VoxEU Column Poverty and Income Inequality

Poor countries or poor people? Development assistance and the new geography of global poverty

Many poor people no longer live in poor countries. Of the 10 countries that contribute most to global poverty, six are middle-income countries. For many aid organisations, ‘middle-income’ means they no longer qualify for the same financial aid. This column argues that such a policy would be failing up to a billion people.

In 1990, 93% of the world’s poor lived in low-income countries (LICs). Now, more than 70% – up to a billion of the world’s poorest people or a “new bottom billion”– live in middle-income countries (MICs) and most of them in stable, non-fragile middle-income countries (see Institute of Development Studies for discussion).  

Many of the world’s poor live in countries that have grown richer in average per capita terms and have been subsequently been reclassified as MICs. According to the World Bank’s Atlas GNI per capita data and country classifications, over the last decade the number of LICs has fallen from 63 to just 35 countries in 2011. 

Most of the world’s poor live in countries that have moved from low- to middle-income country status since 1999 when China graduated to MIC status – notably Pakistan (2008), India (2007), Nigeria (2008), and Indonesia (2003). China is now an Upper MIC as of July 2011. This concentration of the world’s poor in relatively few countries is a key part of the story. Although 28 countries have graduated from LIC to MIC since 2000, about 60% of the world’s poor now live in just five populous new MIC countries – those mentioned above.

Indeed, of the top ten countries by contribution to global poverty only four are LICs – Bangladesh, DRC, Tanzania, and Ethiopia. In other words, most of the world’s poor do not live in countries classified by the World Bank as LICs and most of the world’s poor do not live in fragile and conflict affected states. 

This new geography of global poverty raises some basic questions about aid to help the poor of the world. Although the details vary from scheme to scheme and agency to agency, the current architecture of development assistance, especially grants and concessional loans, ‘graduates’ countries from assistance when they transfer from LIC to MIC status. As a result, all major forms of global official aid are rapidly disengaging from the bulk of the world’s poor. Is this a desirable outcome?

We outline three main reasons to continue aid to new MICs on a case-by-case basis: 

  • First, pockets of poverty call for aid no matter where they occur. It could be argued that the poor in LICs are poorer than the poor in MICs. However, this is not necessarily the case. Especially when dimensions other than income such as health and nutrition are included, the depth of poverty in some MIC enclaves can be quite severe. It can also be argued that the persistence of poverty in a MIC, when the country as a whole now has resources to address poverty, shows a lack of political will and so aid would be wasted. But political inertia can be an issue in LICs as well – poverty reduction performance can be assessed on a per-country basis, without a blanket withdrawal of assistance from all MICs.

  • Second, actions of MICs on carbon emissions, on deforestation, on water usage, etc, have consequences that spill across borders and have a global impact. Bringing MICs into global agreements, which may require them to implement policies that will have negative effects on their growth in the short term, will require resources for compensation. The large sums being discussed on funds for climate change agreements are an indication of this requirement, and of the fact that flows of assistance will be needed for MICs, including those who have recently switched or will soon switch from LIC status. To the extent that large pockets of poverty in MICs exacerbate the negative externalities, the argument for continued support is further strengthened.

  • Third, by engaging with MICs, development agencies gain knowledge, for example on implementation of social safety nets, which can then be useful for development assistance to LICs.

These issues can be seen clearly in the case of assistance from the World Bank’s soft loan agency, the International Development Association (IDA). The threshold for the agency was an income per capita of $1,175 in 2010, compared to the LIC/MIC threshold of $1,005. Access to IDA loans ceases when the country’s GNI per capita exceeds the threshold three years in a row.

If current rules of graduation continue to apply, over the next decade and a half the major sources of concessional finance to poor countries will find themselves disengaged from the bulk of the world’s poor. Moss and Leo (2010) have forecast IDA graduations on the basis of IMF World Economic Outlook growth projections. Their conclusions are sobering. By 2015, IDA will only be serving around 30 countries, mostly in Africa, and many of them in the fragile-states category. If the current level of IDA is maintained, the per capita allocation to the remaining countries will double, while the allocation will of course fall to zero for countries where the bulk of the world’s poor live – not only the top five countries mentioned above but also countries like Vietnam, Ghana, Sri Lanka, Zambia, and Bangladesh.

Faced with this situation, there are two options of moving away from a business-as-usual scenario.

  • Either the donors could declare victory and dramatically scale down contributions to IDA – enough to maintain the current per capita allocation for the countries that do not graduate.

  • Or, in recognition of the fact that poverty persists in MICs, and that the MICs are an integral part of solutions to global problems such as climate change, mechanisms could be developed to continue flows of development assistance to MICs.

We make the case for continued engagement of IDA with MICs. The LIC/MIC threshold is too sharp a cutoff when poverty persists to such a great extent in these countries. A specific proposal (Kanbur 2011) is to consider opening a second window, from the threshold to twice the threshold, where countries would have access to IDA funds but for projects specifically and sharply targeted to pockets of poverty in these countries. Further, these countries could use funds for activities that supported global public goods like climate change mitigation – this could be a separate window by itself.

While the operational details of these windows of continued engagement with MICs need to be worked out – for example, whether the lending terms would progressively harden as countries GNI per capita increased, and how a performance based allocation of IDA across countries would work – it seems clear that something like this transformation of IDA will be needed if it is not to be become irrelevant to countries where the vast majority of the world’s poor will continue to live.

The new geography of global poverty throws into question development assistance policy towards MICs. A policy of cutting, or entirely stopping, development assistance to MICs needs to be examined closely when the bulk of the world’s poor live in these countries. We show that there is no justification for a blanket exclusion of MICs from development assistance. Rather, we argue that the policy has to be crafted on a country-specific basis, taking into account the detailed nature of poverty in each MIC, the role of that MIC in addressing global challenges such as climate change, and the specific institutional and implementation context of development assistance. 


Kanbur, Ravi (2011), “Aid to the Poor in Middle Income Countries and the Future of IDA”.

Kanbur, Ravi and Andy Sumner (2011), “Poor Countries or Poor People? Development Assistance and the New Geography of Global Poverty”, CEPR Discussion Paper 8489, CEPR.

Moss, Todd and Benjamin Leo (2011), “IDA at 65: Heading Toward Retirement or a Fragile Lease on Life?”, Center for Global Development, Working Paper 246, March. 

4,619 Reads