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

updated data set.

Figure 18 compares the impulse responses of all the variables (and vari-
ables derived from it) for the three specification and the US-UK pair: we
have actually chosen the pair which should be most favorable to the Grilli-
Roubini identification, based on the rather orthodox looking results of the
third line in figure 15. Note again, however, that there is a huge price puzzle
for the Grilli-Roubini identification: foreign prices keep on rising for a long
time before returning to zero, following a contractionary foreign monetary
policy shock. For our sign restriction approach, this price puzzle is avoided
by construction.

There are more impulse response functions available for the sign restric-
tion approach both for the GR and the BIG specification. They are shown
in figures 19 as well as 20 (two pages) for completeness and to emphasize,
that our identification strategy is reasonable.

The last line of figure 14 concerns the variance decomposition of the
movements of exchange rates, which we discuss next in subsection 4.3.

4.3 Monetary Policy Shocks and Exchange Rate Volatil-
ity

Figure 21 contains the variance decomposition for the exchange rate move-
ments explained by a US monetary policy shock, while figure 22 contains
the fraction of exchange rate variation explained by foreign monetary policy
shocks.

US monetary policy shocks account for somewhere between 2 and 10
percent of the exchange rate fluctuations at the median estimate, independent
of the horizon and country, for both sign restriction specifications. The
number is smaller at most horizon for the Eichenbaum-Evans identification
except for a rather sharp peak 6 to 8 months after the shock.

For foreign monetary policy shocks, figure 22 delivers a similar result with

21

<<   21   22   23   24   25   26   27   >>

More intriguing information