Heterogeneity and misspecifications in learning

Abstract

In this paper we consider a linear, stochastic, univariate, forward looking model with one lag under adaptive heterogeneous learning. The system is populated by two different types of agents who learn through recursive least squares techniques the parameter values in their forecasting models. The two groups are constrained to have different information sets, one being always a subset of the other. We analyze convergence of these two interacting learning processes under different specifications of the forecasting model, and in one case we find that an heterogenous expectations equilibrium emerges.