0

I wrote my GMM regression for my dynamic panel data set with the xtabond command:

xtabond dependVar var1L(2/2).Var2L(2/2).Var3 L(2/2).Var4 L(2/2).Var5, lags(1) twostep artest(2) vce(robust) noconstant level(90)

Afterwards I wanted to carry out various tests (Sargan, Hansen). Unfortunately, this is not possible with the command in xtabond. So I tried to transform the xtabond into a xtabond2 command (used Roodman D. (2009) "How to do xtabond2: An introduction to difference and system GMM in Stata" for help). But I couldn't figure out how I can get the same results. My xtabond2 command looks the following:

xtabond2 dependVar L(1/1).dependVar L(2/2).var1 L(2/2).var2 L(2/2).var3 L(2/2).var4 var5, gmmstyle(dependVar , laglimits(1 1 )) gmmstyle(var1 var2 var3 var4,  laglimits(2 2)) noleveleq  noconstant level(90) twostep robust

Does anyone have a suggestion on how to convert the xtabond command to xtabond2 to get the same results? Or is there a possibility to use the xtabond command (incl. vce(robust)) for other tests (e.g., Sargan or Hansen)?

Best regards

EcoStudent
  • 1
  • 1
  • You're probably better off asking this on Statalist, I don't think there are many people on SO hanging around that do GMM often – Wouter Nov 11 '21 at 18:38
  • Thanks for the suggestion. I have already asked this on Statalist. But there was no helpful answer. – EcoStudent Nov 12 '21 at 06:54

0 Answers0