Hello I am planning to run a fixed effect model in stata. However I would like to discuss here 1) whether my method is the right way to go and 2) why all my dummy variables are ommited.
My variables are: firmcode year dependentvariables independentvariables industry (50industries in total, each of them I give a specific value （eg. if telecommunication then industry = 1, 0 otherwise, hoping to find how industries are having an effect on dependent variable)
My code:
xtset firmcode year
xtreg dependentvariable idependent variables i.year i.industry, fe vce(robust)
My intention for i.year is to control for any serial correlation that may occur in my panel data ; my intention for i.industry is to control for industry-level effects (for example, firms in specfic industries may have significant effects on my dependent variable). And this is also one variable of interest ; vce(robust) is to adjust for heteroskedasticity.
However after I run the test, all my industry dummies are ommited due to collinearity
note: industry_code1 omitted because of collinearity
note: industry_code2 omitted because of collinearity
note: industry_code3 omitted because of collinearity
note: industry_code4 omitted because of collinearity
note: industry_code5 omitted because of collinearity
note: industry_code6 omitted because of collinearity
note: industry_code7 omitted because of collinearity
I am very comfused why that happened and I cannot analyse the firm-level effects. I wonder if anyone can give me advise on how to resolve that issue.
And I would be quite grateful if anyone can give me advice on how to refine my model overall!
Many thanks in advance!
My variables are: firmcode year dependentvariables independentvariables industry (50industries in total, each of them I give a specific value （eg. if telecommunication then industry = 1, 0 otherwise, hoping to find how industries are having an effect on dependent variable)
My code:
xtset firmcode year
xtreg dependentvariable idependent variables i.year i.industry, fe vce(robust)
My intention for i.year is to control for any serial correlation that may occur in my panel data ; my intention for i.industry is to control for industry-level effects (for example, firms in specfic industries may have significant effects on my dependent variable). And this is also one variable of interest ; vce(robust) is to adjust for heteroskedasticity.
However after I run the test, all my industry dummies are ommited due to collinearity
note: industry_code1 omitted because of collinearity
note: industry_code2 omitted because of collinearity
note: industry_code3 omitted because of collinearity
note: industry_code4 omitted because of collinearity
note: industry_code5 omitted because of collinearity
note: industry_code6 omitted because of collinearity
note: industry_code7 omitted because of collinearity
I am very comfused why that happened and I cannot analyse the firm-level effects. I wonder if anyone can give me advise on how to resolve that issue.
And I would be quite grateful if anyone can give me advice on how to refine my model overall!
Many thanks in advance!
