France 2012 (runoff)
The second round of presidential elections were held in France on May 6, 2012. The first round, covered in extensive detail here, was held two weeks ago on April 22, 2012. The President of France, the head of state in a semi-presidential system, is elected for a five year term which is renewable once. France uses a traditional runoff system, where a candidate must win 50%+1 of the votes in the first round to be elected outright, or else the top two candidates in the first round proceed to a runoff held two weeks later. France has held eight direct presidential elections since 1965, and in none of these eight contests has a candidate ever won an absolute majority of the votes cast in the first round.
In the first round, incumbent President Nicolas Sarkozy of the conservative UMP was distanced by a narrow margin by his main rival, François Hollande of the opposition Socialist Party (PS). Hollande took 28.6% of the vote against 27.2% for Sarkozy, the first time an incumbent president did not place first in a first round ballot. The first round was marked, above all, by the very strong showing of far-right candidate Marine Le Pen, who won 17.9% of the vote. Jean-Luc Mélenchon of the Left Front (FG), an alliance including the old Communist Party, won 11.1% while François Bayrou of the centrist MoDem won 9.1%.
Sarkozy’s troubles were covered in extensive detail in the first round analysis. He saw his only possible path to salvation (re-election) in the form of Marine Le Pen’s voters. By the first round, Sarkozy’s campaign had taken on a very clear right-populist style, which he kept in the two week-long runoff campaign. He stepped up his traditional hard talk on immigration and security (key issues for FN voters) but also played to his strength as an incumbent who is credible, experienced and tested in the handling of economic crises. Sarkozy is a skilled politician, and he was able to roar back from unprecedented lows in second round polling a few months ago (60-40) to make it a fairly close race – although he never broke even 48% in runoff voting intentions. His successful rise in the polls from historic lows to more decent numbers (despite trailing) was due in part to his skill and charisma as a political leader, but also a certain normalization of things. Left-right runoffs in French presidential elections have never been blow-outs, and despite Sarkozy’s chronic weakness, it was never a very likely proposition that he would fall victim to a blow-out victory of the left. Right-wing voters, despite any unease with Sarkozy, floated back towards the candidate of their party.
Sarkozy, however, entered the runoff with a very steep road to climb. He had stabilized at a low of 45-46% support, and the more populaire nature of Marine Le Pen’s electorate made it harder for him to succeed in winning over an overwhelming majority of them. However, it appears as if his campagne au peuple worked out much better than I had expected. He managed to solidify his support with Marine’s first round voters to roughly 50-60%, gaining from voters who had considered abstention. At the same time, he was successful in holding on to a narrow plurality of Bayrou’s centrist voters, despite the very right-wing tone his campaign took on, to the disfavour of certain UMP moderates and other centrists. Despite Bayrou personally endorsing Hollande, a narrow (34-37%) plurality of his voters backed Sarkozy over Hollande (30-34%). Despite any unease with Sarkozy’s right-wing style, Bayrou’s more centre-right electorate seems to have aligned, not entirely but in part, behind the candidate of the right.
Sarkozy needed a blowout victory in the May 2 debate with Hollande to have a chance at actually winning the election. Hollande, never a strong debater, came into the debate as the underdog against Sarkozy, whose clear victory in the 2007 debate against Royal had proved the final blow to Royal’s faltering campaign. However, the May 2 confrontation ended up as a tie. Both candidates were equally aggressive against one another, and traded blows for the entirety of the three hour debate which, in French tradition, often turned into a shouting match or a trite schoolyard fight over a stolen sandwich rather than a serious and competently moderated debate about actual issues. Sarkozy might have gotten a narrow edge out of the debate: a few final polls showed that he had broken his upper limit of 47% and reached 47.5% or even 48%. Odds, however, remained heavily stacked against him.
Results: who, what, where, when and why?
Turnout, abstention and blank votes
Turnout was 80.35% (abstention was 19.65%), which represents a 0.87% increase in turnout from the first round. This is higher than in 2007, when abstention in the runoff was only 16%, but it is a fairly strong turnout for modern standards in a presidential contest in France. Since 1995, with the exception of 2002, there has been no major increase in turnout between both rounds, whereas in the 1980s, there was a slightly more significant increase in turnout in the runoff compared to the first round. As always, there was a two-way road in and out of abstention in the second round this year. On one hand, a minority but still not insignificant amount of voters who had abstained in the first round voted in the runoff. There are a number of reasons: less politicized voters who only vote for the very high-stakes presidential runoff which decides the head of state, but also personal reasons. On the other hand, voters who had voted for unsuccessful candidates in the first round and who did not support any other candidate chose to abstain in the runoff. Most new abstentionists who had voted in the first round came from Marine Le Pen’s electorate. Ipsos says that 24% abstained and 13% voted white or null. Ifop’s final poll said that 25% of her voters did not express a runoff voting intention. Between 15-20% of François Bayrou’s voters and somewhere between 5 and 10% of Jean-Luc Mélenchon’s voters did likewise.
It is interesting to put together our first map at this point: change in turnout between the first round and the second round. The most notable and interesting aspect of this map is the significant increase in every canton in Île-de-France – no canton in that region showed lower turnout in the runoff, most showed significantly higher turnout. Overall, the region has a whole saw turnout increase by a full 3% between the two rounds.
It happens that, on April 22, l’Île-de-France was out on spring break vacations (like in 2002). Schools had closed on April 14 and only returned on April 30, thus a not insignificant amount of voters were out on vacation during the first round but were back home on May 6. Indeed, the heaviest increases in turnout in the region were recorded in fairly affluent suburban areas: those most likely to be out of town for spring break? It is also interesting to note that turnout generally increased, though not by amounts as impressive, in Aquitaine, which is in the same school vacation calendar zone as Paris’ region. Otherwise (excluding Corsica, which votes for reasons fairly unrelated to the mainland), increases in turnout were more patchy. It picked up somewhat in more right-wing rural areas, but there are also rural or mountainous left-wing regions which saw increased turnout.
On the other hand, turnout generally decreased in the Rhône-Alpes region, the greater Toulouse area, most of the Centre, the inner west and continental Brittany and large swathes of Lorraine, Champagne, Ardennes, Picardy and the north. Turnout fell across the Nord-Pas-de-Calais’ mining basin, where Marine and Mélenchon both performed well. In this region, Marine Le Pen attracted a gaucho-lepéniste vote which was more likely to abstain or vote for Hollande than go for Sarkozy. Turnout also declined, though not by very significant amounts, in working-class (left and right-leaning) rural and exurban regions of eastern France where Marine Le Pen had done well (Aisne, Oise, parts of Aube and Haute-Marne, parts of the Doubs and Haute-Saône, Vosges, Meurthe-et-Moselle, parts of Moselle). A similar dynamic was likely at work in reducing turnout in parts of the southeast including the Yssingelais, the Loire, Nord-Isère, northern Ardèche, parts of Ain and the Savoies). Using the raw data from the above map, I calculated a strong negative correlation between increase in turnout and votes for Marine Le Pen in the first round (-0.49: cantons where turnout rose where likely to have at voted below average for Marine Le Pen) In the inner west, it appears as if parts of Bayrou’s electorate might have shied away from voting in the runoff, scared away by a too right-wing Sarkozy but wary of Hollande. On the other hand, it was likely mixed in with exurban Marine voters who also took a similar decision (albeit for different reasons!). There were very little links between decrease in turnout and strong Mélenchon performances: in fact, a lot of the areas where he had done particularly well for a PCF-tradition candidate showed higher turnout (Ariège’s mountains, for example).
While turnout increased, the percentages of voters who cast valid ballots actually decreased from 78% to 75.7%. In the first round, only 1.52% of ballots had been deemed invalid and blank (blanc et nul). In the runoff, this increased to 4.66%, up from 4.2% in the 2007 runoff but down from nearly 6% in the 1995 runoff. The geography of the blanc et nul vote is quite instructive. Firstly, such behaviour is actually not very widespread in urban areas: cities and towns usually show significantly lower numbers of ballots deemed invalid or blank. Rural areas of all sorts are far likelier to cast such votes. Secondly, such behaviour finds itself very limited in favourite son regions. This year, Corrèze really stood out from its neighbors with its significantly smaller percentage of votes blancs et nuls.
In this case, while such ballots were found in high percentages throughout rural areas, there were significant concentrations of high percentages of such votes in the Haute-Saône, the Vosges, Haute-Marne, Meuse, Haute-Loire, Indre and significant parts of the Pas-de-Calais, Aube, Oise, Aisne, Allier, Puy-de-Dôme and the Cher. Marine Le Pen did well in most of these departments, and in those areas her vote tended to be of a sociologically left-wing, working-class/populaire background. This is especially the case in the Vosges, Haute-Saône and the Haute-Marne around Saint-Dizier. Marine Le Pen herself cast a vote blanc, and her political home base in the Pas-de-Calais’ mining basin showed very high numbers of votes blancs et nuls. I calculated a very strong correlation of 0.58 between votes blancs et nuls in the runoff and votes for Marine Le Pen in the first round (percentage-wise, of course).
Without further blabber, the results were as follows:
François Hollande (PS) 51.64%
Nicolas Sarkozy (UMP) 48.36%
Before making any further comments, let us pause and reflect on the most significant aspect of these results. The incumbent President Nicolas Sarkozy, elected by a comfortable margin in 2007, was defeated for re-election and hence became the second French president to lose reelection, joining a club which had since 1981 been a one-man club, led by Giscard, who had been defeated by another Socialist named François – Mitterrand. This in itself is a pretty significant thing, which is made all the more impressive when one considers that Sarkozy made winning the presidency (and presumably holding on to it) the goal of his entire political career, that he had gained a significant international stature and notoriety during his presidency and was known for his political talent, skill and cunning. Just from this standpoint, Hollande’s victory is a significant political feat, regardless of one’s opinion of him or his policies.
François Hollande won, in the end, as he had been expected to since the campaign began. However, his final result proves to be quite disappointing for him. Only one poll in the final stretch showed him lower than 52.5% (Ifop’s final tracker put him at 52%), and he won ‘only’ 51.6%. Because Hollande was the overwhelming favourite right into the very last minutes, he fell victim to the old game of expectations. A great victory, along the lines of Mitterrand’s 1988 trouncing of Chirac (54-46) would have been considered an excellent result and granted Hollande much political legitimacy and increased clout in the transition and honeymoon period. A result along the lines of Sarkozy’s 2007 result (53-47) would still be seen as a success, but would be a mediocre but acceptable result for the UMP. A result between 52 and 53% would have been a nice enough victory, neither great nor particularly bad, and would have allowed the UMP to breathe a bit easier. A result along the lines of Mitterrand’s first win over an incumbent in 1981 (51.76%) would be a disappointing underperformance for the PS, “good enough” in terms of actually winning but weak in the long-term perspective. At the same time, it would allow the UMP to perk its heads up a tiny bit and smile a bit at a defeat which is not that bitter and not that crippling.
Ultimately, and it was fairly surprising, Hollande won by a margin very close to Mitterrand ’81 (in fact, even smaller of a margin). This proved to be a disappointing underperformance for Hollande, who certainly won, but didn’t really win (emphasis on the font!). In the long-term, it could prove to weaken him and shorten the honeymoon period to almost nothing. On the left, a narrow result likely tends to weaken the PS’ clout over its smaller allies and friends (Mélenchon and the Greens, most notably). On the right, while Sarkozy’s defeat gives the UMP a black eye, it is not that bad of a black eye (and it’s only on one eye, you could say) overall. It allows the UMP to perk its mood up a bit and swallow the still bitter pill of defeat a bit more easily. It makes the UMP slightly less vulnerable to either internal explosion or a solid consolidation of Marine Le Pen’s momentum behind the FN.
Compared only to the final polls, which gave Hollande between 53.5% and 52%, what happened? Firstly, it must be noted that it is not at all uncommon or even surprising from an unpopular incumbent to underpoll by a fairly significant margin. I don’t know, in this particular case, if it is a case of French shy Tory, a case of a slightly ashamed Sarkozyst who ultimately decides to “go the safe way” with the devil they know best or voters who hesitated between abstention and Sarkozy but were ultimately remobilized by Sarkozy. In the final days, with things apparently drifting away from Sarkozy with Bayrou personally backing Hollande, the first signs of speculation about the Hollande government post-May 6 and an inconclusive debate, it is possible that you had a rally-round-the-flag effect on the right, with voters remobilized and slightly remotivated behind Sarkozy, an ultimately unsuccessful last straw attempt to stand behind either the candidate as a person or his party/ideological family. The overall evolution of turnout, with the aforementioned increases in a fair number of more conservative rural or mountainous regions, might have given a tiny net boost to Sarkozy. Indeed, it appears as if those who did not vote in the first round but did vote in the runoff apparently backed Sarkozy by a narrow margin (though I haven’t found any stuff from Ifop or Ipsos on this topic).
I know that my good friend and loyal reader of this blog, Antonio V, is eager for an explanation about, in the wider scheme of things, Hollande’s underperformance and Sarkozy’s relative success in the first and second round campaign. First (and foremost), it must be noted that I never bought the 60-40 polls, not even the 56-44 or 55-45 polls. Presidential elections are serious business, and voters are much less likely to make their ballots into middle fingers to the government in power. The right may have lost the 2010 and 2011 ‘mid-term’ elections by very big margins, but not only were we dealing with local elections (played a lot, it must remembered, on regional and local issues) which are nowadays favourable to the left, but also fairly low-stakes elections where unhappy right-wingers could afford to show their displeasure with the government in power without electing, for that matter, a left-wing President in his stead. On the other hand, presidential elections being the high-stakes contests they are, pure protest votes are fairly rare (no, not all votes for Marine Le Pen or Mélenchon are ‘protest votes’ as clueless journalists like to claim!) and, despite everything, the main partisan and ideological families usually end up rallying behind their candidate. Look at results of presidential election runoff since 1965, excepting the screwed up 1969 and 2002 runoffs: they all tend to be fairly narrow. Even Mitterrand’s victory in 1988 was not a real blowout, as he only took 54% and a terribly unpopular and extremely weak opponent still managed 46% of the vote. In 1995, despite an incumbent Socialist president who was deeply unpopular and a PS which had, after 1993 and 1994, been in a terrible state, Lionel Jospin managed to get 47%. Even in 2007, the PS, with a candidate just crippled in a debate, who failed to truly convince her base and who ultimately never came close to weighing up to Sarkozy’s political skill and talent, still won nearly 47% of the vote. To use a case which is very similar to 2012, Giscard lost the 1981 election by a narrow (51.8-41.2) margin despite having been handed a huge slap in the face in the 1977 local elections and being the unpopular leader of a very unpopular government.
The same thing happened this year. Prior to the real campaign, when the quasi-campaign was a weird and terrible bastard child of off-year/mid-term protest sentiment and presidential year serious stuff, Sarkozy was down by significant margins as, while the left backed its candidate with near unanimity in opposition to Sarkozy, the right-wing base was divided and dissatisfied with the incumbent. From the ridiculous heights of 60-40 and even 56-44, a significant narrowing of the gap was a natural phenomenon. True enough, Sarkozy managing to close the gap down to 51.6-48.4 is surprising, but I always had a tough time believing that somebody who has Sarkozy’s charisma, stamina and political skill, talent and cunning would go down by a very big margin.
Secondly, it must be noted that a vote for Sarkozy is not necessarily a Sarkozyst vote, if you get my gist. The election was pretty surely a referendum on Sarkozy – we’ll come back to that – but it is certain that not all voters voted in such fashion. Ipsos’ exit poll showed that 54% of those who voted for Sarkozy in the runoff did so because they wanted him to be President but you still had 46% who voted for him because they didn’t want Hollande for President. 60% of Bayrou-Sarko voters and 70% of Marine-Sarko voters explained their votes in such a way, but you still had a fairly significant 35% of Sarkozy’s first round voters saying they had voted for Sarkozy because they didn’t want Hollande to be President.
This may be a presidential campaign, and we’re all fed the stories about how it’s only a personality contest, but let’s remember that you still have a fair number of partisan voters who vote for their’s party’s candidate in all but the worst of circumstances. Out of Sarkozy’s voters, there were not only hard-core Sarkozyst who were in love with him and enamored by the entirety of his government’s record over the past five years. There were also loyal right-wing voters, who may not have been the most hard-core of Sarkozysts, but who are ideologically and/or traditionally right-leaning or conservative. In the end, whatever their problems with the nuts and bolts of Sarkozy’s record, they were either worried by the prospect of a left-wing victory, returned to their traditional right-wing roots or rallied behind an incumbent who they might consider as imperfect but ultimately – perhaps – a tested, experienced and competent leader. In the debate, Sarkozy, overall, tended to play a lot on the aspect of being a tested, experienced leader.
Finally, as Mitterrand once said, France might be a right-wing country which sometimes elects left-wing governments. While I find the idea of classifying a country of 46 million registered voters as left or right-wing to be downright stupid, there might be a certain truth to it. Perhaps France has a politically conservative inclination? Yet, I shy away from such simplistic and reductionist partisan interpretations. Perhaps France is more structurally right-wing than it appears to be, but it cannot be a convincing explanation.
This election has often been branded by the media and observers as a referendum over Sarkozy more than anything else. In this way, Hollande did not have to worry as much about his own personal image or even the details of his platform, but could instead stand in an enviable spot as the anti-incumbent to an unpopular incumbent, the change candidate against the unwanted status-quo. Hollande exploited this benefit to its maximum, and this advantage allowed him to steer clear of any significant trouble during the campaign. He was never seriously hurt by his comparative weaknesses on issues such as foreign policy and economic/fiscal policy. Even his Achilles’ heel – his lack of a strong ‘presidential stature’ never really hurt him.
Hollande also milked all the benefits of the ‘normal President’ creed he took on from day one. Against an incumbent nicknamed l’agité and known above all for his unconventional, erractic and off-the cuff style, a remarkably short temper and a strong penchant for show and le bling-bling; it appears as if voters were thirsty for a normal president who returned the presidency to its more distinguished traditional stature. Sarkozy was definitely hurt a lot by his image – perhaps even more than his actual policy. The bling-bling for which he became (in)famous contradicted his populist creed adopted in 2007. His proximity to wealth, big money and tycoons hurt him in a country in which money and excessive wealth is generally considered a social taboo or at least frowned upon. Hollande did not promise voters a superhero president, but rather a ‘normal’ president. His normal image allowed him to appear closer to voters, more connected to their problems and perhaps more amiable than Sarkozy (though I’d wager most voters would not fancy having a beer one-on-one with any of the two, for different reasons).
This campaign, from a more political and electoral standpoint, was marked by the clash of two campaign styles. François Hollande, like Mitterrand in 1988, aimed to win the election ‘in the centre’, which means without adopting overly populist rhetoric on economic matter and not running away from institutions such as the European Union. Hollande, however, also allowed himself to go for a more anti-incumbent and more left-leaning sideshow, closer to Mitterrand’s 1981 changer la vie (without the youthful naïve optimism embodied in that famous creed) than to his 1988 France unie centrism. On the right, Nicolas Sarkozy’s avowed campaign strategy was to win through a campagne au people (populism, in this case right-populism). After the first round, I commented on the weaknesses of the Sarkozyst tactic, which found itself hindered by Sarkozy’s record as an incumbent (generally, incumbents do not hope to win re-election on a populist path but generally through consensual centrism emphasizing their experience).
I was probably correct in my observations after the first round, but admittedly the runoff saw some rather different dynamics at play, some of which were probably to be expected.
Firstly, it is now debatable whether Hollande actually ‘won in the centre’. If he did, it was narrow. Even then, ‘winning in the centre’ in his case only takes into account his rhetoric and his campaign’s rationale. It appears as if, when voters were asked, Hollande did not win because of his reassuring moderate image but rather because of his position as the anti-incumbent. Ipsos’ exit poll confirms the narrative about the election being a referendum on Sarkozy. 55% of Hollande’s second round voters said they voted for him to prevent Sarkozy’s reelection while only a minority (45%) said they voted for him because they wanted him to be President. A full 43% of Hollande’s first round voter said they voted for him to defeat Sarkozy, a huge majority (71% and 75%) of Mélenchon and Bayrou’s voters (who voted for Hollande in the runoff) said that defeating Sarkozy was the main reason they backed Hollande.
From a short-term electoral standpoint, the election being a battle between Mr. Sarkozy and Mr. Not-Sarkozy played to Hollande’s advantage. Now, this could prove to be a major weakness for Hollande who was perhaps not elected on his record but rather in opposition to his rival’s record. In this case, the election likely proved to be about Sarkozy’s defeat rather than Hollande’s defeat. This could shorten Hollande’s honeymoon period and lay the groundwork for future nightmares. Hollande enters the Elysée with much less “legitimacy” than he would have if he had won with 53%.
I have to admit that the success of Sarkozy’s campagne au peuple in the runoff, which got confused with a campagne au Front national, proved more successful than I had anticipated it. I had much reserves about the chances for success of such a strategy, given that the FN’s electorate in 2012 was sociologically fairly working-class or lower middle-class, two social categories with which Sarkozy had performed extremely weakly with in the first round. Ultimately, Sarkozy’s performance with Marine Le Pen’s first round voters proved insufficient but still a very good showing for him considering all the things going against him.
According to Ipsos, 50% of Marine Le Pen’s first round voters voted for Sarkozy against 24% who did not vote, 13% who voted for Hollande and 13% who cast an invalid or white vote. According to Ifop, 54% of her voters opted for Sarkozy against 21% for Hollande and 24% who did not express a preference.
However, his right-populist and, some would say, FN-lite campaign in the runoff did not really hinder his chances with Bayrou’s reduced (and more traditionally centrist/centre-right) electorate. Despite their candidate voting for Hollande, 40% backed Sarkozy against 27% for Hollande, 17% who did not vote and 17% who cast an invalid or white vote (Ipsos). Ifop, on the other hand, did show a much more divided centrist electorate: 41% for Sarkozy, 40% for Hollande.
What seems to have happened in the runoff which could serve the explain Sarkozy’s stronger-than-expected showing and Hollande’s Pyrrhic victory is a certain shift away from pure anti-incumbent dynamics to a more traditional, “blank slate” dynamic. The first round was clearly fought on an anti-incumbent dynamic. The runoff, which has always been about ‘eliminating’ after ‘choosing’ two weeks earlier, did not quite shed all the anti-incumbent dynamic (to Sarkozy’s chagrin) but the dynamic edged closer towards a more traditional kind of dynamic which can be styled a “blank slate” dynamic. The election became even more personalized, and Mr. Not-Sarkozy slowly became perceived, in the eyes of voters, as his true self.
This in and of itself was always dangerous for Hollande, whose Not-Sarkozy image was always much stronger than his personal image. It is not to say that he’s a bad candidate, in fact he proved to be a much stronger candidate than anyone would have anticipated a year ago. However, he was more vulnerable to Sarkozy’s antics when voters started judging him based on his platform, his personality and his ideas. Sarkozy certainly pounded on the “risk” which Hollande’s promises of “change” carried. For the small but influential minority of fledgling voters, there was probably an inclination by these voters to go for the ‘devil they know’. This dynamic was probably most pronounced with some of Bayrou’s traditionally centrist and Christian democrat/UDF voters and some far-right voters. Polls report that Sarkozy narrowly won late deciders, though a large majority had decided long before May 6.
There was a bit of a “blank slate” dynamic at work, where both partisan and ideological families rallied around their respective candidates. On the far-right, there was certainly a “blank slate” effect whereby voters opted to back a candidate who, despite a poor record, was ideologically closer to them and shared their similar inclinations for authority, order, nationalism and so forth. This is what a blank slate effect refers to: voters are compelled, by the high stakes of the election, to cast a vote which is based more on their own ideological leanings than any personal disagreement with the candidate of their ideological tradition. An incumbent is sometimes forgiven for his wrongs and for his mistakes by voters hailing from the incumbent’s wider political family. Ultimately, the FN naturally tends to be much closer to the mainstream right in its current incarnation than to to mainstream left in its current incarnation. There is less room in runoffs for protest votes, and much more ideological and partisan votes, as the parties find their solid bases. There was certainly an anti-incumbent effect at work here, and there were quasi-protest votes, but in large parts, the election ended up being fought on traditional partisan and ideological bases.
Anti-Sarkozysm remained a powerful motivator, but ultimately the high waves of anti-Sarkozysm were limited to the ideological left and did not really break the levies and flood over into the centre and the far-right. On the left, however, the waves remained quite high. The best proof is probably the very strong vote transfers towards Hollande coming from Mélenchon’s voters. Despite any weariness or dissatisfaction towards Hollande, the vast majority of Mélenchon’s voters were swept up in the wave of anti-Sarkozysm. Ipsos tells me that 80% of his voters backed Hollande, against only 6% for Sarkozy. Abstention and invalid/white votes were far more limited: only 10% and 4% respectively. There were only very few strongly left-wing voters who, through ideological convictions, refused to vote for the “wimp” Hollande, even over Sarkozy. The mood of the left – the entire left – was radically anti-Sarkozyst. It mobilized the entire left-wing electorate, not just the core PS voters, but even PCF and Green sympathizers. Hollande being able to tap in to the powerful anti-Sarkozyst forces on the left with much ease and keep them mobilized proved to be a major factor in his victory.
Exit Poll Analysis
As I did following the first round, I broke down Ipsos’ study on the sociology of the electorate. I also looked over Ifop’s similar poll, which gives some different results in spot. By way of comparison, the chart below (which is the Ipsos study) compares the performance of each candidate to their performance or that of their party’s candidate in the 2007 presidential runoff, as measured by the 2007 Ipsos sondage jour du vote. For comparisons further back in time, I used Ipsos’ archived data on the 1995 presidential runoff and the Sofres’ archived data on the 1981 presidential runoff. I would advise much caution in the interpretation of some of these results and especially their comparisons with 2007, given the problems inherent in small samples and differing definitions and samples in 2007 and 2012.
There was no significant gender gap, but there was a wider age gap. Hollande performed best with young voters, which have always leaned to the left, while Sarkozy only retained dominance with seniors, winning the 60+ vote by a 59-41 margin. Ifop corroborates the main trend, though reports a less impressive left-wing advantage with young voters (only 56-44 with under 35s) and a less impressive Sarkozyst performance with seniors – 55-45 with voters aged above 65. I would not put too much behind the changes compared to the 2007 results.
The artisans et commerçants (artisans, shopkeepers, small business owners) confirmed their very strong allegiance to the right, with Ipsos pegging their vote at 70% for Sarkozy and Ifop at 67% for Sarkozy. Both of these numbers would represent a significant loss for Sarkozy compared to the 2007 election in which he won a full 82% of their votes, a trend which could signify a certain unease with a president who has proven too elitist and not reformist enough.
Indeed, by their markedly right-wing leanings, artisans and shopkeepers express a strong opposition to the left’s penchant for state intervention and its proximity to salaried employees and trade unions. Artisans and shopkeepers are not generally of the “upper classes”, rather they are a traditional petite bourgeoisie which has lived in constant fear of proletarization and has cultivated a visceral opposition to the left’s historical traditions rooted in Marxist collectivism.
Not quite economically liberal capitalists, the petite bourgeoisie of merchants and shopkeepers is nonetheless fiercely individualistic, egalitarian and instinctively conservative if not reactionary. It founded the base of the Poujadist movement in 1956, and has since been one of the FN’s most prominent factions though they are particularly receptive to Sarkozy’s unorthodox mix of individualism, conservatism and weird hybrid of colbertisme and libéralisme. The 2007 and 2012 rhetoric of individual responsibility, la valeur travail (the ‘value’ or ethic of work and labour) and authority found its most enthusiastic reception with these voters, though some might have fallen out with Sarkozy since 2007 because of his more elitist and liberal penchant for the bling-bling and his improvised economic policies, at times liberal and favouring the privilégiés (privileged upper class) at other times, more Gaullst in its statist and colbertist leanings.
The cadres and professions libérales are a very socially and politically diverse grouping, thus grouping them into a single ensemble – while inevitable (because France hates psephology) – is quite reductive. Overall, Ipsos tells us that Hollande won these professionals with 52%, while Ifop found a much wider 56-44 gap in Hollande’s favour. When breaking down this wide category, it is likely that Ipsos is closer to the mark. Whatever their vote this year, they posted historic numbers for the left, the culmination of a left-wing trend. Sarkozy had won them 52-48 in 2007, which had already been a very weak showing for a right-wing candidate. In 1995, Chirac had won about 65% with these voters and in 1981, an election whose final numbers proveed eerily close to this year’s results, Giscard had handily defeated Mitterrand with these voters, taking 62% of their vote.
If these professionals can be summarized, they are clearly the members of the broad so-called ‘elite’. They are highly educated professionals, upper middle-class in their standing. They range from professions libérales in the private sphere including independent doctors, pharmacists, lawyers, notaries or engineers to upper-echelon public servants (managerial) to secondary education teachers, professors and researchers to artists and journalists to managers, supervisors and other cadres in the private sector. Hollande likely clearly dominated the field with public sector managers and professionals, but also teachers, professors and other researchers. On the other hand, we can probably assume that Sarkozy dominated the field with professions liberals whose independent, non-salaried status makes them hostile to the left and closer to the right’s traditional values of individual responsibility and initiative. Indeed, a poll by Ifop had shown that Sarkozy was by far the favourite with doctors and pharmacists. While some lower-level managers and cadres in the private sector might have voted for Hollande, we can assume that the higher managers and assimilated professionals of the private sector, either because of their wealth or professional status, remained strongly Sarkozyst.
The right’s gains in historically left-wing CSP- categories (employees, ouvriers) have been counterbalanced with the left’s gains over the right in CSP+ categories, most notably these cadres and professions libérales. The progressive transformation of the PS from an old style, SFIO-ish neo-Marxist and working-class party into a socially liberal and social democratic party of the centre-left has been accompanied with major left-wing inroads with professionals, affluent salaried middle-classes and above all with individuals with a so called large ‘cultural capital’ (higher education, intellectual jobs and so forth). These voters – if they can be summarized given their heterogeneity – are socially liberal, pro-European, middle-class, urban and receptive to rhetoric about equality, tolerance and internationalism.
The professions intermédiaires turned heavily against Sarkozy. According to Ipsos, Hollande won them 60-40 while Ifop shows a smaller but still significant gap in the left’s favour at 57-43. In 2007, Sarkozy had lost them by two points (51-49) to Royal. The professions intermédiaires are a wide category, including both private and public sector workers, and ranging from school teachers to foremen to qualified technicians to sales representatives and other intermediate-grade jobs. They are, in their outlook and status, a middle-class between the upper middle-class of cadres and the working-class of ouvriers and employees. Their political outlook cannot be classified with ease. Teachers, social workers and other intermediate-grade civil servants are a core redoubt of the PS and take a social democratic outlook on politics, being closely attached to the values of tolerance, equality and social justice. On the other hand, private sector intermediate-grade salaried employees are likely more right-leaning, and are more concerned about bread and butter issues.
The dramatic shift to the left of this category reflects a finding we had picked up on April 22: Sarkozy has lost considerably with salaried middle-classes of all stripes, and the professions intermédiaires are the best representatives of this type of electorate. He was likely the inadvertent victim of the economic crisis, which has had a large and negative effect on these salaried middle-classes: job loses, the first signs of fiscal austerity, declining purchasing power and rising fuel prices. His controversial pension reform was probably coolly received by these voters, while civil servants and teachers in this category turned virulently anti-Sarkozyst in reaction to job losses in education and his confrontational attitude with public sector unions and the like.
Sarkozy lost employees, as in 2007, but there is disagreement between our two pollsters by how much: Ipsos says he lost them handily, 56-44, while Ifop says he lost them by a margin only marginally bigger than that by which he lost them in 2007 – 52-48 (against 51-49 in 2007). Ipsos is probably closer in this case. Once again, we are dealing with a broad category, which has been generalized to lower-echelon employees and so forth. The ranks of this category range from lower-level public servants (agents de services), clerks, secretaries, police officers, administrative employees, cashiers, clerks, salesmen but also personal service workers (hair dressers, nannies, bartenders, restaurant servers, hotel clerks, concierges).
Their political inclinations are diverse, and probably tend to divide based on their sphere of work (public or private). Public sector employees can be counted on to have been very solidly behind Hollande, but private sector employees might have tended to split more in Sarkozy’s favour. Indeed, private sector and non-salaried employees tend to be far more favourable to the UMP and the FN; they are very concerned about unemployment, job security, purchasing power and cost of living issues. Despite Sarkozy’s poor record on most of these issues, they likely remained more or less resistant, this year as in yesteryears, to the left. Sarkozy’s quasi-nationalist populist appeal might have, for these employees, counteracted Hollande’s consensual anti-Sarkozyst centre-leftism. Disproportionate amounts of this latter kind of employees might be found in the périurbain subi, which we will come to when we look at good ol’ maps.
Ouvriers (manual workers, both qualified and unqualified) are probably, out of all these socioprofessional categories, those about which the most ink has been spilled. Traditionally and historically, ouvriers have formed the backbone of the French left, which, in the glory days of the 50s and late 70s used to command the support of about seven in ten workers. A strong tradition of socialization in a Communist milieu in the immediate post-war era maintained strong familial links of left-wing (and oftentimes, Communist) political orientation. However, since Mitterrand’s election in 1981 and especially since the 1990s, the left has been alarmed at the pace at which their old backbone have been deserting them and flirting for anti-system options, be it the unconventional far-left of Arlette and Olivier or the far-right of Jean-Marie and his daughter. There is a feeling that the left has abandoned its working-class roots and has shifted its style, rhetoric and strategy towards gentrified middle-classes, salaried public employees and the bobos. Indeed, the PS’ style since 1983 has been edging towards either feel-good consensual, moderated toned-down centre-leftism or New Left rhetoric about social justice, equality or tolerance. The Marxist rhetoric about the class struggle, the proletariat and even the mitterrandien creed of changer la vie was left on the side of the road, ready to be picked up by parties to the left or right of the PS.
Le Pen started picking up these votes a plenty in 1995, with his rhetorical shift from the old Poujadist anti-communism and anti-statism of the early 1980s towards his brand of right-populist nationalism which appealed to voters who increasingly felt marginalized, sidelined, abandoned and invisible in the winds of globalization and in the changing face of the PS. The working-class vote which Le Pen and his daughter picked up was politically diverse, but it was a true protest vote – against the bipartisan political system edging (until Sarkozy) towards mushy centrism, globalization, economic stagnation or decline and political and social marginalization. Later on, Sarkozy’s unconventional brand of respectable right-populism in 2007 proved to have a strong appeal to working-class voters (he lost ouvriers by a historically small margin of 54-46).
The ouvriers are a diverse bunch too. It includes qualified and unqualified manual workers but beyond that includes all types of workers from diverse industries: heavy industry, light industry and small manufacturing, mining, metalworking, textile, steelworks, nuclear energy, construction or public works. Carpenters, masons, construction workers, electricians, plumbers, farm workers, gardeners, truck drivers, taxi drivers, butchers, bakers, longshoremen and cheminots all fall under the broad hat of this socioprofessional category. Their social heterogeneity has also resulted in a certain political heterogeneity. Even in the peak days of the late 70s, a sizable 30% of workers were right-leaning, a fact which Mélenchon keenly pointed out during the campaign, also taking the opportunity to call these three in ten right-wing working-class voters “not too bright”.
Hollande won ouvriers with either 58% (Ipsos) or 57% (Ifop) this year, making them about 5-7% more left-wing than the country. 57-58 is neither bad (Royal’s 54-46 was bad, although they remained 7% more left-wing than the country) nor exceptionally good. In fact, it is quite underwhelming when you consider that Hollande won. In 1981, when the margin was similar, Mitterrand had handily trounced Giscard with ouvriers, 67 to 33%, which would mean that Hollande lost about 9-10% of Mitterrand’s 1981 numbers with these voters. In 1995, according to Ipsos, Jospin beat Chirac with ouvriers by a 65-35 margin. Consider, however, that Jospin lost that same election by about 5.5%. We are a long way away from the polls at the turn of the year which let us believe that Hollande could return to the peak of the 1970s with ouvriers and destroy Sarkozy with some 65-35 or 70-30 gap.
Hollande thus underperformed with the working-class, again, while Sarkozy – considering that he lost 4.6% nationally – held up comparatively well. In this case, it is likely a bit of ‘blank slate’ effect where voters must have forgiven Sarkozy some of his mistakes and preferred his tough, “steady hand on the wheel”/devil you know populist rhetoric. It also reflects that protest voting was less prevalent in the runoff than the first round or the 2010-2011 mid-terms. Our maps will teach us instructive lessons about the vote transfers from Marine’s working-class first round base and where Sarkozy’s resistance with the working-class was strongest.
To complete this socioprofessional breakdown, we shift to Ipsos and Ifop’s questions on the employment sector of the interviewee. A traditional private-public gap emerged. The public sector was solidly behind Hollande, which he won with either 63% (Ifop) or 65% (Ipsos), a gain of 8% over Royal on the 2007 Ipsos data. Sarkozy’s policies vis-à-vis education, pensions, healthcare and spending reviews (RGPP) proved extremely unpopular. At times, he gave the impression of picking fights with the public sector and their unions. On the other hand, private sector salaried employees were more divided: they backed Hollande 52-48, after having voted for Sarkozy with 53% in 2007. Their vote is more varied, and can hardly be described as relatively politically homogeneous as that of the wider public sector. However, in general, the private naturally tends more to the right.
The preoccupations, concerns and values of the public and private are similar in some aspects but also quite different in other regards. Public sector employees, notably teachers and healthcare professionals, tend to be concerned about the decline of public services and the “hollowing out” of the state in some areas of social action. Naturally, such concerns lead them to favour the left. The private is more concerned about job security, less concerned about the changing role of the state in society. Finally, Sarkozy heavily dominated with self-employed workers (indépendants), entrepreneurs and employers. He won about 60-61% of their vote, which would be down quite a bit on 2007 (when he won them 77-23!). Like artisans and shopkeepers, these self-employed workers are far more individualistic and have always been extremely resistant to the left and the associated ideas of state intervention.
Unemployed voters backed Hollande 62-38, but for some reason, Sarkozy performed rather strongly compared to 2007 with unemployed voters. These voters traditionally tend to be extremely anti-incumbent, but Sarkozy managed a surprisingly strong result with them this year. The result of a larger number of unemployed voters in 2012 compared to 2007, widening the political composition of this electorate?
Ifop and Ipsos conflict when it comes to vote by diploma. Ipsos shows Sarkozy strongest with those with basic non-BAC certifications (BEPC, BEP, CAP, CEP) and those with a BAC +2. Conversely, Hollande performs strongest at both extremes and in the middle: he does extremely well (59-41) with those with no diploma, and dominates (55-45) with those who have a BAC and also with those with a BAC +3 or higher. Ifop, conversely, tells me that Sarkozy won (53-47) those with no diploma, did worse (55-45) with those who have the BAC and lost narrowly (51-49) with all other categories. Ultimately, education all boils down to income levels and socioprofessional status. In this day and age, the left’s support probably forms something of a parabolic curve, doing best at both ends (those with no diploma and those with post-secondary qualifications).
Ipsos broke down the vote by income levels, though with four large income categories we are a long way from the breakdowns we see in the United States. Ifop didn’t even ask based on income, and Ipsos only started asking based on quantifiable set categories this year; an attitude which reveals the French psyche’s attitude towards wealth and money. The results are hardly surprising: Sarkozy’s support increased as the voter got wealthier. He only won the top category (those earning 3000 euros or more), with 56%, while Hollande had a cross-class appeal with the poorest voters but also the broader middle-class, with whom he made the strongest gains against Royal’s 2007 result (when Ipsos measured income based on interviewee self-identification as rich or poor).
The breakdown of the vote by partisan and ideological self-identification are hardly surprising or interesting. Sarkozy lost all of the minor but still fairly sizable inroads he made with about 10% of left-wing voters in 2007, but Hollande did not make significant gains with right-wingers. FG, PS and Green voters backed Hollande with huge numbers, as did UMP voters for Sarkozy. According to Ipsos, centrists and MoDemites split 62-38 in Sarkozy’s favours, but Ifop has them nearly even at 52-48 for Sarkozy. About eight in ten FN sympathizers who voted did so for Sarkozy.
What was the most important voting determinant? If Ipsos is to be believed, then it was, yes, religion. According to Ipsos, if only Catholics could vote, then Sarkozy would have won reelection handily with 57% of the vote. Even more impressive, regularly practicing Catholics confirmed their strong allegiance to the mainstream right, with 76% of them voting for Sarkozy. The Catholic ethos and values associated with Catholicism has always tended to favour the right. Attachment to values such as the family, social order, the respect of hierarchy and authority, entrepreneurship and individual initiative has bred political conservatism. The left’s shift towards moral liberalism, with Hollande openly supporting gay marriage, likely played some role in further motivating some conservative regularly practicing Catholics to vote for Sarkozy, who at various times during his campaign played on the ”Christian identity” of France.
Occasionally practicing Catholics backed Sarkozy with 62% of the vote, while non-practicing Catholics backed Sarkozy with 54%. Despite the left’s strong inroads into the old Catholic terrains of western France, Brittany and the Massif Central; those voters who identify as Catholics have retained their traditional loyalty to the right, even though those who are ‘Catholics in name only’ tend to be less homogeneous in their voting habits. Indeed, the left’s gains in regions such as Brittany or even the Massif Central have not been the direct result of devout Catholics retaining their religious traditions but switching political allegiances, but rather the result of the secularization of the regional culture and the increasing amount of secular, agnostic or atheist voters in these regions. A geographic analysis confirms the same trends: the most devoutly Catholic regions of departments such as the Ille-et-Vilaine, Morbihan, Manche, Mayenne, Vendée, Lozère, Aveyron, Cantal or Haute-Loire remain solidly left-wing. Rather, the left’s major inroads have come from secularized or secularizing regions – which often tend to be in the sphere of influence of a large urban centre (Rennes, Laval, Caen, Rodez, Aurillac or Le Puy for example).
Those who claimed another religion backed Hollande with 63% of the vote. These other religions have historically included Jews (who tend to be rather right-leaning as of late) and Protestants (who are split between solidly left-wing Calvinists in the south and solidly right-wing Lutherans in Alsace-Moselle). However, in recent years, an increasingly large number of those claiming another religion have been Muslims. Though their cultural ethos is very socially conservative, their social status has led them to favour the left with overwhelming numbers (Hollande likely won over 90% of the Muslim vote).
At the other end of the spectrum, those with no religion gave Hollande 68% of the vote. The irreligious vote has always been very heavily left-wing, because the cultural ethos associated with ‘irreligion’ is naturally left-leaning. Traditionally, the values of social justice, social solidarity, tolerance, equality and moral liberalism has been at the core of the cultural ethos of non-religious voters. However, as the ranks of this electorate have swelled since the 1970s, their vote has become less homogeneously left-wing. In the 1970s, the lack of religion was strongly associated with the left, as it symbolized a rebellion against the established cultural and religious order of sorts. Since then, with the secularization of society, the lack of religion has become far more acceptable and also far more common. Their vote remains solidly left-wing, but not as solidly and homogeneously left-wing as in 1974.
The difference between the two extremes – regularly practicing Catholics and those with no religion was 44% this year. In 1974, the difference was 62% – Giscard won 80% with regularly practicing Catholics while Mitterrand won 82% of the vote with non-religious voters.
The overall map of the runoff is similar to that of the first round. Hollande has a wide geographic base, and raked in some very strong performances in some core left-wing strongholds of the old southwest (breaking 60% or at least 55% in the bulk of them) but also in his native Corrèze (64%).
As in the first round, Hollande’s map, when set against that of Mitterrand in 1981, is much more western and southwestern. All five departments in Brittany backed Hollande, even traditionally conservative Morbihan. In Mitterrand’s era, the left’s only base in the region had been the Côtes-d’Armor and isolated bastions in the Finistère and Loire-Atlantique. Even though Hollande retrieved the old Socialist bastions of the north, Picardy, Upper Normany and Meurthe-et-Moselle, which Royal had lost in 2007, his strongest or more impressive performances are not to be found in those old left-wing strongholds. Rather, they are to be found in the Southwest and Massif Central – most significantly in departments such as the Aveyron, Cantal, Haute-Loire or Lozère, whose political histories have been ones of right-wing dominance.
They are also to be found in urban areas. Hollande won over 55% of the vote in Paris, becoming the first PS candidate to win the French capital. He also won the traditionally moderate bourgeois capitale de Gaules, Lyon, with 53%, when Jospin had won only 41% of the vote in that city in 1995. Lyon, the longtime “capital” of the UDF awarded a significantly higher percentage of the vote to Hollande than Marseille (which he won by the skin of his teeth), the city of Gaston Defferre and a city with long history of Socialist institutional dominance.
He raked in over 60% of the vote in western cities such as Caen, Rennes, Nantes, Quimper, Brest, Niort, Poitiers and La Rochelle. Rennes, Nantes, Brest, Niort or La Rochelle are hardly surprising results, but consider the fact that Caen elected its first left-wing mayor in ages only eleven years ago!
Hollande benefited from a very strong favourite regional and favourite son effect in his semi-native Corrèze. He won 64.9% of the vote, while Jacques Chirac had won 61.4% in 1995. Favourite son votes are much stronger and noticeable in more rural and isolated department, where there is a clear advantage in having a president who is a native son. Locals hope that having their native son as President will encourage local economic development and that the department will be left advantaged by the president’s policy. As a point of comparison, Nicolas Sarkozy never really benefited from any favourite son effect in his urbanized and economically integrated department, the Hauts-de-Seine – in fact, Hollande won 49.5% of the vote in Sarkozy’s native department. Like Chirac, Hollande’s favourite son effect was not confined to Corrèze. It also boosted the natural left-wing vote in surrounding departments, forming a perceptible halo in regions of the Lot, Dordogne, Haute-Vienne, Creuse, Puy-de-Dôme and Cantal which border Corrèze.
Hollande won excellent results in traditionally right-wing departments of the southern Massif Central. He won Cantal, the core of the old pompidolie and a right-wing stronghold for ages, with 51.8%. He took 54.4% in the Aveyron, traditionally a conservative department. In Haute-Loire, he won 51.4%. In the Lozère, which had for ages been a rock-solid conservative stronghold (outside the solidly left-wing Cévennes), Hollande lost by a mere 45 votes to Sarkozy. In all of these cases, Hollande was boosted both by very strong performances in the core left-wing bases of these departments, but also by fairly impressive gains in urban areas and in evolving, secularized regions. The Catholic heartlands of the Aubrac, Pays d’Olt or Margeride remained solidly behind Sarkozy, but Hollande scored impressive results in the demographically evolving Plateau of Saint-Flour, Puy-en-Velay plateau or the Grands Causses (Aveyron).
Hollande was triumphant in all but a small handful of France’s major urban centres. This is the logical culmination of a series of profound transformation in French urban politics. The rural-urban divide and the urban physiology of France have really been turned on its head in the past decades, a shifting reality which has obviously carried deep political repercussions. To talk of a rural-urban cleavage in traditional terms is increasingly misleading in France and other Western countries. Traditional rural areas no longer form a sizable share of France. Areas which may appear, misleadingly, to be “rural” are in fact exurban or suburban areas, where locals do not work in small businesses in their commune of residence but rather commute distances of varying length to their place of work in another commune, oftentimes a major city but also smaller, mid-sized towns which serve as employment centres and focal points of economic activity and social exchange.
In France’s largest cities, the old opposition between bourgeois and proletarian neighborhoods has been progressively weakened albeit not entirely erased. Wealth is no longer a phenomenon constricted to the old central bourgeois neighborhoods, while socio-economic changes since the 1970s mean that the old white working-class, properly speaking, has generally abandoned their old neighborhoods of core urban areas in favour of the suburbs or exurbs. The bobo phenomenon has been a direct result of gentrification of old working-class quarters, a process which is most pronounced in Paris (especially the old working-class east side) but also in other large cities including Lyon, Marseille or Lille.
Rising property prices in core urban areas and old immediate suburbs have meant that those who can afford to live in those municipalities tend to be fairly well-off, highly educated middle-class professionals. At the same time, however, the urban core often tends to have a lower median income than its immediate surrounding suburbs. It is not a population of extremely affluent conservative bourgeois who make up the bulk of the city’s population (example: Paris), but rather younger, middle-class professionals/CSP+ who are fairly well-off – not “filthily rich” (as some would say!), but not poor or deprived either.
Large urban areas also tend to have a younger population, a result of a number of factors including desire for a hip-bobo urban lifestyle, proximity to centres of education and research or a population of younger middle-class professionals and cadres. The lifestyle, culture, values and social makeup of most urban areas in France and around the world tend to be favourable to the left: a mix of secularism, urban progressivism, environmentalism and pro-Europeanism are all urban values which move urban areas, in general, closer to the left.
In France, core urban areas cannot and should not be the subjects of generalization. The reason why Paris voted for the left is quite a bit different from the reasons for which Brest or Le Havre voted for the left. Paris, Lyon, Grenoble, Toulouse, Nantes, Bordeaux, Caen, Rouen, Nancy and Dijon are examples of well-off, highly educated middle-class major cities. Marseille, Montpellier, La Rochelle, Poitiers, Brest, Orleans, Strasbourg, Reims, Metz, Clermont-Ferrand, Limoges, Tours or Amiens fall in an intermediate category – not as affluent, still fairly educated and professional, but with some starker social contrasts. Le Havre, Mulhouse, Roubaix, Cherbourg, Saint-Nazaire and a lot of the cités populaires surrounding Paris are urban areas but have retained a far more traditional, lower income, working-class or industrial makeup despite social changes.
The core, “socially integrated” urban areas of France generally soundly rejected Sarkozy. It is reflective, generally, of two similar and concurrent factors. Firstly, a strong anti-Sarkozyst mood amongst cadres moyens, salaried middle-classes, public and para-public sector employees of all levels and other fairly socially integrated and well-off middle-classes. As the exit polls showed, the “professions intermédiaires” went against Sarkozy by very significant margins. For these voters, a mix of poor economic conditions and unpopular government policies alongside Sarkozy’s chilling right-wing populist turn were likely the main reasons for their strong backing of Hollande. Secondly, employees and lower-level salaried middle classes also tended to reject Sarkozy – especially so in the inner suburbs and the old socialist suburbia. Economic concerns (job loses, purchasing power etc) likely played major roles in these voters’ rejection of Sarkozy.
The famous Red Belts, working-class municipalities with a strong Communist political tradition which form the proletarian hinterland around cities such as Paris, Lyon, Grenoble or Rouen have been changing as well. The Red Belt in the Parisian petite couronne is no longer a centre of heavy industry and large concentration of ouvriers – take a look at the figures for ouvriers in places such as Montreuil (obviously) but also Nanterre and you’ll be surprised at how low or average the numbers are. Some of these buckles of the red belts are gentrifying (Montreuil), but in general they have retained a low-income character but transformed into blighted “inner cities” (to use the term in an American sense) with large immigrant or ‘ethnic’ populations, employed in low-paying service, public or manufacturing jobs. In these type of suburbs, which voted solidly for Royal in 2007 and even more solidly for Hollande in 2012, the candidate of the PS was carried by immigrants, poorer working poor whites (employees) who have retained their family’s political traditions, lower-level public or para-public employees (hospital staff, teachers etc) and some young educated professionals in the gentrified “integrated” inner suburbs.
In the cités populaires of the Parisian basin (Mantes-la-Jolie, Trappes, Garges-lès-Gonesse, Argenteuil, Les Ulis, Grigny, Evry, Corbeil-Essonnes etc) Hollande performed extremely well. Royal had already done quite well in the cités populaires in 2007, where it had been noted that she had been particularly good at motivating young first-time voters of foreign ancestry. In 2012, it appears as if Sarkozy lost most of the gains he had made with poor whites in these areas in 2007 based on his strong appeal on issues of immigration or lower middle-class populism. They returned to their left-wing roots.
On the other hand, Sarkozy performed well in the outer suburbs and exurbia. Marine Le Pen had done extremely well in most of the so-called périurbain subi in 2012, the beneficiary of populist anger by lower-income voters pushed further and further out from the main urban cores by rising inner city property prices and hurt by mortgages, debts and rising fuel prices, and concerned by immigration and insecurity. Sarkozy’s record is probably not as popular as it is in Neuilly-sur-Seine, but his conservative rhetoric likely proved appealing in these areas, where he had done fairly poorly in the first round. His campaign image as the steady hand against the “dangerous change” embodied by Hollande. Sarkozy’s right-wing populism, based, as in 2007, on la France qui se lève tôt and “le vrai travail” were probably much closer to the bread-and-butter concerns of these voters, who do not usually tend to be public or para-public employees in large numbers. Furthermore, Hollande’s centre-left brand of consensual social democratic policies mixed in with left-wing anti-Sarkozyst fodder might have proved less appealing to these voters. At any rate, exurbia has always been a difficult region for the PS. A lot of inhabitants are not employed in the public sector, and they generally tend to frown upon the job security and “fat cat unions” for civil servants which is allegedly defended by the PS. They might feel at odds with the PS’ perceived stylistic bias towards left-wing educated urban middle-classes, the public sector but also immigrants.
Hollande’s performance in working-class areas were far more tepid. While he did join the old left-wing bases of the north, Picardy, Seine-Maritime, Ardennes and Lorraine (lost, in good part, by Royal in 2007) with his core base in the southwest and centre, his comparative performance was actually fairly unimpressive. He did well in rock-ribbed left-wing working-class locales such as the northern mining basin, the Ardennes, the Longwy/Moselle industrial conglomeration, industrial Rouen and Le Havre and the Montbéliard-Sochaux-Héricourt area. But, as we shall see, his performance in these areas often tended to be weaker than Jospin’s results in 1995 – when, must it be noted again, Jospin lost nationally. Sarkozy still dominated throughout right-wing working class areas including, notably, parts of Moselle and Alsace.
Compared to the traditional map of the right, Sarkozy’s map retains its ‘eastern bias’ (first noted in 2007) and its strong correlation (in part) with the general map of the FN in the first round. This map proves that the UMP’s general base has shifted rightwards from the days of chiraquie in 1995-2002 and that a good part of the right’s support in the runoff comes from first round FN voters. Helped by his semi-nationalist populist cultural and stylistic conservatism, Sarkozy held up fairly decently in a lot of eastern France. The second round, thus, was clearly a serious affair without much ‘revolutionary’ votes by angry right-wing voters or protest votes. Even if these voters might have felt let down or disappointed by Sarkozy, they returned home, at the end of the day, to their traditional ideological home. He still lost a sizable share of the most impressive gains he had made in 2007, but he held up well in parts of the grand est and the Rhône Valley/Riviera in the southeast which might have been assumed to prove fairly resistant to Sarkozy based on his record, despite their political conservatism and their electoral penchant for the right.
Historical Geographic Comparisons
The first comparison we can draw is the most obvious one: the 2007 runoff. Royal won 47% of the vote that year as the PS candidate, Hollande won about 51.6% of the vote this year, inferring a national swing of +4.6% in the PS’ favour. Of course, a map of Hollande’s comparative gains (mostly) over Royal is a mirror image of Sarkozy’s loses (mostly) since 2007, which eliminates the need for a second map. A very dark blue denotes cantons where Hollande did worse than Royal in 2007, varying shades of blue indicates cantons where he gained by less than his national average and varying shades of orange-red indicates cantons where he gained by more than his national average.
Obviously, Hollande’s most impressive gains came from the Limousin and his semi-native Corrèze where he improved significantly upon Royal’s performance. This clearly shows an added regional effect for the native son, given that the chiraquien anomaly in traditionally left-wing Corrèze had been almost entirely eliminated by Royal in 2007. Hollande added a favourite son effect to the left-wing base in the Limousin and its neighboring regions. The Lot, Dordogne, Puy-de-Dôme and Cantal are quite telling in this regard. Hollande’s gain vis-a-vis 2007 were extremely heavy in cantons which are immediately adjacent to or fairly geographically close to Corrèze, while his gains remained large though less so the further out you get from Corrèze. The general halo effect is rather interesting.
Another region which is particularly striking is the northern Cotentin and the Cap de la Hague in the Manche. Hollande had already performed very well there in the first round, leading some to come up with the hypothesis that the popular PS/hollandais mayor of Cherbourg, Bernard Cazeneuve, might have boosted the Hollande candidacy in the region. However, I must admit my reluctance to accredit such major regional swings to the work of popular local officials who are big backers of a candidate. Cazeneuve might have had an effect, but I believe that accrediting the result there to his work and popularity is just another incidence of an unfortunate trend to give simplistic and boneheaded answers to complex questions.
The Cotentin is an interesting case. Cherbourg is a working-class Socialist stronghold, and Hollande generally improved by fairly nice margins over Royal in a lot of left-wing working-class areas like Cherbourg. But it is not really a city with a huge suburban influence and hardly the type of urban area which should see a sudden influx in left-leaning suburban families. The Cap de la Hague is also noted for the nuclear power industry at Flamanville, home to one of France’s most famous nuclear power plants. There is a controversial project in the works to expand the nuclear power plant (the EPR project) at Flamanville. Both Hollande and Sarkozy favoured the project, though Hollande wants to close a nuclear power plant in Alsace and reduce France’s dependence on nuclear power. Sarkozy has kind of dragged his feet on the EPR case, so there might be local frustration at the slow pace of the project? Unemployment is high but not excessively so, therefore it is tough to envision particularly profound anti-Sarkozyst anger in a region which, while not as Catholic and hence conservative as the Avranchin and the south of the department, is fairly right-wing.
Elsewhere, the pattern of gains are a bit more patchy. The mining belt of the Nord and parts of the Pas-de-Calais’ mining belt are perceptible. Sarkozy lost heavily in a traditionally working-class and left-leaning part of the Aisne which is east of Laon and south of Saint-Quentin, which notably includes Tergnier, a strongly left-wing cité cheminote. These strong Sarkozyst loses also extend into neighboring parts of the Oise and Somme. Another industrial basin, Montbéliard-Sochaux-Héricourt in the Doubs and Haute-Saône are noticeable. Sarkozy had done quite well in this working-class and usually left-leaning Socialist region in 2007, but Marine Le Pen had done well throughout this declining industrial basin in the first round. Another struggling working-class area where Sarkozy did poorly in the runoff is the Saint-Dizier area in the north of the Haute-Marne.
A general pattern of stronger Hollande gains can be observed in a number of left-wing working-class or industrial areas in eastern or central France. Notably: from Digoin to Autun (in Saône-et-Loire), the ardennois industrial basin, Rouen and Le Havre’s proletarian hinterland, Creil-Montataire (Oise), Dunkerque, Calais, Boulogne-sur-Mer, Maubeuge, the stéphanois mining-industrial basin (including Firminy; Loire), Commentry (Allier), parts of Nièvre, Vierzon (Cher) and Saint-Junien (Haute-Vienne).
He also gained considerably in semi-proletarian, semi-working poor impoverished suburbia or “populaire cities” including Dreux (Eure-et-Loir), Montargis (Loiret), Roubaix/Tourcoing (Nord), Chenôve (Côte-d’Or), Villeurbanne, Vénissieux, Vaulx-en-Velin, Bron (Rhône), Fontaine, Échirolles (Isère), Mantes-la-Jolie (Yvelines), the entire 93, the bulk of the 94, parts of Val-d’Oise (Sarcelles, Garges, Goussainville, Argenteuil, Cergy), Meulan (Yvelines), Grigny/Evry/Corbeil-Essonnes (Essonne) or Savigny-le-Temple.
In the first case, Sarkozy likely had an unusual populist appeal to the white working-class which was traditionally left-wing (remember: Ipsos in 2007 found that 12% and 10% of PCF and PS sympathizers respectively voted Sarkozy in 2007 over Royal), but lost these gains in 2012 for fairly obvious reasons. In the second case, Sarkozy likely appealed to poor whites and lower middle-class whites in these poor and fragile suburban areas which could otherwise have voted FN. Sarkozy had already been a particularly bad candidate for places like the 93, but this year he apparently lost not only the immigrants and visible minorities but also the petits blancs (a term referring to ‘poor whites’/WWC).
Hollande also did quite well, compared to Royal, in a few isolated/mountainous rural cantons where Mélenchon had, in the first round, taken a fair share of traditional PS votes. Did Mélenchon motivate a new, formerly abstentionist electorate and Hollande managed to keep them motivated? Or perhaps the gains in the Basque Country, the Ariège, Pyrénées-Orientales, parts of Aveyron/Lozère, eastern Ardèche and the Monts-d’Arée (Finistère’s solidly left-wing mini-mountains, including the Communist stronghold of Huelgoat) speak to a local concern about the decline of public services in rural areas, regional economic stagnation/decline and the “hollowing out” of the state in those isolated rural areas. Some of these mountainous regions, not all of which are traditionally left-wing (the Basque Country and eastern Ardèche for example) have also suffered from rising local unemployment due to factory closures (especially in parts of eastern Ardèche, hit hard by deindustrialization as of late). But if that was entirely the case, wouldn’t the Vosges have denoted itself by a big swing to the left given its economic troubles and recent job loses?
Where did Sarkozy resist best? Throughout Alsace, with the exception of Strasbourg and Mulhouse, he held up remarkably well and even gained compared to 2007 in a handful of cantons in the Bas-Rhin (seemingly most of the Protestant cantons, but also – ironically – some pretty populaire/ouvrier caché ones too). Even in Moselle, where the UMP had performed like the plague in the 2010 regional elections, Sarkozy’s vote held up very well – losing heavily only in Metz and the metallurgical Moyeuvre-Grande/Fameck/Florange/Gandrange area. In the conservative working-class areas around Forbach, Freyming, Stiring and Carling there were no strong loses for Sarkozy (besides Forbach and parts of Saint-Avold). A similar story in Meuse and parts of Meurthe-et-Moselle. Despite the strong incidences of a far-right semi-protest vote in the first round in a lot of these areas, a lot of first round FN voters – despite being fairly populaire sociologically – came back to the Sarkozyst fold. Another good example of voters “eliminating” in the runoff, opting for the least worst option (for their views) or finally voting for the candidate of their ideological tradition.
In the wealthy countryside of the Sundgau, Champagne, Beaunois (côte viticole of the Côte-d’Or), Chalonnais, Maconnais, Bresse (Ain), Dombes (Ain), Beauce (Loiret) and Sologne (Loir-et-Cher); the Sarkozyst vote held up well. The same phenomenon is observed in more urban affluent areas (Neuilly, Limonest, Meylan, Annecy) and the well-off resort towns (Trouville, La Baule, Les Sables, Royan, Arcachon, the Var/Alpes-Maritimes, ski country). A far more understandable phenomenon. Of all types of voters, those who are most well off are probably those who are the least alienated from the Sarkozyst style of politics.
A final region where Sarkozy’s vote showed the strongest resistance are some of the Catholic Christian democratic lands (inner west, continental Brittany and legitimist Morbihan, Savoies, Flanders, rural Pas-de-Calais, Lyonnais). How can we interpret this performance? The most likely explanation is that Sarkozy had already performed fairly poorly for a right-wing candidate in a lot of those areas (except Savoie and the Lyonnais) in 2007, and that the runoff left had already started taking in almost all the centre-left humanist votes it could, leaving a Bayrouist rump which is far more centrist/centre-right in its political orientation. It could also reflect a stronger overall performance by Sarkozy with the Bayrou/centrist vote, the remaining rump being fairly conservative in temperament and thus perhaps more resistant to Hollande’s ambitious change/anti-Sarkozyst agenda. Add to this list the Deux-Sèvres, where Hollande did not lose votes compared to Royal’s native daughter performance in 2007 but still did not gain as much considering Royal had started hitting the ceiling with her native daughter boost.
The second interesting exercise is to compare Hollande to Lionel Jospin in 1995. Overall, Hollande did 4.2% better than Jospin had in 1995 (he lost to Chirac, taking 47.4% of the vote). We could expect that, like with the 2007 comparison, on a 4% swing, Hollande would have gained on Jospin by varying amounts in this seventeen year period. Is this the case? Far from it.
The map is aesthetically pleasing, with some beautiful patterns and uniform blocs appearing. A clear east-west divide appears: east of an axis defined by the cities of Le Havre, Meaux, Lyon and Perpignan, Jospin – who overall lost the 1995 election – generally did better than Hollande in 2012 – who won the election. Hollande’s ‘loses’ were most pronounced in traditional proletarian areas: the Nord-Pas-de-Calais’ emblematic mining basin, Picardy, most of the ardennois industrial basin, good parts of Lorraine’s two main industrial concentrations, the Haut-Rhin (Cernay, Saint-Amarin, Mulhouse), the Saint-Dié area, Montbéliard-Sochaux, parts of Nord-Isère, Savoie’s Maurienne valley and most of the Marseille industrial waterfront. These results show that while Hollande might have done significantly better than Royal in some working-class areas, he did not come close to match the numbers posted by even a defeated Socialist candidate in 1995!
In these and other regards, the places where Jospin performed better than Hollande are quite eerily similar to those areas where the FN does best. In addition to the bulk of these aforementioned regions, you can add the rest of coastal PACA, the Mediterranean riviera in the Languedoc, the Rhône valley, the Garonne valley, parts of the greater Parisian basin (Eure-et-Loir, Eure, Oise, Aube) and the centre (Loir-et-Cher). In these cases, there are demographic factors at work: population growth along the Mediterranean thanks to seniors and retirees settling in or exurban growth in the outer reaches of the Parisian megapolis (but also in the Garonne valley and outer Toulouse); but also more political factors: rising concern over immigration and security in most of these regions, and of course a strong FN vote.
Let us look at the flip side of the coin: where did Hollande improve the most on Jospin’s performance? Three main areas:
Firstly, the “greater hollandie” – my new name for Hollande’s Corrèze and the favourite son halo effect it has created; mixed in with gains also coming from Christian democratic/Catholic country (Lozèere, Aveyron, Cantal, Haute-Loire). In Tulle, he did a full 36.2% better than Jospin – admittedly, it was the right which had a favourite son halo out of Corrèze in 1995!
Secondly, in western France, strong gains throughout the old Christian democratic/Catholic country. Notice, for example, how Hollande gained a lot in the bocage vendéen, eastern Morbihan, continental Brittany, the Choletais in Anjou, the north of Poitou and the bocage normand but did not gain as much (or even lost) in the plaine et marais, the Baugeois or the Sarthe. We will come back to the reasons for this shift when we look at our final comparison map. In the Pyrénées-Atlantiques, Hollande benefited from a strong “Christian democratic shift” in his favour coupled with a local Bayrouist effect, whereby Bayrou’s native son voters have become far more inclined to vote for the left and follow their leader’s political shifts.
Finally, Hollande gained quasi-universally in urban areas. A few paragraphs above, I described the new makeup of France’s major cities. As a result of shifting demographics which were previously outlined, the main cities have progressively shifted left. The urban effect is huge in Paris (+15.8% – Chirac had a second native son effect in his “other” political base), Lyon (+6.2% – the map is wrong, given that the official results for Lyon in 1995 have been inversed by the useless Interior Ministry), Lille (+9.2%), Grenoble (+11.9%), Strasbourg (+9.2%), Caen (+10.2%), Rouen (+11.4%), Rennes (+10.2%), Nantes (+11.3%), Brest (+10.9%), Bordeaux (+13.2%), Poitiers (+12.4%), Toulouse (+10.7%), Montpellier (+13.8%), Nancy (+12%), Orleans (+10.2%) and Saint-Etienne (+12%). But it is not only confined to those regions. Almost all the isolated “red dots” (showing gains for Hollande) on the above map are urban areas!
The final relevant comparison we can draw up is Hollande against Mitterrand in 1981. Nationally, Mitterrand won the 1981 election with barely more support than Hollande (51.76% vs. 51.64%). A superficial image of overall stability? Indeed! At a local level, there are some huge shifts. Le Figaro drew up a map of a 2012-1981 comparison at a very detailed (communal?) level, which is far more useful than a departmental comparison. Its colour scale could be improved, but it is a very useful map.
The map is amusingly close to the 1995-2012 map, which proves that the most significant shifts in the electoral map happened in the last seventeen years of this thirty-one year gap with Mitterrand’s election to the presidency in 1981. Once again, we clearly see the gains in Auvergne/Massif Central/”greater hollandie“; the gains in the west and the gains in the Pyrénées-Atlantiques.
In the first case, the favourite son explanation can only be correct to a certain extent. 1981 is a better comparison point than 1995 because there was no definite favourite son effect coming out from Corrèze in the runoff. Chirac’s native son voters had not voted for Giscard is disproportionate amounts, and Corrèze voted for the left as it traditionally had in the past. Indeed, on the map on the left, you will notice that the deepest shades of red are not found in the Limousin. Rather, they are found in the southern Massif Central and Puy-de-Dôme. Giscard got a small favourite son boost out of his native Puy-de-Dôme, which shows up in this map. The strong gains registered by the left in the Aveyron, Cantal (especially the monts du Cantal and Saint-Flour plateau), the Haute-Loire, Lozère (notice the lack of gains in the Cévennes but huge gains in the Margeride – a total +9.4% for the PS in the department) and in the haut and moyen Vivarais (Ardèche) all are the result of the “Christian democratic shift”.
This same shift is most visibly shown in the west. The two departments with the sharpest trend to the left were the Ille-et-Vilaine (+9.9%) and Finistère (+9.8%). The Deux-Sèvres (+9.7%), Manche (+9%), Mayenne (+7%) and Maine-et-Loire (+6.4%) also feature prominently on the list of strongest left-wing gains. At a more micro level, it is important to note where the left’s strongest gains came from – the Catholic regions – the Léon, eastern Morbihan, eastern Ille-et-Vilaine, the bocage angevin, the bocage poitevin and the bocage normand. Far less impressive gains from the traditionally republican plaine poitevine, the Baugeois, Sarthe, Perche (Oise), Auge (Calvados) and more modest gains in central Brittany. The Catholic effect is, of course, picked up in the Pyrénées-Atlantiques (especially the Basque Country) but also – amusingly – even in the Pays de Caux, Flanders, Bas-Rhin and parts of Moselle.
The obvious cause: secularization. The religious cleavage remains important, but the structure of the religious cleavage has changed dramatically as voters move away from Catholicism and towards irreligious secularism. This trend has generally gone hand-in-hand with urban growth in a lot of the aforementioned regions. Brittany in particular has moved out of its stereotypical isolation and poverty to become an economically successful, urbanized and educated region. Urban centres such as Rennes or Nantes, cities which the PS started winning in 1977, proved to be the epicentres from which the New Left, driven by moderate social democracy rather than class-struggle Marxism, and influenced by the Christian left tradition, would expand its wings as cities grew, suburbs expanded and isolated clerical rural areas turned into demographically evolving regions. Since Mitterrand’s election in 1981, the left has become far less “scary” for voters of Catholic tradition. The old boogeyman of the Socialist left being a child-eating red atheist monster has died off.
Voters of “Catholic tradition” – which we can define as less clerical, less practicing in these days but still influenced by a Catholic upbringing, environment and political tradition – have shifted pretty dramatically to the left in recent years (though it is a long-term process, begun in the 1980s). In the 1960s and 1970s, the bulk of the “Catholic” vote (practicing + tradition) was solidly right-wing, in part out of the fear of the atheist “Reds”. When the experience of the left in power in 1980s broke those old reflexes and fears of baby-eating communists, secularized voters of Catholic upbringing gradually shifted to the left. After all, despite all that has been said about the Catholic Church being reactionary and so forth, the Catholic tradition often went hand-in-hand with pro-European views (in part, likely, because of the idea of ‘Europe as a Christian project’, which is not uniquely French) and more centrist views on economic matters and social policy; closer to the Christian democratic MRP tradition of the “third-way” between liberalism and socialism than to the right’s traditional liberalism or the Sarkozyst frontiste-appealing right-populism and weird nationalism.
On the other hand, with the notable exception of the Garonne valley, most of eastern France lying east of Le Havre-Meaux-St Etienne-Perpignan shifted to the right over the course of the last 31 years. A lot of explanations in this case, which can also be useful in explaining the patterns in the 1995-2012 map as well.
The biggest shifts to the right all took place along the Mediterranean coast: Gard (-8.7%), Bouches-du-Rhône (-8.9%), Alpes-Maritimes (-9.9%), Vaucluse (-10.7%) and the Var (-11%). A process driven by major demographic changes. Since 1981 (the process had already begun at that point; Giscard did better in 1981 than in 1974 in places such as the Var), the coast has seen a decline of traditional rural socialist traditions (the Var rouge is almost dead, Vaucluse’s old republican traditions are barely perceptible, the rebellious Radical-Socialism of the Languedoc is dying off) and traditional industries (small mining, small-scale wine producers, cooperative farms in rural areas, end of shipbuilding in La Seyne, industrial decline around Marseille). In return, it has seen its population boosted by a huge influx of retirees and seniors, the bulk of whom are either wealthy or very conservative or oftentimes both. An old left-leaning white working-class vote has been replaced by a vote of petits blancs (poorer whites: petite bourgeoisie, lower middle-classes, retired working-class, employees) who are concerned about immigration and security or a more bourgeois conservative vote (from big landowners hiring migrant workers: Vaucluse, or anti-immigration/anti-criminality white retirees and old pieds-noirs along the Riviera). Similar factors can explain the shift in the Nord-Isère, counterbalanced by an opposite shift in the south around Grenoble and the neo-rural mountainous areas.
In other working-class concentrations, such as the Oise, the NPDC mining basin, Ardennes, Aisne, Doubs/Haute-Saône/Belfort, Haut-Rhin and Saint-Dizier; the left has also taken a bit of tumble compared to 1981. 1981, if not 1978, marked the beginning of the decline of the “working-class culture” and the rock-solid 70% vote for the left from the working-class. Mitterrand’s election was followed by a tough period of disillusion with the left from the working-class, hit the hardest by the industrial decline and spike in unemployment which accompanied Mitterrand’s first term. As the PS slowly shifted from the neo-Marxist rhetoric it had long used, even in its SFIO days, towards a social democratic rhetoric closer to that of the SPD in Germany or Labour in the United Kingdom, the working-class vote started shifting away and exploding. Voters felt that the left had abandoned its roots, and they fell out with Mitterrand in large part due to economic decline and a rise in immigration and crime.
The most noted aspect of the end of the “working-class culture” in France was the FN starting to gobble up the support of a good third of the working-class. However, a lot of observers have laid down the FN’s gains with the working-class in the context of a broader right-wing shift in this electorate. The theory goes that the stronger FN vote is only a spin-off result of a broad, general shift to the right observed in the working-class electorate. Rising unemployment, unequal economic decline, differing responses to immigration and crime has led to an “explosion” of the working-class vote, with all ties and bonds holding this electorate together being blown apart by these forces. Carmaux (Tarn) can no longer be counted on to vote quasi-identically to Vénissieux or Saint-Amand-les-Eaux (Nord). This map could give some credence to this theory.
A final pattern of left-wing decline perceptible in this map is the result of outer suburban and exurban growth. This is most striking in the Oise, the southern Aise, parts of Seine-et-Marne and Aube and the confines of the Eure/Vexin Normand; but it also explains part of what is going on in the Garonne valley, northern Haute-Garonne and Nord-Isère. Exurban growth is hardly favourable to the left, as it may tend to tumble old rural socialist strength or old proletarian concentrations (parts of Oise, most notably), but also because the population growth is not of a kind usually receptive to the left.
Below is an attempt at a geographic analysis of vote transfers between the first and second rounds. This is not a very scientific or perfect analysis, but it is a fairly accurate way of easily looking at the regional variations in vote transfers. Basically, the point is to try to compare the “theoretical” base of a candidate and his “actual” result.
For Hollande, the map compares Hollande’s actual result to the total of all first round far-left and left-wing candidates, to which a third (33%) of Bayrou’s vote and a sixth (17%) of Le Pen’s vote is added. This, of course, does not take turnout variations into account and rather amateurishly assumes that all those who voted for a non-PS left-wing/far-left candidate in the first round voted for Hollande: reasonable, but perhaps not 100% accurate.
Overall, the “theoretical” base of Hollande would have been 49.8%, a result which is only 1.8% below what he actually won. The main lessons we can take out of this map:
Mélenchon and “other left” first round voters were, overall, extremely reliable. Indeed, the strongest gains compared to the theoretical national base of 49.8% replicate a part of Mélenchon’s first round map: the old Communist strongholds of the Berry, Bourbonnais, parts of Limousin and Charentes; the unique mélenchoniste mountainous socialism of the Pyrénées, Cévennes and pre-Alps; the old proletarian communism of the north and Picardy; western Brittany and of course the Red Belt in the 93-94. If Hollande underperformed in the first round, it is almost certainly not because Mélenchon’s voters proved to be particularly fickle and unreliable voters. This is hardly surprising.
Bayrou’s voters are another story. This calculation counted only a third of Bayrou’s vote in each individual canton, which is the share of Bayrou’s vote which is assumed to have gone to Hollande in the runoff (averaging Ipsos and Ifop results). In a handful of regions, the theoretical vote proved to be higher than the actual Hollande vote, indicating a clear counter-performance/underperformance on Hollande’s behalf. In the haut-bocage vendéen, the Choletais, the vitréen of Ille-et-Vilaine and parts of Mayenne (cantons coloured in light green) we can safely say that Bayrou’s voters tended to transfer their votes disproportionately to Sarkozy. A similar phenomenon is likely at play in the Aubrac in the Aveyron or the Jura plateau in the Doubs, though turnout shifts could explain things as well. We touched on this point in our comparisons above, with Sarkozy’s vote in the west having shown the strongest resistance in regions where Hollande underperformed or had mediocre transfers from Bayrou.
I repeat the theory I had laid out. In 2007, Bayrou’s vote, even in these conservative regions, was likely boosted by a anti-Royal centre-leftist element which transferred well to Royal in the runoff. This year, that vote shifted back to Hollande by the first round, giving Bayrou an electorate which was likely far closer to the centre-right if not the right overall. For these small-c moderate conservative voters, Hollande’s fairly ambitious agenda might have provoked some fear, and led to a “legitimist”-type vote for the incumbent President.
In other regions, there is also the appearance of poor transfers from Bayrou voters. This is most visible in the Yvelines, Hauts-de-Seine, outer Val-de-Marne and parts of Val-d’Oise; all cantons where the theoretical vote proved to be higher than the actual Hollande vote. This is the likely result of fairly affluent, well-off bourgeois centre-right voters who did not like Sarkozy’s right-wing populism in the first round but who, fairly naturally, transferred their votes to Sarkozy, who is after all the candidate closest to their beliefs and political tradition. The Yvelines, Hauts-de-Seine and Oise confirm this assumption: where did Hollande’s vote fall quite a bit short of the theoretical vote? Neuilly, Saint-Cloud, Boulogne-Billancourt, Le Chesnay, Versailles, Saint-Nom-la-Bretèche, Saint-Germain-en-Laye, Senlis and Chantilly… This is confirmed outside of Paris too: Marcq (Lille), Lyon’s northern suburbs, Meylan (Grenoble) and Haute-Savoie.
Bayrou’s voters proved to be more left-inclined in most of Brittany (including conservative Léon), Normandy, parts of the Massif Central and especially in the Pyrénées-Atlantiques. Bayrou’s native son comes, in part, from traditionally republican or left-leaning regions, thus a transfer to the left is natural. In parts of the Basque Country, however, his voters also transferred solidly to Hollande. Perhaps a result of a mini-regionalist phenomenon, given how Sarkozy seems to be unpopular with the small but potentially electorally significant regionalist community?
Le Pen’s voters did not transfer to Hollande in large numbers, but we assumed a scenario where a sixth of her voters did (which is roughly what happened, according to Ipsos and Ifop). Unsurprisingly, FN voters in PACA, Rhône-Alpes and Alsace-Moselle proved to have the weakest transfers to Hollande. We are, after all, dealing with what is predominantly an electorate with strong right-wing political traditions and identification and, in part, an electorate which is more boutiquier and petit bourgeois than a purely working-class/working poor electorate. In the Var and Alpes-Maritimes, the FN draws votes from fairly middle-class whites who are concerned about immigration and crime, and whose vote is not necessarily a “screw them all” type of protest vote. In the Vaucluse, the FN draws a politically conservative/reactionary vote which is not a “screw them all” vote either.
In the Lyonnais and Ain, mixed in with poor transfers from Bayrou’s voters as well, the périurbain and petit bourgeois electorate the FN grabs proved reticent to Hollande. The périurbain vote in the greater Paris showed marginally better but still quite mediocre transfers overall. Finally, in Alsace, the obliteration of sorts suffered by Hollande is the double whammy coming from poor transfers from a traditionally Catholic, semi-regionalist centre-right Bayrou electorate and a conservative, localist and lower middle-class/rural populaire FN base drawn in good part from the traditional right.
In Moselle, we do note some not-too-shabby transfers (likely from Marine) in the Saint-Avold/Forbach area and strong transfers (likely more from the left) in the metallurgical conglomeration north of Metz. However, the main breeding ground for gaucho-lepénisme this year as in past years was Picardy and the NPDC. While good transfers from Mélenchon’s voters is a big part, the very strong (+6-9%) transfers registered in the mining basin and parts of Aisne and the Somme must have in part from gaucho-lepéniste voters. A working-class electorate, with left-wing traditions, which voted for Marine Le Pen in the first round (often for reasons other than immigration/crime, often as a protest vote) but which opted to return to its left-wing roots in the runoff, further boosted, perhaps, by deep disgust with Sarkozy.
As our look at the Sarkozy transfers will show, Poitou-Charentes and other regions of the inner west and southeast might have had a small gaucho-lepénisme phenomenon at work too. In these parts, the FN vote was probably boosted by an added CPNT/”hunters” element which had not been present in 2002 for the FN. The CPNT vote in these regions, usually high, was often far more left-leaning than the conservative hunters vote in the Somme estuary. It represented a rural protest vote, but one still anchored in local left-wing traditions.
For Sarkozy, the map compares Sarkozy’s actual result to the total first round vote for Sarkozy and Dupont-Aignan, to which 55% of Le Pen’s vote and two-fifths (40%) of Bayrou’s vote is added. These numbers are in line with the observed transfers from both of these candidates. Again, this, of course, does not take turnout variations into account and rather amateurishly assumes that all those who voted for NDA in the first round voted for Sarkozy: reasonable, but perhaps not 100% accurate.
The theoretical vote on this basis is fairly low: only 42.47%, which means Sarkozy outperformed his theoretical base by a full 5.89%. The map is shaded differently to reflect this result: in blue, all gains (only 61 cantons had a higher theoretical vote than actual vote) below the 5.9% average, in red-orange, all gains above the 5.9% average. This means that, obviously, the other left-wing candidates did not see their voters split 100% in Hollande’s favour – Sarkozy like pulled 5-7% of Mélenchon’s voters and 10% of Joly’s voters. It also means that general turnout shifts generally played out in a way favourable to Sarkozy.
Le Pen’s general map is visible, but not entirely: only in coastal PACA, Gard, Hérauult, the Garonne valley, the Rhône valley, Rhône-Alpes, the grand est, Alsace and Oise. In general, throughout the Alpes-Maritimes, Var, western Vaucluse, Bouches-du-Rhône and the eastern Gard, Nord-Isère, the Lyonnais, Ain, Haute-Savoie, parts of Savoie, Alsace-Moselle, Aube, Marne and the périurbain lointain around Paris had some strong transfers to Sarkozy, from which we can infer – generally – that Le Pen’s voters in those regions generally showed good transfers to Sarkozy. In addition, we can likely assume, from the above map, that FN voters in coastal resort areas (Vendée, Royan, Arcachon, Pyrénées-Orientales, Hérault, PACA) transferred well to Sarkozy.
It is interesting to note that the regions where the FN’s first round base proved the most “Sarkozyst” generally replicate the map of the Le Pen vote… in 1988. Interesting, but not especially weird. The FN’s vote in 1988 was beginning to “popularize” in sociological terms, but it remained much more sociologically right-wing and of higher socioprofessional status than it does today. This year, the traditional boutiquier, lower middle-class, petite bourgeoisie and périurbain fractions of the wider FN vote posted strong transfers to Sarkozy. These electorates are, in general, more sociologically right-leaning and their vote for the FN is not entirely a “screw them all” type of vote, as explained above. Certainly the FN vote in PACA and parts of Languedoc and Rhône-Alpes, partly influenced by the pied-noir factor, is quite unlike the newer FN vote in places like Hénin-Beaumont where the FN vote is reflective of a wider social malaise and exasperation.
The old “poujado-frontiste” type of vote (boutiquier, lower middle-classes, petits blancs) has always been if not anti-left-wing at least very resistant towards the left. The wider périurbain vote remains one of the left’s main weaknesses. Finally, despite Sarkozy’s weakness with the working-class as a whole, working-class milieus of conservative tradition (often because they are Catholic, badly unionized or historically hired local workers) showed no particular allergy towards a Sarkozy vote in the runoff: note especially the Yssingelais, Moselle, Cluses-Scionzier or Cernay-Saint Amarin. For many, their Sarkozy vote might not have been a pro-Sarkozy vote rather than an anti-Hollande vote, but the result is the same as far as raw votes are concerned!
On the other hand, Picardy and most of the NPDC (but also parts of Ardennes, the leftist working-class parts of Lorraine, parts of Saône-et-Loire and some leftist banlieues populaires) showed no particular love for Sarkozy. Indeed, the theoretical vote was only marginally lower than the actual Sarkozyst vote, which reflects that Le Pen’s voters generally posted less impressive transfers to Sarkozy. Once again, this is hardly surprising. The bulk of these areas are working-class areas with PS or PCF traditions, and the birthplace of gaucho-lepénisme in 1995. While gaucho-lepénisme properly defined (left-wing voters voting FN in the first round but then for the mainstream left in the runoff) is no longer as prevalent as in the 1990s – a lot of gaucho-lepéniste type voters prefer to sit out the runoff nowadays – it does remain an important phenomenon.
To conclude, Sarkozy’s gains from the FN can be assumed to have been greater in areas which are overall rather right-leaning, and lesser in areas which are overall left-leaning.
Bayrou’s voters in eastern France showed stronger transfers to Sarkozy than their counterparts in Brittany or greater hollandie. In Alsace, Moselle, Marne, the Lyonnais and Haute-Savoie parts of Sarkozy’s large gains over the theoretical right-wing base must have come from Bayrou voters. In Alsace, there appears to be a pretty strong correlation between Sarkozy’s gains in the runoff and Bayrou’s strongest results in the first round. In these conservative regions, a mix of wealth and traditional conservatism likely shifted these first round MoDem voters towards the right in greater numbers than in Brittany, where the centre and the PS have often coincided in a common Christian left upbringing.
Bayrou’s voters in the clerical and extremely right-wing regions of the Aubrac and Margeride likely shifted in larger amounts to Sarkozy as well, but in these regions, a spike in turnout likely proved beneficial to Sarkozy as well (as it did in parts of the Parisian basin). Similarly, lower turnout might have played to Sarkozy’s benefit in parts of the inner west (especially the rural Mayenne, the Choletais, the eastern confines of Brittany, the haut-bocage and marais breton), where the (likely) abstention of fledgling Bayrou voters might have resulted in a more right-wing runoff electorate. However, it is hardly unreasonable to assume that Bayrou’s voters in these regions – all regions with a monarchist and clerical Catholic tradition – might have transferred in a way favourable to Sarkozy as well.
In contrast, Sarkozy struggled in Lower Brittany, a region where Bayrou’s centrist votes were likely socialized in an environment heavily influenced by a Christian left/PSU tradition. In the Limousin, Charentes, Berry, Nivernais, Bourbonnais and the Pyrénées; Sarkozy struggled a lot. Bayrou’s few voters in the greater hollandie were likely attracted to a native son like Hollande. However, his vote share was already quite low in the general region. It is possible that with Le Pen’s voters in these rural areas showing particular resistance to Sarkozy, the right-wing vote in the first round was already approaching Sarkozy’s realistic ceiling in a regional context marked by an extremely strong native son effect which certainly broke old left-right barriers, just like Chirac’s native son effect in the same region was not just a right-wing vote.
In Bayrou’s native Pyrénées-Atlantiques, his own favourite son-influenced voters had bad transfers towards Sarkozy, just like in 2007. Whatever the particular causes, the result of this is quite shocking. The Pyrénées-Atlantiques, long the most right-wing department in Aquitaine and a historically conservative department, gave 57.1% of the vote to Hollande – meaning that Hollande did better in the Pyrénées-Atlantiques than in Gironde, Landes, Gers or Aude!
The immediate next step, in electoral terms, remains the legislative elections – held in June. Since the shift in the electoral calendar in 2000, the legislative elections have been synchronized to take place after a presidential election. This reform has made the presidential election the election, while reducing the stakes of the legislative elections by making them fairly dependent on the results of the presidential election. The obvious hope was to solidify the presidentialist nature of the regime and prevent any further cohabitations. Thus far, voters have proved their legislators correct in their original assumptions. In 2002, the newly created single party of the right – the UMP – won a very large absolute majority on its own, while the PS saved face and began the painful rebuilding after the April 21 rout. In 2007, President-elect Sarkozy’s UMP won the first round easily but voters in the runoff turned the cards around and the vague bleue was more of a blue ripple. The UMP retained its majority, the first legislative majority to win reelection since 1978, but lost a significant amount of seats to the left.
Based on this recent history, the media narrative of these elections is that the left – either the PS alone or the PS alongside its allies – will win the legislative elections. Indeed, it would be quite irrational and downright bizarre on the part of the French electorate to elect a PS president but turn around a return a UMP legislature a month later. The cohabitation idea is not very popular, and the general mood will be to give to President Hollande the means of carrying out the policies he was elected on.
However, early polling has shown a surprisingly small gap between left and right overall. Similarly, the first post-May 6 trends have not indicated a huge honeymoon boost for Hollande or demotivation on the right. I maintain that Hollande will get a short-lived honeymoon, which will likely become apparent after May 16 (when the government is named) and when a tiny sense of optimism and hope sets in until mid-June at the earliest. Voters have likely remained in a presidential mindset, and they will start turning off from politics and elections in a few days time.
Indeed, abstention will be much higher in the legislative elections than the 19-20% abstention registered in the presidential election. In 2007, turnout reached an absolute low at only 60%. While extremely low turnout (below 50%) is unlikely, turnout is unlikely to exceed the high 60s let along the 70% mark. It could yet be a tad higher than in 2007, which seems to be the overall prediction, though I would hardly be surprised if it ended up at 2007 levels or even below the symbolic 60% mark. For the right, the risk is that, following Sarkozy’s defeat, its base is demotivated by an election with low stakes and which seems lost before the battle began. A similar situation had emerged in the 1981 legislative elections, where a close presidential result was followed by a huge PS wave in the snap elections. Hollande will not get a Mitterrand ’81 majority in June, because the euphoria, optimism and hope of changer la vie has not accompanied his election.
The level of abstention is important because the runoff ballot is open to all candidates who have won over 12.5% of registered voters in the first round (unless the seat is filled by the first round). The shock-and-awe idiocy of journalists proclaiming over 350 triangulaires (three-way runoffs) with the FN were based on the low-abstention presidential results. In reality, there will probably not be any more than 50 triangulaires in the best case scenario for the FN. At best, the FN can only hope for 3-4 seats in the best of cases. However, the FN is in a situation to play dangerous tricks on the UMP. In a few constituencies, the FN can hope to place second ahead of either the UMP or the PS and knock them out of contention for the runoff. In other constituencies, there remains a risk for a triangulaires de la mort scenario for the UMP, a la 1997. The FN could very well spoil the UMP and the right’s chances in a handful of constituencies, either through three-way runoffs or poor transfers in the runoff.
The issue of UMP-FN deals has thus emerged, promising headaches for the UMP. Unlike in the past, it seems as if both the UMP and FN’s base favour such deals. The UMP leadership will not sign any formal national deals with the FN, though it seems willing to resist the pressure for traditional fronts républicains with the left to oppose the FN. However, in a handful of local cases, the UMP and FN could end up in agreement on a deal of mutual support in the runoff, agreements similar to traditional PS-PCF deals whereby the party which places second would endorse the one which places ahead of it.
We will come back to the legislative elections and the fun it promises, but for the sake of time, we can work on the assumption that – as things currently stand – the left will win, though the PS will lack an absolute majority on its own and will need the votes of its mini-allies (EELV, PRG) and possibly its picky junior allies (FG/PCF).
In these elections, the left will be “led” by Hollande’s new Prime Minister, the PS mayor of Nantes and longtime parliamentary leader of the PS group, Jean-Marc Ayrault. Ayrault is a competent, tested, experienced and well-known parliamentarian and politician who has the added advantage of having fairly good ties with Germany. He is probably a safe choice for Hollande, which will neither be a yes-man or an overshadowing larger figure, or a potential thorn in the side. The composition of Ayrault’s first government will be interesting to follow. Lots of PS politicos and barons have lined up like good little kids, eager for candy. Like Sarkozy in his cabinets, Hollande will also need to take heed of his party’s various factions and internal families.
President Hollande has a tough presidency ahead of him. The European economic situation, combined with France’s huge public debt, will likely prevent the implementation of the left’s electoral promises (as always). The general mood in France is one of abject pessimism. Voters are resigned towards the high likelihood of some sort of austerity at home, and a lot believe that France could potentially end up like Spain or – gasp – Greece. Thus, the mood which accompanies the election of the second PS president of France is quite unlike the one which accompanied the election of the first PS president, in 1981. The changer la vie hope, euphoria and optimism of 1981 is for the history books. The reaction to Hollande’s election seems to be a widespread “meh” or at best a dismissive “he can’t be as bad as Sarko” impression. The fact that the election was more Sarkozy’s defeat than Hollande’s victory and that Hollande owes his victory to anti-Sarkozysm will certainly come back to haunt the PS and Hollande in the near future, once voters forget Sarkozy and shift their judgement to the new incumbent.
Hollande’s victory was generally read outside of France as an “anti-austerity” vote, alongside the anti-austerity wave in the May 6 election in Greece. In reality, however, austerity was probably a secondary factor behind a local, “franco-french” factor: anti-Sarkozysm. At any rate, it is unlikely that Hollande will single-handedly bend Angela Merkel’s apparent leadership in the EU and the European financial crisis. He will, above all, face important domestic economic and fiscal problems. Before long, he will likely be compelled to adopt austerity policies in some form or another. He will likely attempt to tie any austerity measures with some “tax the rich” measures (likely more symbolic than anything else), and his fiscal policy might be more inclined toward job creation and economic growth than debt/deficit reduction. However, only time will tell what will come of all this.
In the meantime, the UMP will be looking to heal its wounds. Sarkozy’s strong performance on May 6 (comparatively) makes the rebuilding process a bit less tenuous, but the likely absence of Sarkozy from active party politics in the near future opens a new situation. It lays the groundwork for the explosion of longtime private tensions within the UMP and the right as a whole, which will likely soon erupt in open civil war. The UMP’s current leader, Jean-François Copé, is intent on running for the presidency in 2017 and, in the meantime, shaping the UMP into his personal political machine for 2017. He has played his cards well, and he is a rather Machiavellian political figure, who has proven to be politically apt and skillful. However, there is a strong anti-Copé current in the UMP, which could coalesce behind outgoing Prime Minister François Fillon. The expectations seem to be for a Copé-Fillon battle royal, though internal politics in the UMP are not that simple. Fillon is probably the most prominent of the anti-Copé current, but it remains to be whether he can calm the ambitions of the likes of the “young wolves” Nathalie Kosciusko-Morizet or Laurent Wauquiez; but also Alain Juppé.
Yet, the implosion of the UMP remains unlikely, for reasons tied to laws on party financing or the division and dispersion of the centre-right/centre. The UMP will be the place where the reconstruction of the right takes place (with the FN playing a key role as an ‘observer’ of the situation), while the centre will find itself condemned to continued marginalization and “little brother” status behind the UMP.
Only the future will tell what comes of this. Thanks to all readers for reading my posts on the French elections, and tune back for continued number-crunching and analysis of the upcoming legislative elections in the coming weeks and month.