Update: Forced Displacement Fell 5.4M to 117.8M — First Decade Decline, Funding Still Broken

What changed since the research print

Our migration–humanitarian research post closed on a brutal scissors: 123.2 million forcibly displaced at end-2024, 73% of refugees hosted in low- and middle-income countries, UNHCR needs only 37% funded in 2025, and OCHA’s Global Humanitarian Overview appeals 23.4% funded by October 2025. That piece asked who bears costs when people stocks rise and cash falls. UNHCR Global Trends 2025 (data as of 1 May 2026) and the GHO 2026 Mid-Year Review (FTS through 31 May 2026) answer the vintage question: what moved once the first decade-long stock decline landed — and did the funding ledger heal with it?

The stock answer is unambiguous. Forced displacement fell to 117.8 million at end-2025 — −5.4 million / −4% versus the research print — the first meaningful decline in a decade. Conflict IDPs (IDMC) dropped from 73.5M to 68.7M; refugees including UNRWA and other people needing international protection slipped from 42.7M to 41.6M (−3%). Asylum-seeker backlogs moved the other way: 8.4M → 9.0M. Returns of refugees and IDPs jumped +50% to 14.7 million; resettlement and sponsorship arrivals halved to 81,800 against a 2.9 million stated need. Host geography reshuffled: LMIC share of refugee hosting fell 73% → 68%, high-income rose 27% → 29%, and Least Developed Countries rose 23% → 26% (to 9.4 million people).

The cash answer is colder. UNHCR’s Global Report 2025 confirms the research print’s 37% funded year ($3.93B available against a $10.6B budget; 63% gap; expenditure −22% YoY). GHO appeal coverage inched from 23.4% to 24.4% — but requirements were cut from $45.4B to $33.7B (−26%). Coverage did not recover because donors filled the old ask; the ask was prioritized down while funded dollars also fell ($10.6B → $8.2B). Border narratives still treat migration as a rich-country arrival crisis. The vintage delta says the system’s people ledger eased slightly via often adverse returns, while the money ledger remained a structural gap.

The vintage scoreboard

Read the table as one sentence: the people stock finally fell, mostly because returns surged; the funding stock did not heal — it was resized.

The first decade decline is not a soft landing

A −4% drop after a decade of rises invites victory narratives. Global Trends is careful for a reason. Nearly 5.4 million people still fled across borders in 2025 — Sudan, Ukraine, Venezuela, South Sudan, Burkina Faso, Afghanistan, Mali, and Myanmar alone accounted for roughly six in ten of those movements. The stock fell because returns outpaced new displacement, not because conflict stopped generating refugees.

UNHCR flags that many returns occurred under adverse circumstances: insecurity, damaged infrastructure, and weak services in places of origin, plus shrinking protection space in some asylum countries. Afghan returns (~2.9M including ~1.9M refugees) were the largest refugee return corridor and were assessed as largely involuntary, driven by host-country policy shifts in Iran and Pakistan. Syrian returns (~1.3M refugees plus ~2M IDPs) followed the December 2024 political rupture but landed in a country still marked by destruction and fragile services. Sudanese refugee and IDP returns were large even as Sudan remained the world’s largest internal displacement crisis (9.1M IDPs at end-2025, down 2.4M but still enormous).

Composition matters for reading the decline. IDPs drove most of the absolute drop. Asylum stocks rose. That mix is the opposite of a “crisis solved at the border” story: the backlog of people waiting for protection decisions grew while the visible stock of already-displaced people shrank through returns of uneven quality.

Who still hosts — and how the ranking reshuffled

The research print’s host×donor scatter put Iran, Türkiye, Colombia, Germany, and Uganda in the people-heavy quadrant and the United States in the cash-heavy quadrant. End-2025 reorders the people side. Colombia (2.8M) and Germany (2.7M) lead; Türkiye falls to 2.4M (−19%) on Syrian returns; Uganda rises to 1.9M; Iran collapses to 1.7M (−53%) on Afghan returns and reclassification; Chad jumps to 1.5M on Sudanese arrivals; Pakistan slips to 1.3M (−17%).

That reshuffle is not burden-sharing by design. It is policy-driven stock compression in Iran/Pakistan/Türkiye plus neighbour absorption in Chad/Uganda. Low-income countries still host 18% of refugees while representing ~8% of world population and ~0.3% of world GDP. LDCs now host 26% — a +3pp rise and +12% in absolute refugee numbers versus end-2024. High-income countries’ share rose only to 29%. Our earlier refugee hosting burden map remains the right framing: geography of asylum is still neighbour geography, even after a year of record returns.

Funding: coverage flat, ask cut, flexibility worse

The research print already showed UNHCR’s 2025 cliff. Global Report 2025 locks it in: $3.932B available, 37% of needs, expenditure $3.830B (−22%), tightly earmarked funding rising from 24% to 44% of the portfolio. Agencies lost both volume and the ability to move money toward under-attended crises. That is the operational meaning of a “funding gap” beyond the headline percentage.

On the OCHA side, comparing GHO 2025 October to GHO 2026 mid-year is deliberately imperfect — different plan years, different FTS cut dates — but the direction is clear enough for a vintage read. Requirements fell ~26% to $33.7B while coverage only rose ~1pp to 24.4%. People in need in the mid-year frame sit near 252M (targeted 143M), below the research print’s 300M / 181M GHO 2025 framing, consistent with a stripped-back, prioritized appeal architecture rather than a funded recovery. Total humanitarian cash did not refill the old envelope; the envelope was redrawn smaller.

Donor politics still dominate rich-country debate. The OECD DAC aid drop remains the right companion: official development assistance and humanitarian envelopes are under fiscal and political pressure precisely when return sustainability and host-community services need multi-year finance. Private remittances cannot substitute for that public ledger — a point our demographic cash-flow series keeps returning to in a different corridor.

Returns without resettlement is not a durable solutions system

Durable solutions rhetoric usually lists three channels: voluntary return, local integration, and resettlement/complementary pathways. The 2025 vintage is lopsided. Returns hit 14.7M. Resettlement and sponsorship arrivals fell by more than half to 81,800, while UNHCR estimated 2.9M refugees needed resettlement. Arrivals cover roughly 3% of stated need. The United States’ sharp reduction in arrivals is the main arithmetic driver; the systemic effect is that the only large-scale “solution” operating is return — often into fragility.

That imbalance matters for the cost narrative. When high-income states cut resettlement while neighbours host and then push returns, the fiscal and political burden does not disappear. It is reassigned to origin countries with destroyed services and to remaining host communities that still carry protracted caseloads (70% of refugees in protracted situations; 24.9M people under the revised definition). The dashboard’s solutions panel is meant to make that asymmetry visible without treating every return as a protection success.

What the dashboard is for

Use the controls as a vintage desk tool, not a decoration. Meter scope switches the dumbbell set among people stocks, funding meters, and host-share meters. Sort ranks by largest absolute delta, newest level, or name. Region filters host-delta and return-corridor panels. The dual-axis stock×funded chart shows why a falling displacement stock does not imply a healing system: UNHCR’s funded share kept falling into 2025 even as the people stock finally turned. The income-mix scatter shows upper-middle-income hosting losing share (Iran effect) while high-income and LDC shares ticked up.

Caveats

• End-2025 Global Trends figures are official but still subject to revision as governments and partners update registration and estimates
• Return statistics mix voluntary, self-organised, and involuntary movements; large corridors (Afghanistan, Sudan) are hard to classify cleanly
• GHO 2025 Oct vs GHO 2026 May are different plan years and cut datescoverage deltas are directional, not identical instruments
• Income-group and LDC hosting shares exclude some unclassified cases (~3% of refugees in 2025)
• Host country rankings use refugees + other people needing international protection; IDP stocks are a separate ledger
• Reclassifications (especially Afghans in Iran/Pakistan moving between refugee-like and OPNIIP categories) affect levels without always reflecting new physical movements
• Resettlement “need” is a UNHCR planning estimate, not a binding quota

Methodology

Prior research vintage mirrors meters from migration-humanitarian-research-2026: end-2024 Global Trends stocks, GHO 2025 October FTS coverage, and Global Report 2024/2025 funding ratios. Newest stock and host geography follow UNHCR Global Trends 2025 (end-2025; data received as of 1 May 2026), including the first decade decline to 117.8M, host ranking (Colombia, Germany, Türkiye, Uganda, Iran, Chad, Pakistan), income-group and LDC shares, return totals (14.7M), refugee returns (4.4M), and resettlement arrivals (81,800). UNHCR funding uses Global Report 2025 final budget ($10.604B), funds available ($3.932B), funded share (37%), and expenditure change (−22%). GHO 2026 mid-year meters use the Mid-Year Review snapshot as of 31 May 2026: requirements $33.66B, funding $8.21B, coverage 24.4%, people in need 252.1M. Dashboard deltas are arithmetic comparisons of those disclosed prints; mid-path displacement years before 2024 remain labeled estimated where carried forward from the research series.

Unlike a border-apprehension dashboard, this update measures global forced-displacement stocks, host shares, return flows, and humanitarian appeal coverage — the cost ledger outside rich-country arrival headlines.