This will only provide an unbiased estimate if you have more years of data than households (ie T>N where N is the fixed effect dimension). Otherwise, you have the familiar incidental parameters problem in non-linear models. The only fixed effects approach that gives an unbiased estimate here is the conditional maximum likelihood approach used in xtlogit, fe or clogit. It is possible to extend this approach to ordered models but as far as I am aware no one has programmed this in stata. Ferrer-i-Carbonel, A., Frijters, P. (2004), 'The effect of methodology on the determinants of happiness', Economic Journal 114, 641-659 discusses this model in detail and the authors may have code (I believe in GAUSS) available.
Cheers,
Steve
-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of maartenbuis
Sent: Saturday, October 29, 2005 10:17 AM
To: [email protected]
Subject: RE: st: ologit with fixed effect
One way to get a fixed effect model is to add dummies for households
in a regular ologit (gologit2) command.
Hope this helps,
Maarten
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
