KOMMENTAR: World Bank estimates of global extreme poverty rely on many different data sources – among these are the price data that measure differences in the cost of purchasing a bundle of goods across countries.
This measure of PPP – purchasing power parity (reelle lokale købekraft) – is used to ensure that the international poverty line reflects the same real standard of living across countries.
Last year, the International Comparison Program (ICP) released PPP data from 2011, the first global update since the 2005 round.
The release of new PPP data has previously had significant effects on international poverty estimates. For example, when the 2005 PPP data were incorporated in Word Bank poverty estimates in 2008, the global count of poor increased by nearly half a billion people.
Both the World Bank and United Nations have set ambitious goals for ‘ending’ extreme poverty as defined by the 1.25 US dollar (syv til ni DKR) per day line at 2005 PPPs by 2030.
Massive changes in the count of poor resulting simply from using new price data would complicate the responsibility these institutions have to measure poverty, to track progress in reducing it and to explain methodological changes and statistical updates along the way.
In a recent paper on global poverty and the 2011 PPPs, we show that incorporating the 2011 PPP data will not necessarily change the count of the world’s poor as dramatically as seen in earlier revisions.
Indeed, under an approach we view as maintaining a comparable poverty line with the 1.25 US dollar line at 2005 PPPs, the global poverty rate for 2011 increases by less than half a percentage point.
Moreover, the path towards reducing extreme poverty to less than three percent by 2030 remains about as ambitious as indicated by projections using the 2005 PPP data.
Our findings differ from early analysis and commentary which suggested that the 2011 PPP data would decrease significantly the count of poor people in the world.
Setting a comparable poverty line
The international extreme poverty line has been revised several times in the past, but has always been anchored to the principle that it should reflect the national poverty lines of the world’s poorest countries.
The 1.25 dollar line was based on the average of the national poverty lines of 15 of the poorest countries in the world, converted to US dollar at 2005 PPPs.
We estimate that converting the value of same 15 national poverty lines (from the same countries and years) at 2011 PPPs would give an average of 1.82 dollar. An international poverty line of 1.82 /day in 2011 PPPs therefore reflects the average of the same national poverty lines which gave 1.25/day in 2005 PPPs.
We therefore believe this 1.82 dollar line for 2011 is directly comparable to the 1.25 line for 2005, and, as such, can be considered a fixed benchmark for the goal of ending poverty.
Nevertheless, we also note that statistical support for the 1.25 line (and hence also the re-estimated 1.82 line) is limited and prone to measurement error due to the small sample of old poverty lines on which it is based (on average, the 15 lines date from 1997).
To address this, we are also working to establish more robust methods for setting a line based on a much larger sample of national poverty thresholds, for which we present some preliminary findings in our paper.
No adjustments made to the PPPs for China, India and Indonesia
In addition to updating the poverty line with the 2011 PPPs, we change one aspect of the method commonly used in poverty analysis using the 2005 PPPs.
Chen and Ravallion (2010) argue that the 2005 PPPs required rural-urban adjustments for China, India and Indonesia. This was done for good reasons.
For example, it was noted that the 2005 ICP survey in China was confined to only 11 cities and therefore the 2005 PPP were considered to be an urban PPP index for China.
To correct for this, the ratio of urban to rural national poverty lines is used to provide a scaling factor to scale rural consumption such that it is expressed in urban price levels. Similar adjustments were done for India and Indonesia.
The 2011 ICP claims to have achieved a much better coverage of rural areas than the 2005 round, and we therefore argue that the adjustments based on the 2005 data are not valid for use with the 2011 PPP data.
We use the 2011 PPPs without any such adjustments. This is in contrast to initial analyses mentioned above, and a recent working paper from CGD, which leaves in the country-specific 2005-adjustments for China, India and Indonesia.
More work is ongoing to assess the sampling of the 2011 PPPs, which may lead to revisions to how the PPPs are adjusted for rural-urban price differences and biases in data collection.
But we believe the estimates in our paper based on the face value 2011 PPP data serve as an important input the this discussion.
Still one billion poor people and an ambitious goal
Læs videre på
http://blogs.worldbank.org/developmenttalk/world-about-poor-we-thought-a…
Norske Espen Beer Prydtz er økonom i Verdensbanken.
Mere om ham på http://blogs.worldbank.org/team/espen-beer-prydz