Causality between regional stock markets: A frequency domain approach

Nikola Gradojevic, Eldin Dobardzic
2013 Panoeconomicus  
Using a data set from five regional stock exchanges (Serbia, Croatia, Slovenia, Hungary and Germany), this paper presents a frequency domain analysis of a causal relationship between the returns on the CROBEX, SBI-TOP, CETOP and DAX indices, and the return on the major Serbian stock exchange index, BELEX 15. We find evidence of a somewhat dominant effect of the CROBEX and CETOP stock indices on the BELEX 15 stock index across a range of frequencies. The results also indicate that the BELEX 15
more » ... dex and the SBITOP index interact in a bi-directional causal fashion. Finally, the DAX index movements consistently drive the BELEX 15 index returns for cycle lengths between 3 and 11 days without any feedback effect. JEL: C58, G15. Causality between Regional Stock Markets: A Frequency Domain Approach PANOECONOMICUS, 2013, 5, pp. 633-647 ω ω Note: The values of the χ 2 test statistic are given by a solid line. The 5% critical value (5.99) is given by a horizontal dashed line. The null hypotheses are: 1) that CROBEX returns do not cause BELEX 15 returns at frequency ω (left), and 2) that BELEX 15 returns do not cause CROBEX returns at frequency ω (right). Source: Authors' calculations. Figure 1 Causality Tests (CROBEX to BELEX 15 Returns) (Left Panel) Causality tests (BELEX 15 to CROBEX Returns) (Right Panel) Ω Ω Note: The values of the χ 2 test statistic are given by a solid line. The 5% critical value (5.99) is given by a horizontal dashed line. The null hypotheses are: 1) that CETOP returns do not cause BELEX 15 returns at frequency ω (left), and 2) that BELEX 15 returns do not cause CETOP returns at frequency ω (right).
doi:10.2298/pan1305633g fatcat:vyq7fxnal5e37gkomtexq5bc3e