The Health Record · Spring Meetings 2026 · MDB Reform Platform
“Health Works” is the Bank’s flagship health initiative — 1.5 billion people reached with quality, affordable health services by 2030. The Bank’s promotional material cites Mozambique as a success story. IEG’s evaluation of the same portfolio: two projects, $150 million, zero Satisfactory. The gap between the marketing and the evaluation is the accountability problem this paper documents.
The World Bank reports health portfolio performance using the MS+ benchmark — Moderately Satisfactory or above. In the Africa HNP portfolio, that produces a headline rate of 84.6 percent. Using the S+ benchmark — the standard the Bank applies to IFC and MIGA — the rate by commitment is 29.2 percent.
The gap is 55 percentage points — the largest benchmark divergence of any sector in the Africa portfolio, exceeding education (41pp) and transport (44pp). The health portfolio is where the MS+ metric does the most damage to accountability. It is also where Health Works is built on that reported number rather than the honest one.
The World Bank committed $8.4 billion to 78 IEG-rated HNP projects in Sub-Saharan Africa between FY2015 and FY2026. $5.9 billion — 71 percent — went to projects that did not achieve Satisfactory outcomes. By project count: 34.6% S+. By commitment: 29.2% — 29 cents on the dollar. The headline MS+ rate is 84.6 percent, producing a 55-point gap between what is reported and the honest benchmark.
Fourteen countries have two or more evaluated health projects and zero Satisfactory outcomes: 55 projects, $3.0 billion committed. The instrument test is the most important structural finding: MTI failed via DPF, Health fails via IPF. Same countries, same outcome, different instruments. The problem is not the instrument. The Global Fund’s own OIG confirms: both institutions fail in Angola, DRC, and Congo-Brazzaville; the focused Global Fund model succeeds in Kenya while the Bank’s does not. The floor determines what is possible. The delivery model determines the ceiling above it.
Finding 1: The 55-Point Gap — and the Health Works Contradiction
Health Works commits the Bank to reaching 1.5 billion people with quality, affordable health services by 2030. It is backed by a $27 billion global portfolio, 160 projects, 15 National Health Compacts launched in Tokyo in December 2025, $2 billion in co-financing with the Global Fund and Gavi, and $410 million in philanthropic mobilisation. The Bank reports 375 million people already reached.
29%
Honest benchmark
55pp gap
15%
Finding 2: Tanzania — $1.3 Billion, Zero Satisfactory
Tanzania is the single largest health commitment in Sub-Saharan Africa — $1.3 billion — and has zero Satisfactory outcomes. The Basic Health Services Project ($1.062 billion, rated Moderately Unsatisfactory) is the largest single health project failure in Africa. The successor project improved to Moderately Satisfactory but did not reach the Satisfactory threshold. Tanzania will repay both loans in full.
Tanzania Basic Health Services (P099974)
Committed: $1.062bn | Rating: MU
Largest single health project failure in Africa. Facilities built; quality of care not delivered.
Tanzania Health System Strengthening
Committed: ~$250M | Rating: MS
Successor project improved from MU. Did not reach S+. Combined: $1.3bn, 0% Satisfactory.
Finding 3: The Instrument Test — IPF Fails Too
The MTI Zero Club (Part 9) documented 14 countries where Development Policy Financing never achieved Satisfactory health or governance outcomes. The rebuttal: DPF rewards legal compliance over functional change. Investment lending is different — IPFs have procurement, results frameworks, supervision missions, and implementation units.
The Health Zero Club tests this directly. In 14 African countries, 55 health IPFs over up to 36 years produced zero Satisfactory outcomes.
What the IEG 2018 evaluation already confirmed
Access improved in 70% of projects; quality improved in only 46%. Health systems strengthening objectives fell from 42% of closed projects to 27% of open ones — the Bank was retreating from the hard institutional work. Health was the second-most-vulnerable sector to corruption: 191 Integrity VP cases. The 2009 IEG evaluation documented this pattern. The 2018 evaluation confirmed it. The design model has not changed.
Finding 4: Kenya — 36 Years, $694 Million, Never Once Satisfactory
Kenya is not a fragile state. It is lower-middle income, has functioning institutions, a large private health sector, and a Ministry of Health coordinating Global Fund, Gavi, PEPFAR, and bilateral programmes simultaneously. Eight IEG-evaluated health projects. $694 million. 36 years. Not one rated Satisfactory.
| P-Code | Project | Year | Rating | Committed |
|---|---|---|---|---|
| P001312 | Population IV | FY1990 | U | $36M |
| P001339 | Health Rehabilitation | FY1992 | MU | $31M |
| P001333 | Sexually Transmitted Infections | FY1995 | MS | $52M |
| P066486 | Decentralised Reproductive Health | FY2001 | U | $98M |
| P070920 | HIV/AIDS Disaster Response | FY2001 | MU | $50M |
| P081712 | Total War Against HIV/AIDS (TOWA) | FY2007 | MS | $135M |
| P074091 | Health Sector Support | FY2010 | MS | $100M |
| P152394 | Transforming Health Systems | FY2016 | MS | $191M |
| Total: 8 projects, 36 years | 0% S+ | $694M | ||
Finding 5: The Zero Club — 14 Countries, Zero Satisfactory
| Country | Projects | Committed | Context |
|---|---|---|---|
| DRC | 7 | $982M | Fragile. In five Zero Clubs simultaneously (Health, MTI, Transport, Energy, Education). |
| Kenya | 8 | $694M | Non-fragile, stable. 36 years. The instrument-agnostic finding confirmed here. |
| Burkina Faso | 4 | $270M | Lower-middle income. PBF collapse the primary failure mode. |
| Cameroon | 5 | $253M | Non-fragile. Also in Education Zero Club. PBF could not outperform unconditional transfers. |
| Angola | 3 | $131M | Lower-middle income. Global Fund also struggled. The floor has not been reached. |
| Congo-Brazzaville | 3 | $89M | Lower-middle income. Global Fund OIG: limited progress in 10 years. |
| CAR | 2 | $81M | Fragile. External implementation the minimum viable delivery model. |
| Sierra Leone | 3 | $36M | Post-conflict. Three IPFs evaluated. |
| Lesotho | 2 | $21M | Devolution without institutional readiness the primary failure mode. |
| Eritrea | 2 | $32M | Restricted access. Limited supervision possible. |
| Rwanda | 2 | $27M | Exited Zero Club on subsequent projects. Exit is possible. |
| Comoros | 2 | $21M | Small island. HCI declined (-0.004). |
| Guinea-Bissau | 2 | $16M | Fragile. Post-conflict. |
| São Tomé | 1 | $11M | Small island. Institutional capacity the binding constraint. |
| Total | 55 | $3.0bn | Zero Satisfactory. Every dollar will be repaid. |
Finding 6: The HCI Test — The Bank’s Own Metric
The Bank’s Human Capital Index — combining child survival, adult survival, stunting, and education — provides the independent test. If $3 billion in health lending produced transformational results, they should appear here.
| Country | HCI baseline | HCI 2020 | Change | WB Health Commitment |
|---|---|---|---|---|
| DRC | 0.369 (2017) | 0.366 | −0.003 | $982M |
| Angola | 0.361 (2017) | 0.362 | −0.001 | $131M |
| Comoros | 0.409 (2017) | 0.405 | −0.004 | $21M |
| Cameroon | 0.394 (2017) | 0.397 | +0.018 | $253M |
| Burkina Faso | 0.320 (2010) | 0.384 | +0.064 | $270M |
| SSA average (16 countries) | — | — | +0.041 | — |
Cameroon gained only +0.018 on $253M — less than half the SSA average. DRC, Angola, and Comoros declined. Burkina Faso gained +0.064 on $270M but tracking the SSA average, not outperforming it. The Bank can get people to clinics. It cannot ensure adequate care inside.
Finding 7: Six Failure Modes
| Failure Mode | Countries | IEG language (own evaluation) |
|---|---|---|
| Overly ambitious design | Cameroon, DRC, Angola | “Overly complex and risky projects are unwise to embark upon, even in the presence of political pressure.” (Cameroon, FY1995, HU) |
| PBF scaled before evidence | Cameroon, BF, Congo-B, CAR | “Scaling up too quickly can limit the ability to adopt previous learning.” (Cameroon, FY2016, U) |
| M&E without measurement | Kenya, Cameroon, BF | “Data availability is critical to ensure adequate monitoring.” (Kenya, FY2016, MS) |
| Devolution without capacity | Kenya, Lesotho | “Staffing in the PMT was not sufficient to support and supervise 47 counties.” (Kenya, FY2016, MS) |
| Post-conflict state absence | DRC, CAR, Sierra Leone | “The absence of a TTL on the ground can critically affect a project’s ability.” (DRC, FY2004, U) |
| Staff turnover + amnesia | Cameroon, DRC, Kenya | “Frequent team changes, resulting in significant loss of expertise and continuity.” (Cameroon, FY2016, U) |
Finding 8: Two ICRR Deep Dives
Approved at $220M in 2014, grew through five additional financings to $715M. Real gains: ANC visits 36%→65%; immunisation 62%→71%; contraceptive prevalence 6%→27%.
The M&E baseline was the 1984 census — 30 years before the project started. Targets were lowered in 2020; the project rated Substantial against reduced targets but only Modest against originals. Fraud: “High levels of corruption in the province of Equator” — province eventually removed. $906,345 in ineligible expenses. Efficiency: Modest. PBF administrative costs: 26% of PBF spending.
Fraud, a province removed for corruption, a 1984 census baseline, dropped components, lowered targets — absorbed into a single word: Moderately Satisfactory. This is the MS equilibrium at $715 million.
Previous project: PBF in 44 districts. This project: 194 districts simultaneously, plus refugee services, emergency component, GFF investment case — all at once.
The project’s own impact evaluation (De Walque et al., 2021): “outcome differences between the PBF treatment group and the increased financing group were not statistically significant.” The Bank’s preferred health financing instrument could not demonstrate it outperformed unconditional transfers in the same context.
$30M (nearly 20% of expenditure) spent verifying PBF payments. Refugee component: $30M approved, “no data reported and no activities were carried out” — full $30M cancelled. No PDO indicators reported November 2019–June 2022: nearly three years of a $163M project with zero results data. Cameroon will repay in full.
Finding 9: The Global Fund Cross-Institutional Test
Finding 10: The Moderately Satisfactory Equilibrium
The modal rating across 55 Zero Club projects is not Unsatisfactory. It is Moderately Satisfactory.
MS is not failure. It is not success. It is the equilibrium: enough delivery to avoid an Unsatisfactory rating, enough disbursement to sustain the pipeline, insufficient achievement for transformation — but institutionally acceptable. A project rated MS triggers no lending pause, no portfolio review, no mandatory design revision. The next project is prepared. Commitment grows. The lessons are identified, documented, and repeated. The design model does not change.
Where Health Works — and Why
| Country | S+ Rate | Committed | What made the difference |
|---|---|---|---|
| South Sudan | 67.9% | $248M | External implementation via UN/NGO partners. Bounded, measurable objectives. Not dependent on Ministry systems for delivery. |
| Guinea | 62.6% | $151M | Disease-specific focus: malaria, Ebola. Bounded operational endpoints with clear success criteria. |
| Mali | 61.0% | $245M | Narrow nutrition objectives with measurable anthropometric targets. Accountability for specific, quantifiable outcomes. |
| Ethiopia | 50.1% | $320M | Existing Health Extension Programme as the platform. Bank built on a functioning system rather than creating one from scratch. |
| Rwanda | Escaped ZC | — | Strong state discipline. Imihigo performance contracts. Community accountability. Integrated PBF within a functioning state — not as a substitute for one. |
✔ What the successes share
Focused, specific, measurable health objectives. External or strong national platforms as the delivery vehicle. Bounded scope matched to institutional capacity. Disease endpoints rather than systems reform as the primary goal. Functioning accountability mechanisms.
✘ What the Zero Club countries share
Broad health systems strengthening requiring institutional transformation as a precondition. PBF at scale without demonstrated context-specific effectiveness. MS+ absorbs partial achievement. Lessons repeated without design change. Floor below which no model reliably works.
$8.4 billion committed to Health in Sub-Saharan Africa. 84.6% reported successful. 29.2% by the honest benchmark. 55 points between those two numbers — the largest benchmark gap of any sector in Africa. Health Works is built on the first number. IEG’s evaluation record documents the second. The Mozambique success story cites doubled community health workers. IEG rates the same portfolio: zero Satisfactory.
The instrument test removes the most common defence: IPF and DPF fail in the same 14 countries. The Global Fund’s own OIG confirms both institutions struggle where the institutional floor has not been reached — Angola, DRC, Congo-Brazzaville — while focused models succeed where it has (Kenya). The Moderately Satisfactory equilibrium — 53% of Zero Club ratings — is the mechanism that makes $715 million in DRC, $163 million in Cameroon, and $694 million in Kenya all institutionally acceptable without triggering a portfolio review. Health Works promises to be different. The 15 National Health Compacts are the test. Each one should carry an annex: what does the IEG record show for this country, what failure modes were documented, and what specifically will be different this time.