Dear Professor Jpfeifer,
When I see some dynare codes written by others, I have found that they like to transform the original nonlinear equations into exponential forms. May I ask what is the advantage of this?
For example,
y=A*(ksik(-1))^alphah^(1-alpha);
exp(y)=exp(A+alpha*(ksi+k(-1))+(1-alpha)*h); and in steady_state_block
y=log(y_ss)