I am trying to use ivregress and ivreg2.
When I use the option for autocorrelation consistent (bw, hac) I got error messages that the: "sample may not include multiple panels", or "variable NAME not found".
Some of the variables in my initial dataset were constructed by using a firm identifier variable and fiscal year variable, those variables were later been dropped. Now I am trying to regress by using industry identifier variable and calendar year variable. These two variables and all the other in my dataset were constructed by the collapse command (they now represent an average of all the firms in industry I at year T).
Those massage come up with or without using the command tsset before ivregress.
Is it possible that Stata "recall" that some variables were built by using the firm identifier and fiscal year variable?
I will appreciate any help how to overcome this problem.
Thanks,
Netty
When I use the option for autocorrelation consistent (bw, hac) I got error messages that the: "sample may not include multiple panels", or "variable NAME not found".
Some of the variables in my initial dataset were constructed by using a firm identifier variable and fiscal year variable, those variables were later been dropped. Now I am trying to regress by using industry identifier variable and calendar year variable. These two variables and all the other in my dataset were constructed by the collapse command (they now represent an average of all the firms in industry I at year T).
Those massage come up with or without using the command tsset before ivregress.
Is it possible that Stata "recall" that some variables were built by using the firm identifier and fiscal year variable?
I will appreciate any help how to overcome this problem.
Thanks,
Netty
Comment