Announcement

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

  • reghdfe can be used with cross-section data?

    Hi everyone,

    I am modeling a cross-section on migration, so my dependet variable is a count. As I have 558 regions, I included 558 dummies for the destination and 558 dummies for the origen regions. I understand that reghdfe can be used with non linear models as well, but I am not sure if it is suitable for cross-section. Can anyone help me on this?

    Thank you!

    Cirlene

  • #2
    Cirlene:
    the main question is: why creating 558 dummies when, after -reshape-ing your data in -long- format you can benefit from -fvvarlist- notation?
    Kind regards,
    Carlo
    (Stata 18.0 SE)

    Comment


    • #3
      Carlo,
      Thank you for your reply. Indeed, you are right, it is better to use fvvarlist notation.

      But I still have a doubt concerning the use of reghdfe for cross-section count data. Can you give me any tip on this? Thanks.

      Comment


      • #4
        Cirlene:
        if you have cross-sectional data with count as dependent variable, I would consider -poisson-.instead of -reghdfe-.
        Kind regards,
        Carlo
        (Stata 18.0 SE)

        Comment


        • #5
          Abstracting from the count vs continuous debate. Yes, you can use reghdfe in a cross-section. It does not require xtset, instead you just specify the destination and origin variables in absorb() and reghdfe will do the rest.

          Comment

          Working...
          X