Cross-sectional dependence and xtabond2

[Marco Astori]

Dear statalists, I would like to use xtabond2 in order to estimate a dynamic model, but I don't know if standard errors or maybe also parameter estimates could be invalidated by the problem of cross-sectional dependence: I think my dataset has it from xtcdf results related to cross-sectional independence/weak cross-sectional dependence, but I didn't find anything about some correction for maybe xtabond2 s.e. like xtscc for fixed effects. Could it be a big problem?

[David Roodman]

xtabond2 offers multiway clustering of standard errors.

[Marco Astori]

Thank you David Roodman casually didn't get this from the help

[Marco Astori]

I read the help of Stata on xtabond2 but cluster, if I'm not wrong, can't deal with cross sectional dependence, am I right?

[Marco Astori]

I read about cross sectional dependence in 2015 Pesaran's book and I think xtabond2 can deal with it through cluster(time dimension) option, is it correct?