Recent monetary history has been characterized by monetary authorities that appear to shift periodically between distinct policy regimes associated with higher or lower average rates of money creation. As policy regimes are not directly observable and as the rate of monetary expansion varies for reasons other than regime changes, the general public must form beliefs over current monetary policy based on historical realizations of money growth rates. Depending on the parameters governing the behaviour of m onetary policy, beliefs (and therefore inflation forecasts) may evolve very slowly in the wake of actual regime changes, thereby exacerbating the costs of a disinflation policy. The quantitative importance of slowly adjusting beliefs is evaluated in the c ontext of a computable general equilibrium model.
