Re: st: Omitted Fixed Effects Dummy Variables

[Maarten Buis <maartenlbuis@gmail.com>]

On Thu, Sep 27, 2012 at 4:03 AM, Sabin Ahmed wrote:
> I am running a fixed effects regression model with panel data and a
> LOT of county-year and industry-year fixed effects dummy variables,
> taking on a value of (0,1) for each country-year or industry-year
> combination.
>
> I have 9 years, 23 industries, and 61 countries. My country-year fixed
> and industry-year fixed effects dummies for example are as follows:
>
>        - year1country1, year1country2,  ... year9country61.
>
>        - year1industry1, year1industry_2,  ... year9industry23.
>
> I have already dropped all the FE dummies which only have a value of 0
> for all observations.
>
> However, now when I run my regression, some of my dummy variables get
> omitted in the Stata output. For example when I run:

You need to make sure you leave out reference categories. This gets
more complicated as you seem to want to add interactions between year
and country and year and industry. Rather than trying to figure this
out on your own by making your own dummies, I suggest you use the
factor variable notation instead:

I assume you have categorical variables year, country and industry,
than you type
reg y x1 x2 x3 i.year##( i.country i.industry)

-- Maarten

---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany

http://www.maartenbuis.nl
---------------------------------
*
*   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/