Our Programmes



Sign up to our newsletter.

Follow ODI

Poverty is about nutrition, mortality, and other measures that go beyond money

Written by Andrew Norton


​How many people live in extreme poverty? That apparently simple question has profound implications for policies in poor countries, for aid, and for the post-2015 international development goal of poverty eradication. Yet this is a simple question with no easy answer.

Poverty numbers are a construct. Monetary poverty measures like the global US$1.25 a day are based on (hopefully) representative household survey information about what people consume. Because not everything that a household consumes is purchased in the market some things have to be assigned a monetary value – food grown by the household is an example. Once you have derived the number of people living on less than US$1.25 a day, deriving the percentage of the population living in poverty is a matter of basic arithmetic isn’t it?

Well, not exactly. Because the same amount of money buys different amounts of goods and services in different countries, deriving a global poverty number requires the development of a common measure. Enter ‘purchasing power parity’. This is calculated through a massive global exercise called the ‘International Comparison Programme’).  Conducted every 6-7 years to compare what a given amount buys in different countries, the ICP provides data on millions of prices across a wide range of countries (199 in 2011), which is used to provide  harmonised measure of monetary value.

 All kinds of problems are entailed in the exercise: how do you compare the price of millet in Mali with the price of bread in London?  But if you want to have a global poverty line you have to have a common monetary metric. 

Updating  ‘purchasing power parity’ numbers is a technical exercise that has profoundly important implications for our understanding of poverty. The last update  in 2008 (based on 2005 data) created a minor earthquake.  The snappy and intuitive ‘dollar a day’ figure that underpinned the MDG target adopted in of halving poverty  (a figure derived from the World Bank’s 1990 World Development Report) was updated to $1.25 – and poverty numbers rose by some 400 million (mostly because of adjustments to the data on China).

The ICP  earthquake just happened again and is threatening to register an even higher measure on the Richter scale. The  new PPPs (based on 2011 data) indicate that consumption levels in very many countries are higher than previously thought. Their release by the ICP has prompted a major outbreak of blogging by poverty number-crunchers in Washington DC. . 

So what does the new ICP data tell us? First, that the fast growing economies of Asia are much bigger than we thought: China’s economy represents 87% of the US GDP as against 43% in 2005.  Second, that the profile of global poverty also looks different.

How different depends on some tricky technical issues. If you hold the global poverty line at $1.25 as CGD’s Sandefur, Kenny and Dykstra did in an early assessment then the global poverty falls from 19.7% to 8.9%.   Using the same data Homi Kharas and Laurence Chandy of the Bookings Institution did a similar exercise but came up with different results. The reason: they followed the established practice of re-calculating the poverty threshold by using the average poverty lines of the 15 poorest countries – an operation which raises the poverty line to US$1.55.  Factoring in the new PPPs and the revised poverty line produces an extreme poor number of 870 million—a reduction of 343 million from official estimates (with a particularly large reduction in India). 

The rush to crunch the numbers based on the crude first cut of the ICP results is not necessarily a good idea.  The ICP numbers may well need  further adjustment  to capture aspects of the consumption of the poor which may be different from the norm (a point made by Martin Ravallion in a comment on the CGD piece).  There are also questions over the use of the average poverty line of the poorest 15 countries to set the global threshold. There is a rationale for the approach. Below a certain level (around  $2 a day average per capita consumption in 2005) all country poverty lines are more or less the same.  Above that level poverty lines rise and diverge.  As poor countries become richer, the poverty line will move upwards – potentially making poverty eradication a destination that we  never reach. 

This is a debate set to run and the World Bank (which has been treated as the final authority up to now) has wisely indicated that it does not want to rush the work involved. 

My guess is that, when the dust settles, we will still have a ‘global poverty line’. Hopefully, we’ll also emerge with a more realistic view of the real challenges implied by a commitment to poverty eradication, along with a more diverse set of measures for assessing progress. More weight should  be given to wider human well-being indicators, in areas such as such as nutrition, maternal mortality, child mortality, and to  other measures that go beyond a monetary minimum.  And we should remember that the global poverty line is just that – it should not take over at the national level from a process of defining the meaning of poverty that ought to reflect national values and a process of national debate.