This paper evaluates VAR models designed to analyse the monetary polic
y transmission mechanism in the United States by considering three iss
ues: specification, identification, and the effect of the omission of
the long-term interest rate. Specification analysis suggests that only
VAR models estimated on a single monetary regime feature parameters s
tability and do not show signs of mis-specification. The identificatio
n analysis shows that VAR-based monetary policy shocks and policy dist
urbances identified from alternative sources are not highly correlated
but yield similar descriptions of the monetary transmission mechanism
. Lastly, the inclusion of the long-term interest rate in a benchmark
VAR delivers a more precise estimation of the structural parameters ca
pturing behaviour in the market for reserves and shows that contempora
neous fluctuations in long-term interest rates are an important determ
inant of the monetary authority's reaction function. (C) 1998 Elsevier
Science B.V. All rights reserved.