Governments reaffirmed a goal of 1 million resettlement places by 2030 in December 2023 — UNHCR states that resettlement is not a right and that receiving countries offer limited quotas

What is happening?

Resettlement is the transfer of a refugee from the country where protection was first sought to a third state that has agreed to admit that person. UNHCR sets out the terms of it on its own help pages: resettlement is not a right, the countries that resettle offer limited quotas, and the receiving government makes the final decision on every case.

The distance between what that system is asked to do and what it offers is measured every year. In a report published on 2025-06-24, UNHCR projected that about 2.5 million refugees would need resettlement during 2026, against an annual target of 120,000 places for the same year. In a further projection published on 2026-06-16 it put the need for 2027 at about 2.4 million, against a target of about 130,000 places.

The supply side moved the other way. Worldwide resettlement departures exceeded 116,000 in 2024 and then fell to about 37,000 in 2025. UNHCR attributed that fall to policy changes in receiving countries that halted admissions, to stricter criteria, and to processing delays. No source opened here apportions the fall among receiving countries.

One of those policy changes is documented in detail, because the state that made it published the instrument. Executive Order 14163, effective 2025-01-27 at 00:01 eastern time, suspended new entries under the United States Refugee Admissions Program indefinitely, provided for reports every 90 days, and left resumption to a presidential determination that admissions are in the interest of the country. The order gives as its reason that the country lacks the ability to absorb large numbers of migrants, and refugees in particular, into its communities.

The stated ambition has not been withdrawn. At the Global Refugee Forum in December 2023, 168 governments took part and reaffirmed a goal of 1 million resettlement places by 2030, and the European Union announced a further pledge of 61,000 places for 2024 and 2025.

Whose problem is this?

The body that measures the need is not the body that supplies the places, and nothing in the arrangement gives the first any claim on the second.

Where does this problem end?

Four questions sit next to this one and are excluded on purpose. Whether states should admit more refugees at all is a question of national immigration policy, and nothing here takes a position on it. Whether asylum procedures in first countries of refuge are adequate is a separate failure with separate remedies. Whether resettlement or local integration is the better durable solution is contested among the bodies that run both, and this document does not settle it. Whether a state keeps sovereign control over who enters its territory is not in dispute here, because the system the sources describe already rests on that control.

The boundary matters because the number of places on offer and the willingness of any one state to raise it are different questions with different remedies.

What is the state now, and what should it be?

Now

What it should be

The institutions that run this system have set their own targets and those targets are on the record. At the Global Refugee Forum in December 2023, governments reaffirmed a goal of 1 million resettlement places by 2030. The earlier Three-Year Strategy, launched in 2019, set a goal of 3 million people protected by 2028, of whom 1 million through resettlement. The annual targets of 120,000 places for 2026 and about 130,000 for 2027 are the shorter-horizon expression of the same commitment.

What no source opened here supplies is a target for the distance itself. Nothing opened here states how much of the assessed need the system means to cover in a given year, and the annual targets are set well below it.

How big is it?

About 2.5 million people for 2026 and about 2.4 million for 2027. Those are the two most recent published projections of resettlement need, the first from a report published on 2025-06-24 and the second from a projection published on 2026-06-16. They are two consecutive single-year assessments rather than two bounds on one figure, and the population file for this document carries them as a range on that understanding and carries nothing else, because they are the only counts of people, as distinct from counts of places, that the sources opened here supply.

The place counts sit beside that range rather than inside it. The targets were 120,000 places for 2026 and about 130,000 for 2027, and actual departures were more than 116,000 in 2024 and about 37,000 in 2025. This document does not subtract them from the need figure, because the need projections and the departure counts cover different periods and no source opened here performs that subtraction.

The fiscal-year admissions figures for a single receiving state are kept out of this count entirely. They run on a fiscal year rather than a calendar year and they describe one national program, so folding them into a worldwide figure would produce a number that no source supports.

Under what conditions does it arise?

1. The need is assessed centrally and the places are decided one state at a time. UNHCR publishes a projection of how many refugees will need a third country in the coming year, and each receiving government then decides, on its own timetable, how many places it will open. A projection is an input to that decision and not a constraint on it.

2. Resettlement carries no matching duty. On its own help pages UNHCR states that resettlement is not a right, that the countries that resettle offer limited quotas, and that the receiving government makes the final decision. A refugee who meets every criterion can be assessed, submitted and still not travel, with no rule anywhere having been broken.

3. A national program can be halted by a single instrument. Executive Order 14163 suspended new entries under the United States Refugee Admissions Program with effect from 2025-01-27, leaving resumption to a determination made after reports every 90 days. Over the same period worldwide departures fell from more than 116,000 in 2024 to about 37,000 in 2025, a fall UNHCR attributed to policy changes in receiving countries, stricter criteria and processing delays. No source opened here apportions that fall among states.

4. A pledge and a place are counted by different bodies at different times. Pledges are recorded at intervals at the Global Refugee Forum, and whether they are filled is established later. An international refugee aid organization reported that about half of the 16,000 places European Union member states pledged for 2022 were filled. The European Union page opened here announces later pledges and does not address that implementation rate, and no official confirmation of the figure was opened here.

5. Two different things are counted under one goal. The multilateral goal counts resettlement together with complementary pathways, meaning labor, education and family reunification routes opened to refugees. Annual reporting opened here does not consistently keep the two apart, so a figure described as progress can belong to either.

What has been tried?

Two directions have been run at once. One is to state a goal and gather pledges against it, which has produced targets counted in millions and pledges counted in tens of thousands. The other is the annual quota decision inside each receiving state, which is where the places actually come from, and that one moved down between 2024 and 2025.

What was found?

Why is it still unsolved?

Supply shortage — the places are decided one state at a time, and nothing obliges any state to offer them.

The measurement and the supply belong to different bodies. UNHCR counts how many refugees will need a third country in the coming year and publishes the figure in advance. Each receiving government then decides how many places it will open, when it will open them, and which cases it will accept. A projection is an input to that decision and not a constraint on it, which is why the two figures can sit on the same page for years without moving toward each other.

The second part is that the annual target is not built to meet the assessed need. For 2026 the need was projected at about 2.5 million and the target at 120,000 places; for 2027 the figures were about 2.4 million and about 130,000. A year in which every place is filled still leaves most of the assessed need where it was. Meeting the target and closing the distance are different achievements, and only the first of them is measured.

The third part is that the supply can fall faster than any process can answer. Departures exceeded 116,000 in 2024 and were about 37,000 in 2025, and UNHCR gave the reasons as policy changes in receiving countries that halted admissions, stricter criteria and processing delays. None of those requires an amendment to any instrument, because no instrument fixes a floor. A quota that was never owed can be withdrawn without anything being broken.

The fourth part is that the failure produces no event. No deadline passes, no body falls into default, and the annual figure simply arrives lower than the one before it. The goal of 1 million places by 2030 stays on the record and is reaffirmed at intervals, and a goal that nobody is obliged to reach is not missed at any particular moment.

What observation would mean it is solved?

Candidates — (a) worldwide resettlement departures rise toward the need assessed for the same year (b) the annual target is itself set at a level that would meet the assessed need (c) the places stop depending on a voluntary quota, through an allocation rule or an obligation that binds receiving states.

(a) alone can move for the wrong reason. Departures depend on how many cases are submitted, how quickly receiving states process them, and how the need is assessed in the first place. A fall in the assessed need would narrow the same distance without one more person traveling, and the method behind that assessment was not opened here.

(b) alone counts an intention. A target is a number that governments announce, and the record already holds a larger one: 1 million places by 2030, reaffirmed in December 2023. Departures moved down between 2024 and 2025 while that goal stood unchanged.

(c) alone is the hardest to observe and the easiest to imitate. An allocation rule can be written with reservations wide enough that every state keeps the discretion it already had. What would separate a binding rule from the present arrangement is whether a state that offers nothing is in default of anything, and no source opened here describes a mechanism under which one would be.

The three have to be read together, and the first has to be read against the assessment it is measured against.

What is it connected to?

Fills with researchcomplementary pathways and how far they are filled, first-country asylum capacity and the hosting burden it carries, the financing of refugee protection, statelessness as a separate legal status, and the domestic politics of admissions inside receiving states. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• How the need figure is built. The UNHCR projection report is the origin of the 2.5 million and 2.4 million figures and of both annual targets, and it could not be opened here: the host returned an access refusal to every automated request. Nothing opened here explains how a refugee comes to be counted as needing resettlement.
• Which receiving states cut their quotas and by how much. Opened sources describe the fall in departures and give the reasons in general terms only. One receiving state is documented in detail here because that state published the instrument and a legislative research product about it. The others are broken out nowhere opened here.
• Whether need and places are counted on the same basis. One opened source reports about 2.5 million and 120,000 places as applying to the same year. Neither that source nor any other opened here states whether the two quantities are built the same way.
• What the Three-Year Strategy achieved. The final report and the summary describing it both returned access refusals. What is recorded here is the goal and the starting diagnosis, taken from a search summary rather than from an opened page, and no outcome figure was available.
• Whether the 2022 European Union pledge was half filled. That figure rests on one non-governmental account. The European Union page opened here announces later pledges and is silent on the earlier implementation rate, and no official confirmation was opened.
• How the multilateral goal splits between resettlement and complementary pathways. The goal is stated as 3 million protected with 1 million through resettlement, and annual reporting opened here does not consistently keep the two apart.
• What the money did. One opened source records EUR 246 million pledged in support at the December 2023 forum. Nothing opened here says what it funded or how it converts into a place.
• A response from the bodies named here beyond what they publish themselves. UNHCR publishes the assessment, the departure figures and the statement of what the system is, and that is the full extent of its position opened here. For receiving governments, the only statements opened here are the text of the executive order and a legislative research product; no other receiving government was reached.
• The 2026 outcome. No source opened here gives a worldwide departure figure for 2026, so the most recent full year available is 2025.

See the evidence

Two documents that speak for the institutions themselves were opened directly, and they carry the two halves of the claim made here. The UNHCR help pages state what the system is, and the text of Executive Order 14163 shows what a receiving state can do inside it without breaching anything. The official European Commission page supplies the December 2023 goal, the count of participating governments and the European Union pledge, and the legislative research product supplies the fiscal-year admissions figures for one state, which are kept separate throughout because they run on a different calendar and cover one country. Everything quantitative about the worldwide picture comes from two outlets reporting a UNHCR report series that could not be opened at all, which is why the need projections, the annual targets and the departure counts are graded medium rather than high. Those two outlets report adjacent years and so share no single figure to cross-check, but the shape they describe is the same: a need counted in millions, a target counted in the low hundreds of thousands, and departures that fell between 2024 and 2025. Four pages closer to the source could not be opened, two of them published by the bodies that run the system and one by a receiving government, and the reason sits in the rows above instead of being papered over with a substitute address. This is a Path A output, a definition assembled from published documents rather than from a registry feed, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range 2,400,000–2,500,000 As of two consecutive single-year assessments, not an interval on one year: 2.4 million for 2027 published 2026-06-16, and 2.5 million for 2026 published 2025-06-24

Sensitivity The two ends of this range are two consecutive single-year assessments, 2027 at the low end and 2026 at the high end, and not two bounds on one measurement. The width is therefore the distance between two annual projections and is not a measure of uncertainty in either one. Neither projection was opened here; both are taken from outlets citing the UNHCR report series, and the method by which a refugee comes to be counted as needing resettlement was not available anywhere opened here, so the real uncertainty is larger than the narrow span between 2.4 and 2.5 million suggests. Whether the annual projections double count people who wait across more than one year cannot be determined here, for the same reason: the report setting out the method returned an access refusal. This counts people assessed as needing a third country in one year. It does not count refugees who were never assessed, and it is not net of anyone resettled: worldwide departures were more than 116,000 in 2024 and about 37,000 in 2025, against annual targets of 120,000 places for 2026 and about 130,000 for 2027. Those place counts are deliberately not subtracted here, because the need projections and the departure counts cover different periods and no source opened here performs that subtraction. In the other direction, the two projections fall close together and a single revision of the assessment method could move both at once. Fiscal-year admissions figures for one receiving state are excluded entirely, since folding a national fiscal-year count into a worldwide figure would produce a number that no source supports.

Regional breakdown No source opened here breaks the projected resettlement need or the places offered down by region, by country of first asylum, or by receiving country. The only country-level detail opened here covers one receiving state and is compiled on a fiscal year, so it cannot be set alongside the worldwide figures without changing what is being counted.