We develop and implement a limited information diagnostic strategy for assessing the plausibility of monetary business cycle models. Our strategy focuses on a model's ability to reproduce empirical estimates of an actual economy's response to monetary policy shocks. A key input to this diagnostic is a univariate time series representation of the response of money to a shock in monetary policy. We find that a monetary policy shock has only a small contemporaneous effect on the monetary base and M1. Its primary effect is to signal future movements in the money supply. We implement our diagnostic strategy on a limited participation model of money which stresses the importance of credit market frictions in the monetary transmission mechanism.