Priming the Pump for a Successful IDA22

IDA’s next replenishment, launching soon, will be among the most consequential multilateral negotiations this decade and IDA22 deliberations will need to show strong ambition. Today, most low-income countries (LICs) remain heavily dependent on IDA, having been battered by repeated shocks, including COVID-19, the war in Ukraine, the conflict in the Middle East, and numerous disasters exacerbated by climate change (see Figure 1). Meanwhile, bilateral aid is shrinking while poor debt dynamics continue to limit the potential for growth-generating investments.

Despite the anticipated high demand for IDA, we also doubt that donor dynamics will be conducive to another record-setting replenishment. IDA21 hit a record headline target, largely driven by increases in market borrowing, but donor contributions were flat, continuing a downward trend that began fifteen years ago. As a result, IDA stakeholders need to think now about what is needed to drive a successful outcome. With this blog and the accompanying paper, we aim to spur debate well before negotiations launch in early 2027 (IDA22 would take effect for a three-year cycle starting in July 2028). Our hope is that stakeholders will coalesce around a series of reforms to address deficiencies in the current model that have led to a surfeit of IDA-eligible countries above the income threshold, constraining allocations for countries most in need, and support an increased allocation of World Bank resources for concessional finance to help meet rising demand. Meeting these objectives would require changes at both IDA and IBRD, the World Bank’s arm that lends to more creditworthy countries. In our paper, we make the case for three interrelated proposals:

1. a more robust, rules-based IDA graduation process;
2. updated creditworthiness assessments at the IBRD to support an increase in the pace of IDA graduations; and
3. the introduction of a new semi-concessional financing instrument for lower-middle-income countries (LMICs) funded through IBRD net income.

Most IDA borrowers surpass the income threshold

More than half of IDA-eligible countries sit above the FY26 operational income cutoff of $1,325 GNI per capita (see Figure 2). These borrowers are still on IDA's books because they are deemed insufficiently creditworthy for IBRD. The result is a balance sheet increasingly populated by “gap” countries (above the income threshold but not yet creditworthy) and “blend” countries (creditworthy enough for some IBRD borrowing). 

Between FY20 and FY25, gap and blend countries absorbed 42 percent of IDA resources, leaving just 54 percent for IDA-only countries (the remaining 4 percent went to small economies and regional projects) (see Figure 3). Blend countries can borrow from both IDA and IBRD, although in practice, IBRD access is at relatively low levels: between FY20 and FY25, IDA provided an average of $7.3 billion in funding to blend countries per year, compared to only $1.7 billion from IBRD.

IBRD is bigger and more profitable than ever

Unlike IDA, IBRD does not depend on donor largesse and is not capital constrained; its loans are close enough to market rate to make the institution self-sustaining. IBRD’s equity-to-loan ratio is around 21.4 percent, although the minimum threshold is 18 percent, which translates into headroom of about $100 billion. We estimate that IBRD could increase its lending by $10 billion a year over the next five years and remain above the 18 percent threshold.

Moreover, in October 2025, Standard & Poor’s estimated that the major multilateral development banks (MDBs), including the World Bank, possess an extra $600–800 billion in lending capacity that could be deployed over the next ten years without putting their ratings at risk. (Their figures are not broken down by MDB.) In late 2024, Fitch published analysis affirming that twelve MDBs, including the World Bank, could increase lending by a combined $480 billion before they would risk rating downgrades, based on changes to Fitch’s supranationals rating criteria. Among the MDBs, it found that IBRD could increase lending by approximately $117 billion.

IBRD is increasingly profitable too, having brought in a record $2.4 billion in allocable net income in FY25. But transfers to IDA have been falling as a share of net income and IBRD has been channeling more lending to countries above its graduation threshold (i.e., the “graduation discussion income” level or GDI), leading to a decline in the share of funding to lower-middle-income countries since 2018 (see Figure 4).

We propose three reforms to boost support for the poorest borrowers at both IDA and IBRD.

Reform #1: Make IDA graduation rules-based

The first fix is to make the graduation process more transparent and rules-based, starting with the transition from gap to blend status. Today a gap country becomes a blend only when IBRD runs a creditworthiness assessment, a request that must be initiated by the borrower. This policy is leading many gap countries to remain as IDA-only borrowers for too long. Of the thirteen gap countries, ten exceeded the income threshold before 2020, one as early as 2005 (Honduras). Kosovo has never been below the income threshold, and its current GNI of $6,910 is more than five times that threshold (see Table 1).

While the discretionary nature of the transition from gap to blend status is problematic, the real bottleneck is the slow rate of graduation from IDA to IBRD by blend countries. IDA access for blend countries is intended to be temporary, and the country itself is expected to “undertake adjustment efforts designed to establish or strengthen creditworthiness as rapidly as possible.” But in fact, many blend countries have remained on IDA’s balance sheet for decades (see Table 2), and there is no institutional mechanism for countries to establish a transition path. This reality reflects an institutional discrepancy: there are two criteria for IDA eligibility (per capita income and creditworthiness), but numerous criteria for graduation, including GNI per capita, time since a country has exceeded the income threshold, creditworthiness, poverty headcount, population, exports/GDP, life expectancy, urbanization, fragility, institutional development, and resource rents. The implication is that if a country exceeds the IDA cutoff and is creditworthy, other factors justify continued access to IDA funding. This, in turn, implies that IDA eligibility is much more complex than an income threshold and creditworthiness determination. This inconsistency needs to be reconciled.

In addition, we recommend that IDA more systematically advise and track country efforts to graduate and that shareholders be kept apprised of progress (e.g., in the Mid-Term Review). Country partnership frameworks (CPFs) for blend countries should include potential paths to IDA graduation and IDA’s role in setting them on that path. Each pathway should include a set of criteria against which progress is measured (e.g., bond issuances) and a borrowing trajectory that includes a growing share of IBRD resources. We assessed blend country CPFs for discussions of graduation prospects and found that only one—Pakistan—even mentioned graduation.

Reform #2: Revise IBRD’s creditworthiness assessments

The major reason gap and blend countries remain on IDA’s balance sheet is because the World Bank deems them insufficiently creditworthy to borrow partially or exclusively from IBRD. The process for determining creditworthiness is not public but we found that some countries deemed not creditworthy have demonstrated “reasonable” market access as defined by the IMF. Specifically, seven gap countries, which are by definition not creditworthy, have had some degree of market access, as have nine blend countries, which are considered partially creditworthy (see Table 3). We are not suggesting that all sixteen countries are creditworthy but two gap countries—Benin and Honduras— have composite credit ratings of B+, on par with many IBRD borrowers.

Therefore, we recommend that IBRD’s creditworthiness assessments be revised to better take account of real-world developments. In addition to adopting better and timelier market access proxies, this should include giving more weight to default and recovery rates for sovereign loans. Data from the Global Emerging Markets Risk Database (GEMs) consortium of development finance institutions, which includes the World Bank, shows that loan repayments have been consistently near or at 100 percent, with average default rates of only 0.77 percent. High repayment rates are to be expected—the MDBs benefit from preferred creditor status and once a country defaults to an MDB, access to capital markets is usually cut off. As of March 31, 2026, IBRD had $283.6 billion of loans outstanding, of which 0.5 percent were in nonaccrual status, all related to Zimbabwe and Belarus; notably, both were IBRD—not IDA—borrowers at the time of default.

Reform #3: IBRD should offer semi-concessional terms to some borrowers

The third change would expand the volume of concessional resources to help address high demand among low-income countries. It would also address a perennial complaint of IDA graduates: the sudden cutoff of access to concessional finance. The spread between IDA and IBRD varies widely depending on the interest rate environment; the higher the spread between IBRD and IDA, the greater the pain point for new IDA graduates (see Figure 5).

Therefore, our final proposal is that IBRD introduce a new instrument that is less concessional than IDA but more generous than what is currently available. For example, a new instrument could feature a 3 percent interest rate with a 20-year maturity and five-year grace period. The 3 percent interest rate is higher than the 2.1 percent rate that IDA offers but lower than the current IBRD A terms, and is fixed, not variable.

The semi-concessional resources could support new IBRD borrowers and blend countries that remain on IDA’s balance sheet, boosting access to affordable sources of finance across both accounts.

We estimate the cost to IBRD would be around $1 billion per year in net income to mobilize $9–11 billion in IBRD lending, including around $7 billion in IBRD’s new semi-concessional instrument. Given IBRD’s record profitability, this is affordable (see Figure 6).

This proposal represents a major departure from IBRD’s current model and would entail careful vetting with credit rating agencies and shareholders. But it falls well short of a unified balance sheet approach (e.g., the Asian Development Bank model).

Conclusion

IDA stakeholders will face a significant set of challenges next year, stemming in part from a fraught external environment but also from institutional flaws that enable countries to tap IDA resources long after they have moved above the income threshold and despite evidence of sustained market access. Fortunately, there are credible solutions for making both IDA and IBRD more responsive to their poorest borrowers. To recap, these are: 1) a more robust, rules-based IDA graduation process; 2) updated creditworthiness assessments at IBRD to support an increase in the pace of IDA graduations; and 3) the introduction of a new semi-concessional financing instrument for lower-middle-income countries funded through IBRD net income.

We recognize that the implications of moving countries between balance sheets are complex: putting IDA’s more creditworthy borrowers on IBRD’s balance sheet would free up lending for the poorest but also lead to a decline in credit quality and future reflows. This, in turn, would affect IDA’s market borrowing program. IBRD’s credit quality would decline as well, and the potential need for extra provisioning could eat up IBRD’s headroom more quickly. On the upside, the addition of new borrowers for IBRD would create a more diversified portfolio and increase lending and net income (which could be transferred back to IDA or used to finance the new semi-concessional instrument).

What is clear is that IDA’s and IBRD’s balance sheets should no longer be considered in isolation, and each institution should redouble its focus on the poorest countries with the least access to other sources of finance. For IDA, that is LICs and vulnerable LMICs; for IBRD, that is LMICs with demonstrated market access, including blend countries.

The ideal course would be for IDA and IBRD to run a range of scenarios designed to maximize concessional borrowing envelopes and direct funding to the neediest borrowers, without putting their AAA ratings at risk. To enable robust debate and a positive outcome, shareholders should request these for the Mid-Term Review of IDA21, only a few months away.