Linearization
Posted: Sat Apr 06, 2013 7:27 pmHello all,
Well i am facing a simple problem. I want to log-linearize a model so i put all my variables in exp, e.g c = exp(c) and so on. In initial value i then enter c = log(2.5), 2.5 being the steady state value of c and so on. This all is fine but i had a few questions:
1) If i dont transform the variables, can i still do a second order taylor expansion around the steady state? In this case how do i ask dynare to perform the linearization for me?
2) How do i deal with zero steady states?
3) Even if i do the linearizations with the log transformation, how can i get deviations from steady state instead of percentage deviations?
I have attached my codes for the same model:
myrbc1a is a simple model with exp and the code runs fine - the only thing i am concerned with is that i have computed initial values by hand and this could be a pain for larger models.
myrbc1b is the code i tried changing. In this code i tried using the solve_algo command but i think i made a mistake. Any examples of such codes would be appreciated!
Well i am facing a simple problem. I want to log-linearize a model so i put all my variables in exp, e.g c = exp(c) and so on. In initial value i then enter c = log(2.5), 2.5 being the steady state value of c and so on. This all is fine but i had a few questions:
1) If i dont transform the variables, can i still do a second order taylor expansion around the steady state? In this case how do i ask dynare to perform the linearization for me?
2) How do i deal with zero steady states?
3) Even if i do the linearizations with the log transformation, how can i get deviations from steady state instead of percentage deviations?
I have attached my codes for the same model:
myrbc1a is a simple model with exp and the code runs fine - the only thing i am concerned with is that i have computed initial values by hand and this could be a pain for larger models.
myrbc1b is the code i tried changing. In this code i tried using the solve_algo command but i think i made a mistake. Any examples of such codes would be appreciated!