View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Economic SYNOPSES
short essays and reports on the economic issues of the day
2005 ■ Number 20

Foreign Exchange Rates Are Predictable!
Hui Guo
odern economic theory of foreign exchange rates stipulates that the Deutsche mark/U.S. dollar rate, for example, is equal to discounted future fundamentals—e.g.,
aggregate income, interest rates, and monetary aggregates—in both
the United States and Germany. A substantial portion of the variation in these fundamental macroeconomic variables is predictable
across time; therefore, fundamentals should provide important
information about future movements in exchange rates. In an
influential paper, Meese and Rogoff (1983), however, find that a
simple random walk model, in which the forecasted value is the
most recent realization, outperforms various forecasting models,
including those using economic fundamentals as predictors.1
Meese and Rogoff ’s result has inspired numerous empirical
investigations of exchange rate predictability, and their conclusion
has proven to be strikingly robust after 20 years of fresh data and
intensive academic research. In light of seemingly compelling
evidence, some recent authors argue that exchange rates are indeed
unpredictable—possibly because some shocks have a permanent
effect on economic fundamentals. In particular, if people discount
the future very little relative to the present, then exchange rates
could follow a process close to a random walk.
Other economists, however, argue that exchange rates are
predictable and that existing empirical studies suffer from various
misspecifications. For example, some crucial fundamental determinants of exchange rates may have been omitted. Also, many
macroeconomic variables are subject to periodic revisions; therefore, the current vintage data, which have been commonly used
available to investors at the time of forecast. To address these
issues, Guo and Savickas (2005) propose using financial variables, which are broad measures of business conditions and
never revised, to predict exchange rates.2
Guo and Savickas find that a measure of U.S. aggregate idiosyncratic volatility (IV) is a strong predictor of the exchange rates
of the U.S. dollar against major foreign currencies, especially at
relatively long horizons. An idiosyncratic shock to a stock is the
part of the stock return that is not explained by asset pricing
models. To measure IV, Guo and Savickas first estimate idiosyncratic shocks to all (U.S.) common stocks included in the CRSP
(Center for Research in Security Prices) database; they then
aggregate the realized variance of idiosyncratic shocks across
stocks using the share of market capitalization as the weight.
The accompanying chart plots IV from the last quarter of
each year (in natural logarithms, solid line) along with one-year-

M

ahead changes (December 31 to December 31 of the following
year, dashed line) in the Deutsche mark/U.S. dollar rate over the
period 1973 to 1998 and the Euro/U.S. dollar rate over the period
1999 to 2003. The chart reveals a strong positive relation between
IV and changes in the price of the U.S. dollar over the next year.
For example, recent depreciation of the U.S. dollar was preceded
by a sharp decline in IV in the year 2001. Overall, IV accounts
for more than 30 percent of the variation of the Deutsche mark/
U.S. dollar rate; IV also outperforms the random walk model in
out-of-sample forecasting.
The forecasting power of IV is consistent with economic theory. In particular, many early authors have argued that IV is a
proxy for the dispersion of shocks across different sectors; and a
high level of dispersion induces costly sectoral resource reallocation, which reduces output and employment. Indeed, Guo and
Savickas show that IV is a strong predictor of GDP growth, fixed
private business investment, and unemployment rates. Moreover,
they find that a measure of aggregate IV constructed using
German stock price data is also positively related to future dollar
prices of the Deutsche mark. Therefore, although we cannot
completely rule out the possibility of data mining, the forecasting
power of IV appears to provide support for the conjecture that
economic fundamentals are important determinants of foreign
exchange rates. ■
1

Meese, Richard A. and Rogoff, Kenneth. “Empirical Exchange Rate Models of
the Seventies: Do They Fit Out of Sample?” Journal of International Economics,
February 1983, 14(1), pp. 3-24.

Log IV vs. Changes in Deutsche Mark (1973-98)
and Euro Rates (1999-2003) One Year Forward
Changes in Exchange Rates (dashed line)
0.3

Log IV (solid line)
–2

0

–0.3
1973

–3.5

1978

1983

1988

Views expressed do not necessarily reflect official positions of the Federal Reserve System.

research.stlouisfed.org

1993

1998

–5
2003