98 Approvals in a Month: Why Sweden's Citizenship Statistics Collapsed — and Why It Is Not What It Looks Like

The numbers, from Migrationsverket's own monthly table

Migrationsverket publishes a monthly citizenship table on its Swedish citizenship statistics page. As of this writing it shows the following, with the year-to-date column alongside the July column:

Three ratios follow from that table by simple arithmetic — these are our calculations from Migrationsverket's published figures, not agency-published rates:

• July 2026: 98 of 2 825 decided = about 3.5% granted.
• 2026 year to date: 18 332 of 31 054 = about 59% granted.
• January–June 2026, by subtraction: 18 234 of 28 229 = about 64.6% granted.

So the grant rate did not drift. It fell from roughly two in three to roughly one in twenty-nine, inside a single month. That is the number that has been circulating, and on its face it looks like a policy shift of extraordinary severity.

One point of precision before going further, because it matters for how much can be read into the gap. Avgjorda ärenden — cases decided — is not the same as cases refused. It covers every case closed with a decision, which includes grants, refusals, and applications written off or withdrawn. The 2 727 July decisions that were not grants are therefore not 2 727 refusals. Migrationsverket does not publish a refusal count in this table, and this article does not infer one.

What Migrationsverket says is happening

The agency addressed this directly in a news item dated 18 August 2026, headlined "Statistiken för medborgarskap ska tolkas med försiktighet" — the citizenship statistics must be interpreted with caution. Its opening sentence contains the whole mechanism:

"Migrationsverkets statistik för medborgarskap ska tolkas med försiktighet just nu. De första medborgarskapsproven genomfördes 15 augusti och i väntan på resultatet avgör myndigheten framförallt ärenden som avslås."

In English: Migrationsverket's citizenship statistics must be interpreted with caution right now. The first citizenship tests were held on 15 August and, while awaiting the results, the agency is primarily deciding cases that are refused.

The agency then states the constraint that produces this, verbatim:

"Medborgarskapsansökningar där Migrationsverket inväntar resultat på medborgarskapsprov kan inte avgöras just nu."

Citizenship applications where Migrationsverket is awaiting a citizenship-test result cannot be decided at present. And the counterpart:

"Medborgarskapsansökningar som avslås kan fortsätta att avgöras."

Citizenship applications that are refused can continue to be decided.

Those two sentences together explain the 3.5% completely, and they explain it as an artefact of sequencing rather than of severity. A case that fails on a requirement unrelated to the knowledge test — residency, conduct, income, identity — can be refused today, because no test result is needed to reach that conclusion. A case that would otherwise succeed cannot be granted today, because the knowledge requirement is now part of the assessment and the evidence for it does not exist until UHR publishes results in October. The refusals flow through; the approvals queue behind a date.

The consequence is that approvals are largely missing from the numerator rather than being converted into refusals. What looks like a collapse in success is, on the agency's account, a temporary removal of most successful cases from the pool of decidable ones.

Who is actually sitting the test — and why that sharpens the point

One sentence in the same notice is easy to skim past and worth pausing on:

"De som skriver proven är de sökande som uppfyller majoriteten av kraven för medborgarskap."

Those who sit the tests are the applicants who meet the majority of the requirements for citizenship. Migrationsverket does not direct someone to the test as a formality applied to every file. An applicant is directed there when the rest of their case already looks satisfiable and the knowledge requirement is what remains to be established.

Read against the statistics, this tightens the argument considerably. The population whose cases are currently frozen is, by the agency's own description, disproportionately the population whose applications were heading toward approval. That is precisely the group absent from July's 98. It is the strongest reason not to read the July grant rate as an estimate of anyone's odds.

The statistics page carries its own warning

The caution is not confined to a news item that will scroll away. Migrationsverket has placed a standing notice on the statistics page itself, under the heading "Nya regler påverkar statistiken" — new rules affect the statistics:

"De nya reglerna för svenskt medborgarskap, som började gälla den 6 juni 2026, påverkar statistiken. Många ärenden behöver kompletteras utifrån resultaten från medborgarskapsproven innan de kan avgöras. Detta innebär att det kommer vara större variation i siffrorna än tidigare. Statistiken bör därför tolkas med hänsyn till de nya reglerna."

The new rules that took effect on 6 June 2026 affect the statistics. Many cases need to be supplemented on the basis of the citizenship-test results before they can be decided. This means there will be greater variation in the figures than previously, and the statistics should therefore be interpreted with the new rules in mind.

Two things follow. First, the distortion is officially acknowledged at the source, not merely conceded when asked. Second, the agency is forecasting variation — not a one-off dip that reverses cleanly.

Expect the swings to continue — in both directions

Migrationsverket is explicit that the volatility will persist and will run both ways:

"Därför kommer Migrationsverket att ha en mer varierad statistik över medborgarskap – vissa månader visar den lägre grad beviljade och vissa månader högre grad beviljade, beroende på när proven genomförs och myndigheten får resultatet."

Migrationsverket will therefore have more varied citizenship statistics — some months showing a lower proportion granted and some months a higher proportion granted, depending on when the tests are held and when the agency receives the results.

This is the sentence that should govern how the next several monthly releases are read. When the October results reach Migrationsverket, a backlog of held cases becomes decidable more or less at once, and a month's figures will contain approvals that accumulated over a much longer period. That month's grant rate will overstate the underlying rate exactly as July's understates it. Neither figure is a measurement of how hard citizenship has become; both are measurements of when the paperwork moved.

The dates worth having in mind: monthly statistics are published on the 15th — the page states "Månadsstatistiken uppdateras den 15:e varje månad. Om den 15:e är på en helgdag, publiceras statistiken nästa vardag." The next release therefore lands in mid-September and will cover August, a month whose approvals are still blocked. A meaningful reading is not available until the releases that follow the October results.

The queue behind the numbers

The same table records 103 598 open cases — öppna ärenden, applications received that have not yet had a decision. The page defines the term: "Öppna ärenden är ansökningar/ärenden som har kommit in men inte fått beslut än."

Set that against the year's flow. Migrationsverket received 34 697 applications in 2026 and decided 31 054 — so intake has modestly outpaced decisions even before the test-related freeze took hold. The 103 598 figure is the accumulated stock, and the mechanism described above adds to it for as long as approvals cannot be issued. For the wider context on how this queue built up, see our post on the citizenship backlog and on how pending applications are assessed.

Migrationsverket states the consequence for applicants plainly, and it is the practical bottom line of the entire notice:

"För de sökande som skriver medborgarskapsprov innebär det längre väntetid innan deras ansökan om medborgarskap avgörs."

For applicants who sit the citizenship test, it means a longer wait before their citizenship application is decided.

A firmer date for the language test

The 18 August notice also contains the clearest official statement so far on the second half of the knowledge requirement:

"Provet om det svenska samhället erbjuds från och med augusti 2026, medan provet i svenska tidigast kommer kunna genomföras under 2027."

The test on Swedish society is offered from August 2026, while the test in Swedish can at the earliest be held during 2027.

This is worth flagging because it is more specific than what UHR currently publishes. UHR's own page says only "Prov i svenska kommer att införas vid senare tillfälle" — the test in Swedish will be introduced at a later date — with no year attached. Migrationsverket's tidigast under 2027 is a floor, not a scheduled date: it says the language test cannot happen before 2027, not that it will happen in 2027. Anyone planning around a specific month should treat the date as open until UHR publishes one.

How to read the next few months

Four practical rules follow from the agency's own statements.

1. Do not read a monthly grant rate as an approval probability. July's 3.5% is not the chance of a given application succeeding. It is the composition of one month's decidable cases, in a month when most approvable cases were not decidable.

2. Do not read the coming spike as a loosening either. The same distortion runs in reverse once results arrive. A high month will reflect released backlog, not a change in standards.

3. Compare like with like. The informative comparison is not July against June, but a full post-results period against a full pre-reform period — and that comparison cannot be made until well after October.

4. Treat headline coverage of these figures with care. A 3.5% grant rate is a striking statistic that travels easily without the sequencing caveat attached. The caveat is not a spokesperson's gloss; it is published by the agency in two places, and quoted verbatim above.

None of this is an argument that the 6 June framework is easy, or that refusals are not rising. The tightened residency, income, conduct and knowledge requirements are real and are covered in our overview of the 2026 changes. The narrow claim here is only that July's figure does not measure their severity — and that Migrationsverket says so itself.

Sources

• Migrationsverket, "Statistiken för medborgarskap ska tolkas med försiktighet" (18 August 2026). Primary source for the caution on the statistics, the refusals-proceed/approvals-frozen mechanism, who is directed to the test, the longer waiting times, and the "tidigast under 2027" statement on the language test.
• Migrationsverket, "Svenskt medborgarskap" — citizenship statistics. Source for the monthly table (applications received, cases decided, citizenships granted, open cases), the standing "Nya regler påverkar statistiken" notice, and the publication schedule on the 15th of each month.
• Universitets- och högskolerådet (UHR), "Om medborgarskapsprovet". Source for the first civics test having been held in August 2026 and for the undated position on the language test — see our post on the first test and the October results.
• Migrationsverket, "Nya regler för svenskt medborgarskap från 6 juni 2026". The framework underlying the knowledge requirement — see our 2026 changes overview.

The grant-rate percentages in this article are our own arithmetic on Migrationsverket's published figures and are labelled as such; Migrationsverket does not publish grant rates in this table. This article is independent journalism and general information about Swedish public administration. It is not legal advice. Questions about an individual citizenship case should go to Migrationsverket, and questions about the test to UHR.