Abstract

This paper demonstrates that the property of Replication Invariance, generally considered to be an innocuous requirement for the extension of fixed-population poverty comparisons to variable-population contexts, is incompatible with other plausible variable- and fixed-population axioms. This fact raises questions about what constitutes an appropriate headcount assessment of poverty, in terms of whether one should focus on the proportion, or the absolute numbers, of the population in poverty. This observation, in turn, has important implications for tracking poverty and setting targets for its reduction or elimination.