Thank you in advance for your valuable time. I am trying to estimate an open economy DSGE model. When I estimate multivariate convergence diagnostics converges, But I am worried about the posterior plots of standard error of shocks. Does this indicate any problem? How can I improve this? All relevant details are added in the attached word file
shocks.pdf (278 KB)
It’s hard to tell without context. Do you have reason to be suspicious? It just looks as if your posterior was bigger than what the data implies and Bayesian updating was successful because the posterior is different than the prior.
If there is a problem, my guess is the data treatment. It looks as if you are aiming standard errors of 1 representing 1%. So did you make sure your data is in percent and scaled accordingly (by 100)?
Thank you. I haven’t done the scaling. I will do that and check if that works.