The Republican Front Wins in France: So, Was the Polling Wrong?
July 8, 2024
The short answer is: I don’t think so. Let’s take a look at the numbers over the past week and then analyze what happened.
The polling on the eve of the first round1
RN: 35%
NPF: 28%
ENS: 20%
First round results (June 30, 2024)
RN: 33%
NPF: 28.6%
ENS: 20.9%
Early second round polling
RN: 190-250 seats
NPF: 140-200 seats
ENS: 95-162 seats
Late second round polling
RN: 175-205 seats
NPF: 145-175 seats
ENS: 118-148 seats
Exit poll
RN: 135-155 seats
NPF: 170-190 seats
ENS: 135-155 seats
Second round results (July 7, 2024)
Seat allocation:
RN: 142 (24.6%)
NPF: 188 (32.6%)
ENS: 161 (27.9%)
Vote share:
RN: 37.1%
NPF: 26.3%
ENS: 24.7%
One of challenges in following the polling in the French election has been that some of the polling released vote share estimates and others did seat projections. The latter was based on MRP modeling - the type of statistical modeling done for the UK election that was more or less spot on. Polls are an input into MRP modeling, but there are other factors as well. (I hope to explore MRP in another post sometime this summer.) The model allows researchers to project not just how many seats a party will win, but which ones. However, it is a probability model, which is why we seat the seat allocation projections in ranges. You may remember when I wrote that I expected now-former UK PM Rishi Sunak to retain his seat despite fears in his camp that he would lose. That’s because the MRP results I saw for his district gave him a 99% chance of winning even though the projected vote total in that race was close. One reason it was 99% was because of how late in the campaign (I think it was the second to last day) the polling was done. In any case, there was a very slight chance the projection for Sunak to keep his seat would be incorrect.
The French National Assembly is similar to the UK House of Commons and the US House of Representatives in that it is composed of individual members elected directly from constituent districts. While the UK and US use a put first-past-the-post system, France has a modified run-off system that eliminates any candidate not receiving at least 12.5% of the vote in the first round. In the second round, it is a first-past-the-post system: a candidate does not need to get a majority to win. So, the second round of the French election is analogous to the UK general election in terms of making MRP projections. Because the only candidates who win a seat in the first round are those who get over 50% - and there are only a few dozen districts this happens in - the pre-first round polling looked at national vote share to give us an idea of what the race for the second round would look like. Pollsters could not project seat allocations based on first round races.
As was true in the UK, the only thing that matters in the end is seat allocation. It is not the number of votes a party wins; it’s the number of seats. That’s true in the US as well. National party vote share does not translate into seat allocation in Congress - or for the presidential election. The French may be more used to wonky results than the British or the Americans - or maybe they just understand how their system works better. In the UK there is a debate brewing over whether to add some kind of proportional representation component to their elections after the Reform Party finished fifth in seat allocation with five seats while finishing third in the national vote - besting the Lib Dems, who got 71 seats from the election. Looking at yesterday’s results, we might expect to hear RN supporters complaining about the same thing.
While NPF surprised everyone and finished a strong first, RN was relegated to third place. Fascist disaster averted! But that is only true for the seat allocation (which, of course, is all that matters at the moment). RN finished first in vote share, and by a larger number than they received in the first round. Considering so many seats became duels in the second round thanks to the tactical withdrawals by NPF and Ensemble candidates, it’s not surprising that RN increased its share of the vote in the second round. The lower vote share for NPF and higher share for Ensemble also make sense due to how and where the tactical withdrawals occurred. But what is surprising - and troubling - is that the RN vote was almost nine points higher than LPF. So far, I have not heard anyone suggesting changing the electoral system in France but it may come.
But was the polling wrong? As I said at the beginning, I don’t think so. And that is because of two things. First, we don’t know from the first round polling or results what kind of realistic projections could be made for the eventual final seat allocations. So we can’t really say the polling on RN showed that they had a chance at winning 33% of the seats a week ago and blew it. What they did have is 33% of the votes a week ago and got 37% yesterday. But as I explained above, I think that is not really an increase in support as much as movement of vote from eliminated candidates.
Second, a lot happened in the week between electoral rounds. The MRP projections clearly show RN losing support and NPF and Ensemble gaining until we see a reversal of fortune at the exit poll. So it appears there was movement in support. However, it could have been due to the tactical withdrawals. The early MRP projection was released after the withdrawals had been made but it is not clear if in the polling that voters may have known their choices clearly. By the later projection, voters knew. That could have made a difference - but not knowing for sure what is under the hood in the early polling and projection, I can’t say for sure. What almost certainly made a difference was the antifascist campaign during the week.
NPF and Ensemble campaigned on a “Republican Front” to stop RN. As the second round approached the reality of a fascist takeover of the French National Assembly became more real for many. In fact, the BBC interviewed many French this morning who all said they were motivated by the threat of fascism. RN supporters attacked and threatened people across the country in racist and homophobic assaults, taunting that when they were in power they could act with impunity. Even anti-RN candidates were assaulted. This did not help RN in the end, thankfully. But it may help us understand why the vote projection changed so fast for RN.
In the end, this was about expectations. RN acted as if it would win an outright majority, and as the first round showed that probably would not happen they were unable to tamper down expectations, if they even tried. A fear of an RN government likely helped the massive voter turnout (now estimated to be 67%). While RN’s national vote share held steady and should give all of us pause about the party’s chances in the 2027 presidential election, finishing third in seat allocation was a crushing defeat for them.
We can expect RN to try to make the National Assembly as ungovernable as possible to help their presidential candidate - undoubtedly Marie Le Pen again - make the argument that the French need a strong leader and party to run the country. (This is probably why they won’t complain about the system; they don’t need it fairer, they plan to dominate it once they control the presidency.) Jean-Luc Melenchon has announced that NPF will nominate a prime minister for President Macron to appoint and that the bloc will not work with Ensemble. It’s unclear if any of that is true since, while Melenchon’s France Unbowed (LFI) party is the largest in the bloc, he is very unpopular with not just the general public but a lot of folks on the left. He has had a tendency to mouth off about being the left’s leader during the campaign, which his partners have largely ignored publicly. Do not be surprised if Macron picks as prime minister former Socialist President Francois Hollande, who won a seat yesterday with an outright majority against two right-wing opponents. The Socialist Party doubled its seat allocation in this election, finishing a close second to LFI on the left. Macron is probably going to have to appoint someone from NPF, and Hollande might be the safest bet. He’s no one’s first choice, that’s for sure. But that may work in trying to hold the Republican Front together for the next two years.
Key:
RN = National Rally plus allied Republicans (far right)
NPF = New Popular Front (left)
ENS = Ensemble (center)
Smaller parties and blocs are omitted, which is why results do not round to 100% where polling was in vote share or sum to 577 where polling was seat allocation. I am keeping the results in consistent order by party rather than ranking them.