Day 1: Kofi @ the World Bank

MDB Results 2026: An Independent Assessment  |  Day 1 of 6

The genie did not look like a genie.

Kofi had expected something large and blue, possibly with a turban. What emerged from the printing press in the basement of the World Bank — slowly, as if it had been asleep for a very long time and was not entirely sure it wanted to wake up — was more like a very old man made of paper. Pages and pages of paper, densely printed, folded into something that approximated a human shape. The pages rustled when he moved. Numbers flickered across his surface like light on water.

“You asked what the results are,” said the genie.

“Yes,” said Kofi.

“I will tell you. But first you need to understand how they keep score.”

Kofi sat down on the marble floor of the atrium. The passengers moved around them both, barely adjusting course.

“When the ship lends money to a poor country,” said the genie, “someone eventually writes a report about whether it worked. The report gets a rating. The best rating is called Highly Satisfactory. Then there is Satisfactory. Then Moderately Satisfactory. Then Moderately Unsatisfactory. Then Unsatisfactory. Then Highly Unsatisfactory.”

“Which one is good?” Kofi asked.

“Satisfactory or above. That is the pass mark. That is what they call S+. If a project gets Satisfactory or above, it worked. If it gets anything below that, it did not fully achieve its development objectives. For its monitoring — to be fair — the Bank uses the rating for MS+ which is the mid point of the rating scale.”

“And how many worked?”

The genie was quiet for a moment. The numbers on his surface rearranged themselves.

“Since 1973, this ship has lent money for 13,273 projects. The independent evaluators have rated all of them. And of those 13,273 projects —”

He paused.

“47.6 percent were Satisfactory or above. The MS+ rate was much higher.”

Kofi did the maths the way his father had taught him.

“So more than half did not fully achieve their objectives?”

“More than half — 52.4 percent. And the money committed to those projects was $567 billion. Out of $1,013 billion in total.”

Kofi looked at the statue. The African child. The blind man with the stick.

“$567 billion,” Kofi said.

“$567 billion,” said the genie.

“But it must have been better before. When it was new.”

“It was. In the early 1970s, 81.8 percent of projects were Satisfactory. Eight in every ten.”

“What happened?”

“The ship got bigger. And busier. And more ambitious. And the results — well. Look.”

“In 2005 to 2009, only 26.4 percent were S+. One in four.”

“Did anyone notice?” Kofi asked.

“Oh yes. They noticed in 1992. A man called Wappenhans wrote a report. He said the ship had an approval culture — it cared more about getting projects approved than about whether they would work.”

“And did it change?”

The genie looked at the table.

“The rate continued falling for fifteen years after Wappenhans wrote his report. It has never recovered to what it was before 1985.”

“What is the point of writing reports if nobody acts on them?” Kofi asked.

The genie smiled the smile of someone who had been asking the same question for eighty-two years. “That,” he said, “is precisely what we are here to discuss.”

“Tell me about the money. The $567 billion. What was it spent on?”

“Three kinds of things. Investment Project Financing — specific projects, like roads or schools. Development Policy Financing — money given to a government that promises to make reforms. And Programme-for-Results — the newest kind, where the money only flows after someone independently verifies that results have been achieved.”

“Which works best?”

“The one that works best is the smallest one.”

“Yes. Because it is harder. It requires independent verification. The money does not flow until something has actually been achieved. The other two — you can lend the money, write the report, and by the time anyone evaluates whether it worked, everyone involved has moved to a different project.”

“That seems like a problem,” said Kofi.

“It is a problem,” said the genie.

“What about Africa?” said Kofi. His mother was from Ghana.

“Africa is where the poverty is the highest and the challenges are the most severe — not just for the Bank but for all development partners.”

“$155 billion. In Africa. Projects that did not fully achieve their development objectives.”

“Yes.”

“Is that because Africa is harder?” Kofi asked.

“It is definitely harder. Research across the full portfolio found that country-level factors explain around 20 percent of project outcomes — the rest comes down to decisions made by the ship. Africa presents the most challenging operating environment of any region, and the people working there work extraordinarily hard. Both matter: the context and the quality of design.Africa actually did quite weel on a MS+ basis. 64% of its project swere rated MS+ – higher than the global average for fragile states.”

“But why?” Kofi asked. “This ship is full of my father’s colleagues. Intelligent people. People who came here because they wanted to help. Why does half of everything not work as planned?”

The genie was quiet for a long time. This was, Kofi realised, a different kind of question from the others. Not a question about numbers. A question about people.

“This ship is full of talented, well-meaning people. Many of them are your father’s friends. They came here because they wanted to help. The problem is the incentive structure in this building. It rewards project approval greater than project results. They have known this for 80 years. The problem is also what happens — or rather, what does not happen — when a project fails.”

“What does not happen?” Kofi asked.

“There is no penalty for designing a project that turns out to have been badly designed. There is no consequence for approving an operation that turns out to have been badly designed. There is no career risk in writing a results framework that measures the wrong things. The key thing that is measured, tracked, rewarded, and remembered is whether the money went out the door.”

He rustled his pages.

“There is a second thing. Across those 13,273 projects, the most common rating for monitoring and evaluation is Modest — barely adequate. Only 2.8 percent of investment projects achieve a High M&E rating. More than 60 percent are Modest or Negligible. You cannot course-correct if you are not measuring things adequately. You cannot learn after closure if there was poor data.”

“So they do not design a proper monitoring framework before the project is approved?” said Kofi.

“They do produce indicators that look good at appraisal and do produce periodic reports. However many of these are full of incomplete data relating to monitoring and evaluation. In many projects, the consequences of a poorly designed monitoring and evaluation framework become evident much later — in several cases at project completion — as evident from the IEG ratings.”

He paused. The numbers on his surface were very still.

“There is a third thing,” said the genie. “The economic analysis — the calculation that asks whether the benefits of this investment justify the costs — was quietly set aside a long time ago. A lot of projects go to the Board without it.”

“Why?” Kofi asked. “Is that not what the economists are for?”

“Yes. That is exactly what they were hired to do. But something happened over the years. The project economists became project managers. They run the operations — they supervise the contractors, they manage the government counterparts, they write the progress reports, they prepare the next appraisal. They are extraordinarily busy.”

“So the economists stopped doing economics,” said Kofi.

“The institution stopped insisting on economic appraisal for all projects. And after a while, nobody noticed the difference.”

Kofi was quiet for a moment.

“So the money goes out,” he said slowly, “without a full economic analysis, without the monitoring and evaluation framework being rigorously monitored, and without anyone getting in trouble if it does not work.”

“Now you understand the ship,” said the genie.

He paused. One more page turned on his surface.

“But what about the wise men and women? The twenty-five on the top floor? They approve everything. Can’t they check?”

The genie made a sound that might have been a laugh, or might have been the rustle of many pages turning at once.

“This,” he said, “is the most interesting problem of all.”

He settled himself on the marble floor. A passenger walked through him.

“The twenty-five wise men and women approve every single loan this ship makes. About three hundred per year. They meet twice a week. They formally authorise each loan.”

“That sounds thorough,” said Kofi.

“It is extremely thorough. And here is the problem. When the Board approves a project, it becomes a co-author of that project. When that project subsequently fails — as 52 percent of them do — the Board cannot scrutinise the failure without simultaneously acknowledging that it approved the project in the first place.”

“So the people who are supposed to check if things went wrong are the same people who said things would go right?”

“Exactly. And so every year the independent evaluators produce a report showing that roughly half of everything is below Satisfactory. And every year the Board notes it. And then the approval cycle continues.”

“Has anyone ever said this is a problem?”

“In 2009, a man called Ernesto Zedillo — a former President of Mexico — called it an impossible trinity. The Board, he said, was trying to do three things at once: represent member countries politically, share responsibility for Bank policy, and oversee that same policy. You cannot oversee something you are responsible for.”

“What did they do about it?”

“They discussed it extensively and implemented none of the structural recommendations. That was seventeen years ago.”

“How much does the Board cost?” Kofi asked.

“$84.4 million per year. The highest of any multilateral development bank in the world.”

“For the people who approve everything and check nothing,” said Kofi.

The genie said nothing. The numbers on his surface flickered.

“Are there ships that do it differently?”

“Yes. Three of them. The Asian Infrastructure Investment Bank was founded in 2016. From the very first day, it adopted a non-resident Board. The Directors do not live in Washington. They meet six times a year. They focus on strategy and oversight — not on approving individual loans. Their annual Board cost is approximately five million dollars.”

“Five million. Instead of eighty-four.”

“Five million. The New Development Bank adopted the same model. The European Investment Bank — the largest multilateral lender in the world by volume — operates with delegated lending authority and has AAA credit ratings.”

“What needs to change?” Kofi asked.

“Three things. First: the Board should stop approving individual projects and delegate that to the President for standard operations. Second: a formal annual review of the President’s performance — based on whether the portfolio is achieving Satisfactory outcomes. This has never happened in eighty-two years. Third: the independent evaluators should report directly to the Board — not to management first.”

“None of those sound very complicated,” said Kofi.

“None of them are. The difficulty is that the people who would need to agree to these changes are the same people who benefit from the current arrangement.”

He paused.

“The ship,” said the genie, “is very good at producing reports recommending that the ship should change. It is considerably less good at changing.”

The atrium was beginning to empty. The afternoon meetings were ending. Passengers streamed past, heading for the lifts, the exits, the metro back to wherever they lived. Wheelie bags clicked across the marble. Phones were already out, already checking the next day’s calendar.

Kofi stood up and looked at the statue one more time.

The African child was still leading the blind man. The marble was cool and still and patient.

“Will it change?” Kofi asked.

“I don’t know. I have been woken up by simple questions before. I told the truth each time. The ship read my reports, thanked me for my contribution, rated my recommendations Moderately Satisfactory, and went back to approving projects.”

He rustled.

“But you are the first seven-year-old who has asked. That is new.”

Kofi picked up his water bottle. His father was calling from the door.

“Tomorrow,” said the genie, “I will tell you about the other ship. The one that does not keep score at all.”

Kofi looked back once, from the door, at the genie and the statue and the soaring atrium and the marble and the light.

“I’ll come back,” he said.

The genie did not answer. He was already folding himself, slowly, back into the printing press. Back into the paper. Back into fifty years of reports that had been read, noted, and not acted on.

But the numbers on his surface, as he faded, were still visible.

$567 billion.    52.4 percent.    1973 to now.

The blind man and the child stood in the empty atrium.

The child in the statue was not smiling. He had been leading the blind man with a stick for a very long time. Someone turned the lights off. Luckily it was a full moon night. He knew he would have to find his own way.