-20-
In order to investigate the effects of global and idiosyncratic structural shocks on
national variables, we estimate equation (3), using a separate VAR for each country or
region. The variables in levels are again “re-constructed” by calculating the cumulative
sum of the standardized national variables. This is necessary to be fully consistent with
our approach on the global level. Since a national economy can be hit by both global and
idiosyncratic shocks, one has to distinguish between these two effects. We do so by
regressing εxt which consists of global and idiosyncratic disturbances on the global
structural shocks (ut) derived from equation (5). On the basis of the obtained
idiosyncratic components, restrictions can be imposed for exact identification on a
national level (Bagliano and Morana, 2009). Again, we refrain from using a Cholesky
decomposition and implement structural relationships between the respective variables as
follows:
Γ1 |
0 |
0 |
0 |
0 |
0 |
0 ^ |
’ VYR ~ |
YR | ||
0 |
1 |
0 |
0 |
0 |
a26 |
0 |
vP |
zP | ||
0 |
0 |
1 |
0 |
0 |
0 |
0 |
vHP |
ZHP | ||
(6) |
a 41 |
a42 |
0 |
1 |
a45 |
a46 |
0 |
vM' |
= |
zM |
0 |
0 |
a53 |
a54 |
1 |
a56 |
0 |
VSR |
zs | ||
a 61 |
a62 |
a63 |
a64 |
a65 |
1 |
a 67 |
vER |
z | ||
. a 71 |
a72 |
a73 |
a74 |
a75 |
a76 |
1 |
. vSP. |
.z'P. |
with v as the vector of errors in the reduced-form equations and z as the
idiosyncratic structural disturbances. Our restrictions on a national level largely resemble
the assumptions we made for global variables. For example, GDP (YR) and house prices
(HP) are not influenced contemporaneously by other national variables. The inflation
figure (PI) equals consumer prices and is affected only by the real effective exchange rate
(ER) at the same time. The latter variable and share prices (SP) are assumed to respond
contemporaneously to all other variables.
Before presenting the structural impulse response functions, some cautionary
remarks in interpreting the results seem to be appropriate. The identification pattern in
equation (6) is used for all countries included in the sample, although they differ
markedly in economic size. In addition, various institutional settings in national monetary