Empirical process of residuals for regression models with long memory errors

We consider the residual empirical process in random design regression with long memory. We establish its limiting behaviour, showing that its rates of convergence are different from the rates of convergence for the empirical process based on (unobserved) errors.