A result characterizing the effect of temporal aggregation in the frequency domain is known for arbitrary stationary processes and generalized for difference-stationary processes here. Temporal aggregation includes cumulation of flow variables as well as systematic (or skip) sampling of stock variables. Next, the aggregation result is applied to fractionally integrated processes. In particular, it is investigated whether typical frequency domain assumptions made for semiparametric estimation and inference are closed with respect to aggregation. With these findings it is spelled out, which estimators remain valid upon aggregation under which conditions on bandwidth selection.Keywords: long memory, difference-stationarity, cumulating time series, skip sampling, closedness of assumptions * An earlier version of this paper was written while visiting the University of California San Diego and was presented at Texas A&M University, Universidad Carlos III de Madrid, Institute for Advanced Studies, Vienna, and the 3rd ETSERN Meeting, Nottingham. I am grateful to Patrik Guggenberger, Joon Park, Benedikt Pötscher, Philippe Soulier, Jim Stock, Yixiao Sun, and Carlos Velasco for support and many insights. Moreover, I thank two anonymous referees and Peter Robinson for very helpful comments. † RuW,