Announcement

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

  • Cross Sectional Dependence in Panel Data

    Hello everyone. I have a panel data with N=30 and T=20.After testing for cross sectional dependence using xtcsd, pesaran , it showed the presence of cross sectional dependence. After this, does anything else needs to be done to control for it ?

  • #2
    Anuradha:
    you may want to take a look at -xtgls- option.
    Kind regards,
    Carlo
    (Stata 19.0)

    Comment


    • #3
      "xtcsd tests the hypothesis of cross-sectional independence in panel data models with small T and large N " which is not your case.

      On Edit: there is another user written command -xtcd- which can be used for macro panels. I am not familiar with either (xtcsd or xtcd). I just quoted the help file for xtcsd above, and am doing the same for xtcd here.
      Last edited by Eric de Souza; 08 Oct 2020, 09:14.

      Comment


      • #4
        Thank you Carlo Lazzaro Eric de Souza . I cannot use xtcsd pesaran the. xtgls would be helpful right?

        Comment


        • #5
          Anuradha:
          -xtgls- was actually developed for T>N panel datasets. However, since your T dimension does not seem negligible, I would take a look at it.
          Kind regards,
          Carlo
          (Stata 19.0)

          Comment


          • #6
            In my case T<N. But its not a case of small T and large N I think?

            Comment


            • #7
              Anuradha:
              yes, you're dealing with N>T panel dataset that shows with a relevant T: hence cross-sectional correlation may be an issue (as it proved to be in your case) that should be addressed.
              Kind regards,
              Carlo
              (Stata 19.0)

              Comment


              • #8
                Thank you Carlo Lazzaro . what do I do next ?

                Comment


                • #9
                  Anuradha:
                  -panels(correlated)- available from -xtgls- specifies a heteroskedastic error structure with cross-sectional correlation.
                  Check whether it fits your research needs.
                  Kind regards,
                  Carlo
                  (Stata 19.0)

                  Comment


                  • #10
                    Thanks a lot again Carlo Lazzaro

                    Comment

                    Working...
                    X