This paper presents an empirical evaluation of the targeting outcomes of the Samurdhi programme. It uses data from three sources. The first is the 1999 Sri Lanka Integrated Survey (SLIS). This survey is based on a multi-stage stratified random sample and was carried out by the Sri Lanka Business Development Centre (SLBDC). The second source is a qualitative examination of key anti-poverty programmes at household and community levels. This is based on a series of research instruments using methods of rapid ruralappraisal and participatory rural appraisal that elicit the views of community members, particularly those who are economically disadvantaged. The third source is government documents and research reports, prepared by various agencies. The conclusion is that, based on the empirical analysis of the distributional outcomes, Samurdhi does not emerge as an efficient transfer programme. It is modestly successful in reaching the intended beneficiaries, but it transfers a large portion of its resources to the non-poor. Moreover, the non-randomness of its targeting errors indicates that the programme would need extensive re-design, in order to improve its efficiency.