Poverty targeting programs can in principle be assessed as projects using the tools of either cost-benefit or cost-effectiveness analysis. In practice where quantitative assessments have been made they are principally a form of cost-effectiveness calculations based on costs per unit of benefit received by the poor. Many poverty programs appear high cost due to the targeting problems of undercoverage (as some of the poor are excluded) and leakage (as some beneficiaries are not poor). Problems also arise for a number of reasons including lack of funding, difficulties in identifying who the poor are, technical errors of design and weak governance. This paper discusses conceptual issues and draws on empirical evidence from five country surveys commissioned by the ADB Institute. It concludes that poverty targeting remains important but schemes must be modest and narrowly focused to avoid past difficulties. The underlying country studies for PRC, India, Indonesia, Thailand and the Philippines will be published in full by Edward Elgar for the ADBI in early 2005 under the title Poverty Targeting in Asia.