st: Too many omitted dummy variables w/ ivreg / use xtivreg instead?

[Saba Movahedi <saba.movahedi@gmail.com>]

Hi,

I'm running a regression on stata with the following variables: gini
coefficients, malaria (prevalence of malaria), gdp, employment, life
expectancy, ground-frost levels, year-dummies (18 years), and
country-dummies (149 countries).

I attempted to run the following basic IV-regression:

ivreg gini (malaria = gfrost) employment, lifeexpect, gdp, country-dummies ,
year-dummies

The issue which comes up is that in my results 17/18 year dummies are
omitted and approximately 120/145 country dummies are also omitted. I
believe that the omission of the country dummies might come as a result of
lacking observations for certain years. However, I'm still not sure why
there is only ONE year dummy included instead of only one being dropped.

I thought of running it as panel data instead and therefore using: xtivreg
gini (malaria = gfrost) employment, lifeexpect, gdp

and dropping the dummies as they would no longer be needed but apparently
instrumental variables cannot be used with panel data? Is this correct?

Any ideas? I'd be very grateful for any help at all!

Best,

Saba M.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/