ODI Global · Strategy Critique · MDB Reform Platform
Evidence Without Accountability: A Critique of ODI Global’s 2026–2031 Strategy
How the development sector’s leading think tank avoids the most important question in development finance
I. The Plan
ODI Global’s 2026–2031 strategy diagnoses a world in which multilateralism is imperilled, democratic norms are under strain, and the costs of instability fall on those with the least agency. It promises robust evidence, equitable partnerships with Global South institutions, coalition-building, and narrative-shaping across four priorities: institutions and power, equity and rights, crisis and resilience, and sustainable economic transformation.
ODI will leverage its London and Brussels offices, scale its advisory arm, and use the Fellowship Scheme’s 1,300-strong alumni network. Impact will be assessed through “plausible narratives, drawing on metrics and indicators where possible, and inference where necessary.”
The flagship contribution of the past strategy period is the MDB client survey — nearly 900 officials across 125 countries — and ODI’s role in shaping the G20 MDB reform agenda. ODI now serves as Secretariat of the IDA technical working group ahead of the IDA21 mid-term review. Because ODI occupies such a central convening role in MDB reform debates, its analytical choices — what it measures, what it asks, and what it omits — directly shape the information environment in which replenishment and scaling decisions are made. Its silence on the issues raised below is consequential rather than incidental.
II. The Admission
In a candid 2025 keynote, ODI’s Chief Executive Sara Pantuliano described inheriting an institution that faced near-bankruptcy in 2019 — a project-based, donor-driven contract research house, chasing large contracts that offered little space for innovation or independence. Staff fell from 250 to 190. The budget dropped from £40 million to £32 million. Major funding streams were shed.
Charity Commission filings confirm the restructuring. Total income fell from £34.96 million (2020–21) to £27.44 million (2021–22) before recovering to £34.81 million (2024–25). Government contracts collapsed from £4.80 million to under £200,000. ODI Global Advisory revenue dropped from £10.20 million to £3.38 million. Approximately 78 percent of ODI’s income comes from charitable research funded by bilateral and multilateral donors — FCDO, EU institutions, Nordic agencies, and the MDBs themselves — with the Fellowship Scheme supported by FCDO, Australia’s DFAT, the Global Fund, and WHO.
This means that for much of the previous decade, ODI’s enormous output was driven less by a theory of change than by the need to meet contract deliverables. Pantuliano’s admission is valuable. The question is whether the reinvention addresses the fundamental problem.
III. The Evidence Gap
The new strategy avoids the most consequential evidence question in development finance. Not because the evidence does not exist, but because the institutional incentives run against confronting it. Throughout this section, “delivery failure” refers to failure to achieve stated development objectives as assessed under the World Bank Group’s own independent evaluation methodology.
The Survey: What Was Asked and What Was Not
ODI’s client survey is presented as “the most comprehensive, comparative assessment currently available” of how client countries view MDB performance. It covers 11 MDBs across 125 countries. Its findings have informed G20 Evolution Roadmap discussions, shareholder debates, and replenishment framing.
The survey’s questionnaire architecture systematically privileges questions about institutional relevance, demand, and client satisfaction over questions about whether MDB-funded projects actually achieved their development objectives. The survey asks government officials to rate MDBs on: relevance of roles and functions; perceived effectiveness; alignment with country priorities; quality of financing terms; demand for future grants and loans; quality of policy advice; coordination among MDBs; and length of the project cycle. What the survey does not ask — in any form — is whether the projects and programmes financed by those MDBs produced Satisfactory development outcomes as independently evaluated.
The respondent pool itself creates predictable response dynamics. The survey population consists largely of finance ministry, planning ministry, and line ministry officials who manage ongoing MDB financing relationships — actors institutionally embedded within those relationships, who negotiate replenishment access and manage active portfolios. Their incentives to report positively on relevance, alignment, and future demand are structural, not personal. The survey design does not adjust for this.
The consequence is that the survey operationalises “effectiveness” primarily as usability, responsiveness, and financing availability — not as demonstrated development outcomes. It implicitly equates stronger client demand with institutional success, without testing whether increased financing has produced successful results. In a context where IDA has doubled its lending volume while its Satisfactory outcome rate has remained flat at 31 percent for two decades, the distinction between demand and delivery is not academic.
What the Survey Didn’t Tell Respondents
The IEG Master Database — the most comprehensive longitudinal record of development project performance in existence — shows what respondents were never presented with (Brar, 2026). Using the Satisfactory-or-above (S+) threshold — the standard historically treated as the meaningful pass threshold across the Bank Group’s evaluation architecture (see Annex A of the full paper):
The sectoral data is worse. The Macroeconomics, Trade and Investment Global Practice — the largest IDA portfolio at $26.9 billion — achieves Satisfactory outcomes 11.7 percent of the time. Development Policy Financing in IDA fragile states achieves 10.8 percent. Sub-Saharan Africa, which receives 70 percent of IDA resources, committed $98 billion over the period, of which $71.5 billion — 73 percent — went to below-Satisfactory projects. IFC projects in fragile states collapsed from above 60 percent Satisfactory to 11 percent in CY2020–22.
A finance ministry official in an IDA-eligible country answering the ODI survey had no access to this information. A survey that claims to be the most comprehensive assessment of how client countries view MDB performance, while withholding the most comprehensive assessment of how MDB projects actually performed, is not a diagnostic instrument. It is an exercise in managed perception — and one whose conclusions are readily deployable in support of arguments for expanding MDB lending volumes, precisely the institutional positioning of the organisation that produced it.
The 91 Percent Problem
The information gap extends into IDA’s own governance. The IDA21 Deputies Report — on which donor governments committed $23.7 billion in pledges, leveraged to $100 billion total — states that IEG ratings show “satisfactory performance in 91 percent of IDA-financed operations.” This uses the Moderately Satisfactory or above (MS+) threshold, which includes every project that partially achieved its objectives. The actual S+ rate is approximately 31 percent.
ODI, which serves as Secretariat of the IDA technical working group and positions itself as the leading independent voice on MDB reform, has never flagged this gap — not in its survey, not in its reform publications, and not in its advisory work with shareholders. For an institution whose strategy is built around “robust and compelling evidence to inform our ideas and recommendations,” the omission is difficult to reconcile with the claim.
Why This Has Persisted
The persistence of the MS+ benchmark and the survey’s design choices are not accidental. They are sustained by a political economy in which no major actor has strong incentives to redefine success downward. Shareholders prefer expandable institutions with positive headline metrics; replenishment politics reward optimism; management’s credibility rests on demonstrated results; and think tanks that depend on the same shareholders for funding face significant institutional risk in challenging the reporting framework.
ODI’s position illustrates the dynamic precisely. The survey was funded by the Open Societies Foundation and the Gates Foundation, reviewed in draft by staff from eight MDBs including the World Bank, and disseminated through channels that amplify its reform-friendly conclusions. The new strategy, for all its eloquence about independence, does not resolve this conflict. It points toward more narrative building, more coalition convening, more advisory services — activities that keep ODI useful to the institutions whose performance it might otherwise examine.
What This Costs
When reform debates centre on capital adequacy rather than delivery capability, the consequences are operational. Instruments that consistently fail — DPF in fragile states at 10.8 percent — continue to be deployed at scale because the failure rate is invisible in headline figures. Proven alternatives — Programme-for-Results at approximately 61 percent — remain a small fraction of the portfolio. Implementing partners with strong delivery records are excluded from IDA’s highly centralised delivery architecture.
At 31 percent Satisfactory, with the Bank’s largest Global Practice succeeding one time in ten and DPF in fragile states failing nine times in ten, “the question that every implementing partner, every concessional financier must ask is: why should 100 percent of IDA resources flow exclusively through a delivery platform that fails at this rate?” (Brar, 2026).
IV. The Larger Problem
This note uses ODI as the case study, but the problem is systemic. MDB reform discourse increasingly privileges politically functional metrics over outcome-accountable evaluation, allowing expansion and replenishment to proceed without confronting underlying delivery performance. Once demand, satisfaction, coordination, and institutional relevance become proxies for success, actual delivery performance disappears from the conversation entirely.
Related Analysis
The Survey That Did Not Ask the Right Question — The earlier response to ODI’s survey report, with sector-level IEG data for Africa.
IDA: Why Is the World’s Largest Concessional Fund Delivering Satisfactory Outcomes on Only 31% of Its Portfolio? — The full IDA performance analysis. $117 billion below standard.
The FCV Strategy Submission — Including Annex B: The Case for S+ as the Correct Performance Benchmark.
Institutional Power Architecture and Portfolio Distortion — Why MTI at 11.7% is a structural problem, not a performance outlier.
Why the System Does Not Learn — A game theory analysis of the World Bank’s institutional equilibrium.