The RATS Software Forum

I am running a Bayesian VAR using mixed estimation with variables that are integrated of order 1 as endogenous variables and variables that are stationary as exogenous variables. One of the stationary exogenous variables is a dummy variable, while the other exogenous variables are relevant from a theoretical point of view. If I wish to estimate the Bayesian VAR for this model with normal Inverted wishart as the prior, does RATS allow me to do the same? ( I wish to keep the flat prior on the exogenous variables). Is it also possible to model the Hierarchical prior(Korobilis (2012)) Thanks in advance.

sanjeev wrote:I am running a Bayesian VAR using mixed estimation with variables that are integrated of order 1 as endogenous variables and variables that are stationary as exogenous variables. One of the stationary exogenous variables is a dummy variable, while the other exogenous variables are relevant from a theoretical point of view. If I wish to estimate the Bayesian VAR for this model with normal Inverted wishart as the prior, does RATS allow me to do the same? ( I wish to keep the flat prior on the exogenous variables).

Yes. I'm not sure why you think there's an issue.

sanjeev wrote: Is it also possible to model the Hierarchical prior(Korobilis (2012)) Thanks in advance.

Could you please provide a full reference?

When i tried to run the Gibbs VAR programme with these additional variables (exogenous) I get the following error message.

## MAT14. Non-invertible Matrix. Using Generalized Inverse for SYMMETRIC.
The Error Occurred At Location 63, Line 6 of loop/block
So I was wondering if the problem is with the data set or the programmes and that is why i sent the query. Your response has helped.

The reference for the Hierarchical prior is is Korobilis, D. (2012). “VAR Forecasting Using Bayesian Variable Selection,” Journal
of Applied Econometrics.

See https://estima.com/forum/viewtopic.php?f=7&t=1267

thanks a ton for the help.