Thirty years ago, Nanak Kakwani provided elegant nonparametric formulae for the point elasticities of measures of poverty with respect to changes in the mean of the distribution of income, thus analytically linking the poverty measures to key macroeconomic aggregates. Numerous insights are found in Kakwani’s elasticities. However, the literature on poverty and growth since then has revealed that the impacts of economic growth on poverty, as observed in practice, can be substantially lower than suggested by Kakwani’s elasticities; the reasons include rising inequality, measurement errors, discrepancies between surveys and national accounts, and changing ideas about what “poverty” means in specific contexts. Nor should Kakwani’s elasticities be treated as structural parameters. Rather, they can vary over time and place, and in systematic ways that merit closer attention.

ACCESS PAPER