Short panel, random-effects model with heteroskedasticity, autocorrelation and cross sectional independence problems

[Juan-Jose Najera-Sanchez]

Hi everyone,
I'm working with a short panel (N=95, T=6). I have carried out the robust Hausman test (Wooldrige, 2002) and the result suggests that I have to use a RE model. Also, I already have done the Wooldridge test for autocorrelation in panel data and the result suggests that I have to handle an autocorrelation problem. The lrtest suggests that my model has a heteroskedasticity problem. Finally, Pesaran's test of cross sectional independence suggests that the data has this problem also. I have been reading and I have found that "xtpcse vardep varindep, c(ar1)" could be suitable with all of these problems but I know that this command is more suitable for long panels. I would be grateful if anybody has some advice. Thank you in advance.

[ThongKe Club]

I only use robust option to fix heteroskedasticity
command: xtreg depvar [indepvar], re/fe robust

[Juan-Jose Najera-Sanchez]

Thank you for your response. However, I think that this option deals with heteroskedasticity problem but no with the others two (autocorrelation and serial correlation). xtpcse seems to be able to solve all of the problems but I'm not sure if it is suitable for short panels. Any other hint?

[ThongKe Club]

sometime, i used to xtgls command to fix all. You can try do it.

[Carlo Lazzaro]

Juan-Jose:
- clustered robust standard errors account for both heteroskedasticity and/or autocorrelation in N>T panel dataset;
- -xtpcse- is usually applied to T>N panel datatsets; in fact, an option explicitly deployed for dealing with correlation across panels for N>T panel datasets is, as far as I know, unavailable;
-xtgls- is a pooled OLS estimator for T>N panel datasets.

[ThongKe Club]

Juan-Jose:
- clustered robust standard errors account for both heteroskedasticity and/or autocorrelation in N>T panel dataset;
- -xtpcse- is usually applied to T>N panel datatsets; in fact, an option explicitly deployed for dealing with correlation across panels for N>T panel datasets is, as far as I know, unavailable;
-xtgls- is a pooled OLS estimator for T>N panel datasets.

thank for your support, can you help me once more ?

https://www.statalist.org/forums/for...command-thanks

[Juan-Jose Najera-Sanchez]

Thank you for your comments. In this time, I was looking for literature and I have found two papers that are very interesting in this situation. FYI, the references are:
Reed, W. R., & Ye, H. (2011). Which panel data estimator should I use?. Applied economics, 43(8), 985-1000.
Moundigbaye, M., Rea, W. S., & Reed, W. R. (2018). Which panel data estimator should I use? A corrigendum and extension. Economics: The Open-Access, Open-Assessment E-Journal, 12(2018-4), 1-31.
Specifically, the second one runs some simulation with xtgls and xtpcse in a N>T situation.
Best regards.