AAA Credit. Unrated Development Effectiveness. · Zero Club Series · MDB Reform Platform
The World Bank’s credit ratings tell investors that its bonds are safe. They tell borrowers, taxpayers, and donor governments almost nothing about whether its lending achieves development results. After two Disbursement Disconnect papers documenting $361 billion deployed to non-satisfactory operations, the gap between those two statements deserves examination.
A credit rating answers one question: will the institution repay its bondholders on time and in full? For the World Bank, the answer is an unqualified yes. The rating rests on capital adequacy, liquidity, preferred creditor status, and shareholder support. On every one of those dimensions the assessment is correct, and this paper does not dispute it.
The mechanism that secures the rating is the sovereign guarantee. Every IBRD loan and IDA credit is a legal obligation of the borrowing government, which is bound to repay regardless of whether the project the money financed achieved anything. The loan performs as a financial instrument whether or not the project performs as a development intervention.
The critical implication: the same guarantee that justifies the AAA also severs the link between development performance and financial consequence. A project can fail completely and the loan is still repaid, the capital ratio is still strong, the rating is still triple-A. The paper names that silence, measures it, and proposes it be disclosed.
The World Bank holds triple-A credit ratings from all three major agencies, correctly assigned on the dimensions they measure. This paper does not argue that the AAA is wrong. It argues that the AAA is incomplete — and that the incompleteness matters more now than at any point in the institution’s history. S&P’s October 2025 criteria revision has materially expanded WBG lending headroom: IBRD’s risk-adjusted capital ratio rose from 24.1% to 40.1% in a single cycle; IDA’s rose from 58% to 90%. S&P estimates the revision could potentially unlock $600 to $800 billion in additional sovereign lending capacity across the MLI sector as a whole. The case for that expansion rests on capital ratios and shareholder support. It says nothing about whether the institutions deploying that capital are producing development outcomes.
S&P’s own shareholder support analysis states that World Bank loans are “highly scrutinised for impact” and that “impact is measured and transparently reported.” S&P does not verify this. The Disbursement Disconnect evidence shows it is not accurate: ERR coverage is zero, QAG was disbanded in 2014, and $257 billion in IBRD went to non-satisfactory operations. The paper proposes not a downgrade but a disclosure.
The paradox, the sovereign guarantee, the rating-dimensions table, the Moody’s word count, the Caa1-to-Aaa mechanics, the management-accounts finding, the S&P contradiction, the criteria revision, and the full 11-point proposal to the agencies, management, and the IDA21 MTR.
The Paradox: AAA Credit, Persistent Delivery Failure
For most institutions, financial failure and operational failure are connected. When a commercial bank’s loans fail to perform, provisions rise, capital ratios erode, the credit rating deteriorates. Bad loans cost money. The World Bank is structured differently. Its loans do not fail financially — they fail developmentally.
A financial failure is when a borrower cannot service its debt. By this standard IBRD has had 27 non-accrual events in 80 years — a lifetime default rate of 0.7 percent. A development failure is when a project does not achieve its stated objectives. By this standard $361 billion across both windows went to operations that did not achieve a satisfactory outcome. These are two entirely different metrics. Rating agencies measure the first. Nobody in the rating framework measures the second.
The Sovereign Guarantee Does Two Things at Once
Every IBRD loan and IDA credit carries a sovereign guarantee — a legal obligation to repay regardless of the development outcome. Without it the AAA would be unjustifiable. But the guarantee insulates every party in the disbursement chain from the development outcome. No one loses money when a project fails developmentally.
| Project | Country | IEG outcome | Disbursed | Loan repaid? |
|---|---|---|---|---|
| Eskom P116410 | South Africa | Moderately Unsatisfactory | ~$3.75bn | Yes |
| NH28 P077856 | India | Unsatisfactory | $615.7M (99.3%) | Yes |
| Vietnam PIR 1 P117723 | Vietnam | Unsatisfactory | $500M (100%) | Yes |
| Brazil Env DPL P095205 | Brazil | Unsatisfactory | $1.3bn (100%) | Yes |
The Rating Agencies Are Evaluating the Wrong Risk
Moody’s and S&P reward capital adequacy, liquidity, shareholder support, preferred creditor status, and callable capital. They do not evaluate whether projects succeed, whether development outcomes are achieved, or whether lending creates development value. Yet development effectiveness is the institution’s stated purpose.
| Dimension | Moody’s / S&P | MDB Reform evidence |
|---|---|---|
| Repayment risk | AAA | AAA — confirmed: 0.7% lifetime default rate |
| Capital adequacy | AAA | Confirmed: IBRD 40% / IDA 90% RAC after Oct 2025 revision |
| Liquidity | AAA | Confirmed: IBRD liquid assets 119% of FY26 target |
| Development risk | Not rated | Weak — 46.4% (IBRD) / 65.8% (IDA Africa) non-S+ |
| Appraisal quality | Not rated | Weak — ERR coverage: 0% (IDA Africa since 2015; IBRD since 2020) |
| Outcome record | Not rated | Weak — QAE below-standard: 91.6% failure rate |
| Economic analysis | Not rated | Collapsing — 70% to zero over five decades |
| Value for money | Not rated | Unknown — no ERR, no systematic cost-effectiveness |
| Development effectiveness | Not rated | $361bn deployed to non-satisfactory operations across both windows |
Inside the Moody’s Reports: What They Count — and What They Don’t
The two Moody’s credit opinions — IBRD (February 19, 2026) and IDA (February 21, 2025) — run to roughly 25 pages of analysis. Read systematically against the Disbursement Disconnect evidence, they produce a precise accounting of the gap.
What Moody’s assesses — exhaustively
Capital adequacy: IBRD leverage 3.94x, usable equity $73.6 billion, E/L ratio 21.6% against 18% floor; IDA leverage 1.1x, usable equity $204.2 billion. Asset performance as repayment: IBRD NPA ratio 0.5%, 27 non-accrual events in 80 years; IDA NPA ratio 0.4%. Preferred creditor treatment: IBRD lifted from average borrower B1; IDA lifted from weighted average Caa1, the lift mechanical in both cases. Shareholder support: the primary downgrade trigger in both reports.
What Moody’s does not assess — at all
Zero mention across both opinions of IEG outcome ratings, the share of disbursements going to non-satisfactory operations, ERR coverage or its collapse, Quality at Entry ratings, the disbandment of QAG in 2014, or project success rates. The phrase “development effectiveness” does not appear in either report. “Development” appears exclusively as an institutional descriptor — never as a performance dimension.
The gap confirmed in management’s own accounts
The same separation appears in the Bank’s own financial reporting. Both the IBRD and IDA Management’s Discussion & Analysis for the nine months ended March 31, 2026 — prepared under US GAAP and subject to independent auditor review — formally name “development outcome risk” as a risk category, in identical language: “The risk of [IBRD’s/IDA’s] operations not meeting their development outcomes (development outcome risk) in [IBRD’s/IDA’s] lending activities is monitored at the corporate level by Operations Policy and Country Services (OPCS).”
How Caa1 becomes Aaa: the five-step lift
| Step | What it adds | Moody’s assessment |
|---|---|---|
| 1. Average sovereign borrower | Caa1 baseline | B2 to Caa1 WABR across borrowing members |
| 2. Preferred creditor treatment | +4 to +5 notches | PCT lifts DACQ to ‘baa’ |
| 3. Portfolio diversification | Additional support | 78 countries, low crisis correlation |
| 4. Capital adequacy | Low leverage buffer | Leverage 1.1x, usable equity $204.2bn |
| 5. Callable capital + shareholder support | Final lift | Aa3 weighted-average shareholder rating |
| IDA institution | Aaa | Scorecard range: Aaa–Aa2, lifted to Aaa |
IDA’s return on equity is negative (-1.9% in fiscal 2024) — the institution loses money — and the rating is unaffected because equity erosion is gradual and triennial donor replenishments offset what is consumed. Development effectiveness contributes zero to the calculation. A portfolio producing $104 billion in non-satisfactory disbursements is, from Moody’s perspective, indistinguishable from one producing none — because the relevant metric is NPA ratio (0.4%), not development outcome.
Development Effectiveness Enters Through the Back Door
S&P’s 2025 Supranationals edition contains a dedicated article on US shareholder support. In explaining why the US is expected to remain supportive, S&P asserts that MLI loans are highly scrutinised for impact, that there are very few inefficiencies in project outcomes involving taxpayer funds, and that impact is measured and transparently reported. S&P uses these claims as inputs to its shareholder-support assessment — the highest-sensitivity variable in the rating — without verifying them.
“Highly scrutinised for impact”
Of IBRD projects approved since 2020, not one carries a calculated ERR. The Quality Assurance Group was disbanded in 2014. IEG’s own 2010 evaluation found 80% of task team leaders said cost-benefit analysis was sufficient to “tick a box.”
“Very few inefficiencies in outcomes”
$257 billion in the IBRD portfolio went to non-satisfactory operations. In IDA Africa, 65.8% of disbursements went to non-satisfactory operations. Neither figure is “very few inefficiencies.”
“Impact is measured and reported”
IEG evaluates outcomes ex-post, often years after disbursement. The ex-ante instrument — the ERR — has reached zero coverage. What is measured is outcomes after the money has moved; the prospective analysis of whether it should move has disappeared.
The analytical move
An unverified development-effectiveness premise sits at the heart of the rating’s most sensitive variable. The agencies accept management’s narrative as an input; the evidence contradicts it. Verifying the premise requires no methodology change.
The S&P Criteria Revision: Capital Endorsed, Development Unexamined
S&P’s October 2025 methodology update significantly increased risk-adjusted capital ratios across the MLI sector. For the World Bank specifically, IBRD’s RAC ratio rose from 24.1% to 40.1% and IDA’s rose from 58% to 90% — both in a single revision cycle. S&P estimates the criteria change could potentially unlock $600 to $800 billion in additional sovereign lending capacity across the MLI sector as a whole, with many institutions able to expand sovereign exposure by 35 to 70 percent, subject to other constraining factors. The capital case is real and correctly made.
What the criteria revision does not examine is whether the institutions deploying that expanded capacity are producing development outcomes. For the World Bank, the Disbursement Disconnect evidence answers that question: $361 billion has gone to non-satisfactory operations; ERR coverage in both windows has reached zero; and the Bank’s own management accounts classify development outcome risk as a category delegated to an operational policy unit with no financial metric attached. S&P assessed the capital grounds for expansion exhaustively. The development grounds were not assessed at all.
Conclusion
IBRD and IDA deserve their AAA ratings on the dimensions those ratings measure. Bondholders are safe. Capital ratios are strong. Preferred creditor status has held for 80 years.
The gap can be named, measured, and disclosed. The rating agencies have the platform to name it. The IDA21 Mid-Term Review is the next formal accountability gate. Ajay Banga, presiding over an IBRD whose risk-adjusted capital ratio has just doubled and an IDA whose has risen by a third, has both the balance sheet and the moment to close it. The proposal is disclosure, not downgrade.
What We Are Proposing
The three papers in this series — the IDA Africa Disbursement Disconnect (Part 15), the IBRD Disbursement Disconnect, and this paper — converge on the same institutional problem. The proposals below are aligned across all three. They are proposals for disclosure and accountability — not for downgrade and not for reduced lending.
Series Context
IBRD Disbursement Disconnect
4,535 projects, $554bn disbursed, 46.4% to non-satisfactory operations — $257bn. The IBRD evidence removes the fragility explanation: the same patterns appear in middle-income borrowers with stronger institutions and access to alternative financing.
IDA Africa Disbursement Disconnect (Part 15)
2,576 projects, $158bn disbursed, 65.8% to non-satisfactory operations. The disbursement system does not respond when supervision flags failure. $104bn deployed to non-S+ operations.
Quality at Entry in IDA (Part 14)
When design quality and M&E are both weak — 52% of the portfolio — 5.4% of projects succeed. The 65-percentage-point QAE gradient is virtually identical across both IBRD and IDA windows.
FCV Strategy Assessment
“Sound Strategy, Broken Platform.” The strategy is coherent; the delivery platform that must execute it is the same one producing the Zero Club outcomes.
Complete analysis: the paradox, sovereign guarantee, rating-dimensions table, Moody’s word count, Caa1-to-Aaa five-step lift, management-accounts finding, S&P contradiction, criteria revision, 91% problem, and full 11-point proposals. Full methodological note with all source citations.
The World Bank’s AAA is correctly assigned on the dimensions the rating agencies measure: capital, liquidity, preferred creditor status, shareholder support. What the rating does not measure is whether the lending achieves development results. The sovereign guarantee makes the AAA possible and, at the same time, makes development failure financially invisible. The loan is repaid whether or not the project works.
The evidence that the gap is real is not contested and not external: $361 billion deployed to non-satisfactory operations in the Bank’s own evaluation record; zero occurrences of “development effectiveness” across both Moody’s opinions; an unverified impact claim embedded in S&P’s most sensitive rating variable; development outcome risk named and left unmeasured in the Bank’s own audited quarterly accounts; and a 91% performance figure reported to IDA Deputies that is three times the honest S+ number. The proposal is disclosure, not downgrade — name the gap, measure it, and report it alongside each annual credit review.