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

36,18-80”. Rather we wish to leave the response of the exchange rate as
agnostically open as possible, since that is the variable of focus.

Following the arguments of Sims and Uhlig (1991), we use a thoroughly
Bayesian procedure. Thus, we provide posterior distributions regarding the
parameters of interests - like the time and the size of the peak response -
rather than robust 90 percent coverage bands. The sign restrictions imposed
take center stage in this paper, as they are key to identification and should be
subjected to debate and scrutiny. We also exploit this Bayesian perspective
to ask questions concerning the risk a Bayesian investor faces when betting
on the UIP violations in 2. We calculate an implied Sharpe ratio and compare
it to Sharpe ratios conventionally observed on e.g. equity markets.

As a benchmark and similar to Faust and Rogers (2003), we apply our
identification method to the VAR specifications used by Eichenbaum and
Evans (1995) and Grilli and Roubini (1995, 1996). When using conventional
identification methods increasing the number of variables in a VAR implies
a rising number of assumptions which become increasingly difficult to jus-
tify. Adding possibly important variables, while using identification via sign
restrictions, is typically rather straightforward, however: thus we do.

Analyzing the US-German, the US-UK and the US-Japanese bilateral ex-
change rates, we now avoid the price puzzles of Eichenbaum-Evans or Grilli-
Roubini by construction. Nonetheless, the delayed overshooting puzzle and
the forward discount as well as the exchange rate puzzles are still there and
they are sizeable. However, the quantitative features are different. The peak
appreciation happens during the first year after the shock for the US-German
and the US-UK pair, and during the first two years for the US-Japan pair.
This is consirably faster than the three-year horizon found by Eichenbaum-
Evans. The forward discount bias puzzle comes with risk. The implied
Sharpe ratio for a Bayesian investor can be as high as 2.5, which is five times
as high as the annual Sharpe ratio for US stock markets. The exchange rate

<<   4   5   6   7   8   9   10   >>

More intriguing information