Visa Fee vs Cost Markup Index (2026)

What gets into the index

A route enters this table only if the government itself publishes both numbers — the fee and its own estimate of what delivering the service costs — and only if both come from the same dated publication. Nothing here is estimated, modelled or inferred. That rule is what keeps the index defensible, and it is also why it is short: 64 routes across exactly two countries.

Countries with no published cost are listed as having no published cost. They are never assigned an imputed or estimated one, and they are never ranked. Canada publishes a per-application cost figure only for citizenship, and that figure is a 2010–11 estimate; New Zealand publishes programme-level totals but no per-application cost; Australia and Ireland publish nothing comparable. All four are shown below the index with what they do publish.

5 further rows were excluded even though a figure exists on both sides — because the fee is a base rate before mandatory add-ons, because no matching fee could be verified, or because the charge is a levy rather than a payment for processing. Those exclusions are listed rather than made silently.

Why no currency conversion

Amounts are never converted between currencies in this index. A markup is a ratio, and ratios compare across countries perfectly well on their own; converting the underlying pounds and dollars would add an exchange rate, an as-of date and FX noise without improving the comparison. Native amounts and the dimensionless multiple sit side by side in every row, and absolute gaps are only ever compared within a single currency.

How the UK builds a unit cost — and why it moves

The UK published unit cost is a full-cost estimate: direct costs, local and central overheads (accommodation, HR, Finance, IT), depreciation, cost of capital employed, and other immigration and nationality connected factors such as operational policy, apportioned across services using operational business planning productivities.

The stated settlement cost has been restated repeatedly: 13 April 2023 £491, 4 October 2023 £646, 9 April 2025 £523, 8 April 2026 £310. The fee rose at every one of those points; the cost went up, then down, then down again by 40.7% in a single year. A 41% fall in a stated cost is not a plausible efficiency gain.

The Home Office states that unit costs "may be influenced by changes in the way that applications in certain routes are processed from year to year, for example where additional checks are introduced or required, or by changes elsewhere within the overall system which impact on the weighting calculations and therefore the amount apportioned to any individual service." A unit cost is an apportionment output, not a measured cost, and a fall in the stated figure is not evidence of processing getting cheaper.

Read the unit cost as an apportionment output, not a measurement, and do not read the series as a trend in processing efficiency.

The two publications the UK issues, and why only one is usable

The fee table most people find and cite — "Home Office immigration and nationality fees, 8 April 2026", published under "UK visa fees" — contains no unit cost column at all. It has Current Fee, New Fee, Change and % Change and nothing else. The unit costs live in a separate publication, "Visa fees transparency data", in a table titled "Table with details of fees and unit costs". Every UK figure in this index comes from the second one, where the fee and the cost sit on the same row of the same table.

That is why the £523 figure keeps circulating. It was the published settlement unit cost from 9 April 2025 to 7 April 2026, and it was superseded on the day the £3,226 fee took effect. Pairing the new fee with the old cost gives 6.17× instead of 10.41× — it overstates the cost and understates the markup at the same time. The honest £523-era statistic is £3,029 against £523: 5.79×, with 17.3% of the fee accounted for by cost.

Why the US rows are not quite like the UK rows

US fees are current — set by the 2024 fee rule and still codified at 8 CFR 106.2 in August 2026 — but the unit costs are FY2022/23 forecasts published in November 2023. UK fees and costs come from the same April 2026 table; US ratios are a 2024 fee divided by a three-year-old forecast. Every US row is flagged for this, and the two countries are not strictly like-for-like without the caveat.

USCIS states that its unit costs "are only estimates. They are not prior year actual expenses. USCIS does not track actual costs by immigration benefit request." Appendix Table 4 divides projected cost by fee-PAYING receipts rather than total workload, so forms with heavy fee waivers carry a mechanically higher per-paying-applicant cost.

Appendix Table 4 of the November 2023 supporting documentation is used throughout, because it is the actual costing-model output and the document the final rule's own footnotes cite. Its figures differ from the numbers quoted in the rule's preamble text by roughly 3–5% — the preamble says $1,150 for naturalisation where the appendix it cites says $1,115 on a blended basis. Appendix Table 5, which divides by total workload rather than fee-paying receipts, gives $768 for the same form and is the better answer to "what does it cost to process one application", where Table 4 answers "what does each paying applicant have to cover".

Proposed fees are excluded from the ranked index entirely. The June 2026 proposal to raise US naturalisation to $1,330 is a notice of proposed rulemaking, not law, and it is shown separately and labelled as such.

What a markup is, and what it is not

A high multiple is not a measure of profit. The UK surplus is legally authorised — the Immigration Act 2014 permits fees set above cost by reference to the wider benefits of the service — and it funds routes deliberately priced below it. This data proves that directly: 10 of the 46 UK routes here are loss-making against the government's own cost, including student sponsor licences at 29.9% cost recovery, six-month visit visas losing £24 each, and administrative review — the mechanism for challenging a Home Office decision — priced below cost.

The US runs the same logic in reverse: naturalisation at 0.67× under-recovers by $377 per applicant as a deliberate choice to keep citizenship affordable, funded by other forms. The June 2026 proposal is an explicit move to end that.

No aggregate surplus figure appears anywhere on this page. Neither government publishes application volumes alongside these tables, so any total would be an invention.

The historical series, and its limits

The UK settlement fee series runs from £155 in 2003 to £3,226 in 2026 — 20.81 times, 1981.3%, a nominal compound rate of 14.1% a year over 23 years. Every point is taken from the statutory instrument or official table that made it.

That headline needs a caveat: the £155 charged in 2003 was a general leave-to-remain postal fee covering variation of leave and indefinite leave alike, not a settlement-specific fee, and the in-person fee that year was £250. No dedicated settlement fee existed before 2010. The stricter like-for-like series starts there: £900 to £3,226, 3.58 times over 16 years.

All growth figures are nominal. No inflation adjustment appears anywhere on this page because no CPI series is carried in this dataset; any real-terms line would have to be sourced separately first.

Route mapping is a judgement where the table has no label

The UK table has no line called "spouse" or "partner". Overseas partner entry clearance is mapped to the published "Route to Settlement" category and in-country partner extensions to "Leave to remain – Other". Those are mappings based on the fee categories and enabling provisions, not Home Office labels, and every affected row says so in its own note. Where a common name and a published label diverge, the published label is the one shown in the table.