Global production rationalization typically leads to a heavy volume of intercompany fund flows. By transferring only a netted amount of these payments, significant savings can be realized in the form of reduced foreign exchange spreads, float, and other transaction costs. However, if this netting is done by paper and pencil, opportunities to reduce transfer costs may be missed, especially where there is a complex pattern of cross-border transactions. The purpose of this paper is to show how mathematical programming can be used to design a netting system capable of minimizing the total costs involved in settling interaffiliate accounts.