Small-boat arrivals: recorded spending and a labelled scenario
Two views that must not be confused. The first is what has been spent, as published. The second is a calculation from stated assumptions, and it says so on every line.
Companion to the daily arrivals ledger. Labels on this page: observed a published count; official estimate a published figure that is itself an estimate; derived measure computed by this site from published inputs by a stated method; assumption chosen by this site because nothing published settles it; scenario an output of the calculator; unavailable no defensible basis exists.
1. Recorded spending
These are system-wide totals for people in the asylum system, not the marginal cost of a newly arrived person, and not a cost of small-boat arrivals specifically: the asylum population includes people who arrived by every route. The containment column exists so that no two rows are added together.
| Programme | Period and basis | Amount | Coverage and overlaps |
|---|---|---|---|
| Home Office: Asylum accommodation and support (excluding UASC) official estimate | 2024-25 (financial year); paid | £3.40bn | Accommodation, cash support and dispersal grants for people with asylum claims; excludes unaccompanied asylum-seeking children. contains: Direct spending on asylum accommodation; Cash support to people seeking asylum; Grants to local authorities for dispersal accommodation. inside: Asylum support expenditure (all); Total asylum-system spending NAO, An analysis of the asylum system, December 2025, Figure 5 note 2 (actual costs, rounded) |
| Home Office: Direct spending on asylum accommodation official estimate | 2024-25 (financial year); paid | £2.70bn | Payments to accommodation providers (hotels, dispersal, initial and large sites). inside: Asylum accommodation and support (excluding UASC); Asylum support expenditure (all); Total asylum-system spending NAO, December 2025, Figure 5 note 2 |
| Home Office: Cash support to people seeking asylum official estimate | 2024-25 (financial year); paid | £220m | Weekly allowances (section 95, section 4). inside: Asylum accommodation and support (excluding UASC); Asylum support expenditure (all); Total asylum-system spending NAO, December 2025, Figure 5 note 2 |
| Home Office: Grants to local authorities for dispersal accommodation official estimate | 2024-25 (financial year); paid | £247m | Grant funding to councils hosting dispersal accommodation. inside: Asylum accommodation and support (excluding UASC); Asylum support expenditure (all); Total asylum-system spending NAO, December 2025, Figure 5 note 2 |
| Home Office: Asylum support expenditure (all) official estimate | 2024-25 (financial year); paid | £4.00bn | Asylum support including UASC grants (around 636 million pounds) and accommodation transformation. contains: Asylum accommodation and support (excluding UASC). inside: Total asylum-system spending NAO, December 2025, summary (around 4.0 billion pounds on asylum support) |
| Home Office and Ministry of Justice: Total asylum-system spending official estimate | 2024-25 (financial year); estimated | £4.90bn | NAO estimate from actual costs: support, casework (around 367 million), Irregular Migration Intake Unit (nearly 75 million), MoJ tribunals and legal aid. The NAO says it does not fully capture local-authority costs. contains: Asylum support expenditure (all) NAO, December 2025, paragraph 1.14 and Figure 5 |
| All departments (FCDO reporting): ODA in-donor refugee costs official estimate | 2025 (calendar year); accrued | £2.40bn | OECD DAC accounting classification of eligible first-12-month costs of supporting refugees and asylum seekers in the UK, across departments. Provisional; final figures due autumn 2026. Overlaps the Home Office figures above on a different period and basis. overlaps (different period and basis): Asylum accommodation and support (excluding UASC); Asylum support expenditure (all); Total asylum-system spending FCDO, Statistics on International Development: provisional UK ODA spend 2025, 9 April 2026 (2,395 million pounds; 2,827 million in 2024) |
| All departments (FCDO reporting): ODA in-donor refugee costs official estimate | 2024 (calendar year); accrued | £2.83bn | As record 7, prior year, final. FCDO, 9 April 2026 |
Calculations this page will not make. Arrivals multiplied by a whole-system average cost, presented as the cost caused by arrivals. Cash support plus hotel costs plus a total that already contains both. The ODA in-donor figure added to the Home Office figures. The NAO says its system total does not fully capture local-authority costs; nothing here fills that gap by apportionment.
2. How it is funded
| Appropriation | Programme and recipient | Category, period, basis | Amount | Note |
|---|---|---|---|---|
| Parliamentary Supply Estimate, Home Office Home Office | Asylum accommodation and support to: Accommodation providers (asylum accommodation and support contracts) | Accommodation; 2024-25; paid | £2.70bn | NAO Figure 5 note 2. Provider-level split not stated in the document; not invented. source |
| Parliamentary Supply Estimate, Home Office Home Office | Asylum support to: People seeking asylum (section 95 / section 4 allowances) | Cash support; 2024-25; paid | £220m | Weekly rate as at December 2025: 49.18 pounds self-catered, 9.95 catered (NAO paragraph 1.4). source |
| Parliamentary Supply Estimate, Home Office Home Office | Dispersal accommodation grants to: Local authorities | Grants to local authorities; 2024-25; paid | £247m | NAO Figure 5 note 2. Per-council amounts are a separate (council-level) document; not apportioned here. source |
| Parliamentary Supply Estimate, Home Office Home Office | Unaccompanied asylum-seeking children to: Local authorities | Grants and incentive payments; 2024-25; paid | £636m | NAO Figure 5 note 3. Outside the 3.4 billion accommodation-and-support figure; inside the 4.0 billion. source |
| ODA budget (cross-government, reported by FCDO) Home Office and others | In-donor refugee costs to: Not a payment: an accounting classification of eligible spend already in departmental budgets | Classification; 2025; accrued | £2.40bn | Provisional. Never added to the Home Office rows. source |
3. Related populations, kept separate
| Measure | Latest value | Period | Unit |
|---|---|---|---|
| Detected small-boat arrivals (people), calendar year to date observed | 20,935 | 1 Jan to 20 October 2026 | a flow of arrival records |
| Other detected illegal-route arrivals (air, UK detections, ports) observed | 4,712 | Year ending June 2026 (as published) | recorded detections |
| Asylum seekers in receipt of Home Office support (all types) observed | 93,293 | Jun 2026 (as published) | people at a date |
| Of which accommodated in hotels observed | 16,021 | Year ending June 2026 (as published) | people at a date |
| People in immigration detention observed | 2,152 | 30 Jun 2026 (as published) | people at a date |
Each row is its own dataset with its own definition. The supported caseload is not the number who arrived by boat; the detention count is not a count of offences; other-route detections are published yearly and have no daily values.
4. Scenario calculator scenario
This is not a forecast. It is the arithmetic of a stated set of assumptions, run forward from 30 June 2026, the latest date on which the supported caseload is observed. Arrivals for months the ledger already covers are actual; later months continue past seasonality at a chosen level. The calculator never lets an arrival stay in support indefinitely, and it carries the people already in support at the origin, because a projection of only new arrivals would omit the existing demand.
In the central case the whole supported caseload is 86,674 at 12 months and 85,367 at 60 months, of which the cohort that entered from small-boat arrivals after the origin is 17,129 and 32,360. Modelled support spending is £1.77bn in the first year (£198m for that cohort) and £8.70bn cumulatively over five years (£2.56bn for the cohort). Those numbers move with every row of the assumption table, which is why the table comes first and the cases are shown side by side.
The arithmetic
modelled_arrivals[t] = actual (where the ledger has the month) else level x seasonal_share[month] small_boat_intake[t] = modelled_arrivals[t] x share_entering_support other_intake[t] = other_route_claims_per_year / 12 x share_entering_support exits[t] = caseload[t-1] x monthly_exit_hazard (applied to every cohort) supported_caseload[t] = caseload[t-1] + small_boat_intake[t] + other_intake[t] - exits[t] modelled_support_spend[t] = person_days[t] x (hotel_share x hotel_cost + (1 - hotel_share) x dispersal_cost + cash_per_day) small-boat cohort = the same recurrence run for post-origin small-boat intake alone
| Parameter | Value | Source or method |
|---|---|---|
| Origin date observed | 30 June 2026 | The scenario origin is the latest date at which the supported caseload is observed (quarterly 'as at' date). Months between the origin and the ledger's data-current-to date use actual arrivals from the ledger; later months use the seasonal assumption. |
| Baseline supported caseload at origin observed | 93,293 people | Home Office, Asylum summary tables YE June 2026, Asy_09b, as at 30 June 2026 |
| Share of supported caseload in hotels (held constant) observed | 17.2% | Asy_00a 'Of which accommodated in hotels' as at 30 June 2026 (16,021) over total supported (93,293). Holding it constant is an assumption; the share has been falling. |
| Hotel accommodation, per person per night official estimate | £134 (range £127 to £148), 2023-24 prices | NAO, Investigation into asylum accommodation, March 2024, paragraph 2.5: hotel cost expected to fall from 148 to 134 pounds per person per night including VAT; range 127 to 148 between April 2023 and January 2024 |
| Dispersal and other accommodation, per person per night derived measure | £31.00 | (direct accommodation spend 2024-25 minus hotel person-days times hotel unit cost) divided by non-hotel supported person-days, using the as-at December 2024 caseload (112,187 supported, 38,054 in hotels) as the year's level. An implied average across dispersal, initial accommodation and large sites, at a 2023-24 hotel rate. Not a contract price. |
| Cash support, per person per day (weighted by hotel share) official estimate | £6.06 | NAO December 2025 paragraph 1.4: 49.18 pounds a week self-catered (7.03 a day); 9.95 a week catered |
| Share of small-boat arrivals claiming asylum observed | 98.4% | IER_02d: of 40,595 small-boat arrivals in 2025, 32,023 claimed asylum only and 7,929 claimed asylum with an NRM referral (98.4%) |
| Share of arrivals entering supported accommodation assumption | lower 70%, central 85%, higher 95% | Not published. Support is available to claimants who would otherwise be destitute (NAO paragraph 1.4); the take-up among small-boat claimants specifically is not a published statistic. Shown as a range and labelled an assumption. What would settle it: A Home Office statistic or FOI response giving the share of small-boat asylum claimants in receipt of section 95/98 support within three months of arrival. |
| Other-route asylum claims entering support, per year observed | 51,942 claims a year, times the share entering support | Asylum summary tables YE June 2026: 85,891 people claimed asylum in the year ending June 2026 (Asy_00a), of whom 33,949 had arrived by small boat (Asy_01c); the remaining 51,942 claimed after arriving by other routes. Held flat at this level, divided by 12, and multiplied by the same share-entering-support as small-boat claimants. The supported caseload is fed by claims from every route. Modelling small-boat intake alone cannot reproduce it, which the first back-test showed (errors of 22 to 73 per cent); this flow is what closes the gap. |
| Monthly exit hazard from support derived measure | lower-cost case 10.0%, central 7.0% (fitted), higher-cost case 4.0% a month | Calibrated: the monthly hazard at which the central case, started from the observed supported caseload on 30 June 2025 (106,075) and fed the arrivals the ledger actually recorded plus the other-route claims level, lands exactly on the observed caseload on 30 June 2026 (93,293). The lower-cost and higher-cost cases move the fitted hazard by the spread below. Because the central hazard is fitted, the central back-test closes by construction; the back-test is a consistency check of the other cases and of the share assumption, not an independent validation. Case labels follow their effect on cost: 'higher' means slower exits. A single hazard is applied to everyone in support regardless of route or stage; decisions, grants, refusals, returns and move-on periods are not modelled separately because the published tables do not give support exits by reason. |
| Arrivals level derived measure | trailing 12 months at origin: 33,374; cases 0.75x, 1.0x, 1.25x; actual ledger months used first: 2026-07, 2026-08, 2026-09 | Trailing 12 months of detected arrivals from the ledger at the origin, scaled by case: lower 0.75, central 1.0, higher 1.25. Monthly shape is each calendar month's share of the trailing 36 months of arrivals (historical seasonality continuation, not a prediction). |
| Seasonal shape (share of a year's arrivals by month) derived measure | Jan 3%, Feb 3%, Mar 9%, Apr 8%, May 9%, Jun 10%, Jul 11%, Aug 12%, Sep 13%, Oct 9%, Nov 6%, Dec 6% | Trailing 36 months of the ledger at origin. A continuation of past seasonality, not a prediction of it. |
| Inflation unavailable | none applied (nominal at stated vintages) | Nominal, at stated vintages. Unit costs are 2023-24 (hotel) and 2024-25 (derived dispersal); no uplift applied, so spending is understated in later years to the extent prices rise. |
Results
| Horizon | Case | Additional small-boat arrivals | Supported caseload at horizon | of which small-boat cohort | Support spend, final 12 months | of which small-boat cohort | Cumulative support spend | of which small-boat cohort |
|---|---|---|---|---|---|---|---|---|
| 12 months | no further small-boat arrivals scenario | 0 | 69,545 | 0 | £1.57bn | 0 | £1.57bn | 0 |
| lower scenario | 23,259 | 57,739 | 9,693 | £1.41bn | £125m | £1.41bn | £125m | |
| central scenario | 28,574 | 86,674 | 17,129 | £1.77bn | £198m | £1.77bn | £198m | |
| higher scenario | 33,890 | 123,362 | 26,478 | £2.19bn | £281m | £2.19bn | £281m | |
| 24 months | no further small-boat arrivals scenario | 0 | 59,617 | 0 | £1.27bn | 0 | £2.85bn | 0 |
| lower scenario | 48,290 | 48,145 | 12,862 | £1.04bn | £245m | £2.45bn | £370m | |
| central scenario | 61,948 | 85,889 | 26,272 | £1.74bn | £473m | £3.52bn | £671m | |
| higher scenario | 75,607 | 146,712 | 47,631 | £2.77bn | £804m | £4.96bn | £1.09bn | |
| 36 months | no further small-boat arrivals scenario | 0 | 55,467 | 0 | £1.14bn | 0 | £3.99bn | 0 |
| lower scenario | 73,321 | 45,440 | 13,756 | £937m | £278m | £3.39bn | £648m | |
| central scenario | 95,322 | 85,561 | 30,094 | £1.73bn | £590m | £5.25bn | £1.26bn | |
| higher scenario | 117,324 | 161,001 | 60,575 | £3.12bn | £1.13bn | £8.08bn | £2.21bn | |
| 48 months | no further small-boat arrivals scenario | 0 | 53,732 | 0 | £1.09bn | 0 | £5.07bn | 0 |
| lower scenario | 98,352 | 44,676 | 14,008 | £908m | £288m | £4.30bn | £936m | |
| central scenario | 128,696 | 85,424 | 31,692 | £1.73bn | £639m | £6.97bn | £1.90bn | |
| higher scenario | 159,041 | 169,745 | 68,497 | £3.35bn | £1.33bn | £11.43bn | £3.54bn | |
| 60 months | no further small-boat arrivals scenario | 0 | 53,006 | 0 | £1.06bn | 0 | £6.14bn | 0 |
| lower scenario | 123,383 | 44,461 | 14,079 | £900m | £291m | £5.20bn | £1.23bn | |
| central scenario | 162,070 | 85,367 | 32,360 | £1.72bn | £659m | £8.70bn | £2.56bn | |
| higher scenario | 200,758 | 175,095 | 73,344 | £3.48bn | £1.45bn | £14.91bn | £5.00bn |
"Spend, final 12 months" is the modelled support spending in the twelve months ending at the horizon; "cumulative" is from the origin. Both cover accommodation and cash support only, at the stated unit-cost vintages, and exclude casework, tribunals, legal aid, unaccompanied children, local-authority costs not grant-funded, and anything else. They are therefore not comparable with the £4.9bn system total above and should not be set beside it.
Back-test
Starting from the observed caseload of 106,075 on 30 June 2025 and feeding in the 33,374 arrivals the ledger actually recorded over the following twelve months, the same machinery gives the caseloads below for 30 June 2026. The Home Office observed 93,293.
| Case | Modelled caseload | Error against observed |
|---|---|---|
| lower | 64,416 | -28,877 (-31.0%) |
| central | 93,266 | -27 (-0.0%) |
| higher | 129,298 | +36,005 (+38.6%) |
The central exit hazard (7.0% a month) is fitted so that the central case lands on the observed figure; its near-zero error is by construction, not a validation. The lower and higher rows show what the spread in the hazard and in the share entering support does over one year. The back-test checks the caseload arithmetic only; it cannot check the spending side, because no published figure gives support spending for the year ending June 2026.
What would change the answer most
- The exit hazard: at 6% a month a cohort's mean stay is about 17 months; at 15% it is under 7. It is the least settled parameter and the most powerful.
- The hotel share: every person moved from a hotel to dispersal at the stated rates saves about £103 a night in this calculation.
- The share entering support, which no published statistic gives for small-boat claimants. A single FOI answer would replace an assumption with an observation.
- Unit-cost vintage: the hotel rate is a 2023-24 figure and no inflation is applied.
Download
- costs.json: the recorded spending, funding flows and parameter sources, hand-curated with citations.
- costs-results.json: every month of every case, the back-test and the assumptions as run.
- context.csv: the related-population series with source hashes.
Corrections to anything on this page are made in place and logged in the corrections log.