The author sets up a simplistic agent-based model where agents learn with reinforcement observing an incomplete set of variables. The model is employed to generate an artificial dataset that is used to estimate standard macro econometric models. The author shows that the results are...