While Granger causality results show qualitative relationship between the variables, variance decomposition analysis provides quantitative measure to these causal relationships representing how much the movement in a sector index can be explained by other indices in terms of the percentage of the forecast error variance of that sector index. Results for variance decomposition analysis for periods ranging from 1 day, 5 days, 10 days and 15 days ahead of forecast for all indices are summarized in Table 7. The findings indicate that for all sector indices, high percentage of forecast variance is influenced by the innovations in the past variance of the same sector. The analysis concentrates on 15-day ahead forecast result in order to see the contribution of each index in explaining the percentage of forecast error variance. For instance, after 15 days 99.36% of the variation in Banking sector is explained by its own innovations whereas changes in Telecommunication; Chemistry, Petrol and Plastic; Trade; Basic Metal; Holding and Investment only explain 0.07%, 0.14%, 0.10%, 0.05% and 0.25% of the variance in the Banking sector, respectively. On the other hand, Banking sector explains approximately 63% of the variation in Chemistry, Petrol and Plastic sector, 79% of the variation in Holding and Investment sector, 45% of the variation in Telecommunication sector which is consistent with causality results. The dominant role of the Banking sector entails that changes in the Banking sector index may potentially be used in forecasting movements in other sector indices corroborating a radical departure from the weak form efficiency. cheapest payday loans
This paper analyzes short and long-run relationships among the major sector indices of ISE, using daily data. The empirical results from Engle and Granger two-step cointegration tests indicate that there is long-run relationship between all sector index pairs. On the basis of cointegrated variables, the Granger causality test through VECM is used to check the direction of the short and long-run relationship. The significant lagged error correction terms suggest the bidirectional causality in the long-run for all sector indices except Banking-Holding and Holding-Commerce index pairs. For the short-run causality channels, Banking granger causes Chemistry, Holding and Communication. Moreover, the empirical results suggest the existence of short-run bidirectional causality between Holding and Commerce whereas unidirectional causality from Chemistry to Commerce. The leading and dominant role of the Banking sector, confirmed by the results of variance decomposition test, can be explained by 41% share of the total market capitalization of the ISE-100 index.
The findings of this paper provide important information particularly for investors and portfolio managers in optimal portfolio allocation decision across sectors. The findings show that there are no potential portfolio diversification benefits through investing in the sub-sector indices in Turkish stock market.

Table-7: The results of the Generalized Variance Decomposition Analysis

Explained by
Period XBANK XILTM XKMYA
XBANK 1 100 0.00 0.00
5 99.36 0.07 0.14
10 99.36 0.07 0.14
15 99.36 0.07 0.14
XILTM 1 45.07 54.92 0.00
5 44.99 54.71 0.00
10 44.99 54.71 0.00
15 44.99 54.71 0.00
XKMYA 1 63.05 1.24 35.70
5 62.71 1.27 35.51
10 62.71 1.27 35.51
15 62.71 1.27 35.51
XTCRT 1 49.49 1.53 4.71
5 49.15 1.60 4.97
10 49.15 1.60 4.97
15 49.15 1.60 4.97
XMANA 1 56.80 1.78 5.63
5 56.67 1.78 5.63
10 56.67 1.78 5.63
15 56.67 1.78 5.63
XHOLD 1 79.09 1.29 3.79
5 78.75 1.28 3.78
10 78.75 1.28 3.78
15 78.75 1.28 3.78
XTCRT XMANA XHOLD
XBANK 1 0.00 0.00 0.00
5 0.10 0.05 0.25
10 0.10 0.05 0.25
15 0.10 0.05 0.25
XILTM 1 0.00 0.00 0.00
5 0.01 0.06 0.20
10 0.01 0.06 0.20
15 0.01 0.06 0.20
XKMYA 1 0.00 0.00 0.00
5 0.10 0.13 0.26
10 0.10 0.13 0.26
15 0.10 0.13 0.26
XTCRT 1 44.26 0.00 0.00
5 43.90 0.01 0.35
10 43.90 0.01 0.35
15 43.90 0.01 0.35
XMANA 1 0.12 35.65 0.00
5 0.16 35.63 0.10
10 0.16 35.63 0.10
15 0.16 35.63 0.10
XHOLD 1 0.57 1.04 14.20
5 0.65 1.16 14.35
10 0.65 1.16 14.35
15 0.65 1.16 14.35