0
$\begingroup$

If I know about an omitted variable in the regression and the data for it is available. But I also know that this omitted variable is endogenous, because the dependent variable and the omitted variable cause each other (simultaneity). Should I include the omitted variable and does that introduce new problems for my variable of interest?

Improve this question
$\endgroup$
3
  • 2
    $\begingroup$ Is your goal prediction, explanation, or an attempt to find causality? $\endgroup$
    – Peter Flom
    Commented Jan 4 at 11:08
  • $\begingroup$ Attempt to find causality $\endgroup$
    – Marlon Brando
    Commented Jan 4 at 14:48
  • 1
    $\begingroup$ You will have problems in either case, either omitted variables bias when omitting or simultaneity bias when including. If you include it, you can try if you can find an instrumental variable for the endogenous regressor. $\endgroup$
    – Christoph Hanck
    Commented Jan 12 at 13:37

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.