Announcement

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

  • Reversed Causality in Labor forceparticipation regression

    Dear all,

    I plan to investigate the relationship between labor market participation and heart diseases. My dependent variable is a Dummy variable for the labor market participation (=1 if working, 0 otherwise) and my main independent variable is a Dummy variable for having a heart disese (=1 if heart disease, 0 otherwise). Therefore, I want to estimate a individual and time fixed effects logit model. LPM is not appropriate because I want to use the predicted probabilities later on.

    As simultaneous causality is likely because having a heart disease should influence the labor supplly decision, also the decision to work or not might change an individuals lifestyle and thus influence the probability of getting a heart disease, so I want to use IV to account for the simultaneous causality.

    Of my best knwoledge there is no command for a logit IV fixed effects model in Stata. Is there any conceptional mistake in a logit IV fixed effects model, or can I just run both IV stages by hand?

    Thankds for your help.

    Best

  • #2
    You could use cmp or do it in GSEM. Generally, doing the first stage by hand results in incorrect errors in a subsequent estimation. I have seen this for 2SLS but assume it also holds for logit.

    Comment

    Working...
    X