Dear Sir/Mam,
My data is panel data, secondary data. The results show that fixed effects model is appropriate. But breusch pagan/ Cook Weisberg heteroskedastic test shows that data is heteroskedastic. Should the data be homoskedastic to continue multiple linear regression. If yes kindly provide me solutions. Thank you
Heteroskedasticity is not a big problem in panel data. However, using fixed effects cannot solve endogeneity among the variables. In this world, its hardly to find pure exogenous regressors that favour fixed effects model.
I would strongly suggest to use dynamic panel model (Ardl_PMG), or system GMM where lagged dependent variable is used as a regressor in the modelling approach.
Thanks
For heteroskedastic panel data, you can use the "vce(robust)" option following a fixed effects regression to give valid standard errors and hypothesis tests.
outliers are important source of heteroskedasticity. If you model is heteroskedastic. I suggests you log your data as a means to re-scale and to reduce outlier. If this does not work you may consider using estimation techniques which are less sensitive to outliers such as robust least squares.
As you are following static panel data procedure and have used fixed effect model. Now to remove heteroskedasticity you can apply FGLS in model it will also take action against correlation. But keep in mind that it will only work while your dataset has T> N in it.
- Badges
- Science topic
- Similar topics
- Correction