New to Dynare (and DSGE!)

Hello all, I am a grad student attempting to use Dynare to speed me along with my dissertation. Please forgive me if my question(s) are remedial or are answered elsewhere.

My current issue is that I am attempting to use Dynare 4.1.1 to solve a model that I have already reduced and linearized. In my calculations, I’m using the format for the variables:

y(t) = log Y(t) / Yss

The manual states that if I add (linear) after the word model in the model section, I can solve a linearized model in Dynare. However, it states that I should either double the variable for deviation from steady state or type exp(repeatedvariable) if I want Dynare to log linearize. My instinct tells me that for my notation I could use (linear) and simply type ‘log (repeatedvariable)’ for each variable, but I’m not certain. Am I on the right track?


If your model is already loglinearized, i.e. your equations are written in terms of y(t) as specified in your equation, you just have to specify model(linear) and you should be fine. What do you mean with doubling?
If you put an exp() before your variable, Dynare does a substitution to define all variables as their logs. But if you already log-linearized, you would do it twice.

Thank you, that is what I thought. When I said double I meant typing the variable name twice (e.g. yy instead of y). I’d like to post a follow up question, however. Since I’ve already log linearized my model, will Dynare be able to calculate steady states? That is, can I use the command ‘steady;’ in the initval section or do I need to specify the initial values myself?

Thanks again!