In October 2026 Czechia held municipal elections and elections to a third of the Senate, the upper house of parliament, a year after the October 2025 general election to the Chamber of Deputies. ANO is the populist movement of Andrej Babiš; SPOLU (Together) is a centre-right coalition of ODS, KDU-ČSL and TOP 09; STAN (Mayors and Independents) is a centrist party; Piráti is the Czech Pirate Party; SPD (Freedom and Direct Democracy) is a far-right party.
Of the people who voted for ANO last October, roughly half did not vote in this year's municipal and Senate elections. In the 14 cities and 8 Senate districts where counting has finished and where the results can be compared polling district by polling district with the 2025 general election, the share of last year's ANO voters who did not turn out this year ranges from 38% to 76% (median 54%). Among SPOLU voters, no more than 20% stayed away in 20 of the 22 places, and the median is 7%.
In the same places, turnout fell from a median of 66% in 2025 to 40%, and on average 41% of last year's voters stayed at home. ANO voters were therefore missing 13 percentage points more often than average, SPOLU voters 34 points less often. Individual votes cannot be linked to individual people, so we use polling-district results and two independent methods (ecological inference and constrained regression) to estimate how last year's voters split between this year's candidates and lists and those who did not turn out. The method and its weaknesses are described in the box at the end.
For each of the 22 places, the chart shows two lines: a dark one for ANO voters and a lighter one for SPOLU voters. The ends of each line are the estimates from the two methods, and the dot is their average. ANO voters sit well to the right everywhere, from Plzeň (38%) to Havířov (76%); SPOLU voters are below 16% everywhere except Louny (39%) and Kladno (23%).
The gap between the two groups exceeds 30 percentage points in 21 of the 22 places. The exception is Louny, where the methods disagree about SPOLU voters (33% and 46%) and the gap against ANO comes out at 14 points. In Havířov, Most, Ostrava-město, Ústí nad Labem and Karviná the gap is 65 to 75 points.
The chart does not say that ANO voters are generally harder to mobilise. Turnout in general elections is 26 percentage points higher than in municipal and Senate elections, and people relate differently to local lists than to national parties. Nor does it follow from the chart that the remaining half voted for ANO: in Ústí nad Labem, the ANO list received only 2,200 to 3,000 of the movement's 16,800 voters from last year.
The ranges for SPD and Pirate voters are wide, but the direction is clear
According to the regression, a median of 82% of SPD voters stayed at home, and 66% according to ecological inference; for Pirate voters the figures are 67% and 50%. Regression and ecological inference agreed within 15 percentage points in only 6 of the 22 places for SPD and in 9 for the Pirates, compared with 21 for ANO and 16 for SPOLU. For both camps we therefore give a range rather than a single number: for SPD roughly two-thirds to four-fifths, for the Pirates a half to two-thirds. Parties with few voters in each polling district, or with a distribution similar to other parties, are hard to separate in polling-district data, which is why the estimates for these two groups are the least precise.
Prague: 50,000 to 57,000 of the movement's 127,000 voters from last year backed the ANO list, the ranges for other parties are wider
In Prague, 127,000 people voted for ANO last year, after adjusting to this year's electoral register. This year, the ANO 2011 list, which won 14.4% of the vote, received 52,800 of them according to the regression and 56,300 according to ecological inference; the range of the two methods is 49,800 to 57,300. Between 60,700 and 76,600 of them did not turn out (regression 71,000, ecological inference 62,600), so about half.
The table shows the ranges for the largest flows. Where the two methods disagree, the range runs from the lowest lower bound to the highest upper bound of their 90% intervals. It is wide, but the true value probably lies within it, provided the assumptions described in the box at the end hold. It is narrowest for ANO and STAN. SPOLU voters split mainly between the ODS and TOP 09 coalition (22.1% of the vote) and STAN (23.2%, first in Prague), with the rest scattered. The difference between the methods is largest for the Pirates: according to the regression, the Pirate list received 2,500 votes from the party's voters of last year, according to ecological inference 24,500.
| Last year's voters | This year | Range (votes) | Regression / ecological inference |
|---|---|---|---|
| ANO (127,000) | ANO list | 49,800 – 57,300 | 52,800 / 56,300 |
| ANO | did not vote | 60,700 – 76,600 | 71,000 / 62,600 |
| SPOLU (217,000) | ODS and TOP 09 | 71,900 – 89,700 | 86,500 / 74,800 |
| SPOLU | STAN | 49,100 – 66,700 | 60,900 / 51,800 |
| SPOLU | Praha sobě | 7,700 – 31,100 | 15,600 / 28,700 |
| SPOLU | Pirates | 1,000 – 25,200 | 22,400 / 2,200 |
| STAN (85,500) | STAN | 29,200 – 43,500 | 35,400 / 40,800 |
| STAN | Pirates | 6,800 – 16,500 | 13,700 / 9,000 |
| STAN | Praha sobě | 4,200 – 24,900 | 17,800 / 6,400 |
| Pirates (107,000) | Praha sobě | 35,800 – 48,900 | 42,500 / 37,700 |
| Pirates | Pirates | 0 – 26,600 | 2,500 / 24,500 |
| Pirates | did not vote | 40,100 – 70,100 | 62,300 / 42,500 |
| SPD (33,500) | did not vote | 21,400 – 33,500 | 31,100 / 22,600 |
Prague is the largest unit we calculated: 1,107 polling districts and 909,000 voters. Ecological inference converged worst there of all places (R-hat 2.9).
In Ústí, centre-right voters from 2025 mostly backed one joint list
In Ústí nad Labem, the Jedno Ústí (One Ústí) list won 51.3% of the vote. According to the abbreviation on the ballot, it is backed by the Greens, STAN, the Pirates, TOP 09 and others. It received 4,700 to 5,300 votes from last year's SPOLU voters, 3,000 to 3,800 from STAN voters and 2,700 to 3,400 from Pirate voters. The coalition of KDU-ČSL and ODS (SPOLU pro Ústí, 7.1%) received only about 1,050 votes from last year's SPOLU voters, the same in both methods. Of last year's 6,700 SPOLU voters, roughly three-quarters went to the joint list and just under a sixth to the KDU-ČSL and ODS list.
By contrast, of last year's 16,800 ANO voters, 11,900 to 12,500 did not turn out and 2,200 to 3,000 voted for the ANO list. At 70%, Ústí has the third-highest share of non-voting ANO voters of the 22 places, after Havířov (76%) and Most (70%).
In Kladno, almost two-thirds of last year's ANO voters stayed at home; in Žďár, where ANO fielded no candidate, slightly less than half
In Kladno, 23,900 people voted for ANO last year, after adjustment. Between 4,400 and 6,200 of them, a fifth to a quarter, voted for Aleš Gerloch, the ANO candidate. Where the rest of his votes came from cannot be reliably broken down from these data. According to both methods, 62% to 63% of last year's ANO voters stayed at home, roughly 15,000 people. Adéla Sucharda Šípová (according to the candidate list a senator and lawyer, candidate of a coalition whose abbreviation includes the Greens, STAN, the Pirates, TOP 09 and Senátor 21) received 8,200 to 8,700 votes from last year's STAN and Pirate voters. Petr Paták (ODS, secondary school head) received 4,800 to 6,700 from SPOLU voters.
In Žďár nad Sázavou, ANO had no candidate. There were 23,100 ANO voters from last year here: according to the estimates, Michal Šmarda (SOM+Volt+HR, mayor of Nové Město na Moravě) received 4,900 to 6,100 votes from them, Radek Koten (SPD, member of the Chamber of Deputies) 4,100 to 4,900, and roughly half did not turn out. Josef Klement (KDU-ČSL, senator) received 6,500 to 7,500 votes from SPOLU voters and Radek Černý (ODS, orthopaedic surgeon) 5,600 to 6,700. Second place between Klement and Černý was decided by a single vote (11,373 to 11,372), as described in our comparison with the prediction (in Czech).
In Kladno, the run-off on 16 and 17 October will be between Sucharda Šípová and Paták, who won 14,766 and 14,112 votes in the first round. The difference is 654 votes, whereas Gerloch's 10,600 voters and the roughly 15,000 ANO voters from last year who stayed at home are many times larger groups.
How we calculated it and what we know about it
We work with polling-district results from the 2025 general election and from the 2026 municipal and Senate elections (open data at volby.gov.cz) in places where counting has finished. We drop polling districts where the number of registered voters changed by more than 15% compared with last year, because their boundaries or numbering have probably changed. In Plzeň and Ostrava this applies to 26% to 28% of polling districts, and the estimates there do not include that part of the city. For the municipal elections, we convert votes for lists into votes of voters (the list's share multiplied by the number of envelopes cast).
We make the estimate using two methods: ecological inference (pyei, multinomial-dirichlet-modified model) and constrained regression with a bootstrap over polling districts. The sampling in ecological inference did not converge; R-hat comes out at 1.7 to 2.9, so its intervals are probably too narrow. We therefore give the range of both methods: where they overlap it is narrower, and where they disagree it runs from the lowest lower bound to the highest upper bound of the two 90% intervals and is wide. In a simulation with a known truth, the 90% regression range covered the true value in 87% to 89% of cases, but only if voter behaviour does not depend on the composition of the polling district. If it does, the ranges are probably too narrow.
