The effects of monetary policy on real economic variables have been debated for some time. This debate became more intense after the discovery of the Phillips curve which appeared to show a stable trade-off between inflation and unemployment. This curve in its original form has now been abandoned and debate has centered around the question of a short run trade-off. It is this question, do short-run trade-offs exist and if so, why, and what affects their length, that this dissertation addresses.

Chapter II explores this question in a model where the Federal Reserve does not have full credibility among all the agents in the economy and where beliefs are endogenous. It is shown that when the Federal Reserve announces a new monetary policy rule, temporary nonneutrality of money can result if some agents are skeptical of the Fed's intentions to follow the announced rule or if some agents merely believe some other agents are skeptical whether or not they truly are. The magnitude of the trade-off depends on the proportion of agents who are skeptical and how different the old and new rules are. The length of time the trade-off exists depends on how skeptical the agents are. The more skeptical they are, the longer it takes the Fed to convince these agents that it is following and will continue to follow the announced rule.

Chapter III develops an empirical model to determine if the evidence supports the existence of short-run trade-offs in general and the credibility implications in particular. A Bayesian Vector Autoregression is used for the estimations. It is shown that short-run trade-offs do exist and do vary in length and magnitude. These variations are related to the implications of the credibility theory. It is found the degree of skepticism and the proportion of agents who are skeptical could have caused these short-run trade-offs to vary in length and magnitude.