EViews.com

Hello,

I know that Eviews does not give the option to calculate the confidence intervals for the IRFs of BVARs, but there must be ways of calculating these manually since I've seen them presented in the results of papers.

Any guidance on how to manually calculate CI's for BVARs using an Eviews program?

I'm using Eviews 9.5

Thanks.

Upgrade to EViews 11, where you can compute them (via MCMC).

I've just got the upgraded Eviews 11 at my university. Can you tell me how I can calculate the impulse standard errors using MCMC? I still see that standard errors for impulse responses are disabled in BVARs and VECs.

I'm estimating a BVAR using Litterman priors.

Keep the simulations in a new page, then compute the quantiles of the simulations you want.

Ok, thanks. Technical question: I could use the standard deviation of each simulation at one period x 1.96 to get the +- standard errors for 95%?

Technically, yes, but I suspect a Bayesian would tell you to use the 95% quantiles.

Can you confirm that EViews 11 University Edition includes the MCMC feature to simulate the CIs for the IRFs of the BVAR. If not, please let me know which specific version will work.

Confirmed