How to obtain cluster robust standard errors in spxtregress?

[Sachin Gupta]

Hello, I am using spxtregress to estimate a spatial model with panel data. The method does not allow the option vce(robust) or vce(cluster). Is there a way to obtain standard errors that are clustered at the cross-sectional units? thank you.

[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.

Often, if Stata doesn't provide a standard option there is a good reason for it. I don't know this to be the case in spatial models, but it is a possibility. I don't know if the weighting matrix handles this, or if you can restructure your data to get it estimated by xtreg..

[Simeon Haefele]

Hi Sachin,

I currently have the same question - did you find any information on it elsewhere in the meantime?

Best,
SImeon

[Sachin Gupta]

hi Simeon, I did not get an answer. I did look at Lee and Yu (J of Econometrics 2010) that the procedure is based on. The estimator is a quasi-maximum likelihood estimator. As best as I could tell from the paper, the standard errors as computed in the paper do not require clustering. But I could never determine whether the spxtregress follows the procedure in the paper to compute standard errors. If you find out more I would love to hear as well. Sachin

[Brenden Beck]

I have this question, too. It seems like there ought to be a way to manage serial autocorrelation in spatial models.