Withdrawals between rounds, France 2024
The 2024 French runoffs, the 224 candidates who withdrew, the transfer rates between rounds, and what the vote counts do and do not show.
n France's 2024 legislative election, third-placed candidates stood down before the second round so the vote against the Rassemblement National could pool behind one rival. What did that do, and what settled the outcome?
Ministère de l’Intérieur, 2024 results
The official Interior Ministry results files for both rounds of the 2024 legislative election, constituency by constituency.
The RN bloc's share of valid votes went from 33.35% to 37.14% between rounds, while its vote count fell from 10,689,896 to 10,132,261. The second round had fewer valid votes to divide: 27,279,713 against 32,057,946.
Of the 258 runoff seats where the RN bloc led after round one, it won 104. The left led 137 and won 120; the presidential bloc led 66 and won 66; the right led 31 and won 29.
224 withdrawals, but none of them was the round-one leader. A withdrawal removes a third-placed candidate, so the leader going into the runoff was unchanged in every constituency where one happened.
Across 151 RN-led runoffs after a withdrawal, RN votes rose 113.9% between rounds and the opponent's rose 175.5%. Across 96 with no withdrawal, RN rose 113.7% and the opponent 164.3%.
CEVIPOF (2024)
Eight Sciences Po researchers reading the second round the day after it happened, with the survey behind the numbers.
Unsupported block: pub.terrain.quote
Unsupported block: pub.terrain.quote
Fondation Jean-Jaurès (2024)
A Fondation Jean-Jaurès note using a 10,000-person election survey to estimate how voters of one bloc moved to another between rounds.
Unsupported block: pub.terrain.quote
Unsupported block: pub.terrain.quote
Unsupported block: pub.terrain.quote
Le Grand Continent (2024)
A seat projection, run before the second round, of what would happen with no withdrawals, with left-only withdrawals, and with transfers halved.
Unsupported block: pub.terrain.quote
The one published seat model that separates the effect of withdrawals from the effect of transfers does so using vote-transfer matrices from 2022, not 2024, and its author flags it to be taken with caution.
Mișcoiu & Gherghina (2024)
A peer-reviewed article that interviewed 29 of the candidates who stood down, asking why they did it.
Unsupported block: pub.terrain.quote
Gougou (2024)
A peer-reviewed election review that frames the second round as a rejection of the radical right at the ballot box.
Unsupported block: pub.terrain.quote
Ipsos-Talan, via France Info (2024)
A news report of an Ipsos-Talan survey giving the second-round transfer rates by voter group.
Unsupported block: pub.terrain.quote
LCP (2024)
A running count of the announced withdrawals four days before the second round, with their party breakdown.
Unsupported block: pub.terrain.quote
224 candidates who had qualified for the second round were not on the ballot in it: 135 from the left, 81 from the presidential bloc. Contemporaneous counts of the announced withdrawals at the time ran lower, at 217 and 210.
Synthesis
Whether the withdrawals caused the result is not settled by the vote counts. The 175.5%-against-164.3% gap is not a measured effect: the withdrawal and no-withdrawal groups differ by construction, and the decomposition is not identifiable from the observed data.
It is not clear how much of the outcome sits with the candidates who stood down and how much with the voters who transferred, since both moved in the same direction.
The three counts of the withdrawals (210, 217, 224) differ because one counts announced withdrawals before a deadline and the others count candidates absent from the final files. Which figure belongs in a description of what happened has not been settled.
Linked notes
/notes/vox-and-the-frame
/notes/canada-1993-right-split
/notes/preference-flows
/notes/what-counts-as-evidence
Did you enjoy this article?
Recommend it — Standard Reader surfaces well-loved writing to more readers across the network.