Up Front

What Do New Price Data Mean for the Goal of Ending Extreme Poverty?

Homi Kharas and Laurence Chandy

Every country in the world has a poverty line—a standard of living below which its citizens are considered poor. Typically, the richer the country, the higher the poverty line is set. The average poverty line of the poorest 15 countries in the world is used to define the global extreme poverty line—a minimum standard of living that everyone should be able to surpass. In 2005, this global poverty line was set at $1.25 per person per day. The Millennium Development Goals set out to halve the share of people living below this global minimum by 2015, and the successor agreement on sustainable development goals promises to finish the job and end extreme poverty everywhere by 2030.

How feasible is the goal of ending extreme poverty, and where is poverty concentrated? Our understanding of these issues has just been jolted by new data on prices (known as Purchasing Power Parities) for goods and services in every country in the world.

Prices are important for determining the number of poor people because when prices are low, households can afford to buy more. Price levels are also important in comparing poverty across countries. If two households in two countries have the same income levels using market exchange rates, the household in the country with lower prices will have a higher standard of living.

With the new price data now available, two big adjustments must be made to previous estimates of global poverty. First, the purchasing power of each household in every country must be reassessed, based on the new prices. Second, the global poverty line must be recomputed so that it represents roughly the same standard of living as it did before, equivalent to the poverty lines of the world’s poorest countries.

The new price data come from the International Comparison Program—perhaps the largest statistical exercise ever undertaken, involving surveys of prices for hundreds of goods and services in nearly 200 countries. The program estimates prices for 2011, and include many methodological and operational improvements on the 2005 round.

Table 1 summarizes how our understanding of the size and location of extreme poverty in 2010—the most recent year for which official global poverty estimates are available—might change as a result of the new price data. As a point of reference, column A reproduces official estimates, which use relative prices from the 2005 survey round and are updated for local inflation. This is the basis for the prevailing view that 1.2 billion people or 20.6 percent of the developing world lives in extreme poverty, with almost two-thirds of those in Asia.

Table 1.


Column B provides revised estimates in which the purchasing power of households is updated using the new price level data for 2011. This shows that the number of people living below $1.25 a day (measured in 2005 dollars) has been revised down to under 600 million because prices have been found to be much lower in most countries than was previously thought.

However, recall that the global extreme poverty line does not reflect a fixed level of dollar consumption power, but is anchored to a level of consumption power set in the developing world—the average value of national poverty lines in the world’s poorest countries. In column C we attempt to recalibrate the poverty line so that it represents the same living standards as it did previously. We cannot quantify this change precisely at present but our best estimate is that this would bring the global extreme poverty line to $1.55 (again measured in 2005 dollars). With this correction, the total number of extreme poor would be just shy of a billion.

There is one more adjustment to make. India and Nigeria are respectively home to the largest and third largest populations of extreme poverty in the world, so these countries are critical to global poverty numbers. They are also countries where existing global poverty estimates rely on problematic data. We replace India’s 2009/10 living standards survey, which was conducted at the height of an extreme drought, with a 2011/12 survey conducted when conditions were more representative. This lowers the estimate of poor people in India by 50 million. Similarly, we replace Nigeria’s 2009 survey, the results of which are riddled with inconsistencies, with a more reliable 2011 survey. This lowers the estimate of the number of poor people in Nigeria by 25 million. Column D presents our final results for which the global number of extreme poor is 870 million—a reduction of 343 million from official estimates—or 14.8 percent of the developing world.

The changes in the new estimates of extreme poverty are not equal across countries. Table 2 identifies those countries that have undergone the biggest revisions in extreme poverty based on our calculations. In terms of numbers of people, India stands out. Its population of extreme poor drops by 220 million, accounting for two-thirds of the total global reduction. Other large Asian countries also seem to have fewer poor people than was previously thought. If we instead focus on changes in countries’ poverty rates—the population share that is classified as extremely poor—then the breadth of change across different countries becomes apparent. Thirteen countries, of which seven are African, see their poverty rates revised downward by more than 10 percentage points, with the largest downward revision in Nigeria. At the other end of the spectrum, Chad sees its extreme poverty rate ratcheted up by 32 percentage points.

Table 2.



Table 3 shows how the composition of the global extreme poor has changed according to various country groupings. Among regions, sub-Saharan Africa’s share has risen most prominently, despite a downward revision in its poverty rate. The share of the world’s extreme poor situated in low income countries has risen, as has the portion living in fragile states.

Table 3.


The changes in poverty numbers that we’ve described are profound. The effect of previous rounds of the International Comparison Program on global poverty estimates have been compared to earthquakes and based on our calculations, this round seems no less dramatic. What does it mean for the movement to end extreme poverty?

  1. More achievable. The goal to end extreme poverty within a generation is noble and exhilarating. However, many commentators and experts have, in good faith, questioned whether it is feasible. For our part, we believe the goal is highly ambitious but eminently achievable. The new price data reinforce this belief, and indicate that we are even closer to the finish line.
  2. More focus. With a smaller number of people estimated to be in extreme poverty, global poverty reduction efforts must become more targeted if they are to be effective. Revisions to country estimates and the composition of the global poor should inform these efforts. This suggests that greater focus should be given to low income countries, sub-Saharan Africa, and fragile states—three groups whose share of global poverty has risen in the new calculations.
  3. More humility. Although the new round of price comparisons has been carried out with enormous care, the size of the changes are a reminder that while our knowledge about the number of people living in extreme poverty and their location is improving, it remains far from complete. Global poverty estimates are subject to large error. Furthermore, previous research of ours has shown that there is a large concentration of people whose living standards almost exactly match the global poverty line; this means that even small changes in the underlying data are likely to bring about large changes in poverty estimates. Thankfully, the International Comparison Program is moving to a rolling system for updating relative prices within the developing world, so this should be the last earthquake of this scale. We look forward to the World Bank’s official update of global poverty estimates to account for the new price data in the coming months. Yet even after this update, we should be more humble when discussing global poverty and the language we use ought to reflect this. The empirical foundation for poverty numbers is not as solid as it should be given the gravity of the issue and our common interest in improving the lives of the poorest people on the planet.