This paper investigates how well the gravity model explains various cross-border flows that may lead to knowledge spillovers. It turns out that the model works well for trade and telephone traffic, but less satisfactorily for merger and acquisition flows.