I linear models we know that residual of equations should be zero. But my question is that in non-linear models residuals of equations should be zero or not?
I don’t seem to understand the question. Do you mean theoretically? Of course the residuals of equations should be zero, no matter how the equation looks like, it just means that the left hand side has the same value as the right hand side.
Or do you mean specifically in dynare when you have non-linear equations and it shows you the residuals of your equations in steady state (the resid command)? Even then the residuals should be zero (or numerically very close), which means that you have provided the correct steady state, that was either given via the steady state block or initial values.
My question is that in non-linear DSGE models residuals of equations should be zero or not?
Is there any different between linear and non-linear DSGE models about residuals??
I know in linear models residuals of equations should be zero, but what about non-linear DSGE models??
Like I said, no difference. They should also be zero.
The residuals are the “error” in your steady state computation. They should always be zero. The difference is in the actual steady state values: for linearized models, the steady state values are generally zero. For nonlinear models, they are regularly non-zero .
Thanks so much professor.