05-12-2015 04:05 PM
I am running proc glm's absorb function to control for firm fixed effects with my panel dataset. Is it possible for me to simultaneously 1) cluster standard errors by firm, and 2) adjust standard errors for heteroskedasticity using proc glm?
Alternatively, is there any other procedure in SAS that allows me to 1) control for firm fixed effects, 2) adjust for heteroskedasticity, and 3) cluster standard errors simultaneously? I have several thousand firms in my sample, so manually constructing firm indicator variables and using proc reg/proc surveyreg isn't a feasible option.