How we read Italian polls
How we turn Italy's polls into a forecast for the Camera dei Deputati (400 seats, 201 for a majority): the 33 polls of the last 30 days, corrected for how Italy's final polls have missed before. Today's average: FdI 27.4, PD 20.5, M5S 12.9, FI 7.1.
Our rule everywhere: measure how far final polls missed at past elections, take off half the average miss, and keep the correction only if it would have made past forecasts better when each election is predicted without knowing its result. Then run the shares through the country's own electoral law, which in Italy also needs an assumption about who stands with whom.
The seats now follow the new electoral law
Since the law's final approval on 8 October 2026 our seats are worked out under it: every Camera seat shared out nationally by vote (Hare quota, largest remainder) among coalitions on 10% and lists on 3%, plus the best list below 3% inside each coalition; 70 extra deputies (35 senators) for the coalition that comes first with 42%, capped at 220 (113); the seats abroad, in Valle d'Aosta and in Trentino-Alto Adige shared in proportion to the national vote. The Senate uses the same national shares; the law shares its seats region by region.
The polls, the correction test and the simulations below are unchanged. The collegio step still runs, but only to show who leads in each former collegio on the maps; the 2022 check and the collegio tests describe the old law, which elected the current Parliament. The new law →
Why we correct
Averaged over 2001–2022 the final polls missed by: FdI −0.3, PD +1.7, M5S −5.7, Lega −0.4, FI −0.4, AVS +1.0 points (positive = too high). Predicting each election with the correction learned from the others cut the average miss per party from 1.92 to 1.82 points (closer at 4 of 6 elections), so we do it. Most of the gain is the M5S, which final polls have understated every time. The test →
Polls ask about lists, but a third of the seats go to coalitions' candidates. Until the lists are filed, weeks before the vote, nobody knows the line-ups, so we show three and make one our projection:
| Line-up | Who with whom | Centre-right | Centre-left |
|---|---|---|---|
| Centre-left 'campo largo' with the M5S (our projection) | PD, M5S, AVS, +Europa and Italia Viva together, as in the regional elections since 2024; FdI, Lega, FI and Noi Moderati together; Azione and Futuro Nazionale alone. | 143 | 224 |
| Campo largo; Futuro Nazionale with the centre-right | As the default, but with Futuro Nazionale inside the centre-right coalition. | 224 | 165 |
| The 2022 line-up | As in 2022: the M5S alone; PD, AVS and +Europa together; Azione and Italia Viva together; Futuro Nazionale alone. | 223 | 96 |
- Average: each pollster's latest poll within 21 days of the newest, equal weight per pollster.
- Collegi: each list's 2022 vote in every collegio, scaled by its national change; Azione and Italia Viva split the 2022 joint list in their national ratio; Futuro Nazionale, new since 2022, takes its national share everywhere. Each coalition's candidate gets the sum of its lists; the most votes wins.
- Proportional: 245 seats by largest remainder nationally among coalitions with 10% (and a list with 3%) and lists with 3%; within a coalition, among its lists with 3%; a list's votes count for its coalition from 1%. The SVP keeps its minority-rule seat.
- Abroad: the 8 seats of the Circoscrizione Estero are held at their 2022 result.
- Chances: 5,000 simulations; each list moves by its typical final-poll miss nationally and a little more in each collegio.
Proportional or uniform? 1994–2013
The Camera elections of 1994–2013 were fought on the same 26 circoscrizioni. Replaying each from the one before with the true national change, and counting only pairs from 2000 on (the model uses no earlier election; 2 older pairs are shown below but not used), proportional swing (what the forecast uses) had the smaller share error at 2 of 3 pairs (1.5 against 1.9 points per list) and named the largest list in 62 of 78 circoscrizioni (uniform 62). So the forecast stays proportional.
| Pair | Largest list right, proportional | … uniform | Share error, proportional | … uniform |
|---|---|---|---|---|
| 2008 → 2013 | 14 / 26 | 14 | 1.5 | 2.1 |
| 2006 → 2008 | 23 / 26 | 23 | 1.6 | 2.3 |
| 2001 → 2006 | 25 / 26 | 25 | 1.3 | 1.2 |
| 1996 → 2001 not used by the model | 24 / 26 | 24 | 1.6 | 2.0 |
| 1994 → 1996 not used by the model | 22 / 26 | 21 | 2.2 | 2.5 |
Lists by lineage: the Ulivo parties and the PD as one, Forza Italia, AN and the PdL as one, the parties to the left of the PD as one. Source: Ministero dell'Interno, Eligendo open data (list votes by comune, summed by circoscrizione); Valle d'Aosta and the vote abroad are left out.
The seat law on the 2022 votes
Given the official 2022 list votes and the 2022 line-up, our calculation gives: Centre-right 237, Centre-left 85, M5S 52, Third pole 21, SVP 3, Others 2. The official result was centre-right 237, centre-left 85, M5S 52, Azione–Italia Viva 21, SVP 3, others 2: every collegio and every proportional seat by list match.
How the local model did
The collegi come from moving each list's previous vote by its national change (proportional swing). To test that step on its own, we took the 2018 list votes rebuilt on the 2022 collegi, moved each list by its true national change to 2022, added the lists new in 2022 (Azione–Italia Viva, Impegno Civico) at their national share everywhere, summed each 2022 coalition and compared the winners with the real ones. No polls are involved. Valle d'Aosta, which has no lists, is left out.
| 2022 collegi won | Projected from 2018 | Actual |
|---|---|---|
| Centre-right | 131 | 121 |
| Centre-left | 11 | 12 |
| M5S | 2 | 10 |
| SVP | 2 | 2 |
| Others | 0 | 1 |
Right winner in 128 of 146 collegi. Average error in coalition share: 3.2 points (against the candidates' share); in list share: 2.1 points.
* in a city split between collegi (see below).
Where proportional swing needs care
- The M5S in the south. Its vote halved nationally between 2018 and 2022, but it fell much less in Campania, Puglia, Calabria and Sicily, where it held 8 collegi that a uniform proportional fall gives to the centre-right.
- The six cities split between collegi (Rome, Milan, Naples, Turin, Genoa, Palermo). Our 2018 votes there are the city-wide result spread by population, so their 18 collegi all start from the city average; 8 of the misses are there, including the two largest, both in Rome.
- Local candidates and lists. Cateno De Luca's Sud chiama Nord won Messina, which no swing of national lists can find; and new lists (Azione–Italia Viva in 2022, Futuro Nazionale now) are spread evenly, though their vote is not.
9 of the 18 misses had a projected lead under 5 points, on a swing as large as FdI's from 4% to 26%. For this reason the simulations also move each list in every collegio at random. The same step is used for the Senate. The Senate →
The Senate law on the 2022 votes
Given the 2022 Senate list votes and the 2022 line-up, our Senate calculation gives: Centre-right 115, Centre-left 44, M5S 28, Third pole 9, Others 2, SVP 2. 74 of 74 collegi uninominali match the real winners, and the 122 proportional seats by list match exactly. The official result was centre-right 115, centre-left 44, M5S 28, Azione–Italia Viva 9, SVP 2, Sud chiama Nord 1, MAIE 1 (abroad; our "others" 2 are these last two).
Correcting final polls
Each election is predicted with the correction worked out from the other elections only, so it never sees the result it is tested on. Correcting Italian polls was closer at 4 of 6 and further out at 2 (average miss per party 1.82 against 1.92 for the polls as published). So our projection uses the correction.
| Election | Polls' average miss | Corrected | Verdict |
|---|---|---|---|
| 2001 | 1.19 | 1.42 | Correction further out |
| 2006 | 1.19 | 0.83 | Correction closer |
| 2008 | 1.52 | 2.03 | Correction further out |
| 2013 | 3.14 | 2.70 | Correction closer |
| 2018 | 2.95 | 2.42 | Correction closer |
| 2022 | 1.55 | 1.52 | Correction closer |
- Candidates matter in collegi (in 2022 a local list won Messina); we move list votes only.
- A list's collegi are our split of its coalition's wins by vote share; the real split is agreed between the parties.
- The electoral law may change before the vote; if it does, so will this calculation.
Figures as of . Published by Polling Forecast.



