Testing for Heteroskedasticity using a mi set panel data

[Lars Kollmeier]

Hello,

I use stata 16 and I have an unbalanced panel for 45 countries from 1990 to 2018. My analysis is aimed at explaining emissions (dependent var) with gdp per capita and inequality (gini coefficients from Solt, Frederick. 2020. "Measuring Income Inequality Across Countries and Over Time: The Standardized World Income Inequality Database." Social Science Quarterly. SWIID Version 9.1, May 2021.)
The inequality estimates and their associated uncertainty are represented by 100 draws from the posterior distribution.

My baseline regression code using the logs of gini, gdppc, squared gdppc and the interaction gdp*gini:
mi estimate, post: xtreg logco2 loggini_disp loggdppc loggdppc_sq loginteraction

Now, I am unable to conduct several tests for cross sectional dependence, heteroskedasticity, serial correlation (and unit root tests). The particular test commands used are:

• xttest2
• xttest3
• xtserial
• xtcsd, pesaran abs

My error messages are somehow similar (and I think might be related to one and the same problem (the mi set?), that is why I post several test commands).
1.
mi estimate, post: xtreg logco2 loggini_disp loggdppc loggdppc_sq loginteraction, fe
xtcsd, pesaran abs
--> Error: no; data are mi set
--> same error was displayed using xtserial

2.
mi estimate, post: xtreg logco2 loggini_disp loggdppc loggdppc_sq loginteraction, re
xttest0
--> Error: no observations

3.
mi estimate, post: xtreg logco2 loggini_disp loggdppc loggdppc_sq loginteraction, fe
xttest2
--> Error: option e not allowed
--> same error was displayed using xttest3



I hope, I was able to explain my problem well. Any help on what I have to do in order to perform these tests with my mi set data correctly are highly appreciated.

I want to thank you a lot in advance, and excuse for any inconvenience. It is my first post in this forum and it is my first time working with mi set data.

Best
Lars