What is difference between PROC GLM and the TSCSREG procedure

[aminkarimid]

Hello everybody,

I want to run a regression between volume and time, including a fixed effect for each firm (firm_code). So, I use two ways to estimate parameters which are shown below:

Proc glm DATA=Sampledata;
	class time firms;
	model volume = time firms / solution;
run;

And another approach is:

proc tscsreg data= Sampledata;
	model volume = time1 time2 time_3 time_4
time_5 time_6/ fixone;
	id firms datetime;
run;

Time1, time2, ... are dummy variables based on time.

However, estimated parameters and their t values are so different in two procedures.

So, what is the problem? because both of them use OLS to estimate and these are same procedures to consider firm fixed effect.

Thanks in advance

[aminkarimid]

Hello everybody.
I have no response until now.
Is my question hard or vague to answer?
Please tell me to correct it if needs be.
Thanks