. (2009) carried out empirical study examining interdependence structure amongst the four stock market indices in Jordan Stock Exchange. The authors have supported the concept of cointegration amongst the four stock market sector indices (including financial, industrial, service, and genera), whereas the financial sector was the most influential one. One of the most recent studies was carried out by Karmakar (2010) who has explored return and volatility effects between small and large stocks in the National Stock Exchange in India. Based on the study results the author has concluded that there were significant asymmetric return spillovers from the large stocks’ portfolios to small stocks’ portfolios (Walid, 2012). Research Objectives and Research Questions The research objective is to provide empirical evidence on the interdependence structure amongst the sector-specific indices of the Qatar Exchange, including Banking and Financial Institutions, Insurance, Industrial indices, and Services. The research questions which should be investigated cover the following three key issues: the short-run causal relationships amongst the sectors, the long-run relationships amongst the sectors, and the relative degree of exogeneity/endogeneity of each sector (Walid, 2012). Research Design (Methods and Methodology) In order to achieve research objective, there was employed the econometric analyses of three models: Johansen’s multivariate cointegration, Granger’s causality, and generalized forecast error variance decomposition. Johansen’s multivariate cointegration analysis was chosen for assessment of the existence of a long-run equilibrium relationship amongst the sector indices. This specific type of analysis was chosen due to the fact the both differences and levels of variables would be represented and thus, information included in levels would not be lost as in other cases. Granger’s causality analysis enables the author to explore whether X causes Y and to see how much of “current Y can be explained by past values of Y and then to see whether adding lagged values of X can improve the explanation” (Walid, 2012:474). Generalized variance decomposition analysis was used to fill the gap created by the Granger causality analysis as VAR VECM do not allow to measure the degree of exogeneity amongst the stock indices beyond the same period of time and the relative strength of the Granger causal chain. Variance decomposition analysis was applied to measure the extent to which the variation of sector indices could be explained by innovations from other sector indices and thus to reveal the level of endogeneity or exogeneity of a certain sector index (Walid, 2012:475). If the forecast error variance of the sector index is mainly explained by innovations in other sector indices, then it should be considered endogenous, and if it is explained by its own innovations, then – exogenous. Data Collection and Analysis For this study there were used both daily and weekly stock index closing prices for the four sectors of the Qatar Exchange, including the following: Banking and Financial Institutions, Insurance, Industrial indices, and Services. The historical data of performance of stock indices was taken for the period from January 2, 2008 to April 7, 2011. Due to the fact that Fridays and Saturdays at the Qatar Stock Exchange were weekends, every Tuesday was a day of closing prices for weekly time series. The total sample size has been fixed at 802 daily observations, and 165 weekly observations. The data for all four respective sector indices was taken from Qatar Exchange corporate website (Walid, 2012). Empirical Analysis Results of Johansen’s cointegration analysis In order to determine potential cointegrating relationships amongst the four sector indices, there was used a multivariate test developed by Johansen. The results of Johansen’s multivariate cointegration analysis based on both daily and weekly frequency data have shown the existence of a single cointegrating factor across the identified sector indices with significance level of 5 percent. Thus, the author has interpreted these results as a confirmation of the hypothesis that the sector indices of the Qatar Exchange are interrelated as their movements are correlated amongst each other, at least in the long-term perspective (Walid, 2012: 479). These findings are consistent with the findings published by Arbela´ez et al. (2001) for the Columbian Stock Exchange, Wang et al. (2005) for the Chinese stock markets, Patra & Poshakwale (2008) for the Athens Stock Exchange, and Al-Fayoumiet al. (2009) for Jordan Stock Exchange. Thus, the authors provided additional evidence of the fact that various sectors in one country depend on different market forces and country-specific factors to certain degree. Results of Granger’s causality analysis Granger’s causality analysis was used to determine the direction and strength of information transmission between sector indices. Block exogeneity Wald test was used to examine “the joint significance of each of the other lagged endogenous variables in each equation” as well as “the joint significant of all other lagged endogenous variables” (Walid, 2012: 479). Due to this analysis there were revealed the following findings: The Insurance sector did not appear to Granger cause any of the other three sectors of the Qatar Exchange; Services sector did not appear to Granger cause the other sectors, except the Banking and Financial sectors; Banking and Financial sector was found to Granger cause the Insurance sector; There was identified significant causal relationships between the pairs of Industry Banking and Financial sectors, and between the pairs of Banking and Financial sectors and Services sectors; Industry sector had a significant causal factor in relation to price changes of both Services and Insurance sectors; However, the results of weekly data analysis were somewhat inconsistent with daily frequency data analysis, as all pairwise Granger causal relationships were found to be unidirectional. Industry sector was found to Granger cause the other sectors. Still, the Banking and Financial sector was found to Granger cause to both Services and Industry sectors. Results of generalized variance decomposition analysis Generalized variance decomposition analysis was utilized in order to determine the relative exogeneity or endogeneity of each sector under the study. Testing of the daily frequency data has shown that both Banking and Financial and Industry sectors were increasingly influential on each other. Further, analysis has shown that these two sectors were the most exogeneous sectors in the Qatar Exchange, whereas Industry sector had 75,35 percent, and Banking and Financial sector had 75,14 percent of its forecast error variance. However, Industry sector was found to be the most endogenous sector with 73,73 percent. The findings also have enabled to define the importance of innovations in the Banking and Financial sector on the forecast error variances in other sectors. Insurance and Services sectors were found to have minimum capacity to influence any of the other sectors under the study. These findings support the idea that Banking and Financial sectors play dominant role over the other sectors in the Qatar Exchange, and therefore, the most influential role in the Qatari stock market. Conclusion The main objective of the current study was to provide empirical evidence on the interdependence structure amongst the sector-specific indices of the Qatar Exchange, including Banking and Financial Institutions, Insurance, Industrial indices, and Services. The research questions, which were investigated, covered the following three key issues: the short-run causal relationships amongst the sectors, the long-run relationships amongst the sectors, and the relative degree of exogeneity/endogeneity of each sector. The data used in the study were daily and weekly closing price indices for the four respective sectors of the Qatar Exchange for the period from January 2, 2008 up to April 7, 2011, retrieved from its corporate website. In total, sample sizes were comprised of 802 daily observations and 165 weekly observations. In order to investigate the long-run relationships amongst the sector indices, there was utilized the Johansen’s multivariate cointegration analysis. The results of trace statistic and Max-eigenvalue statistic analysis of daily and weekly data have provided evidence of a unique cointegration relationship amongst the four respective sector indices of Qatar Exchange. Further, the analysis was continued by the Granger’s causality analysis aiming to detect direction and strength of information transmission across the sectors. The analysis has shown that the Banking and Financial Institutions sector was the most influential one in terms of price predictability of the other three sectors, while the Services sector was the least influential sector. Next, the study analyzed each sector for relative endogeneity/exogeneity factor with a help of the generalized forecast error decompoisition analysis. Based on the findings of the analysis of daily data, the authors concluded that the Services (SER) and Insurance sectors (INS) were the most endogeneous sectors, while the Banking and Financial Sector (BFI) as well as Industrial Sector (IND) were the most exogeneous ones. The results of analysis of the weekly data have supported this conclusion in relation to BFI and SER sectors. However, the results of relative exogeneity and endogeneity of the INS and IND based on daily frequency data were inconsistent with the results based on weekly frequency data. This inconsistency in results was referred to the unique properties of the daily and weekly frequencies. The results of this empirical study have supported the findings published by Patra and Poshakwale (2008) for the Athens Stock Exchange, Arbela´ezet al.(2001) for the Colombian Stock Exchange, and by Al-Fayoumiet al.(2009) for the Jordan Stock Exchange. The current study has not only supported the theory of existence of the long-run cointegration relationships amongst the sector indices in one country, but also identified banking and industry sectors as the most important and influential sectors in terms of its effect on the return and volatility of other sectors. Finally, it has been suggested that the study should have practical implications for both institutional and individual investors considering cross-sector diversification strategy. As the sector indices of the Qatar Exchange were interdependent on the other sectors within domestic economy, the potential gains could be below initial expectations. The findings of the study might also be helpful for policymakers for setting up relevant precautionary measures and developing appropriate strategies for managing unanticipated shocks from BFI sectors. References: Walid A. (2012). On the interdependence structure of market sector indices: the case of Qatar Exchange. Review Of Accounting And Finance, 11(4), 468-488. doi:10.1108/14757701211279204. Read More