Monetary Policy and Radical Uncertainty

In a world of radical uncertainty the frequency distributions of economic variables deviate from the normal distribution and typically exhibit fat tails. We show that this feature is obtained in simple models where agents have cognitive limitations and fail to understand the underlying model. Although the model is simple, we obtain great complexity. We analyse the implications for monetary policy. We show that in such models the central bank bears a much greater responsibility to stabilize an otherwise unstable system than in mainstream models that assume Rational Expectations. We also question the use of impulse responses to exogenous shocks when the distribution of these impulse responses is not normal.