Nonlinear and linear DSGE

Dear dynare team, could you please tell me what is the difference between Nonlinear DSGE models and Linear DSGE models? Thanks a lot. My best.

Hi,

In general a DSGE model is a non linear set of stochastic difference equations. These equations are the necessary conditions of the model’s agents optimization problems. Tackling directly the stochastic non linear equations is very difficult, that’s why we often replace the original problem by an approximation, most often a first order linear approximation around the deterministic steady state (ie we replace each non linear equation by its tangent at the deterministic steady, using a first order Taylor approximation). Depending on what you ask to the model, this approximation will be more or less accurate.

Best,
Stéphane.

1 Like