Announcement

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

  • #16
    I have 48 countries across a period of 25 years (on average) with a quarterly frequency (i.e. 100 periods in total). Is your strategy still feasible?
    Indeed, does this solution address only contemporaneous correlation or serial correlation as well?

    Comment


    • #17
      Marcus:
      your actually dealing with a T>N panel dataset.
      See -xtregar- and -xtgls- instead of -xtreg-.
      Kind regards,
      Carlo
      (Stata 19.0)

      Comment


      • #18
        Thank you Carlo, I’ll go for this solution.
        I read the help file and the documentation, but I’m still unsure about the last question for xtregar: does this solution address only contemporaneous correlation or serial correlation as well?

        Comment


        • #19
          Marcus:
          -xtergar- options manages different mechanisms of autocorrelation, but not heteroskedasticity.
          Kind regards,
          Carlo
          (Stata 19.0)

          Comment


          • #20
            Sorry, bt I have a similar problem with a lag variable. My data are described as the following:
            HTML Code:
            . describe zRE zage family_firm zn_employees year
            
            Variable      Storage   Display    Value
                name         type    format    label      Variable label
            ----------------------------------------------------------------------------------
            zRE             double  %9.0g                 Std ROA
            zage            float   %9.0g                 Standardized values of age
            family_firm     byte    %10.0g                FAMILIAR_1
            zn_employees    float   %9.0g                 Standardized values of n_employees
            year            double  %10.0g                
                         
            And, when I run the dynamic regression:
            I have a panel data
            HTML Code:
            . xtset company year
            
            Panel variable: company (weakly balanced)
             Time variable: year, 2006 to 2019
                     Delta: 1 unit
            HTML Code:
            . xtabond2 zRE l.zRE zage 1.family_firm zn_employees  i.year , gmm(l.zRE, lag(2 2)) iv(zage 1.family_firm zn_employees i.year ) twostep robust arte(3)
            The result is:
            HTML Code:
            No observations.
            r(2000);
            Because do not regnonize the lar on the variable zRE:
            HTML Code:
            . list RE zRE l.zRE in 1/10
            
                 +--------------------------+
                 |                        L.|
                 |     RE         zRE   zRE |
                 |--------------------------|
              1. |   5.78    -.061369     . |
              2. |   .161   -.6198423     . |
              3. |   .475   -.5886338     . |
              4. |   .276   -.6084124     . |
              5. | -3.822   -1.015713     . |
                 |--------------------------|
              6. |  1.407   -.4960021     . |
              7. |  2.912     -.34642     . |
              8. |   1.07   -.5294966     . |
              9. |  2.344   -.4028736     . |
             10. |  4.468   -.1917689     . |
                 +--------------------------+
            Please, I appreciate some help

            Comment


            • #21
              Sorry I notice what the problem is: I was only one valor of Panel variable for each time variable.

              Comment

              Working...
              X