Professor in Practice in LSE Department of International Development Mark Lowcock and Research Assistant with the Center for Global Development Bernat Adrogué look at the World Bank’s International Development Association (IDA) in advance of IDA20 which will aim to support countries to recover from the COVID-19 pandemic and to transition to green energy. This article, originally published by the Center for Global Development, considers problems with IDA’s model and ways that it could mobilise resources more effectively.
IDA (the International Development Association), the World Bank’s concessional fund, was set up in 1960 to provide affordable finance to countries with the smallest economies, lowest per capita incomes, and lowest creditworthiness. The goal was to help those countries to grow faster and more equally and thereby sustainably to reduce poverty. IDA can fairly claim to have made a significant contribution to global poverty reduction over recent decades. And its choices are important, because it is the largest single source of international finance for economic and human development in the world’s poorest countries.
IDA is financed by donor contributions (effectively grants to the World Bank, though in principle, were IDA ever wound up, the remaining funds would go back to the donors), topped up more recently by repayments from borrowing countries of loans taken out long ago and borrowing against the equity IDA has built up over the decades. The first fund agreed in 1960 amounted to $900 million. It has subsequently been topped up 18 times, roughly every three years. Each replenishment brings not just more money but also adjustments to IDA’s policies and priorities.
A new replenishment (IDA20) is under negotiation now, and due to be finalised next month. So it’s timely to ask what would be desirable in the next phase.
The world has changed since IDA was established. In 1960, most of the global population was living in extreme poverty. Most people were hungry much of the time, many children died in infancy, most did not go to school, and many of their mothers died in childbirth. Average life expectancy was 52 years. Now, fewer than 10 percent of people live in extreme poverty and average life expectancy is 72.
The pandemic has dented some of these achievements. There is a growing and dangerous divergence between the low-income countries IDA was intended to serve, especially the subset mired in conflict and humanitarian crisis, and the rest of the world. Some of this predates the pandemic, but COVID-19 has made it worse.
Many countries no longer need IDA, but some need it more than ever. Is it still fit for purpose? Does it focus on the right countries? Does it allocate its resources to them in the best way? Does it offer them the right products? Does it prioritise the most salient issues? Does it have systems and processes that maximise its effectiveness? Do the people responsible for its decisions have the right skills and experience? And how big should it be given today’s challenges?
Is IDA focused on the right countries?
The original rationale for IDA—focusing on the poorest countries who can’t get finance elsewhere—still makes sense. One implication is that, as they develop, IDA recipients should “graduate” (including to borrow from the World Bank’s main lending vehicle, the International Bank for Reconstruction and Development, or IBRD).
Between IDA’s founding and 2020, 37 countries graduated. Another 9 graduated but suffered development reverses and re-entered. (Other re-entering graduates, like the Philippines and Egypt in 1990s, later graduated again.) Eighty-two countries were recipients in 2019, some of whom received no new money, just disbursements from previously approved allocations. There are 30 fragile and conflict-affected states among the current recipients.
Formally, the World Bank’s graduation threshold for IDA is an annual per capita income of $1,205, but it is not applied. Of the 74 countries in principle eligible for new IDA resources, at least 49 have per capita incomes above the threshold. (It might be 51. There is not enough data on Eritrea or Syria to be sure where they currently stand.)
There may be an argument for countries with per capita incomes above the threshold retaining access to IDA if they can’t get resources elsewhere. But for a good number, that’s not the case. A glance at data from credit ratings agencies (S&P, Moody’s, and Fitch) suggests that some 25 IDA recipients have credit ratings in the B- to BB- range—comparable, in other words, with the likes of Brazil, South Africa, Turkey, Oman, and Bosnia.
Time lags in the IDA system reinforce the issue. Countries moving beyond IDA still get disbursements for years after graduation. India, for example, which graduated in 2014, received more than $1 billion in 2019. Substantial disbursements five years after graduation is common.
The effect of retaining in IDA relatively creditworthy countries which have developed sufficiently to exceed the graduation threshold is to reduce the resources available to countries with more severe problems (which are also much further from creditworthiness).
Does IDA allocate its resources in the best way?
IDA is more transparent than some other funds (for example, IBRD) on its process for allocating resources among eligible countries. Each eligible recipient receives an allocation significantly based on a formula linking a country’s population, GNI per capita, and a “country performance rating.”
The rationale for the country performance rating, which was introduced in 1977, was that resources should be focused where they could make the biggest impact on IDA’s goals, and that a country’s economic management and policies, and the quality of its governance and institutions, affected how well resources could be used.
The consequence of this approach is to skew resources away from fragile and conflict-affected states, where public sector management and institutions are typically weak, towards countries with fewer problems. Currently, criteria covering public sector management and institutions (for example, quality of financial management, revenue administration, and quality of public administration) account for more than two-thirds of the country performance rating. So a country could have excellent fiscal, exchange rate, monetary, debt management, and business regulation policies, but still end up with a poor country performance rating and hence a low IDA allocation.
A country’s IDA allocation is the pot from which commitments to projects and programmes are made, with decisions taken by World Bank management in consultation with the authorities in the recipient country. What matters, however, is not just the allocation or commitment of resources but whether they are actually disbursed. IDA has grown over the last 15 years and disbursements to non-fragile recipients have nearly doubled. Disbursements to fragile countries, however, have grown much less: in 2019 they were 44 percent higher than in 2007. So the neediest countries are losing out on this score too.
One of the reasons disbursements to fragile countries lag is that IDA’s model is based on lending to governments. At the same time, IDA rightly wants to uphold high fiduciary standards and minimise leakage through corrupt government systems. But relying on government systems while insisting on high fiduciary standards too often proves unworkable: IDA approves proposals which then get stuck and fail to be implemented, exacerbating the ex-ante resource allocation biases.
The result of all these features is that IDA is less focused on the neediest of its recipients than are other providers of official development assistance (ODA) to those countries. Attempts to remedy this in negotiations on replenishments over the last 10 years have not so far succeeded. While (in terms of 2019 disbursements) 31 percent of IDA resources went to fragile countries, 41 percent of ODA for IDA countries went to the fragile group.
Does IDA have the right products?
Most IDA resources are deployed through projects and budget or adjustment financing. Projects and adjustment financing are good, but one of the main things poor countries need is faster help in a crisis. The pandemic illustrated that. While G20 countries were able to throw 20 percent of their national incomes in 2020 into often novel schemes to protect their economies and citizens against the contraction COVID-19 imposed, low-income countries could mobilise only 2 percent of their (much smaller) incomes.
IDA’s tool kit for crisis finance is inadequate. Once a country’s IDA allocation is determined in each replenishment round, there is limited scope to change it to respond to new problems. In 2010, under the 16th replenishment of IDA, a crisis response window was created. But its funds too are modest: $2.5 billion of the $82 billion available under IDA19. The scope to adjust the use of resources within the fixed country allocation in the light of a new crisis is also limited. In 2011 (some years after the global financial crisis) the World Bank board approved a new mechanism allowing IDA countries to access up to 5 percent of their undisbursed balances and reallocate them for crisis response. That is too small: shocks in fragile low-income countries typically require larger adjustments.
IDA also needs to be more flexible in choosing delivery partners, given the track record of slow disbursement to governments of fragile and conflict-affected countries. One option is greater use of other intermediaries with proven delivery capacity. IDA has experimented with this in a few countries (including Yemen and South Sudan), mostly partnering with UN agencies and nongovernmental organisations. It would benefit the populations of such countries for this to be done on a larger scale.
Another issue is the terms on which IDA finance is available. Historically, IDA has been a lender, with low interest rates and long repayment periods. More recently, however, allocations have increasingly been on grant as well as loan terms, as Scott Morris and others have noted. The need for more generous terms was growing even before the pandemic; the pandemic has made it more acute, not least in order to address the divergence problem mentioned earlier. For the recipients, the balance between grants and loans depends on the extent to which they are in debt distress. Whether the level of debt distress should be the sole or main determinant of the terms IDA offers is debateable. The terms could also vary according to what the resources will be used for. There is, for example, an argument for greater use of IDA’s ability to provide grants in financing global public goods, like vaccinating against COVID-19 or hosting large numbers of refugees.
Does it prioritise the most salient issues?
IDA’s country-driven operating model means that sectoral priorities within each country’s allocation are significantly determined by the authorities in the recipient country. It would be undesirable to move too far from that approach.
In the immediate future, the top priority should be to contain the damage from the pandemic and then to support recovery. But as we have said, there are important questions as to which countries IDA should focus these efforts on.
As others from CGD have argued (Amanda Glassman and Masood Ahmed), there is (separate from the question of terms just discussed) a good case for IDA to play a stronger role on global public goods, in particular health and climate. That could include a substantial and over time growing ring-fenced window, which could also earmark more resources for regional projects. That might, however, be more an issue for IDA21 than IDA20.
Conversely, as Charles Kenny has argued, recent experiments in using IDA resources to subsidize the International Finance Corporation should be dropped for the time being.
Does IDA have systems and processes that maximise its effectiveness?
IDA has a sophisticated project cycle and highly skilled technical staff, but incentives are skewed towards project approval rather than implementation or the delivery of outcomes. World Bank staff observably like analysing and debating the issues. They appear less motivated by delivering results promptly. One of the most common private complaints of ministers in IDA-eligible countries is that projects take too long to approve and implement.
Bureaucracies tend to accrete processes—and therefore delay—and the World Bank board and management have often made that worse by adding new requirements in response to high-profile problems (for example, corruption scandals). It may be time for a root-and-branch process review motivated by the goal of delivering faster results in challenging environments.
As noted above, IDA can be too reliant on government mechanisms for delivery—to the disadvantage of those with good policy intent but poor delivery capacity. The prevailing fiduciary framework makes it harder for IDA to be effective in challenging environments. The solution is not to weaken that but to reduce risks through different choices of delivery modalities and implementation partners.
Do the people taking decisions have the right skills and experience?
Our view is that many development agencies are too centralised, and they would make a bigger impact if they delegated more of their work to field staff, allowing them broader scope (within an appropriate centrally determined framework) to respond flexibly and with agility to local circumstances and events. That critique applies to the World Bank: too much operational detail is determined in Washington by the board and senior staff.
These concerns have been recognised to a degree. Replenishment negotiations over the last decade have seen increased debate on the numbers of people the World Bank has dealing with fragile and conflict-affected states; whether they have the rights skills, aptitudes, level of authority, and incentives; and whether they are in the right places. But progress has been limited and incremental. Bigger changes are required.
How big should IDA be?
IDA20 has been brought forward by a year as part of an agreement under which use of IDA19 resources has been front-loaded to help recipients cope with the pandemic. A respectable case can be made for a large replenishment. However, IDA has over the decades built up large equity which is sitting underused on its balance sheet, as Clemence Landers argues.
IDA is substantially financed from the ODA budgets from which countries providing most of its resources pay their contributions.
Our starting point is that ODA should be more concentrated in the countries with the biggest poverty and development problems and least access to capital markets.
It follows that how big IDA should be in the future ought to depend on how well it supports people in those countries relative to other uses of ODA.
IDA has much to be proud of. But it is now underperforming in the countries with the biggest challenges. If it wants to retain its preferred status as a beneficiary of donor resources in the future, it needs a better offer to the neediest countries: not just in terms of the share of the money they get, but the products, processes, systems. and staff too.
Our guess is that IDA20 will come in at around $95 billion, to be committed over three years from July 2022 (so around $31.5 billion a year). That will compare with IDA commitments of around $35 billion a year for the preceding two years. New donor contributions will represent a smaller share than in previous replenishments: for IDA20 they may amount to $25 billion or so. More of the total will come from accumulated equity and leveraging the balance sheet, which the donors will push for. We are, however, not optimistic that there will be much progress on the policy switch to focus more on the neediest and most fragile of IDA’s recipients. But these issues are not going away.
Ahmed, M. (2021). Don’t Lose Sight of the Real Business of the IMF-World Bank Annual Meetings. Center for Global Development. https://www.cgdev.org/blog/dont-lose-sight-real-business-imf-world-bank-annual-meetings
Glassman, A. (2021). Four Pandemic Asks for IDA20 Replenishment at the World Bank. Center for Global Development. https://www.cgdev.org/blog/four-pandemic-asks-ida20-replenishment-world-bank
Kenny, C. (2021). Go Big on IDA, but Leave Out the Private Sector Window. Center for Global Development. https://www.cgdev.org/blog/go-big-ida-leave-out-private-sector-window
Landers, C. (2021). IDA-20: Donors Must Go Big, and IDA Must Too. Center for Global Development. https://www.cgdev.org/publication/ida-20-donors-must-go-big-and-ida-must-too
Morris, S., Landers, C., Gardner, A. (2020). More World Bank Borrowers Will Need Grants, Not Loans. As a Result, More World Bank Donors Will Need to Pony Up. Center for Global Development. https://www.cgdev.org/blog/more-world-bank-borrowers-will-need-grants-not-loans-result-more-world-bank-donors-will-need
The views expressed in this post are those of the author and in no way reflect those of the International Development LSE blog or the London School of Economics and Political Science.
Photo: The World Bank’s ‘IDA Impact: A Call to Action’ meeting in 2019. Credit: World Bank Group on Flickr.