Robustness of GM-tests in autoregression against outliers

The article deals with properties of GM-estimators and GM-tests for linear hypotheses in AR(p)-processes when the observations contain outliers. In particular, we obtain the marginal distribution of test statistics, which allows us to prove the robustness of these GM-tests. The scheme of data contamination by additive single outliers with the intensity $O(n^{-1/2})$, where $n$ is the data level, is considered.