Does anyone know how to use the estimated model to predict 100 years?

Dear all.
I want to predict 100 years by using estimated dsge-var model on dynare.
However, I do not know how to do them on dynare code.
so that, anyone teach me how to do them.


Do you need the forecast from the VAR or the DSGE-model?

jpmodel5.mod (12.8 KB)

Thank you, jpfeifer.
I need to forecast consumption tax, labor income tax, and capital tax on DSGE-VAR model as attached file.
If forecast can use DSGE-VAR, i want to use it. However, if dsge-var cannot use that, I will use dsge-model.

I want to forecast the effect of tax in case of increase or decrease during 50 years(it means 50years * 4 quater = 200 quaters). By setting scenario of following.(estimation is 1981 1Q to 2018 4Q)

setting consumption tax in 2019 1Q to 10%, and 2025 1Q to 12% and 2030 1Q to 15%.
or maintaining them 12% from 2025.
And compare them with the case not increasing consumption tax.
I want to compare the line of effecting japan economy.

In addition, If you know how to use estimation and simulation result which I had already estimated and simulated, I want know that because I am still doing them for MCMC’s sampling must be converged.


Bayesian impalse response is not succeed by the determinism of the Hessian matrix is impaired because My model is too complex and observation data is too combined many shocks because of japan economy is too singularity. Normal impalse response is succeed.

Hi, DSGE-VAR forecasts are not implemented in Dynare. But I think we provide the auto-regressive matrices somewhere under oo_. So it’s not difficult to produce a point forecast from the VAR (iterating on these matrices). For building predictive densities you would need to dive deeper into Dynare (sampling in the posterior distribution using the MH draws, computing the auto-regressive matrices and the forecast for each draw).

I don’t understand your last statement.


Thank you for stepan-a.
I know forecast cannot use at dsge-var model.