This paper tests a necessary condition for the neutrality of money in
a framework that imposes only weak restrictions on the money supply pr
ocess. It extends Bernanke's [1986] work by weakening the set of just-
identifying restrictions and by providing a statistical test of the ov
eridentifying restrictions. Instead of specifying a structural model t
o identify primitive shocks, I deduce the impact effects of structural
money shocks under the neutrality hypothesis and then test whether th
e system maintains neutrality as it propagates these impact effects. T
he tests reject neutrality for both the M1 and the monetary base.