Modification of Huber-White standard errors to account for clustering

[Oliver Simkovic]

Hi,

I am currently examining the effect of banking crises on life expectancy and have panel data from 130 countries over 45 years. I would like to use a modification of the Huber- White sandwich estimator of varience to correct for clustering on the country level. I found the following link (see below) that refers to Rogers (1993) and the _robust function in Stata, however as I don't have a background in econometrics I am having dificulties figuring out what exactly I need to do. Does someone have experience with this and could help me out?

http://www.stata.com/support/faqs/st...cs/references/

Thank you in advance!

Oliver

[daniel klein]

Perhaps I am missing the point, but you usually merely specify vce(cluster country) as an option to your estimation command, and that is it. StataCorp. has already implemented clustered standard errors for virtually all estimation commands, where it makes sense.

If you want something different, please provide more information.

Best
Daniel