Announcement

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

  • Regression results

    Hi, I am new to regression and still reading books about it.

    I am regressing debt ratio against profitability, size etc. dy are the years, and dc are life cycle of firms, and ind are industries.

    Seeing this results, is there anything I should be concerned about?
    Actually the thesis that I refer to, has the same number of dy and dc but only one industry dummy, but here I created more industry dummies so I am not so sure.

    Appreciate if anyone could help. Thank you.

    Code:
     
    
    . reg totaldebtmarketvalue profitability size salesgrowth tangibility nontaxdebt liquidity dy1 dy2 dy3 dy4 dy5 dy6 
    > dy7 dy8 dy9 dc1 dc2 dc3 dc4 ind1 ind2 ind3 ind5 ind6 ind7 ind8 ind9 ind10
    
          Source |       SS           df       MS      Number of obs   =     4,820
    -------------+----------------------------------   F(28, 4791)     =     86.42
           Model |   86.829243        28  3.10104439   Prob > F        =    0.0000
        Residual |  171.918517     4,791   .03588364   R-squared       =    0.3356
    -------------+----------------------------------   Adj R-squared   =    0.3317
           Total |   258.74776     4,819  .053693248   Root MSE        =    .18943
    
    -------------------------------------------------------------------------------
    totaldebtma~e |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
    --------------+----------------------------------------------------------------
    profitability |  -.4817474   .0236452   -20.37   0.000    -.5281029    -.435392
             size |   .0331574   .0020903    15.86   0.000     .0290595    .0372554
      salesgrowth |   .0018625   .0011357     1.64   0.101     -.000364    .0040891
      tangibility |   .0822787   .0151707     5.42   0.000      .052537    .1120203
       nontaxdebt |  -.5633339   .1152644    -4.89   0.000     -.789305   -.3373628
        liquidity |  -.0106057   .0004338   -24.45   0.000    -.0114562   -.0097552
              dy1 |  -.0598851   .0122223    -4.90   0.000    -.0838463   -.0359238
              dy2 |  -.0835679   .0122264    -6.84   0.000    -.1075372   -.0595986
              dy3 |  -.0755893   .0122124    -6.19   0.000    -.0995312   -.0516475
              dy4 |  -.0789071   .0122277    -6.45   0.000     -.102879   -.0549352
              dy5 |  -.1193103   .0122344    -9.75   0.000    -.1432953   -.0953253
              dy6 |  -.1279573   .0122498   -10.45   0.000    -.1519726    -.103942
              dy7 |  -.1339691   .0122584   -10.93   0.000    -.1580011    -.109937
              dy8 |  -.1303951   .0122818   -10.62   0.000     -.154473   -.1063171
              dy9 |   -.142779   .0123201   -11.59   0.000     -.166932   -.1186259
              dc1 |   .0554025   .0098219     5.64   0.000      .036147     .074658
              dc2 |   .0427338   .0076047     5.62   0.000     .0278252    .0576424
              dc3 |    .012387   .0081153     1.53   0.127    -.0035227    .0282967
              dc4 |   .0677144   .0112457     6.02   0.000     .0456678    .0897611
             ind1 |  -.0521258    .007409    -7.04   0.000    -.0666509   -.0376007
             ind2 |   .0116828   .0165759     0.70   0.481    -.0208136    .0441792
             ind3 |  -.2127064    .021915    -9.71   0.000      -.25567   -.1697429
             ind5 |  -.1892949   .0134049   -14.12   0.000    -.2155746   -.1630151
             ind6 |   .0076983    .009576     0.80   0.421     -.011075    .0264716
             ind7 |  -.1085723   .0105618   -10.28   0.000    -.1292782   -.0878665
             ind8 |  -.1258606   .0166844    -7.54   0.000    -.1585697   -.0931516
             ind9 |  -.0617338   .0150073    -4.11   0.000    -.0911551   -.0323126
            ind10 |  -.0443932   .0220523    -2.01   0.044    -.0876259   -.0011605
            _cons |   .1446462   .0284044     5.09   0.000     .0889605    .2003319
    -------------------------------------------------------------------------------

  • #2
    Larissa:
    I'm afraid your model is basically flawed.
    You performed an OLS on panel data (ie, a sample of firms measured repeadety along a given span of time) but without acknowledging that the observations belonging to the same firm are not independent (which means: please use -cluster- standard errors if you cannot help going (pooled) OLS).
    Even more important, it is rare (although possible) that -xtreg- (that you should spot and read in Stata .pdf manual) is outperformed by a (pooled) OLS when it comes to panel data.
    In sum, your regression model, as it is, is not reliable and can be easily questioned by any decent econometrician/applied economist.
    Kind regards,
    Carlo
    (Stata 18.0 SE)

    Comment


    • #3
      Originally posted by Carlo Lazzaro View Post
      Larissa:
      I'm afraid your model is basically flawed.
      You performed an OLS on panel data (ie, a sample of firms measured repeadety along a given span of time) but without acknowledging that the observations belonging to the same firm are not independent (which means: please use -cluster- standard errors if you cannot help going (pooled) OLS).
      Even more important, it is rare (although possible) that -xtreg- (that you should spot and read in Stata .pdf manual) is outperformed by a (pooled) OLS when it comes to panel data.
      In sum, your regression model, as it is, is not reliable and can be easily questioned by any decent econometrician/applied economist.
      Hi Mr Carlo,

      Thank you for your reply. If I have to use OLS, I read that "you should -cluster- your standard errors on your -panelid",
      May I know what is the panel id in my case?

      Comment


      • #4
        Larissa:
        I would guess -industries-.
        Kind regards,
        Carlo
        (Stata 18.0 SE)

        Comment


        • #5
          You may also consider using factor notation for the use of dummies, instead of creating them.

          Please read the FAQ. There you'll find advice to register with name and family. Thanks.
          Best regards,

          Marcos

          Comment


          • #6
            Originally posted by Carlo Lazzaro View Post
            Larissa:
            I would guess -industries-.
            Thanks for letting me know my model's flaws. I was replicating my senior's model. I will read more about it.
            Apologize as I am still new to this.

            Comment


            • #7
              Larissa:
              no need for apologizing: we all were newbyes at a given point in time.
              Consider also Marcos' wise advice in revising your regression model.
              Kind regards,
              Carlo
              (Stata 18.0 SE)

              Comment


              • #8
                Originally posted by Marcos Almeida View Post
                You may also consider using factor notation for the use of dummies, instead of creating them.

                Please read the FAQ. There you'll find advice to register with name and family. Thanks.
                Saw it, thanks for the advice!

                Comment

                Working...
                X