The centre-right has shed 6.5 points in eighteen months — and its lead is gone
With the latest harvest from Italy's official political polling registry now fully digested into our database, we recalculated quarterly voting-intention averages over the last eighteen months. What comes out is not noise: the combined share of Fratelli d'Italia, Lega and Forza Italia falls by 6.5 points between Q1 2025 and summer 2026, and the coalition's lead over the centre-left — nearly 7 points at the start of the period — disappears entirely.
- −6.5points lost by the centre-right (FdI+Lega+FI) from Q1 2025 to Q3 2026
- +0.2change for the centre-left (PD+M5S+AVS) over the same period
- 0.0gap between the two blocs today, down from 6.7 points in early 2025
- 447non-suspect national polls behind these averages
The gap closes — by subtraction
suspect by our quality checks are excluded. *Q3 2026 is still a partial quarter (25 polls).Through the end of 2025 the picture is static: the centre-right oscillates between 46.7% and 47.4%, the centre-left between 40.5% and 41.1%. The movement starts in Q1 2026 and accelerates in Q2: −1.8 points, then −2.6, then another −2.1. The red line, meanwhile, does not move by a tenth.
That is the least intuitive part of the finding: the centre-left is not gaining. The gap closes entirely because the other side is falling.
Who loses, and how much
All three governing lists retreat, but not proportionally: Lega loses 2.4 points off a base of 8.5 — more than a quarter of its own support — against the 9% lost by Fratelli d'Italia. In relative terms Lega takes the hardest hit by some distance.
On the other side the Democratic Party is also down 1.1 points; the centre-left's flat total is held up by the Five Star Movement alone.
| Quarter | FdI | Lega | FI | PD | M5S | AVS | N |
|---|---|---|---|---|---|---|---|
| Q1 2025 | 29.4% | 8.5% | 9.4% | 22.6% | 11.6% | 6.3% | 84 |
| Q2 2025 | 29.9% | 8.5% | 9.0% | 22.4% | 12.3% | 6.3% | 56 |
| Q3 2025 | 29.1% | 8.6% | 8.8% | 22.2% | 12.5% | 6.7% | 51 |
| Q4 2025 | 30.0% | 8.4% | 8.8% | 21.9% | 12.7% | 6.5% | 84 |
| Q1 2026 | 29.6% | 8.3% | 9.2% | 22.7% | 13.1% | 7.1% | 67 |
| Q2 2026 | 27.7% | 6.7% | 8.1% | 21.8% | 12.6% | 6.7% | 80 |
| Q3 2026* | 26.7% | 6.1% | 7.8% | 21.5% | 12.6% | 6.4% | 25 |
It is not a sample-composition artefact
When an average moves, the first hypothesis to rule out is that the mix of pollsters behind it moved instead: if the houses that estimate a party lower happen to publish more in a given quarter, the average drops without anything changing in the electorate.
So we compared each house against itself, between the last quarter of 2025 and polls fielded from June 2026 onwards.
| House | FdI Q4 ’25 | FdI from Jun ’26 | Δ | Lega Q4 ’25 | Lega from Jun ’26 | Δ |
|---|---|---|---|---|---|---|
| SWG | 31.2% | 27.5% | −3.7 | 8.3% | 5.5% | −2.8 |
| Termometro Politico | 30.1% | 27.8% | −2.3 | 8.6% | 6.4% | −2.2 |
| Tecnè | 30.9% | 29.2% | −1.7 | 8.5% | 6.5% | −2.0 |
| Demopolis | 30.1% | 28.3% | −1.8 | 8.7% | 6.6% | −2.1 |
| BiDiMedia | 29.1% | 22.8% | −6.3 | 8.3% | 5.6% | −2.7 |
The decline shows up in all five houses with enough volume for the comparison, on both lists. On Fratelli d'Italia the magnitude varies a lot (−1.7 to −6.3), but the sign is the same everywhere; on Lega the convergence is tighter still, between −2.0 and −2.8 across the board. That is the kind of regularity a shift in pollster mix does not produce.
Where the points are going
Part of the answer is a list that did not exist in the archive eighteen months ago. "Futuro Nazionale", the list associated with Roberto Vannacci, appears in national polling from Q1 2026 and grows: 4.3% in Q1, 4.2% in Q2, 5.6% in the current quarter. A transfer inside the right-wing bloc is the simplest reading, and it points the same way as the FdI and Lega losses.
The rest cannot be reconstructed from voting-intention figures alone: registry polls report undecided and abstention shares under definitions that are not comparable across houses, and we have not aggregated them.
- Q3 2026 is partial. It rests on 25 polls against the 51–84 of full quarters: the last point of every series is the most fragile.
- March 2026 is an anomalous month. In the monthly averages the whole roster comes out inflated (PD 23.8%, M5S 14.6%, AVS 8.2%): a subset of polls with a narrowed list and redistributed shares. We have not corrected it — we flag it.
- The date is not the fieldwork date. Every poll is placed on its fieldwork end where available, not on registry insertion — which follows a median 3 days later, but more than two weeks later in 10% of cases.
- Party labels are still raw. The canonical party register for the official source is incomplete: the aggregation here is by label matching. We cover this in Inside the data.
- The centre-right goes from 47.2% to 40.7% over six quarters; the centre-left sits at 40.7%.
- The gap between the blocs closes by subtraction, not by opposition growth.
- The decline appears within every house: it is not a composition artefact.
- Lega loses more than a quarter of its own support in relative terms — the hardest hit of the three.
- The current quarter is still partial: September polling is what confirms it.
Source: ElectionLab polling archive — official register of political and electoral polls (database under art. 8 c.3 of Law 28/2000, Presidency of the Council of Ministers) and the onData dataset (CC-BY-4.0). Averages computed on national polls with data_quality other than suspect. Data as of 29 August 2026. Explore the full series in the Polls section.
— ElectionLab Studio