Abstract
This paper investigates the short-run and long-run dynamics among the major sectoral stock indices of the Istanbul Stock Exchange over the period 1997-2011. Long-run relationship among these indices is analyzed by using both conventional Engle and Granger (1987) and Johansen-Juselius (1990) cointegration tests, causal relationship through Vector Error Correction Model (VECM). Likewise, variance decomposition analysis is employed to partition the variance of the forecast error of one sector index into proportions attributable to shocks in each sector index in the system including its own. The findings suggest that all sectors show consistent and strong evidence of a long-run relationship, the short-term causal relationships between the sector indices are considerably limited and, where they exist, especially unidirectional. The variance decomposition analysis confirms that even though a high percentage of error variance is accounted for by the innovations in the same index, innovations in the variance of returns in the Banking sector are able to explain, on average, 63% innovations in the variance of the Chemistry, Petrol and Plastic, 57% of the Basic Metal and 79% of the Holding and Investment sector returns. These results indicate that the Banking sector is the most influential sector in the ISE.