Problems in steady state linearization

Stochastic simulations
1

I am new using Dynare, so I could appreciate any help.
I have some problem in Dynare when I state de model(linear) block, it does not recognizes the variables I am putting even though I defined it in the first block.

This is the fileAmpliacion.mod (2.6 KB) . If anyone could help me I would appreciate it very much. Thanks.

2

You are referring to objects which are not defined. Dynare does not know what “ce_ss” and the other “_ss” terms are without you telling it. I suppose these are the coefficients from linearizing your nonlinear model around its non-stochastic steady state, which are necessarily functions of the steady state values for the endogenous variables. You must calculate those somehow, either analytically or numerically.

See the Dynare Reference Manual’s section on Model Declaration and Steady State, the discussion on the Dynare forums here, or Pfeifer (2018) A Guide to Specifying Observation Equations for the
Estimation of DSGE Models available here.

1 Like
3

Thanks for your response, you are completely right. Those “_ss” are variables in steady state that result from the log-linealization, but I do not know if that should be there. Because of this two problems arise:

  1. I have tried to put the steady state in the model block in the “#” form, but it says that I can not define more than once the “_ss” variables, so I am not sure which ones to put.
  2. I am trying to replicate a 2001 paper An Estimated Canadian DSGE Model with.pdf (236.8 KB), but I have certain problems trying to state that the estimation must be done by maximum-likelihood.

I appreciate any help, thanks.

4

Before I respond to your specific questions, I should say I get the sense that are not looking through the reference manual, other supporting documents, or previous forum posts before posting. I would strongly recommend trying to help yourself first by carefully reviewing these documents before posting. Otherwise you are wasting your own valuable time waiting on people to respond when the answers are easy to find on your own.

I have tried to put the steady state in the model block in the “#” form, but it says that I can not define more than once the “_ss” variables, so I am not sure which ones to put.

Without a .mod file I can only guess at what you have done. You can not include the “_ss” as variables in the var block and then again as model locals in the model block. See Listing 3 in A Guide to Specifying Observation Equations for the
Estimation of DSGE Models and the discussion surrounding it for an example of proper implementation of this approach.

I am trying to replicate a 2001 paper An Estimated Canadian DSGE Model with.pdf (236.8 KB), but I have certain problems trying to state that the estimation must be done by maximum-likelihood.

I do not know what “certain problems” means. In general to estimate using ML you just choose the appropriate mode compute algorithm, as outlined in the Dynare documentation on Estimation.