11-17-2015 10:17 AM
Hi, I'm working with proc surveyreg to get standard errors adjusted for heteroskedasticity and clustered by day of
announcement. In some specifications I'm hoping to introduce firm fixed effects in order to control for firm heterogeneity. Do you know any way to accomplish this using proc surveyreg? It's correct to just include the firm identification variable in the class statement of the proc surveyreg?