Climate finance: Earning trust through consistent reporting: Executive summary
Executive Summary
Overview
What is climate finance? Despite the importance of this question, and recent claims that developed countries[1] have reached the goal agreed in 2009 to provide US$100 billion of climate finance,[2] there is no agreement on how it should be defined. This has not only led to doubts about the true value of finance raised, but also contributed to a breakdown of trust between developed and developing country Parties to the United Nations Climate Change Convention (UNFCCC). This year the Parties come together to agree on the New Collective Quantified Goal (NCQG) to replace the US$100 billion goal, and the question of how progress towards it should be measured must be at the centre of the debate, not an afterthought.
This report combines insights from climate finance experts (from national governments, the OECD, UNFCCC and civil society organisations), as well as analysis from Development Initiatives (DI) to highlight where consistency is lacking in climate finance reporting – over time, within and between donors. It sets out recommendations that we hope will help providers adopt the best-practice practical steps they can take to drive consistency in their reporting and help build a trusted and transparent climate finance data system. These include:
- greater consistency in how climate finance is defined and reported,
- better shared review and auditing processes,
- better reporting on impact, to build a stronger understanding of whether funding is enough and well targeted, and
- approaches to remove the capacity constraints that might limit the implementation of these changes.
It will also be of interest to countries that themselves receive climate finance and advocates looking to ensure the system works for everyone, focusing minds on the first-best solution to the problem: a single clear definition, supported by a transparent reporting system.
Key findings
Inconsistent measurement has eroded trust and made it difficult to track progress
A measure of climate finance should allow us to track progress over time, compare how providers are helping partner countries achieve their climate goals, and establish how close we are to meeting needs. However, inconsistency in measurement means that the current estimates of climate finance fail on each count:
- Lack of consistency over time: The lack of a common definition of climate finance has allowed providers to adapt their methodologies for tracking it in order to meet political targets more easily. Even where providers have not explicitly changed their methodologies, the subjectivity involved in assessing how much of a project should count as climate finance has allowed reporting to become much looser over time. This means that we do not know by how much climate finance provision has increased. While the Organisation for Economic Co-operation and Development (OECD) estimates that US$115.9 billion[3] was provided in 2022, compared to US$52 billion in 2013, much of this increase may have come from reporting changes. For example, we know that when the US$100 billion goal was agreed in 2009, a large share of bilateral aid was not screened for its impact on climate. That means that developed countries must have been spending more on climate finance than the data shows. This might be a good thing, but it nevertheless implies that the increase since then has been smaller than official numbers suggest. Furthermore, there is evidence suggesting that more projects would have included as climate finance in the past if judged by today’s standards.[4]
- Lack of consistency across providers: Different countries count similar projects in different ways, and so their estimates of climate finance provision are not comparable. This makes it impossible to accurately assess which countries are providing climate finance support that matches their means. For example, some countries counted Covid-19 response measures as adaptation, whereas most did not. Some countries are counting core contributions to certain trust funds as having a principal climate focus, while others do not deem such contributions relevant to climate finance. More systematically, when Development Initiatives and other researchers used machine learning models to assess climate finance against a common benchmark, it showed dramatic variation in the quality of reporting across countries.
- Lack of comparable measure of need: We know that the need for climate finance is huge. But because there are so few restrictions on what activities providers can claim as climate relevant, we cannot have confidence that the finance reported is really addressing those needs.
There is even a lack of consistency in the data countries self-report to different organisations. Differences in the point of measurement for reporting to the UNFCCC and OECD mean that although these databases contain information on largely the same projects, they do not provide a consistent picture of climate finance.
The Paris Agreement requires data on climate finance to be reported to the UNFCCC, and so it should be a more authoritative and comprehensive source. However, the information currently submitted to the UNFCCC by providers is insufficient to determine why some projects are included as climate finance, how much has actually been spent towards commitments made, or even what it means to have made a commitment. This lack of detail means it is rarely used.
By contrast, data reported to the OECD is carefully curated but lacks crucial information on flows beyond official development assistance (ODA) and on the share of project expenditure reported to the UNFCCC as climate finance.
A common definition is the first-best approach, but other steps can be taken that would improve consistency
The best way to ensure that climate finance is being measured consistently across time and providers is to agree on a common definition that is based on a common understanding of need – but agreeing on a definition that has sufficient detail would need huge political will. In submissions to the UNFCCC Standing Committee on Finance (SCF), developed countries have argued that a common definition is unnecessary. However, other definitions that guide providers in specific circumstances do exist and add value, such as the ASEAN Taxonomy for Sustainable Finance, or the Climate Bonds Taxonomy.
This report therefore explores ways in which consistency in reporting could be improved in the immediate absence of a definition, based on interviews with a number of climate officials from developed countries, ex-officials and development finance experts at the OECD. These recommendations have been developed to be politically and technically feasible (some are already being implemented by some provider agencies). While they are second-best to a full, detailed definition, we hope that they could improve the state of reporting, and so heighten the impact of the NCQG, while discussions on definitions continue.
Recommendations
1. Transparency requirements need to be expanded
Three basic additions to the UNFCCC transparency requirements would give greater clarity:
- Countries that calculate the percentage of a project's funding that is counted as climate finance (climate share) on a case-by-case basis should be required to report the percentage of the total project value that is included as climate finance.
- Countries should not only agree on a standard, stricter definition of ‘commitment’ to engender trust that such commitments will lead to actual spending, but also track this spending against corresponding commitments made (in a way easily distinguishable from new commitments, to avoid double counting).
- Countries should provide links to project documentation for projects over a certain value. And where projects are also reported to the OECD, countries should provide project codes to allow linking between databases.
2. Climate finance assessment should be as granular as possible
Aggregated assessments make it hard to know if the amount of climate finance counted reflects the true nature of the project. More granular assessments that assess the climate focus of individual activities, or even transactions, would give a better estimate and correct for the imprecise marker system.
3. Parties should consider using novel techniques to ease capacity constraints
Many countries face significant capacity constraints when it comes to quality assuring the climate finance data submitted to the UNFCCC, as countries often embark on thousands of projects in any given year. In addition, each country measures climate finance by its own unique standard which reduces comparability.
Individual countries could train natural language processing models (NLP – a form of machine learning)[5] on a selection of projects that have been manually classified according to their climate focus, creating an automated process to identify questionable marking decisions. Additionally, a centralised body, such as the SCF, could train an NLP model on a wider selection of projects across all countries, creating a common benchmark for assessing reporting from individual countries.
Such models would not replace human judgement, but could facilitate quicker checking by highlighting projects whose descriptions do not accord with their markings.
4. The existing UNFCCC peer review process should be strengthened
There is an existing process for peer reviewing the transparency and completeness of the Biennial Reports countries submit to the UNFCCC. This is valuable but focuses on ensuring that projects are consistent with COP agreements.
The UNFCCC should expand review requirements to include an assessment of the quality of climate finance reported. For example, the review team could scrutinise a sample of projects to highlight where reporting approaches are out of line with common practice or capturing projects with questionable relevance, and ensure that the projects are consistent with the requirements outlined in partner countries’ nationally determined contributions[6] or national adaptation plans[7] (which set out Parties’ commitments for reducing greenhouse gas emissions and their adaptation needs respectively).
5. Countries should report on impact estimates before and after implementation where possible
The quantity and quality of climate finance are both essential concerns. Given that finance is scarce, relative to the size of the problem, countries need to ensure it is as impactful as possible. Publishing better information on ex-post impact will help assess this, but the more fundamental question is whether a project should be counted as climate finance in the first place. If a country cannot explain, ex-ante, the impact a project will have on climate finance goals, it should not be able to count it.
6. There should be official guidance from the UNFCCC on a case-by-case approach to assessing the ‘climate share’ of projects
Most countries employ the OECD’s Rio marker system to identify their climate finance, which uses a three-point scale to grade climate focus. However, this was never intended to be a quantitative system of measurement – most countries decided to use it is because it was a readily implementable system that was already integrated with other reporting.
Individually, some countries already use case-by-case calculations (which assess the climate share of each project rather than using the marker’s three-part scale). But despite the widespread view among CSOs and many of the specialists we interviewed that this approach is preferable, there is no common methodology.
The Rio markers have a handbook that gives substantial guidance on implementation, yet there is nothing equivalent for the case-by-case approach. This could easily be developed from the internal guidelines that exist in countries already using the approach. It would enhance transparency and potentially make the system more attractive to providers.
Conclusion: Towards greater consistency
These recommendations will not solve all the problems associated with climate finance. Many of these issues are highly political: the difficulty of spending money abroad (especially when domestic problems and debt-burdens have mounted); the fact that many climate finance projects are more about promoting domestic firms[8] or exporting technology; and the fact that many countries have an interest in reporting climate finance figures that show them in a good light.
But the first step in improving the quantity and quality of finance is understanding the current landscape, and this requires consistent measurement across time and providers. These, largely technical, recommendations that emerged from our conversations with officials would create greater consistency in reporting: an essential step towards more, and more effective, climate finance.
Notes
-
1
The terms ‘developed’ and ‘developing’ appear frequently in both COP texts and the Paris Agreement – we repeat them here to match those documents. There is a glossary that defines other key climate finance terms used in this report available at https://devinit.org/resources/climate-finance-earning-trust-through-consistent-reporting/glossary-of-climate-finance-terms/Return to source text
-
2
UN Statistics Division, Background note on the USD 100 billion goal in the context of UNFCCC process, in relation to advancing on SDG indicator 13.a.1. Available at: https://unstats.un.org/sdgs/tierIII-indicators/files/13.a.1_Background.pdfReturn to source text
-
3
OECD, 2024. Climate Finance Provided and Mobilised by Developed Countries in 2013–2022. Available at: https://www.oecd.org/environment/climate-finance-provided-and-mobilised-by-developed-countries-in-2013-2022-19150727-en.htmReturn to source text
-
4
You can read this and more quantitative analysis in Chapter 3.Return to source text
-
5
Natural language processing models (NLPs) are types of machine learning models that identify patterns in text data and can link these patterns to different categories (among other things). For an example of their application, see Development Initiatives, 2023. Is climate finance wrongly reported by over a billion dollars per year? Available at: https://devinit.org/blog/climate-finance-wrongly-reported-ai-world-bank-fcdoReturn to source text