OLS--Learning in Non-Stationary Models with Forecast Feedback

In this study we consider a linear model with forecast feedback in which boundedly rational agents are learning the parameter values of the rational expectations equilibrium by the OLS learning procedure. We show strong consistency of the OLS estimates under much weaker assumptions on the involved time series than the ones usually employed. This result extends the boundedly rational learning approach to models including non-stationary time series, like processes with polynomial trends or unit root autoregressive processes, and indicates that the idea that agents can learn only stationary rational expectations equilibria is misleading.