In response to calls for finer targeting of spending to alleviate poverty in developing countries, this paper discusses the principles of targeting. The ideal solution is that all transfers go to the poor. However, this is unrealizable because of three factors: (a) the costs of administration and data collection; (b) individual responses and incentive effects; and (c) considerations of political economy. The best strategy will probably lie somewhere between the two extremes - the ideal solution and universal intervention - mediated by these three considerations. Two types of targeting, although short of the ideal may be useful in certain contexts. With statistical targeting (using indicators), programs target key indicators such as a region, occupation, or the crops grown. Self-targeting uses differences in needs, tastes, or incomes as a device for achieving self-selection by only the poor into poverty alleviation programs. Real progress in understanding how targeting works best can be made only through country specific research that quantifies the costs and benefits of targeting using data that has increasingly become available for many developing countries - and research that is sensitive to the political realities of reform.
