 |
Predicting Inflation With the Term Structure SpreadSharon Kozicki |
Abstract It is tempting to interpret empirical evidence in a number of recent studies as
suggesting that term structure spreads help predict future inflation over moderate
horizons of 3 to 5 years. This paper argues that common measures of the predictive power
of the term structure spread for future inflation are misleading. In particular, R2s for
estimated inflation-change equations can drastically overstate the predictive power of
spreads. The paper explains why the overstatement is likely to be particularly large in
countries whose monetary authorities have strong reputations for credibly targeting a
stable inflation rate. Results from an empirical analysis of data from eleven
industrialized countries suggest that the level of the short-term real rate may be more
useful for predicting inflation than the term structure spread, possibly because changes
in short-term real rates provide clearer measures of changes in the stance of monetary
policy. Sharon Kozicki is a senior economist at the Federal Reserve Bank of Kansas City.Barak Hoffman provided excellent research assistance. The author is grateful to Vance
Roley, seminar participants at the Federal Reserve Bank of Kansas City and the Bank of
Canada, and attendees of the November 1997 Federal Reserve System Committee Meeting on
Macroeconomics for comments and suggestions. Special thanks to Timothy Morley for
assistance in compiling the data. The views expressed in this paper are those of the
author and not necessarily reflect the views of the Federal Reserve Bank of Kansas City or
the Federal Reserve System.
|