Italy General electionby 22 Dec 2027in 396 days Centre-left majority 63% No majority in the Camera 0% Calendar Track record Research
New electoral law Our forecast now uses Italy's new electoral law, approved on 8 October 2026: no single-member seats, and 70 extra deputies and 35 extra senators for a coalition that comes first with 42% of the vote. The maps show which coalition leads in each former collegio; those are no longer seats. What changed →
Research · Italy
EN IT

Simulating Italian polling error: where the chances come from

Abstract. Seat ranges, the chance that a coalition reaches 201, the chance that none does and each list's chance of clearing 3% come from 2,000 simulated elections under the default line-up, each list moved by its typical final-poll miss and each collegio by a local factor. The M5S's spread is the widest on the site. Because the PD's and the M5S's errors have run in opposite directions, drawing them independently makes the centre-left's total more uncertain than the record suggests.

Polling Forecast · updated 9 November 2026 · 6 min read · Italy, simulation, uncertainty, majority, hung parliament, collegi, M5S, independent errors

Summary

The seat ranges, the chance that a coalition reaches 201, the chance that none does and the chance that each list clears 3% all come from 2,000 simulated elections. In each one, every list's national share is moved by a random error the size of its typical final-poll miss; every list's vote in every collegio is moved again by a small local factor; and the electoral law (the new law since 8 October 2026) is run under the default line-up. The outputs are counts across the 2,000 runs.

Each list's error is drawn independently. In Italy that matters more than elsewhere. The M5S's typical miss is more than six points, the largest of any party on the site, and the PD's more than three, and at every election both contested their errors ran in opposite directions. Drawn independently, they make the centre-left's total, which contains both under the default line-up, more uncertain than the record suggests. We say below by how much.

What has changed since this paper was written

Updated 9 November 2026. The rest of the paper is as first published, except where a sentence below says it was corrected.

  • The new electoral law is in force (8 October). The Chamber gave the new law its final approval on 8 October 2026, by 227 votes to 164, and the Italian forecast has used it everywhere since. As the forecast applies it to national shares:

  • Seats are shared in proportion to the vote, by Hare quota and largest remainder, among coalitions on 10% or more and lone lists on 3% or more. Inside a coalition, lists on 3% or more, plus the best list below 3%, share its seats; every list's votes count towards the coalition's total.

  • The bonus. The coalition or list that comes first with at least 42% of the valid votes gets 70 extra deputies and 35 extra senators, up to caps of 220 deputies and 113 senators.

  • The Senate is worked out the same way on national shares (the law allocates its seats by region). Seats abroad, Valle d'Aosta and Trentino-Alto Adige are shared in proportion to the national vote, an approximation.

  • The collegi no longer elect anyone. The maps and collegio pages show who leads each former collegio.

Where this paper describes the Rosatellum, it describes the law under which the 2022 Parliament was elected: it remains the basis of the check against 2022 and of the by-elections in this Parliament. /it/new-law shows the seats under both laws side by side.

  • Only elections from 2000 shape the forecast (28 September). Earlier elections stay in the long view and on the history pages but no longer feed any test, correction or error size. In Italy the tests were already on elections since 2001, so nothing changed.
  • Every outcome (3 October). The seats in every simulated election are shown as histograms at /it/outcomes.

One run, step by step

  1. Move each list's national share. Add to each list's share a random error drawn from a normal distribution with the standard deviation below. A share cannot fall below 0.05%. Others take what the named lists leave, with a floor of 0.5%.
  2. Project the collegi. Scale each list's 2022 vote in each collegio by its new national change, exactly as for the projection (From votes to seats: the Rosatellum), multiplying each list's vote in each collegio by its own random local factor.
  3. Count the seats. The collegio winners under the line-up; the proportional seats from the new national shares; the seats abroad as in 2022.
  4. Record each list's and each coalition's seats, whether each list is over 3%, the winner of each collegio, and whether the largest coalition has 201.

The random number generator has a fixed seed, so the same polls give the same chances.

National error

The standard deviation for list \(p\) with share \(S_p\) is

\[ \sigma_p = \begin{cases} \max(0.3,\ \text{RMS}_p) & \text{a list measured at past elections,}\\ 0.4 + 0.12\,S_p & \text{Futuro Nazionale,}\\ \max(0.3,\ 0.3 + 0.1\,S_p) & \text{any other list with no record,} \end{cases} \tag{1} \]

where \(\text{RMS}_p\) is the root mean square of the list's final-poll errors at past elections (Measuring Italian polling error, 2008–2022). In run \(r\),

\[ S^{(r)}_p = \max\!\left(0.05,\ S_p + \sigma_p\, z_{r,p}\right), \qquad z_{r,p} \sim \mathcal{N}(0,1) \text{ independently for each } p. \tag{2} \]

Today's values:

ListOur betElections measuredTypical final-poll miss (RMS)Standard deviation used
FdI27.030.720.72
PD19.262.772.77
M5S15.536.706.70
FI7.161.791.79
FN6.30–1.16
Lega5.962.422.42
AVS5.431.421.42
Azione3.00–0.60
IV2.40–0.54
+Europa1.30–0.43
NM1.10–0.41
Live The simulation's national error: standard deviation drawn for each list (points) Computed from the database, polls to 19 November 2026.

The M5S's spread, more than six points, comes from its misses of 10.4, 4.8 and 2.0 points; the PD's, more than three, from misses of up to 4.7. FdI, polled closely at three elections, has the smallest spread among the large lists. Futuro Nazionale gets a spread that grows with its share and is a little wider than that given to other new lists: about 1.3 points at 7.6%. Azione, Italia Viva, +Europa and Noi Moderati have no record of their own and get about half a point.

Two features of this design are worth stating plainly.

The spread does not shrink with the correction. The M5S's polls are corrected for half its average miss (Reading Italian polls: the correction and its test), and then drawn with a spread equal to its whole typical miss, which includes the part the correction has already taken off. That errs on the wide side.

Independent draws and the centre-left. At 2013, 2018 and 2022 the polls overstated the PD and understated the M5S. Their errors ran in opposite directions each time, so their sum, the part that matters for a coalition containing both, was missed by less than either: by about six points in 2013 and by less than a point in 2018 and 2022, a root mean square of about 3.4. Drawn independently with spreads of 3.3 and 6.7, their sum has a standard deviation of about 7.5. Under the default line-up the simulation therefore makes the centre-left's total considerably more uncertain than the record of the two parties together would. Under the 2022 line-up, with the M5S alone, the same independence makes the gap between the M5S and the centre-left less uncertain than the record. The British simulation avoids this by replaying whole past misses (Simulating UK polling error: where the chances come from); with four Italian elections we have not done so here.

Local error

Each list's projected vote in each collegio is multiplied by its own factor,

\[ \hat{u}^{(r)}_{k,l} \leftarrow \hat{u}^{(r)}_{k,l} \times \max\!\left(0.2,\ 1 + 0.08\, u_{r,k,l}\right), \qquad u_{r,k,l} \sim \mathcal{N}(0,1), \tag{3} \]

so that collegi do not all move in lockstep. The factor applies to the SVP's vote, held at 2022, and to Futuro Nazionale's even share as well. Its size, 8%, was chosen, not estimated. The proportional seats use national shares and are not affected by it.

What comes out

Coalition seats and majorities

For each coalition, the 80% range is the band within which its seats fell in the middle 80% of runs. In each run, the largest coalition is counted as having a majority if it has 201 seats or more; if it has fewer, the run counts as "no majority".

CoalitionCollegiSeats80% rangeMajority
Centre-left0224138–22463%
Centre-right0143140–22437%
FN02318–30–
Azione0100–13–
No majority0%
Live Our projection by coalition (Centre-left 'campo largo' with the M5S) Computed from the database, polls to 19 November 2026.

The majority chance is defined for the largest coalition in each run, so the chances of the coalitions and of "no majority" add up to 100%. It is a chance for pre-election coalitions under the assumed line-up (Who stands with whom: the coalition line-up assumption): a government formed after the vote from lists that stood apart is not counted.

Lists and the 3% line

Each list's seat range is counted the same way, with its collegi shared among its coalition's lists as described in From votes to seats: the Rosatellum. The "Clears 3%" column is the share of runs in which the list's national share is 3% or more. For a list inside a coalition that reaches 10%, 3% is the line for sharing in the coalition's proportional seats; for a list alone it is the line for any proportional seat.

ListCoalitionPollsOur betSeatsProportionalCollegiClears 3%
FdICentre-right27.427.094940over 99%
PDCentre-left20.519.21011010over 99%
M5SCentre-left12.915.58282097%
FICentre-right7.17.12525099%
FNalone6.56.323230over 99%
LegaCentre-right5.85.92020089%
AVSCentre-left6.15.42929096%
Azionealone3.13.01010052%
IVCentre-left2.42.41212013%
+EuropaCentre-left1.31.30000%
NMCentre-right1.11.14400%
Live Lists: polls, our bet and seats (collegi are our split of each coalition's wins) Computed from the database, polls to 19 November 2026.

Collegi

For each collegio the simulation records the share of runs each coalition won it, which the collegi pages and the map show. The least certain are those where the centre-right and the centre-left are within a few points of each other, and, because the M5S's spread is so wide, those where the M5S's vote is largest.

What we know about the simulation

We can say how it was built, not yet how well it performs: four elections are too few to check how often results fall inside the ranges. Four points matter.

The spreads come from four elections, three for the M5S. A miss larger than 2013's, or the M5S overstated for the first time, would fall outside what the spreads suggest.

Independent draws make the centre-left's total too uncertain under the default line-up, for the reason above.

The local factor is a guess, as in Germany and Spain.

Everything holds the line-up fixed. The largest source of uncertainty about the next Camera, who stands with whom, is not in the simulation at all; it is shown by the three line-ups side by side.

Limitations

The simulation turns one projection into chances with the least machinery the evidence allows: a normal draw per list, a local factor per list and collegio, and the seat law under one line-up. It does not model linked errors, regional differences in polling error, turnout, candidates, or the line-up. The polling-day version adds drift for the time left (From today to polling day in Italy). What we don't know yet: limitations of the Italian forecast collects the rest.

References

Jennings, W. and Wlezien, C. (2018). Election polling errors across time and space. Nature Human Behaviour, 2(4), 276–283.

How to cite

Polling Forecast (2026). Simulating Italian polling error: where the chances come from. Retrieved 21 November 2026 from https://www.pollingforecast.com/research/italy-simulation. Live figures change as new polls arrive; the date above is when the text was last revised.

You may also like