The latest World Bank estimates suggest that the percentage of the developing world’s population living below $1.25 a day declined from 52% in 1981 to 22% in 2008. While this indicates that there is still a long way to go in poverty reduction, the progress is encouraging. Moreover, we now also appear to be in a much better position to make such statements. The numbers above, by my colleagues at the Department Research Group, are based on over 850 household surveys for nearly 130 developing countries, representing 90% of the population of the developing world. By contrast, they used only 22 surveys for 22 countries when the first such estimates were reported in the 1990 World Development Report .
Yet despite these improvements, the “snapshot” representations of poverty still hide a lot of unresolved issues, both conceptually and in terms of data availability. This deserves much more attention than the world is currently giving it. Consider, for instance, two individuals who are currently at the same level below the consumption-based poverty line. One person has spent his whole life in poverty, while the other has just dropped below the poverty line for the first time. Both individuals most likely face quite different prospects of moving out of poverty in the future. Labeling the two individuals equally poor would seem an inadequate characterization of their poverty status. Yet that’s exactly what the numbers reported above do. Ditto much of the literature on poverty trends.
Individual lifetime measures of poverty, which account for both the duration and the sequencing of poverty spells in a person’s life time, are needed to properly assess a person’s poverty status. That’s because staying in poverty for a long time may cause permanent damage and early childhood poverty, in particular, has proven to be especially detrimental, damaging one’s cognitive development and leading to lower lifetime earnings.
However, to construct conceptually richer lifetime poverty measures, we would need panel data collected with sufficient frequency, over sufficiently long time periods and on a sufficiently large geographic scale. But we continue to lack straightforward household consumption surveys that are conducted regularly and on a comparable basis.
That hampers our ability to monitor the evolution of poverty, especially in developing countries, even when the attention is limited to static “snapshot” poverty measures derived from cross-sectional consumption surveys. It is quite sobering to realize that the $1.25-a-day poverty numbers for 2008, as reported above, were derived from a smaller base (82% of the population in the developing world) than in 2005, when 89% of the population in the developing world was covered by household surveys.
While there is a continuing need to invest more in the regular collection of high-quality, comparable consumption data , there is also a need to develop calculation methods to overcome coarse, as well as missing, expenditure data. One possibility is prediction techniques that can track poverty by making use of other indicators that are more comparable and more regularly available. That can also circumvent the need for price deflators, which have often been found to be wanting. But empirical validation is needed to ensure that they are also accurate.
The need for such empirically accurate prediction techniques raises further questions about the practical relevance of striving toward conceptually more precise poverty measures. In other words, the assumptions underpinning the development of conceptually richer and more accurate poverty measures should be empirically supported and lead to meaningfully different rankings of groups. Otherwise, it might be better to focus scarce attention and resources on improving imputation methods and data collection efforts instead, at least in the short run.
The issues discussed above are the focus of the latest Special Issue of the Journal of Economic Inequality, “Measuring Poverty Over Time ”. It will be launched on Thursday 14 June 2:00 pm at an event  hosted by the Poverty and Inequality Measurement and Analysis Thematic Group of the World Bank and co-organized with the Journal of Economic Inequality and the World Institute for Development Economics Research of the United Nations University (UNU-WIDER). Two of the papers in the issue propose concrete lifetime measures of poverty that incorporate the incidence, duration and timing of past poverty experiences. They also compare how different treatments of these dimensions affect overall poverty comparisons across different groups in the US.
Two other papers examine empirically whether snapshot measures of poverty are indeed deficient. They analyze whether past poverty spells still affect the likelihood of escaping poverty or subjective assessments of current well-being, even after we control for the person’s current poverty or income status. The last two papers address the more mundane issue of how to measure poverty, even snapshot poverty, when income and consumption data are incomplete or missing altogether. The good news is that the results suggest that it can actually be done.
Together, the findings in these papers indicate that despite substantial progress in measuring poverty, many theoretical and practical issues remain unresolved. It will be curious to see which implications the poverty experts will take away at the event on Thursday, as the findings also clearly suggest that we need to do more to measure poverty over time. The event, which is open to the public, can also be followed via webcast.
“Measuring Poverty Over Time”
Special Issue of the Journal of Economic Inequality
Thursday, June 14, 2012
2pm – 4:30pm ET
Room MC2-800, World Bank Main Complex, 1818 H Street, NW, Washington DC 20433
The event is open to the public. Guests who want to attend in person should RSVP via e-mail to email@example.com  with your full name and organization/affiliation.
To view a live webcast, click on the following links:
Blackberry (mobile) rtsp://wbmswebcast5.worldbank.org:1935/live/myStream1