91,789 Swedes Voted From Abroad, a Record. The Roll They Voted From Drops People on a Ten-Year Timer.

On 25 September the Swedish Election Authority published the number nobody outside the diaspora was waiting for.

91,789 Swedish citizens voted from abroad in this year's election. In 2022 the figure was 75,312. It is a record, and it is a rise of nearly 22 percent in four years, in a country where domestic turnout moved by less than a point.

The authority also noted that the number may still creep up, because late ballots from abroad were still being registered when it published.

This is the follow-up we promised when we reported the preliminary count on 15 September. It contains the final figures, what the overseas vote actually looked like, and three numbers in our own earlier coverage that turned out to be wrong.

The final result

The Election Authority established the Riksdag result on 19 September. Turnout was 84.9 percent — 6,834,413 votes cast from 8,051,355 registered voters.

• Socialdemokraterna 28.02% — 99 seats
• Moderaterna 19.85% — 70 seats
• Sverigedemokraterna 17.48% — 62 seats
• Vänsterpartiet 8.40% — 30 seats
• Centerpartiet 7.03% — 25 seats
• Kristdemokraterna 6.17% — 22 seats
• Miljöpartiet 6.12% — 22 seats
• Liberalerna 5.34% — 19 seats

The bloc arithmetic held exactly where election night left it: 176 seats to 173, the left bloc on 49.57 percent of the vote against 48.84. A majority is 175.

So the headline finding of our first piece survived the count. The 300,000 late and overseas ballots that were still outstanding on election night moved two seats between parties — one from the Social Democrats to the Left Party, in effect — and not one across the line between blocs.

What we got wrong

Three figures in our 15 September piece were drawn from preliminary or secondary sources and were wrong. We are correcting them here rather than quietly editing them.

Turnout did not fall. We reported roughly 80 percent and called it a second consecutive decline. The final figure is 84.9 percent, against 84.21 in 2022. Turnout rose. Election-night percentages understate turnout for exactly the reason this article is about: the ballots counted last are disproportionately the ones cast from somewhere else.

The Social Democrats took 99 seats, not 100, and the Left Party 30, not 29. Both were preliminary at the time.

The electoral roll figure was badly wrong. We wrote that 370,271 Swedes abroad were on the roll. The correct number is just over 210,000. The Election Authority said so in March, said so again when it posted the voting cards in July, and Svenskar i Världen — the organisation that exists to represent Swedes abroad — gives the same figure. The increase of 41,146 on 2022 that we reported alongside it was right, which should have been the tell: 41,146 is 23 percent of 179,000, and it is not 23 percent of 370,271. We should have checked the arithmetic against itself.

How the overseas vote was actually cast

The breakdown is worth having, because it is the part that determines what a diaspora organisation in Minneapolis or Toronto can usefully do next time.

Of the 91,789 votes from abroad, 56,132 were cast in person at Swedish embassies and consulates — 289 of them worldwide — and 35,657 were postal votes.

In other words, the in-person route, the one with fixed dates set locally and mostly advertised in Swedish, carried nearly three votes in five. The postal route carried the rest.

The five busiest missions were London (4,467), Oslo (4,288), Copenhagen (3,722), Torrevieja on the Spanish coast (2,416) and Brussels (1,567). Not one is in North America. The Swedish electorate abroad is, overwhelmingly, a European and Nordic one — which is part of why the North American arrangements get the attention they do, and part of why they do not change.

Just over 1,000 votes cast from abroad never reached a Swedish municipality at all: the voter could not be identified, was not entitled to vote, or the ballot arrived too late.

The number that is easy to miscite

A figure that circulates in coverage of Swedes abroad is that around 13 percent of their postal votes get thrown out. It is a real number attached to the wrong election.

In the 2019 European Parliament election, 2,953 postal votes from abroad were rejected out of 21,226 — 13.9 percent. In the 2022 general election the figure was 2,164, or 6.3 percent. General elections do considerably better than European ones, and the gap is large enough that using the higher number to describe a Riksdag election overstates the failure rate by more than double.

The reasons, in both cases, are administrative rather than political: the ballot arrives late, a witness signature is missing, the personal identity number is incomplete, the voter has been struck from the roll, or the envelope was posted inside Sweden, which is not allowed.

The roll is shrinking by attrition, and the state knows it

Here is the part that ought to concern anyone running a Swedish-American or Swedish-Canadian institution.

A Swedish citizen who emigrates stays on the electoral roll for ten years automatically. After that they must re-register, whether or not they have voted. Voting itself counts as re-registration for a further ten years — so one missed cycle is survivable and two, depending on when you left, are not.

In March 2026 the Election Authority wrote to 82,000 Swedes abroad who had either dropped off the roll or were about to. Against a roll of roughly 210,000, that is close to two letters for every five people on it.

The roll grew anyway — by 41,146, to a record — which is the encouraging half of the story and suggests the letters, and the work organisations like Svenskar i Världen do, are landing. But the mechanism underneath is unchanged. It removes people quietly, on a ten-year timer, and the people it removes first are those whose administrative connection to Sweden has already thinned to a forwarding address. Those are precisely the people a heritage organisation is otherwise best placed to reach.

And there is still no government

Eighteen days after the final result was established, Sweden has not formed one.

Magdalena Andersson was given the mandate by Speaker Andreas Norlén on 18 September. On 28 September she handed it back, after the Left Party declined to hold the bloc together in the vote for Speaker and she concluded there was, in her words, no possibility for her to form a government. Norlén reopened consultations on 29 and 30 September and has not reassigned the mandate.

Under Swedish rules the Speaker may put up to four candidates to a vote. If all four fail, an extraordinary election follows automatically, to be held within three months.

Which brings the two halves of this article together, and is the reason it matters here rather than only in Stockholm.

The overseas voting window for a Swedish election opens about six weeks before polling day. Postal material could not be prepared before 30 July for a 13 September vote; missions began taking ballots on 20 August; every North American window had closed by 4 September. A snap election held inside three months would compress that entire calendar — and it would run through the same 289 missions, with the same locally set hours, in November or December.

91,789 Swedes abroad managed it this time, with a full summer of notice. Nobody should assume the same number would manage it again on six weeks.

Final results, turnout and the overseas breakdown are from Valmyndigheten, the Swedish Election Authority: the result-established notice of 19 September 2026 and the overseas-voting release of 25 September 2026. Electoral-roll figures are from Valmyndigheten's notices of 27 March and 7 July 2026 and from Svenskar i Världen. Postal-vote rejection rates are from Svenskar i Världen's analysis of Valmyndigheten data. Government-formation reporting is from The Local and the Irish Times, 28–30 September 2026. Our earlier coverage — "Sweden Just Reversed Its 2022 Election. Exactly." (15 September) and "Sweden Votes on Sunday Week" (4 September) — has been corrected where noted above. Photograph: the American Swedish Institute, Minneapolis, by Tony Webster, licensed CC BY 2.0, via Wikimedia Commons.