Announcement

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

  • robust panel model

    hello everyone, i want to run my data panel model with robust estimation because the assumption of normality is violated, I have run the code below

    "xtrobreg mm pop_y_nonhidro x19 x1 x12 x4 x47 x53, usave"

    with the result

    converting data ... done
    preparing data for subsampling ... done
    enumerating 20 candidates 0%....20%....40%....60%....80%....100%
    refining 5 best candidates ..... done
    iterating RLS ........... done
    computing standard errors ... done

    Pairwise-differences MM regression (85% efficiency)

    Number of obs = 176
    Wald chi2(6) = 22.96
    Prob > chi2 = 0.0008
    Breakdown point = 50
    M-estimate: k = 3.4436898
    S-estimate: k = 1.547645
    Scale = 3.791e-07

    Group variable: Kode Number of groups = 8
    Time variable: Tahun Group size: min = 22
    avg = 22
    max = 22

    (Std. err. adjusted for 8 clusters in Kode)
    ------------------------------------------------------------------------------
    | Robust
    pop_y_nonh~o | Coefficient std. err. t P>|t| [95% conf. interval]
    -------------+----------------------------------------------------------------
    x19 | 1.07e-09 1.49e-08 0.07 0.944 -3.41e-08 3.62e-08
    x1 | -1.04e-08 6.01e-09 -1.73 0.126 -2.46e-08 3.79e-09
    x12 | -4.29e-09 4.36e-09 -0.98 0.358 -1.46e-08 6.02e-09
    x4 | 3.55e-11 9.86e-12 3.60 0.009 1.22e-11 5.88e-11
    x47 | 3.04e-07 1.76e-07 1.73 0.128 -1.12e-07 7.19e-07
    x53 | 2.38e-07 1.28e-07 1.87 0.104 -6.37e-08 5.40e-07
    ------------------------------------------------------------------------------
    Hausman test of MM against S: chi2(6) = 15.83801 Prob > chi2 = 0.0147


    I have some questions:​​​​​​
    1. Is the syntax & the steps that I did correct?
    2. How I defined this is FEM/CEM/REM model?
    3. How can I display the ui for FEM, F test, R2?
    4. Can I custom the efficiency?

    sorry, I'm still newbie in stata

  • #2
    Arief:
    normality is a weak requirements of the residual(s) distribution.
    If your regerssand is continuous, go -xtreg,fe- first with cluster-robust standard errors.
    Kind regards,
    Carlo
    (Stata 19.0)

    Comment


    • #3
      Thanks Mr. Lazzaro

      Comment


      • #4
        Arief:
        Carlo is enough. Thanks !
        Kind regards,
        Carlo
        (Stata 19.0)

        Comment

        Working...
        X