Help on Fixed time effect and Industry effect!

[michelleli]

Hi:
I am new to Econometrics/Eview. I have a question regarding the fixed time effec and industry effect.
I am running a logit model on unbalanced panel data for the period betwen 1997 and 2008 for 16 industries. In order to control for year effect and industry effect, I would like to insert year dummies and industry dummies in my regression. I firstly inserted all the year dummies. When I include all the year dummies, the Eviews only produce the coefficients for the variables. The standard errors, z values and prob. are shown NA. However, when a drop a few year dummies, all the statistics are shown. I wonder what I have done wrong. Another question is how to control for the industry effect?

Cheers
Michlle

[startz]

Did you remember to drop the constant when you put in the dummies?

[michelleli]

Yes, I did.

[startz]

You might want to post your workfile, including the equation you estimated, to see if anyone can scope out the problem.

[EViews Gareth]

Did the equation report any warnings?

[Alex410]

HI,

I am also a beginner and have a similar problem. I have a company/year unbalanced panel and want to control for time and company effects. A recent article (Petersen, 2009) suggests clustering standard errors by time and company, but I have no clue how this could be done in eviews. Given that the "white period" option is not offered in binary regressions, I guess that including time (year) dummies and re-running the regression would not provide robust company standard errors?

Thanks
Alex