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

policy shock.

We furthermore either impose that interest rates do not fall or, alterna-
tively, impose that they do not fall more than the interest rate in the other
country. These two choices reflect two prototypes of the game played be-
tween the two monetary authorities. Consider a surprise rise in the interest
rate of some country A of interest, which may be contemporaneous with a
surprise rise of interest rates in some other country B. With the first inter-
pretation, one can think of the central bank of country A as a Stackelberg
leader: if both interest rates go up, then it and not necessarily the central
bank of country B was the ultimate cause. In the second, the central bank
of country A is viewed as a follower: deviations from the interest rate set by
the other central bank of country B are viewed as the key surprise, to which
the central bank in country A reacts quickly.

In principle, the game played between two monetary authorities could
be rather complicated. There is really no good reason to a priori rule out
e.g. a game in which central banks alternate as to who is using the higher
interest rate. And it is easy to think of considerably more complicated games.
We view the two possibilities investigated here as two particularly plausible
benchmarks. To impose the commonly held view that US monetary policy
is leading and other countries are following, we shall impose the restriction
i ≥ 0 when identifying US monetary policy shocks, but i* — i ≥ 0, when
identifying foreign monetary policy shocks. For the BIG VAR, one can also
think of these choices as reflecting a causal ordering of the monetary policy
choices, with the US ordered first. The identification restrictions imposed
are summarized in table 1.

For the restriction horizon, we have used k = 0, . . . , K = 11, i.e. one
year, throughout. Choosing shorter restriction leaves too much room for
spurious effects, while imposing a longer horizon imposes an implausibly
long duration for the liquidity effect. Uhlig (2005) contains some discussions

15

<<   15   16   17   18   19   20   21   >>

More intriguing information