Targeting Performance under Self-selection and Administrative Targeting Methods

Abstract

We evaluate the contributions of self-selection and administrative targeting to program targeting performance using unique household survey data collected for the evaluation of a Mexican social program that has acted as a regional prototype. Both forms of targeting improve targeting performance, but administrative selection based on proxy-means testing plays a crucial role in reducing total program coverage (to meet a budget constraint) while maintaining high coverage among the lowest income groups. Its importance is reinforced when existing knowledge barriers to participation are addressed, reflecting the high application rates among higher-income groups conditional on knowledge. Differentiating transfers based on household demographics also substantially improves targeting performance reflecting the linking of transfers to the number of children. Although expansion of the program to reduce the remaining undercoverage is likely to come at the expense of targeting performance, this trade-off can be reduced by improving the underlying proxy-means algorithm and linking transfer levels directly to proxy-means scores. The latter should also improve the contribution of self-selection to targeting performance. (c) 2009 by The University of Chicago. All rights reserved.