note: index_msa omitted because of collinearity

[sahar shabani]

hi friends.
i have faced a problem. in my regression , this notification confused me.
my regression:
xi: areg inv RE_value_msa index_msa cash qm i.ageq*index_msa i.roaq*index_msa i.assetq*index_msa i.state*index_msa i.sic2*index_msa , a(gvkey) cl(id2)
my problem:
note: index_msa omitted because of collinearity
how to solve it? why is it created?
please help me.
thanks alot.

[Carlo Lazzaro]

Sahar:
Some comments about your post:
- you can safely get rid of -xi:- because -areg- supports -fvvarlist- notation;
- collinearity: there's nothing you can but changing your model specification (it may be that -index_msa is colinear with -gvkey-);
- please do not double-post (see that FAQ). Thanks.

[Phil Bromiley]

To add to Carlo's comments, with your kind of data, a panel estimator (xtreg) is sometimes preferred to areg although these give the same results if you have the same panels and error structures but xtreg gives you more options (like random effects). You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. In your case, we don't know the structure of your data - I would assume you have multiple observations per firm over time?

It is sometimes worth figuring out why you have colinearity - it might reflect a variable that is acting differently than you expect or something else unintended. If you want to understand what is generating the colinearity, the first thing is to areg index-msa against all the other x's in the model. This may illuminate the source of your problem. As Carlo points out, if you only have one value for index_msa for each panel (i.e., index_msa doesn't vary by year), then it will be colinear with the fixed effects.