When method matters : toward a resolution of the debate about Bangladesh's poverty measures

Measurement problemshave confounded recent attempts to assess Bangladesh's progress in reducing poverty. The issues at stake, though poorly understood, are common in poverty measurement. The authors review the issues, recommend an operational approach to resolving them with available data, and present new estimates of various poverty measures on a consistent basis for 1983-92. They then examine proximate causes of the changes in Bangladesh's poverty measures and possible implications for future assessments of the country's progress in reducing poverty under alternative growth paths. The authors argue that poverty measurement requires both normative value and judgement and assumptions about behavior to interpret available - invariably imperfect - data. Of interest for policymakers is how much bearing the analyst's choices have on key conclusions. They use the case study of Bangladesh to illustrate that those choices sometimes - but not always - affect qualitative conclusions about the sectoral structure of poverty. There appear to be considerable discrepancies among recent estimates of poverty measures for Bangladesh - even when the same survey data and a similar specification of food-energy requirements are used. The author's contend that all recent estimates are questionable from one viewpoint or another, although some problems are more worrisome than others. They propose a new series that, they hope, preserves the best features of previous work and eliminates the others. They believe that their estimates are more consistent over time and space than some others in the literature - consistent in the sense that two people with the same command over basic consumption needs will be treated the same way by the poverty measures.