Announcement

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

  • Help with xtdcce2 command

    Dear community,

    I'm trying to estimate the CS-DL model as in the Chudik et. al (2013) using the xtdcce2 command as explained by Dr JanDitzen in "Estimating long run effects in models with cross-sectional dependence using xtdcce2". But whenever I try to estimate I get the following error;

    uneven dimensions
    } is not a valid command name
    r(199);

    However, the command that I'm trying to run is following;
    xtdcce2 d.y dp d.gd d.(dp d.gd), cr(d.y dp d.gd) cr_lags(0 3 3) fullsample

    If I change the cross-sectional lag structure by defining a number rather than a numlist then the model is estimatable such that;
    xtdcce2 d.y dp d.gd d.(dp d.gd), cr(d.y dp d.gd) cr_lags(3) fullsample

    But I don't want to take the cross-sectional lag of the dependent variable. What should I do?
    Besides, I have been able to estimate the CS-ARDL without any problem. Thanks.

    Regards,
    Kumail
    Last edited by Kumail Dhudra; 29 Oct 2023, 22:27.

  • #2
    Ho Kumail,

    thank you very much for your interest in xtdcce2.

    Which version of xtdcce2 are you using? You can find out by typing "xtdcce2, version".

    The error you obtain is unusual in the sense that it seems that you are loosing some of your data. Do you have any missings or an unbalanced dataset? Also which Stata version do you use?

    Comment


    • #3
      Hi Jan Dizten,

      Thanks for the reply.

      I'm using xtdcce2 version 4. As indicated by "xtdcce2, version" command.

      I have tried different datasets but the same error continues. The dataset is strongly balanced, but it has some missing observations. I'm using the dataset from Chudik et al. (2013) available at Kamiar Mohaddes' personal webpage. Currently, I'm using StataMP 17. Thanks

      Comment


      • #4
        Have you updated it from GitHub?

        I am assuming that there is something with the missing observations. Can you please send me the dataset by mail?

        Comment


        • #5
          Thanks, Jan Ditzen. I uninstalled the previous package and have now reinstalled it from GitHub. The problem is now solved, as there are no more errors, and the model is estimatable. I believe the issue was with the Stata package, as I had previously downloaded xtdcce2 using the 'ssc install xtdcce2' command. The GitHub package seems to work flawlessly.

          Can you also help me clarify another thing: In CS-DL the dependent variable is always taken in difference form just like CS-ARDL? Thanks.
          Last edited by Kumail Dhudra; 02 Nov 2023, 10:47.

          Comment


          • #6
            You can download the latest version by typing

            Code:
             
             net from https://janditzen.github.io/xtdcce2/
            Please let me know if that works.

            Best,

            Jan

            Comment


            • #7
              I have download the latest version using the following command

              Code:
              net install xtdcce2 , from("https://janditzen.github.io/xtdcce2/")
              And it works. Thanks alot.

              Comment

              Working...
              X