Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Linear state space

    I have been largely benefited from this forum every time I post a question or I am looking for a question someone had previously asked.

    For today, my model is a TVP-PVAR in a normal linear state space model composed of the State Equation and the Measurement Equation. . There is a large amount of data, therefore I'm using numerous variables in the model. My model equations are the ones given below and for a panel after linearization panel where according to Ciccareli and Canova (2013) can be estimated as such. My model of equations are therefore defined .

    Click image for larger version

Name:	b5M60.png
Views:	2
Size:	9.3 KB
ID:	1692495

    Where X ̃t=XtΞ and ut=Xt′+ut with Ut∼N(0,(I+σ2Xt′Xt))×Σ



    Then this can be estimated a normal var or a bayesian var where the time variation is given by the unit root. Whether then the estimation is made by classic tolls or Bayesian vars is indifferent as long a the likelihood is well defined

    I know that in Stata 17 there are some tools for estimating bayesian vars but could anyone of here be of help on how to procedure, eventually with a code. I

    can provide data as long as I can find a good solution

    I would immensely greatfull for any help
    Attached Files

  • #2
    https://www.stata.com/new-in-stata/bayesian-VAR-models/

    Comment


    • #3
      Thanks, but those seem to be stationary models.
      How could I introduce in those time varying parameters, a latent unit root process?
      Can those still fit for panels Bayesian Vars ?
      Last edited by Mario Ferri; 08 Dec 2022, 00:57.

      Comment


      • #4
        Good question. I'm not a Bayesian scholar, and very few people on here are. I would just Google "bayes time series stata" and see if someone's released something new, it looks like new stuff has come out recently, but I'm not skilled enough to judge if it's suited for your purposes

        Comment


        • #5
          Originally posted by Jared Greathouse View Post
          Good question. I'm not a Bayesian scholar, and very few people on here are. I would just Google "bayes time series stata" and see if someone's released something new, it looks like new stuff has come out recently, but I'm not skilled enough to judge if it's suited for your purposes
          Thanks Jared. Nothing found till now for Stata. I think will have to write to stata support to ask directly.

          Comment

          Working...
          X