Targeting the poor: evidence from a field experiment in Indonesia

Developing country governments often lack verifiable household income information to identify beneficiaries for social programmes. This report analyses results from a randomised controlled trial in Indonesia to better understand how targeting could be improved. The study randomised three targeting methodologies – proxy means testing (PMT), self-targeting and community targeting – across 600 villages.

This impact evaluation by Vivi Alatas, Abhijit Banerjee, Rema Hanna, Ben Olken, Matt Wai-poi and Ririn Purnamasari shows that compared to PMT, self-targeting identifies poorer beneficiaries and has lower administrative costs. Community targeting, on the other hand, does less well at identifying the poor but results in beneficiaries being more satisfied with the programme.