The TSCSREG Procedure
The TSCSREG procedure provides combined time series cross-sectional regression analysis.
Using the TSCSREG procedure you can estimate the regression parameters under three common error structures:
-
Fuller & Battese method (variance component model)
-
Parks method (autoregressive model)
-
Da Silva method (mixed variance component moving-average model)
In addition, you can specify any number of models, and you can estimate one-way and two-way fixed and random effects models. Fixed effects and random effects models handle both balanced and unbalanced data. The TSCSREG procedure offers linear hypothesis testing for regression parameters and R-squared measure for all models, F specification test for fixed effects models, and the Hausman specification test for random effects models. It provides a variety of estimates and statistics, including
-
underlying error components estimates
-
regression parameter estimates
-
standard errors of estimates
-
t tests
-
correlation matrix of estimates
-
covariance matrix of estimates
-
autoregressive parameter estimate
-
cross-sectional components estimates
-
autocovariance estimates
Statistics and Operations Research Home Page | SAS/ETS Software