xtabond2: interpretation of Arellano-Bond test using forward orthogonal deviations

[Ralph Ross Camelo]

I've been trying to understand the interpretation of the Arellano-Bond test when using FOD for system GMM. Is it the same? Normally you would expect AR(1) to be significant but since the data may be from a future time period, is this still the case? Many thanks for any help you can provide.

[Sebastian Kripfganz]

The Arellano-Bond test is always computed for the first-differenced residuals, even if the estimator uses forward-orthogonal deviations. Thus, the interpretation is unchanged.