Abstract:
We develop a framework to assess when social programmes should prioritize expanding coverage versus improving targeting accuracy. Rather than treating universalism and targeting as a binary choice, we derive a practical rule-of-thumb based on observed eligibility rates and programme quotas to indicate when marginal welfare gains from expanding programme access may outweigh those from improved targeting. Applying the framework to Mozambique's national social pension program, we find that zero targeting effort may be optimal in over 60% of administrative areas, and reallocating existing resources using a binary RESbased rule could raise welfare by 13%. Our results support a differentiated approach to targeting: prioritize programme expansion in underserved high-poverty (rural) areas, and improve targeting effort where poverty is much lower (urban areas).