Cross sectional dependence

[Mary SMV]

Good Morning,

What is the difference between xtpcse and xtscc?

My data:

- N>T;
- Problems: cross sectional dependence, heteroskedasticity and autocorrelation (the tests were performed n fixed effects model)

1) I was using xtcss depvar indpvars, fe, but I read that this "solution" is indicate for T>N and I don´t know if I need to do more autocorrelation tests to check Ma (q) or if xtserial is enough.

2) Can I ignore cross sectional dependece and just correct heteroskedasticity and autocorrelation?

3) I made some research in order to understand better this topic and I also found the command xtpcse depvar indpvars, corr (ar1). Can I use xtpcse for N>T? In this case the model is no longer fixed effect?

Thanks.

Regards,
Mary

[Carlo Lazzaro]

Double post.
See, with reply, https://www.statalist.org/forums/for...nel-regression.