Advances in Risk Management

HELENA CHULIÁ ET AL. 319

1

0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

10 19 28 37 46 55

Days

A. A negative shock in the S&P500

AVIRF S&P500

64 73 82 91 100 1

0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

10 19 28 37 46 55

Days

B. A negative shock in the IBEX35

AVIRF S&P500

64 73 82 91 100

1

0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

10 19 28 37 46 55

Days

C. A negative shock in the S&P500

AVIRF IBEX35

64 73 82 91 100 1

0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

10 19 28 37 46 55

Days

D. A negative shock in the IBEX35

AVIRF IBEX35

64 73 82 91 100

Figure 16.3AVIRF to negative unexpected shocks from the

VAR-asymmetric BEKK. Total period (dashed lines display

the 90% confidence interval)

There is no significant impact coming from positive shocks (Figures 16.2-D,
16.4-D and 16.6-D) and, after September 11, negative shocks have a signifi-
cant effect that takes a very long time to die out due to its persistence. This
asymmetric effect can explain why, when estimation results where analyzed
in the previous section, we concluded that the IBEX35 was not affected by
its own unexpected shocks. This occurs when unexpected shocks are taken
as a whole. In the case of the S&P500, there also exists an asymmetric effect,
which is observed in all analyzed periods.
The AVIRF figures depend on the estimated coefficients coming from
matrix G. In all periods, the figures confirm the existence of an asymmetric
effect of positive and negative shocks on own conditional variances (related
to coefficientsg 11 andg 22 ). A negative shock has a stronger impact than a
positive one. However, this asymmetric effect does not exist in the case of
shocks coming from another market (coefficientsg 12 andg 21 ).