1.
Introduction
This paper investigates whether exchange rates respond to only the surprise component of actual
monetary policy changes and whether failure to disentangle the surprise component from the monetary
policy news announcement leads to an underestimation of the impact of the news, or even to a false
acceptance of the hypothesis that the news has no impact on exchange rates. In addition, we examine
whether the exchange rate price adjustment associated with the monetary policy change is instantaneous
or delayed.
Recent empirical contributions by Andersen, Bollerslev, Diebold and Vega (2003), Faust, Rogers,
Swanson and Wright (2003), Evans and Lyons (2005) and Simpson, Ramchander and Chaudry (2005)
find that exchange rates react to monetary policy surprises, i.e. to the unexpected component of a change
in the monetary policy stance.1 This finding is in line with the predictions of the standard asset pricing
models of exchange rate determination where exchange rates are viewed as forward-looking asset-prices.2
A common characteristic of all these studies, however, is that they focus their analysis of news
effects on only one type of variable, namely the surprise component of a news announcement.3 Rather
than focusing on only the surprise component of news which, by construction, is a simple linear function
of the actual announcement and the expected component, our analysis makes use of the three variables
(the surprise component, the expected component, and the actual announcement) separately.4
In doing so we are able to address not only whether exchange rates respond to the surprise
component of news but also, at the same time, whether exchange rates respond to the expected component
1 These papers as well as ours are related to an older literature on the effectiveness of anticipated versus
unanticipated monetary policy on output and unemployment. While the recent literature employs survey data or
market-based measures of expectations for distinguishing between anticipated and unanticipated monetary policy
innovations, the older literature relies on output from econometric models of these policy innovations. Important
older contributions include Barro (1977), Barro and Hercowitz (1980) and Mishkin (1982).
2 Engel and West (2005) revisit the “asset-market” approach to exchange rate determination and conclude that
“exchange rates and fundamentals are linked in a way that is broadly consistent with asset pricing models of the
exchange rate”.
3 Recent papers by Bernanke and Kuttner (2004), Craine and Martin (2003) and Flannery and Protopapadakis (2002)
follow the same approach in regards to using only the surprise element (the difference between announcement and
expectation) of news as the explanatory variable when measuring the response of the stock market to monetary
policy changes and other macroeconomic news.
4 We are only investigating the impact of US monetary policy news and, therefore, we do not rescale our news
variables by their respective standard deviations.