The gap between promises and payments

In 2009, rich countries pledged at the UN climate talks to mobilise US$ 100 billion a year in climate finance by 2020. The idea was that the countries that did the most to cause climate change, through past and present emissions, should help the countries hit hardest by it.

Rich countries missed the target in 2020 and 2021, and it became a sore point in UN climate negotiations.

In 2024, the OECD reported that the goal had finally been met in 2022. Its latest figures indicate that the goal was also met in 2023 and 2024, based on data that mostly reflects commitments.

Climate finance is hard to track because the OECD’s data focuses on promised money, not money actually paid. Payments get recorded, but often as separate entries with no link to the promise. Four development banks, including the IFC, publish no project payments at all.

ONE Data developed an innovative methodology to connect disparate datasets to follow data on commitments through to disbursements.

Using Splink, an open-source tool built by the UK Ministry of Justice to link records across datasets on prison, court and parole records, ONE Data traced each project from the original promise to the payments made against it. It then checked those links against development banks’ own payment records and the International Aid Transparency Initiative, a public database of aid projects.

ONE Data also used AI to read the descriptions of more than 300,000 climate-marked project records and flag those that state no climate aim.²

Our data shows that actual disbursements from developed countries amounted to US$ 60.4 billion of public climate finance.³

The chart below begins with the OECD’s headline figure and removes the money we cannot trace to a payment.

So developed countries paid US$ 60.4 billion against the US$ 101.6 billion in public climate finance that the OECD counts, a gap of US$ 41 billion.

Commitments always run ahead of disbursements, given that money paid in a given year often settles promises made in previous years, and not every promise is paid in full. But reported finance has exceeded payments every year since 2020.

Governments and EU institutions have paid the largest share of climate finance in recent years. Many of them are now cutting aid.

Much of climate finance is counted as ODA, and 64% of bilateral climate finance was provided by countries that have since cut, or plan to cut, their ODA budgets.⁶

This leaves climate finance vulnerable to cuts, or risks that other forms of development assistance could be cut to preserve climate finance.

Climate is also accounting for a growing share of aid. It made up one-fifth of rich countries’ bilateral aid in 2024, up from one-eighth in 2015, not counting aid to Ukraine or money spent inside donor countries. For France it was 37%, Germany 36%, and Japan 35%.

The mix of climate finance is shifting too. From 2021-2022 to 2023-2024, those three countries cut new adaptation commitments by 31% after inflation, while their mitigation commitments rose 37%, mostly because of Japanese lending in 2023.⁷

Part of the shift may reflect projects being labelled differently.

From 2021 to 2024, the World Bank paid 42% of the climate finance African countries received. The money came through its two main lending arms: the International Development Association, which funds low-income countries, and the International Bank for Reconstruction and Development, which lends to middle-income countries. The World Bank’s climate finance to African countries averaged US$ 5.4 billion a year, more than four times what France, the next-largest provider, paid.

The World Bank provided more than half of all climate finance in 16 of 54 African countries. On 29 June 2026, the Bank dropped its target to spend 45% of its financing on climate. IDA’s climate target covers only its current funding round, which runs to 2028.

Loans are an increasingly important part of adaptation finance for the least developed countries. They represented 40% of the adaptation finance paid to least developed countries in 2024, up from 34% in 2020. Most of the loans are concessional (79% of money lent in 2024, down from 94% in 2020)⁸, but the Paris Agreement asks donors to consider the need for public and grant-based adaptation finance for these countries.

There are two main cases for where climate finance should go: to the countries most exposed to climate change and least able to adapt, or to countries like Indonesia and South Africa that are big emitters, to help them cut emissions.

From 2021 to 2024, half of the climate finance paid to named countries went to 14 countries. India alone received 13.7%.

Half the people in developing countries living on less than US$ 3 a day live in eight countries, which receive just 25% of climate finance payments and 23% of adaptation finance.⁹ These countries are Nigeria (93 million people in extreme poverty), the Democratic Republic of Congo (87 million), India (74 million), Pakistan (51 million), Ethiopia (45 million), Tanzania (32 million), Yemen (27 million), and Mozambique (27 million). India alone receives more than half of that climate finance. The other seven, home to 43% of people living in extreme poverty, receive 11% of climate finance.

The COP28 Declaration on Climate, Relief, Recovery, and Peace called for more finance for fragile states. Yet, countries the OECD rates as highly or extremely fragile get little more than half as much climate finance as non-fragile developing countries of the same population.¹⁰

In fragile countries, one dollar in seven arrives via UN agencies and NGOs, compared with one in 33 elsewhere. Where donors suspended lending to governments after a coup or takeover, more of their climate finance went through NGOs.¹¹

Energy has moved up the climate agenda but the proportion of climate finance flowing to it has declined.

At COP28, countries pledged to triple renewable energy capacity by 2030, and in 2024 the World Bank and African Development Bank launched Mission 300 to connect 300 million Africans to electricity.

But energy’s share of new climate finance fell from 26% to 16% between 2015-2017 and 2022-2024, or from 29% to 21% counting policy loans coded to energy, which this analysis classifies as budget support.

Before inflation, climate commitments to named countries more than doubled, from US$ 33.0 billion a year in 2015-2017 to US$ 72.6 billion a year in 2022-2024, while climate finance for energy rose by only 34%, from US$ 8.6 billion to US$ 11.5 billion.

A single very large infrastructure project can outweigh a whole sector.

One high-speed railway in India accounted for more climate finance than all the projects donors coded as off-grid solar and mini-grids in developing countries combined. From 2021 to 2024, loans from the Japan International Cooperation Agency, known as JICA, paid US$ 3.8 billion in climate finance for the Mumbai-Ahmedabad line. Over the same period, all of the projects that donors coded as off-grid solar or mini-grids received US$ 1.8 billion.¹² Rail cuts emissions, so it counts as climate finance.

Climate finance paid for the Mumbai-Ahmedabad high-speed railway, 2021-2024

US$3.8bn

Climate finance paid for off-grid solar and mini-grids in developing countries, 2021-2024

US$1.8bn

In 2024, donors committed US$ 19.7 billion in aid marked for adaptation alone. For about half of that money, the project descriptions donors sent to the OECD made no mention of any adaptation aim.

ONE Data read donors’ own documents for the 40 largest of those projects, worth US$ 5.4 billion. Twenty-two of them, worth US$ 2.8 billion, describe no climate purpose at all. That’s at least 17% of the US$ 16.5 billion of adaptation-only aid that donors reported project by project for 2024.¹³

One is a US$ 636 million Korean-financed toll-road tunnel to Indonesia's new capital, which Korea’s own IATI record marks as having no adaptation aim.