testing for moderation using interation terms
Posted: Mon Aug 25, 2014 9:50 am
Hi All,
I am using interaction terms and dynamic panel GMM estimator to test for moderation. However, I keep getting AR(2) although I deepen the lags. Without the interaction terms there is no evidence of serial correlation. How can I remedy this? Am I correct to use interaction terms with the AB GMM estimator?
I am using interaction terms and dynamic panel GMM estimator to test for moderation. However, I keep getting AR(2) although I deepen the lags. Without the interaction terms there is no evidence of serial correlation. How can I remedy this? Am I correct to use interaction terms with the AB GMM estimator?