Measuring poverty in a growing world (or measuring growth in a poor world)

This paper empirically explores the contradictions and inadequacies of measurements of growth and inequality between national accounts and surveys. It explores the reasons as to why there is divergence of results from national accounts and survey-based estimates of consumption. The paper aims to provide a sharper characterization of the measurements and offer possible remedies.

Findings include:

• consumption, measured from surveys, frequently grows less rapidly than consumption measured from the national accounts. Consistent with this general relationship, the ratio of the two magnitudes is highest in the poorest countries, and lowest in the richest. Within countries as diverse as China, India, the United Kingdom and the United States, the ratio falls over time as real income increases
• non-compliance almost certainly explains some of the discrepancies between surveys and national accounts, and whilst there are other problems with the surveys, there are serious quality issues with the national accounts estimates of consumption and GDP
• although inadequacies have been identified, there is essentially no choice but to use the surveys, because only they provide direct measures of the living standards of the poor
• an initiative is needed to provide a set of consistent international protocols for survey design, as well as deeper study into the effects of non-sampling errors, particularly non-compliance. The work program cannot be circumvented by using national accounts statistics to measure poverty