Where Integrity Risks Enter Climate Finance

Climate finance integrity is often approached through the lens of financial misuse: bribery, fraud, embezzlement, collusion, or diversion. Those risks are real. The evidence reviewed for this note, however, shows a wider set of vulnerabilities. Integrity can be weakened before implementation begins, during procurement and delivery, through poor participation or weak institutional capacity, in the systems used to measure results, and in the way institutions respond when warning signs or evaluation findings emerge.

This note reviews evidence from climate-finance governance research, multilateral evaluation portfolios, public-investment and procurement research, documented project cases, and institutional review material. The sources are deliberately treated according to what they can and cannot establish. Portfolio evaluations are used to identify recurring weaknesses across defined groups of projects. Country studies and case databases illustrate mechanisms and contexts, not global prevalence. General public-investment evidence is used only where it helps explain a risk mechanism that is also relevant to climate finance.

Across that evidence, four recurring patterns are visible. First, integrity risk is distributed across the lifecycle rather than concentrated at the point of disbursement. Second, corruption is only one category of integrity weakness; design quality, institutional capability, participation, monitoring, and evidence can also determine whether climate outcomes merit confidence. Third, formal compliance and positive aggregate ratings can coexist with weaknesses in particular dimensions. Fourth, the usefulness of controls depends on whether institutions can convert information into challenge, corrective action, and learning.

This evidence also helps explain one dimension of The Integrity Gap: formal safeguards may exist while weaknesses in design, implementation, evidence, or institutional response still reduce confidence in how climate finance is delivered.

Two additional cautions shape the analysis. First, integrity is used here as a governance concept, not as a synonym for any adverse outcome. Poor technical performance, delay, or disappointing results do not automatically constitute integrity failures. Second, the note distinguishes allegation, evaluation finding, estimate, and institutional conclusion. These terms are not interchangeable and are preserved as such throughout.

2. Before implementation: design and allocation can shape integrity outcomes

The project lifecycle does not begin when finance is disbursed. Before implementation, decisions have already been made about what problem is being financed, which locations or groups are prioritized, which institution will implement the intervention, what assumptions underpin the design, and how success will be measured. Those decisions can determine much of what follows.

The Adaptation Fund Technical Evaluation Reference Group (AF-TERG) provides one of the clearest portfolio-level views. Its second synthesis reviewed 12 final evaluations from projects and programs implemented across Asia, Latin America and the Caribbean, and Sub-Saharan Africa. Among the recurring issues were project-design weaknesses, including inadequate stakeholder analysis, insufficient needs or capacity-gap analysis, over-ambitious geographic scope or targets, and weak links between activities and intended outcomes. AF-TERG therefore recommended stronger due diligence during project development to connect stakeholder, beneficiary-needs, capacity, and inclusion analysis to the theory of change and risk-management arrangements.

This is not evidence of corruption. It points to a different integrity problem: a project can be formally approved and competently administered while still resting on assumptions that are insufficiently tested. Where the evidence base is weak, the most affected groups are poorly understood, or the causal link between activities and outcomes is thin, formal approval does not cure the underlying weakness.

Allocation also creates a distinct category of risk because climate resources are scarce and selection decisions create winners and losers. U4’s 2025 review of climate governance argues that climate finance is not always allocated solely according to mitigation potential or vulnerability; political and economic considerations can also shape flows. U4 treats this as an area requiring stronger integrity controls and further research, not as proof that particular allocations are corrupt.

Country-level evidence from Bangladesh shows what more direct distortion can look like. Transparency International Bangladesh (TIB) reported governance weaknesses in project formulation, including inadequate identification of target areas and insufficient attention to local needs. Its 2025 study also reported alleged bribery and illegal transactions during project approvals. Those findings are specific to the Bangladesh system and should not be generalized. They nevertheless illustrate why allocation integrity matters: if influence changes which projects receive finance, the integrity consequence can arise before implementation begins.

3. Procurement and contracting: climate purpose does not remove ordinary exposure

Once a project is approved, procurement converts program choices into contracts, counterparties, and financial commitments. Climate projects frequently require civil works, equipment, technical services, consultants, technology, land-related services, or specialized contractors. Procurement therefore becomes a point where discretion, information asymmetry, commercial incentives, and public resources meet.

The OECD Climate Adaptation Investment Framework makes this connection explicitly. It identifies fairness and trust as part of the enabling environment for adaptation investment and notes that public officials must address conflicts of interest, corruption, and fraud, particularly in public procurement, which is a major source of finance for adaptation investments. The framework also recommends that procurement policies account for benefits over the lifecycle of the investment, rather than focusing narrowly on upfront transactions.

The broader OECD Anti-Corruption and Integrity Outlook 2026 explains why procurement is structurally exposed: transactions are complex; large sums are involved; public and private actors interact closely; and supply chains can be fragmented. It also highlights an implementation problem. Legal provisions are often established, but risk-based tools, data integration, supplier integrity measures, and proactive detection remain uneven across countries. This evidence is not climate-finance-specific. Its relevance is structural: climate-financed procurement operates through many of the same public-investment mechanisms.

Climate-specific evidence then shows how these mechanisms can materialize. TIB reported alleged irregularities in contractor appointment and collusion in tendering and contractor selection in the Bangladesh climate-finance system it studied. Transparency International’s Climate & Corruption Case Atlas documents an illustrative Kenyan geothermal case in which senior managers of the state-owned Geothermal Development Company were charged over an allegedly inflated equipment-transport contract. Transparency International reports that a contract worth about US$19.5 million was awarded for work that the same contractor had reportedly provided elsewhere for approximately US$230,000.

The Kenyan case should not be treated as representative of renewable-energy procurement. Its analytical value is narrower. It demonstrates that a climate-positive asset and a compromised procurement process can coexist. The environmental purpose of an investment does not itself protect competition, pricing, conflicts of interest, or value for money.

4. Implementation: integrity depends partly on institutional capability

Implementation is often where integrity risk becomes most visible. Funds can be diverted, goods can be substituted, delivery can be falsely certified, or activities can differ materially from the approved design. But the evaluation evidence shows another class of implementation weakness that should not be mislabeled as corruption: institutions can lack the capacity to operate the systems on which credible delivery depends.

AF-TERG identified recurring operational and logistical challenges across the final evaluations it reviewed. These included high turnover or weak capacity among implementing partners, delayed release of funds, cumbersome procurement, inadequate staffing, poor decision-making that contributed to delays, and weak reporting. These findings matter because controls do not operate themselves. A procurement policy, monitoring plan, or risk register can exist on paper while insufficient staff, weak coordination, or low technical capability prevents it from functioning effectively.

The distinction is important for integrity analysis. A low-capacity implementing entity is not necessarily an unethical one. But capability becomes an integrity concern when it materially weakens the institution’s ability to steward resources, detect deviations, produce reliable information, or correct problems. In that sense, institutional readiness is not merely an efficiency variable; it can be part of the integrity environment surrounding the project.

The Bangladesh study adds a different kind of implementation evidence. TIB reported fund embezzlement during implementation among the corruption risks identified in the BCCT portfolio. Again, that is a country-specific finding. Placed alongside the evaluation evidence, it helps distinguish two pathways through which implementation can fail: deliberate misuse on one hand, and inadequate institutional capability on the other. They require different controls, but both can weaken the connection between approved finance and intended climate outcomes.

5. Beneficiary participation and inclusion: when process is part of the outcome

Climate adaptation makes the relationship between process and outcome especially visible. Many adaptation interventions depend on local knowledge, behavioral uptake, maintenance, collective action, or sustained use by the communities they are intended to support. Participation is therefore not always a procedural accessory. In some projects it forms part of the causal pathway through which resilience is supposed to improve.

AF-TERG reported lack of beneficiary participation in eight of the 12 evaluation reports reviewed. In some projects, insufficient participation contributed to non-acceptance of technologies, products, or livelihoods introduced by the intervention. The synthesis also identified insufficient gender and youth perspective in planning and implementation and recommended stronger beneficiary-needs, stakeholder, capacity, and inclusion analysis during project development.

GEF Independent Evaluation Office evidence points in the same direction from a different portfolio. Its 2023 Annual Performance Report examined projects seeking behavioral change and found that successful projects tended to assess stakeholder needs and design interventions around barriers to adoption. The report emphasized participatory, systems-based engagement in projects that achieved intended behavioral outcomes.

The integrity implication is precise. A consultation can occur and still fail to perform the function for which consultation exists. Attendance can be documented while relevant knowledge, distributional effects, or barriers to uptake remain absent from the project design. Formal participation and substantive participation are therefore different questions.

6. Monitoring and reporting: activity becomes evidence only through credible measurement

Climate finance ultimately depends on assertions: emissions were reduced, vulnerability declined, resilience improved, ecosystems recovered, technologies were adopted, or intended beneficiaries were reached. Those assertions require evidence. Monitoring is therefore not simply an administrative requirement attached to implementation; it is the bridge between activity and a credible assertion of climate impact.

The Adaptation Fund synthesis found weak monitoring and evaluation systems in 10 of the 12 final evaluation reports reviewed. Contributing factors included poor design, lack of mechanisms to measure improvement, delayed institutionalization of M&E, inadequate uptake of monitoring tools, lack of dedicated M&E personnel, minimal budget provision, and the absence in some projects of systems capable of tracking longer-term adaptation results. AF-TERG noted that some projects had to commission additional data collection at midterm and cautioned that retrospective data collection brings its own limitations.

GEF IEO evidence provides a second lens. Its 2023 Annual Performance Report found that almost 45 percent of projects intended to promote behavior change lacked indicators capable of tracking that change. The report also found that 69 percent of the wider project cohort was rated satisfactory for M&E design and 66 percent for implementation of M&E plans. These figures should not be read as contradictory. They show that an M&E system can be broadly satisfactory while still leaving important outcome questions imperfectly measured.

Keeping the two evidence sources separate matters. They examine different portfolios and answer different questions. They should not be combined into a single statistic. Together, however, they support a common analytical point: the ability to establish what changed depends on indicator design, data systems, monitoring capacity, and time horizon.

This is where evidence quality becomes an integrity issue. A reported result does not become more credible simply because it appears in a project document. Confidence depends on whether the underlying measurement can distinguish outputs from outcomes, whether data are complete and reliable, and whether the monitoring system is capable of detecting weak or null results as well as positive ones.

Monitoring produces information. Integrity systems still require a way to challenge that information, surface concerns, and act when something appears wrong. This is the point at which verification, complaints channels, audit, whistleblowing, evaluation, management escalation, and investigation interact.

Transparency International’s climate-finance guidance has long emphasized that people affected by climate projects need safe and accessible channels to raise grievances or corruption allegations throughout the project cycle, from allocation through implementation and evaluation. The significance is not simply that a complaints mechanism exists. Accessibility, referral, mandate clarity, protection, follow-up, and the ability to produce remedy determine whether it is useful in practice.

The UNDP/GEF Standards and Labels for Promoting Energy Efficiency in Russia project illustrates the distinction between detection and response. The project ran from 2010 to 2017 with about US$7.8 million in GEF financing and was intended to strengthen energy-efficiency standards and reduce greenhouse-gas emissions. UNDP later commissioned an independent external review of the project’s governance, risk management, controls, fiscal management, conflicts of interest, and allegations of fraud. The resulting January 2021 report, Systems and Silos, documented repeated warning signs during implementation, beginning with a 2012 management audit that found major deficiencies in procurement, and concluded that management actions appear to have been severely deficient along multiple dimensions. In its words, the failures were not caused by a lack of warnings but despite the warnings. The review’s value here is institutional: it shows how controls can generate warnings without those warnings producing timely corrective action.

The case is not evidence about climate finance generally. It is useful because it isolates a mechanism: an institution can possess multiple control channels and still fail if information is not escalated, responsibilities are unclear, warnings are repeatedly returned to the same actors, or management response is weak. Detection and response are separate capabilities.

Evaluation systems face a related challenge. In its review of evaluation quality, AF-TERG found that the utility dimension was among the weaker areas: several reports did not clearly identify the main users of the evaluation or what follow-up would ensure uptake of lessons and recommendations. The Adaptation Fund now uses management action trackers to record agreed actions, responsibilities, timelines, and implementation progress after evaluations. That is important because evaluation only strengthens integrity if findings are converted into accountable follow-up.

Climate-finance projects have administrative endpoints. Climate outcomes often do not. Adaptation makes this particularly clear because resilience may evolve after implementation, and sustainability depends on whether institutions, communities, budgets, maintenance arrangements, and behavioral changes persist.

The Adaptation Fund conducts ex post evaluations three to five years after selected projects close specifically to assess how impacts have evolved and whether outcomes are being sustained. Its final-evaluation synthesis also identified the absence of M&E systems capable of tracking longer-term adaptation results as a recurring problem. This creates a basic tension: institutions need administrative closure, while some climate outcomes require time to become visible.

Positive project ratings therefore require interpretation. AF-TERG found that around 90 percent of the projects and programs in the second synthesis with overall ratings were assessed between Moderately Satisfactory and Highly Satisfactory. The same synthesis nevertheless identified recurring weaknesses in design, participation, M&E, implementation capacity, and evaluation utility. There is no contradiction. Aggregate ratings combine multiple dimensions. A project can perform strongly overall and still have consequential weaknesses underneath the headline rating.

This is precisely why close-out should not be treated as a final integrity verdict. The relevant questions are more granular: was the purpose achieved; are the benefits likely to endure; were weak results visible; were unresolved risks carried forward; and did lessons alter future practice?

9. Corruption evidence: useful, but easy to overstate

There is a temptation in climate-finance integrity research to accumulate dramatic corruption cases and allow the cases to stand in for the system. That produces compelling storytelling but weak inference.

Transparency International’s Climate & Corruption Case Atlas currently contains more than 100 publicly reported cases across climate policy, renewable energy, adaptation, land, and climate-related finance. Its 2024 analytical report drew lessons from 80 documented cases. The Atlas is valuable because it shows the range of reported behaviors and enabling conditions. Transparency International also cautions that the database is neither comprehensive nor systematic. Cases are drawn from public reporting, complaints mechanisms, national chapters, literature, and other sources; they are not a representative sample of global climate finance.

The same discipline is necessary when using the Bangladesh figures. TIB reported that 891 BCCT projects were approved between 2010 and 2024 with an estimated allocation of US$458.5 million, and it estimated that US$248.4 million was lost through corruption. That is a striking estimate from a specific national study.

This distinction matters for CFIN’s wider research agenda. The evidence is strong enough to justify a lifecycle lens and a broader conception of integrity than corruption alone. It is not yet strong enough to measure how much integrity failure exists across global climate finance or to rank lifecycle stages by prevalence. Those are research questions, not conclusions.