Imagine reading a paragraph like this somewhere in a magazine, a newspaper or a blog:
“In recent years several African countries have made strong progress on poverty reduction. Driven by solid growth, national poverty rates in Rwanda have fallen by 14 percentage points over the past decades, and countries such as Tanzania (poverty headcount of 33 percent) and Uganda (25 percent) seem in a position to almost eradicate poverty within the next decade-and-a-half. Despite this, the extent of extreme poverty, as defined by the World Bank’s $1.25 a day line, remains daunting. Rwanda’s extreme poverty rate (63 percent) is still 18 percentage points higher than its overall national poverty rate (45 percent), and in Tanzania twice as many people live in extreme poverty (68 percent) than in poverty (33 percent).”
Weird, right? And yet it is exactly this situation depicted in the below figure that poverty economists (well, at least one of them) had to explain repeatedly over the past year to Government, development partners and even other economists.
To make sense of this confusion, one first needs to know that there are national and international poverty lines. National poverty lines are usually estimated based on the so-called cost-of-basic-needs method . This method estimates, for each country, the monetary value of the level of food and non-food consumption that is deemed ‘absolutely necessary’ to sustain human life. The monetary value (in local currency) of this combined food and non-food bundle is the national poverty line. Households with expenditures lower than the national poverty line are considered poor, and the percentage of these households in the total population is the national poverty headcount (red circles in Figure 1). Households with expenditures below the food poverty line (the monetary value of the minimal required food bundle) are considered extremely poor and the percentage of these households in the total population is the extreme poverty headcount. The national poverty lines should be used to examine the extent and evolution of poverty within a given country. 
As the national poverty lines are country-specific, they can differ substantially across countries. For instance, the poverty threshold per capita per month (in 2005 dollars and adjusted for differences in purchasing power) amounts to $30 in Rwanda, $19 in Tanzania, almost $85 in Kenya, $52 in Sierra Leone, $45 in Gambia, $56 in Ghana, etc. (data from Ravallion, Chen and Sangraula ). To address these wildly different national poverty lines, an internationally comparable benchmark was needed. Voila the international poverty line, which was first set at “one dollar a day”  and was later upgraded to the less catchy “dollar point twenty-five a day ” (yellow circles in Figure 1), which roughly corresponds to the average of the poverty lines of the 15 poorest countries in the world. To have a less frugal benchmark, economists also use the $2 a day line as an “upper” international poverty line. These international poverty lines should be used to compare levels of poverty across countries.
This means that for each country there are four distinct measures of poverty: Poverty based on the national poverty line, extreme poverty based on the national food poverty line, poverty according to the international $2 a day line, and poverty according to the international $1.25 dollar a day line. For Rwanda for instance this gives the following:
The current confusion is the result of the recent labeling of the international $1.25 a day poverty line as “extreme poverty”. This means that in any given country there are two extreme poverty headcounts which can be quite far apart (in Rwanda’s case, 24.1 percent based on the national food poverty line compared to 63.2 percent for the international extreme poverty line of $1.25 a day).
The original papers introducing the international dollar a day line (Ravallion, Datt and Van de Walle, 1991) and revisiting it (Ravallion, Chen and Sangraula, 2009) do not label the line as a measure of “extreme” poverty (in fact, the first paper refers to the $ a day line as the “upper poverty line”). Rather, it is called the “international poverty line”, useful for cross-country comparisons of the extent of poverty. Recently however, this distinction got blurred, and the $1.25 a day line is now widely known, both within and outside the Bank, as extreme poverty . This has led to the situation depicted in figure 1, with national poverty based on the cost of basic needs (red circles in figure 1) in many countries being lower than extreme poverty based on the international $1.25 a day line (yellow circles in figure 1).
The $1.25 a day line is useful for comparing poverty across countries and for measuring progress towards a common goal (most notably the World Bank goal of ending extreme poverty by 2030). For the sake of clarity however, we should clearly label distinct concepts with distinct terms. Given the common use of $1.25 a day as extreme poverty nowadays, it is probably a bad idea to go and change this. So should we give another label to the extreme poor (the ones below the extreme (food) poverty line)? Looking forward to your reactions…
And, by the way, there is some good news from the current confusion: We will probably eradicate poverty sooner than extreme poverty!