Page 1 of 1

Steady State Calculation

PostPosted: Thu Jul 28, 2011 12:42 pm
by jhkemp
Hey,

The attached model gives me the following error when trying to run: STEADY: convergence problems

I understand that this could be because the values in initval are wrong, or that the model has no steady state.

What is the best/simplest way to calculate the steady state values for a model if not doing it by hand (so as to use these as inputs in initval)?

Re: Steady State Calculation

PostPosted: Sat Jul 30, 2011 4:52 pm
by wjgatt
There are many posts on this subject, do a search.

I'm no expert, but these are the options I consider when working with a model and when having convergence problems:

1) You could solve by hand the simple variables, such as the steady state interest rate. Also make sure that some variables are not initialised to 0 if they form part of a multiplication or division.

2) Use the 'resid(1);' option just before 'steady;' to see the residuals for each equation; for example you have large residuals in eqns 3,4 and 6. Use this info to change the initial values or the parameter values.

3) You can also choose between the different solvers available, refer to the user guide.

4) If these don't work I would suggest using the program developed and kindly distributed by a member (mathwiz): 'dynareFindSteadyState', you basically copy your model up to the model specification in another mod file, and let the program find the exact values for you. See the link:

viewtopic.php?f=2&t=1424


If that also doesn't work, you might have chosen bad values in your calibration, change the ones which are more controversial in the literature first.

I used dynareFindSteadyState on your model, but it could not find the solution. You might want to consider option 2.