ABSTRACT
Another complementary approach to bounded rationality is heterogeneous expectations models as, for example, introduced in Brock and Hommes (1997, 1998) and Branch and Evans (2006). Agents endogenously switch between different forecasting rules, ranging from simple heuristics to more sophisticated strategies, based upon their relative performance. Notice that in both adaptive learning and heterogeneous switching models, the learning is endogenous and agents will adapt to policy changes, so that these models, at least to a first order approximation, mitigate the Lucas critique.
