I’m now handling with a non-linear DSGE model. I need to use Bayesian Estimation to estimate some parameters. However, these parameters will also be used to calculate the steady state value for some endogenous vars.

For example,in a non-linear model, to calculate the var consumption, I need to set the consumption habitat hh= 0.6 (or other number) first. But I also need to use Bayesian Estimation to estimate the consumption habitat hh.

Currently, I set the hh=0.6 to calculate the steady state for the model, and then put the steady state values into the mod.file, then run the Bayesian Estimation in the mod.file, which means the hh used to calculate the steady state values is not the one after Bayesian Estimation (so it might not be so capable of reflecting the data performance) but one value I set by merely referring to some literature. I write an external file using a very widely used method to calculate the steady state for the model. ()

//******************************************************************//

SO WILL it be possible to set the consumption habitat parameter to be the one after Bayesian Estimation to calculate the steady state for the model?

Will it be possible to use the parameters values(after bayesian estimation ones) to calculate steady state?

//******************************************************************//

What I perceive is, to do the Bayesian Estimation, I need to know the steady state, which means you have to set the parameters needed for the calculation first; meanwhile, I want to use the parameter’s bayesian estimation value to calculate the steady state.

Is it nearly impossible to meet the requirement of Using the parameter’s Bayesian estimation to calculate the steady state for the non-linear model？ (As you need to know the steady state before Bayesian Estimation, and to know the steady state, you need to set the parameters value first.)

Appreciate any explanation or any help.

Many thanks,

Amy