DSGE-Modelling : When Agents are Imperfectly Informed

DSGE-models have become important tools of analysis not only in academia but increasingly in the board rooms of central banks. The success of these models has much to do with the coherence of the intellectual framework it provides. The limitations of these models come from the fact that they make very strong assumptions about the cognitive abilities of agents in understanding the underlying model. In this paper we relax this strong assumption. We develop a stylized DSGE-model in which individuals use simple rules of thumb (heuristics) to forecast the future inflation and output gap. We compare this model with the rational expectations version of the same underlying model. We find that the dynamics predicted by the heuristic model differs from the rational expectations version in some important respects, in particular in their capacity to produce endogenous economic cycles