There is considerable effort to reduce the software budget devoted to the maintenance of applications systems. This effort will have the effect of improving productivity of development and maintenance programmers. This means that for a given system over a given time period, the amount spent on software maintenance can be reduced significantly. The reduction might even reach the eighty per cent sometimes shown in the literature. We support and applaud the efforts to improve software maintenance procedures. Despite this type of reduction, it is not certain that organizations will spend less on maintenance relative to development. It is likely that the opposite will occur as more systems are being supported. This paper shows how that can happen through the derivation of some simple cost equations.