CAF 2.0: The Next Challenges for MDB Capital Adequacy

The independent review of the multilateral development banks’ capital adequacy frameworks commissioned by the G20, better known as the CAF report, has—with strong support from shareholders—driven remarkable progress in advancing more efficient use of multilateral development bank (MDB) capital. The G20 tasked the expert panel that conducted the review (of which I was a member) with answering a basic question: Can the existing capital of the MDBs support larger volumes of lending for development without jeopardizing strong MDB credit ratings (which protect shareholder capital and translate to low costs for developing country borrowers)?

After careful analysis, our answer was yes—current capital could safely support hundreds of billions of additional MDB lending deployed over time. That has, of course, been good news for both the shareholders who are the main sources of MDB capital and the shareholders who borrow from these institutions.

As they have been given access to more data on MDB credit performance, credit rating agencies have strongly concurred. They estimate massive volumes of additional lending headroom if key CAF report recommendations are implemented—especially the appropriate valuation of callable capital and more accurate risk weights grounded in actual MDB credit track records. Most recently, S&P put the potential additional lending capacity of MDBs as a group at $600-800 billion. Of course, actual deployment of additional lending capacity will depend on the degree to which MDB management and shareholders take up these recommendations and other parts of the MDB reform agenda. Most institutions still need to set and implement leverage targets that value callable capital appropriately and that reflect shareholder-defined risk tolerance.[i]

Just as importantly, how—and how much—headroom is used will also be shaped by the wider development finance landscape: shifting demand for MDB finance, depressed cross-border private finance flows, widespread debt sustainability problems, ongoing exogenous shocks, and sharp cuts in aid flows. The MDB response to these factors will have important implications for MDB finance allocation, instruments, and, therefore, capital requirements.

If MDBs are to use their scarce public funding to meet the most urgent development challenges, the inescapable reality is that they will have to take on more risk in at least three ways: more finance for riskier countries, allocating more finance to private clients (vs. governments), and deploying riskier instruments. These changes have direct implications for capital holdings—any bank’s central tool (along with liquidity holdings) for managing risk.

Below, I outline the issues and pose key questions that help define the next challenges for MDB capital adequacy.

Sovereign vs. non-sovereign MDB finance

As global growth slows, private foreign investment in EMDEs slumps, and domestic private investment rates also trend downward, EMDEs increasingly look to MDBs to help them finance private-sector-led growth. Both the World Bank and the African Development Bank have made job creation a central priority—jobs which must be driven by private sector development. Poorer countries especially will continue to need MDB public finance that expands their fiscal space for infrastructure and social investments. And promoting private investment requires effective use of MDB loans to governments to support investment climate reforms.

But overall, it is likely that MDB commitments and portfolios will evolve toward larger shares of finance to the private sector. The AIIB, for example, has announced a private finance target of 50 percent of annual commitments by 2030. While finance to the private sector is often riskier than financing governments, it can also be more profitable. But we are not in a position to compare relative risk-adjusted returns for MDBs at a disaggregated level. MDBs have been making more disaggregated data available on sovereign and non-sovereign credit defaults and recoveries from the GEMs database, but shareholders and other stakeholders do not have access to similar data on financial returns.

• As finance to the private sector occupies a growing share of MDB portfolios, what are the implications for MDB capital adequacy, leverage ratios, and financial performance transparency?
• What credit rating targets are best suited to institutions or parts of institutions that focus mostly on finance to the private sector?

MDB finance instruments and risk

Disappointment with MDBs’ private finance mobilization performance, and persistent capital market gaps, point to the need for a shift in MDB finance products, from a concentration in senior lending in hard currency to more risky instruments, including early-stage finance, first loss coverage, guarantees, equity, subordinated debt, local currency finance and hedging products. Riskier instruments require more capital and reduce leverage.

• How should the shift toward riskier instruments be managed in MDB capital adequacy frameworks? Is risk managed more efficiently by creating more purpose-built off-sheet balance vehicles or by more balance sheet consolidation to diversify risk?
• What additional data do stakeholders need on the financial performance of different instruments in order to make the right decisions on capital needs?
• How can MDBs accelerate progress toward systematically collaborating in ways that better share these risks across institutions and economize on the need for additional capital?

MDB capital sources

Remarkably, MDB shareholders have no standardized criteria for decisions to pursue general capital increases or to judge the right size for increases. Nor do all institutions have regular capital reviews. Rather shareholder decisions are generally the result of global or regional developments that build political support for increases: stretched MDB balance sheets after the 2008-2009 global financial crisis, regional conflict affecting major borrowers (e.g., Russia’s war on Ukraine), expanded countries of operations for regional institutions (e.g., the EBRD), or when an institution—with shareholder support--adopts a new strategy that requires more capital (e.g. IDB Invest).

Decisions are heavily influenced by domestic politics in the largest shareholding countries. While the logic of MDB capital increases to increase climate finance capacity (without sacrificing finance for development), for example, is recognized by many shareholders and experts, we have not seen capital increases in most institutions in recent years, as popular support for aid in rich countries shrinks. Finance shortfalls are particularly acute for countries that need concessional finance. As proposed in a recent piece by CGD colleague Clemence Landers, the World Bank in particular could source more IDA finance internally from its own balance sheet through net income transfers from the IBRD. But the traditional MDB approach for hard loan windows, including those that lend to the private sector, has been to use retained earnings to add to capital, thereby reducing the frequency of requests to shareholders for capital increases.

• Should the current ad hoc model for MDB hard window capital increases be replaced by regular capital reviews based on common capital adequacy benchmarks and rigorous needs assessments (as recommended in the CAF report), combined with more demanding reviews and targets for MDB performance and impact?
• How much should MDBs rely on internal income for adding to hard window capital vs. adding to grants and concessional finance?
• Should donors consolidate the large number of earmarked and fragmented trust funds in MDBs and redeploy some portion of the funds as capital so that it can be leveraged[ii]?
• Should private investors be permitted to put equity in MDBs and, if so, how should they participate in governance?
• How much of a role should hybrid capital play across different MDBs, and how can they better coordinate to build the asset class and reduce borrowing costs?

Exogenous shocks, temporary debt relief, and MDB liquidity

Most poor countries are in, or at risk of, debt distress, a problem that has persisted well past the pandemic, especially as growth trends make it harder for countries to grow out of their debt problems. We know that climate and conflict-related shocks and future pandemics are inevitable and that poor countries are particularly vulnerable, with limited countercyclical capacity.

In response, MDBs struggle to meet shareholders' expectations for large net positive MDB finance flows to poor countries. This situation has prompted calls for MDBs to participate in debt restructurings, especially as debt owed to MDBs rises as a share of poor country debt stocks. But that approach places major demands on shareholders to cover MDB losses and undercuts preferred creditor treatment. A more feasible approach is to include temporary debt service suspension clauses in MDB loans, which the IDB and World Bank have begun to do. Expanded use of such clauses covering a broader range of exogenous shocks would offer more meaningful temporary relief, but could affect MDB liquidity positions, especially if suspended payments are concentrated in particular regional banks.

• The CAF report raised but did not pursue an effort to promote central bank liquidity lines to MDBs. Should this issue be revisited in the current risk-filled environment?

Pooled MDB assets and risk transfers

A growing number of MDBs are taking up the CAF recommendation to pursue risk transfers and securitizations that offload performing assets to private investors and insurers.[iii] The benefit is not only freeing up their own capital. It is also a more efficient approach to crowding in private investors at the portfolio level and at scale, rather than transaction by transaction. And as better data on risk become available and these assets become familiar to markets, the need to include a concessional tranche in these transactions will likely diminish. But some MDBs continue to view such transactions as only relevant to capital-constrained institutions or go to the market with relatively small offerings, as the World Bank just did. This slows the development of the market, limits its size, and reduces the diversification benefits of assets pooled across institutions.

• How can MDB management and shareholders collectively accelerate development of the MDB asset class by standardizing assets and pooling across institutions?
• Should all MDBs incorporate the originate-to-share model as a core feature of their institutional strategies and capital adequacy frameworks?

These are not questions at the margin; they lie at the heart of the MDB finance model. Ignoring them at a time of increasingly scarce public development resources is no longer tenable. This is a time of rapid change in global banking models. Multilateral development banks and their shareholders should embrace, not seek to escape, the evolution imperative.

[i] See the document labeled 2025 CAF Progress Report under the documents of the G20 International Financial Architecture Working Group using this link: IFA WG.

[ii] The World Bank, for example, has 229 standalone trust funds and 67 large, multi-donor “umbrella” trust funds currently holding $18.9 billion in trust: in fiscal year 2024, trust funds received $7.2 billion in contributions from over 100 donors and disbursed $10.1 billion. For a discussion of the issues, see here.

[iii] Some argue that risks transfers should focus on non-sovereign assets. Concerns have been raised about risk transfers involving sovereign assets, which could be construed as conferring preferred credit treatment to private investors, though that risk is mitigated in cases where MDBs remain the lenders of record.