Populus polling analysis of the French election
By James Kanagasooriam and Claudia Chwalisz for ‘Populism across the spectrum – left, right and centre’ by Matthew Elliott.
2016 was a year of unprecedented political change. The year saw the UK voting to leave the European Union and the election of Donald Trump as President of the United States.
The upcoming French elections pose a similar change in global politics, and we could well see voters again steer away from the established parties and candidates to insurgent alternatives.
Now, on the eve of the first round of the presidential election, The Legatum Institute offers a full report offers an overview of France’s current political situation.
Read on for Populus’ thoughtful analysis, written by James Kanagasooriam and Claudia Chwalisz, which forms part of the full report.
The French presidential is a two-round system. If no candidate wins over 50 per cent, there is a run-off between the top two candidates two weeks later. The polls for both the first and second round are therefore important indicators of who is likely to win. The first round will take place on Sunday 23rd April; voting for the second round is on Sunday 7th May.
In 2012, the presidential election was a tight race between the traditional centre-left and centre-right candidates: François Hollande of the Socialist Party (SP) and Nicolas Sarkozy of the Union for a Popular Movement (now called the Republicans—LR). The polls were rather accurate for both the first and second round, predicting the order of victory and the final results within the margin of error.
2017, however, has seen the traditional duopoly of the left and right challenged by a new progressive movement launched by Emmanuel Macron, En Marche! (EM), and Marine Le Pen’s far-right National Front (NF). Only 1-3 per cent behind these two front runners are the Republican candidate, François Fillon, ex-Prime Minister under Nicholas Sarkozy’s presidency, and Jean-Luc Mélenchon (Left Front), a pan-Communist candidate who has benefited from a late surge following the television debates. These four candidates account for around 85 per cent of vote share in the first round polls.
Other candidates include Benoît Hamon (Socialist Party) and a host of minor candidates, some of whom also ran in 2012—Nicolas Dupont Aignan, Jean Laselle, Nathalie Arthuad, Philippe Poutou, Jacques Cheminade and François Asselineau—none of whom is expected to win more than 4 per cent in the first round.
Over the past few weeks, a number of key campaign trends that have taken shape, which we detail in the rest of our analysis. As an overview, these trends are:
Le Pen’s lead during the campaign:
She has been ahead in the first round polls for months and is the favourite to be one of the two candidates in the run-off. Only the most recent polls show her grip on first place weakening.
The emergence of Macron:
The former economy minister launched his movement En Marche! one year ago. As Fillon’s fake jobs scandal began to unfold in late February 2017, Macron gained momentum and has held a stable lead over all other candidates bar Le Pen. But the lack of historical precedent and his voters’ higher levels of uncertainty about their vote choice put into question the true strength of his support.
Fillon’s scandals and polling stability:
Despite the scandal regarding his wife’s alleged fake employment and the judicial inquiry into it, Fillon has been shown to have a strong support base of around 20 per cent of the population. These are primarily older, retired, Catholic and wealthy voters. It remains to be seen whether this is his ceiling or if he can gather enough support from moderate centre-right voters tempted by Macron to make the second round of voting.
Above: Figure 4a: First round poll-of-polls (up to and including 17/04/2017)
The Socialists’ collapse and Mélenchon’s late surge: The TV debates seem to have helped far-left candidate Mélenchon to the detriment of Hamon. The latter has conceded defeat by unofficially endorsing Mélenchon, who has a chance of making it to round two if his surge of support continues to grow. His chances should not be overstated, however.
There is an enthusiasm gap between his supporters and those of Le Pen and Fillon, which when taken into account, places Mélenchon firmly behind in fourth place. At the last election, Mélenchon also experienced a late surge of support in the polls, but only won 11 per cent on election day.
The First Round
On the eve of the first round of voting, little more than four per cent separates the first and fourth placed candidates, meaning that mathematically there are 24 possible permutations as to the order of the top four candidates.
Taking into account all of the campaign polls, Marine Le Pen looks to have the best chance of making the second round. She has polled either first or second in almost every first round poll for months, as is clear in Figure 4a.
The identity of her challenger is most likely to be Emmanuel Macron, but a Fillon-Le Pen run-off cannot be dismissed, even though it has been written off as less likely by much of the press. Mélenchon’s recent surge in the polls to around 19 per cent means his presence in the second round is no longer unimaginable either, but it remains less likely than Macron or Fillon facing off against Le Pen.
A second round contest without Le Pen would be a surprising outcome and would constitute an enormous polling error, given that she has polled either first or second in over 95 per cent of public first round polls in 2017.
Large polling errors have been a feature of a number of recent elections, though. Given the possibility of herding (discussed later in this chapter), Le Pen’s presence in the run-off is highly probable, rather than certain (an important difference).
Ultimately, there are numerous factors such as turnout, high levels of uncertainty, the presence of a new political movement, the TV debates, and the enthusiasm gap—not to mention any significant events that potentially occur between now and election day—which will influence the result.
Turnout and motivation
Turnout in French presidential elections is typically very high, around 80 per cent. The polls at this time in the 2012 electoral cycle were predicting turnout at around that level.
In 2017, first round polls are suggesting that turnout will be between 65-75 per cent—a significant drop compared with previous elections.
Those who voted for Marine Le Pen or Nicolas Sarkozy in 2012 are the most likely to say that they will definitely vote this time, as opposed to those who voted for François Hollande, François Bayrou or Jean-Luc Mélenchon.
Older people, the retired and those working in intermediary occupations are also more likely to say they are certain to vote, which is unsurprising given general voter turnout trends in Western democracies.
Although he is third in most polls, these factors are likely to benefit Fillon on polling day. These are the demographics which support him in the highest numbers, a reason often cited for why opinion polls leading up to the Republican primary failed to predict the scale of his victory.
Reaching older, retired people is more difficult online, the medium that most pollsters have been using during this election cycle.
At this point, four in ten people who say they are definitely going to vote have not yet made up their minds about who to vote for.
This compares with 30 per cent of those certain to vote not having decided at a similar stage in 2012. The lack of an incumbent in the race, the emergence of a new political movement and the scandal which scarred a candidate previously perceived as a ‘safe’ choice have likely all contributed to people’s indecision.
Amongst uncertain voters (Figure 4b), 37 per cent are waiting to be fully convinced, 36 per cent say there is not a candidate who meets their expectations, 24 per cent say they need more information, 20 per cent are hesitating between two candidates, 14 per cent say they will decide at the last moment and 10 per cent are waiting to watch the TV debates.
The emergence of En Marche!
One of the likely reasons why there is more uncertainty surrounding 2017’s election is the emergence of Macron’s political movement, En Marche! While the strength of Le Pen’s support has destabilised the centre-right, Macron has broken up traditional progressive coalitions in the centre
Above: Figure 4b: Voter uncertainty: Reasons why the undecided are yet to make up their minds (BVA 5-7 April)
and on the left. François Bayrou, leader of the centrist Democratic Movement, hesitated for months about running in the presidential race before deciding to back Macron right before candidate declarations were due.
The former Socialist prime minister, Manuel Valls, ran in the Socialist Party primaries on a moderate centre-left platform. After losing to the more left-wing candidate, Hamon, he later endorsed Macron. A number of other senior Socialists have gone down the same route.
Furthermore, a recent study by Cevipof, Ipsos-Sopra Steria and the Fondation Jean-Jaurès15 of 1,500 Socialist Party sympathisers found that Emmanuel Macron is their preferred candidate. 42 per cent of them are planning to vote for Macron, 38 per cent for Hamon and 15 per cent for Mélenchon.
The voter flows between the three are fluid, and numerous voters might be waiting to see which of the three (though really, which of the two, as Hamon has little chance) offers the best chance to avoid a Le Pen-Fillon run-off.
While the support from heavyweight political figures lends Macron some credibility, it also presents him with three challenges. First, he claims to be neither of the left nor the right, but thus far his key endorsements have only come from the left.
This is more likely down to career protectionism rather than lack of support on the centre-right. Fillon still stands a good chance of winning; moderate right-wing figures tempted by supporting Macron do not want to risk their careers in case of a Fillon victory. They are leaving their options open.
Second, the support of key politicians also makes it harder for Macron to distance himself from the current government, leading him to be labelled as “Emmanuel Hollande” and forcing him to spend time rebutting the suggestion that he represents continuity with the status quo.
Third, despite these endorsements, only 21 per cent of French people think that Macron best represents the ideas and values of the left,16 so he ends up in a position where those on the left see him as too right-wing and those on the right see him as too left-wing.
For all these reasons, the addition of En Marche! to this year’s election throws up numerous questions. Macron’s movement is new; he is relatively young (in French political terms).
This worries some, particularly older voters, who are concerned that chaos will ensue if he is elected. As there is no precedent, it also remains to be seen how many people will come out for him on voting day, particularly as a large part of his base is among the young, who express a lesser conviction to vote in the first place.
The fact that he claims to be from the centre whereas his strongest support comes from figures on the left may also harm his chances of winning over enough moderate centre-right voters.
TV debates and the Mélenchon surge
The televised leader debates are a new factor to consider in this election. It is the first time that there have been live debates before the first round of voting.
Traditionally, these have only taken place between the first and second round. Although not a precise guide, the TV debates for both the Republican and Socialist primaries seem to have had an impact on the outcome of each contest. Neither Fillon nor Hamon were seen as the frontrunners until after these debates, held right before polling day.
While it is difficult to disentangle the effects of the debates from other factors, the fact that millions of French people watched them and that there were such volatile swings in the polls after each of them suggests that they likely changed some minds.
The first TV debate, featuring the top five contenders—Macron, Le Pen, Fillon, Mélenchon and Hamon—took place on 20th March. 9.8 million people watched the three-hour spectacle, reaching a peak of 11.5 million viewers.
Two weeks later, a second debate with all 11 candidates was watched by only 1.75 million people. A third debate planned for three days before the election has been cancelled because not all candidates promised to participate.
The main beneficiary of the debates seems to have been Mélenchon, who has seen his support rise from around 12 per cent to 17-19 per cent in the most recent surveys.
It remains to be seen whether he can sustain this dramatic surge. Mélenchon seems to be benefiting at the expense of Hamon, unsurprisingly as the two draw deeply from the same demographic pool of voters. Whether as an early admission of defeat or because of a gaffe, Hamon also unofficially endorsed Mélenchon, saying he will vote for him if Mélenchon gets through to the second round.
As a note of caution, a recent Ifop poll highlights that Mélenchon worries 62 per cent of French people and that 54 per cent think he does not have the stature of a president. At the same point in the 2012 election cycle, Mélenchon experienced a similar rise in support to around 15-17 per cent in the polls, only to receive 11 per cent on voting day. Nevertheless, a Mélenchon rise has made the battle for second place in the first round even less predictable than it was before.
Top: Figure 4c: First round voting headline numbers (IFOP, 10/04/2017)
Above: Figure 4d: Per cent of candidate support that are “certain” of their vote (IFOP, 10/04/2017)
Figure 4e: First round voting headline numbers amended by certainty of vote choice (IFOP, 10/04/2017)
The enthusiasm gap
The final factor which makes the first round such an open race is the enthusiasm gap. As witnessed in the recent US election and the UK’s referendum to leave the EU, the varying strengths of support for the different ‘sides’ and candidates was significant. Donald Trump and Leave voters were more passionate in their support and easier to turn out to vote than Hillary Clinton and Remain voters.
In France, polls suggest a similar discrepancy.
For a long time, around 80 per cent of Le Pen’s supporters have said they are certain they will vote for her; this figure has now climbed to 86.5 per cent in the latest IFOP poll on 10th April (Figure 4d). Around 70 per cent of Fillon’s supporters— despite the scandals—have consistently remained committed in their vote for him, reaching a high of 79.7 per cent (Figure 4d). Even when the press was focused on nothing but his judicial inquiry, his support never dropped below 17.5 per cent in the polls, and has since been recovering. On the other hand, Macron’s supporters have been less sure of their choice.
For many months, only about 50 per cent of his supporters said they were certain about him; at this point, this has risen to 67 per cent. A similar number, 64 per cent, are sure of their vote for Mélenchon. By way of comparison, at this point in 2012, 80 per cent of both Hollande and Sarkozy’s supporters were saying they would definitely vote for their candidate.
Testing the “enthusiasm gap”, we have multiplied the reported poll numbers for each candidate by the percentage of that candidate’s supporters who say they will not change their mind about who they are voting for. Such treatment would increase Le Pen’s vote in the first round to 29 per cent (+5 per cent from her polling) and Fillon’s vote to 20.6 per cent (+2.1 per cent). However, it crucially reduces Macron’s vote by -1.5 per cent.
This would leave little more than one per cent difference between Macron and Fillon. As such, inadequate attention has been paid to the possibility of Fillon slipping through into the final round, despite his recent adverse press coverage. The recently considered Mélenchon scenario appears least likely when supporter certainty is taken into account— he drops a further 4.5 per cent behind Fillon (Figure 4e).
The Second Round
Given that there are four candidates polling close to, or in excess of, 20 per cent in the first round, there are six possible scenarios for the second round run-off.
Despite this massive level of uncertainty regarding who will enter the final round of voting, the polls taken in aggregate strongly suggest that one of these candidates will be Marine Le Pen.
The likeliest scenario of the possible six is a Macron-Le Pen battle. They have been neck and neck in the first round polls for a few weeks, although the space between the “front-runners” and Fillon / Mélenchon has narrowed recently. A Fillon-Le Pen run-off is the next likeliest scenario. For the purposes of this second round analysis, we assess in depth only these two most likely scenarios, and briefly cover a Mélenchon-Le Pen run-off.
The three possible second round scenarios which would not involve Marine Le Pen have not been analysed due to the paucity of polling around them. However, given how close the first round polling is, we do not rule out the “tail-risk” of a second round run-off that does not involve Marine Le Pen.
Le Pen versus Macron
While they have been more or less tied at around 23-24 per cent in first round polls, the second round polls have consistently shown Macron leading Le Pen by about 15-20 points (see Figure 4f). Is this plausible?
What we know about Macron’s core vote is that, as a representative of a new movement, his coalition is naturally built from other parties’ bases and support structures. There is a strong demographic similarity to François Bayrou’s (centrist Democratic Movement) and Eva Joly’s (Green) voters in 2012 (see Figure 4g), who collectively won no more than 15 per cent of the vote in any French region in the first round of voting.
To win with around 60 per cent of the votes in the second round, as polls suggest, Macron would need to draw support amply from both left and right. There are historic precedents for such broad-reaching support across party lines. In 2002, Jacques Chirac won only 20 per cent in the first round, but went on to get 82 per cent of the votes in the second against Jean-Marie Le Pen.
However, demographically and attitudinally, it is by no means certain that conservative, Catholic Fillon voters would back an anti-Le Pen candidate like Macron so heavily, or that left-wing voters who might share Le Pen’s protectionist and statist stance in old white working class areas would vote overwhelmingly for an economic liberal like Macron either.
For the 60-40 Macron-Le Pen scenario suggested by most national polls to be a reality, Macron would need to win three out of every four voters who backed candidates defeated in the first round.
Put another way, looking at the regional polls (by Ipsos), Macron would have to increase his vote share by between 30-42 per cent in each region of France (Table 4a). For her part, Le Pen’s second round vote share increase would need to be restricted to between 9-15 per cent depending on the region (Table 4b).
Figure 4f: Second round poll-of-polls: Macron versus Le Pen (up to and including 17/04/2017)
Figure 4g: Macron’s support is strongly correlated with François Bayrou and Eva Joly’s 2012 first round scores
Figure 4h: Second round voter flows in a Le Pen-Macron run-off (1 block = 1% round voters) (IFOP 10/04/2017)
Table 4a: Macron: Increase in support from first to second round by region in Metropolitan France (Ipsos, 14-17 March)
Table 4b. Le pen: Increase in support from first to second round by region in Metropolitan France (Ipsos, 14-17 March)
Figure 4i: Second round poll-of-polls: Fillon versus Le Pen (up to and including 17/04/2017)
Indeed this could be possible. Only 43 per cent of French people think that the National Front is a party of government, down seven points from a year ago.18 The same poll also finds that a minority of French people support Le Pen’s key campaign promises related to Europe—only 22 per cent want to leave the euro and return to the franc. The numbers who see the National Front as a danger to democracy have also grown over time—58 per cent in 2017 compared to 46 per cent five years ago.
Furthermore, if key figures in the Republican party, like Sarkozy, choose to endorse Macron rather than ‘ni ni’—neither—then this might sway a portion of undecided voters away from the temptation of Le Pen.
However, it is important to reserve some scepticism about Macron’s ability to reach far more undecided voters than Le Pen until the first round of voting actually occurs.
Le Pen versus Fillon
Until the beginning of March, the polls were indicating that a contest between Fillon and Le Pen was the most likely outcome.
Since the scandal surrounding the alleged fake employment of his wife and children came out, Fillon’s support has remained steady at around 19 per cent, | 25 failing to build the momentum that has accompanied Macron.
However, 25 per cent of those who currently say they will vote for Macron or Le Pen also say that they could change their mind in favour of Fillon according to the latest BVA poll.19 One third of those who voted for Sarkozy in 2012 have thus far refused to give a voting intention or have claimed they will abstain.
As election day approaches, Fillon will have scope to rally these voters and is well placed to gain an additional few per cent. Furthermore, for reasons to do with turnout, demographics and enthusiasm, outlined earlier in this analysis, Fillon has has a better chance of reaching the second round than received wisdom suggests.
If he does so, the size of a likely Fillon victory over Le Pen has narrowed from 20 per cent in January to only 10 per cent, on average, most recently. Some polls show an even narrower margin.
While the ‘republican front’—the collusion of left- and right-wing voters to block the National Front—has held up thus far, this would be its biggest test yet. With his socially conservative views on gay marriage and abortion and his economically liberal programme, Fillon alienates large sections of the left-wing vote who would find it difficult to ‘pinch their nose’ and vote for him. Abstention rates in the second round are thus predicted to be much higher, compared with Le Pen’s highly-motivated base.
Moreover, far-left voters have more in common with Le Pen than Fillon, particularly when it comes to the economy and Europe, so it would not be surprising to see a good number of them swing that way.
Le Pen versus Mélenchon
This remains the least probable of the Le Pen-based scenarios for the moment, but it cannot be ruled out. As our analysis later on this chapter indicates, Hamon and Mélenchon draw deeply from similar pools of voter groups. Therefore, as Hamon’s fortunes have faded, it makes sense that many have drifted into Mélenchon’s camp.
However, Mélenchon’s dramatic increase in support over the course of the campaign needs to be tempered by the fact that his support base is younger, and contains a sizeable chunk of voters who are not habitual voters. His polling strength may not translate into electoral strength at the ballot box.
There are only a handful of public polls which have tested the Mélenchon-Le Pen run-off, a recent poll from Élabe20 has Mélenchon at 61 per cent and Le Pen at 39 per cent. It seems Mélenchon would be better placed to rally voters against Le Pen than Fillon.
Figure 4j: Second round voter flows in a Le Pen-Fillon run-off (1 block = 1% of 1st round voters) (IFOP 10/04/2017)
Can we trust the polls?
We have conducted correspondence analysis on aggregated polls to better understand the similarities and differences between different parties and their respective political positioning.
This statistical technique spatially represents the political landscape in which parties operate. Where parties are presented opposite each other, this represents great difference. Where they are in closer proximity, it indicates that there are large political and demographic similarities among their supporters. The axes are approximate, but the horizontal one is left-right and the vertical one is globalist-nationalist.
The correspondence analysis highlights that Conservatives and Nationalists appear to operate in distinct political spaces. The former are most markedly defined by being 65+ and retired. The latter have their demographic roots in the bottom left quadrant—among less educated, rural workers.
Notably, the moderation of Le Pen’s vote over time is there to see. The correspondence analysis indicates that average National Front voters are similar to both private and public sector workers. Le Pen has significantly changed the party, moving its support base closer to that of other parties.
It is also clear from this analysis why Mélenchon is able to siphon votes off Hamon and 2012 Hollande supporters, given their demographic similarities of appealing to younger people, students and those in the public sector.
Polls appear to be quite favourable to Macron among far-left voters, considering their demographic dissimilarity from Macron, who appeals most to the highly-educated, professional executives, urbanites and the self-employed. His position near the centre demonstrates that he is well-placed to also gather support from a wide range of ages and the private sector.
Macron and Le Pen’s opposite positioning point to a similar ‘open/closed’ divide that we have witnessed in both the UK’s referendum to leave the EU and the US elections. Well-educated, urban internationalists are directly opposite lesser educated, rural workers in their political support.
Have French pollsters been herding?
French polls have historically been excellent, coming close to the final result for elections involving a large number of candidates.
This, however, is no guarantee of future success. One of the striking features of the first round polling for this election is how tightly clustered and similar the numbers have been from different French pollsters over the course of the campaign.
We will never know how much of this convergence, if any, is down to a reluctance to call “through the gate” any two candidates into the final round.
This possibility means that despite the surfeit of polling data, observers should be prepared for (but not expect) a result that may differ from the narrative constructed from the polls.
Figure 4k: Correspondence analysis (Aggregated IFOP polls in March 2017)
Figure 4l: correspondence analysis (Aggregated IFOP polls in March 2017)
Note: Correspondence plots of tabulations for age, levels of education, working status, geographic area on aggregated IFOP polls for March.
One of the many aspects of this French presidential cycle that has been commentated on is the level of similarity of Le Pen’s populist movement to both the leave side in the EU referendum and Donald Trump’s victorious presidential bid.
The reality is more complicated than simple equivalence or dissimilarity.
Below we analyse the similarities and differences in a “scorecard”.