11-12-2013 12:58 PM
I’m trying to replicate an Eviews pool equation estimation on SAS, but I can’t have equal results.. Someone can help me?
I used different procedures, (i.e. proc panel, proc tscsreg, proc reg, proc model), but results are always different from the Eviews ones.
Data: Unbalanced sample of panel data (Time series cross-section data), with 12 time series and 14 cross-section
Estimation method: Least Squares (no instrumental variables)
Estimation model: Pooled EGLS
Only common coefficients and not cross-section or period specific coefficients
No fixed and random effects
Variance structure: Cross-section heteroskedasticity. The estimates of the variances are then used in a weighted least squares procedure to compute a GLS weighting transformation, and then reestimate in iterative form until the coefficients and weights converge (CROSS SECTION WEIGHTS)
Estimates of coefficient standard errors and covariances: Cross-section SUR (PCSE) with degrees of freedom correction.