Testing for cross-sectional dependence in a panel data model with serial correlation

[Victoria Consolvo]

I'd like to test the residuals of a panel regression for cross sectional dependence, but I know that there is serial correlation in the data. Is there any functionality in stata that can perform a test that is valid under these conditions? (For example, xtcsd is not valid in this case.)
Thanks!

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

I don't know of such a test. Do you know of a theoretical treatment solving this problem? If so, start a new thread explaining what you want to calculate and giving enough data so folks have something to work with.

[JanDitzen]

Victoria, out of my head, I remember that most tests for cross-sectional dependence require no serial correlation. The background for this assumption is two fold, first to make sure that the errors are iid which is important for derivation of the distribution of the test statistics, resp. for the behaviour of the cross-correlations and secondly it would be difficult to differentiate between serial and cross correlation.

My approach would be to make sure that your errors are not serially correlated. As Phil pointed out, if you tell us what kind of data you use, there might be some further help possible.