Learning, Information and Heterogeneity

Liam Graham (University College London) 

Riccardo Faini CEIS Seminars

Riccardo Faini CEIS Seminars
When

Friday, October 21, 2011 h. 12:00-13:30

Where
Aula B - Primo piano
Description

Most DSGE models assume full information and model-consistent expectations. This paper relaxes both these assumptions in the context of the stochastic growth model with incomplete markets and heterogeneous agents. Households do not have direct knowledge of the structure of economy or the values of aggregate quantities; instead they form expectations by learning from the prices in their market-consistent information sets. The economy converges quickly to an equilibrium which is similar to the equilibrium with model-consistent expectations and market-consistent information. Learning does not introduce strong dynamics at the aggregate level, though more interesting things happen at the household level. At least in the context of this model, assumptions about information seem important for aggregates; assumptions about the ability to form model-consistent expectations less so.

 VIEW PAPER