1
$\begingroup$

I am doing a survey study on factors related to child health on around 7000 individuals. All dependent variables are binary and exploratory variables are binary and categorical.

One of the explanatory variables suffered from endogeneity bias (education is correlated with error term). I used ivprobit to overcome endogenity bias. Now I checked few published study related to child health. Almost all used instrumental variables and fixed effects models.

I need to know — can I use only instrumental variable regression alone? Or must I have to use both fixed effects as well after IV to overcome endogeneity and heterogeneity?

$\endgroup$
3
  • $\begingroup$ Can you please elaborate a bit more on your problem/dataset? From what you write it is unclear what you are asking. $\endgroup$
    – Knarpie
    Commented Mar 19, 2018 at 14:17
  • $\begingroup$ I have a survey data of 7000 individuals nested into 500 (clusters) primary sampling units. One of my explanatory variable suffered from endogeneity problem. i have one endogenous and one instrumental variable. I applied IV regression using probit. in first stage endogenous co variate was regressed on explanatory variable. in 2nd stage of iv, predicted probability of endogenous covariate was used as explanatory variable with other explanatory variable , i have one endogenous and one instrumental variable. $\endgroup$
    –  Atif
    Commented Mar 19, 2018 at 17:04
  • $\begingroup$ i checked some published papers. which assume endogeneity problem in their study. and found they almost all applied fixed effects models if they use instrumental variable. so my question is that is it possible to apply only IV regression when we assume endogeneity problem in our study or we must to have apply fixed or mixed effects alongwith IV $\endgroup$
    –  Atif
    Commented Mar 19, 2018 at 17:06

1 Answer 1

1
$\begingroup$

You can use the instrumental variable alone if it is unconditionally exogenous and relevant. However, if you need to condition on fixed effects to motivate that the instrument is exogenous, then you need to perform the instrumental variable estimation conditioning on fixed effects.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.