Yes, it's equivalent - but you only need 3 dummies because you need to
use one of the country as the reference or else you'll have
multicollinearity in your model. The reasonable results you are getting
now might not be so reasonable afterall.
Cordula Stolberg wrote:
Dear Statalisters,
I have a rather basic questions about fixed effects models. I am
estimating a fixed effects model, where my dependent variable is a trade
specialisation index for four countries with respect to six other
countries (basically I have countries 1, 2, 3, and 4 and I create an
index for each of the countries in relation to countries 5 - 10). I then
regress that index on a number of independent variables, one of them
being investment flows from the six countries into each of the four
countries.
When I run the regression with "xtreg, fe", Stata creates "too many"
fixed effects, as it takes each country-by-country trade index as a
fixed effect (thus I end up with 24 fixed effects dummies). I however
want only four fixed effects, namely one dummy each for countries 1 to
4. Is it ok if I run the regression only with "reg" and add the four
country dummies manually? When I did that the results seemed very
reasonable, but I'm not sure if that's appropriate to do.
Your help would be very much appreciated.
Cordula
--
Ada Ma
Research Assistant
Department of Economics
University of Aberdeen Business School
Edward Wright Building F55
http://www.abdn.ac.uk/economics/
Tel: +44 1224 273417
Fax: +44 1224 272181
Email: [email protected]
University of Aberdeen Open Day - Tuesday 24th August 2004
Visit our website at: www.abdn.ac.uk/openday or freephone 0800 0271495
*
* 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/
