We live in uncertain times. We do not know what will happen tomorrow, next quarter, or next year not only because the world and economy experience unexpected events but also because the world keeps changing. People, both as consumers and as entrepreneurs, keep learning about this ever-changing environment. Economic policy should take this into account. At the annual meeting of Society for Computational Economics we presented the paper “The dangers of macro-prudential policy experiments: initial beliefs under adaptive learning” that tackles this very issue. Importantly, using the example of macro-prudential policy, we show that policymakers should not only take into account that people learn but also that it is a lengthy process. Despite the conference being virtual, we benefited from many comments.

While this paper was predominantly theoretical we also presented our results on taking models with learning to the data. Our empirical strategy involves combining theoretical specification of the learning process with the computational power of numerical maximization. Our results shed light on how quickly people learn and how much they revise their perception about the world.