DIY: Measuring Global, Regional Poverty Using PovcalNet, the Online Computational Tool behind the World Bank’s Poverty Statistics


This page in:

World Bank Group President Jim Yong Kim recently announced ambitious goals to end poverty and boost shared prosperity, with a target to reduce the percentage of absolute poor – those living on or less than $1.25 a day (in 2005 PPP) – to 3 percent by 2030. The Bank, he said, will also focus on expanding opportunities for those living at the bottom 40 percent of the income or consumption distribution in each country.

That exercise requires credible statistics to back it up. And for more than 12 years now, the source of World Bank’s poverty statistics is the online computational tool PovcalNet, which I manage with a group of World Bank researchers, with inputs from all six regions and many countries’ statistical offices. PovcalNet is used to calculate the Bank’s official global poverty measures, using poverty lines of $1, $1.25 or $2 a day. It has provided poverty estimates for the first Millennium Development Goal to halve the 1990 poverty rate by 2015 and various World Bank flagship reports, such as the World Development Report, World Development Indictors and Global Monitoring Report.

More importantly, PovcalNet allows users to do their own calculations of global and regional poverty measures, using the same methodology as we do, but different poverty lines or country and regional groupings, or alternative purchasing power parity. In 2012 alone, PovcalNet completed 7 million analytical computations for its users around the world.

PovcalNet was created in 2001, in response to increased demand for international poverty estimates, especially after the Millennium Development Goals were established. Many people criticized the World Bank’s poverty estimates without understanding how these numbers were calculated. Some complained that the Bank overestimated poverty to seek more donor funding, while others said the Bank underestimated poverty to show progress in poverty reduction.

With PovcalNet, we made our poverty data and methodology transparent, so people can duplicate and monitor our results, and better understand the methodology and the limitations of underlying data. Along with our operational colleagues, we work with each region and country to improve the consistency and reliability of the data and results.

What’s unique about PovcalNet is that, unlike other web applications, PovcalNet not only provides static statistics – pre-calculated poverty and inequality estimates at regional or country level ¬– but lets the user create new poverty measures at different poverty lines or using alternative purchasing power parity. With a special computing engine, calculations of all distributions and aggregations at user-specified levels are performed in a split second. Compared with other online tools, where the user’s request is executed manually and results are sent by email, PovcalNet offers on-demand statistics without delay. Along the way, we have provided technical support to millions of our users every day.

In recent years, we have added more indices on inequality and improved the output function to meet clients’ needs. For example, you can do simulations to explore what growth rate your country or region needs to reach a certain rate of absolute poverty. Or, with the overall survey mean and decile distribution provided in PovcalNet, you can easily compute the mean income or consumption for the bottom 10%, 20%, 40%, and so on, for each country or household survey in our database.

PovcalNet is a prime example of how the World Bank can serve as a knowledge bank and it demonstrates that open research and open data can increase our institution’s credibility. We hope you will find it useful.


Shaohua Chen

Lead Statistician, Development Research Group, World Bank

Join the Conversation

Sanat Mishra
April 30, 2013

It is really praiseworthy for the endeavors taken by WB in supporting the people living on the fringes. Essentially, the poor people live in complex, diversified and risk prone (CDR) agro-eco situations, thereby requiring location-specific and regionally differentiated strategies involving the community. Hence, any research project should have enough space for the community to participate and share their experince.
Sanat Mishra

Charu N Thakur
May 09, 2013

The exercise is massive and no doubt sher numbers are mind boggling. But two things;
1. The time taken to compile, analyse and intetrpret data, may lead to change in the numbers. Especially in a country like India where population and poverty statistics change very hour.
2. Poverty indices deal with household income, but the individual expenditure varies a great deal within one family. Female malnutrition in India is not so much as to extreme poverty, as. It is due to lower social status of women in general. Thus a family not 'poor' as per the standard definition may have poverty stricken members.