Nigeria Power Sector · World Bank Group · MDB Reform Platform
How Not to Do PPPs in Africa: Lessons from the World Bank’s Nigeria Power Sector Interventions, 2009–2019
A Critical Analysis of NEGIP (ICR P106172) and PSGP (ICR P120207)
This paper examines a decade of World Bank power sector interventions in Nigeria through two projects that collectively deployed or guaranteed over $900 million: the Nigeria Electricity and Gas Improvement Project (NEGIP, 2009–2018, rated Moderately Unsatisfactory) and the Power Sector Guarantees Project (PSGP, 2014–2019, rated Moderately Satisfactory). The argument is that both ratings obscure a systemic institutional failure — and that the PSGP in particular represents a case study in what happens when a development institution faces no consequence for getting it wrong.
The Setup: A Broken Market
Nigeria’s fundamental power sector problem at the time of PSGP approval was not insufficient generation capacity. By 2014 the country had over 10 GW of installed generation — and was delivering barely 4,000 MW to the grid. That 60 percent utilisation gap was not a mystery: a transmission system that could not wheel available power to demand centres, and distribution companies collecting 50–60 percent of the energy they received while disconnecting feeders to areas with high losses.
Nigeria did not need more generation capacity behind a broken distribution sector. It needed investment in DISCOs and transmission. The Bank’s Project Appraisal Document acknowledged distribution sector weakness in footnotes. It treated generation as the binding constraint. That diagnosis was wrong, and the consequences are documented in this paper.
The Ignored Precedent: NEGIP (2009–2018)
The Nigeria Electricity and Gas Improvement Project was a $300 million IDA credit plus $600 million in Partial Risk Guarantees, designed to unlock private investment in gas supply. Of the seven to ten gas supply agreements envisaged — with Shell, Chevron, ExxonMobil, Addax, and others — exactly one was ever signed. $488.2 million of the $600 million PRG authority was cancelled. The gas guarantee component’s efficacy was rated Negligible by IEG.
The one deal completed — a gas supply agreement between Accugas and the Calabar Generation Company — illustrates the structural trap. Accugas had a take-or-pay obligation requiring payment for 131 million standard cubic feet per day of gas, regardless of how much the plant actually dispatched. The Calabar plant was dispatched at approximately half capacity because the DISCOs were rejecting load rather than absorbing the electricity. NDPHC was paying for gas it could not use, under a guaranteed contract, while being unable to collect from the distribution companies that had failed to take the power.
The NEGIP ICR’s own lesson: “payment guarantees provide comfort to the seller but do not make the buyer more creditworthy.” This lesson was available to the PSGP preparation team in 2013–2014. The PSGP PAD does not reference NEGIP. The Bank proceeded to approve a second, larger PRG programme for the same sector in the same country with the same structural defects.
The Azura Deal: A Structurally Defective Bargain
The PSGP approved a $237 million IBRD Partial Risk Guarantee to backstop the Azura-Edo IPP — a 459 MW open-cycle gas turbine plant near Benin City, with a total project cost of $877 million, or approximately $1,910–1,950 per kilowatt of installed capacity. The OECD benchmark for an OCGT plant in 2016 was $978 per kilowatt. The Azura cost was approximately twice the benchmark. The project was an unsolicited proposal: the PSGP ICR acknowledges there was “no evidence found by the ICR team of any formal process in the Government for evaluating unsolicited proposals like the Azura project.” The Bank backed an overpriced, non-competitively procured project, then reviewed the EPC construction contract and called it due diligence.
The Azura PPA was structured as a take-or-pay arrangement with a USD-denominated bulk tariff of US¢9.14 per kilowatt-hour. NBET was obligated to pay Azura regardless of how much power the DISCOs absorbed and regardless of how much NBET collected from those DISCOs. The creditworthiness was provided by the FGN sovereign indemnity and the World Bank’s Partial Risk Guarantee. When the naira lost 30 percent of its value against the dollar in June 2016, NBET’s real payment obligations surged. The 2015–16 tariff shortfalls were eventually to be funded under the Power Sector Recovery Programme — a further reform lending operation the Bank approved in 2020, six years after the PSGP.
The Three-Role Conflict
What makes the Nigeria power sector case categorically different from a standard sequencing failure is the structure of the World Bank Group’s financial exposure. On the Azura-Edo transaction, the WBG was playing three roles simultaneously:
The PSGP ICR acknowledges the consequence in a paragraph that deserves to be read in full: the Bank’s guarantee structure required “periodically reminding the Government to honor its guaranteed payment obligations for specific guarantees to the private developer,” and this scrutiny “may strain the relationship with government counterparts, and be at odds with the World Bank’s ability to pursue overall dialogue on sector viability.” The Country Director was making monthly visits to the Finance Minister to ensure NBET’s payments were met and the PRG was not called. The Bank had been converted from policy adviser to debt collector.
The Dispatch Inversion: Nigeria Pays for Power It Cannot Use
Nigeria operates approximately 2,638 MW of installed hydropower across Kainji, Jebba, Shiroro, and Zungeru. These plants run on water. Their marginal cost of generation is between $3 and $5 per megawatt-hour. Under any rational merit-order dispatch regime, hydro is the first call on the Nigerian grid — dispatched ahead of every gas-fired plant in the country.
The take-or-pay structure of Azura and Calabar destroys this logic. Because NBET must pay the full capacity charge for both plants regardless of dispatch, the System Operator faces a contractual obligation that overrides the merit order. Azura and Calabar cost between $110 and $170 per megawatt-hour all-in. They are the most expensive plants on the Nigerian grid. The take-or-pay structure means that when the System Operator must choose between dispatching hydro and dispatching the IPPs, the contractual arithmetic favours the IPPs: if Azura runs, NBET pays for energy it can use; if Azura does not run, NBET still pays the capacity charge in dollars while the hydro megawatt-hours flow for nothing.
The combined annual invoice for Azura and Calabar is approximately $660 million. Nigeria’s four hydro stations at comparable generation volumes would cost approximately $15–20 million per year in operations and maintenance. The differential — approximately $640 million annually — is the cost of the dispatch inversion. As the naira has depreciated from ₦360 to over ₦1,500 per dollar, the naira cost of this dollar drain has quadrupled. The volume of electricity has not changed.
The Timeline of Institutional Failure
The Approval Chain: Who Signed Off
The paper names, with documentary evidence, the individuals who carried these projects to approval. The argument is not that they acted in bad faith. It is that the institutional system gave them no structural reason to confront the evidence that was already in front of them.
NEGIP (P106172) — Board Approval June 2009
Country Director, Nigeria: Onno Ruhl. Statement at approval: “We are especially excited about the prospect that our support to the power sector might help solve the perennial problem of generation capacity lying idle whilst Nigerians stay without light.”
At the time of NEGIP’s approval, Nigeria already had substantial installed generation capacity being underutilised — not primarily because of gas shortages, but because of transmission constraints and the financial collapse of the distribution sector. The problem acknowledged in 2009 was still present at PSGP project closure in 2019.
PSGP (P120207) — Statutory Committee Report, April 16, 2014
Signatories: Senior VP & Group General Counsel, IBRD; Makhtar Diop, Regional Vice President, Africa; FRN Nominee.
Task Team: Country Director: Onno Ruhl; Sector Manager: Paul Noumba Um; Project Team Leader: Peter J. Mousley.
Makhtar Diop signed the Statutory Committee Report as the Regional Vice President responsible for the Africa Region, and in that capacity was the senior officer who carried the PSGP to the Board for approval. He subsequently became CEO of IFC. The paper documents that the PSGP was submitted to his office for factual review in March 2026. No response has been received.
PSGP Executive Board Chair Summary, May 1, 2014
The Board acknowledged “the high-risk high-reward nature of this transformational proposal” and noted that “achievement of the Project’s development impact requires long-term financial viability of the sector, including a sound regulatory framework, and a reduction in the high level of technical and commercial losses in the power system.”
ATC&C losses were running at 35 percent at approval. The Board stated their reduction was a prerequisite for the project’s development impact. They did not reduce. They remained at 35–50 percent throughout implementation. No consequence followed.
The Replication Risk
The most troubling dimension of this story is not what happened in Nigeria. It is what the Bank and IFC are doing with the lessons. IFC is actively structuring comparable PRG-backed, USD-denominated, take-or-pay IPPs across West and East Africa — in Ghana, Côte d’Ivoire, Senegal, and elsewhere. The Azura “template-making contracts” (PSGP ICR, para. 47) are being applied to markets whose distribution sectors may be even weaker than Nigeria’s was in 2014.
The PSGP ICR was published in 2021. The NEGIP ICR was published in 2020. Neither document references the other. The institutional ecosystem — preparation teams, country departments, sector practice groups — appears to operate with no systematic mechanism for ensuring that the lessons documented in one project’s ICR are applied in the design of a subsequent project in the same sector. The ICR is written. The lessons are recorded. And then nothing happens.