Table 2.
Directional spillovers (in %).
| To market i | From market j |
|||
|---|---|---|---|---|
| EM | DM | US | From Others | |
| EM | 33.335 | 32.636 | 34.029 | 66.665 |
| DM | 31.603 | 34.641 | 33.756 | 65.359 |
| US | 32.135 | 32.541 | 35.324 | 64.676 |
| Contribution to others | 33.335 | 32.636 | 34.029 | 66.665 |
| 65.57% | ||||
| Net Contribution (To – From) Others | -2.926 | -0.182 | 3.109 | |
Notes: This table reports the variance decompositions for estimated TVP-VAR model. Variance decompositions are based on 10-step-ahead forecasts with the ordering as shown in the column heading, i.e. the (i, j) value is the estimated contribution to the variance of variance premium forecast error of market i coming from innovations to variance premium of market j. A lag length of order 1 was selected by the Bayesian information criterion.
The lower right corner (in bold) indicates the level of the total spillover index in %. The last row ‘Net Contribution’ indicates the total sum of net directional pairwise spillovers, expressed as a negative value (net recipient) and a positive value (net transmitter), respectively.