Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Question about panel data and time-invariant binary endogenous variable
From
Alfonso Sanchez-Penalver <[email protected]>
To
"[email protected]" <[email protected]>
Subject
Re: st: Question about panel data and time-invariant binary endogenous variable
Date
Mon, 24 Feb 2014 10:25:22 -0500
Hi Anna,
Yes there is another way. You can model this as a fixed effects model without including the time invariant endogenous variable. Fixed effects is appropriate when we expect unobserved heterogeneity captured in the errors to be correlated with the explanatory variables. Not including the variable you mention would have it captured by the errors, and since it's endogenous this would make the errors correlated with the explanatory variables. Since your variable is time invariant it would be dropped with a fixed effects estimation, so I believe that it would work perfectly in your case. See -help xtreg-, and the option you want is -fe- for fixed effects.
Alfonso Sanchez-Penalver
> On Feb 24, 2014, at 10:10 AM, "Veksler, Anna" <[email protected]> wrote:
>
> Hello Dear Stata experts,
>
> I have a question about panel data and time invariant endogenous variable. Right now I have a panel data with three observations (for second, fourth and sixth quarters) for every person and I have a binary endogenous variable that takes on 0 or 1 and doesn't change over time. Right now I am modeling it using iv regress. Is this a right thing to do? Are there any other options?
>
> Thanks a lot.
>
> Anna Veksler
>
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/faqs/resources/statalist-faq/
> * http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/