New Evidence on the Puzzles. Results from Agnostic Identification on Monetary Policy and Exchange Rates.

There, he has argued against the large-scale models popular at that time
by stating that “the connection between ... models and reality - the style in
which ’identification’ is achieved for these models - is inappropriate, to the
point at which claims for identification in these models cannot be taken seri-
ously.” The advantage of reduced-form vector autoregressions is that it does
not need these incredible identification restrictions at all, but for structural
vector autoregressions, some identification for the structural shocks is obvi-
ously needed. In line with the quote by Sims, we believe that some of the
work using structural VARs similarly are in the danger of employing identifi-
cation restrictions which are not a priori plausible, while the sign restriction
approach allows a much more direct connection with believes derived from
theory and the empirical application.

We thus follow the lead of Faust and Rogers (2003), who likewise use sign
restrictions to narrow down the set of reasonable identifications for monetary
policy shocks. In contrast to these authors, we view the sign restrictions as
a means to identifying monetary policy shocks rather than an aid in ro-
bustness analysis. Most of the Faust-Rogers restrictions are on-impact only,
complemented with very few and hand-selected extra sign restrictions at later
horizons. By contrast, we impose sign restrictions on the impulse responses
for key monetary policy variables for several periods after the shock. We
view this as plausible. For example, to implement the benchmark view that
inflation slows down after a monetary tightening, an on-impact drop in prices
followed by a much larger subsequent rise ought to be ruled out as a possible
response to a contractionary monetary policy shock. This is exactly what we
do here. Thereby, we can narrow down the range of possible monetary policy
shocks considerably, as has already been argued in Uhligs (1998) discussion
of Faust (1998), and indeed this will turn out to be true here as well. On
the other hand, we do not impose Faust-Rogers-type shape restrictions that
“the exchange rate response falls between lags 1-2,2-3,3-4,4-6,6-12,12-18,18-

<<   3   4   5   6   7   8   9   >>

More intriguing information