We explore methods for confronting model misspecification in macroeconomics
. We construct dynamic equilibria in which private agents and policy makers
recognize that models are approximations. We explore two generalizations o
f rational expectations equilibria. In one of these equilibria, decision-ma
kers use dynamic evolution equations that are imperfect statistical approxi
mations, and in the other misspecification is impossible to detect even fro
m infinite samples of time series data. In the first of these equilibria, d
ecision rules are tailored to be robust to the allowable statistical discre
pancies. Using frequency domain methods, we show that robust decision-maker
s treat model misspecification like time series econometricians. (C) 2001 A
cademic Press.