Diamonds in the Rough? : Repurposing Multi-Topic Surveys to Estimate Individual-Level Consumption Poverty

Traditional per capita measures of poverty assign the same poverty status to individuals living in the same household and overlook differences in living standards within households. There has been a long-standing need for a tool that enables poverty measurement at the individual level, while avoiding overly complex estimation techniques and, if possible, using readily available household survey data. An ordinary least squares-based strategy was recently introduced to estimate individual resource shares. This paper presents the theory behind this approach in an accessible fashion for those interested in individual-level consumption poverty measurement using existing household survey data. The strategy's assumptions are compared with the assumptions of the prevailing per capita approach to deriving poverty estimates. The empirical analysis presents competing individual-level poverty estimates in four diverse countries under the individual resource shares strategy versus the per capita approach. The results suggest that poverty is underestimated under the per capita approach. There is further evidence that women may be poorer than men, and that children and the elderly are disproportionately affected by poverty. However, the pursuit of the individual resource shares approach reveals cross-country heterogeneity in the extent of increase in headcount poverty estimates, and in the direction of change in headcount poverty estimates for men and women. The paper concludes with suggestions for further methodological research in this area