Dynare forums

Hi dear all, I was wondering if it is possible to simulate OLG models by Bayesian method in Dynare, I did not find any paper on this topic until now, but I found some examples using Calibration method (Heer, B. and A. Maussner (2012)), how can the Bayesian estimation for DSGE model be applied to OLG model?

Best regards,
Frank

If the OLG model is stochastic, it needs to be differentiable and have a recursive time-invariant representation. In that case, you can use perturbation techniques and estimate it. An OLG model of this type is not different from any other smooth DSGE model and it can be estimated using standard Bayesian techniques.

jpfeifer wrote:If the OLG model is stochastic, it needs to be differentiable and have a recursive time-invariant representation. In that case, you can use perturbation techniques and estimate it. An OLG model of this type is not different from any other smooth DSGE model and it can be estimated using standard Bayesian techniques.

Thanks for your prompt response!
I'm trying to investigate the effect of population aging on economic growth, it seems DSGE model with represent agent is not available.
so i was developing a stochastic OlG model with various shocks, but to my knowledge, most of the OLG models were solved by theoretical proofs or simulated by calibration method, i did not find any paper on stochastic OLG model estimated by Bayesian techniques until now, consequently, i dared not use OLG model.
Would you know if there is any paper on stochastic OLG model estimated by Bayesian techniques?

sincerely !

I have no answer. You could look at https://github.com/davidrpugh/pyeconomics/wiki/OLG-Models and http://perso.uclouvain.be/david.delacroix/data.html. Maybe you can find something useful there.

Dear Johannes,

Could I ask if it would be also right to use Linear-approximation method instead of perturbation techniques to solve an OLG model and use Bayesian method to estimate that model?

Thanks in advance,
Catherine

What do you mean? A linear approximation corresponds to first order perturbation.

jpfeifer wrote:What do you mean? A linear approximation corresponds to first order perturbation.

Thank for your reply.

What I want to do is to estimate an OLG model with several shocks using Bayesian estimation.

I am a little bit confused about this post.... Do you mean I can estimate the OLG model just like estimating general DSGE model?

Catherine

You can use Dynare to estimate any stationary recursive linear(ized) model. If your OLG model falls into this class, the answer is yes.