Source: Authors’ calculations based on Bloomberg data. The Granger causality was considered for 1 lag, 5 lags (a typical trading week), and 20 lags (the average number of trading days within a month).ĬSE Main Market Granger causes CSE InvestmentĬSE Investment Granger causes CSE Main Market For Cyprus, only the main market indices were considered since it has the largest contribution to the stock market capitalization.
For Croatia only two indices were retained for analysis due to the fact that all the sector indices were launched recently and the data are available only since Feb. The total return (TR) indices were eliminated due to their short history for all the exchanges. The period under investigation differs for each of the four exchanges. The data for the indices were provided by Bloomberg for the available period between Jan. The p-values for each F-test are reported in table format. An F-test is then used to determine whether the coefficients of m-lag values of the reference index are jointly equal to zero.
The null hypothesis that the reference index does not Granger-cause the chosen index is accepted if and only if no lagged values of reference are retained in the regression. R Index ( t ) = a 0 + a 1 R Index ( t − 1 ) + ⋯ + a m R Index ( t − m ) + b 1 R BET-Fi ( t − 1 ) + ⋯ + b m R BET-Fi ( t − m ) + ε t The other results of this study further reveal that English-speaking African countries exhibit (1) unidirectional causality from investment to economic growth in the short and long run in all the three models, and (2) unidirectional causality from FD2 to investment in the short run. Other results show that in French-speaking African countries, there is (1) unidirectional causality from investment (GFCF) to FD1, FD2 and FD3 only in the short run and (2) bidirectional causality between investment and FD2 in the short run.
Although contrary to expectations, these results are not unusual (see, among others, Shan et al., 2001 Nyasha & Odhiambo, 2015). However, the study failed to find a causal relationship between economic growth and financial development when FD2 is used as a proxy. The results reported in Table 3.4, Panel B, show that English-speaking African countries exhibit a distinct unidirectional causal flow from economic development to financial development only in the short run when FD1 and FD3 are used to proxy financial development (Panel B, Models 1 and 3). However, when FD2 is used to measure financial development, there is short- and long-run unidirectional Granger-causality from financial development to economic growth. The results reported in Table 3.4, Panel A, reveal that in French-speaking African countries, there is unidirectional Granger-causality from economic growth to financial development but only in the short run when financial development is proxied by FD1 and FD3. ∗ ∗∗ and ∗∗∗ denote significance at 10%, 5% and 1% respectively. ECT coefficients and t-stats in parentheses for long-run relation. Chi-squared statistics and P-values in parentheses for short-run. Note: Null hypothesis of ‘no Granger causality’ between dependent and independent variable. Panel B: English-Speaking African Countries Panel A: French-Speaking African Countries
0 kommentar(er)
