topic sas regression fixed effect (by industry and year) in SAS Programming

sas regression fixed effect (by industry and year)

spasico — Fri, 21 Feb 2020 15:25:05 GMT

I'm new to SAS. I'm doing research for stock market, and I want to run a regression with fixed effect by industry and year at the same time(i.e: if there are four industry and ten years, I want to generate forty regression results for each industry in each year)

I run the code as following, which the result isn't what I want obviously.

/////

proc glm data=dac;
class industry year;
model y = x1 x2 x3 industry year / solution;
run;
quit;

/////

Is there any way to solve my problem?

Re: sas regression fixed effect (by industry and year)

PaigeMiller — Fri, 21 Feb 2020 15:34:00 GMT

I run the code as following, which the result isn't what I want obviously.

Since Industry and Year are class variables, you do get different intercepts for each industry/year combination, but the slopes of X1 X2 and X3 are constant. If you want different slopes for each industry/year combination:

model y = x1*industry*year x2*industry*year x3*industry*year/ solution;

That's one way to get the slopes for each industry/year combination, there are other MODEL statements that do the equivalent but are parameterized differently.

Re: sas regression fixed effect (by industry and year)

spasico — Fri, 21 Feb 2020 16:14:28 GMT

now I feel like I'm stupid....
This does help! thanks a lot to @PaigeMiller