Announcement

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

  • How to capture the coefficients of fixed effects

    Dear all!
    I am currently doing a study on audit team style, and I want to capture the impact of the team on the audit program through multiple fixed effects. one audit program is be called assess .
    My STATA code is reghdfe ASSESS CONTROL, absorb(industry Year team) vce(cluster STKCD)
    The literature I drew from is
    Audit Outcomes/Audit Fees = X + ∑ Year + ∑ Client + ∑ Firm+ ∑ Team + ∑ Individual auditor + εit
    So how can i use stata to capture the coefficient of team FIXED effect ?and gen a new_variable which use the coefficient of team fixed effect
    and i don`t know my regression model and stata code is correct or not ,it is so hard for me
    Have a good day!

  • #2
    Code:
    sysuse auto, clear
    reghdfe mpg price weight, absorb(ferep89 = rep78)
    From help file:
    "However, be aware that estimates for the fixed effects are generally inconsistent and not econometrically identified."

    Comment


    • #3
      IF you want them all

      Code:
      reghdfe mpg price weight, absorb(rep78 foreign, savefe)

      Comment


      • #4
        Originally posted by George Ford View Post
        Code:
        sysuse auto, clear
        reghdfe mpg price weight, absorb(ferep89 = rep78)
        From help file:
        "However, be aware that estimates for the fixed effects are generally inconsistent and not econometrically identified."
        THANK U very very much ! i also use this code ,but my worry is this code whether include Year fixed effect and industry fixed effect ,and how to add the vce(cluster stkcd) or whether i should add it .
        Have a good day!

        Comment


        • #5
          If the coefficients on team are of interest, then I'd estimate them directly.

          Code:
          sysuse auto, clear
          
          g FE = .
          reg mpg price bn.rep78 , nocons
          forv i = 1/5 {
              replace FE = _b[`i'.rep78] if rep78==`i'
          }
          Or just use reghdfe and absorb, if the properties of the FE are not of great concern.

          Code:
          reghdfe ASSESS CONTROL, absorb(industry Year teamfe = team) cluster(STKCD)
           g newvar = teamfe + whatever
          I have no idea what STCKD is, so can't say much about the cluster. But read this:
          HTML Code:
          https://blogs.worldbank.org/en/impactevaluations/when-should-you-cluster-standard-errors-new-wisdom-econometrics-oracle

          Comment

          Working...
          X