[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Choice of clustering of standard errors

From   Alice <>
Subject   st: Choice of clustering of standard errors
Date   Tue, 21 Jul 2009 10:50:53 +0000 (GMT)

Dear Statalist,

I have a panel of bilateral trade flows that varies across industries and over time. Therefore the data vary across exporters, importers, industries and years.

I am also including
- exporter/sector/year fixed effects
- importer/sector/year fixed effects
- pair/sector fixed effects
- pair/year fixed effects

in addition to my variables of interest (that vary across pairs, sectors and years).

I need to cluster my standard errors but I am unsure about the dimension around which I should cluster. I have tried to cluster around pair/sectors, among other choices, but Stata tells me that my variance matrix is nonsingular.

Is there any standard way of deciding around which dimensions to cluster? Any suggestions?

Many thanks.



*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index