MDB Results 2026 · A Series
Kofi and the Six Ships
A seven-year-old boy visits Washington with his father. He asks the genies what they do. They answer from the reports.
Six multilateral development banks. Nine pieces (Day 0 through Day 7, with the IMF carrying five parts). Six genies made of numbers. Kofi is seven years old and takes everything literally — which turns out to be exactly the right qualification for reading evaluation reports. Each piece has its own Kofi story and its own analytical note with the underlying data. The series was timed to the Spring Meetings of the MDB shareholders, April 21, 2026. Start at Day 0.
The Series — MDB Results 2026
0
Scene Setter · Washington DC · Published
The Boy Who Thought the Bank Was a Cruise Ship
Lunch in the atrium. A boy with a notebook. A genie who was asleep.
Kofi and his father arrive in Washington. Kofi sees the World Bank building and assumes it is a cruise ship. He is not entirely wrong. He has lunch in the atrium, meets another child, and finds a genie who has been waiting a very long time for someone to ask the right questions.
World Bank · Published
The Cruise Ship
$577 billion lent. The emperor is a very busy man.
53% of what the ship carries arrives at its destination in working order. The Board spends $84 million a year approving loans it cannot then independently evaluate because it approved them. The S+ rate. The 50-year decline table. The 80/20 finding. Four structural layers. Three reforms.
IMF · Part 1
The Ship That Does Not Keep Score
Zero project ratings in 149 countries.
The IMF’s IEO produces thematic evaluations — valuable, but not a systematic rating of whether the Fund’s work delivers results. The five MDB evaluation offices compared. Why the absence of project ratings is not a technicality. You cannot fix what you do not measure.
IMF · Nigeria · Part 2
Nigeria and the Africa-Wide Emergency Funding
$3.4 billion. Gone in months. Not rated.
The largest single-country RFI in Sub-Saharan Africa. Disbursed in weeks with no performance conditions. The IEO evaluation of emergency lending in Sub-Saharan Africa named nineteen countries. Nigeria — the largest recipient — was not among them. The Idris arrest. What was omitted and why it matters.
IMF & World Bank · Part 3
The Cold War Across 19th Street
$750M–$1.1bn per year. Same Finance Ministers. Same week. 35 years of concordats.
Two institutions, one street, contradictory advice to the same Finance Ministers. Five domains where both institutions do the same work. Five countries where contradictory advice cost real money — Nigeria, Bangladesh, Ethiopia, Ghana, Pakistan. The 35-year timeline of failed coordination. Six actions that would actually change the architecture.
IMF · AFRITAC · Part 4
The Homework It Does Do — And Why the Score Has Not Moved
$200M in technical assistance. Five regional centres. Not rated.
AFRITAC teaches treasury reform, tax administration, debt management. Allen Schick identified seven basics a finance ministry must do before any of that matters: payroll, procurement, internal control, cash management, basic reporting. The SSA PEFA data shows the basics are still not done. The score has not moved because the sequence is wrong.
IMF · Governance · Part 5
Europe’s Chair
The Managing Director has always been European. The President has always been American. Both are customs, not rules.
Since 1944. The IMF Managing Director: always European. The World Bank President: always American. A rule you can change at a meeting. A custom you can only change by deciding to be embarrassed by it. What changed at the WTO in 2021. What the Spring Meetings could look like if the shareholders chose to act on it.
IFC · Published
The Palace Casino
Croupiers in Washington. Tables in Kinshasa.
The private sector arm of the World Bank group. 11% satisfactory in fragile states. 62% to 17% — the decline in IFC success rates in FCS states over fifteen years. The PSW: $2.5 billion to go where the private sector would not, used to replace the chips the IFC would have played with anyway. That is a refund, not additionality.
African Development Bank · Published
The Dhow
94% satisfactory. Every year. Through COVID. Through conflict. Through everything.
The task manager who wrote the rating was wrong, and kept it. The smile that does not change. Medupi. The adjusted plausible range: 61–75%. IDEV validates a sample and produces no independent rating. The gap is unpublished. Five named cases where the documented number and the official number tell different stories.
Asian Development Bank · Published
The Atakebune
Nine presidents. Fifty-nine years. One nationality.
The ship that holds ground. Konnichiwa, Kofi-san. Sayonara, Kofi-san. A rule you can change at a meeting. A custom you can only change by deciding to be embarrassed by it. The best evaluation architecture in development banking — IED validates everything, signs its name to the disagreement, publishes the gap. And a 12-point management-IED divergence that is not closing.
Inter-American Development Bank · Published
The Carnival Ship
Management: 81% satisfactory. OVE: 53%. Same projects. Same year. A decade. Neither reconciled.
Kofi hears the music before he sees the building. The marble is very expensive. The twenty-six borrowing countries of Latin America built this institution and run it. On Friday evenings the whole continent calls and the difficult conversation moves to Monday. They even evaluated the evaluation — sixteen years after launch, and found the cultural change did not occur. Poco a poco. Todavía no.
7
The Full Picture · Published
Six Ships. One Ocean.
Kofi closes his notebook. His mother has something to say.
No genie today. Kofi reads back through six days and tells his father five things he would change. The evaluation ranking — all five offices placed. Are they going to get better? Slowly, then faster. And then Chinwe — his mother, from Benin City — says: I am going home. And I want to take you with me. Next stop: Nigeria.
The evaluation ranking from the series: ADB IED first (validates all, signs the gap), IDB OVE second (validates all, too detached), World Bank IEG third (Board co-authors loans it evaluates), AfDB IDEV fourth (sample only, no independent rating), IMF IEO last (zero project ratings in 149 countries).
All underlying datasets are on the Data page.