in 10-year bond rates and industry returns. To this end, the
following VAR model o f order K is estimated:
Dft = f t , + £ « * + + < d 3 )
k= 1 /c=1
o f f = f e + £ « _ * + j y S D f a + u«> (14)
k = 1 k=1
where D'jRt is the HTW-based d e ta il component o f the
series o f changes in 10-year bond yields at scale zj, is the
HTW-based d e ta il component o f the series o f returns o f the
industry considered a t scale r j. The number o f lags o f the
model, K, is chosen according to the Schwarz in formation
c rite rio n and Akaike information c rite rio n .In order to investigate whether industry equity returns
Granger cause movements in 10-year government bond
yields, we test whether the null hypothesis that all of the
(ft parameters are zero can be statistically rejected. And
vice versa, in order to ascertain whether changes in 10-year
bond rates Granger cause industry equity returns we test
whether the null hypothesis that all of the parameters
are zero can be statistically rejected. Both null hypotheses
are tested by using a F-test.
The findings of the Granger causality tests at different
time horizons, reported in Table 2, are broadly consistent
with those of the wavelet analysis outlined above in terms of
the multiscale nature of the interest rate-stock market link
and the industries with higher and lower levels of connection
with 10-year government bond rates. Thus, the evidence of
Granger causality is in general quite weak at shorter horizons.
However, the causality is dramatically strengthened at
long horizons. Significant bidirectional causal relationships
between movements in 10-year bond yields and industry
returns are found at longer horizons for most industries,
confirming that the association between 10-year bond yield
fluctuations and Spanish stock market increases with the
investment horizon.
อัตราพันธบัตรอายุ 10 ปีและผลตอบแทนอุตสาหกรรม in 10-year bond rates and industry returns. To this end, the
following VAR model o f order K is estimated:
Dft = f t , + £ « * + + < d 3 )
k= 1 /c=1
o f f = f e + £ « _ * + j y S D f a + u«> (14)
k = 1 k=1
where D'jRt is the HTW-based d e ta il component o f the
series o f changes in 10-year bond yields at scale zj, is the
HTW-based d e ta il component o f the series o f returns o f the
industry considered a t scale r j. The number o f lags o f the
model, K, is chosen according to the Schwarz in formation
c rite rio n and Akaike information c rite rio n .In order to investigate whether industry equity returns
Granger cause movements in 10-year government bond
yields, we test whether the null hypothesis that all of the
(ft parameters are zero can be statistically rejected. And
vice versa, in order to ascertain whether changes in 10-year
bond rates Granger cause industry equity returns we test
whether the null hypothesis that all of the parameters
are zero can be statistically rejected. Both null hypotheses
are tested by using a F-test.
The findings of the Granger causality tests at different
time horizons, reported in Table 2, are broadly consistent
with those of the wavelet analysis outlined above in terms of
the multiscale nature of the interest rate-stock market link
and the industries with higher and lower levels of connection
with 10-year government bond rates. Thus, the evidence of
Granger causality is in general quite weak at shorter horizons.
However, the causality is dramatically strengthened at
long horizons. Significant bidirectional causal relationships
between movements in 10-year bond yields and industry
returns are found at longer horizons for most industries,
confirming that the association between 10-year bond yield
fluctuations and Spanish stock market increases with the
investment horizon.
การแปล กรุณารอสักครู่..