Multiple cluster SE correction panel data

[Iris.S]

Hello,
I have a panel data set with where 20 countries (panel variable) rate 49 countries on 6 variables (correlated). No time variable is included. I have to correct for these multiple clusters. How do I do this regress this data set correct? Do I use the xtreg function with random / fixed effects model and with vce (cluster)? Or the cgmreg function?

Thank you in advance!

[Jorge Eduardo Perez Perez]

What regression model are you estimating? What are the dependent and independent variables?
If I understand correctly you have 49 observations, is that right?

[Iris.S]

I have 20 countries (panel variable) all ratings the same 49 countries on 6 independent variables.
This gives me a total of a 1000 observations. I included a small sample of my data set in the attachment.

depV(ij) = b0 + b1*indepV1(ij) +b2*indepV2(ij) +b3 *indVep3(ij) +.............+b6 *indepV(ij) + e(ij)

There is no time variable.
I have been looking at fixed/random effects model with a vce (cluster) and at Mixed Models (xtmixed).
However I do not know which regression method is suitable for this dataset.

Thank you in advance!

[Jorge Eduardo Perez Perez]

I see. So you have clustering in two dimensions.
If you want to be agnostic about the structure of the clustering, I would go with the user written command cgmreg that can handle clustering in multiple dimensions.

[Iris.S]

Thank you! No additional vce (cluster) functions are required?

[Jorge Eduardo Perez Perez]

No, you don't require any additional functions other that what the command is going to ask you for. You will have to specify the dimensions of the clustering. See the help for cmgreg

[Iris.S]

Thank you Jorge.
But I cannot download the cmgreg function anywhere. Do you know where I can find it?
I have Stata 12