Distributed generation (DG) offers anattractive alternative to T&D expansion. Instead of expanding existing substations, building new transmission lines, and building new substations, DG can be used to accommodate new load growth and provide relief for overloaded components. Additional attractive features include: low capital cost, low fuel cost, low O&M cost, modularity, ease in siting and short lead times. This paper presents a successive elimination algorithm capable of identifying the number, size and location of distributed generators needed to optimally defer T&D expansion. This algorithm is applied to a utility test system and results are discussed.