Heteroscedasticity in panel data
1. Professor Najid Iqbal commented as such > One can use modified wald test for group
heteroskadesticity,, autocorrection by Wooldrigde, and for cross sectional dependence by
2. Hausman test is applicable when there is homoscedasticity
Professor Abu Subhi commented that the Hausman test could be misleading if
heteroscedasticity is present. Try other alternatives; 1. Mundlak approach or 2. -xtoverid- after
using -xtreg, re vce(cluster panelid)-.
3. How to remove hetrocedasticity problem from panel data?
Professor Saad Baloch commented as such>In case of panel data, you can use robust method
to remove hetero problem.
4. How to handle heteroscedasticity in panel data?
Professor Ali Mohamed commented as such>> If you have a heteroskedasticity problem, in
STATA, you can use the white robust standard error to correct heteroskedasticity as well as
you can use the xtgls
5. How to remove heteroscedasticity from panel data?
Professor Suborno Aditya commented as such >> If we have only heteroskedasticity problem,
using robust option should do. We must Check whether heteroskedasticity is gone using xttest3
or hettest after estimating the equation with robust option. However, if heteroskedasticity is
coupled with serial correlation or cross section dependence, different methods e.g. PCSE,
FGLS, Newey West or Driscoll Kraay should be used. Changing the data form to remove
heteroskedasticity is not a good idea or it may create additional complexity. Also there is no
certainty that heteroskedasticity will go away if form is changed e.g. Log transformation. Even
having log transformed data, you may have heteroskedasticity present.
Meo School of Research
Heteroscedasticity in the panel data
Hossain Academy Note