Hello researchcers
Can you help me to understand that the DOLS model is suitable for cross sectional dependence panel data or not. The sample of my research is CSD through individuals (nationals).
If not, which model i can use for cointegration estimator for CSD panel data with T=18, N=25 (hesterokedasticity, serial correlation and endogeneity as well).
Thank you so much.
Sau Mai
Dear Sau Mai,
As far as I know, DOLS estimator is not suitable for cross sectional dependence panel data. To overcome the problem of cross-sectional dependence in panel data analysis, you might consider using some estimators like AMG, FGLS, PCSE, Driscoll-Kraay standard error regression, Dynamic seemingly unrelated regression, or CUP-FM, CUP-BC.
Best regards,
Quocviet Bui
Dear Sau Mai,
I think STATA software is not appropriate for CUP-FM and CUP-BC estimators. You might consider using GAUSS software. The GAUSS code for these estimators can be found at the author's website: https://sites.google.com/site/chihwakao/programs.
Regards,
Quocviet Bui
Dear Sau Mai, SUR model can be run by using STATA. Note, however, that I talked about the DSUR (Dynamic seemingly unrelated regression) model, not SUR model. About the second question, to my best knowledge, FGLS and PCSE can deal with autocorrelation, heteroskedasticity, and cross-sectional dependency, not endogeneity problem.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Sectioning