The PANEL procedure is similar to other regression procedures in SAS software. Suppose you want to regress the variable Y on the regressors X1 and X2. You specify the dependent variable first, followed by an equal sign, followed by the list of regression variables, as shown in the following statements:
proc panel data=sample;
id state date;
model y = x1 x2;
run;
One advantage of using PROC PANEL is that you can incorporate a model for the structure of the error terms. It is important to consider what type of model is appropriate for your data and to specify the corresponding option in the MODEL statement. The following model estimation options are supported: POOLED, BTWNG, BTWNT, FIXONE, FIXONETIME, FIXTWO, FDONE, FDONETIME, FDTWO, RANONE, RANTWO, PARKS, DASILVA, HTAYLOR, AMACURDY, DYNDIFF, and DYNSYS. The methods that underlie these estimation options are described in this same order, beginning with the section Pooled Regression (POOLED Option).
The following statements fit a one-way random-effects model with variance components estimated by the Fuller-Battese (FB) method:
proc panel data=sample;
id State Date;
model Y = X1 X2 / ranone vcomp = fb;
run;
You can specify more than one estimation option in the MODEL statement, and the analysis is repeated for each specified method. You can use multiple MODEL statements to estimate different regression models or to estimate the same model by different methods.
The DYNDIFF and DYNSYS options cannot be combined with other estimation options in the MODEL statement. If you want to fit a dynamic panel model and perform some other estimation (such as one-way random effects), specify multiple MODEL statements.