Administrators, consumers, and policy makers are increasingly interested in supported employment as a way of helping persons with severe mental illness get and keep competitive jobs. However, in an atmosphere of increased expectations for performance and declining public financing, administrators want to know the costs and benefits of different approaches before they reallocate scarce treatment or rehabilitative dollars. This article discusses the net benefits of two approaches to supported employment that were compared in a randomized trial: Individual Placement and Support (IPS) and Group Skills Training (GST). The authors analyze costs and benefits from societal, government, and consumer perspectives. Although a previous analysis showed that IPS participants were significantly more likely to find work, worked more hours, and had higher earnings, net benefits of the two programs were not significantly different. The authors also discuss some of the strengths and weaknesses of cost-benefit analysis in mental health care and suggest future directions for policy and research.