Four Years Out: What the UN's SDG Report 2026 Says About a Scorecard Running Out of Money

Research Desk is Exereco's strand for analysing studies from public-interest bodies rather than companies. This week's subject is the United Nations' flagship progress check on the world's shared development agenda, The Sustainable Development Goals Report 2026, produced by the UN Department of Economic and Social Affairs and published in July 2026. It is a long report that covers every one of the 17 goals. We have not tried to summarise all of it. Instead we go deep on one thread that runs through its opening chapters and that, in our view, deserves more attention than it usually gets: the scorecard of how the 139 best-measured targets are performing, and the squeeze on the money and the data that make that scorecard possible.

The report opens, as UN reports tend to, with a foreword from the Secretary-General, António Guterres, who begins by looking back to the moment, a little over a decade ago, when the world embarked on a bold effort to fulfil the 2030 Agenda for Sustainable Development. The introduction that follows is balanced in a way that is worth noticing. It lists real gains and it lists real failures, and it does not pretend that one cancels the other. Reading the two lists side by side is the quickest way to understand where the agenda stands with four years to go.

On the gains side, the report says electricity now reaches 92% of the world's population, with a third of that supply powered by renewable sources. Internet access has risen from 40% to 74%. New HIV infections and AIDS-related deaths have each fallen by roughly a third. Fifty-nine countries have eliminated at least one neglected tropical disease, and 134 countries have met the target for reducing child mortality. Social protection now covers over half of the global population for the first time, and disaster-related deaths have fallen by 65% compared with the previous decade. These are not small numbers. They describe the kind of change in ordinary lives that the goals were designed to produce, and the report is right to put them first.

The other list is harder to read. According to the report, one in ten people still lives in extreme poverty, and food insecurity affects 2.3 billion people. About 2.1 billion people, roughly one in four, lack safely managed drinking water. More than 150 million children are stunted. Maternal mortality is nearly three times the global target, and everyday discrimination is still felt by one in five people. In 2025, global temperatures reached 1.43°C above preindustrial levels. Violent conflict has forcibly displaced 118 million people. And, in the sentence that frames the rest of this article, official development assistance collapsed by a record 23% in 2025, while the annual SDG financing gap in developing countries is around $4 trillion.

The report card on 139 targets

The most useful single piece of analysis in the introduction is the scorecard. The report looks at the 139 targets for which there are reliable historical trends, which is a subset of the full framework of 17 goals, 169 targets and more than 230 indicators, and sorts them by how they are moving. The headline is that 36% are on track or making moderate progress. Forty-nine percent are moving too slowly. And 15% have slipped below their 2015 baselines, meaning the world is further from those targets today than it was when the goals were adopted.

The chapter on monitoring breaks the picture down more finely, and the finer breakdown is where the story sits. Fifteen percent of the 139 targets are on track. A further 21% show moderate progress. Thirty-two percent are advancing only marginally, 17% are stagnating, and the final 15% are regressing. The report notes that the marginal and stagnating categories together make up the "nearly half" that the introduction describes as moving too slowly.

It is worth dwelling on what these categories mean, because it is easy to read past them. On track is the best outcome, and it describes only about one target in seven. Moderate progress is better than nothing but, by the report's own framing, not enough to say a target will be met. Marginal progress and stagnation are different from regression, but in practical terms they describe the same experience for the people concerned: a target that is not moving fast enough, or at all, towards the finish line. Together those two categories account for 49% of the tracked targets, which makes slow or flat progress the single most common outcome in the whole scorecard. More targets are stuck than are moving, and nearly as many are going backwards as are on track.

There is a second point buried in the arithmetic. The scorecard covers 139 targets with sufficient trend data, not all 169. The report is explicit that some areas have trend data covering more than 80% of their indicators, while fewer than a third of indicators have sufficient trend data for gender equality, sustainable cities, climate action, and peace and justice. In other words, the areas where we know least are not a random selection. They include several of the most politically and environmentally urgent goals. A scorecard can only grade the targets it can see, and the targets it cannot see well are not necessarily the ones doing better.

A decade of building the measuring stick

To appreciate what is at stake in the data, it helps to see how far the measuring system has come. The report describes a decade of monitoring in which the SDG indicator database grew from 115 indicators in 2016 to 233 in 2026. The number of data records rose from about 330,000 to 3.2 million. The introduction frames this as a change from a world in which data were lacking for half of the indicators to one with a global database of 3 million data points covering nearly every indicator.

The quality of the foundations improved as well as the quantity. Nearly 40% of indicators initially lacked internationally agreed methodologies. Today every indicator has one. The share of indicators with good country coverage rose from roughly a third to almost 70%. Statistical performance indicator scores rose from 59 to 70 between 2016 and 2023. These are the unglamorous achievements that rarely make headlines, and they matter, because a target without an agreed method of measurement is a target that cannot be held to account.

This is the part of the story that makes the scorecard credible. When the report says that 15% of targets are regressing, it is not offering an impression. It is reporting the output of a system that took ten years and a great deal of public money to build. The gains on the scorecard and the gains in the data are connected: it is only because the measuring stick got better that we can now say, with some confidence, which targets are in trouble.

The money is moving the wrong way

The same chapter then turns to financing, and the numbers here are stark. According to the report, official development assistance fell by 23.1% between 2024 and 2025. That is the largest annual drop on record, and it returned aid to the levels seen at the start of the 2030 Agenda. The report adds that further cuts are expected in 2026. Set against an annual SDG financing gap in developing countries of around $4 trillion, a fall of that size in one year is not a rounding error.

Aid is not only a transfer of resources for schools, clinics and power lines. It is also, in many low-income countries, part of the financing for the unglamorous machinery of government statistics. The report draws this connection directly. Surveys by the Inter-Secretariat Working Group on Household Surveys, covering low- and middle-income countries, show funding cuts affecting SDG monitoring, household surveys, administrative data systems and dissemination platforms. Put simply, the same cuts that make development harder also make it harder to see whether development is happening.

This is the squeeze at the heart of the report's argument. Development finance is falling just as the scorecard shows that the agenda needs more effort, not less, and the instruments used to assess whether the effort is working are being starved by the same decision. A reader could be forgiven for expecting a fall in aid to show up first in the outcome numbers, such as poverty, hunger or child mortality. Our reading of the report is that it may also show up in a less visible way, in the data systems that would have reported those outcomes. A household survey that is not conducted this year does not produce a bad number. It produces no number at all.

Who owns the data

The report's treatment of data sovereignty is where the analysis becomes most specific, and it deserves a careful reading. It notes that, in 2025, 135 countries had national statistical plans under implementation, but only 59% of those plans were fully funded. In sub-Saharan Africa, only 15% of countries reported full funding. The report also cites a 2023 analysis finding that only about a third of low- and lower-middle-income countries manage their own microdata repositories.

Those two figures matter for different reasons. A statistical plan that is not fully funded is a plan that will be trimmed, delayed or dropped when budgets tighten, and the report suggests budgets are tightening. A country that does not manage its own microdata repository is a country that depends on someone else to store and, in practice, to control access to its own records. That is the sense in which the report uses the word sovereignty: not as a political slogan, but as a description of who holds the data, who pays for it, and who decides what it is used for.

The coordination figures reinforce the point. In the 2021 survey on the Cape Town Global Action Plan, only 17% of countries said they were satisfied with their capacity to coordinate the data ecosystem. In low- and lower-middle-income countries, the figure was 8%. In 2024, high community participation in rural drinking-water policy was reported in only 30% of cases. And only 29% of responding countries reported statistical programmes for internally displaced persons, against more than 70% for persons with disabilities and for youth. The people who are hardest to count, such as those displaced by the 118 million-strong wave of conflict the report describes, are the least likely to have a programme designed to count them.

There is a quiet strength in the statistical profession's own legal framework that the report also records. A 2023 review found that 93% of the 167 national statistical laws include a confidentiality clause. It is a reminder that official statistics are built on trust between citizens and the state, and that trust is a form of infrastructure that is cheap to maintain and expensive to rebuild.

Where artificial intelligence fits, and where it does not

The report's chapter is titled "SDG monitoring in the AI era", and it would be a mistake to skip over the technology question, because it is the obvious response to a budget squeeze. If surveys are expensive and funds are short, can machines fill the gap? The report's answer is measured. A meta-analysis of about 1,800 articles confirms that artificial intelligence is already applied across the full survey life cycle. The tools are real and they are in use.

But the report does not treat AI as a substitute for funded national systems. Its way forward sets out three priorities: people-centred data systems, AI held to statistical transparency standards, and institutional and financial sovereignty for national statistical offices. The report's closing passage on the subject ends on the point that answers matter even when the system delivering them is no longer the national statistical office itself. We read that as a way of naming the risk. If the answers about development increasingly come from systems that national institutions do not run, fund or fully understand, then the scorecard becomes harder to challenge, and the countries being scored have less say in how they are scored.

None of this is an argument against using new tools. It is an argument that tools need institutions around them, and that institutions need money. The report's emphasis on funding national statistical plans and keeping the offices that run them at the centre of the system is consistent with that reading.

What the numbers add up to

Taken together, the report's figures describe a loop. The scorecard shows that most of the best-measured targets are not moving fast enough, and that a meaningful minority are going backwards. That should argue for more resources and better targeting. Instead, aid has fallen by a record amount, with more cuts expected, and the cuts are already reaching the household surveys, administrative systems and dissemination platforms that tell governments and donors where to target. The poorest countries, where plans are least likely to be fully funded and where data repositories are least likely to be nationally controlled, are the most exposed.

It is also worth being clear about what the report does not say. It does not say the goals have failed, and it does not say the progress of the past decade was an illusion. The gains in electricity, connectivity, health and disaster mortality are presented as real. What it says is that the pace of the whole is too slow, that the funding environment has worsened sharply, and that the tools used to observe the whole are themselves at risk. That is a more specific and, in some ways, more useful message than a general call for ambition.

WHY IT MATTERS

You cannot manage what you cannot measure, and the SDG Report 2026 is a reminder that measurement is not free. Every one of the figures in this article exists because someone funded a survey, trained a statistician, agreed a method and kept the records. When aid falls by a record 23.1% in a single year and the report warns of further cuts, the risk is not only that fewer clinics and classrooms get built. It is that we lose sight of where the agenda is failing, at the exact moment the scorecard says that 15% of targets are going backwards and only 36% are on track or moving moderately.

For anyone working in sustainability, ESG, public policy or development finance, that should change how public data is treated. Public statistics are often handled as a free background resource, the stable ground on which reports, investment screens and policy models are built. This report suggests the ground is shifting. With 135 countries running statistical plans and only 59% of them fully funded, and only 15% of sub-Saharan African countries reporting full funding, the baseline that others lean on is thinner than it looks. Protecting the data is not a technical side issue to the goals. It is part of delivering them, because it is what lets us tell the difference between a target that is on track and one that has quietly stalled. With four years left, the measuring stick needs defending as much as the targets themselves.