Announcement

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

  • Test for endogeneity in presence of continuous interaction using IVREG2H

    Hello,

    I have gone through multiple posts/replies on Statalist regarding endogeneity but I am still stuck so pardon my ignorance here. I would very much appreciate help on this-

    I have a panel dataset (firms and years). I have 5 independent variables X1-X5. X1-X3 are the controls and I have an interaction term as well. All variables are continuous. X1-X3 and Y vary with time. X4 and X5 do not vary with time.
    I am running an OLS regression with double-clustered robust standard errors using the -IVREG2- command in Stata as shown below-

    ivreg2 Y X1 X2 X3 X4 X5 X4*X5 , cluster(firm years)

    I want to

    1) check if X3 and X4 are possibly endogenous.
    2) If they are, then I want to use -IVREG2H- to generate instruments for them.

    As you will notice that I also have an interaction term involving the 2 potential endogenous variables. How do I test and correct for endogeneity in such a case?

    ivreg2h Y X1 X2 X3 (X4 X5 X4*X5=),cluster(firm years) gmm2s <- something like this? However, in this case, I am basically considering the interaction of X4 and X5 as an altogether new endogenous variable which is not making sense to me.

    Thank you so much in advance!
Working...
X