Solving for Steady State Analytically

Steady state
1

Greetings Dynare Experts,

I am solving a model for steady state analytically. I am able to write two equations in terms of two variables C and Y (after couple substitutions ) and then use fsolve to solve for the two variables in two unkowns. After that I willl be able to determine all other variables.

My concern that when I used fsolve to get C and Y I have not used their relevant closing condition equations….Shouldn’t I substitute first closing conditions equations and wage clearing conditions in all equations before I solve for the two variables and substitute them in the rest ?

Thank you for your help

2

That is hard to tell. Often, the budget constraint is required to solve for consumption and output. But there are cases where that is not required. You can always try whether your computed steady state solves the model. If there is a residual in the budget or resource constraint, you know you made a mistake.

1 Like
4

I have checked the forum and saw that reducing the tolerance for solving steady state is

options_.dynatol.f=5e^-3
options_.solve_tolf= 5e^-3

Are these tolerances for Function only ? Shouldn’t we impose tolf and tolx together ?

Do these tolerances apply for solve_algo=0 only or also for solve_algo=1 too?

Best,

5
  1. options_.dynatol.f is for the perfect foresight solver, not steady state finding.
  2. Yes, typically you want to set tolf and tolx together.
  3. solve_tolf, i.e., the options set by steady(tol) governs all solve_algo.
6

Thank you alot for your help. Can I ask one more thing ? Is there any dsge models that uses multiple Philips curves? namely, food and non food Philips curves?

7

There are surely models with multiple sectors out there that feature multiple Philips curves.

8

Could you refer to any examples or replications in dynare ?

9

There should be many examples in the Macroeconomic Model Database.

10

Thank you