Wald (joint) test in system GMM using xtabond2

[beth cooper]

I need to carry out a Wald (joint) test on the inequality variable coefficient(s) in my regression using system GMM. I also have other regressors which I am using e.g. fertility ect. Whenever i execute the system gmm regression using xtabond2, next to the wald test and the p-values columns for such I just get a dot with no numbers.

If I need to carry out the Wald test to test the joint significance of all the inequality coefficients (only) what can I use? and why am I getting issues?

[Sebastian Kripfganz]

You simply need to use the test postestimation command.

[beth cooper]

How do I do that? when I write test it just says 'last test not found'

[Sebastian Kripfganz]

Assuming that x1 x2 are your inequality variables, then directly after the estimation you type the following:

test x1 x2

[beth cooper]

thank you, how do I interpret a wald test? Is it that if the p - value number is low, the joint significance is higher?

[Sebastian Kripfganz]

The smaller the p-value, the less compatible the data is with your null hypothesis of joint irrelevance. Say, if p=0.01, then there is only a 1% chance in repeated samples to observe such an extreme value of the test statistic if the null hypothesis was actually correct. This provides confidence in rejecting the null hypothesis, unless you choose a very cautious significance level of less than 0.01.